Abbreviations and symbols
carbonyl telechelic natural rubber
epoxidized liquid natural rubber
hydroxyl telechelic natural rubber
polydispersity index (M̅w/M̅n)
parts per hundred rubber
degree of polymerization at beginning t=0, according to Eq. (5)
degree of polymerization at the reaction time t, according to Eq. (5)
number-average molecular weight
weight-average molecular weight
Polyurethane is a commodity plastic that has been used in many applications such as building and construction, thermal insulations, transportations, sport equipment, and footwear due to its desirable lightweight, excellent thermal insulating and mechanical properties , , , , , , . In the chemical structures of polyurethane, urethane repeating units are produced by the reaction of isocyanates and polyols, which contain hydroxyl groups at the chain ends , .
Nowadays, polyols are readily derived from biomass feedstocks (e.g. natural rubber (NR) and vegetable oil) , , , , , , . NR is one of the most well-known polymers that can be functionalized into biobased polyols , ,  due to its structure that contains unsaturated bonds. Scheme 1 shows a four-step reaction pathway for preparing biobased polyurethane from NR that includes three intermediates: epoxidized liquid natural rubber (ELNR), carbonyl telechelic natural rubber (CTNR), and hydroxyl telechelic natural rubber (HTNR). However, the use of virgin NR in Scheme 1 yields non-uniform properties of the resulting biobased polyurethanes due to the broad and high molecular weight of the natural polymer , . To produce suitable polymeric materials, it is necessary to reduce the molecular weight of virgin NR before preparing biobased polyurethane.
In the chemical degradation process, oxidizing agents (e.g. hydrogen peroxide , , , periodic acid , , potassium persulfate , , and diphenyl disulfide , , ) were used for molecular weight reduction. Among these oxidizing agents, hydrogen peroxide (H2O2) is preferable due to its environmentally friendly characteristic , . However, H2O2 is typically supplied in an aqueous solution with water up to 70 wt.%, while NR is only soluble in organic phases (e.g. toluene or xylene solvents). Thus, the NR molecular weight reduction with H2O2 as an oxidizing agent is limited due to the mass transfer of H2O2 from the aqueous phase into the organic phase in the biphasic aqueous-organic systems.
The use of CO2 in CO2-expanded liquid (CXL) reaction systems is a promising technique for green and sustainable processes , , ,  because it allows one to adjust both polarity and transport properties by altering pressure and temperature. In the triphasic CO2/organic solvent/H2O systems with H2O2, the addition of CO2 can (i) promote the formation of oxidative peroxycarbonic acid intermediate by the reaction of CO2 and H2O2 , ,  and (ii) promote the solubility of small molecules in the organic phase (e.g. H2, O2, and oxidizing agent) , ,  by lowering the viscosity and interfacial tension of the biphasic aqueous-organic systems . The peroxycarbonic acid intermediate can be used as an effective oxidative agent in the epoxidation of olefins in the biphasic systems reported in literature , , , . Thus, chemical degradation in the triphasic CO2/toluene/H2O systems with H2O2 could improve the efficiency of the NR molecular weight reduction.
The first objective of this work is to study the enhancement of NR molecular weight reduction in triphasic systems of CO2/toluene/H2O using H2O2 as an oxidizing agent and toluene as an organic solvent for NR dissolution. The second objective of this work is to prepare biobased polyurethane obtained from the NR molecular weight-modified polymers. A comparative study on the NR molecular weight reduction in biphasic water-toluene systems without the addition of CO2 was made as a basis for assessing the results. Reactions were performed to evaluate the activation energy at a constant CO2 pressure of 12 MPa. Gel permeation chromatography (GPC), Fourier-transform infrared spectroscopy (FT-IR), and proton nuclear magnetic resonance (1H-NMR) analyses were used to analyze the products.
2 Materials and methods
Virgin NR STR-5L (Standard Thai Test Rubber) with dirt less than 5% was used as received without further purification. Liquid CO2 (purity≥99.95%) was supplied by United Industrial Gases (Thailand). Hydrogen peroxide (H2O2) with a purity of 30 wt.% was obtained from Fluka Chemical. Toluene, methanol (AR grade), tetrahydrofuran (AR grade), and chloroform (HPLC grade) were purchased from RCI Labscan Co. (Thailand). Periodic acid (H5IO6) and sodium borohydride (NaBH4) were purchased from ACS Reagent Chemical Co. Dibutyltindilaurate (DBTDL) as a catalyst was obtained from Air Products and Chemicals, and toluene diisocyanate (TDI) from IRPC Public Company Limited (Thailand) was used in the preparation of biobased polyurethanes.
2.2 NR molecular weight reduction
NR molecular weight reduction was performed in the presence of CO2 at a constant pressure of 12 MPa over a temperature range of 60°C–80°C and at reaction times of 1 h–10 h using H2O2 concentrations of 10–30 parts per hundred rubber (phr). These conditions were adopted from the literature that did not consider experiments in the presence of CO2 , , .
In the experimental apparatus (Figure 1), a solution of NR (1% w/v in toluene) of 5 ml was loaded into a high-pressure vessel (15 ml, Taiatsu Techno/Japan SUS316) and mixed with H2O2 using a magnetic stirrer (550 rpm, g≈9.98 m·s−2). The high-pressure vessel was then placed in a water bath at a constant temperature of 60°C. The residual air inside the vessel was subsequently flushed out with CO2 from the cylinder. CO2 was delivered into the high-pressure vessel by a syringe pump (ISCO 260D, USA) until the pressure inside the vessel rose to 12 MPa. CO2 was then discharged from the vessel after reaching the reaction time. Comparative experiments on NR molecular weight reduction using similar conditions were made without the addition of CO2. The obtained ELNR products in both cases were analyzed with GPC, FT-IR, and 1H-NMR techniques.
2.3 Preparation of biobased polyurethane
ELNRs obtained from the NR molecular weight reduction were subsequently functionalized to CTNR and HTNR as shown in Scheme 1. CTNR was prepared by adding 1.1 mol equivalent of periodic acid (H5IO6) to 0.4 mol·l−1 of the ELNR solutions [Eq. (2), Scheme 1]. The solutions were stirred at a rotational speed of 720 rpm at 60°C for 6 h in a glass reactor with a magnetic stirrer. The CTNR product was neutralized with sodium hydrogen carbonate and sodium chloride.
Subsequently, the resulting CTNR was functionalized to the HTNR solution by adding excess sodium borohydride (5 mol equivalent) at 60°C for 6 h [Eq. (3), Scheme 1]. The obtained HTNR product was then hydrolyzed with 10 ml of cool water, purified by a sodium chloride solution, and dehydrated by magnesium sulfate. The molecular weight and functional groups of the obtained HTNR were analyzed by GPC, FT-IR, and 1H-NMR.
The HTNR was then used as a raw material for preparing HTNR-based polyurethane by a one-shot method ,  [Eq. (4), Scheme 1]. A solution of HTNR in THF at a concentration of 0.5% (w/v) was mixed with dibutyltindilaurate (DBTDL) as a catalyst at a [DBTL]/[OH] ratio of 0.045, before adding isocyanate into the mixture at a [NCO]/[OH] ratio of 1.2 at 60°C. After stirring for 20 min, the resulting HTNR-based polyurethanes were cast onto an aluminum substrate and then cured at 40°C for 48 h. The functional groups of HTNR-based polyurethanes were analyzed with FT-IR.
2.4 Analysis methods
Gel permeation chromatography (GPC): The number-average molecular weight (M̅n), weight-averaged molecular weight (M̅w), and polydispersity index (PDI, M̅w/M̅n) of the samples were characterized by GPC. Before the measurements, the GPC apparatus was calibrated using 0.1% (w/v) polystyrene standard solutions (PS, molecular weights of 30–900 kg·mol−1) in chloroform. The sample was prepared as a 0.1% (w/v) chloroform solution filtered through a 0.2-μm Teflon (Millex) filter. The GPC system had an HPLC pump (Spectra System, model P2000), a GPC column (Shodex, GPC KF-80M with two columns), and a UV-Vis detector at a fixed wavelength (λ) of 254 nm (Lab Alliance, model 201, USA). Chloroform (HPLC) was used as the mobile phase, and the flow rate of the mobile phase was 1 ml·min−1 in the column maintained at 40°C. A sample (20 μl) was injected into the column in the measurements.
FT-IR and 1H-NMR: The functional groups in the chemical structure of NR, ELNR, CTNR, HTNR, and PU were analyzed using FT-IR and 1H-NMR. A dried sample was mixed with potassium bromide tablets and then dried at 60°C for 2 h to remove moisture. The FT-IR spectra were taken in transmission mode with an FT-IR spectrometer (Thermo-Nicolet Avatar 360 Multi Bounce, USA). Each sample was dissolved in chloroform-d (CDCl3), and tetramethylsilane (TMS) was used as the internal standard solution. The 1H-NMR spectra were measured with an1H-NMR spectrometer (Bruker, 300 MHz, USA).
3 Kinetic studies
The rate constant (k) and activation energy (Ea) were estimated for assessing the enhancement of NR molecular weight reduction by CO2. The rate constants were estimated from second-order reaction equations as follows , , ], :
where DPn(t) and are the degree of depolymerization at the reaction time t and the beginning t=0, respectively. The M̅n and M0 are the number-average molecular weight of NR and repeating unit of isoprene (68 g·mol−1). The Ea was determined by the Arrhenius equation as follows:
where A is the frequency of collisions in the correct orientation, and R is the universal gas constant.
4 Results and discussion
4.1 NR molecular weight reduction
Figure 2A–C and Table 1 show molecular weight reduction in terms of M̅n values of ELNR with CO2 and without CO2. With an increase in reaction time (Figure 2A–C), the M̅n values of ELNR in both cases tended to decrease exponentially during the first two hours and then gradually plateaued. Trends in decrease of M̅n values of ELNR with CO2 (Figure 2A–C) were larger than those without CO2, especially at low temperature.
Trends of NR molecular weight reduction (Figure 2A–C) in both cases were similar to the second-order reaction rate of depolymerization [Eq. (5)]. The k [Eq. (5)] and Ea [Eq. (7)] were calculated from the M̅n values (Table 1). Figure 2D–F shows a linear relationship between degree of depolymerization (1/DPn(t)) and reaction time, according to Eq. (5). The rate constants for NR molecular weight reduction were estimated from the slopes in Figure 2D–F and are tabulated in Table 2. The k values for chemical degradation in both cases (Table 2) showed an increase with temperature, while the k values for experiments made in the presence of CO2 were higher than those in the absence of CO2.
The Ea can be determined from the slopes in Figure 3, according to Eq. (7). The Ea value for molecular weight reduction in the presence of CO2 with H2O2 was 38 kJ·mol−1, which was lower than that in the absence of CO2 (121 kJ·mol−1) or methods in the literature such as chemical degradation with other potassium persulfate oxidizing agent (77 kJ·mol−1)  and thermal decomposition of NR (203 kJ·mol−1) . Due to lower oxidative activity, Ea value from chemical degradation with H2O2 was lower than that from potassium persulfate, however, combination of H2O2 with CO2 allowed more effective than that using potassium persulfate.
Figure 4 shows a proposed process for NR molecular weight reduction in triphasic CO2/organic/H2O systems. The triphasic systems (Figure 4) include the CO2 phase, the aqueous phase containing H2O2 solution, and the organic phase of NR solution in toluene solvent. As CO2 can be soluble in both aqueous and organic phases, CO2 can cause (i) the formation of peroxycarbonic acid in the aqueous phase by the reaction of CO2 and H2O2, as reported in the literature , , ,  and (ii) the promotion of the mass transfer of H2O2 to react with NR in the organic phase due to a reduction in viscosity in the organic phase, as reported in the literature , , , , and interfacial tension between the aqueous and organic phases according to the molecular dynamics simulation .
To elucidate the effect of peroxycarbonic acid formation on NR molecular weight reduction, the NR molecular weight reduction was carried out in the H2O-CO2 system without the addition of H2O2. The M̅w and PDI values of the NR obtained in the H2O-CO2 system (50°C, 12 MPa, and 5 h) were 8.3×105 g·mol−1 and 3.4, respectively. The NR molecular weight reduction in the H2O-CO2 system was less effective than that with the addition of H2O2, according to the higher M̅w values (Table 1) obtained with CO2. These results were consistent with studies on epoxidation of cyclohexene in H2O-CO2 systems  in that no reaction occurred.
The reason for the lack of reaction in the H2O-CO2 system without the addition of H2O2 could be due to unsuitable conditions such as pH for epoxidation because pH higher than 7 is generally required for epoxidation reactions , . However, the reaction of CO2 with H2O can generate carbonic acid (pH≈3–4)  in the H2O-CO2 systems, while the pH values in the aqueous solution with the addition of H2O2 in the presence or absence of CO2 are typically higher than 7 , .
CO2 can reduce the viscosity of organic solvents , , ,  and polymer solutions , , , while the effect of viscosity reduction is negligible in aqueous solutions . Thus, the main reasons for efficient NR molecular weight reduction in the triphasic systems with H2O2 as the oxidizing agent can be thought of as (i) the formation of oxidative peroxycarbonic acid intermediate that synergically reduces the NR molecular weight along with H2O2 and (ii) the promotion of mass transport as CO2 can reduce both organic solution viscosity and interfacial tension that would enhance the degradation reaction and increase in the O2 transport.
The addition of H2O2 up to 30 phr (Figure 5) caused an increase in M̅n and polydispersity index due to an excess amount of the oxidative agent that promoted retrogressive repolymerization among the free radical chain ends , . Therefore, the chemical degradation using H2O2 at 20 phr in the presence of CO2 was used to prepare the ELNR, and the biobased polyurethane as shown in Scheme 1 is discussed in the following section.
4.2 Chemical structures of NR from molecular weight reduction
The chemical structures of the ELNR obtained from NR molecular reduction were characterized with 1H-NMR and FT-IR (Figures 6 and 7). The 1H-NMR spectra of both the NR and ELNR (Figure 6A and B) showed the olefinic proton of cis-1,4-polyisopren unit at 5.10 ppm, the methyl proton at 1.67 ppm, and the methylene proton at 2.16 ppm next to the C=C bond. After the molecular weight reduction, small signals of the ELNR (Figure 6B) were detected from the methine proton adjacent to the epoxide ring at 2.70 ppm and the methyl group (-CH3) adjacent to the epoxide unit at 1.30 ppm. The epoxide content of the ELNR was estimated from methods in the literature , ,  by considering the integration areas of signals of 2.70 and 5.10 ppm and determined to be 5.7%. Figure 7 shows the FT-IR spectra of the NR and ELNR that exhibited the C=C stretching of polyisoprene at 1665 cm−1. The ELNR (Figure 7) showed a weak signal of the epoxide ring (C-O-C) at 870 cm−1. The 1H-NMR and FT-IR results obtained in this work were in accordance with the literature , , .
4.3 Preparation of biobased polyurethane
The obtained ELNR was functionalized to CTNR and HTNR [Scheme 1, Eqs. (2) and (3)]. The chemical structures of CTNR and HTNR were confirmed by FT-IR and 1H-NMR (Figures 6 and 8). The FT-IR spectra of the C=O stretching of CTNR at 1720 cm−1 (Figure 8) were observed, which were consistent with the 1H-NMR results (Figure 6C) wherein there appeared new peaks for the aldehyde proton (at 9.80 ppm), the methylic proton in the ketone end groups (at 2.13 ppm), and the CH2 in the α and β terminal carbonyl (-CH2) groups (between 2.20 and 2.60 ppm).
The FT-IR spectra of the OH stretching in the HTNR at 3440 cm−1 (Figure 8) were consistent with 1H-NMR results (Figure 6D) as they showed the appearance of new peaks for methylic protons adjacent to a secondary alcohol (at 1.20 ppm) and two peaks corresponding to CH (3.80 ppm) and CH2 (3.68 ppm) adjacent to alcohol groups at the chain ends. Thus, the FT-IR and 1H-NMR results showed that the hydroxyl functional groups were located at the chain ends of the HTNR.
The functionality of the HTNR was estimated by an 1H-NMR method reported in the literature . The HTNR obtained in this work had two functionalities, indicating that there are two hydroxyl groups at the chain end of the HTNR structure so that the prepared HTNR material can be used for preparing the biobased polyurethane. The biobased polyurethane synthesized from the HTNR was confirmed by FT-IR analysis (Figure 8) that exhibited absorption wavenumbers of 3300 cm−1 for N-H stretching and 1720 cm−1 for C=O stretching.
High-pressure (12 MPa) CO2 improves the efficiency of NR molecular weight reduction in water-toluene biphasic systems with hydrogen peroxide as the oxidizing agent. The role of CO2 in the reaction system seems to be to improve mass transport in the toluene phase and to enhance the oxidation rate through H2O2 transport and formation of peroxycarbonic acid intermediate. CO2 lowers the activation energy for the NR molecular weight reduction using H2O2 oxidation by a factor of about three in the water-toluene-NR reaction system. The ELNR obtained a molecular weight reduction that can be used subsequently to prepare biobased polyurethane.
This work was financially supported by the National Research Council of Thailand (NRCT).
Li Y, Luo X, Hu S. Introduction to Bio-based Polyols and Polyurethanes. Bio-based Polyols and Polyurethanes, Springer International Publishing: Berlin, 2015. Google Scholar
Amnuaysin T, Buahom P, Areerat S. J. Cell. Plast. 2015, 52, 585–594. Google Scholar
Shi J, Jiang K, Zou H, Ding L, Zhang X, Li X, Zhang L, Ren D. J. Appl. Polym. Sci. 2014, 131, 40298. Google Scholar
Saetung A, Rungvichaniwat A, Campistron I, Klinpituksa P, Laguerre A, Phinyocheep P, Pilard J-F. J. Appl. Polym. Sci. 2010, 117, 1279–1289. Google Scholar
Duereh A, Smith RL. J. Supercrit. Fluids, 2017, doi.org/10.1016/j.supflu.2017.11.004. (in press). Google Scholar
About the article
Published Online: 2018-09-29
Published in Print: 2019-01-28
Citation Information: Green Processing and Synthesis, Volume 8, Issue 1, Pages 288–296, ISSN (Online) 2191-9550, DOI: https://doi.org/10.1515/gps-2018-0092.
©2019 Walter de Gruyter GmbH, Berlin/Boston. This work is licensed under the Creative Commons Attribution 4.0 Public License. BY 4.0