DSpace at VNU: Formation and stability of carbon nanotube network in natural rubber: Effect of non-rubber components tài...
Accepted Manuscript Formation and stability of carbon nanotube network in natural rubber: Effect of nonrubber components H.H Le, T Pham, S Henning, J Klehm, S Winer, K.-W Stưckelhuber, A Das, X.T Hoang, Q.K Do, M Wu, N Vennemann, G Heinrich, H.-J Radusch PII: S0032-3861(15)30126-9 DOI: 10.1016/j.polymer.2015.07.044 Reference: JPOL 18004 To appear in: Polymer Received Date: February 2015 Revised Date: June 2015 Accepted Date: 24 July 2015 Please cite this article as: Le HH, Pham T, Henning S, Klehm J, Winer S, Stưckelhuber K-W, Das A, Hoang XT, Do QK, Wu M, Vennemann N, Heinrich G, Radusch H-J, Formation and stability of carbon nanotube network in natural rubber: Effect of non-rubber components, Polymer (2015), doi: 10.1016/ j.polymer.2015.07.044 This is a PDF file of an unedited manuscript that has been accepted for publication As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Formation and stability of carbon nanotube network in natural rubber: Effect of nonrubber components H H Le1,2*, T Pham3, S Henning4, J Klehm4, S Winer1,5, K.-W Stưckelhuber1, A Das1, RI PT X T Hoang6, Q K Do7, M Wu8, N Vennemann8, G Heinrich1,5, H.-J Radusch9 Leibniz-Institut für Polymerforschung Dresden, Hohe Str 6, D-01069 Dresden, Germany Institut für Polymerwerkstoffe e.V (IPW), An-Institut der Hochschule Merseburg, D-06217 SC Merseburg, Germany Hochschule Albstadt-Sigmaringen, D-72458 Albstadt-Ebingen, Germany Fraunhofer IWM, Walter-Hülse-Str 1, D-06120 Halle/S., Germany Technische Universität Dresden, Institut für Werkstoffwissenschaft, D-01062 Dresden, M AN U Germany University of Technology – National University Ho Chi Minh City, Viet Nam Institute of Chemistry, Vietnamese Academy of Science and Technology, Hanoi, Vietnam University of Applied Sciences Osnabrück, D-49076 Osnabrück, Germany Polymer Service Merseburg, D-06217 Merseburg, Germany AC C EP TE D * Corresponding author Tel.:+493461462741, Fax:+493461463891 E-mail address: le-haihong@ipfdd.de (H.H Le) ACCEPTED MANUSCRIPT ABSTRACT The formation and stability of carbon nanotube (CNT) network in natural rubber (NR), deproteinized NR (DPNR) and polyisoprene (IR) compound were investigated by means of RI PT the method of the online measured electrical conductance in a whole process from processing and rolling over pressing/cross-linking to post-stretching The kinetics of CNT flocculation was described and explained by taking into consideration the depletion force considered as SC driving force and the thickness of the bound rubber layer considered as hindering factor The presence of linked phospholipids in NR and DPNR improves the rubber-filler interaction of M AN U CNTs through the cation-π bonding that hinders the filler flocculation The absence of the cation-π bonding in CNT/IR compound and the related thin layer of bound rubber are the reason for the strong tendency of flocculation of CNTs in IR even at room temperature The effect of pressing time, temperature and cross-linking reaction as well as mechanical TE D deformation on the formation and stability of CNT network in NR compounds was also AC C Introduction EP investigated and discussed by taking into consideration the role of the linked phospholipids Rubber nanocomposites containing carbon nanotubes (CNTs) have been widely researched in order to impart tire treat compounds some functionalities like gas barrier, flammability resistance and electrical conductivity [1-6] It is well-known that for a given polymer system, structural change of filler network plays a crucial role in determining the dynamic and mechanical properties of the filled rubber compounds [7] The breakdown and reformation of the filler network during cyclic strain would cause additional energy dissipation, i.e higher fuel consumption of tire treat compounds It has been demonstrated that when filler is well dispersed in the polymer, the aggregates tend to re-agglomerate during storage and ACCEPTED MANUSCRIPT vulcanization of the uncured compound, especially at high filler loading [8] According to the theory proposed by Wang et al [8], the filler network formation in the polymer matrix is mainly determined by the attractive force between filler aggregates and the interaction between polymer molecules, as well as the interaction between filler and polymer In contrast, RI PT according to Meier and Klüppel [9] the short ranging van der Waal interactions cannot explain the attractive force acting over several nanometers Instead, the driving force for the flocculation process can be attributed to the so called depletion effect, which has long been SC known to drive phase separation in colloid-polymer mixtures Its nature is not determined by microscopic interaction such as van der Waals interactions between like objects The M AN U properties of this force are determined by the tendency of the whole system to increase its entropy as stated by the second law of thermodynamics, which leads to the force which favours clustering of objects immersed in such a suspension [10] Natural rubber (NR) is considered as one of the most important bio-based polymers TE D possessing excellent chemical and physical properties, such as outstanding elasticity and flexibility [11-15] NR was proposed to consist of two trans-1,4 isoprene units, about 1000– 3000 cis-1,4 isoprene units with α- and ω-terminal [11-13] The α-terminal group of NR was EP postulated to consist of monophosphate and diphosphate groups, which are linked with AC C phospholipids via hydrogen bonding as a predominant linkage On the other hand, the ωterminal of NR molecule was postulated to be a modified dimethylallyl group linking to a functional group, which is associated with the proteins Effect of non-rubber components in NR like proteins and phospholipids on storage hardening and gel formation of NR during accelerated storage under various conditions was investigated by Yunyongwattanakorn et al [16] Effect of non-rubber components on vulcanization kinetics of NR was studied by Wang et al [17] and Smitthipong et al [18] The results showed that the rate of scotching period and curing period of NR are greater than that of natural rubber extracted with acetone, and the activation energies of both periods of NR are lower than those of extracted NR Dierkes et al ACCEPTED MANUSCRIPT [19] found the influence of non-rubber components on the properties of silica filled NR compounds The presence of a small quantity of proteins makes the silane less efficient for improving dispersion and filler polymer coupling, and thus negatively influences the final properties of the rubber material Recently, the selective wetting and localization of CNTs in RI PT binary SBR/NR blends and ternary styrene butadiene rubber (SBR)/nitrile butadiene rubber (NBR)/NR blends were investigated in our works [20,21] It was found that NR molecules dominantly wet CNT surface, while SBR nearly has no contact with CNTs The non-polar NR SC is able to compete with the polar NBR with respect to the CNT wetting A possible explanation could be related to the effect of the non-rubber components presented in NR, M AN U which are believed to enhance the interaction between NR and CNTs [21] In the present work we demonstrated the impact of the non-rubber components on the formation and stability of the CNT network in NR In particular, the CNT flocculation and network destruction during mixing, milling and pressing/cross-linking as well as post- TE D stretching was characterized by means of the method of the online measured electrical conductance The comparison between CNT/NR, CNT/DPNR and CNT/IR compounds with respect to their electrical conductivity addressed the role of the non-rubber components in EP formation and stability of filler network in natural rubber The Wang-theory and the theory of depletion force were taken into account in order to explain the experimental results received AC C in the present work Materials and experimental 2.1 Materials and mixture preparation Natural rubber (NR) (Standard Malaysian Rubber (SMR 10), Weber & Schaer GmbH) with a nitrogen content of 6.0 wt% and deproteinized natural rubber (DPNR) (Pureprena, Malaysian Rubber Board) with a nitrogen content of 0.08 wt% as well as synthetic polyisoprene (IR) ACCEPTED MANUSCRIPT (Cariflex JR 309, Shell Chemical Co.) with 95 wt% cis-1,4 isoprene units were used as rubber matrix NR and DPNR were masticated by means of a two-roll mill in order to reduce their Mooney viscosity to the range of IR Multi-walled carbon nanotubes (CNTs) (NanocylTM NC7000, Nanocyl S.A., Belgium) were used as filler According to the provider Nanocyl RI PT possesses an average radius rCNT of nm and a broad length distribution with several nanotubes up to 10 µm The amorphous carbon content is about wt% and an impurity of 10 wt% is detected as metal oxide SC For preparation of CNT filled compounds and blends an internal mixer (Rheocord 300p, ThermoHaake) was used by keeping the following mixing conditions: initial chamber M AN U temperature TA of 50 °C, rotor speed of 75 rpm and fill factor of 0.68 A conductivity sensor system was installed in the chamber of the internal mixer to measure the electrical signal of the conductive mixtures Three rubber compounds and two blends of 50/50 NR/IR and 50/50 DPNR/IR were prepared with phr CNTs with and without curing additives For an effective TE D dispersion of CNTs in rubber matrix an ethanol-assisted mixing process (wet mixing) according to our previous works was applied [20,21] EP 2.2 Experimental determination of filler wetting in rubber compounds and blends For the investigation of the rubber-filler gel of the compounds and blends, for example from AC C NR and IR, 0.1 g of each raw mixture was stored for seven days in 100 ml cyclohexane at room temperature The rubber-filler gel was taken out and dried up to a constant mass The rubber content in the gel LNR and LIR as well as LB(NR/IR) as a measure for the wetting behaviour of CNT surface by NR and IR as well as NR/IR blend, respectively, is determined according to Eq [22] L= m − m1 ⋅ cCNT m2 (1) ACCEPTED MANUSCRIPT The mass m1 is corresponding to the rubber compound before extracting m2 is the mass of the rubber-filler gel, which is the sum of the undissolvable rubber part and the mass of CNTs cCNT is the mass concentration of CNTs in the single rubber mixture or binary blends For NR/IR blend the rubber layer LB(NR/IR) is the rubber part in the rubber-filler gel and consists of RI PT two contributions according to Eq LB ( NR / IR ) (t ) = LB ( NR ) (t ) + LB ( IR ) (t ) (2) SC LB(NR) and LB(IR) can be determined by means of a calibration curve For creation of the M AN U calibration curve, blends with different NR/IR ratios were prepared and investigated by FTIR according to the procedure described in our previous work [22] FTIR spectral were recorded by use of a FTIR spectrometer S2000 (Perkin Elmer) equipped with a diamond single Golden Gate ATR cell (Specac) The peaks at 1376 cm-1 and 888 cm-1 were taken for calculation of the ratio of the surface under peak ANR/AIR The correlation between the ANR/AIR ratio and the TE D given NR/IR mass ratio can be established, and based on this correlation the ratio LB(NR)/LB(IR) of the rubber-filler gel can be determined experimentally EP The selective wetting of CNTs in NR/IR blend can be calculated according to the AC C following equations: S B ( NR ) (t ) LP LB ( NR ) (t ) = ⋅ S B ( IR ) (t ) LP NR LB ( IR ) (t ) (3) S B = S B ( NR ) (t ) + S B ( IR ) (t ) (4) IR SB(NR) and SB(IR) are the CNT surface fractions wetted by NR and IR component of blend, respectively t is the mixing time SB is the total filler surface fraction wetted in blend LPNR, LPIR and LPB(NR/IR) are the saturated rubber content in the gel of the single compounds and ACCEPTED MANUSCRIPT blend, respectively They can be determined from extraction experiments of the samples at a long mixing time of about 30 The average thickness d of the rubber layer bonded to the CNT surface can be calculated d= m2 − m1 ⋅ cCNT ρ R m1cCNT sCNT (5) SC ρR is rubber density and sCNT is the specific surface of CNTs RI PT using Eq M AN U 2.3 Characterization Optical microscopy - Optical microscopy was used to characterize the CNT macrodispersion The ratio of the surface of non-dispersed agglomerates to that of the image, A/A0, is a measure TE D for the filler macrodispersion Transmission electron microscopy (TEM) - Ultrathin sections with approximately 50 nm thickness cut from compression-molded plates with a diamond knife (35° cut angle, EP DIATOME, Switzerland) at -120 °C using a cryo-ultramicrotome RMC PowerTome PT-PC with CRX cryo-chamber (RMC, Tucson) were used for transmission electron microscopy AC C (TEM) The sections were collected on carbon coated copper grids The specimens were investigated by means of a LEO 912 Omega EFTEM at an accelerating voltage of 120 kV using the zero loss mode Offline electrical conductivity - Measurement of electrical conductivity was carried out at room temperature by means of a multimeter 2750 (Fa Keithley) for high conductive samples and an electrometer 6517A (Fa Keithley) for low conductive ones The shape of the test specimen was a rectangular strip whose ends were coated by silver paste in order to receive a ACCEPTED MANUSCRIPT good contact with the electrodes Conductance was measured for unstrained samples and deformed ones, whose conductance was simultaneously measured during the tensile test Online measurement of the electrical conductance during the cross-linking process in the RI PT compression mould - The setup for electrical measurement during vulcanization was described in our previous work [23] The cavity positioned in the middle of a PEEK plate has a diameter of 45 mm and a thickness of mm The electrodes were made from two aluminum SC foils and insulated from its surroundings by two PTFE foils A multimeter 2750 (Keithley) was used for measurement of conductance of the composites A pressure of bar was kept M AN U constant at 130 °C for the whole pressing process Determination of surface energies - Sessile drop contact angle measurements on a sheet of uncured rubbers were conducted with the automatic contact angle meter OCA 40 Micro, TE D DataPhysics Instruments GmbH (Filderstadt, Germany) The surface energies were calculated from the results of these wetting experiments For this purpose a set of test liquids with different surface tension was used Surface energy calculations were performed by fitting the EP Fowkes equation [24] AC C Tensile test – Stress-strain testing in combination with electrical conductivity measurement was performed according to ISO 37 using a tensile tester Z005 (Zwick/Roell) with a crosshead speed of 200 mm min-1 at room temperature The test specimens had a thickness of mm and an initial gauge length of 50 mm All data presented are the average of five measured specimens for each sample Temperature scanning stress relaxation experiment (TSSR) - The thermoelastic behavior was tested with a computer driven testing device of the TSSR meter type made by Brabender [25] ACCEPTED MANUSCRIPT The difference in rubber-filler interaction causes the different thickness of the rubber layer bound to the filler surface The layer thickness of the bound rubber was determined by means of Eq Setting m1CNT/NR = 0.1 g, m2CNT/NR = 0.0376 g and m1CNT/IR = 0.1 g, m2CNT/NR = 0.0106 RI PT g, which were determined from extraction experiment, and ρR = 0.94 g/cm3, cCNT = 4.65 wt%, sCNT = 280 m2/g into Eq the thickness d of the rubber layer bound to the CNT surface can be determined An average layer thickness dCNT/NR = 26.8 nm for CNT/NR compound and dCNT/IR SC = 4.8 nm for CNT/IR compound were determined that emphasizes the better rubber-filler M AN U interaction of CNTs to NR compared to IR 3.5 Kinetics of filler flocculation: a correlation between the depletion forces and the rubberfiller interaction TE D Upon milling at room temperature CNTs are separated from each other and each CNT is surrounded by a bound rubber layer with a thickness d, i.e the distance between two nearest CNTs should be δCNT/NR = 2×dCNT/NR = 53.6 nm for CNT/NR compound and δCNT/IR = 2×dCNT/IR EP = 9.6 nm for CNT/IR compound Assuming that the end of CNTs and the bound rubber layer form a sphere with a radius RCNT/NR=rCNT + dCNT/NR for CNT/NR compound and RCNT/IR=rCNT AC C + dCNT/IR for CNT/IR compound as illustrated in Fig 5a When the spheres touch each other, the distance l becomes zero The ratio FdCNT/NR/ FdCNT/IR can be calculated using Eq 14 by taking into consideration Eq 13 FdCNT / NR R CNT / NR + rr = FdCNT / IR R CNT / IR + rr rCNT + d CNT / NR + rr = CNT / IR + rr rCNT + d (14) ACCEPTED MANUSCRIPT Setting the CNT radius rCNT = nm, gyration radius rr = 20 nm [35] as well as rubber layer thickness dCNT/NR= 26.8 nm and dCNT/IR = 4.8 nm into Eq 14 we obtain FdCNT/NR/ FdCNT/IR = 3.0 It is clear from Eq 14 that the extent of the depletion force Fd is determined by the rubber layer thickness d and thus by the rubber-filler interaction The depletion force between two RI PT ends of CNTs is higher for CNT/NR compound Under influence of attractive depletion force Fd two CNTs approach to each other, i.e the inter-particle gap δ is decreased and the in- between rubber layers are squeezed out that causes an entropic repulsive force Fr The inter- AC C EP TE D M AN U SC particle gap δ at the stationary state is determined by the balance between Fd and Fr Fig Inter-particle gap δ between two CNTs determined by the balance of attractive depletion force Fd and repulsive force Fr in the (a) fast and (b) slow process According to Fig 2a, upon rolling at room temperature CNT/NR compound shows a conductance of about 1×10-9 mS at 5000 sec corresponding to an inter-particle gap δCNT/NR = ACCEPTED MANUSCRIPT 6.5 nm That means the strong depletion force in CNT/NR compound causes a large gap reduction of CNTs from 53.6 nm to 6.5 nm However, the gap of 6.5 nm is still too large for an effective transport of electrons and consequently the conductance remains in low level for CNT/NR compound as seen in Fig 2a In contrast, the lower depletion force in CNT/IR RI PT compound reduces the narrow gap from 9.6 nm to 3.4 nm, which is a typical cluster-cluster distance of a conductive filler network [28] Thus, the conductance of CNT/IR compound with thinner bound rubber layer increases strongly even at room temperature During pressing SC at 130 °C as shown in Fig 2b the entropic force Fd increases and the bound rubber layer becomes softer As a result the gap reduces from 6.5 nm to 3.5 nm for CNT/NR compound M AN U and from 3.4 nm to 3.0 nm for CNT/IR compound that causes an increase of conductance, strongly in CNT/NR and slightly in CNT/IR compound 3.6 Effect of cross-linking reaction on filler network formation TE D The effect of cross-linking reaction on the electrical conductance and related filler network formation has been reported in literature [23,36,37], however, the results seem not to be generalized Gesprache et al [36] stated that the conductivity of CB filled rubber compounds EP increases during cross-linking process strongly up to a cross-linking time t70 and slowly from t70 to t90 Zhang et al [37] found that the conductivity of silicone rubber loaded with CB is AC C improved with increasing cross-linking density, when CB loading is lower than the percolation threshold At a CB loading above the percolation threshold, conductivity decreases first and then increases with cross-linking density In our previous work [23] the conductivity of CB filled ethylene octane copolymer decreased, when the cross-linking reaction occurred It was discussed that the de-agglomeration of CB network is the main reason for the decrease of conductance during cross-linking In the present work the effect of cross-linking reaction on the electrical conductance evolution and related filler networking of CNT in rubber will be investigated for three ACCEPTED MANUSCRIPT different CNT/NR compounds The first CNT/NR compound does not contain any curing additives; the second one contains the sulphur system A with N-cyclohexylbenzothiazole-2sulfenamide (CBS) as accelerator; the third one contains the sulphur system B with CBS and additionally an amount of diphenylguanidine (DPG) as second accelerator The development RI PT of the conductance of CNT/NR compounds with and without sulphur recorded in the compression-mould at 130°C is presented in Fig 6e along with the torque separately recorded by a rheometer The conductance of the compound without curing additives increases strongly SC from 10-10 mS to about 20 mS due to the flocculation of individual CNTs as discussed above AC C EP TE D M AN U (curve 1) Fig TEM images of CNT/NR compound before (a,b) and after cross-linking at 130 °C (c,d) and electrical conductance evolution recorded in compression-mould of CNT/NR compounds without and with different curing packages (e) ACCEPTED MANUSCRIPT The conductance of the second compound with the sulphur system A (curve 2) increases similarly to that of the first compound At a pressing time of 1200 sec the conductance of the second compound reaches a plateau value Taking a look at the torque curve of the second RI PT compound it reveals that the conductance stop to increase prior to the onset of cross-linking The same result is observed for the third compound with CBS and DPG (curve 3) With two accelerators the third compound shows the earlier onset of the cross-linking reaction at 800 SC sec At this time the conductance stops to increase From this result, it is obvious that the fast process of flocculation of the investigated compounds takes place very fast and is not M AN U influenced by the cross-linking The slow process of flocculation is sensitive to the scotching process A slight reduction of mobility of rubber chains during scotching is strong enough to stop the diffusion and flocculation of filler particles in the slow process even if the crosslinking process not yet occurs TE D The network formation of CNT in rubber can be representatively visualized by TEM images of CNT/NR compounds before and after pressing process In Fig 6a and b individual tubes of CNTs of the compound before pressing are seen to be separately distributed EP throughout the matrix Upon pressing CNTs seem to re-aggregate to form a continuous network as seen in Fig 6c and d AC C The flocculation is accelerated for the first compound without sulphur, when the pressing temperature increases from 130°C to 150 °C and then 170 °C as seen in Fig 6e (curve 1) The high temperature causes flocculation of CNTs even in cross-linked sample, however, the effect is small (curve and 3) Upon cooling to room temperature the conductance decreases only slightly, that emphasizes that the huge increase of conductance taking place during heating in the compression-mould is caused by the irreversible flocculation process of filler 3.7 Destruction of CNT network in NR compound during post stretching ACCEPTED MANUSCRIPT Upon pressing of CNT/NR, CNT/DPNR and CNT/IR compound, which not contain any curing additives, the plates were taken out from the cavity for further investigation Five samples of each compound were cut from the plate as illustrated in Fig 7a The offline RI PT measurement of electrical conductance of samples was performed and shown in Fig 7b The samples 1, 2, and positioned at the edge of the plate of CNT/NR and CNT/DPNR compound show clearly different values, which are several orders of magnitude lower than EP TE D M AN U SC that of sample placed at the centre AC C Fig Position of samples cut from the compression-moulded plate of CNT/NR, CNT/DPNR and CNT/IR compound (a) and electrical conductance of samples cut from different positions (b) In contrast, all the samples of CNT/IR compound present similar values of conductance Thus, it appears that during pressing at high temperature a weakly bonded CNT network develops in CNT/NR and CNT/DPNR compound, which is easily destroyed by taking out from the cavity The samples located at the edge of the plate are deformed much more than that located at the centre The network structure of CNTs in IR is stable, thus, the conductance ACCEPTED MANUSCRIPT remains unchanged at high level In order to make clear the effect of deformation on the stability of CNT network, the relative conductance with respect to the conductance of undeformed sample was measured during stretching of samples cut from the centre of M AN U SC RI PT uncross-linked CNT/NR and CNT/IR compound and is shown in Fig 8a Fig Relative conductance of uncross-linked CNT/NR and CNT/IR compound (a) and cross- TE D linked CNT/NR compound (b) in dependence on strain The results show that the uncross-linked CNT/NR compound has a high sensitivity of EP conductance-strain response, while the conductance of uncross-linked CNT/IR compound nearly remains unchanged with deformation The change in electrical conductivity due to AC C mechanical deformation of CNT/polymer composites can be rationalized in term of changes in the CNT network configuration, i.e enlargement of cluster-cluster gap [38-40] The sensitive conductivity-deformation response of CNT/NR compound compared to that of CNT/IR compound is surely related to the better rubber-filler interaction, which enables an effective force transfer from the matrix to filler leading to a strong destruction of CNT network Fig 8b represents the effect of cross-linking on the conductance-deformation response of CNT/NR compound Upon cross-linking the mobility of rubber chains is reduced and the CNTs are blocked in position and as a result the filler network structure is stabilized ACCEPTED MANUSCRIPT That is why cross-linked CNT/NR compound shows an improvement in stability of electrical properties against deformation as seen in Fig 8b The temperature dependent stability of the CNT network in cross-linked CNT/NR and CNT/IR compound can well be characterized by the temperature scanning stress relaxation RI PT experiment (TSSR) The relative stress S(T)/S(T=23 °C) of both cross-linked compounds changes differently with increasing temperature because of different thermally induced structural processes as seen in Fig 9a The related relaxation spectral H(T), which is SC determined by Eq 15 [41,42] and shown in Fig 9b clearly presents different temperature dS (T ) H (T ) = −T dT β = T − 23°C =const t M AN U dependent processes (15) According to Vennemann et al [41,42] the slight increase of stress in the range up to 40 °C is caused by the entropy elastic forces of the rubber chain The processes taken place between TE D 40 °C and 105 °C are attributed to the debonding process of the bound rubber from the filler surface The scission of rubber chains is the main process leading to the strong decrease of the stress in the temperature range above 105 °C In the present work, it is obvious that the stress EP of CNT/NR compound decreases in the temperature range between 40 °C and 105 °C, while it AC C changes insignificantly for CNT/IR compound Because the CNT-NR interaction is much stronger than the CNT-IR interaction, the stronger decrease of the stress of CNT/NR compound cannot be attributed to the debonding process of NR from the CNT surface EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C Fig Relative stress S(T)/S(T=23 °C) and relaxation spectrum H (T) of an cross-linked CNT/NR and CNT/IR compound According to our previous work [43] the stress of the CB filled rubber compound obtained from an isothermal stress relaxation experiment decays strongly at about 90 °C The interparticle gaps between filler aggregates are smaller than approximately 3.0 nm The glassy thin polymer-layer in these gaps can keep the filler network stable against the impact of applied stresses The decay of stress may be related to the softening of the thin polymer layers ACCEPTED MANUSCRIPT connecting two CB aggregates, when the test temperature exceeds the glass temperature TG of the immobilized layer The thin polymer layer leaves its glassy state and becomes softened that makes the CB network looser and more instable against the mechanical impact In such state, the elastic stress induced by the CB network decays rapidly In this work, the thinner RI PT inter-particle gaps of CNT/IR compound leads to the higher TG of the immobilized layer und thus, to a more stable network of CNTs compared to that of CNT/NR compound SC Conclusions On the basic of experimental results it could be concluded that the extent of the filler M AN U flocculation in filler rubber compounds is determined by the balance between the attractive force, i.e depletion force, and the repulsive force, which is dependent on the thickness of the rubber layer bound to the filler surface The effect of the non-rubber component identified as linked phospholipid, on the formation and stability of the CNT network in natural rubber was TE D found and explained by the thickness of bound rubber layer The cation-π bonding of phospholipids of NR and DPNR with CNT surface improves the rubber-filler interaction and thus causes a thick bound rubber layer, while it is absent in CNT/IR compound That is the EP reason for the strong tendency of flocculation of CNTs in IR even at room temperature The AC C network formation of CNTs in NR takes place strongly during the first stage of pressing process at high temperature It found to be stopped already during the scotching prior to the cross-linking reaction The good rubber-filler interaction in uncross-linked CNT/NR compounds causes the instability of CNT network under mechanical testing, which can be stabilized through cross-linking process ACCEPTED MANUSCRIPT Acknowledgements The authors wish to thank the Deutsche Forschungsgemeinschaft (DFG) (Project Nr LE 3202/1-1) and Vietnam National Foundation for Science and Technology Development RI PT (Nafosted) (Grant number 104.02-2014.90) for the financial support Keywords: electrical conductivity, carbon nanotubes, filler network, nanocomposites, rubber Bokobza L, Rahmani M, Belin C, Bruneel J-L, El Bounia N-E Blends of carbon blacks M AN U [1] SC References and multiwall carbon nanotubes as reinforcing fillers for hydrocarbon rubbers J Appl Polym Sci 2008;46:1939-1951 [2] Bokobza L Multiwall carbon nanotube elastomeric composites: A review Polymer 2007;48:4907-4920 [3] Park SM, Lim YW, Kim CH, Kim DJ, Moon WJ, Kim JH, Lee JS, Hong CK, Seo G TE D Effect of carbon nanotubes with different lengths on mechanical and electrical properties of silica-filled styrene butadiene rubber compounds J Ind Eng Chem 2013;19:712-719 [4] Rattanasom N, Prasertsri S Relationship among mechanical properties, heat ageing EP resistance, cut growth behaviour and morphology in natural rubber: Partial replacement of clay with various types of carbon black at similar hardness level Polym Test [5] AC C 2003;28:270-276 Zhou XW, Zhu YF, Liang J Preparation and properties of powder styrene-butadiene rubber composites filled with carbon black and carbon nanotubes Mater Res Bull 2007;42:456-464 [6] Liu Y, Li L, Wang Q, Zhang X Fracture properties of natural rubber filled with hybrid carbon black/nanoclay J Polym Res 2011;18:859-867 [7] Payne AR, in “Reinforcement of Elastomers”, Kraus G, Ed., Intersience Publishers, New York, 1965, Ch.3 [8] Wang MJ Effect of polymer-filler and filler-filler interactions on dynamic properties of filled vulcanizates Rubber Chem Technol 1998;71:520-589 ACCEPTED MANUSCRIPT [9] Meier JG, Klüppel M Carbon Black Networking in Elastomers Monitored by Dynamic Mechanical and Dielectric Spectroscopy Macromol Mater Eng 2008;293:12-38 [10 ] Götzelmann B, Evans R, Dietrich S Depletion forces in fluids Phys Rev 1998;E 57:6785-6800 [11] Tarachiwin L, Sakdapipanich JT, Ute K, Kitayama T, Bamba T, Fukusaka E Structural RI PT characterization of alpha-terminal group of natural rubber Decomposition of branchpoints by lipase and phosphatase treatments Biomacromolecules 2005;6:1851-1857 [12] Tarachiwin L, Sakdapipanich J, K Ute, T Kitayama, Y Tanaka, Structural characterization of alpha-terminal group of natural rubber Decomposition of branch- SC points by phospholipase and chemical treatments Biomacromolecules 2005;6:18581863 M AN U [13] Tanaka Y, Mori M, Ute K, Hatada K Structure and Biosynthesis Mechanism of Rubber from Fungi Rubber Chem Technol 1990;63:1-7 [14] Kawahara S, Kakubo T, Nishiyama N, Tanaka Y, Isono Y, Sakdapipanich JT Crystallization behavior and strength of natural rubber: Skim rubber, deproteinized natural rubber, and pale crepe J Appl Polym Sci 2000;78:1510-1516 [15] Kawahara S, Kakubo T, Suzuki M, Tanaka Z Thermal Properties and Crystallization TE D Behavior of Highly Deproteinized Natural Rubber Rubber Chem Technol 1999;72:174180 [16] Yunyongwattanakorn J, Tanaka Y, Kawahara S, Klinklai W, Sakdapipanich J Effect of Non-Rubber Components on Storage Hardening and Gel Formation of Natural Rubber EP During Accelerated Storage under Various Conditions Rubber Chem Technol 2003;76:1228-1240 AC C [17] Wang PY, Wang YZ, Zhang BL, Huang HH Effect of non-rubber substances on vulcanization kinetics of natural rubber J Appl Polym Sci 2012;126:1183-1187 [18] Smitthipong W, Tantaherdtam R, Rungsanthien K, Suwanruji P, Sriroth K, Radabutra S, Thanawan S, Vallat M, Nardin M, Chol MK Effect of Non-Rubber Components on Properties of Sulphur Cross-linked Natural Rubbers Adv Mater Res 2013;844:345-348 [19] Dierkes WK, Sarkawi SS, Noordermeer JWM The challenges of silica-silane reinforcement of natural rubber In: 28th International Conference of the Polymer Processing Society (PPS-28), 11-15 December, 2012, Pattaya, Thailand (pp - 4) [20] Le HH, Parsekar M, Ilisch S, Henning S, Das A, Stöckelhuber K-W, Beiner M, Ho CA, Adhikari R, Wießner S, Heinrich G, Radusch H-J Effect of Non-Rubber Components ACCEPTED MANUSCRIPT of NR on the Carbon Nanotube (CNT) Localization in SBR/NR Blends Macromol Mater Eng 2014;299:569-582 [21] Le HH, Sriharish MN, Henning S, Klehm J, Menzel M, Frank W, Winer S, Das A, Stưckelhuber K-W, Heinrich G, Radusch H-J Comp Sc Technol 2014;90:180-186 [22] Le HH, Ilisch S, Heidenreich D, Wutzler A, Radusch H-J Kinetics of the Phase RI PT Selective Localization of Silica in Rubber Blends Polym Comp 2010;31:1701-1711 [23] Le HH, Ali Z, Uthardt M, Ilisch S, Radusch H-J Effect of the Cross-Linking Process on the Electrical Resistivity and Shape-Memory Behavior of Cross-Linked Carbon Black Filled Ethylene-Octene Copolymer J Appl Polym Sci 2011;120:2138-2145 SC [24] Fowkes FM Additivity of intermolecular forces at interfaces I Determination of the contribution to surface and interfacial tensions of dispersion forces in various liquids J M AN U Phys Chem 1963;67:2538-2541 [25] Fuchs F Grenzen aufzeigen – anisotherme Spannungsrelaxionsmessung Kautsch Gummi Kunstst 2006;59:302–303 [26] Stöckelhuber KW, Das A, Jurk R, Heinrich G Contribution of physico-chemical properties of interfaces on dispersibility, adhesion and flocculation of filler particles in rubber Polymer 2010;51:1954-63 TE D [27] Le HH, Hoang XT, Das A, Gohs U, Stöckelhuber K-W, Boldt R, Heinrich G, Adhikari R, Radusch H-J Carbon 2012;50:4543-4556 [28] Meier JG, Klüppel M Carbon Black Networking in Elastomers Monitored by Dynamic Mechanical and Dielectric Spectroscopy Macromol Mater Eng 2008;293:12-38 EP [29] Sichel EK, Gittleman JI, Sheng P Transport properties of the composite material carbon-poly(vinyl chloride) Phys Rev B 1978;18:5712-5716 AC C [30] Simmons JG Generalized formula for the electric tunnel effect between similar electrodes separated by a thin insulating film J Appl Phys 1963;34:1793-1803 [31] Asakura S, Oosawa F On Interaction between Two Bodies Immersed in a Solution of Macromolecules J Chem Phys 1954;22:1255-1256 [32] Asakura S, Oosawa F Interaction between Particles Suspended in Solutions of Macromolecules J Polym Sci 1958;33: 183-192 [33] Fukushima T, Aida T Ionic liquids for soft functional materials with carbon nanotubes Chem Eur J 2007;13:5048-5058 [34] Ma JC, Dougherty DA The cation-π-interaction Chem Rev 1997;97:1303-1324 ACCEPTED MANUSCRIPT [35] Fetters LJ, Hadjichristidis N, Lindner JS, Mays JW Molecular weight dependence of hydrodynamic and thermodynamic properties of well-defined linear polymers in solution J Phys Chem Ref Data1994;23:519-640 [36] Gersprache M, Nikiel L, Yang HH, O´Farell CP, Schwartz GA, Flocculation in carbon black filled rubber compounds, Rubber Division, American Chemical Society , RI PT Cleveland Ohio, Oct 16-19 2001, Paper No 20 [37] Zhang J, Feng S Effect of Crosslinking on the Conductivity of Conductive Silicone Rubber J Appl Polym Sci 2003;89:3471-3475 [38] Zeng Y, Liu H, Chen J, Ge H Effect of strain on the electrical resistance of carbon SC nanotube/silicone rubber composites J Wuhan University of Technology, Mater Sci Ed 2011;26:812-816 M AN U [39] Oliva-Avilésa AI, Avilésb F, Seidelc GD, Sosaa V On the contribution of carbon nanotube deformation to piezoresistivity of carbon nanotube/polymer composites Comp Part B: Eng 2013;47:200-206 [40] Wang L, Xu C, Li Y Piezoresistive response to changes in contributive tunneling film network of carbon nanotube/silicone rubber composite under multi-load/unload Sensors and Actuators A: Physical 2013;189:45-54 TE D [41] Vennemann N, Wu M, Heinz M Thermoelastic properties and relaxation behavior of SSBR/silica vulcanizates Rubber World 2012;246:18-23 [42] Vennemann N, Wu M, Heinz M Experimental investigations and development of a model for the description of the thermoelastic properties of carbon black filled SBR- EP Vulcanizates Kautsch Gummi Kunstst 2011;64:40-46 [43] Le HH, Heidenreich D, Kolesov IS, Ilisch S, Radusch H-J Effect of Carbon Black AC C Addition and Its Phase Selective Distribution on the Stress Relaxation Behavior of Filled Thermoplastic Vulcanizates J Appl Polym Sci 2010;117:2622–2634 ACCEPTED MANUSCRIPT Highlights Description of the kinetics of CNT network formation in NR, deproteinized NR and isoprene (IR) CNT network formation determined by the depletion forces and rubber layer bound to CNT surface Effect of linked phospholipids on CNT wetting and network formation in rubber compounds EP TE D M AN U SC RI PT Effect of linked phospholipids on CNT network stability AC C • • • • ... CNTs in IR even at room temperature The effect of pressing time, temperature and cross-linking reaction as well as mechanical TE D deformation on the formation and stability of CNT network in NR... demonstrated the impact of the non-rubber components on the formation and stability of the CNT network in NR In particular, the CNT flocculation and network destruction during mixing, milling and. .. role of the non-rubber components in EP formation and stability of filler network in natural rubber The Wang-theory and the theory of depletion force were taken into account in order to explain