the influence of non-rubber constituents on performance of silica reinforced natural rubber...
TRANSCRIPT
European Polymer Journal 49 (2013) 3199–3209
Contents lists available at SciVerse ScienceDirect
European Polymer Journal
journal homepage: www.elsevier .com/locate /europol j
The influence of non-rubber constituents on performance ofsilica reinforced natural rubber compounds
0014-3057/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.eurpolymj.2013.06.022
⇑ Corresponding author. Tel.: +31 53489 2529; fax: +31 53489 2151.E-mail address: [email protected] (J.W.M. Noorderm-
eer).
S.S. Sarkawi a,b, W.K. Dierkes a, J.W.M. Noordermeer a,⇑a University of Twente, Elastomer Technology and Engineering, PO Box 217, 7500 AE Enschede, The Netherlandsb Malaysian Rubber Board, RRIM Research Station, Sg. Buloh, 47000 Selangor, Malaysia
a r t i c l e i n f o
Article history:Received 16 October 2012Received in revised form 7 June 2013Accepted 13 June 2013Available online 1 July 2013
Keywords:Natural rubberSilicaSilaneProteinPayne effectTyre
a b s t r a c t
An in-rubber study of the interaction of silica with proteins present in natural rubber showthat the latter compete with the silane coupling agent during the silanisation reaction; thepresence of proteins makes the silane less efficient for improving dispersion and filler–polymer coupling, and thus influences the final properties of the rubber negatively. Fur-thermore, the protein content influences the rheological properties as well as filler–fillerand filler–polymer interactions. Stress strain properties also vary with protein content,as do dynamic properties. With high amounts of proteins present in natural rubber, theinteractions between proteins and silica are able to disrupt the silica–silica network andimprove silica dispersion. High amounts of proteins reduce the thermal sensitivity of thefiller–polymer network formation. The effect of proteins is most pronounced when nosilane is used, but they are not able to replace a coupling agent.
� 2013 Elsevier Ltd. All rights reserved.
1. Introduction
In recent years, the increasing demand for low-energyconsuming and low rolling resistance tyres has lead togrowing use of silica in tread compounds. Four essentialelements in silica–rubber technology: the rubber polymer,a special type of silica, an effective coupling agent and theappropriate mixing technology are interconnected inexpanding the magic triangle of tyre technology: the com-promise between rolling resistance (wet), traction andwear. Compared to carbon black, mixing silica compoundsinvolves many difficulties due to the large polarity differ-ence between silica and rubber. A bifunctional organosi-lane such as bis(triethoxysilylpropyl) tetrasulphide(TESPT) or its disulphide equivalent is commonly used ascoupling agent in enhancing the compatibility of silicaand rubber, by chemically modifying silica surfaces andeventually creating a chemical link between silica aggre-
gates and the rubber chains [1] as illustrated in Fig. 1. Com-plications arise during mixing silica compounds as severalchemical reactions need to take place, all at their appropri-ate time slots during rubber processing, namely the silicaand silane reaction or silanisation, silane–rubber couplingand crosslinking between the rubber chains [2].
The high-dispersion silica technology, as it is used to-day, employs mainly solution-polymerised synthetic rub-ber, and is still not commercially feasible with naturalrubber [3]. It was postulated that non-rubber constituentscontained in natural rubber such as proteins compete withthe coupling agent for reaction with the silica during mix-ing, so disturbing its reinforcement action [3]. However, nosupporting evidence is available on this subject.
Commercial Natural Rubber (NR) comes from the milkysap or latex that exudes from the rubber tree, Hevea Brasil-iensis, which coagulates on exposure to air. Hevea latexconsists of rubber hydrocarbon for about 30–45 wt% non-rubber constituents for about 3–5 wt%, and the rest iswater. The non-rubber constituents comprise of proteins,amino acids, carbohydrates, lipids, amines, nucleic acids,as well as other inorganic and mineral components [4].
Fig. 1. Silica-silane-rubber coupling.
Table 1Protein content of natural rubber’s used.
Rubber type Nitrogen content(wt%)
Protein content(wt%)
NR (SMR 20) 0.21 1.3DPNR
(pureprena)0.07 0.4
Skim rubber 2.06 12.9
3200 S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209
The work by Tanaka and coworkers has revealed that thefundamental structure of a linear NR chain consists of along sequence of 1000–3000 cis-1, 4 isoprene units, withat the a- and x-chain ends specific other groups [5,6].The a-terminal is composed of mono- and diphospategroups linked with phospholipids by hydrogen or ionicbonds [7]. The x-terminal entails two trans-1, 4 isopreneunits [8] and a modified dimethylallyl unit linked withfunctional groups, which is associated with proteins toform crosslinks via hydrogen bonding [9,10]. Both non-rubber constituents, i.e. proteins and phospholipids arepresumed to be the origin of branching and gel formationin NR [11]. These secondary structures play a significantrole in the strain-induced crystallization of unvulcanisedand vulcanised natural rubber [12,13].
The protein content of NR varies upon its source andmethods of production. The nitrogen content of the NR isrelated to the protein level. It is generally accepted thatthe conversion factor from nitrogen content to protein con-tent is 6.25 [4,14]. Typical raw NR has a nitrogen content inthe range of 0.3–0.6%. Many attempts have been made topurify NR from the non-rubber constituents such as pro-teins. One of the most successful attempts is the ‘Depro-teinised Natural Rubber’ (DPNR) which is characterisedby its very low nitrogen, ash and volatile matter contentscompared to the equivalent commercial NR. DPNR is pro-duced via treatment of natural rubber latex with bioen-zyme (proteinase), which hydrolyses the proteins presentinto water soluble forms [15]. In addition, deproteinisationof natural rubber is also achieved in the latex stage withurea in the presence of a surfactant, such as sodium dode-cyl sulphate [16,17]. On the other hand, during concen-trated latex production, the serum phase aftercentrifugation contains 5–10% of the total rubber, andmany of the non-rubbers. This is coagulated with sulphuricacid to produce skim rubber with a low dirt content andlight colour and that is relatively cheap. Skim rubber hasa high protein content [18], where the nitrogen contenthas values in the range of 1.5–2.5%.
Gregg and Macey [19] have demonstrated that theinsoluble non-rubber constituents in NR account for thedifferences in properties between compounded NR andcompounded synthetic polyisoprene. This non-rubbermaterial is mostly proteins and responsible for the higher
modulus, faster scorch time and higher tear strength ofNR. The protein is postulated to act as a reinforcing fillerat low concentration (3–4 wt.%) and as a cure activator.Othman and Hepburn [4] have shown that the presenceof proteins from B-, C-serum and proteolipids did not sig-nificantly affect the elastic modulus of rubber vulcanisate.However, the presence of its hydrolysed constituents, ami-no acids, gave a marked increase in the modulus of vulca-nisates, in particular alanine and arginine, the basic andneutral amino acids, respectively [20].
In the present paper, the influence of non-rubber con-stituents in NR, particularly proteins, on the properties ofsilica-filled NR compounds in the presence and absenceof coupling agent is illustrated. In order to demonstratethe variation of proteins content in NR, DPNR and skimrubber are selected in comparison with normal NR. The fil-ler–filler and filler-to rubber interactions of silica rein-forced NR compounds at varying mixing dumptemperatures is highlighted.
2. Experimental
2.1. Materials
Natural rubbers with different protein contents werecompared in this study. The rubbers with their proteincontents based on nitrogen estimations are summarisedin Table 1. The nitrogen contents of the three rubbers weredetermined by the semi-micro Kjeldahl procedure carriedout by the Materials Characterisation Unit, Malaysian Rub-ber Board (MRB).
The compound formulation used throughout this inves-tigation is shown in Table 2. Highly dispersible silica, Ultra-sil 7005 from Evonik with CTAB surface area of 164 m2/gwas used. The other compounding ingredients were usedas obtained from the respective sources, as indicated inTable 2.
2.2. Compounding
The compounds were mixed in 2 steps. The mixing pro-cedure is described in Table 3. The first step was doneusing a laboratory internal mixer Brabender Plasticoder350S lab station with a capacity of 390 ml. The fill factorof the mixer was fixed at 70% and the rotor speed usedwas 60 rpm. The starting temperature of the mixing cham-ber and rotor was varied from 70 to 120 �C in order to ob-tain variable temperature histories and dumptemperatures. After mixing for 14 min, the batches weresheeted out on a Schwabenthan Polymix 80T80 � 300 mm two-roll mill, under a tight nip with 10
Table 2Compound formulation.
Ingredients Source phr
Natural rubber (various types) Malaysian rubberboard
100a
Silica ultrasil 7005 Evonik 55Silane, TESPT Evonik 5Zinc oxide Sigma Aldrich 2.5Stearic acid Sigma Aldrich 1Processing oil, TDAE H & R Europe 82,2,4-trimethyl-1,2-dihydroquinoline
(TMQ)Flexsys 2
Sulphur Sigma Aldrich 1.4N-cyclohexyl-2-benzothiazyl
sulphenamide (CBS)Flexsys 1.7
Diphenyl guanidine (DPG) Flexsys 2
a For skim rubber, the formulation was adjusted to 112 phr to take intoaccount the high protein content.
Table 3Mixing Procedure.
Time (min) Action
Step 10 Add rubber (mastication)1 Add ½ silica and ½ silane5 Add ½ silica, ½ silane and oil9 Sweep11 Add zinc oxide, stearic acid and TMQ14 Dump
Step 2After 24 h the curatives were mixed on the two roll mill with 14 min
mixing time
S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209 3201
passes to improve the dispersion of silica. After 24 h, thecuratives were mixed on the two-roll mill in the secondstep, with 14 min mixing time.
2.3. Characterisation methods
Mooney viscosity was measured at 100 �C with a Moo-ney viscometer 2000E (Alpha Technologies) using the largerotor (ML(1 + 4)) for productive compounds and the smallrotor for non-productive masterbatches (MS(2 + 4)). Vulca-nisation curves were measured using a Rubber ProcessAnalyzer (RPA 2000) from Alpha Technologies, under con-ditions of 0.833 Hz and 2.79% strain over a period of 30 minat a temperature of 150 �C. The Payne effect was measuredprior and after cure in the RPA 2000 as well. Before cure thesample was heated to 100 �C in the RPA and subsequentlysubjected to a strain sweep at 0.5 Hz. The Payne effect wascalculated as the difference between the storage modulus,G0 at 0.56% and G0 at 100.04% strain. The Payne effect aftercure was measured after vulcanisation in the RPA 2000 at150 �C for 10 min and subsequent cooling to 100 �C, mak-ing use of the same strain sweep conditions.
Wolff’s filler structure parameter, af was determinedfrom the ratio between the increase in vulcameter torqueof the filled compounds and that of the unfilled gum [21]:
Dmax � Dmin
Domax � Do
min� 1 ¼ af
mf
mpð1Þ
where Dmax�Dmin is the maximum change in torque for thefilled rubber, Do
max � Domin is the maximum change in torque
for the unfilled gum rubber, mf/mp is the weight ratio of fil-ler to polymer, af is a filler specific constant which is inde-pendent of the cure system and closely related to themorphology of the filler
The bound rubber content (BRC) measurements wereperformed on unvulcanised samples by extracting the un-bound rubber with toluene at room temperature for sevendays in both normal and ammonia environment. Theammonia treatment of BRC was done to obtain the chem-ically bound rubber as ammonia cleaves the physical link-ages between rubber and silica [22,23]. The amount of BRC(g/g filler) was calculated by:
BRCðg=g fillerÞ ¼ wdry �winsolubles
wo �wfiller;phrwtotal;phr
ð2Þ
where wo is the initial weight of the sample, wdry is the dryweight of the extracted sample, winsolubles is the weight ofinsolubles (mainly filler) in the sample, wfiller,phr is the totalfiller weight in phr and wtotal,phr is the total compoundweight in parts per hundred rubber (phr). The physicallyBRC was taken as the difference between untreated BRCand ammonia treated BRC.
Vulcanisates were prepared by curing the compoundsfor their respective t95 (time to reach 95% of torque differ-ence in the curemeter) at 150 �C using a Wickert laboratorypress WLP 1600/5*4/3 at 100 bar. Tensile properties of thevulcanisates were measured using a Zwick Z020 tensiletester according to ISO-37. The hardness of the cured sam-ples was determined according to DIN-53505. The tan del-ta (G00/G0) at 60 �C was measured using the RPA 2000 byapplying a frequency sweep at 3.49% strain after first cur-ing in the RPA at 150 �C.
An apparent crosslink density was determined by swell-ing a vulcanised sample in toluene. The vulcanised sampleof about 0.2 g was cut from a sheet with a thickness of2 mm and immersed in 50 ml toluene at room temperaturefor 72 h. The solvent was renewed after 24 h. The samplewas removed, blotted quickly with filter paper andweighed in a tared weighing bottle. The samples were im-mersed in acetone for 30 min to remove the remaining sol-vent. The samples were collected and left for 24 h at roomtemperature in a fume hood before the dried weight wasmeasured. The swelling value Q, defined as grams of tolu-ene per gram of rubber hydrocarbon, was calculated as:
Q ¼ wswollen �wdried
wo � 100wtotal;phr
ð3Þ
where wswollen is the swollen weight, wdried is the driedweight, wo is the initial weight and wtotal,phr is the total for-mula weight in phr. The apparent crosslink density wascalculated as the reciprocal swelling value, 1/Q.
3. Results and discussion
3.1. Processability of silica-filled NR compounds
In terms of processability of the masterbatches, NR andDPNR are comparable, but skim rubber has a lower
0
10
20
30
40
50
60
70
80
90
100 120 140 160 180
MS
(2+4
) @
100°C
Dump Temperature,°C
(a) masterbatches
0
10
20
30
40
50
60
70
80
100 120 140 160 180
ML
(1+
4) @
100° C
Dump Temperature,°C
(b) compounds
Fig. 2. Mooney viscosities of: (a) masterbatches after 1st mixing, and (b) compounds after 2nd mill mixing of silica-filled natural rubber at varying proteincontents: (d) 0.4% (DPNR); (j) 1% (NR); (D) 12% (SkimRubber).
3202 S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209
viscosity. In Fig. 2, the increase in viscosity of the master-batches with rising mixer dump temperature up to a tem-perature of 150 �C is a combination of the hydrodynamiceffect and silanisation rate of the silica. More silica ishydrophobised by TESPT when the dump temperature israised, and this results in a higher compatibility betweensilica and rubber and consequently increment of the vis-cosity. However, the viscosity of the masterbatches of NRand DPNR start to decrease above the optimum dump tem-perature, but in the case of skim rubber it levels off. Oneexplanation is the degradation of the NR chains at highertemperatures, which seems to be inhibited by a high pro-tein content.
Once the curatives are added to the compounds, the vis-cosities drop to processable levels, mainly due to the remil-ling step. In spite of the overall lower Mooney viscosities ofthe skim rubber masterbatches after the first mixing step,the Mooney viscosities of the compounds with curativesafter mill mixing are almost comparable with those ofthe NR and DPNR compounds. A comparable Mooney
0
2
4
6
8
10
12
14
16
0 5 10 15 20
Tor
que,
dN
m
Time, minutes
Without silane
DPNR (0.4%)
NR (1%)
SkimRubber (12%)
Fig. 3. Comparison of vulcanisation curves of silica
viscosity between NR, DPNR and skim rubber compoundsis due to many factors which include silica dispersion,the interaction of polar curatives with silica, the remillingstep and reduction of chain entanglements or pseudo net-work of rubber chains.
3.2. Effect of protein on filler–filler interaction
The influence of proteins in NR on the silica–silica inter-action can be clearly observed from the vulcanisationcurve as depicted in Fig. 3. The clear two-step curve forNR–silica and DPNR–silica compounds without silane isdue to the silica flocculation or re-agglomeration [24,25]and strong silica networking. With high amounts of proteinpresent in the compound, the silica–silica interaction isdisrupted and this is shown with no sign of flocculationat the beginning of vulcanisation for the skim rubber–silicacompound without silane. A longer scorch time andslightly faster cure time is also observed for the skim
0
2
4
6
8
10
12
14
16
0 5 10 15 20
Tor
que,
dN
m
Time, minutes
With silane
DPNR (0.4%)
NR (1%)
SkimRubber (12%)
compounds with silane and without silane.
0
10
20
30
40
50
60
70
80
90
100
100 120 140 160 180
Cur
e R
ate
Inde
x
Dump Temperature, °CDPNR (0.4%)
NR (1%)
(b)
0
2
4
6
8
10
12
14
16
100 120 140 160 180
ΔTor
que,
dN
m
Dump Temperature, °CDPNR (0.4%)
NR (1%)
SkimR (12%) SkimR (12%)
(a)
Fig. 4. Comparison of the cure characteristic of NR–silica–TESPT compounds with different protein contents as influenced by mixing dump temperature: (a)torque difference and (b) cure rate index.
S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209 3203
rubber compound indicating an effect of protein onvulcanisation speed as well.
The use of a silane, TESPT in this case, in the NR–silicacompound results in less pronounced silica flocculationand this is demonstrated by a small initial torque rise atthe beginning of vulcanisation (Fig. 3). As compared tothe silica compounds without silane, the flocculation of sil-ica in the compounds with TESPT is small due to hydropho-bation of the silica surface by TESPT. The effect of proteinon the cure behaviour of the silica compounds fades withthe presence of TESPT.
The thermal history and in particular the dump temper-ature has been shown to be a parameter of paramountimportance in mixing silica and rubber in presence of TES-PT as coupling agent in order to achieve proper silanisationof silica and to avoid premature scorch reactions [26–28].The strong influence of dump temperature on the NR–sil-ica–TESPT compound is reflected in the torque difference(D torque) and cure rate index as shown in Fig. 4. At highdump temperature, the NR–silica–TESPT compound exhib-its a considerable reduction in D torque and a suddenremarkable increase in the cure rate index. It seems thatwith increasing silanisation, the NR compound shows ashorter cure time due to enhancement in cure efficiencywith the use of TESPT. This leads to anomalous behaviourof the cure rate index for the NR compound mixed at high-er temperature. Moreover, with higher silanisation filler–filler interactions are reduced and this leads to lower max-imum torque. DPNR–silica–TESPT also shows a decrease inD torque with increasing dump temperature, but to a les-ser extent than that of NR. On the other hand, the high pro-tein content skim rubber–silica–TESPT compound shows adifferent behaviour where D torque is not influenced bythe dump temperature. This may indicate a strong influ-ence of the proteins in the silica compound and suggeststhe contribution of a silica-protein network in the curedcompounds.
Filler–filler interaction is commonly measured by theso-called Payne effect: the drop in storage modulus in a dy-namic mechanical test when the strain (deformation) is in-creased from low (0.56%) to a high value (100%) at constantfrequency and temperature. The storage modulus of filledrubber drastically decreases as the strain increases. Thisis the result of breakage of physical bonds between filleraggregates, for example van der Waals, hydrogen bondsand London forces.
The use of silica without silane modification in rubberresults in a high Payne effect due to strong interaggregateinteraction of silica. With TESPT modification, the Payne ef-fect of the silica-filled compounds is greatly reduced as thesilica surface is hydrophobised by TESPT, and the silica–sil-ica network is disrupted as shown in Fig. 5. What is inter-esting in this study is that the same effect can be seen withprotein. With a high amount of protein present in the rub-ber, the Payne effect of the silica compound without silaneis lowered. There is a relation between the amount of pro-tein and the decrease of silica–silica interaction. This indi-cates a strong affinity of the proteins towards silica, as wellas the role of proteins in hydrophobising the silica surface.The interaction between proteins and silica most likelycome from hydrogen bonding between silanol groups onthe surface of the silica to amide and carbonyl groups ofproteins. In the presence of a silane, the effect of proteinbecomes more pronounced in the Payne effect after vulca-nisation. An increase in the Payne effect at higher proteincontent indirectly indicates competition between proteinand silane for the silica surface. This results in reductionof silanisation efficiency and consequently less hydropho-bation of the silica surface.
For silica compounds with silane, the Payne effect of NRand DPNR compounds decreases sharply with increasingdump temperature, as is also seen in synthetic rubber/sil-ica compounds and taken as a sign of reaction and conse-quent hydrophobation of the silica by the silane coupling
0
0.1
0.2
0.3
0.4
0.5
0.1 1 10G' a
t 0.
56%
-G
' at
100%
str
ain,
MP
a
Protein content, wt. %
(a) Before vulcanisation
No silane
silane
0.0
0.4
0.8
1.2
1.6
2.0
2.4
0.1 1 10G' a
t 0.
56%
-G
' at
100%
str
ain,
MP
a
Protein content, wt. %
(b) After vulcanisation
No silaneSilane
Fig. 5. Influence of proteins on the Payne effect of NR–silica compounds: (a) before vulcanisation (b) after vulcanisation; (s): silica compounds with silaneTESPT; (N): silica compounds without silane.
3204 S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209
agent [26,27]. No effect of mixing temperature is perceivedon filler–filler interaction for the skim rubber compound,Fig. 6; this is observed for the unvulcanised as well as forthe vulcanised compounds. This again indicates a stronginterference of the proteins in skim rubber with the fil-ler–filler network. For skim rubber, the silica–silica net-work is not influenced by the dump temperature even inthe presence of silane because silanisation is hindered. Itcan be seen in Fig. 6 that the Payne effect of the vulcanisedskim rubber compound is higher than for the NR and DPNRcompounds for the higher dump temperatures. The pro-teins in the skim rubber prevent the modification of the sil-ica surface by the silane coupling agent. The logicalexplanation is that the interaction between silica and pro-tein overrules the coupling agent and that the protein isshielding the silica surface.
3.3. Effect of protein on rubber-to-filler interaction
The difference between the NR compounds with vary-ing protein content is further illustrated in Fig. 7, coveringthe effect of mixing temperature on Wolff’s filler structure
0.00
0.05
0.10
0.15
0.20
0.25
0.30
100 120 140 160 180
G' a
t 0.
56%
-G
' at
100%
str
ain
(MP
a)
Dump Temperature, ºC
(a) Before vulcanisation
DPNR (0.4%)NR (1%)SkimR (12%)
Fig. 6. Payne effect of silica compounds with TESPT at varying protein contents
parameter, af. For compounds without silane, af is muchhigher than for those with TESPT, except for skim rubber.The af value for skim rubber without silane is even lowerthan the typical values obtained by Wolff for carbon blackcompounds: Table 4 [29]. As observed earlier with the Pay-ne effect, af is reduced with increasing dump temperaturefor NR and DPNR compounds containing silane. This is dueto the increased hydrophobation of the silica by the silani-sation reaction at higher mixing temperatures. Betterhydrophobation leads to a decrease in silica–silica interac-tion and consequently results in reduced af. The af of NRmixed till high dump temperature is comparable to thetypical value obtained by Wolff for TESPT-modified-silicacompounds based on NR: Table 4. The DPNR compoundshows a higher af than the NR compound, indicating a dif-ferent type of filler and rubber network in the two com-pounds. For the high protein content skim rubbercompound, the af value is much smaller than for the NRand DPNR compounds, and is constant regardless of thechanges in the dump temperature. This corresponds withthe results of its Payne effect described above. The non-rubber constituents or proteins in the skim rubber again
0.0
0.5
1.0
1.5
2.0
2.5
100 120 140 160 180
G' a
t 0.
56%
-G
' at
100%
str
ain
(MP
a)
Dump Temperature, ºC
(b) After vulcanisation
DPNR (0.4%)NR (1%)SkimR (12%)
in natural rubber: (a) unvulcanised samples and (b) vulcanised samples.
0
2
4
6
8
10
12
14
100 110 120 130 140 150 160 170 180
Wol
ff's
fille
r st
ruct
ure
para
met
er, α
f
Dump temperature, °C
DPNR (0.4%)
NR (1%)
SkimR (12%)
NR no-silane
DPNR no-silane
Fig. 7. Effect of dump temperature on Wolff’s filler structure parameter, af, of silica reinforced natural rubber with varying protein contents.
Table 4Typical Wolff’s filler structure parameters (af)29.
Filler af
Carbon blacka 1.86Unmodified silicab 5.65TESPT treated silica 1.84
a Corax N110.b Ultrasil VN2.
S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209 3205
plays the main role in the formation of the filler structure,and mixing temperature again has little influence on thecompound properties.
The filler to rubber interaction of silica-filled NR withvarying protein content can also be judged on basis ofthe chemically and physically bound rubber as illustratedin Fig. 8. The chemically bound rubber content of theNR–silica-TESPT compound increases with increasingdump temperature up to 150 �C, and above 150 �C it tendsto decrease slightly. This can be explained by the higherrate of silanisation. At 150 �C, there is saturation in theamount of TESPT which has reacted and the surface ofsilica is completely covered. Hence, there is no further
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
100 120 140 160
BR
C, g
/g s
ilica
Dump Temperature, °C
(a) Chemically BRC
Fig. 8. Comparison of: (a) chemically and (b) physically bound rubber content(DPNR); (j) 1% (NR); (D) 12% (SkimRubber).
increase in chemically bound rubber above 150 �C. Thelow protein DPNR compound also shows an increase inchemically bound rubber with increasing dump tempera-ture, but has an optimum at 155 �C dump temperature.The skim rubber compound has no chemically bound rub-ber at all at low dump temperatures below 140 �C. Thisclearly indicates that no silanisation occurs and that theinteraction of protein and silica in skim rubber is purelyphysical of nature. Surprisingly, above 150 �C, the skimrubber compound has a constant, and high chemicallybound rubber content comparable to DPNR. This suggeststhat some part of the silica surface is silanised althoughat lower efficiency due to competition between proteinand silane as discussed in the filler–filler interactions. An-other explanation of the higher chemically bound rubber ofskim rubber is the effect of crosslinks generated by sulphurreleased from TESPT at the high mixing temperature. Thispremature crosslinking is possibly accelerated by proteinin the case of skim rubber.
In Fig. 8(b), the increase in physically bound rubber forthe NR–silica compound at higher dump temperatures canbe explained by the completion of the silica–TESPT cou-pling. Additional interactions above 150 �C between the
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
100 120 140 160
BR
C, g
/g s
ilica
Dump Temperature, °C
(b) Physically BRC
of NR–silica–TESPT compounds with varying protein content: (d) 0.4%
0
10
20
30
40
50
60
70
80
90
0 20 40 60 80
MS
(2+4
) @
100
ºC
Chemically Bound Rubber, %
DPNR (0.4%)
NR (1%)
SkimR (12%)
Fig. 9. Mooney viscosity of NR–silica–TESPT masterbatches as a functionof chemically bound rubber.
Table 5Bound rubber content (BRC) of silica compounds without silane.
Rubber type Physically BRC (%) Chemically BRC (%)
NR 57 0DPNR 45 0Skim rubber 51 0
3206 S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209
non-hydrophobised silica surfaces and rubber are physicalof nature. For the low protein content DPNR compound,the physically bound rubber is lowest throughout thewhole dump temperature range, but also still shows aslight tendency to grow at higher dump temperatures. Fur-ther, a strong interaction between protein and silica can beseen in the high physically bound rubber values for theskim rubber at low dump temperatures. Above dump tem-perature of 150 �C, an equal level of physically bound rub-ber of skim rubber and DPNR is observed indicating thephysical interactions caused by proteins are destroyed.
Most of the bound rubber formed in a NR–silica–TESPTcompound is chemically attached. This is obviously due tothe hydrophobation of the silica surface as a result ofsilanisation with TESPT. The increase in silica-TESPT cou-pling consequently results in more filler-to-rubber interac-tion. This corresponds well with the lower Payne effect ofthe silica compounds with TESPT. The trends seen in chem-ically bound rubber correlate surprisingly well with theMooney viscosities of the masterbatches with the excep-tion of skim rubber: Fig. 9. Apparently, when more chem-ically bound rubber is formed, the molecular motion ofthe rubber chains is restricted and this results in higherMooney viscosity. This emphasizes the importance ofchemical coupling of the rubber to the silica surface. Inthe skim rubber, the increase in the chemically bound rub-ber seems to be constrained, as the silanisation is hinderedby protein, and hence no changes in the Mooney viscosityare observed.
Without silane in the compounds, there is still silica–rubber interaction, as indicated by the physically boundrubber in Table 5. There are many factors that could con-tribute to the formation of bound rubber content, whichinvolves physical adsorption, chemisorption and mechani-cal interaction. The results demonstrate that the proteinscontained in skim rubber do interact with the silica, mainlyvia hydrogen bonding, make it less hydrophilic and thusincrease the rubber–silica interaction. A comparison of fil-ler-to-rubber interactions in compounds with silane andwithout silane is graphically represented in Fig. 10. How-ever, no chemically bound rubber was obtained for the sil-ica compounds without silane after ammonia treatmentbecause all the hydrogen bonds are destroyed. This meansthat without silane in the compound, only loosely or phys-ically bound rubber is formed. This again indicates that the
Fig. 10. Illustration on the filler-to-rubber interaction for a silica com
interaction of silica with NR in the absence of couplingagent is not as strong as compared to covalent bonds in acompound with silane. A slightly higher physically boundrubber content for the NR than for the skim rubber com-pound relates to the effect of higher molecular weight ofNR. Many studies have established increase of bound rub-ber and preferential adsorption onto reinforcing fillerswith higher molecular weight polymers [30–32]. A longerrubber chain will have more inter-aggregate connections,and this will contribute to more bound rubber formation.In addition, filler morphologies also a contributing factorto formation of bound rubber. The stronger silica–silicainteraction in the NR compound without silane as shownby the higher Payne effect increases the bound rubber con-tent. This is in accordance with the occluded bound rubbermodel where the rubber is trapped in the filler aggregate.
pound: (a) with silane and (b) with protein and without silane.
0
4
8
12
16
20
24
28
32
100 120 140 160 180
Tens
ile S
tren
gth,
MP
a
Dump Temperature,ºC
0
100
200
300
400
500
600
700
100 120 140 160 180
Elo
ngat
ion
at b
reak
, %
Dump Temperature,ºC
0
4
8
12
16
20
100 120 140 160 180
Mod
ulus
at
300%
(M
300)
, MP
a
Dump Temperature,ºC
30
35
40
45
50
55
60
65
70
75
80
100 120 140 160 180
Har
dnes
s, S
hore
A
Dump Temperature,ºC
(a) (b)
(c) (d)
Fig. 11. Comparison of physical properties of silica–TESPT–NR vulcanisates with different amounts of protein content: (d) 0.4% (DPNR); (j) 1% (NR); (D)12% (SkimRubber); and silica compounds without silane: (+) NR; (e) DPNR; (�) Skim Rubber.
0.0
0.1
0.2
0.3
0.4
0.5
100 110 120 130 140 150 160 170 180
1/Q
Dump Temperature, ºC
DPNR (0.4%)NR (1%)SkimR (12%)
Fig. 12. Apparent crosslink density of silica–TESPT reinforced naturalrubber vulcanisates with varying protein content: (d) 0.4% (DPNR); (j)1% (NR); (D) 12% (SkimRubber).
S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209 3207
3.4. Physical and dynamic properties of silica reinforced NR
The use of TESPT as a coupling agent improves the vul-canisate properties of silica-filled compounds. Vulcanisateswithout silane exhibit inferior tensile strength than thosewith silane as depicted in Fig. 11. NR–silica–TESPT vulcani-sates mixed till higher dump temperatures exhibit a slightdrop in tensile strength, elongation at break and modulivalues. On the other hand, the effect of dump temperatureis less pronounced for DPNR–silica–TESPT vulcanisates,showing more constant tensile strength. This relates tothe crosslink density of the vulcanisates as shown inFig. 12. The DPNR–silica–TESPT vulcanisates exhibit arather constant crosslink density while NR–silica–TESPTvulcanisates show a reduction in crosslink density at high-er dump temperature. The decrease in the elongation atbreak for NR and DPNR when the dump temperature ex-ceeds 150 �C is most probably caused by the degradationof natural rubber chains under high temperature. NR andDPNR vulcanisates have comparable hardness and showquite a decrease in hardness with increasing dump tem-perature. The skim rubber vulcanisates perform overall
much worse compared to NR and DPNR vulcanisates interms of physical properties, due to lack of silica–rubbercoupling and lower molecular weight of the polymers tostart with. Skim rubber shows higher modulus and
0
1
2
3
4
5
6
7
100 120 140 160 180
M30
0 / M
100
Dump Temperature,ºC
0.00
0.04
0.08
0.12
0.16
0.20
100 120 140 160 180
Tan
δat
60°
C
Dump Temperature,ºC
(a) (b)
Fig. 13. Comparison of: (a) reinforcement index M300/M100 and (b) tand at 60 �C for natural rubber vulcanisates with varying protein content: (d) 0.4%(DPNR); (j) 1% (NR); (D) 12% (SkimRubber); and silica vulcanisates without silane: (+) NR; (e) DPNR; (�) Skim Rubber.
3208 S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209
hardness. This correlates with the highest crosslink densityof skim rubber compared to NR and DPNR as shown inFig. 12. The higher crosslink of skim rubber is also in agree-ment with the earlier discussion on the premature cross-links generated by sulphur released from TESPT duringhigh mixing temperature, which results in higher chemi-cally bound rubber. This premature crosslinking is possiblyaccelerated by protein in the skim rubber.
Commonly, two characteristic properties of cured com-pounds are accepted as indications for the rolling resis-tance of tyres made thereof: the reinforcement index orM300/M100 and the loss angle tand at 60 �C. The higherthe M300/M100 or the lower the tand at 60 �C, the lowerthe rolling resistance expected in real tyre performance.Fig. 13 illustrates both indicators of rolling resistance ofthe silica-filled vulcanisates. Both, NR–silica and DPNR–sil-ica have a superior reinforcement index as compared toskim rubber–silica vulcanisates. With increasing mixingtemperature, the reinforcement index of the NR vulcani-sate is somewhat improved. The processing conditionshave no influence on the reinforcement index M300/M100 of DPNR and skim rubber vulcanisates, where skimrubber is substantially lower than NR and DPNR. Thesedata seem to indicate that the correlation between rein-forcement index and rolling resistance only applies forNR, but not for DPNR and skim rubber. This also relatesto the constant crosslink density of DPNR and skim rubbervulcanisates.
All natural rubber vulcanisates show a strong decreasein tand at 60 �C with increasing dump temperature regard-less of the amount of protein in the rubber. Improvementin tand at 60 �C can still be achieved with higher mixingtemperatures, like with synthetic rubber. This must obvi-ously be the result of more coupling of silica to the rubberwith greater silanisation efficiency at high temperatures.This is further supported by the high tand at 60 �C for silicacompounds without silane. Due to low protein content, theDPNR vulcanisates with high dump temperatures exhibitthe lowest tand values at 60 �C. This actually relates wellwith a higher chemically bound rubber content of DPNRthan of NR. Still with all the protein contained in skim rub-ber, the tand at 60 �C is significantly lowered by increasingmixing temperature, and only marginally worse than for
NR and DPNR. It is further quite surprising that the rein-forcement index and tand at 60 �C show opposite behav-iour in the present NR systems, compared to syntheticrubber based silica-filled compounds [33].
4. Conclusions
Coupling agent and proteins show a complicated antag-onistic effect in silica reinforcement of natural rubber. Theeffect of proteins is most pronounced when no silane isused in NR–silica compounds. When high amounts of pro-teins are present in NR, the interactions between proteinsand silica are able to disrupt the silica–silica networking.The temperature development is an important parameterin mixing NR–silica with the aid of TESPT as couplingagent, as silica–silica interaction is reduced through silani-sation at sufficiently high mixing temperatures. This isclearly the case for NR and low protein content rubberDPNR. However, mixing temperature has little influenceon the properties of a high protein-content skim rubbercompound. Consequently, the hydrophobation of the silicasurface by silane may be partially hindered due to silica-protein interactions.
Acknowledgement
The authors would like to express gratitude for financialsupport from the Malaysian Rubber Board for this project.
References
[1] Meon W, Blume A, Luginsland H-D, Uhrlandt S. In: Rodgers B, editor.Rubber compounding: chemistry and applications. NewYork: Marcel Dekker Inc.; 2004 [chapter 7].
[2] Noordermeer JW, Dierkes WK. In: White J, De SK, Naskar K, editors.Rubber technologist’s handbook, vol. 2. Shawbury (UK): SmithersRapra Technology; 2008. p. 59–96.
[3] Sarkawi SS, Dierkes WK, Noordermeer JWM. Energy-saving tyresbased on natural rubber. In: EU-PEARLS 2010 meeting: the future ofnatural rubber, Montpellier, France; 2010.
[4] Othman AB, Hepburn C. Influence of non-rubber constituents onelastic properties of natural rubber vulcanizates. Plas, RubberCompos Proc Appl 1993;19:185–94.
[5] Tanaka Y, Tarachiwin L. Recent advances in structuralcharacterization of natural rubber. Rubber Chem Technol2009;82:283–314.
S.S. Sarkawi et al. / European Polymer Journal 49 (2013) 3199–3209 3209
[6] Tarachiwin L, Sakdapipanich J, Ute K, Kitayama T, Bamba T, FukusakiE, et al. Structural characterization of a-terminal group of naturalrubber. 1. Decomposition of branch-points by lipase andphosphatase treatments. Biomacromolecules 2005;6:1851–7.
[7] Sakdapipanich J. Structural characterization of natural rubber basedon recent evidence from selective enzymatic treatments. J BiosciBioeng 2007;103:287–92.
[8] Eng AH, Kawahara S, Tanaka Y. Trans-isoprene units in naturalrubber. Rubber Chem Technol 1993;67:159–68.
[9] Nawamawat K, Sakdapipanich J, Mekkriengkrai D, Tanaka Y.Structure of branche-points in natural rubber. Kautsch GummiKunstst 2008;61:518–22.
[10] Mekkriengkrai D, Sakdapipanich J, Tanaka Y. Structuralcharacterization of terminal groups in natural rubber: origin ofnitrogenous groups. Rubber Chem Technol 2006;79:366–79.
[11] Tarachiwin L, Sakdapipanich J, Tanaka Y. Structure and origin oflong-chain branching and gel in natural rubber. Kautsch GummiKunstst 2005;58:115–22.
[12] Amnuaypornsri S, Sakdapipanich J, Toki S, Hsiao BS, Ichikawa N,Tanaka Y. Strain-induced crystallization of natural rubber: effect ofproteins and phospholipids. Rubber Chem Technol 2008;81:753–66.
[13] Amnuaypornsri S, Nimpaiboon A, Sakdapipanich J. Role ofphospholipids and proteins on gel formation and physicalproperties of NR during accelerated storage. Kautsch GummiKunstst 2009;62:88–90.
[14] Loadman MJR, Wake WC. Analysis of rubber and rubber-likepolymers. Dordrecht: Kluwer Academic Publishers; 1988. p. 96.
[15] Barlow FW. Rubber compounding: principles, materials andtechniques. New York: Marcel Dekker; 1988. p. 9–28.
[16] Kawahara S, Klinkai W, Kuroda H, Isono Y. Removal of proteins fromnatural rubber with urea. Polym Adv Technol 2004;15:181–4.
[17] Klinklai W, Saito T, Kawahara S, Tashiro K, Suzuki Y, SakdapipanichJT, et al. Hyperdeproteinized natural rubber prepared with urea. JAppl Polym Sci 2004;93:555–9.
[18] Salamone DJC. Polymeric materials encyclopedia. Boca Raton(FL): CRC Press; 1996. p. 4556.
[19] Gregg EC, Macey JH. The relationship of properties of syntheticpoly(isoprene) and natural rubber in the factory: The effect of non-Rubber constituents of natural rubber. Rubber Chem Technol1973;46:47–66.
[20] Othman AB, Hasma H. Influence of Hevea proteins and amino-acidson properties of natural rubber. In: Proceedings international rubbertechnology conference; 1988. p. 166–77.
[21] Wolff S. Möglichkeit einer neuen charakterisierung derwirkungsweise von ruben in 1, 5-polyenen. Kautsch GummiKunstst 1970;23:7–14.
[22] Polmanteer KE, Lentz CW. Reinforcement studies – effect of silicastructure on properties and crosslink density. Rubber Chem Technol1975;48:795–809.
[23] Wolff S, Wang M-J, Tan E-H. Filler–elastomer interactions Part VII:study on bound rubber. Rubber Chem Technol 1992;66:163–76.
[24] Lin CJ, Hergenrother WL, Alexanian E, Böhm GGA. On the fillerflocculation in silica-filled rubbers Part I. Quantifying and trackingthe filler flocculation and polymer–filler interactions in theunvulcanized rubber compounds. Rubber Chem Technol2002;75:865–90.
[25] Mihara S, Datta RN, Noordermeer JWM. Flocculation in silicareinforced rubber compounds. Rubber Chem Technol2009;82:524–40.
[26] Reuvekamp LAEM, ten Brinke JW, van Swaaij PJ, Noordermeer JWM.Effects of time and temperature on the reaction of TESPT silanecoupling agent during mixing with silica filler and tyre rubber.Rubber Chem Technol 2002;75:187–98.
[27] Kaewsakul WK, Sahakaro K, Dierkes WK, Noordermeer JWM.Optimization of mixing conditions for silica-reinforced naturalrubber tyre tread compounds. Rubber Chem Technol2012;85(2):277–94.
[28] Sarkawi SS, Dierkes WK, Noordermeer JWM. In: Proceedings ofmalaysia polymer international conference 2011. Bangi-Putrajaya,Malaysia; 2011.
[29] Wolff S. Reinforcing and vulcanization effects of silane Si 69 in silica-filled compounds. Kautsch Gummi Kunstst 1981;34:280–4.
[30] Villars DS. Studies on carbon black III theory of bound rubber. JPolym Sci 1956;21:257–71.
[31] Kraus G, Gruver JT. Molecular weight effects in adsorption of rubberson carbon black. Rubber Chem Technol 1968;41:1256–70.
[32] Meissner B. Theory of bound rubber. Rubber Chem Technol1975;48:810–8.
[33] ten Brinke JW, van Swaaij PJ, Reuvekamp LAEM, Noordermeer JWM.The influence of silane sulfur and carbon rank on processing of asilica reinforced tyre tread compound. Rubber Chem Technol2003;76:12.