Reinforcement of elastomers: Ef­fect of polymer/filler interactions on the dynamics of elastomeric chains at the interface* **

S le w to w te n y nr 4 lipiec - sierpień 1999 r. TOM 3

A . V id a l* a n d B . H a id a r *

Reinforcement of elastomers: Ef­

fect of polymer/filler interactions on the dynamics of elastomeric chains at the interface* **

Molecular dynamics o f polymer chains in elastomer/filler blends can provide indepth information on the processes taking place at the polymer/

loading material interface and in the elastomeric phase surrounding the reinforcing particles (the interphase) as well A fraction o f the elastomer is in a glass-like state, even at temperatures in large excess o f the polymer T . Taking advantage o f physical ageing processes and using solid state NMR, one may quantify the amount o f polymer which participates to the interphase and show that interfacial interactions are selective and dependent on the filler/polymer couple under investigation.

Key words: reinforcement, interaction, interphase area, testing methods

Wzmocnienie elastomerów: W pływ oddzia­

ływania polimer/napełniacz na dynamikę łańcuchów elastomerowych na granicy faz

Dynamika molekularna łańcuchów polimerowych w mieszaninach elastomer/napełniacz może dostarczyć wnikliwych informacji o procesach przebiegających na granicy fa z polimer/napełniacz oraz w fazie elastomeru otaczającego cząstki napełniacza. Frakcja elastomeru jest w stanie podobnym do szklistego, nawet w temperaturze o wiele wyższej niż Tg polimeru.

Nowoczesne metody badań, wykorzystujące procesy fizycznego starzenia i NMR ciała stałego, stwarzają możliwość ilościowej oceny polimeru uczestniczącego w tw orzeniu obszaru m iędzyfazow ego i p o zw a la ją na obserw acje współoddziaływań w tym obszarze. Oddziaływania te są selektywne i zależą od polimerów i napełniaczy użytych do badań.

Słowa kluczow e: w zm acnianie, w spółoddziaływ anie, obszar międzyfazowy, metody badań

♦Institut dc Grimie des Surfaces ct Interfaces - CNRS

15, rue Jean Starcky - BP 2488 68057 MULHOUSE Cedcx - France, E-mail : A.Vidal@univ-mulhouse.fr

**Paper presented at the Int. Conference ELASTOMERS’98, Warsaw, 13-15 October, 1998

TOM 3 lipiec-sierpień 1999 r. S ta A tw t& U f, nr 4

Introduction

It is well agreed by now that the driving forces which control the so-called reinforcem ent effect, i.e.

the improvement o f properties associated with the blen­

ding o f fillers with elastom ers, are the result o f the interactions which can take place at the polym er/lo- ading m aterial interface. However, in spite o f the lar­

ge am ount o f w ork w hich has been devoted to the un­

derstanding o f the processes w hich m ay occur w ithin a volume in the vicinity o f the surface o f the dispersed solid (the interphase) im portant questions still rem ain unanswered; particularly on the point o f view o f the exact nature o f the interactions involved, and the di­

stance from the filler surface at w hich they influence the matrix. In this paper, inform ation on the structure and properties o f the interphase betw een a filler and an elastomeric m atrix will be obtained from the study o f the m olecular dynam ics o f the corresponding poly­

meric chains. In this respect two m ajor questions will be stressed. On the one hand, how im m obilized are the elastomeric chains w hich are interacting with the solid surface? On the other hand, is a particular struc­

tural group o f the rubber chains more specifically af­

fected by the filler surface than any other ?

Filler/elastomer interactions:

the interphase

The blending o f a filler w ith an elastom er is known to be associated w ith the occurence o f two ty­

pes o f interactions w hich m ay be either purely m echa­

nical or m ay result from m ore com plex processes in­

volving for example physical or chemical interactions.

The former, known as ’’occluded rubber” , results from the absorption during the com pounding o f part o f the elastom eric m atrix in the voids o f the filler par­

ticles. This m olecular occlusion results in a shielding o f the corresponding material from the strain supported by the bulk o f the polymer in the compound under stress and, as a consequence, in an increase o f the effective volume o f the filler dispersed in the rubber [1],

The latter is the so-called ’’bound rubber” which corresponds to the quantity o f elastom er w hich can­

not be recovered from an uncured filler/elastom er blend upon solvent extraction. It is know n to be very much dependent on the surface activ ity o f the filler [2-4], This bound ru b b er to g eth er w ith the n e ig h ­

bouring entangled elastom er molecules makes up a po­

lym er shell o f finite thickness which surrounds the fil­

ler particles and forms the interphase. Previous works have pointed at the importance o f its occurence for the observation o f reinforcing effects. Indeed, the com ­ pounding o f a styrene-butadiene rubber (SBR) w ith a reinforcing carbon black (N326) the surface o f which has been deactivated by grafting o f hexadecyl chains (N 326C |6) yields a m aterial exhibiting a severely de­

pressed reinforcing ability as assessed from the 100 % m odulus o f the corresponding m aterials (Table 1).

Table 1. Stress strain properties o f SBR compounded with different carbon blacks [5]

Tabela 1. Naprężenie przy wydłużeniu 100% SBR za­

wierającego różne sadze [5]

Sample a 100 MPa

SBR 0,9

SBR/N326 2,1

SBR/N326C16 1,0

Filler/elastomer interphase:

structure and characteristics

Several experim ental techniques allow ing the obtention o f inform ations on the interphases and their features have been selected: dynamic m echanical te­

sting, differential scanning calorimetry, relaxational behavior o f strained samples, solid state N M R spec­

trometry. All o f them can allow an approach o f the interphase through the m olecular m obility o f polym er chains located in the vicinity o f the filler particles.

Dynam ic m echanical testing

The dynamic m echanical properties o f unfilled and silica filled SBR sam ples were studied using a M etravib viscoanalyzer. Storage moduli E ' and loss tangents (tan8) were m easured at 5 Hz in a -150°C to +80°C tem perature range. As expected, upon incre­

asing filler loading, E ' was shown to increase. Co­

nversely, the area under the tan8 peak which is in rela­

tion with the amount o f elastom er undergoing the glass transition appeared to decrease (Figure 1). Com pari­

son o f the experimental values with the theoretical ones obtained by m ultiplying the area measured for the pure

S fa A tM t& ity nr 4 lipiec-sierpień 1999 r. TOM 3

SBR by the polym er fraction in the com pounded m a­

terial, indicates that the form er are system atically lo­

w er than the latter. Such a result suggests that a part o f the polym er in a filled sam ple is not detected by dyna­

mic m echanical testing as passing through the therm al transition w ith in the con sid ered tem perature zone [6].This could be interpreted as due either to a part o f the rubber trapped in the filler particles and therefore not being subm itted to the m echanical sollicitation (i.e.

occluded rubber) or to part o f the chains im m obilized in a w ay such as to be practically in a glass-like state over the w hole tem perature range (i.e. bound rubber).

In order to settle betw een both possibilities, differen­

tial scanning calorim etry (DSC) appeared to be the technique o f choice.

sim ilar Tg as that o f the pure SBR, calorific capacity changes (ACp) norm alized for the polym er fraction only were shown to be dependent on the type o f sam ­ ple (Figure 2). I f the ACp values obtained w ith the filled rubbers are com parable to that o f the unfilled elastomer, the values obtained w ith the extracted sam ­ ples are consistently low er [6]. Such results suggest again that part o f the elastom er is not detected by DSC as undergoing the therm al transition. However, the com parison o f these inform ations with those yielded by dynam ic m echanical testing suggests that, since D SC-analyzed sam ples are in no way m echanically strained, the elastom er w hich is not detected as pas­

sing through the therm al transition within the consi­

dered tem perature zone can only be associated with the form ation o f bound rubber.

Fig. 1. Variation o f tanSpeak area vs filler loading ratio (experimental data, ■ ; theoretical data, -) Rys. 1. Zależność powierzchni piku tanSod stopnia na­

pełnienia (dane doświadczalne ■, dane teoretyczne -)

Differential scanning calorimetry

Experiments were performed on a M ettler DSC30 calorimeter. Values were considered out o f the second run from -150°C to +180°C at a 2°C/m in tem perature increment. From the obtained therm ograms, providing inform ations on first (melting, crystallization,...) and particularly second order transitions exhibited by the m aterials under study, the evolution o f the glass trans­

ition characteristics versus filler loading was exam i­

ned. W hile, on the average, filled compounds exhibit

Fig. 2. Variation o f ACp vs filler loading ratio fo r unfil­

led, filled ( □ ) and extracted-filled SBR (■ )

Rys. 2. Zależność ACp od stopnia napełnienia SBR:

nienapełnionego i napełnionego ( □ ) oraz wyekstra­

howanego napełnionego (■ )

It is thus clear that a part o f the elastomer, which is in the close vicinity o f the filler surface and partici­

pates to the form ation o f the interphase, is strongly im m obilized and exhibits a glass-like behavior. H o­

wever, gaining a better insight on what is the interpha­

se requires the obtention o f other informations such as the assessm ent o f the amount o f material which par­

ticipates to its form ation, a more precise determ ina­

tion o f its properties and an appreciation o f the degree o f m obility o f the corresponding polym er chains, par­

ticularly in order to evidence the occurence o f any

TOM 3 lipiec-sierpień 1999 r. S ta & ta tK en y nr 4

specificity o f interactions. Two different approaches were retained to provide these informations: a relaxational approach which takes advantage o f the non-equilibrium behavior exhibited by a confined polym er which, unli­

ke a free polym eric m aterial, is anisotropic in nature and has specific local tem perature dependence, and a spectral approach (solid state N M R spectrom etry) w hich could allow the assessm ent o f the m olecular m obility o f polym er chains and as a consequence p ro­

vide structural, therm al and m orphological inform a­

tions about the interphase.

Relaxational approach

It takes advantage o f the so-called physical age­

ing process, known to happen in glassy polymers. Upon application o f a static deform ation the storage m odu­

lus o f these m aterials first decreases suddenly, then increases continually (linear increase on a logarith­

mic time scale w ith a slope identified as p ) [7,8]. If such a behavior is unobservable when dealing with pure elastomers, it is no longer the case w hen study­

ing filled rubber samples. Such a result implies that at least a fraction o f the m atrix exhibits a low segmental m obility and hence behaves as a glassy polym er with respect to physical ageing phenom ena [10]. W ithin the limits o f this assum ption, this tool can be used to de­

tect the restriction o f m obility exhibited by the poly­

mer fraction which surrounds the filler particles and

Fig. 3. Carbon black/SBR blends evolution o f ju vs fil­

ler loading ratio

Rys. 3. Zależność jd mieszaniny sadza/SBR od stopnia napełnienia

as such participates to the interphase. In this respect the crucial param eter is p, the rate o f increase o f the modulus after deform ation. Its evolution versus filler volume fraction (<|)) points at different behaviors de­

pending on filler loading (Figure 3). At low (j), the in­

fluence o f the interphase on the value o f p results from particles acting individually. At intermediate filler lo­

adings, particles coated w ith an imm obilized rubber shell start to interact over long distances. At high lo­

adings, as a consequence o f the form ation o f a tridi­

m ensional filler-interphase network, the m obility o f a large part o f the rubber m atrix is severely depressed, although the corresponding m acrom olecules may not be directly immobilized on the filler surface. From this curve, one m ay define a critical filler loading ratio, (j)c, for w hich particles and their interphases start to inte­

ract. Thereafter, it is possible to estim ate the volum e fraction o f the interfacial region (c))ir) as being the dif­

ference betw een (J)c and a theoretical value, (J)0, a case in w hich there are no filler/m atrix interactions and no m atrix m odification at the interfacial region (natural com pactness limit). §c and (j)ir values obtained for a reinforcing silica and a m odel filler (glass beads) are reported in Table 2.

Table 2. Critical filler loading and immobilized SBR values [9]

Tabela 2. Krytyczny stopień napełnienia i zawartość SBR związanego [9]

Filler _Oi,

Silica 0,09 0,34

Glass beads 0,41 0,02

It appears that the values o f the critical filler lo­

ading ratio and o f the am ount o f im m obilized rubber, as well, are strongly dependent on the reinforcing abi­

lity o f the filler. Indeed, the use o f a non-reinforcing filler is associated w ith high (J)c values and a small frac­

tion o f the m atrix being involved in the surface re­

gion.

U sing the relaxational approach, it is thus p ossi­

ble to obtain inform ations on the properties o f the po­

lym er chains w hich are in close vicinity o f a solid sur­

face. Their mobility is strongly depressed and one m ay roughly quantify the am ount o f polym er w hich parti­

cipates to the interfacial domain. In order to confirm these observations and obtain a better understanding

nr 4 lipiec-sierpień 1999 r. TOM 3

o f the nature o f the interactions w hich occur in the filler/m atrix interphase and how they contribute to its formation, solid state N M R spectrom etry w as used.

Solid state N M R approach

At tem peratures well above the glass transition o f the elastomer, 2H solid state spectra obtained with carbon black-filled deuterated polybutadienes exhibit a double peak (Pake doublet) w hich is characteristic o f glassy or strongly imm obile deuterium together with a single and narrow peak associated with mobile 2H (Figure 4) [10]. Such a result confirm s that part o f the elastomeric chains are in a glass-like state, while at the same tem perature the pure elastom er is indicating the presence o f highly m obile deuterium . It is clear that these im m obilized rubber segments participate to the form ation o f the interphase.

Fig. 4. NMR spectra o f 2H-polybutadiene (a) polybutadiene/carbon black blends (b)

Rys. 4. Widmo NMR 2H-polibutadienu (a) i mieszani­

ny 2H-polibutadien/sadza (b)

Using a new approach, 'H high resolution high speed solid state N M R, inform ations m ay be obtained as for the specific effects exerted by the filler on the

different chemical species constitutive o f the polym er chains. In this respect, it was shown that it is possible, by spinning the sam ples at m agic angle and at a 15- kHz frequency, to obtain well resolved spectra and to assess from relaxation m easurem ents the evolution o f the intensity o f each peak versus relaxation time. In the case o f SBR/carbon black blends (filler loading ratio in the 0 to 100 phr range) previously extracted so as to get rid o f any free p o ly m er chains and e n h an ­ ce as a co n seq uen ce the answ er o f the interphase (bound rubber) we show ed that the plots o f the trans­

verse m agnetization relax atio n functions for the dif­

ferent types o f pro to n s (aro m atic, m ethine, m eth y ­ lene) could be fitted w ith a sum o f two exponential decay functions corresponding to short (T2s, relatively im m obilized m aterial) and long (T21, less im m obili­

zed m aterial) relaxation times, respectively. Following the evolution o f T2s versus carbon black loading, it was evidenced that the olefinic part o f the butadiene segments is the polym er m oiety which, w hatever the filler loading ratio, is the m ost affected by the carbon black surface (Table 3).

Table 3. Effect o f carbon black on T2sfo r aromatic, methine and methylene protons [11]

Tabela 3. Wpływ sadzy na T2s protonów aromatycz­

nych, metynowych i metylenowych [11]

Carbon black loading, phr

Ta , ms

Aromatic 1H Methylene 1H Methine 1H

0 0,50 0,22 0,75

20 0,52 0,17 0,12

50 0,48 0,16 0,14

80 0,45 0,13 0,19

Since butadiene polym erizes according to two different m icrostructures (1,2- and 1,4-), use o f a se­

ries o f polybutadienes (PB) o f identical m olecular w eight but different and well controlled 1,4- content indicates, from the com parison o f the slopes o f the lines relating T 2| to 1,2-polybutadiene content, that in the presence o f carbon black (filler loading ratio: 50 phr) the 1,2- fractions (vinyl protons) are more imm o­

bilized than the 1,4- m oieties (1,4- methylene protons) (Table 4).

This specificity o f interaction was confirmed by studying the evolution o f the ACp o f the samples ver­

sus polybutadiene m icrostructure. One m ay expect o f course this specificity o f interactions to be dependent

TOM 3 lipiec-sierpień 1999 r. nr 4

Table 4. Comparison o f the slopes o fT 2l vs 1,2-poly­

butadiene content fo r vinyl 1,2- and methylene 1,4- protons in the case o f pure, filled and extractedfilled- polybutadiene [12]

Tabela 4. Porównanie nachylenia linii zależności T2l od zawartości 1,2-polibutadienu dla protonów 1,2-wi- nylowych i 1,4-metylenowych w przypadku polibuta- dienu czystego, napełnionego i wyekstrahowanego na­

pełnionego [12]

Slope of T2I vs1,2-polybutadiene sample ____________ 5 !? !!!!____________

• =CH2 1,2-PB •C H j-1 ,4-l

Pure PB -0,34 -0,31

Filled PB -0,19 -0,27

Extracted filled-PB -0,04 -0,09

on the type o f polym er used but it appeared that it also depends on the elastom er/filler couple under investi­

gation. Thus, in the case o f SB R/silica com pounds the effect o f the filler is felt, as seen by N M R spectrom e­

try, at the level o f all branched fractions (vinyl and aromatic groups as well). N evertheless, assessing the intensity o f the silica/elastom er interactions through flow m icrocalorim etry, using as solute m olecules re­

presentative o f the different polym er moieties, we were in a position to show that the stronger therm al effects are obtained w ith arom atic com pounds, w hat w ould indicate that the adsorption o f SBR on the surface o f silica w ould preferentially proceed through its arom a­

tic m oieties [13].

Conclusion

In a filled elastom eric m atrix, the effect o f fillers is not limited to surface processes but is extending to the polym er phase surrounding the reinforcing partic­

les (the interphase). Dynam ic m echanical testing and differential scanning calorim etry show that part o f the rubber, certainly the m ost affected by the presence o f the filler, is not detected as passing through the glass transition tem perature, thus indicating a fraction o f the elastomer is in a glass-like state even at tem peratures

well above the polym er nom inal T . Studying the rela­

xation behaviour o f strained sam ples (physical ageing process) and using 2H and 'H solid state NM R, one may be in a position to evidence the presence o f the interphase, to quantify the am ount o f polym er which is thus affected and to show that the interactions which take place at the polym er/filler interface are se le c ti­

ve and depen dent on the fille r/p o ly m e r couple u n ­ der inv estigatio n. The natu re o f the co rresp o n ding m o d ificatio n s and the driving forces b eh ind them are, however, not clearly understood.

References

1. A. I. Medalia, „Filler Aggregates and their Effects on Dynamic Properties o f Rubber Vulcanizates in „Le renforcement des ćlaslomeres ”, CNRS Ed., Paris, pp. 63-79 (1975)

2. A. K. Sircar and A. Voet, Rubber Chem.Technol., 41, 973 (1970)

3. J. B. Donnet, M.J. Wang, E. Papirer and A. Vidal, Kaut. Gummi Kunstst.,19. (6), 510 (1986)

4. Y. C. Ou, Z.Z. Yu, A. Vidal andJ.B. Donnet, Rubber Chem. Technol.,6Z 834 (1994)

5. A. Vidal andB. Haidar, Angew. Makromol. Chemie, 20212m.(3632), 133 (1992)

6. M. Wagner, B. Haidar, A. Vidal, L. Delmotte, L. La- douce and Y. Bomal, Proc. IRC 97, Nuremberg, pp.65-67 (1987)

7. B. Haidar and T.L. Smith, Polymer, 21(10), 1904 (1990)

8. B. Haidar and T.L. Smith, Macromolecules, (1990)

9. B. Haidar, H. Salah Derradji, A. Vidal Papi­

rer, Macromol. Symp.,

1 0

10. B. Haidar, A. Vidal and E. Papirer, Proc. Eurofil- ler 97, Manchester, pp. 239-250 (1997)

11. N. K. Dutta, N. Roy Choudhury, B. Haidar, A. Vi­

dal, L. Delmotte and J. M. Chezeau, Polymer, 2 1 4293 (1994)

12. R. Mansencal, PhD Thesis, Univ. Haute Alsace, Mulhouse (1996)

13. M. Wagner, PhD Thesis, Univ. Haute Alsace, Mul­

house (1997)