Cardanol: a green substitute for aromatic oil as a plasticizer in natural rubber

Abstract

The grafting of cardanol onto natural rubber in the latex stage was carried out successfully at room temperature using cumene hydroperoxide and tetraethylene pentamine. The grafted natural rubber was characterized by FTIR, NMR and GPC. The grafting of cardanol onto natural rubber resulted in a 45.6% increment in the molecular weight without affecting the molecular weight distribution. The Taguchi method was used to optimize the grafting conditions to achieve maximum yield in terms of percent grafting and grafting efficiency. The optimal parameter combination was found to be an initiator concentration of 2 phr, a cardanol concentration of 10 phr, a reaction temperature of 35 °C and a reaction time of 10 hours. The percent grafting was found to be 8.25% and the grafting efficiency was 82.5% for the optimum parameter combination. Analysis of variance (ANOVA) method was used to evaluate the percentage contribution of the different control factors on percent grafting and grafting efficiency. Cardanol concentration was found to be the most dominant parameter on grafting efficiency, while initiator concentration was found to play the dominant role on percent grafting. The cardanol grafted natural rubber (CGNR) was found to have a higher molecular weight, a lower Mooney viscosity, a lower Wallace plasticity number and a higher cure rate as compared to the unmodified natural rubber. The physico-mechanical properties of the CGNR vulcanizates were at par with or even better than the gum natural rubber vulcanizates. The rheological characteristics exhibit a better flow behavior as compared to the unmodified natural rubber. Differential scanning calorimetry and dynamic mechanical analysis demonstrate a lowering of the glass transition temperature of the CGNR as compared to the raw natural rubber. This confirms the plasticization effect of the cardanol when grafted onto the natural rubber.

---

Introduction

Aromatic oils continue to predominate as the plasticizers and process aids in the rubber industry despite their carcinogenic effect due to the presence within them of polycyclic aromatic hydrocarbons. However, because of growing concerns about environmental and health protection related issues, there is a need to look for alternative sources of plasticizers mostly from renewable resources. Thus, it is appropriate to search for an environmental friendly substitute for these hazardous aromatic oils. Cardanol, chemically known as m-pentadecenyl phenol, obtained by the double vacuum distillation of cashew nutshell liquid (CNSL) is an agricultural renewable resource and a by-product of the cashew industry. In the last few decades, CNSL has been found to be a very desirable substitute for commercial phenolic compounds in the resin industry, especially because of its sustainability, low cost, large availability and biodegradability.

The phenolic moiety of cardanol, along with its aliphatic side chain containing 15 carbon atoms in the meta position to the hydroxyl group, renders it amenable to a variety of chemical reactions. Moreover, the long aliphatic side chain can vary its functionality by being a saturated hydrocarbon, either a monoene, diene or triene (non-conjugated) as shown in Fig. 1. Cardanol and its derivatives have a wide range of applications in the form of brake linings, surface coatings, paints, and varnishes because of their bifunctional moiety and high chemical reactivity. Recently cardanol has been used in the polymer and rubber industries as a multifunctional additive. It has been reported that cardanol may be used along with lignin based compounds for the synthesis of polyurethanes that exhibit good thermal and mechanical properties.¹ Calo et al.² have reported a novel benzoxazine prepolymer derived from cardanol, which is employed in the synthesis of phenolic resins that exhibit good thermal properties and flame retardant characteristics as well as improved mechanical properties and greater molecular design flexibility.

Fig. 1 Structure of cardanol.

Ravichandran et al.³ have reported the synthesis of polycardanol in aqueous media, which is a non-toxic, low-leaching, non-halogenated fire retardant polymer. Bai et al.⁴ have also reported the preparation of polycardanol by a solvent free grinding polymerization technique which has a higher thermal stability with degradation commencing at 360 °C. The oxidative polymerization of cardanol using a fungal peroxidase from Coprinus cinereus (CiP), produced an oil soluble polymer, and this polymer may be used as a glossy coating material.⁵

Applications of CNSL, along with its major constituents, cardanol and their hydrogenated derivatives, have also been explored. They can act as new green larvicides,⁶ which may be considered as new alternatives to combat the spreading of dengue. Vasapollo et al.⁷ have presented an overview of the developments in olefin metathesis involving cardanol or cardanol-derivatives for the preparation of new cardanol based olefins and hybrid materials, combined with porphyrins, phthalocyanines and fullerenes. More et al.^8–10 have reported the synthesis of processable polyamides, aromatic polyazomethines and poly(amideimide)s having pendant pentadecyl chains which were synthesized from cardanol. The presence of an alkyl side chain and flexible ether linkage in the polymer backbone imparts a greater segmental mobility to the polymer and enhances its solubility. It also marginally reduces its glass transition temperature.

Since cardanol is an important natural renewable resource containing a phenolic group, it possesses fair antioxidant properties. Derivatives of cardanol have also shown equal promise as anti-oxidants in the stabilization of gasoline.¹¹ Maria et al.¹² have studied the antioxidant properties of phosphorated cardanol on the mineral oils NH10 and NH20. Lomonaco et al.¹³ have reported that in concentrations as low as 1%, CNSL and its main components, i.e. cardanol and cardol and their thiophosphorylated derivatives,¹⁴ act as antioxidants for polymethylmethacrylate with good antioxidant activities. Mazzetto et al.¹⁵ have also reported the synthesis and characterization of four phosphorylated esters derived from hydrogenated cardanol and their applications as antiwear additives for diesel and antioxidant additives for mineral oils. They have also reported on the synthesis of phosphorylated compounds¹⁶ from cardanol and proposed their application as antioxidants for biodiesel which are comparable with butylated hydroxytoluene (BHT). Rios et al.¹⁷ have explored the thermal behavior of phosphorylated derivatives of hydrogenated cardanol and butylated cardanol, and suggested the possibility of the application of these compounds as antioxidants for petrochemical products like lubricant oils.

Menon et al.¹⁸ have reported that natural rubber (NR) modified with phosphorylated cardanol as a plasticizer is superior to that obtained with di-ethyl-hexyl-phthalate as a plasticizer in terms of higher tensile properties, better flame retardancy and resistance to thermo-oxidative degradation. Also, phosphorylated cardanol has been established as an effective plasticizer for ethylene–propylene diene rubber,¹⁹ polychloroprene and polybutadiene rubber,²⁰ natural rubber/EPDM tercopolymer blends²¹ and LLDPE/EVA copolymer blends.²² Greco et al.^23,24 have also used cardanol acetate and epoxidated cardanol acetate, the esterified derivatives of cardanol, as efficient plasticizers for PVC.

Natural rubber is a well-known renewable resource and is obtained from the tree Hevea brasiliensis in the form of a milky white fluid. Because of its high unsaturation (each repeat unit contains one double bond in its structure) it is less resistant to oxidation, ozone, weathering, and various chemicals and solvents in comparison with other synthetic rubbers. Also processability for a good surface finish and dimensional stability is poor. Hence, the chemical modification of NR is essential to overcome some of its drawbacks. Graft copolymerization is one such technique used to modify natural rubber.

Menon et al.²⁵ have established that cardanol and its derivatives act as good plasticizers in NR. They have confirmed the multifunctional activity of cardanol in NR. They have also reported that cardanol and its derivatives incorporated into the rubber act as plasticizers, process aids, cure promoters, antioxidants and tackifiers.²⁶ However, the incorporation of cardanol and its phosphorylated derivative which are highly viscous, is a tedious and a time consuming process. Besides, they cause cure-retardation due to the absorption of activators and thus additional doses of ZnO are needed to compensate for the loss.²⁶ In order to overcome this problem, Vikram and Nando²⁷ have accomplished the grafting of cardanol onto the NR backbone by a solution technique. They established that cardanol and its derivatives act as better non-mineral plasticizers than the aromatic oils not only by better processability, but also by imparting better tensile properties, higher thermal stabilities, flame retardancy and age resistance properties.²⁸

NR in the latex stage would be a more viable option both technically and commercially. Hence, the present work has been focused on grafting in the latex stage. Grafting is usually conducted using free-radical emulsion polymerization, in which the initiator utilized in the polymerization is either a redox initiator or a thermal initiator.^29–31 Arayapranee et al.³² have reported that a redox initiation system consisting of organic hydroperoxide and tetraethylene pentamine (TEPA) is insensitive to oxygen and can work well at a high pH. Hence, it is preferable to carry out reactions involving NR latex with ammonia added for preservation purposes. Moreover, redox initiators can work at low temperatures, which is advantageous in terms of lower energy consumption and the prevention of thermally induced termination or a depolymerization process.³³ Kochthongrasamee et al.³¹ have reported that a cumene hydroperoxide (CHP)/tetraethylene pentamine (TEPA) initiator is the best redox system as compared to other redox initiators (i.e., TBHPO–TEPA and K₂S₂O₈–Na₂S₂O₃ initiators) for the graft copolymerization of methyl methacrylate onto NR latex, because the solubility of CHP in the oil phase is highest giving the highest grafting efficiency.

In our previous work, the grafting of cardanol onto natural rubber in the latex stage was undertaken successfully using the redox initiator system potassium persulfate/sodium thiosulfate³⁴ and the optimum conditions for grafting were found to be at a higher temperature of 65 °C. The present study focuses on the grafting of cardanol onto natural rubber in the latex stage using cumene hydroperoxide (CHP)/tetraethylene pentammine (TEPA) with the objective of lowering the temperature of the reaction. Moreover, the present paper emphasizes the optimization of the grafting parameters by an economic and viable experimental strategy based on Taguchi's parameter design.³⁵ The effect of grafting of cardanol onto the natural rubber backbone has been evaluated by studying its processability characteristics, physico-mechanical properties and the thermal characteristics of its vulcanizates as compared to natural rubber vulcanizates prior to grafting.

Experimental

Materials

Natural rubber latex (60.02% dry rubber content) was supplied in kind by The Rubber Board, Kottayam, India. Cardanol was procured from M/S Satya Cashew Chemicals Limited, Chennai, India. The initiator cumene hydroperoxide (CHP) was procured from E-Merck, India and tetraethylene pentamine (TEPA) was obtained from Sigma-Aldrich, India. Sodium dodecyl sulfate, the anionic surfactant was obtained from E-Merck, India. Zinc oxide, stearic acid, N-(1,3-dimethylbutyl)-N′-phenyl-P-phenylenediamine (6-PPD), 2-mercapto benzothiazole disulfide (MBTS), tetramethyl thiuram disulfide (TMTD) and sulfur were of the commercially available grades. Other solvents and reagents from E-Merck were used directly without further purification.

Grafting of cardanol onto natural rubber latex

The grafting was carried out by following the process described in an earlier publication of the authors.³⁴ Natural rubber latex (5 g, DRC 60.02%) was placed in a three-necked flask. Then 10 mL of 10 wt% potassium hydroxide solution and an emulsifier sodium dodecyl sulfate (1 phr) were added and the mixture was stirred. The cardanol was made into an emulsion by mixing mechanically with a 10% aqueous solution of sodium dodecyl sulfate. The cardanol emulsion thus prepared was added to the natural rubber latex mixture and stirred for at least one hour and purged with nitrogen for 15–20 minutes. Then the initiator cumene hydroperoxide was added followed by tetraethylene pentamine after an interval of 15 minutes (the ratio of CHP/TEPA was 1 : 1). The reaction was carried out at different temperatures, for different times with constant stirring at 300 rpm. After the reaction was over, the cardanol grafted natural rubber latex was coagulated and washed with distilled water several times and dried under vacuum at 70 °C till a constant weight was achieved. Then the dried coagulum was Soxhlet extracted with methanol to remove any unbound cardanol present in the rubber coagulum. Grafting parameters such as percent grafting and grafting efficiency were calculated gravimetrically using eqn (1) and (2).

Preparation of rubber vulcanizates

The formulation used for the preparation of the rubber vulcanizates is given in Table 1. All ingredients, except the curatives were mixed with rubber in a Haake Rheomix OS (Germany) mixer for four minutes at a set temperature of 120 °C and a rotor speed of 60 rpm. Then the curatives were added on a two-roll mixing mill and mixed for 4 minutes at room temperature. The cure characteristics of the rubber compounds were determined at 150 °C by a rheometer (Monsanto R100) as per ASTM D-2084-11. Then, the rubber compounds were compression molded into sheets at 150 °C using a hydraulic hot press (model David Bridge) according to their respective optimum cure times.

Table 1 Formulations for the preparation of the rubber vulcanizates

Sample code	Natural rubber (coagulated from latex)	Cardanol grafted natural rubber	ZnO	Stearic acid	6-PPD	TMTD	MBTS	Sulfur
NR	100	—	5	2	1	0.2	0.8	2.5
CGNR	—	100	5	2	1	0.2	0.8	2.5

---

Characterization methods

(a). Characterization of the raw NR and CGNR. IR spectroscopy of natural rubber (NR) and cardanol grafted natural rubber (CGNR) was carried out using a FTIR-spectrophotometer (model spectrum RX-I, PerkinElmer Life and Analytical Sciences, Massachusett, USA) in the range 700–4000 cm⁻¹. The samples were dissolved in chloroform and then a film was cast on the KBr disk. ¹H NMR spectra of natural rubber (NR) and cardanol grafted natural rubber (CGNR) were recorded on a Bruker 500 MHz NMR spectrometer using CDCl₃ as a solvent and tetramethyl silane as an internal standard.

The molecular weights of the natural rubber (NR) and cardanol grafted natural rubber (CGNR) were determined by gel permeation chromatography (GPC) (Agilient 1260 Infinity GPC instrument) using THF as an eluent at a flow rate of 1 mL min⁻¹ and narrow disperse polystyrene as a calibration standard. The polymer solutions were passed through three PLgel 10 μm MIXED-B columns (300 × 7.5 mm) connected in series, which were preceded by a PLgel 10 μm guard column (50 × 75 mm). A RI detector was used to record the signal. Before injecting the polymer solution into the GPC instrument, it was thoroughly filtered using a regenerated cellulose filter of pore size 0.2 μm. Intrinsic viscosity measurements of the rubbers were carried out using an Ostwald viscometer with benzene as the solvent at 30 °C.

The Mooney viscosity and Mooney scorch time (t₅) were determined as per ASTM D-1646-07, by using a Mooney viscometer (MK 3, Negretti Automation Aylesbury, England). The plasticity number was determined by using a Wallace rapid plastimeter MK-II.

The viscoelastic properties of the raw cardanol grafted natural rubber and the ungrafted rubber were studied using a Rubber Process Analyzer (RPA 2000, Alpha Technologies, Ohio, USA). The dynamic frequency sweep test was carried out at a strain of 6.98% (0.5° of arc) and a temperature of 100 °C in the frequency range 0.30–33 Hz. The dynamic strain sweeps of the samples were studied at a frequency of 0.33 Hz and at a temperature of 100 °C in the strain range 0.7–1256%.

The glass transition temperatures of the rubbers were measured with a differential scanning calorimeter (TA instruments, DSC Q100) in the temperature range −100 °C to +100 °C at a heating rate of 10 °C min⁻¹. Thermogravimetric analysis was carried out by using a Thermogravimetry Analyzer (TA instruments, TGA Q50) from ambient temperature to 600 °C at a heating rate of 10 °C min⁻¹.

(b). Characterization of the NR and CGNR vulcanizates. Tensile properties were measured with the help of a Hounsfield Universal testing machine (model H10KS), at a crosshead speed of 500 mm min⁻¹ as per ASTM D-412-06 (method A). The tear strengths of the specimens were determined as per ASTM D-624-00. Hardness was measured as per ASTM D-2240-05, using an indentation hardness tester (type shore A). A compression set test at constant strain was carried out according to ASTM D-395-03 (method B). The rebound resilience was measured using a Dunlop tripsometer as per BS 903: part A8. A heat build-up test of the specimens was measured using a Goodrich flexometer (Ferry machine co. Kent, Ohio, USA) at 50 °C, as per ASTM D-623-07. The abrasion resistance was determined by using a DuPont abrader as per ISO 4649:2010 (method A). The fatigue life of the rubber vulcanizates were studied with the help of a Monsanto Fatigue-to-Failure Tester (FTFT) as per ASTM D4482-11 at an extension ratio of 2.0.

The crosslink density was measured by using an equilibrium swelling method using benzene as the solvent. The Flory–Rehner equation³⁶ was employed to calculate the crosslink density as given in eqn (3).

−ln(1 − v_r) − v_r − χv_r = 2v_sη(v_r − 2v_r/f) (3)

where, v_r is the volume fraction of the rubber in the swollen sample, v_s is the molar volume of solvent, χ is the rubber–solvent interaction parameter, η is the crosslink density of the rubber (mol cm⁻³) and f is the functionality of the crosslinks (being 4 for the sulfur curing system).

Dynamic mechanical properties of the vulcanizates were determined with the help of a Dynamic Mechanical Analyzer DMA Q800 (TA Instruments, Lukens Drive, Newcastle, Delaware). The measurements were done under tension mode in the temperature range from −100 °C to +100 °C at a heating rate of 3 °C min⁻¹ with 0.1% strain and 1 Hz frequency.

(c). Taguchi method. The Taguchi method is an experimentally established technique used in the design of experiments for achieving optimum conditions to improve quality performance based on a mathematical approach.³⁵ It is highly effective in studying the effects of multiple factors on the deliverables with a minimum number of experiments. It also determines the factors which have a greater influence than others, by using an analysis of the variance technique.

The Taguchi method drastically reduces the number of experiments that are required to model the response function compared with the full factorial design of experiments. Hence, it is a technique for designing experiments to investigate processes where the output depends on many factors (variables or inputs) without resorting to studying all possible combinations of values which is a tedious and uneconomical process. Scheme 1 represents the major steps of implementing the Taguchi method.³⁵

Scheme 1 Scheme of the major steps of implementing the Taguchi method.

In the Taguchi method, the results of experiments are analyzed to achieve the following objectives: (1) to find out the best or optimal conditions for the product or process, (2) to identify the contribution of individual factors, and (3) to estimate the response under optimal conditions.³⁷

In the present work, the Taguchi method has been employed to study the effect of the four control parameters viz. initiator concentration (A), cardanol concentration (B), reaction time (C), and reaction temperature (D), each set at three different levels 1, 2 and 3 as shown in Table 2. With four factors each at three levels, the full factorial design requires 3⁴ = 81 runs or experiments to be carried out. However, with the help of the Taguchi method using the L9 orthogonal array only 9 runs/experiments need to be carried out as shown in Table 3. The response variables chosen are percent grafting and grafting efficiency.

Table 2 Control factors and levels

Code	Control factors	Levels
		1	2	3
A	Initiator concentration (phr)	1	2	3
B	Cardanol concentration (phr)	5	10	15
C	Reaction temperature (°C)	35	50	65
D	Reaction time (hours)	6	8	10

---

Table 3 Grafting conditions: Taguchi L₉ orthogonal array layout

Run no.	Initiator concentration (phr)	Cardanol concentration (phr)	Reaction temperature (°C)	Reaction time (hours)
1	1	5	35	6
2	1	10	50	8
3	1	15	65	10
4	2	5	50	10
5	2	10	65	6
6	2	15	35	8
7	3	5	65	8
8	3	10	35	10
9	3	15	50	6

---

The Taguchi method employs a generic signal-to-noise (S/N) ratio which measures the effect of noise factors on performance characteristics. A larger S/N ratio represents better quality characteristics and less variation. There are primarily three categories of S/N ratios: “smaller-is-better”, “larger-is-better” and “nominal-is-best”. The selection principles of the S/N ratio depend on the goal of the design.

In the present study, since both the response variables, the percent grafting and the grafting efficiency are intended to be maximized, hence the “larger-is-better” approach is adopted for which the S/N ratio is calculated as follows:³⁵

where, y_i is the observed data and n is the number of observations.

Results and discussion

IR spectroscopy study

The FTIR spectrum of natural rubber (Fig. 2a) shows very important absorption peaks, at 2960–2854 cm⁻¹ due to aliphatic C–H stretching, at 1634 cm⁻¹ due to aliphatic C=C stretching, at 1448 and 1375 cm⁻¹ peaks due to the C–H bending vibration, at 1260 cm⁻¹ for C–C stretching and at 801 cm⁻¹ for the =C–H bending vibration. However, the FTIR spectrum of cardanol grafted natural rubber (Fig. 2b) shows an additional peak at 3446 cm⁻¹ which has been attributed to the –OH stretching vibration of the phenolic moieties present in the cardanol. This infers that the double bonds present in the side chain of cardanol have taken part in the grafting reaction leaving the phenolic moiety intact. This observation is similar to the previously reported one of the authors when using potassium persulfate/sodium thiosulfate as the initiator system³⁴ for the grafting reactions with NR.

Fig. 2 IR spectrum of (a) NR and (b) CGNR.

NMR spectroscopy study

The ¹H NMR spectrum of NR (Fig. 3a) shows a singlet resonance signal at 5.16 ppm corresponding to the unsaturated methyne proton. The signal at 2.08 ppm is attributed to the methylene protons and the singlet resonance signal at 1.71 ppm may be due to the methyl protons. The spectrum of CGNR (Fig. 3b) shows signals at 1.73, 2.08 and 5.12 ppm corresponding to the NR backbone. In addition, it shows a multiplet at 6.70–7.18 corresponding to the aromatic protons due to the presence of the phenolic moiety in cardanol that is absent in the ¹H NMR spectrum of NR. Hence, it may be presumed that the unsaturation present in the cardanol must have taken part in the grafting reaction leaving the phenolic moiety intact.

Fig. 3 ¹H NMR spectrum of (a) NR and (b) CGNR.

The grafting sites on the natural rubber backbone in a free radical reaction may be generated through two possible routes (Scheme 2) viz. abstraction of α-methylenic hydrogen (route-A) or the addition to the double bond in the polyisoprene backbone (route-B) that leads to formation of polyisoprene macro-radicals.³⁸ Bulky initiator radicals such as the cumyloxyl radical (generated from cumene hydroperoxide) tend to favour abstraction of the α-methylenic hydrogen from the allylic carbon (5th carbon) as compared to the addition reaction. Moreover, addition of the polymeric radicals to the trisubstituted double bonds on the polyisoprene is slow, hence abstraction is more likely to be favoured than addition.³⁹ Among the allylic protons present in the NR backbone, the most labile one is the –CH₂– group present in the fifth position in comparison with the –CH₃ group present at the second position of the isoprene moiety. This is because of the existence of a maximum number of hyperconjugate structures for the radical which forms upon the loss of a hydrogen atom. Hence, the preferred grafting site at the main chain backbone of natural rubber is the carbon which is at the fifth position. The probable mechanism of the grafting of cardanol onto natural rubber in the latex stage and the structure of CGNR are given in Scheme 3.

Scheme 2

Scheme 3

Optimization of the reaction conditions using the Taguchi method

The response variables such as percent grafting and grafting efficiency are calculated by using eqn (1) and (2) and the results are shown in Table 4. The signal-to-noise ratio (S/N) for each series of experiments has been calculated using eqn (4) and the results are presented in Table 4. The response to each factor at its individual level has also been calculated by averaging the S/N ratios of all the experiments at each level for each factor. In order to evaluate the influence of each factor on the yield, the S/N ratio for each factor is computed. The S/N ratio for a single factor can be determined from the average values of the S/N ratios at different levels. For example, the mean S/N ratio for the reaction temperature at level 1 can be calculated by averaging the S/N ratios for the experiments 1, 6 and 8. The mean S/N ratio for every factor at different levels has also been calculated similarly. Fig. 4 and 5 represent the effect of the four control factors on the percent grafting and the grafting efficiency respectively. Monitoring the S/N ratios at the different levels of the control factors for higher values led to a conclusion that the combination A₂, B₂, C₁ and D₃ of the control factors yields the maximum percent grafting, while the maximum grafting efficiency is achieved with the combination A₂, B₁, C₁ and D₃ of the control factors.

Table 4 Response variables and results for the S/N ratios

Run no.	Percent grafting (%)	Grafting efficiency (%)	S/N ratio for percent grafting	S/N ratio for grafting efficiency
1	5.01	96.36	13.99	39.68
2	4.70	47.39	13.45	33.51
3	5.19	34.11	14.30	30.66
4	5.44	87.21	14.71	38.81
5	6.87	72.39	16.74	37.19
6	7.61	56.18	17.63	34.99
7	3.50	70.55	10.88	36.97
8	6.61	68.60	16.40	36.73
9	3.45	23.36	10.75	27.37

---

Fig. 4 Effect of control factors on percent grafting.

Fig. 5 Effect of control factors on grafting efficiency.

The present work focuses on the grafting of cardanol onto NR by using a CHP/TEPA system unlike the earlier published work³⁴ with the objective to carry out the reaction at ambient temperature which is usually 35 °C in the southern part of the country throughout the year. Both the initiating systems give similar results regarding the grafting yield and grafting efficiency. However, the optimized reaction temperature for grafting was 65 °C in the earlier published work and for the current system it is 35 °C. This will be beneficial from a commercial point of view for the grafting of NR latex on a large scale for industrial use. It will save energy and reduce production costs significantly as no heating of the latex is required in this case.

Analysis of variance (ANOVA) results

The analysis of variance (ANOVA) method has been performed to evaluate the influence of the relative control factors on the grafting reaction. From the analysis, it becomes easier to identify the effectiveness of the four control factors on the percent grafting and the grafting efficiency. ANOVA has been established based on the sum of the square (SS), the degree of freedom (D), the variance (V), and the percentage of the contribution to the total variation (P) which can be calculated as follows:^40,41

The total sum of squares SS_T can be calculated as:

where, m is the total number of the experiments and n_i is the S/N ratio at the ith test.where, SS_p denotes the sum of squares for the tested factors, p represents one of the tested factors, j the level number of this specific factor p, t the repetition of each level of the factor p, and S_ηj the sum of the S/N ratio involving this factor and level j.where, V_p is the variance of the tested factors and D_p is the degree of freedom for each factor.

SS^′_p = SS_p − D_pV_e (8)

where, P_p is the percentage of the contribution of each individual factor to the total variation.

The analysis of variance (ANOVA) for the response variables viz. percent grafting and grafting efficiency has been performed and the results are shown in Table 5 and Table 6 respectively. From Table 5, it is evident that the control factor of initiator concentration stands as the most significant factor for the percent grafting with the percentage contribution of up to 45.1% followed by the reaction temperature which contributes up to 32.8%. From Table 6, it may be inferred that cardanol concentration is the most significant factor for the grafting efficiency which contributes up to 72.2%. The percentage contributions due to all the four factors have been shown in Fig. 6. Reaction time is found to be the least significant factor for both percent grafting and grafting efficiency. The order of the effect of the control factors for percent grafting is: initiator concentration, reaction temperature, cardanol concentration, and reaction time as presented in the last column of Table 5. While as far as grafting efficiency is concerned, it is observed that cardanol concentration has the highest influence and reaction time has the minimum effect of up to only 1.4% for overall grafting efficiency. Both initiator concentration and reaction temperature however have almost nominal effect on the grafting efficiency.

Table 5 ANOVA table for percent grafting

Factors	Degree of freedom (D)	Sum of squares (SS)	Variance (V)	Corrected sum of squares (SS′)	Percentage contribution (P, %)	Rank
Initiator concentration (A)	2	7.497	3.75	7.50	45.1	1
Cardanol concentration (B)	2	3.0075	1.50	3.01	18.1	3
Reaction temperature (C)	2	5.4542	2.73	5.45	32.8	2
Reaction time (D)	2	0.6537	0.33	0.65	3.9	4
Error	0	0	—	0	—	—
Cor total	8	16.61		16.61	100.00	—

---

Table 6 ANOVA table for grafting efficiency

Factors	Degree of freedom (D)	Sum of squares (SS)	Variance (V)	Corrected sum of squares (SS′)	Percentage contribution (P, %)	Rank
Initiator concentration (A)	2	501.12	250.56	501.12	11.0	3
Cardanol concentration (B)	2	3293.3	1646.66	3293.31	72.2	1
Reaction temperature (C)	2	699.77	349.89	699.77	15.4	2
Reaction time (D)	2	64.28	32.14	64.28	1.4	4
Error	0	0	—	0	—	—
Cor total	8	4558.48	569.81	4558.48	100.00	—

---

Fig. 6 Percentage contribution of the control factors for percent grafting and grafting efficiency.

As per the ANOVA results, if the percentage error (P_e) contributes to lower than 15% of the total variance, there is no significant defect in the experimental design. On the other hand, if the percent contribution of the error exceeds 50%, then there is a significant defect in the design of the experiment and it must be re-designed. As seen from Tables 5 and 6, the percentage error (P_e) appears to be 0%. This implies that there is no significant factor that has been missed out from the experimental design and thus, the design of the experiment is correct.

Confirmatory test

A correlation between the input control factors and the yield such as percent grafting and grafting efficiency has been established by using the multiple linear regression analysis. Linear regression is performed with the help of “Minitab 15” software.

The regression equation for percent grafting is obtained as follows:

Percent grafting (%) = 8.09 − 0.022 × A + 0.107 × B − 0.0362 × C − 0.043 × D (10)

The regression equation for grafting efficiency is obtained as follows:

Grafting efficiency (%) = 143 − 1.28 × A − 4.31 × B − 0.471 × C − 1.26 × D (11)

where A = initiator concentration (phr), B = cardanol concentration (phr), C = reaction temperature (°C), D = reaction time (hours).

The final step in the Taguchi method is to confirm the experimental results. Once all the control factors have been optimized, then the experiment is performed at the optimum level of each of the control factors as shown in Table 7. The experiments were performed in duplicate and the results are recorded in Table 8 for confirmation and for comparison with the theoretical values calculated by using the regression eqn (10) and (11) for percent grafting and grafting efficiency respectively.

Table 7 Set of control factors for the confirmation tests

Test	Initiator concentration (phr)	Cardanol concentration (phr)	Reaction temperature (°C)	Reaction time (hours)
1	2	5	35	10
2	2	10	35	10

---

Table 8 Confirmation tests and their comparison with regression model

Test	Percent grafting (%)			Grafting efficiency (%)
	Experimental	Predicted	% Error	Experimental	Predicted	% Error
1	5.91	6.88	14.09	90.37	89.81	0.62
2	8.25	7.42	11.18	82.5	68.26	20.86

---

Moreover, it is observed (from Table 8) that for a cardanol concentration of 5 phr, the percent grafting is less but the grafting efficiency is more. However, when the cardanol concentration is increased to 10 phr, percent grafting increases and grafting efficiency decreases. Hence, considering both percent grafting and grafting efficiency, the optimum combination of parameters is determined as the cardanol concentration of 10 phr having all other control factors the same as shown in Table 7.

From Table 8 it is observed that the error in the calculations may vary from 11.18% to 14.09% for percent grafting and from 0.62% to 20.86% for grafting efficiency. This establishes the fact that the multiple regression equation derived from eqn (10) and (11) correlate with the percent grafting and grafting efficiency to a reasonable degree of approximation.

GPC analysis of NR and CGNR

The GPC trace before and after the grafting reaction is shown in Fig. 7. A marginal shift of the GPC trace for the CGNR towards the higher molecular weight region in comparison to NR is observed which indicates that there is an increase in the molecular weight due to the grafting of cardanol onto the natural rubber backbone. The molecular weight (M_n) for NR is found to be 2.83 × 10⁵ g mol⁻¹. The molecular weight of CGNR is found to be 4.12 × 10⁵ g mol⁻¹ which shows a 45.6% increase in the molecular weight as a result of the grafting of cardanol onto NR. The PDI value for NR is 3.14 and that of CGNR is 2.90 which indicates that the molecular weight distribution for both NR and CGNR is almost the same.

Fig. 7 GPC traces of NR and CGNR.

Processability characteristics

The processability characteristics of cardanol grafted natural rubber and the ungrafted natural rubber and the physico-mechanical properties of their vulcanizates are given in Table 9. Intrinsic viscosity practically refers to the hydrodynamic volume of the macromolecules in the solvent.⁴² The intrinsic viscosity of the grafted rubber is found to be higher than that of the unmodified rubber. This may be due to an increase in the hydrodynamic volume resulting from the grafting of cardanol onto the natural rubber backbone. The Mooney viscosity of the virgin CGNR is found to be lower than that of NR which is a clear indication of the plasticization effect of the cardanol in NR because of the long aliphatic side chain of cardanol. CGNR shows a higher plasticity than that of NR as indicated by its lower plasticity number.

Table 9 Processability characteristics and physico-mechanical properties of the vulcanizates

Properties	NR	CGNR
1. Intrinsic viscosity (dL g^−1)	0.845	0.971
2. Mooney viscosity, ML (1 + 4)@100 °C	89	78
3. Plasticity number	47	39
4. Optimum cure time @ 150 °C, t90	4 min 30 s	3 min 37 s
5. Scorch time, t2	2 min 45 s	1 min 45 s
6. Cure rate (s^−1)	0.95	1.33
7. Mooney scorch time @ 120 °C, t5	11 min 30 s	6 min 15 s
8. Tensile strength (MPa)	20.36	23.34
9. Elongation at break (%)	998	1268
10. Modulus at @100% elongation (MPa)	0.65	0.71
11. Modulus at @200% elongation (MPa)	1.14	1.14
12. Modulus at @300% elongation (MPa)	1.67	1.49
13. Tear strength (N mm^−1)	30.20	34.64
14. Hardness (shore A)	39	34
15. Compression set (%)	3.61	3.26
16. Rebound resilience (%)	72.60	76.31
17. Goodrich heat buildup, ΔT, after 25 min (°C)	3	3
18. Abrasion loss (cm^3 h^−1)	1.63	1.38
19. Monsanto fatigue-to-failure (kC)	394	450

---

Rheological measurements by rubber process analyzer

(a). Frequency sweep. Fig. 8 shows the dependence of storage modulus (G′) and complex viscosity (η*) on the angular frequency resulting from the dynamic frequency sweep measurements. The G′ value for CGNR is higher than that of NR which may be due to the higher hydrodynamic volume of CGNR resulting from the grafting of cardanol onto the natural rubber backbone. The initial complex viscosity is higher for the grafted rubber (Fig. 8b). With an increase in frequency the complex viscosity gradually decreases for both the rubbers, however, the extent of the decrease is more in the case of CGNR and at very high frequency both have almost the same complex viscosity. It may be due to the fact that at higher frequency disentanglements of the polymer chains may have taken place and the extent of disentanglement is higher in the grafted rubber as a result of the plasticizing effect of the long aliphatic side chain present in the cardanol.

Fig. 8 Dynamic frequency sweep of (a) storage modulus, G′ (b) complex viscosity, η* of the raw cardanol grafted natural rubber and ungrafted natural rubber.

(b). Strain sweep. Fig. 9 shows the dependence of storage modulus (G′) and complex viscosity (η*) on the strain amplitude in the dynamic sweep measurements. The plateau region of the dynamic strain sweep experiment where elastic modulus is independent of shear strain is termed as the linear viscoelastic region of the polymers. The magnitudes of the storage modulus, G′ of all the polymers exhibit a linear region (Newtonian plateau) at low strains and a nonlinear region (non-Newtonian plateau) at high strain amplitudes. At low strain (up to 28% strain) the storage modulus and complex viscosity for CGNR are higher than that of NR which may be due to the increase in hydrodynamic volume imparted by the grafted cardanol onto the polymer chain as explained earlier. However, with an increase in strain, there may be disentanglement of the polymer chains from the scheduled coiled and twisted state which leads to a decrease in the storage modulus and in the complex viscosity. It is seen that the extent of the decrease is more in the grafted rubber as a result of the plasticizing effect of the long aliphatic side chain of cardanol in natural rubber.

Fig. 9 Dynamic strain sweep of (a) storage modulus, G′ (b) complex viscosity, η* of the raw cardanol grafted natural rubber and ungrafted natural rubber.

Cure characteristics

The cure characteristics of the vulcanizates are listed in Table 9. It is seen that CGNR shows a higher cure rate than that of the NR vulcanizates. This may be due to the participation of the unsaturation present in the side chain of cardanol in the curing reaction.²⁶ The scorch time is found to be lower for the CGNR vulcanizates.

Crosslink density

The crosslink density for the ungrafted and the cardanol grafted natural rubber vulcanizates have been found to be 7.61 × 10⁻⁵ mol cm⁻³ and 8.36 × 10⁻⁵ mol cm⁻³ respectively. The higher crosslink density in the CGNR vulcanizates has been explained as due to the active participation of the unsaturation present in the side chain of cardanol in the vulcanization reaction. This has been amply substantiated by Menon et al.²⁶ where phosphorylated cashew nutshell liquid added physically to NR was vulcanized by sulphur and the accelerator.

Physico-mechanical properties

The physico-mechanical properties are listed in Table 9. CGNR shows a higher tensile and tear strength than that of the NR vulcanizates. This may be due to the higher crosslink density in the CGNR vulcanizates which accounts for the enhancement in the tensile properties. The hardness of CGNR is lower than that of NR. This may be due to the plasticizing effect of cardanol which makes the grafted rubber softer. CGNR exhibits a compression set of 3.26% which is 9.6% lower than that of the NR vulcanizates. This lower compression set may be because of the higher elastic behaviour of the cardanol grafted rubber which undergoes a fast recovery upon removal of the force. The CGNR vulcanizates show a better rebound resilience than that of the NR vulcanizates. However, the heat buildup is revealed to be the same in both the vulcanizates. The CGNR vulcanizates show a better abrasion resistance than that of the NR vulcanizates. The resistance to fatigue and failure is enhanced in the case of the CGNR vulcanizates. This may be due to the increase in flexibility imparted by the cardanol as an internal plasticizer.

Differential scanning calorimetry (DSC)

Fig. 10 shows the differential scanning calorimetry thermograms for the ungrafted and the cardanol grafted natural rubber. The glass transition temperature, T_g of ungrafted NR is found to be −62 °C while the T_g of CGNR is found to be −66 °C. The decrease in T_g of the grafted rubber confirms the plasticization effect of cardanol within NR inherently. It is well established that plasticizers increase the free volume due to the facilitation of molecular mobility of the polymer chains. This leads the glass transition, T_g, towards a lower temperature for the plasticized polymers.

Fig. 10 DSC curve for cardanol grafted rubber and the ungrafted natural rubber.

Thermogravimetric analysis (TGA)

The TGA thermograms of NR and CGNR are shown in Fig. 11. The thermograms overlap on each other amply suggesting that there is no change in the thermal stability of NR on grafting with cardanol.

Fig. 11 Thermogravimetry analysis of cardanol grafted natural rubber and the ungrafted natural rubber.

Dynamic mechanical analysis (DMA)

Fig. 12 shows the effect of the dynamic storage modulus (E′) and the loss factor (tan δ) as a function of temperature for NR and CGNR in the temperature range of −100 °C to +100 °C. The dynamic storage modulus (E′) of CGNR is higher than NR below the T_g. Above the T_g, the difference is very small, and the trend is also reversed.

Fig. 12 Effect of dynamic storage modulus (E′) and loss factor (tan δ) as a function of test temperature for CGNR and NR vulcanizates.

However, there is a marginal shift in the maximum loss tangent (tan δ_max) of CGNR towards the lower temperature (−46.2 °C) as compared to NR (−41.6 °C) which confirms the findings from DSC of the plasticizing effect of cardanol in NR. In general, plasticizers lower the T_g by configuring between the polymer chains, thus altering the polymer intermolecular interactions and enhancing chain mobility. The peak height is slightly increased for CGNR indicating a marginally higher damping behaviour of CGNR. This may be due to an increase in the number of crosslinks in the case of cardanol grafted natural rubber.

Conclusions

The grafting of cardanol onto natural rubber in the latex stage has been accomplished successfully by using the redox initiator system of cumene hydroperoxide and tetraethylene pentamine. FTIR, NMR and GPC results confirm the grafting of cardanol onto natural rubber backbone. The grafting results in a 45.6% increase in the molecular weight without affecting the molecular weight distribution. The Taguchi method provides a simple, systematic and efficient tool to evaluate the effect of the four different control factors on the response variables of grafting: percent grafting and grafting efficiency. The optimum combination of the parameters are found to be an initiator concentration of 2 phr, a cardanol concentration of 10 phr, a reaction temperature of 35 °C and a reaction time of 10 hours considering both percent grafting and grafting efficiency. The percent grafting is found to be 8.25% and grafting efficiency is 82.5% for the optimum parameter combination. The CHP/TEPA initiator system enables the grafting of cardanol onto natural rubber latex at a much lower temperature (ambient temperature) in comparison with the earlier method reported. This method is expected to be techno-economically feasible and suitable from a scale-up perspective.

The analysis of variance technique provides the percent contribution of the different control factors on the percent grafting and the grafting efficiency. Initiator concentration has the most dominant effect on the percent grafting to the extent of 45.1% while, cardanol concentration has the highest effect on the grafting efficiency to the extent of 72.2%. The experimental results of percent grafting and grafting efficiency are found to be in good agreement with the predicted values as derived from regression analysis. The processability characteristics of the grafted rubber studied by a rubber process analyzer exhibit higher complex viscosity and storage modulus at lower frequency and under lower strain in comparison to the ungrafted rubber. The physico-mechanical properties of the CGNR vulcanizates have an edge over those of the unmodified rubber because of the higher crosslink density. The plasticizing effect of the cardanol upon grafting onto NR has been confirmed from the DSC and DMA results. Thus cardanol grafting onto NR imparts an inherent plasticizing effect in addition to acting as a cure promoter.

Acknowledgements

Miss Sunita Mohapatra is grateful to the Council of Scientific and Industrial Research (CSIR), New Delhi, India for the award of an individual Senior Research Fellowship. The authors would like to acknowledge the Rubber Board, Kottayam, India for supplying natural rubber latex for this research work. The authors would like to acknowledge the National Chemical Laboratory, Pune, India for the NMR facility.

References

1. T. T. M. Tan, Polym. Int., 1996, 41, 13–16.
2. E. Calo, A. Maffezzoli, G. Mele, F. Martina, S. E. Mazzetto, A. Tarzia and C. Stifani, Green Chem., 2007, 9, 754–759.
3. S. Ravichandran, R. M. Bouldin, J. Kumar and R. Nagarajan, J. Cleaner Prod., 2011, 19, 454–458.
4. W. Bai, X. Xiao, Q. Chen, Y. Xu, S. Zheng and J. Lin, Prog. Org. Coat., 2012, 75, 184–189.
5. Y. H. Kim, K. Won, J. M. Kwon, H. S. Jeong, S. Y. Park, E. S. An and B. K. Song, J. Mol. Catal. B: Enzym., 2005, 34, 33–38.
6. D. Lomonaco, G. M. Pinheiro Santiago, Y. S. Ferreira, A. M. Campos Arriaga, S. E. Mazzetto, G. Mele and G. Vasapollo, Green Chem., 2009, 11, 31–33.
7. G. Vasapollo, G. Mele and R. Del Sole, Molecules, 2011, 16, 6871–6882.
8. A. S. More, S. K. Pasale and P. P. Wadgaonkar, Eur. Polym. J., 2010, 46, 557–567.
9. A. S. More, P. S. Sane, A. S. Patil and P. P. Wadgaonkar, Polym. Degrad. Stab., 2010, 95, 1727–1735.
10. A. S. More, A. S. Patil and P. P. Wadgaonkar, Polym. Degrad. Stab., 2010, 95, 837–844.
11. T. N. Castro Dantas, M. S. G. Dantas, A. A. Dantas Neto, C. V. D'Ornellas and L. R. Queiroz, Fuel, 2003, 82, 1465–1469.
12. M. Riosfacanha, S. Mazzetto, J. Beserracarioca and G. Debarros, Fuel, 2007, 86, 2416–2421.
13. D. Lomonaco, F. J. N. Maia and S. E. Mazzetto, J. Therm. Anal. Calorim., 2013, 111, 619–626.
14. D. Lomonaco, F. Cangane and S. Mazzetto, J. Therm. Anal. Calorim., 2011, 104, 1177–1183.
15. S. E. Mazzetto, L. D. M. Oliveira, D. Lomonaco and P. A. Veloso, Braz. J. Chem. Eng., 2012, 29, 519–524.
16. D. Lomonaco, F. J. N. Maia, C. S. Clemente, J. P. F. Mota, A. E. Costa and S. E. Mazzetto, Fuel, 2012, 97, 552–559.
17. M. A. de Sousa Rios and S. E. Mazzetto, Fuel Process. Technol., 2012, 96, 1–8.
18. A. R. R. Menon, C. K. S. Pillai and G. B. Nando, Eur. Polym. J., 1998, 34, 923–929.
19. A. R. R. Menon, Iran. Polym. J., 2003, 12, 305–313.
20. A. R. R. Menon and L. L. Y. Visconte, J. Appl. Polym. Sci., 2006, 102, 3195–3200.
21. A. R. R. Menon, T. A. Sonia and J. D. Sudha, J. Appl. Polym. Sci., 2006, 102, 5123–5130.
22. A. R. R. Menon and C. K. S. Pillai, Iran. Polym. J., 2002, 11, 85–91.
23. E. Calò, A. Greco and A. Maffezzoli, Polym. Degrad. Stab., 2011, 96, 784–789.
24. A. Greco, D. Brunetti, G. Renna, G. Mele and A. Maffezzoli, Polym. Degrad. Stab., 2010, 95, 2169–2174.
25. A. R. R. Menon, C. K. S. Pillai and G. B. Nando, J. Appl. Polym. Sci., 1999, 73, 813–818.
26. A. R. R. Menon, C. K. S. Pillai and G. B. Nando, J. Appl. Polym. Sci., 1994, 51, 2157–2164.
27. T. Vikram and G. B. Nando, J. Appl. Polym. Sci., 2007, 105, 1280–1288.
28. T. Vikram and G. B. Nando, Kautsch. Gummi Kunstst., 2005, 644–649.
29. M. Schneider, T. Pith and M. Lambla, J. Appl. Polym. Sci., 1996, 62, 273–290.
30. W. Arayapranee and G. L. Rempel, J. Appl. Polym. Sci., 2008, 109, 1395–1402.
31. T. Kochthongrasamee, P. Prasassarakich and S. Kiatkamjornwong, J. Appl. Polym. Sci., 2006, 101, 2587–2601.
32. W. Arayapranee and G. L. Rempel, J. Appl. Polym. Sci., 2008, 110, 2475–2482.
33. Handbook of vinyl polymers: radical polymerization, process, and technology, ed. I. Reetz, M. K. Mishra and Y. Yagci, CRC Press/Taylor & Francis, Boca Raton, 2nd edn, 2009.
34. S. Mohapatra and G. B. Nando, Ind. Eng. Chem. Res., 2013, 52, 5951–5957.
35. R. K. Roy, Design of experiments using the Taguchi approach: 16 steps to product and process improvement, John Wiley & Sons, Inc., Newyork, 2001.
36. N. Sombatsompop and C. Kumnuantip, J. Appl. Polym. Sci., 2003, 87, 1723–1731.
37. A. Salahi, T. Mohammadi, M. Nikbakht, M. Golshenas and I. Noshadi, Desalin. Water Treat., 2012, 48, 27–37.
38. P. C. de Oliveira, A. M. de Oliveira, A. Garcia, J. C. de Souza Barboza, C. A. de Carvalho Zavaglia and A. M. dos Santos, Eur. Polym. J., 2005, 41, 1883–1892.
39. J. F. Anstey, N. Subramaniam, B. T. T. Pham, X. Lu, M. J. Monteiro and R. G. Gilbert, Macromol. Symp., 2000, 150, 73–84.
40. Y. Ma, H. Hu, D. Northwood and X. Nie, J. Mater. Process. Technol., 2007, 182, 58–64.
41. P. J. Ross, Taguchi techniques for quality engineering, second edn, McGraw-Hill Companies, Inc., USA, 1996.
42. P. Rani, S. Mishra and G. Sen, Carbohydr. Polym., 2013, 91, 686–692.

---