Influence of Nanogels on Mechanical, Dynamic Mechanical, and Thermal Properties of Elastomers
© to the authors 2009
Received: 6 October 2008
Accepted: 27 January 2009
Published: 13 February 2009
Use of sulfur crosslinked nanogels to improve various properties of virgin elastomers was investigated for the first time. Natural rubber (NR) and styrene butadiene rubber (SBR) nanogels were prepared by prevulcanization of the respective rubber lattices. These nanogels were characterized by dynamic light scattering, atomic force microscopy (AFM), solvent swelling, mechanical, and dynamic mechanical property measurements. Intermixing of gel and matrix at various ratios was carried out. Addition of NR gels greatly improved the green strength of SBR, whereas presence of SBR nanogels induced greater thermal stability in NR. For example, addition of 16 phr of NR gel increased the maximum tensile stress value of neat SBR by more than 48%. Noticeable increase in glass transition temperature of the gel filled systems was also observed. Morphology of these gel filled elastomers was studied by a combination of energy dispersive X-ray mapping, transmission electron microscopy, and AFM techniques. Particulate filler composite reinforcement models were used to understand the reinforcement mechanism of these nanogels.
Virgin polymers, especially elastomers have inherently low stiffness and strength. In order to overcome these obvious limitations and to expand their applications in different fields, particulate fillers, such as carbon black, silica, glass, calcium carbonates, carbon nanotubes, nano clays etc. are often added to polymer. Particulate fillers modify physical and mechanical properties of polymers in many ways. Use of carbon black for improving reinforcement properties of an elastomer has been studied extensively in numerous investigations [1, 2]. Amongst the nonblack fillers, mostly silica provides the best reinforcing properties . In the last decade, it has been shown that dramatic improvements in mechanical and other properties can be achieved by incorporation of a few weight percentages (wt%) of inorganic exfoliated clay minerals consisting of mostly layered silicates in polymer matrices [4–10]. These are better known as polymer nanocomposites. Similar enhancements in various properties have also been reported with other types of nanofillers e.g. multiwalled carbon nanotubes and layered double hydroxides [11, 12].
Although not strictly categorized as filler, use of gels to improve various physical properties of elastomers, with an added advantage of superior processability, can be found in the prevailing literature [13–15]. Kawahara et al.  have reported the effect of gel on green strength of natural rubber. In most of the above work, the authors have used physically crosslinked or entangled network gels. However, our recent preliminary work with chemically crosslinked nanogels and quasi-nanogels has revealed that addition of these gels leads to a considerable improvement in processability, mechanical, and dynamic mechanical properties of virgin natural rubber (NR) and styrene butadiene rubber (SBR) [17–19]. Optimization of these as-prepared crosslinked gels has been carried out by measuring various physical properties including crosslink density and the optimum level of gel loading has been determined from the rheological properties of the gel filled systems [17, 19]. However, the extent of property enhancement upon the addition of chemically crosslinked gels varies with the nature of matrix and gels. In the present work, our aim was to improve the deficiency in virgin NR property by using SBR nanogels and vice versa. For example, we have attempted to improve the thermal stability of NR using SBR gels which have inherently better thermal stability, without sacrificing any other properties. Similarly, green strength of SBR can be improved greatly by using the relatively high strength NR gels. For this purpose, NR and SBR latex nanogels having gradient of crosslink density and different particle sizes were prepared by sulfur prevulcanization technique and thoroughly characterized. These latex gels were then intermixed with neat NR and SBR lattices at different loadings. Finally, influence of these chemically crosslinked gels on mechanical, dynamic mechanical, and thermal behavior of virgin elastomers was studied in detail along with an extensive morphological study, for the first time.
High ammonia centrifuged natural rubber (NR) latex having 60% dry rubber content (DRC) was provided as free sample by the Rubber Board, Kottayam, India. Sulfur, zinc oxide (ZnO), and zinc diethyl dithiocarbamate (ZDC), all in 50% aqueous dispersion, were also obtained from the same source and used as received. Styrene butadiene rubber (SBR) latex having 30% total solid content (T.S·C) and 30% bound styrene content, with a pH of 10.5 was generously received as gift sample from the Apar Industries, Ankeleswar, India. Toluene (LR-grade), potassium hydroxide (KOH), and potassium laurate (KC12H23O2) were procured from s.d. Fine Chemicals, Mumbai, India. Doubly distilled water was obtained from indigenous source.
Preparation of Sulfur Prevulcanized Latex Gel and Gel Filled Rubber
Formulations for sulfur prevulcanization
Ingredients (dry wt basis)
NR latex (60%)
SBR latex (30%)
10% potassium laurate
50% sulfur dispersion
50% ZDC dispersion
50% ZnO dispersion
Intermixing of gel filled raw rubber samples was carried out by adding a given amount of a particular type of NR latex gel to virgin SBR latex and vice versa, followed by gentle stirring (200–300 rpm) for 1 h at 25 ± 2 °C. Then, these were cast and dried following the above-mentioned procedure. These gel filled raw rubber films were used for further testing.
Control natural rubber latex and styrene butadiene rubber latex were designated as NR and SBR, respectively. Individual NR and SBR gels were expressed as NS a and SBS a , respectively, where ‘a’ represents the ratio of sulfur to accelerator used in the prevulcanization recipe. NR gel mixed SBR systems were denoted as SBNSa/b, where ‘a’ has the same notation as stated above and ‘b’ is the amount (phr) of prevulcanized NR gel added into the SBR latex. Similarly, SBR gel filled NR latex systems were noted as NRSBSa/c, where ‘a’ has the same meaning as stated above and ‘c’ is the amount (phr) of prevulcanized SBR gel added into the NR latex.
Characterization of Gelled Latex Samples and Measurements of Various Properties of Gel Filled Rubbers
whereW1is the initial weight of the polymer andW2,the weight of the insoluble portion of the polymer. The results reported here are the averages of three samples.
where DsFfAwAsρr, and ρs are deswollen weight of the sample, fraction insoluble, sample weight, weight of the absorbed solvent corrected for swelling increment, density of rubber, and density of solvent, respectively.
Dynamic light scattering (DLS) technique was used for the measurement of particle size of gels and their distribution. Before testing, the latex samples were diluted to 0.1 g/L concentration level using doubly distilled water. The DLS studies were carried out in Zetasizer Nano-ZS (Malvern Instrument Ltd, Worcestershire, UK) with a He–Ne laser of 632.8 nm wavelength. The data were analyzed by in-built machine software. The mean hydrodynamic particle diameter (Zavg) was directly obtained from the machine software (as per ISO 13321).
The energy dispersive X-ray sulfur (S) mapping of the gel filled raw rubber systems was recorded in Oxford ISIS 300 EDX system (Oxford Instruments, Oxfordshire, UK) attached to the JSM 5800 (JEOL Ltd., Tokyo, Japan) scanning electron microscope operating at an accelerating voltage of 20 kV. The scan size in all the specimens was 10 square microns with a 200× magnification. The white points in the figures denote sulfur signals.
The morphology of the gel particles, as well as the gel filled matrices was analyzed with the help of atomic force microscopy (AFM). AFM studies were carried out in air at ambient conditions (25 °C, 60% RH) using multimode AFM, from Veeco Digital Instruments, Santa Barbara, CA, USA. Topographic height and phase images were recorded in the tapping mode AFM with the set point ratio of 0.9, using silicon tip having spring constant of 40 N/m. The cantilever was oscillated at it resonance frequency of ~280 kHz. Scanning was done at least 3 different positions of each sample and the representative images were taken. The latex gel samples were diluted several times before testing with doubly distilled water. A drop of this diluted sample was placed on a freshly cleaved mica surface which was allowed to dry before taking the image. In the case of gel filled matrices, very thin cast film samples were used for morphology. Due to the difference in their elastic modulus, one of the phases appears darker (NR) and the other one brighter (SBR) in all the AFM micrographs.
The gel filled rubber samples for transmission electron microscopy (TEM) analysis were prepared by ultra-cryomicrotomy using Leica Ultracut UCT, at around 30 °C below the glass transition temperature of the compounds. Freshly cut glass knives with cutting edge of 45° were used to get the cryosections of 50-nm thickness. The microscopy was performed using JEM-2100 (JEOL Ltd., Tokyo, Japan) operating at an accelerating voltage of 200 kV.
For the measurement of mechanical properties of the neat matrix, individual gels and gel filled matrices, tensile specimens were punched out from the cast sheets of 1 mm thickness, using ASTM Die-C. The tests were carried out as per the ASTM D 412-98 method in a universal testing machine, Zwick Roell Z010 (Zwick Roell, Ulm, Germany), at a crosshead speed of 500 mm per min at 25 ± 1 °C. TestXpert II software (Zwick Roell, Ulm, Germany) was used for data acquisition and analysis. The average of three tests is reported here. The experimental error was within ±1% for tensile strength and modulus, and within ±3% for elongation at break values.
Dynamic mechanical properties of gels, as well as gel filled rubbers were measured as a function of temperature using the Dynamic Mechanical Analyzer DMA Q800 (TA Instruments, Luken’s Drive, New Castle, DE, USA). The measurements were taken under film-tension mode in the appropriate temperature range with a heating rate of 3 °C/min and at 1 Hz frequency. The peak value of Tan δ curves was taken as the glass transition temperature (Tg). Thermal Advantage software (TA Instruments, Newcastle, Delaware) was used for data acquisition and analysis.
Thermogravimetric analysis (TGA) of gel filled systems was done using TA Instruments (Luken’s Drive, New Castle, DE, USA) TGA-Q 50. The samples (10 ± 2 mg) were heated from ambient temperature to 700 °C in the furnace of the instrument under nitrogen atmosphere at a flow rate of 60 mL/min. The experiments were done at 10 °C/min heating rate and the data of weight loss versus temperature were recorded online in the TA Instrument’s Q series Explorer software. The analysis of the thermogravimetric (TG) and derivative thermogravimetric (DTG) curves was done using TA Instrument’s Universal Analysis 2000 software version 3.3B. In the present study, the temperature corresponding to 5% weight loss was taken as initial degradation temperature (Ti) and the temperature corresponding to the maximum rate of degradation in the derivative thermogram was considered as peak degradation temperature (Tmax). The experimental error limit was within ± 1 °C.
Results and Discussion
Characterization of Crosslinked Nanogels
Various properties of the gels
Z-avg diameter (nm)
Gel content (%)
Crosslink density × 104(gmolcm−3)
Young’s modulus (MPa)
E′ at 25 °C (MPa)
The values of gel content and crosslink density for all the crosslinked gels are tabulated in Table 2. With the increase in sulfur to accelerator ratio, both gel content and crosslink density increase for SBS as well NS gel systems. SBS0.5has a gel content of 89%, which increases up to 97% in SBS3. A similar trend is also observed for crosslink density (0.8 × 10−4gmol/cc for SB0.5to 2.4 × 10−4gmol/cc for SB3). A comparable increase in gel content and crosslink density is observed for NR gels. The increment in gel content and crosslink density values with increasing sulfur to accelerator ratio can be attributed to the formation of sulfide linkages between the molecules, which lead to a three-dimensional network structure. However, as the sulfur to accelerator ratio increases from 2 to 3, the increase in the amount of crosslinking tends to level off, as evident from gel content and crosslink density values of SBS2/SBS3and NS2/NS3systems. This is because of the saturation of sites available for crosslinking. Although the gel content values are quite close for both SBS and NS types of gels at any given sulfur to accelerator ratio, SBR gels show almost double the amount of crosslinking than their NR gel counterparts. Because of the nano size of SBR latex particles compared to the NR latex, higher available surface area in nano latex particle leads to the efficient diffusion of these curing agents during prevulcanization and hence higher amount of crosslinking.
The effect of sulfur crosslinking is also very pronounced on the mechanical properties of different gels as compared to their virgin counter parts. The mechanical properties of the gelled lattices are reported in Table 2. The maximum tensile stress of the control SBR latex (SB), which is only 0.29 MPa, shows many fold increase after sulfur crosslinking. The elongation at break (EB) value of neat SBR is 700%, which decreases considerably upon crosslinking to 360% in SBS3. The tensile strength (TS) increases steadily, while the EB value decreases consistently with the increase in amount of sulfur in the system. Increase in T.S. and reduction in EB values are related to the introduction of greater number of crosslinks initiated by the sulfide linkages. In the case of NS series of gels, TS value increases by more than 10 times from 1.86 MPa in NR to 18.9 MPa in NS3gel with a concomitant decrease in EB from 1400% in NR to 1120% in NS3. The trend in Young’s modulus (Ey) values is very similar to that of TS. However, SBR gels have comparatively higher values ofEythan the NR gels.
The dynamic mechanical properties of different gels as compared to that of neat rubber strongly reflect the influence of crosslinking. With the increase in sulfur to accelerator ratio, tan δ peak (considered asTghere) shifts toward higher temperature (Table 2). It is worth mentioning here that the neat NR has aTgof about −56 °C and that of SBR is −39 °C. Hence, considerable increase inTgwith the introduction of crosslinking in the rubber matrix can be seen along with the broadening of tan δ peak height (not shown here). In the case of NR gels,Tgshifts by more than +7 °C (from NR to NS3), while for SBR gels, there is a +8 °C shift from SBR to SBS3. The increase inTgvalues with the progressive increase in sulfur to accelerator ratio can be ascribed to the restriction imposed on the chain movement due to the crosslinking, as there is lesser number of free chains available to execute unrestricted segmental motion. The storage modulus (E′) values at 25 °C are also reported in Table 2for the gels used in this study. As in the case of tensile modulus,E′ also increases steadily with increase in amount of crosslinking. SBS0.5has anE′ value of 0.8 MPa, which increases by more than threefold to 2.63 MPa for SBS3.Similar observations are also noted for NR gels; however, the level of increment in modulus values is less as compared to SBR gels.
These nanogels were subsequently used as viscoelastic fillers for the inter mixing study i.e. NS gels were added to SBR matrix and SBS gels were mixed with NR at a given concentration, to investigate their effect on the morphology, mechanical, dynamic mechanical and thermal properties of raw SBR and NR.
Morphology of the Gel Filled Rubbers
Effect of Gels on the Tensile Properties
Tensile properties of gel filled samples
Young’s modulus (MPa)
Modulus at 300% elongation (MPa)
Elongation at break (%)
where Ec and Em are Young’s modulus of composite and matrix, respectively and Φ is the volume fraction of the fillers. The constant 2.5 is applicable for spherically shaped particles.
where υm is the matrix Poisson ratio taken as 0.5 here. The equation is based on the assumption that the Young’s modulus of the particulate inclusions (Ef) is greater than that of the matrix (i.e. Ef ≫ Em).
Effect of Gels on the Dynamic Mechanical Properties
Figure 8(inset) also illustrates the temperature dependencies of loss tangent of SBR nanogel filled NR systems. With the addition of 16 phr SBS1gel in NR,Tgof NR shifts towards higher temperature by 4 °C, accompanied by steady reduction in tan δ peak height. It is very interesting to mention here that upto 8 phr (~7.4 wt%) of SBS1gel loading in NR generates singleTgcorresponding to NR. However, at 16 phr (~13.9 wt%) gel loading, two distinct peaks can been seen easily (one with a broad shoulder peak at −32.8 °C for SBS1gel). This could be attributed to the macro phase separation of gels with matrix at relatively higher loading. Similar trend is also observed for NS1filled SBR systems (see Figure S4 of Supplementary Information). For example, in SBNS1/16, a small peak appears at −53 °C for NS1along with another one at −33.3 °C for SBR. With addition of NS1gel in SBR,Tgshifts from −39.2 °C in SBR to −32.0 °C in SBNS1/8.The presence of crosslinks in the raw rubber matrix hinders the segmental motions of the polymer chains and therefore,Tgis progressively shifted to higher temperature with the increase in gel loading.
Effect of Gels on the Thermal Properties
The DTG plots of SBS1nanogel filled NR also clearly demonstrate the improvement in thermal stability as shown in the inset of Fig. 9. There is 3 °C shift ofTmaxto higher temperature with addition of 16 phr SBS1gel in neat NR and there is a significant reduction in the rate of decomposition in the presence of the gels at major degradation step (from 2.09%/°C in NR to 1.91%/°C in NRSBS1/16). It can be noted here that prominent 2nd peak in the DTG plots for NRSBS1/8and NRSBS1/16is due to the degradation of SBS1gels. Similar two stage degradation can also be seen in the NS1gel filled SBR systems (see Figure S6 of Supplementary Information). Addition of NS1gels considerably suppresses the rate of decomposition in case of neat SBR (from 1.64%/°C in neat SBR to 1.37%/°C in SBNS1/16).
The use of chemically crosslinked nanogels to improve various properties of virgin elastomers has been reported for the first time. Following conclusions can be drawn from the present work. Sulfur prevulcanized nanosized latex gels have been prepared and characterized using various methods. The morphology of gel filled NR and SBR systems has been studied using X-ray dot mapping, TEM, and AFM. These show that the gels are evenly distributed at low loadings, while they tend to form agglomerates at relatively higher loadings. SBR nanogels have greater tendency for agglomeration.
Addition of chemically crosslinked nanogels also considerably improves the tensile strength and modulus of the gel filled rubbers compared to the pristine one. The tensile strength (or maximum stress) and Young’s modulus increase, whereas EB decrease with the increase in nanogel loading for NR and SBR matrices. The reinforcement ability of the gels depends on their crosslinking densities. Guth and Kerner particulate reinforcement models have been used to understand the reinforcement behavior of these gels.
Presence of nanogels has shifted theTgof neat elastomers towards higher temperature with an concomitant increase in storage modulus. Interestingly, 16 phr gels loaded samples showed two peaks in their tan δ versus temperature plots. Addition of SBR nanogels in neat NR has given rise to better thermal stability for the latter.
The authors would like to thankfully acknowledge the financial assistance provided by Department of Atomic Energy (DAE), Board of Research in Nuclear Sciences (BRNS), Mumbai vide sanction no. 2005/35/4/BRNS/516 dated 08-06-2005 and also to Mr. Pradip K. Maji for the AFM measurements.
- Kraus G: Reinforcement of Elastomers. Interscience, New York; 1965.Google Scholar
- Donnet JB, Bansal RC, Wang MJ: Carbon Black Science and Technology. Marcel Dekker, New York; 1993.Google Scholar
- Waddell WH, Evans LR: Rubber Chem. Technol.. 1996, 69: 377. COI number [1:CAS:528:DyaK28XlvV2iurs%3D]View ArticleGoogle Scholar
- Lan T, Pinnavaia TJ: Chem. Mater.. 1994, 6: 2216. COI number [1:CAS:528:DyaK2cXmvVWrtrg%3D] 10.1021/cm00048a006View ArticleGoogle Scholar
- Giannelis EP: Adv. Mater.. 1996, 8: 29. COI number [1:CAS:528:DyaK28Xht1Wjsbc%3D] 10.1002/adma.19960080104View ArticleGoogle Scholar
- Vaia RA, Giannelis EP: Macromolecules. 1997, 30: 7990. COI number [1:CAS:528:DyaK2sXnsVSksLs%3D] 10.1021/ma9514333View ArticleGoogle Scholar
- Cho JW, Paul DR: Polymer (Guildf). 2001, 42: 1083. COI number [1:CAS:528:DC%2BD3cXntF2ktLg%3D] 10.1016/S0032-3861(00)00380-3View ArticleGoogle Scholar
- Ray SS, Okamoto M: Prog. Polym. Sci.. 2003, 28: 1539. COI number [1:CAS:528:DC%2BD3sXoslOhsbk%3D] 10.1016/j.progpolymsci.2003.08.002View ArticleGoogle Scholar
- Varghese S, Karger-Kocsis J, Gatos KG: Polymer (Guildf). 2003, 44: 3977. COI number [1:CAS:528:DC%2BD3sXkt1Wlsb4%3D] 10.1016/S0032-3861(03)00358-6View ArticleGoogle Scholar
- Ganguly A, Bhowmick AK: Nanoscale Res. Lett.. 2008, 3: 36. COI number [1:CAS:528:DC%2BD1cXlslejsLY%3D]; Bibcode number [2008NRL.....3...36G] 10.1007/s11671-007-9111-3View ArticleGoogle Scholar
- Singh RP, Singh D, Mathur RB, Dhami TL: Nanoscale Res. Lett.. 2008, 3: 444. COI number [1:CAS:528:DC%2BD1cXhsVyhtrvF]; Bibcode number [2008NRL.....3..444S] 10.1007/s11671-008-9179-4View ArticleGoogle Scholar
- Acharya H, Srivastava SK, Bhowmick AK: Nanoscale Res. Lett.. 2007, 2: 1. COI number [1:CAS:528:DC%2BD2sXhsVSgtL8%3D]; Bibcode number [2007NRL.....2....1A] 10.1007/s11671-006-9020-xView ArticleGoogle Scholar
- Hofman W: Rubber Chem. Technol.. 1964, 7: 85.Google Scholar
- Bhowmick AK, Cho J, MacArthur A, McIntyre D: Polymer (Guildf). 1986, 27: 1889. COI number [1:CAS:528:DyaL2sXjsVelug%3D%3D] 10.1016/0032-3861(86)90177-1View ArticleGoogle Scholar
- Nakajima N, Collins EA: J. Rheol.. 1978, 22: 547. COI number [1:CAS:528:DyaE1MXntlCg]; Bibcode number [1978JRheo..22..547N] 10.1122/1.549488View ArticleGoogle Scholar
- Kawahara S, Isono Y, Sakdapipanich JT, Tanaka Y, Aik-Hwee E: Rubber Chem. Technol.. 2002, 75: 739. COI number [1:CAS:528:DC%2BD38XpvVSqsbY%3D]View ArticleGoogle Scholar
- Mitra S, Chattopadhyay S, Bhowmick AK, Appl J: Polym. Sci.. 2008, 107: 2755. COI number [1:CAS:528:DC%2BD1cXhs1Gjsr4%3D]Google Scholar
- Mitra S, Chattopadhyay S, Bharadwaj YK, Sabharwal S, Bhowmick AK: Radiat. Phys. Chem.. 2008, 77: 630. COI number [1:CAS:528:DC%2BD1cXjvF2muro%3D]; Bibcode number [2008RaPC...77..630M] 10.1016/j.radphyschem.2007.10.006View ArticleGoogle Scholar
- 19. S. Mitra, S. Chattopadhyay, A.K. Bhowmick, Rubber Chem. Technol. 81(5), 842 (2008)View ArticleGoogle Scholar
- Sperling LH: Introduction to Physical Polymer Science. John Wiley & Sons Inc., New York; 1992.Google Scholar
- Bhowmick AK, Hall MM, Benarey H: Rubber Products Manufacturing Technology. Marcel Dekker, New York; 1994.Google Scholar
- Sanguansap K, Suteewong T, Saendee P, Buranabunya U, Tangboriboonra P: Polymer (Guildf). 2005, 46: 1373. COI number [1:CAS:528:DC%2BD2MXnsVOmtA%3D%3D] 10.1016/j.polymer.2004.11.074View ArticleGoogle Scholar
- Ho CC, Khew MC: Langmuir. 1999, 15: 6208. COI number [1:CAS:528:DyaK1MXks1Sqtr0%3D] 10.1021/la981601vView ArticleGoogle Scholar
- Yerina N, Magonov SN: Rubber Chem. Technol.. 2003, 76: 846. COI number [1:CAS:528:DC%2BD3sXovFWntLw%3D]View ArticleGoogle Scholar
- Maiti M, Bhowmick AK: Polymer (Guildf). 2006, 47: 6156. COI number [1:CAS:528:DC%2BD28XnsFOktbo%3D] 10.1016/j.polymer.2006.06.032View ArticleGoogle Scholar
- Pramanik M, Srivastav SK, Samantaray BK, Bhowmick AK: Polym. J. Sci. Part B Polym. Phys.. 2002, 40: 2065. COI number [1:CAS:528:DC%2BD38XmslChsL4%3D] 10.1002/polb.10266View ArticleGoogle Scholar
- Sadhu S, Bhowmick AK: Polym. J. Sci. Part B Polym. Phys.. 2004, 42: 1573. COI number [1:CAS:528:DC%2BD2cXjsV2hsrc%3D] 10.1002/polb.20036View ArticleGoogle Scholar
- Smallwood HJ: Rubber Chem. Technol.. 1948, 21: 667.View ArticleGoogle Scholar
- Guth E: J. Appl. Phys.. 1945, 16: 20. COI number [1:CAS:528:DyaH2MXpvFKh]; Bibcode number [1945JAP....16...20G] 10.1063/1.1707495View ArticleGoogle Scholar
- Kerner EH: Proc. Phys. Soc. B. 1956, 69: 808. Bibcode number [1956PPSB...69..808K] Bibcode number [1956PPSB...69..808K] 10.1088/0370-1301/69/8/305View ArticleGoogle Scholar
- Maiti M, Mitra S, Bhowmick AK: Polym. Deg. Stab.. 2008, 93: 188. COI number [1:CAS:528:DC%2BD1cXotl2mtA%3D%3D] 10.1016/j.polymdegradstab.2007.10.007View ArticleGoogle Scholar