J. MATER. CHEM., 1994, 4(8), 1189-1193 Interfacial Chemistry and Mechanical Effects of a Multifunctional Processing Additive on Carbon Black Filled Rubber Robert H. Bradley,* Enshan Sheng, Ian Sutherland, Philip K. Freakley and Hanafi Ismail+ University of Technology, Loughborough, Leicestershire, UK LE?I 3TU The interfacial effects of a multifunctional additive (MFA), the n-tallow-propane-l,3-diamine salt of carboxylic acid, on carbon black filled rubber have been studied. The surfaces of three normal-cure-rate carbon blacks, of differing nitrogen BET surface areas, were characterised by X-ray photoelectron spectroscopy (XPS) and vapour-phase chemical derrvatis- ation and found to contain very few functional groups. The MFA has been found to decompose at ca.120 "C and the decomposition to generate diamine and a carboxylic acid species. Bound rubber, determined by o-xylene extraction, was found to decrease with the addition of MFA and a limiting bound rubber value was obtained at the MFA loading that corresponds to a monolayer coverage of the carbon black surface. The reduction of bound rubber with the addition of MFA is attributed to the release of rubber, immobilised within carbon black agglomerates, as a result of improved dispersion. The mechanical properties of the rubber were also found to improve with the addition of MFA and again this was attributed to the dispersing effect of the MFA on the carbon black. Optimum mechanical properties were observed at an MFA loading which approximately corresponds to a monolayer coverage of the carbon black surface.It has been shown that the n-tallow propane-1,3-diamine salt of a carboxylic acid having the general formula [RNH;(CH2),NH:][R'C0,1, enhances the mechanical properties of carbon black filled natural and synthetic rub- ber~.'~~This surfactant has been termed a multifunctional additive (MFA) since it is observed that it can function as a processing aid, a filler dispersant, a cure accelerator and a mould-release agent. To use this MFA effectively in the rubber industry, it is important to understand the resultant interfacial effects between the filler and the elastomer and also the effects of the MFA on the bulk mechanical properties of the rubber. In this study XPS and FTTR were used to characterise the surfaces of a series of carbon blacks and the MFA itself.The blacks chosen have differing specific surface areas, measured by nitrogen BET, as shown in Table 1; this allows the amount of carbon surface available for interface formation to be varied. Bound rubber has also been assessed since this gives a measure of the degree of interaction between the carbon black surface and the rubber and how this interaction is modified by the presence of the MFA. In this study bound rubber is regarded as the rubber component of an unvulcan- ised mix, which is strongly associated with the filler surface and which cannot be removed by a specific period of solvent extraction. This bound rubber may include rubber that is strongly adsorbed onto the black surfaces and that is physi- cally trapped within the voids of the carbon black agglomer- ates; the latter material is sometimes termed 'immobilised rubber'.A schematic diagram showing the terminology used in this work is given in Fig. 1. o-Xylene, which is a good solvent for natural rubber, was used as the solvent for bound t On study leave from Universiti Sains Malaysia, 11800 Minden, Penang, Malaysia. Table 1 BET specific surface areas (by N, adsorption) and XPS analysis results of several carbon blacks carbon black N115 N330 N550 atom% surface area/m2 g-' C 0 S 139.5 98.9 0.6 0.5 78.0 98.6 0.9 0.5 38.9 98.5 0.9 0.6 carbon black agglomerate im'mobilised rubber Fig. 1 Schematic representation of immobilised rubber within a carbon black agglomerate.The agglomerate is composed of loosely bound aggregates of fused carbon black primary particles rubber determination in this study. The effects of the MFA on the mechanical properties of vulcanised rubber were also investigated. Experimenta1 Materials Table 2 lists the compounding materials, their manufacfurers and the levels used. Various levels of the MFA, expresxed as parts per hundred parts of rubber by weight (pphr), were incorporated to study the concentration effects. Sulfur is added as a cross-linking, i.e. curing, agent. Zinc oxidt: and stearic acid act together as a cure activator, CBS (N-cyclohex- ylbenzothiazole-2-sulfenamide) is a cure accelerator and Flectol-H ( 1,2-dihydro-2,2,4-trimethylquinoline)is an an tioxi- Table 2 Formulation used in rubber mixing material manufacturer formulation ( pphr) natural rubber (SMR20) Malaysia carbon blacks Cabot zinc oxide Union Minere Oxide (Belgium) stearic acid Caldic UK Ltd.sulfur Schill & Seilscher CBS Monsanto Flectol H Monsanto MFA, EN444 Akzo 100.0 50.0 5.0 3.0 2.5 0.5 1.o 0.0-5.( 1 1I90 dant. Mixing was carried out using a Francis Shaw K1 Intermix. After it had been mixed, each batch was sheeted off to ca. 3 mm thickness, which is suitable for subsequent preparation of mechanical test specimens. Carbon black BET specific surface areas were determined by nitrogen adsorption at 77 K using a Micromeritics automated ASAP2000 multi- point BET instrument.Analytical-grade o-xylene, for bound rubber measurement, was supplied by Fisons Chemicals. Vapour-phase derivatisation reagents, i.e. trifluoroacetic anhydride (99 + YO),trifluoro-ethanol (99 +%), pyridine (99+YO)and 1,3-di-tert-butyl-carbodiimide (99%) were all obtained from Aldrich. XPS, FTIR and SEM XP spectra were recorded on a VG ESCALAB MK1 spec- trometer. Surface compositions were calculated from elemental peak areas, after subtraction of a Shirley-type background, using Scofield photoelectron cross- section^.^ Correction was made for the inelastic mean-free path of the photoelectrons in the solid,6 for transmission of the energy analyser7 and for the angular asymmetry of the photoemissions.' Chemical derivatisation of hydroxy and carboxylic acid groups was performed at controlled vapour pressures of trifluoroacetic anhydride and trifluoroethanol respectively, in a vacuum frameg according to the following: 00 0 n 0 Derivatisation of hydroxy groups was allowed to proceed for 2 h.Carboxylic acid groups were derivatised for 12 h in the presence of pyridine and 1,3-di-tert-butylcarbodiimidesince previous work had established that no further increase in surface fluorine concentration with reaction time occurred after 10 h using poly(acry1ic acid) as a model surface." A Nicolet 20 DXC FTIR spectrometer was used to record IR spectra. The MFA was mixed with KBr and pressed into a disk. IR spectra were recorded at different temperatures, which were controlled using an in situ temperature cell.Scanning electron microscopy (SEM), used to examine the carbon black dispersion, was carried out using a Cambridge Stereoscan 360 instrument. Measurement of Bound Rubber The bound rubber was measured by immersing ca. 0.25 g of an unvulcanised rubber mix, diced into pieces, into 50 cm3 of o-xylene in a wide-mouthed glass bottle. The bottle was then sealed and left for 10 days at room temperature. It had previously been shown that under these conditions the meas- ured bound rubber decreased with extraction time up to ca. 10 days, after which no further decrease was observed. After this period of extraction the rubber pieces were taken out and thoroughly dried in a vacuum oven at 40 "C and thermogravi- metrically analysed using a Stanton Redcroft TG-750.For these measurements a 10mg sample was purged in nitrogen and then heated at a rate of 50°C min-l and the loss in weight was monitored with a chart recorder. Each sample was first heated in nitrogen and then air. The weight loss in nitrogen corresponds to rubber and processing aids, which are normally difficult to resolve in the pyrolysis curve, whilst the weight loss in air corresponds to carbon black. The ratio between the two weight values (ie.grams of bound rubber J. MATER. CHEM., 1994, VOL. 4 per gram of carbon black) was taken as the bound rubber in this study. Results and Discussion Surface Characterisation of Carbon Blacks Surfaces of three carbon blacks were studied.XPS results show that all of the surfaces are effectively carbon with very low levels of oxygen and sulfur as shown in Table 1. Diffuse- reflectance Fourier-transform infrared (DRIFT) spectra were also recorded, but no functional groups were detected, suggest- ing that the oxygen- and sulfur-containing species may only be present in the near-surface region of the carbon black structute since XPS has a much shallower sampling depth (ca. 50 A) than DRIFT (ca. 1pm). High-energy resolution C 1s spectra were recorded and deconvolved to remove the broad- ening effects of the A1-Ka X-ray line shape. The spectra so obtained show no evidence for chemically shifted peaks, which indicates that the near-surface concentration of the individual functional groups is below the detection limit of the experi- ment.To increase the sensitivity and to identify these individ- ual species, chemical derivatisation was used. The reactions described label the OH and C02H each with three fluorine atoms, which, since the photoionisation cross-section of F is a factor of four greater than that of C (which determines the sensitivity in the C 1s spectrum), should give a ca. 12-fold increase in sensitivity. Surface hydroxy and carboxylic acid concentrations can then be estimated from fluorine atomic levels. The results obtained are shown in Table 3. The method reveals some OH groups on the carbon black surfaces, but there appear to be very few C02H groups.Although the derivatisation technique for acid groups has been found to work on polymer surfaces," it may not be applicable to carbon black surfaces either because of the differing reactivity of the C02H when attached to the carbon surface, which comprises largely graphitic ring structures, or because the acid groups are positioned within surface pores that are too small to admit the tagging agent and allow chemisorption to occur. The lack of reactivity of carbon black surface C02H groups toward trifluoroethanol was confirmed using a sample of the N330 black, which had been oxidised using nitric acid (increasing the surface oxygen concentration from 0.9% for the pre-oxidised sample to 8.0% after oxidation). The C 1s peak envelope from this material contained a prominent peak at a shift of +4.5eV from the main C-C/C-H peak (at 284.6eV), which is consistent with the presence of C02H.However, no acid groups were detected using the derivatis- ation technique described. Thermal Stability of MFA The thermal stability of the MFA (EN444) was studied using IR with an in situ temperature cell. Spectra (between 1000 and 1900 cm-l) taken at differing temperatures, are shown in Fig. 2. They show that the MFA gives structures characteristic of a typical amine salt of carboxylic acid, and is thermally stable up to ca. 120°C. Thereafter changes in peak structure Table3 Surface OH and C0,H concentrations of several carbon blacks as determined by vapour-phase derivatisation carbon black N115 N330 N550 concentration (atom%) of surface oxygen 0.6 0.9 0.9 concentration (atom%) of surface oxygen present as OH 0.3 0.2 0.6 concentration (atom%) of surface oxygen present as C0,H 0.0 0.2 0.0 J.MATER. CHEM., 1994, VOL. 4 I\ . L----L 100°C 1900 1675 1450 1225 1000 wavenum ber/cm-' Fig. 2 IR spectra of MFA at various temperatures and intensity occur as the temperature is raised and thermal decomposition proceeds. This is characterised by the decrease of peaks at 1558 cm-' (NH;/NHl) and 1400 cm-' (COT), and the increase, at a wavenumber of 1714 cm-',of the C02H peaks. There is little change in spectra recorded at tempera- tures & 170"C, suggesting that above this temperature the detectable decomposition is effectively complete.According to the spectral data the decomposition of the MFA appears to proceed as follows: RNH,f(CH,),NH;( RC0,)2+RNH(CH2)3NH2 +2R'C02H It has been demonstrated that the diamine can act as a vulcanisation activator or accelerator, while the carboxylic acid can act as an internal flow additive and also a mould- release agent.' Bound Rubber Measurement Fig. 3 shows the effects of the MFA loading on bound rubber of N330 filled compound, determined by o-xylene extraction. -5" I \" 2 0.41 0123456 EN444 loading Fig. 3 Effect of MFA loading on bound rubber of N330 determined with o-xylene extraction at room temperature It is notable that the bound rubber decreases with increasing MFA concentration, suggesting that the MFA weakens the interfacial interaction between the carbon black and the rubber molecules. Fig.4 shows SEM micrographs of rubber mixes which contain differing levels of MFA. They show that the control mix (no MFA) contains a larger number of undispersed carbon black agglomerates, and that the d isper-sion of these improves with the addition of MFA up to 2 pphr. Further increases in MFA loading above 2 pphr do not appear to give any corresponding improvement in carbon black dispersion. Therefore it is concluded that the reduced inter- action between the carbon black and the rubber, owing to the presence of the MFA, aids the breakdown of carbon black agglomerates at the mixing stage. The decrease of bound rubber with the addition of MFA could then be explained in terms of the release of immobilised rubber situated within the carbon black agglomerates (Fig.1). The amount of this immobilised rubber, which is measured as part of bound rubber, depends not only on the structure of the carbon black but also on the efficiency of breaking down the agglomerates during mixing. Further increases in MFA loading above 2 pphr have no effect on bound rubber. The precise mechan- ism(s) of this interfacial energy modification and the mode of carbon black attrition is currently being studied in detail and the findings of this work will be reported in future communi- cations. Using the BET surface area, shown in Table 1,50pphr of N330 would have an area of 3900m2. We calculate the approximat: projected geometric area of the MFA molecule to be 293 A2 and therefore a loading of ca.1.9g (which corresponds to approximately 2pphr) is required to give a monolayer coverage of the black assuming that the molecules are adsorbed onto the black surface flat rather than end-on. Mechanical Properties The tensile strength and 300% modulus of N330 filled rubber vulcanisate are plotted against the MFA levels in Fig. 5. They show that with the addition of MFA, both properties are improved by up to 10%. This is believed to be due to the improved dispersion of carbon black. For N330 the mechan- ical properties increase with MFA loading and reach maxima at ca. 2 pphr which, as shown, corresponds to an approximate monolayer coverage of the carbon black surface.Further increases in MFA loading are not accompanied by correspond- ing increases of the carbon black dispersion. Excessive MFA addition may produce a weak boundary layer at the interface between the rubber and the carbon black, possibly by forma- tion of a multimolecular layer, or may result in modification of the bulk properties of the rubber, leading to a reduction in mechanical properties. Behaviour similar to that described above has been observed for the other two carbon blacks (N115 and N550), which are of different surface areus. In these instances the peaks in mechanical properties shift to the MFA loadings which correspond to the respective monolayer equivalents of MFA. o-Xylene extracts of unvulcanised N330 filled rubber mixes were analysed using FTIR and peaks from the MFA were found to appear as the MFA loading exceeded 2pphr and progressively increase in intensity as more MFA was added to the mix.This suggests that the excess of MFA molccules are weakly bound at the carbon black/rubber interface and/or are incorporated into the bulk rubber where they can be physically extracted by the o-xylene. Conclusions Only low levels of oxygen and sulfur were detected on the carbon black surfaces by XPS. Low levels of hydroxy groups J. MATER. CHEM., 1994, VOL. 4 Fig. 4 SEM micrographs of rubber mixes with various MFA loadings (pphr): (a) 0.0, (b)0.3, (c) 1.0, (d) 2.0, (e) 3.0. (f)5.0 are indicated by chemical derivatisation, but the method used to tag carboxylic acid functionalities appears to be suspect.The chemical inertness of the carbon black surface, probably due to the presence of large areas of basal plane, suggests that the interaction between the carbon black and the MFA is probably physical and of the dispersion type. The MFA has been found to decompose at ca. 120"C,and the decomposition is complete at ca. 170°C. Diamine and carboxylic acid are generated by the decomposition process. The diamine can subsequently function as a vulcanisation activator or acceler- ator, while the carboxylic acid can function as an internal flow additive and a mould release agent. Bound rubber, measured in o-xylene, appears to decrease with the addition of MFA to a limiting value of ca.2 pphr for N330. This corresponds to an approximate monolayer coverage for this carbon black. The decrease in bound rubber was attributed to the release of immobilised rubber located within carbon black agglomerates which are shown by SEM to be better dispersed and broken down when the MFA is present. No further improvement in dispersion was observed after a mono- layer coverage of the carbon black surface (cn. 2 pphr for N330). The mechanical properties of vulcanised N330 filled rubber were improved by incorporating the MFA in the formulation. These properties appear to be optimal at MFA levels that correspond to monolayer coverage of the blacks. These improvements in mechanical properties appear to be J. MATER.CHEM., 1994, VOL. 4 27 I I19 Fig. 5 EffecL of MFA loading on tensile strength and 300% modulus of N330 filled vulcanised rubber attributable to the improved dispersion of carbon black in the rubber matrix. Again the optimum level was found to correspond approximately to the amount needed for a mono- layer coverage of the carbon black surface. The authors acknowledge the financial assistance of SERC (ES) and the Malaysian Public Service Department/Uni\ ersiti Sains Malaysia (HI). They also thank Mr. F. Page of IPTME/LUT for operating the SEM instrument. References 1 C. Hepburn and M. S. Mahdi, Plast. Rubber Process. Appl.. 1986, 6, 247. 2 C. Hepburn and M. S. Mahdi, Plast. Rubber Process. Appl.. 1986, 6, 257. 3 C. Hepburn and M. S. Mahdi, Plast. Rubber Process. Appl.. 1986, 6, 267. 4 C. Hepburn and M. S. Mahdi, GB Patent PCT.GB84/00148, 16 May, 1984. 5 J. H. Scofield, J. Electron Spectrosc. Relat. Phenom., 1976,8. 129. 6 M. P. Seah and W. A. Dench, Surf. Interface Anal., 1979,1, .!. 7 M. P. Seah, Surf Interface Anal., 1980,2,222. 8 R. F. Reilman, A. Msezane and S. T. Manson, J. Eltaron Spectrosc. Relat. Phenom., 1976,8, 389. 9 E. Sheng, Ph.D. Thesis, Loughborough University of Tech-nology, 1992. 10 R. P. Popat and I. Sutherland, unpublished results. Paper 4/00295D; Received 17th January, 1994