Study of the mechanical, thermal properties and flame retardancy of rigid polyurethane foams prepared from modified castor-oil-based polyols

Industrial Crops and Products 59 (2014) 135–143

Contents lists available at ScienceDirect

Industrial Crops and Products journal homepage: www.elsevier.com/locate/indcrop

Study of the mechanical, thermal properties and flame retardancy of rigid polyurethane foams prepared from modified castor-oil-based polyols Meng Zhang a,∗ , Hui Pan b , Liqiang Zhang a , Lihong Hu a , Yonghong Zhou a,∗ a Institute of Chemical Industry of Forestry Products, CAF, National Engineering Lab. for Biomass Chemical Utilization, Key Lab. of Forest Chemical Engineering, SFA, Key Lab. of Biomass Energy and Material, Nanjing 210042, PR China b Institute of Chemical Engineering, Nanjing Forestry University, Nanjing, Jiangsu 210037, PR China

a r t i c l e

i n f o

Article history: Received 14 December 2013 Received in revised form 5 May 2014 Accepted 10 May 2014 Keywords: Castor-oil-based polyols Thermal properties Flame retardant Polyurethane foams

a b s t r a c t Po1lyurethane foams (PUFs) were prepared using modified castor-oil-based polyols (MCOs). In the first stage, castor oil (CO) was converted into monoglycerides and diglycerides by alcoholysis with glycerol and pentaerythritol. Next, the polyester polyols were synthesised by condensation with the alcoholysis of CO and phthalic anhydride. The chemical and physical properties, foaming behaviour and miscibility with other components of the MCOs were studied by mechanical testing, Fourier transform infrared (FTIR) spectroscopy, gel permeation chromatography (GPC) and thermogravimetric analysis (TGA). The results showed that the components of the MCOs and the foaming behaviour of the foams prepared from the MCOs were similar to those of commercial polyester polyol PS-3152. The reaction activities esterification modified CO polyols were higher than those of alcoholysis modified CO polyols, due to the higher relative content of primary hydroxyl groups. The MCOs and CO had higher thermal stability and better miscibility with polyether polyol 4110 and the physical blowing agent cyclopentane than PS-3152. The properties and flame retardancy of PUFs prepared from MCOs were studied by mechanical testing, TGA and cone calorimetry. The results indicate that the PUFs prepared from castor-oil-based polyester polyols with a reasonable distribution of soft and hard segments had better mechanical properties and thermal conductivities than the PS-3152-based PUF5. Additionally, the MCO-modified PUFs exhibited much higher thermal stability during the pyrolysis process. The cone calorimetry results showed that adding flame retardant ammonium polyphosphate (APP) into PUFs can significantly decrease their heat release rate (HRR), total heat release (THR) and mass loss. These test results indicate that APP has a better synergistic effect with phthalic anhydride polyester polyols than long-chain fatty polyols. All of these unique properties of the MCO-modified rigid PUFs were correlated to the structures of these PUFs. This study may lead to the development of a new type of polyurethane foam using castor oil. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Polyurethane foams (PUFs) are widely used in furniture, packing, insulation and automobile manufacturing. PU materials in the current market are usually made from petro-based polyether or polyester polyol and polyisocyanate through urethane linkages. These petro-based PU products have limited bio-degradability when discarded after use and thus pose an environmental problem (Corcuera et al., 2010; Krämer et al., 2010). Therefore, biobased materials obtained from renewable resources are receiving

∗ Corresponding authors. Tel.: +86 25 85482520; fax: +86 25 85413445. E-mail addresses: [email protected] (M. Zhang), [email protected] (Y. Zhou). http://dx.doi.org/10.1016/j.indcrop.2014.05.016 0926-6690/© 2014 Elsevier B.V. All rights reserved.

considerable attention for use in an increasing number of applications (Campanella et al., 2009; Belgacem and Gandini, 2008; Williams and Hillmyer, 2008) from a social, environmental and energy standpoint due to the increasing emphasis on issues concerning waste disposal and the depletion of non-renewable resources. Vegetable oil is composed of triglycerides of long-chain fatty acids. The most common chain lengths in these fatty acids are 18 or 20 carbon atoms, which can be either saturated or unsaturated at the double bonds located at the 9th, 12th and 15th carbons. They are relatively low-cost materials and offer a priori the possibility of biodegradation. Bio-based materials derived from natural oils, such as castor, palm, canola and soybean oils, have been used to synthesise polyols, which can be used as raw materials in the preparation of bio-based polyurethane foam (Hablot et al., 2008; Petrovic, 2008; Sharna and Kundu, 2008; Xu et al., 2008).

136

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

The preparation of various polyurethane products has been extensively studied in recent years, especially those containing biomass-originating raw materials (Hatakeyama and Hatakeyama, 2005; Hatakeyama et al., 2005; Hatakeyema et al., 2005; Kurimoto et al., 2000, 2001). Moreover, it has become clear from previous studies that biomass-based PUs are biodegradable. Castor oil is not edible, does not compete with food and has free secondary hydroxyl groups. Approximately 90% of CO is ricinoleic acid, and the remaining 10% is comprised of oleic and linoleic acids. Ricin oleic acid (C18:1) has an hydroxyl functional group at the 12th carbon, which provides CO with a hydroxyl value of between 160 and 180 mg KOH/g and makes it a favourable raw material for PU products. There are several papers related to the elaboration of PUs based on castor oil, such as interpenetrating polymer networks (Niranjan et al., 2009), elastomers (Oprea, 2010), coatings (Trevino and Trumbo, 2002), adhesives (Somani et al., 2003), and some semi-rigid PU foams that have potential uses in thermal insulation (Ogunniyi et al., 1996; Chian and Gan, 1998; Maznee et al., 2001; Siwayanan et al., 1999) and other interesting applications. Wang et al. (2008) prepared a polyester-polyol (MACO) by chemical modification of hydroxyl groups with maleic anhydride. Then, biodegradable semi-rigid PU foams were synthesised and exhibited mechanical properties comparable to those of a foam derived from commercial polyether. Mazo et al. (2012) studied the kinetics of transesterification and condensation of castor oil with maleic anhydride using conventional and microwave heating. Although it has been reported that polyurethane obtained directly from CO has good water resistance, desirable mechanical properties and high thermal stability (Ogunniyi, 2006; Yeganeh and Hojati-Talemi, 2007), limitations have also been reported. Due to the low hydroxyl value and the presence of secondary hydroxyls, it has poor reactivity, a slow curing time, low flame retardancy and low miscibility with other components, and it shrinks readily. The modification of CO to increase its hydroxyl value and hard segment composition to improve the rigidity and crosslink density of the final PU products is needed (Baser and Khakhar, 1993; Pan et al., 2008). The conversion of triglyceride to monoglyceride via glycerolysis is an alternative to epoxidation for the introduction of hydroxyls to CO. By reaction with glycerol or pentaerythritol, two primary hydroxyl groups will be introduced to a monoglyceride molecule derived from CO (Noureddini and Medikonduru, 1997; McNeill and Berger, 1993). The benzene rings could be introduced to the modified castor-oil-based polyol (MCO) as hard segments by esterification with phthalic anhydride. Thus, in the present work, four different polyols were obtained from CO and its immediate derivative from the alcoholysis of CO and reaction with phthalic anhydride. The flame retardant ammonium polyphosphate (APP) was added to the foam to investigate the flame retardancy of PUFs. The properties and flame retardancy of PUFs prepared from MCO were also characterised and compared with traditional PUF based on the polyester polyol PS-3152. The purpose of this study was to synthesise modified castor-oil-based polyols with an appropriate proportion of hard and soft segments and good miscibility with polyether polyols and an environmentally friendly low-VOC foaming agent cyclopentane. In turn, the development of such materials would expand the potential applications of PUFs and polyurethanebased CO, especially in flame retardancy. 2. Experimental 2.1. Materials Castor oil (hydroxyl value: 163 mg KOH/g) was purchased from Aldrich Chemical Co., Inc. (USA). Polyester polyol (PS3152) was kindly provided by Stepan Jinling Chemical Co., Ltd.

(Nanjing, China). Polyether polyol 4110 (sucrose and diethylene glycol as initiator, hydroxyl value: 403 mg KOH/g, viscosity at 25 ◦ C: 3600 mPa s) was obtained from Jiangsu Qianglin Bio-energy Co., Ltd. PM-200 as the polyaryl polymethylene isocyanate (PAPI) was obtained from Yantai Wanhua Polyurethane Co., Ltd. (Shandong, China). The NCO group content was 30.3% by weight. N,N-dimethyl cyclohexylamine (DMCHA) was used as a catalyst and was kindly supplied by Aladdin Chemical Co., Ltd. Polysiloxane-polyether copolymer (AK8805) was used as the surface-active agent and purchased from Jiangsu Maysta Chemical Co., Ltd. APP was a product of the Zhenjiang Xingxing Flame Retardant Co., Ltd. 2.2. Preparation of castor-oil-based polyols Approximately 20–40 g of glycerol or pentaerythritol was placed in a 250-ml three-neck round-bottom flask equipped with a watercooling condenser and a thermometer. Accurately weighed CO was added to the flask in a mole ratio of 1/2. Ca(OH)2 (0.2 g) was then added into the mixture as a catalyst. The alcoholysis reactions were carried at 230 ◦ C for 3 h under an N2 atmosphere. The alcoholysis reactions with glycerol and pentaerythritol gave products labelled as MCO1 and MCO2, respectively. The alcoholysis products were then modified by phthalic anhydride at a OH/COOH ratio of 3:1 with tetrabutyltitanate as the catalyst. The reaction was maintained at 180 ◦ C for 2 h, increased to 220 ◦ C at a rate of 10 ◦ C/h and then kept at this temperature. The reaction was stopped when the acid value of the reactants was less than 1.5 mg KOH/g. Two MCO-based polyols (MCO3 and MCO4) could be prepared from MCO1 and MCO2, respectively. The reaction schemes of the alcoholysis and condensation are shown in Fig. 1. 2.3. Preparation of castor-oil-based PUFs Table 1 lists the composition of castor-oil-based rigid PUFs. All materials except for isocyanate were first mixed well in a plastic beaker. Next, isocyanate was added into the beaker and mixed at 3000 rpm for 8 s. To measure the foaming behaviour of the foaming system, approximately 35 g of the total formulation was poured into a 500-ml beaker. The rest of the total formulation was quickly poured into an open mould (200 mm × 200 mm × 300 mm) to produce free-rise foam for the measurement of the physical properties. After the rising was completed, the foams were kept in an oven at 70 ◦ C for 24 h to complete the polymerisation reaction. Different specimen sizes were cut from the cured foam to assess its physical mechanical properties. The final dimensions of the specimens were achieved by rubbing the foams with fine emery papers. 2.4. Characterisations of the MCOs and PUFs 2.4.1. Moisture content, acid value, hydroxyl value and relative content of primary hydroxyl The moisture content of the MCOs was determined by the Karl Fischer method using a CBS-1A model moisture content meter (Beijing Chao Sheng Co., China). Table 1 Composition of investigated PUFs. Samples

PUF1

PUF2

PUF3

PUF4

PUF5

PUF6

PUF7

PUF8

MCO1 MCO2 MCO3 MCO4 PS-3152 AK8805 DMCHA Cyclopentane APP PM-200

100 0 0 0 0 2 1 25 0 110

0 100 0 0 0 2 1 25 0 110

0 0 100 0 0 2 1 25 0 110

0 0 0 100 0 2 1 25 0 110

0 0 0 0 100 2 1 25 0 110

0 100 0 0 0 2 1 25 10 110

0 0 0 100 0 2 1 25 10 110

0 0 0 0 100 2 1 25 10 110

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

137

O O

OH

O

HO

O O

OH

+

HO

O

OH

ºC

Ca(OH)2

OH

pentaerythritol

CO O

HO OH

O

OH

OH

2

O O

OH

O

O O

OH

MCO4

tetrabutyl titanate

HO HO

MCO2 O O

OH

O

O O

OH

O

OH

+

ºC

2 HO

Ca(OH)2

HO pentaerythritol

CO

O

O OH

HO

O HO HO

O HO

O

O O

O

O

O

tetrabutyl titanate

O

OH O

OH

OH

MCO1 O

MCO3

Fig. 1. Alcoholysis and condensation reaction schemes.

The acid value was measured according to GB 2895-82. Briefly, a mixture of 2 g of the polyol sample and 40 ml of alcohol was titrated with 0.5 N sodium hydroxide solution to the equivalent point using a pH meter. The acid value (mg KOH/g) was calculated according to Eq. (1). Acid value =

(C − B) × N × 56.1 , W

(1)

where C is the titration volume of the sodium hydroxide solution at the equivalent point (ml), B is the volume of the blank solution (ml), N is the normality of the sodium hydroxide solution and W is the weight of the polyol sample (g). The hydroxyl value was measured according to GB 7193.2-87. Approximately 1 g of the polyol sample was reacted with phthalate reagent (25 ml) at 105 ◦ C for 50 min, followed by the addition of 75 ml of alcohol. Next, the mixture was titrated with a 1 N sodium hydroxide solution to the equivalency point. The phthalate reagent was created by completely dissolving 150 g of phthalic anhydride and 24.2 g of imidazole in 1000 g of pyridine. The hydroxyl value (mg KOH/g) of the sample was calculated according to Eq. (2): Hydroxyl value =

(B − A) × N × 56.1 + Acid Value W

(2)

where A is the volume of the sodium hydroxide solution required for titration (ml), B is the volume of the blank solution (ml), N is the

normality of the sodium hydroxide solution and W is the weight of the polyol sample (g). The relative content of primary hydroxyl of polyol was determined by the 19 F NMR spectrum. All 19 F NMR experiments were performed on a BRUKER Advance spectrometer at 600 MHz, and the sample was dissolved in deuterated chloroform with a concentration of ∼1% (w/v), the method of test and analysis according to Wu et al. (1988). The relative content of primary hydroxyl groups in the polyols was calculated according to Eq. (3): Relative content of primary hydroxyl =

A × 100% A+B

(3)

where A is the relative area of primary hydroxyl and B is the relative area of secondary hydroxyl. 2.4.2. Fourier transform infrared (FTIR) spectroscopy FTIR spectra were recorded on a Nicolet Magna-IR 550 instrument. Solid PUF samples were ground into powder, mixed with KBr and pressed into a pellet. Liquid samples were sandwiched between two plates of KBr. The spectra were collected at a resolution of 4 cm−1 using 64 scans in the range of 4000–400 cm−1 . 2.4.3. Gel permeation chromatography (GPC) GPC was performed using a column set comprised of two 5 ␮m × 30 cm × 0.78 cm Phenogel mixed bed columns (Phenomenex). Tetrahydrofuran (THF) was used as the mobile phase at

138

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

a flow rate of 1 ml/min. Monodispersed polystyrene standards were used for instrument calibration. The column temperature was maintained at 35 ◦ C. The samples were dissolved in THF (0.2 wt%) and filtered before injection. The data were analysed by Polymer Laboratories Caliber software. 2.4.4. Thermal analysis The thermal analysis of MCOs and PUFs was carried out on a thermo-gravimetric analyser (STA 409, NETZSCH, Germany). Approximately 1–5 mg of each sample was encapsulated in an aluminium pan and heated at 120 ◦ C for 2 min to remove any absorbed moisture. Next, they were heated from 50 ◦ C to 800 ◦ C at a heating rate of 10 ◦ C/min under a nitrogen atmosphere. 2.4.5. Physical and mechanical properties The densities of PUF samples were measured according to ASTM D1622-03. The size of the specimens for density measurements was 30 × 30 × 30 (mm × mm × mm), and the average value of three samples was reported. The compressive strength of the PUF samples was recorded at 10% strain parallel to foam rising according to ASTM D1621-00 under ambient conditions. The size of the specimen for compressive strength testing was 50 × 50 × 25 (mm × mm × mm), and the strain rate was fixed at 2.5 mm/min. The average strength of five measurements per sample was reported. 2.4.6. Thermal conductivities Thermal conductivities of PUFs were measured using a thermal conductivity analyzer (DZDR-PL) according to the ASTMC518 standard test method. The size of the specimen was 50 mm × 50 mm × 20 mm (length × width × thickness). 2.4.7. Cone calorimetry measurements The cone calorimetry test was performed using an FTT2000 cone calorimeter according to ISO5660-1. Each 20 × 100 × 100 (mm × mm × mm) specimen was wrapped in aluminium foil and exposed horizontally to an external heat flux of 35 kW m−2 . All samples were run in duplicate, and the average value was reported. 3. Results and discussion 3.1. Characterisation of castor-oil-based polyols 3.1.1. Physical properties The alcoholysis of CO by glycerol and pentaerythritol occurs in the presence of Ca(OH)2 at elevated temperature. In this stage, one triglyceride can be theoretically converted to 3 monoesters by glycerol or pentaerythritol. The result is an increase in the number of hydroxyl groups per molecule. However, the products were a mixture of monoesters and diesters as well as the unreacted triglyceride and free glycerol or pentaerythritol in practice. The alcoholysis products could react with phthalic anhydride to convert the free hydroxyls to phthatic half-esters. Phthalic anhydride was introduced to increase the molecular weight and hard segment composition. The use of pentaerythritol instead of glycerol in the same synthetic route, shown in Fig. 1, offers certain advantages, such as the presence of more hydroxyl groups and therefore more reactive sites for condensation, which should result in a higher cross-link density for the final polymers. Ca(OH)2 was reported to be an effective catalyst for the glycerolysis of castor oil. It forms a soap with the free fatty acids in the oil and increases the solubility of glycerol or pentaerythritol in the oil and the monoester content and reduces the triglyceride content. For rigid PUFs, the required of hydroxyl value of polyols is between 200 and 500 mg KOH/g. As shown in Table 2, the hydroxyl values of the alcoholysis products (MCO1 and MCO2) were higher than those of the condensation products (MCO3 and MCO4).

Table 2 Physical properties of CO-based polyols and PS-3152. Samples

Hydroxy value (mg/g KOH)

Acid value (mg/g KOH)

Viscosity (mPa s 25 ◦ C)

Moisture content (%)

MCO1 MCO2 MCO3 MCO4 PS-3152

418 464 270 308 310

0.80 0.64 1.21 1.49 1.40

1120 3150 6240 10,800 2400

0.11 0.08 0.15 0.l2 0.10

Because the pentaerythritol has one more hydroxyl than the glycerol, the hydroxyl values of MCO1 and MCO2 were 418 and 464 mg KOH/g, respectively. These values were lower than the theoretical hydroxyl value because of secondary reactions including the formation of ether bonds between hydroxyl groups, producing double bonds between the hydroxyl group and the H on the ␣-C. The hydroxyl value of MCO4 was similar to that of the commercial polyol PS-3152. Overall, the acid values of the MCOs were below 1.50 mg KOH/g. The acid values of MCO1 and MCO2 were slightly lower than those of MCO3 and MCO4 because the former contain unreacted phthalic anhydride. The acid values of MCO3 and MCO4 were similar to that of the commercial polyol PS-3152. Moisture content is a very important parameter for polyols, as it interferes with the foam formulation and properties of PUFs. It is known that water, as a chemical foaming agent, can react with isocyanate to increase the cell size and produce thinner walls, which leads to a decrease in the mechanical properties. Therefore, the moisture content of MCOs is required to be less than 0.15%. As shown in Table 2, the viscosity of the MCOs increased from MCO1 (1120 mPa s) to MCO4 (10,800 mPa s). The viscosities of MCO2 and MCO4 modified with pentaerythritol were higher than those of MCO1 and MCO3 modified with glycerol, respectively. Because MCO2 and MCO4 had a greater average hydroxyl functionality, they readily formed a macromolecular network structure in the esterification process. The miscibility of polyol with other components, especially the physical blowing agent and other polyols, is very important for preparing PUFs. This property affects the cell size, mechanical properties and thermal conductivity of the PUFs. Due to the ozone depletion potential (ODP) and global warming potential (GWP) problems of CFC and HCFC blowing agents, halogen-free aliphatic hydrocarbons and environment-friendly cyclopentane (ODP and GWP equal to 0) are the most promising and representative candidates. The miscibility of polyester polyols with polyether polyols has always been problematic. Therefore, the miscibility of MCOs and PS-3152 with commercial polyether polyols was investigated. The MCOs were mixed with 4110 (1/1, w/w) and cyclopentane (7/3, w/w) at 5 ◦ C for 24 h to observe their miscibility. The results showed that the MCO mixtures were transparent and had good miscibility with polyether polyol 4110 and cyclopentane. However, the mixtures of PS-3152 were delaminated. The relative content of primary hydroxyl of pentaerythritol alcoholysis of castor oil MCO2 and the corresponding polyester polyol MCO4 were determined by the 19 F NMR spectrum (seen in Fig. 4). And, the relative contents of primary hydroxyl groups and secondary hydroxyl groups in the polyols have been calculated. The results showed that the relative contents of primary hydroxyl of MCO2 and MCO4 were 74.4% and 83.3%, respectively. Due to the higher reactivity of primary hydroxyl groups, so that the foaming reaction rate of polyester polyol of castor oil based PUFs would be higher than those of alcoholysis polyol of castor oil. 3.1.2. Foaming behaviour The foaming behaviour of PUFs is important in the calculation of the required proportions of each substance in the foam

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143 Table 3 Foaming behaviour parameters of foam for polyols. Samples

Cream time (s)

Full cup time (s)

String time (s)

Tack free time (s)

End of rise time (s)

PUF1 PUF2 PUF3 PUF4 PUF5

38 36 28 26 28

45 42 38 37 39

51 51 48 46 45

66 62 53 52 50

94 90 65 60 55

production to assure a completely filled mould. The foaming behaviour parameters of cup foam for polyols are listed in Table 3, including the cream times, full cup times, string times, tack-free times and end of rise times of the different polyols (MCOs and PS3152). It can be seen that the foaming reaction rates of PUF3 (65 s), PUF4 (60 s) and PUF5 (55 s) were slightly faster than those of PUF1 (94 s) and PUF2 (90 s). This difference could be due to the presence of primary hydroxyl groups at the end of the macromolecular chains. In general, the reactivity of a foam system mainly depends on the hydroxyl activity of polyols. The effect of diffusion caused by the viscosity on the initial speed of the reaction is small. The cream times of PUF 3 (28 s) and PUF 4 (26 s) were shorter than those of PUF 1 (38 s) and PUF 2 (36 s), whereas the foaming rate and wire drawing time of the polyester-based foam systems were approximately the same, indicating that their reactivities were almost equivalent. In the early stage of the reaction, the reactivity was mainly determined by the activity of polyol hydroxyl. However, the viscosity and the benzene ring content also slightly affected the reaction. When the system reached the gel point, the reaction was controlled by diffusion due to the rapid increase of the viscosity. Because of the rigidity of the benzene rings, the motion of the macromolecular chains in polyol-based MCOs was limited. Therefore, the curing time of PUF-based MCOs was longer than that of PUF5. The MCO-based foams exhibited viscoelasticity for a longer period of time, and the cells were more easily deformed. 3.1.3. GPC The chromatogram for the initiator CO was also provided for comparison (Fig. 2). Different chromatograms were observed for the alcoholysis and condensation products of CO. The data clearly show that the triester peak becomes much smaller after alcoholysis and condensation. The alcoholysis products are mainly composed of the monoester and diester; a small peak at 14.0 min was observed due to slightly higher molecular weight of the oligomers relative to the polyol. The yields of the alcoholysis products of MCO1 and MCO2 were 85.9% and 81.7%, respectively. The peak observed at

139

14.6 min was attributed to the CO, while the peaks at approximately 15.2 and 16.1 min corresponded to the diester and monoester, respectively. The diester and monoester contents were 33.0% and 42.8% for MCO1 and 33.8% and 34.3% for MCO2, respectively. The peak at approximately 14.0 min was attributed to ester products with weights ranging from 1000 to 4000. The chromatograms of MCO3 and MCO4 showed that the monoester and diester contents significantly decreased and the higher-molecular-weight (2200–4600) content increased dramatically by 78.7% and 86.7% after condensation for MCO3 and MCO4, respectively. For MCO3, the relative content of monoester decreased from 42.8% to 2.4%, which is ascribed to the condensation with phthalic anhydride. The diester content was reduced from 33.0% to 7.2%. The degree of condensation of the diester was less than that of the monoester because of its steric hindrance. The triglyceride content was reduced from 14.1% to 10.4%, which could be partly explained by the condensation with phthalic anhydride. For MCO4, the monoester and diester were substantially involved in the condensation reaction with phthalic anhydride. It can be seen from Table 1 that the viscosity of MCO4 was higher than that of MCO3 because MCO4 with its greater average hydroxyl functionality, readily formed a macromolecular network structure in the condensation process. 3.1.4. FTIR The Fourier transform infrared (FTIR) spectra of MCO and CO are shown in Fig. 3. The spectra of all of the CO-based polyols were similar. MCO showed a strong, broad band between 3600 cm−1 and 3200 cm−1 . The intensity of this band increased after alcoholysis and condensation because of the introduction of more hydroxyls. The band at 1730 cm−1 can be ascribed to ester carbonyl. However, it shifted to 1740 cm−1 and increased in intensity due to the introduction of the benzene ring after condensation in MCO3 and MCO4. In addition, the absorption peak at 1280 cm−1 for MCO3 and MCO4 is attributed to the stretching vibration of C O end hydroxyl in ester bonds. The bond at 1100 cm−1 is assigned to the characteristic absorption of the secondary hydroxyl of CO. The strong absorption peak at 1046 cm−1 for the MCOs is due to the primary hydroxyl after alcoholysis. No such peak was observed in the spectrum of CO. 3.2. PUFs 3.2.1. Physical and mechanical properties The physical properties of rigid PU block foams prepared with castor-oil-based polyols and PS-3152 are listed in Table 4. For

150

CO

MCO4 MCO3

diester

50

monoester

triester

MCO2

0

MCO1

MCO1 MCO2

Transmmittance(%)

Response/mW

100

MCO3 MCO4

-50

CO -100 10

12

14

Retention time/min Fig. 2. GPC results for CO and MCOs.

16

18

4000

3500

3000

2500

2000

1500

Wavenumbers (cm-1) Fig. 3. Infrared spectra of CO and MCOs.

1000

500

140

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

Primary hydroxyl

Primary hydroxyl

Secondary hydroxyl

Secondary hydroxyl MCO4

MCO2 -75.5 ppm

-76.0

-75.0

-75.5

-76.0

ppm Fig. 4.

19

F NMR spectrum of MCO2 and MCO4.

similar foam densities, the 10% compression strengths in the parallel and vertical directions (relative to the foaming rise) for PUF3 (196.5 kPa) and PUF4 (198.8 kPa) are higher than those for PUF1 (172.6 kPa) and PUF2 (178.8 kPa) because of the introduction of harder segments in the form of benzene rings. Table 4 also indicates that the 10% compression strengths in both directions for PUF3 and PUF4 are also higher than that for PUF5 (192.3 kPa). In this case, it seems that PUF3 and PUF4 were composed of more flexible segments than PUF5, enhancing the toughness of the PUFs and giving rise to better mechanical properties. However, this enhancement may also be due to the density difference (38.1 kg/m3 for PUF4, 38.3 kg/m3 for PUF4 and 37.5 kg/m3 for PUF5). Furthermore, the dimensional stability of PUF3 and PUF4 foam systems is similar or somewhat better than that of the control system, especially at high and low temperatures (100 and −30 ◦ C). All of these experimental results indicate that the reasonable distribution of soft and hard segments of castor-oil-based polyester polyols not only enhanced the mechanical properties but also improved some general physical properties of the PU foams. Table 4 shows the thermal conductivities of different PUF samples. In general, the thermal conductivity of PUF depends on cell size, ratio of close to open cell and the miscibility of polyol with blowing agent cyclopentane. It can be seen from Table 4, the thermal conductivities of MCO-based PUF samples are 0.021–0.022 W/(m K), which are lower than that of PS-3152 based PUF5. The possible reason is that the MCOs-based polyols have good miscibility with cyclopentane than polyester polyol PS-3152.

3.2.2. TGA The TGA curves of polyols and PUFs are presented in Figs. 5 and 6, respectively. The onset temperatures of the MCOs were higher than that of polyol PS-3152, apparently. In general, for polyols, their thermal decomposition primarily occurred in the range of 210–475 ◦ C. MCO1 and MCO2 have lower onset temperatures than

CO MCO4 MCO3 MCO2 MCO1 PS-3152

100 TGA Weight (%)

-75.0

50

0

200

400

600

800

Temperature/ Fig. 5. TGA curves of CO, MCOs and PS-3152.

MCO3, MCO4 and CO, probably because their structures contain more hydroxyl groups. In addition, MCO3 and MCO4 have more hard segments, including benzene rings and ester bonds. CO had the highest decomposition temperature of 353 ◦ C. For PUFs, there are two stages in the pyrolysis process for these systems. The weight loss in the first stage was dominated by the degradation of the polyol components and that in the second stage was governed by the degradation of the isocyanate components. The onset temperatures for MCO1-based PUF1 (280 ◦ C) and MCO2-based PUF2 (280 ◦ C) were lower than those of MCO3-based PUF3 (300 ◦ C) and MCO4-based PUF4 (300 ◦ C). However, they were higher than that of PS-3152-based PUF5 (258 ◦ C). The thermal stability of the PUFs at this stage was in accordance with the corresponding polyols. At 50%

Table 4 Physical properties of rigid PUFs.a Samples

Thermal conductivity (W/(m K))

Core density (kg/m3 )

10% compression strength (kPa) Parallel to foam rise direction

Vertical to foam rise direction

172.6 178.8 196.5 198.8 192.3

56.9 64.7 88.4 89.7 82.8

Dimensional stability (%)b L

W

H

L

3.12 2.02 1.13 1.02 1.42

−1.65 −1.05 0.24 0.45 −0.65

−3.61 −2.17 −0.32 −0.12 −0.61

(100 ◦ C) PUF1 PUF2 PUF3 PUF4 PUF5 a b

0.022 0.022 0.021 0.022 0.023

37.2 37.6 38.1 38.3 37.5

4.16 3.36 1.46 0.87 2.14

All foams were prepared at ambient temperature (25 ◦ C) and tested after storing at ambient temperature for two days. W, L, and T, wide, long, and thick direction of foam, respectively.

W

H

−1.43 −0.83 −0.23 −0.11 −0.03

1.63 1.25 0.12 0.34 0.63

(−30 ◦ C)

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

141

14 100

12 10 2

THR (MJ/m )

Weight(%)

PUF4 PUF3 PUF2 PUF1 PUF5

50

8 6

PUF6 PUF7 PUF8

4 2 0 -2 0

400

50

150

200

Time (s)

800

Temperature (

100

)

Fig. 8. Total heat release of PUF6, PUF7 and PUF8.

Fig. 6. TGA curves of PUFs.

mass loss, the temperatures of MCOs and CO were higher than that of PS-3152. The temperatures for PUF2 (449 ◦ C) and PUF3 (448 ◦ C) were slightly higher than that of PUF1 (440 ◦ C) and lower than that of PUF4 (463 ◦ C). The temperatures of MCO-based PUFs were significantly higher than that of PUF5. The reason was that the long fatty chain had a better thermal stability than that of PS-3152 at this stage. The residues (1.73–1.82%) of MCO3, MCO4 and PS-3152 CO at 800 ◦ C were higher than those of MCO1, MCO2 and CO (1.23–1.37%). The residue of PUF4 (15.68%) was higher than those of PUF1 (9.15%), PUF2 (10.95%), PUF3 (12.85%) and PUF5 (15.56%). This is due to a barrier effect of the impact char layer and introduction of benzene ring hard segments could enhance the thermal stability of the PUFs. The ester bonds, benzene and isocyanate rings chain segments with higher thermal stability separated and limited the decomposition of those segments with lower-thermal-stability chain segments, thereby increasing the thermal stability of the polyols and PUFs. Castor-oil-based polyols were suitable for the preparation of polyurethane foam with a higher decomposition temperature. 3.2.3. Cone calorimetry measurements Cone calorimetry was extensively employed to investigate the effect of the proposed flame retardant systems on the behaviour

of polyurethane foam when subjected to a heat flow because its results correlate well with those obtained from large-scale fire tests and can be used to predict the combustion behaviour of materials. Cone calorimetry, which is based on the oxygen-consumption principle, is a small-scale test that provides results in good agreement with those of large-scale flame tests (Lefebvre et al., 2004). Cone calorimetry allows the quantitative analysis of the flammability of materials by investigating such parameters as time to ignition (TTI), heat release rate (HRR), total heat release (THR) and mass loss rate (MLR). A commercial fire retardant, ammonium polyphosphate (APP) with a high phosphorus content (31.0–32.0%) and nitrogen content (14.0–15.0%), was added to the PUFs to investigate their flame retardant effectiveness. The fire behaviours of PUF6, PUF7 and PUF8 with the addition of APP were characterised by cone calorimetry. The detailed data are illustrated in Figs. 7–9 and Table 5. The TTI (time to ignition) and combustion time are important flame-retardancy parameters for polymer materials. It can be seen from Table 5 that the TTIs of the PUF6 (3 s) and PUF7 (3 s) samples were the same as that of PUF8 (3 s). This could be attributed to the cellular structure of the PUF. It is noteworthy that the combustion times of the PUF6 and PUF7 samples increased from 125 s to 210 s and 155 s compared to that of PUF8, which indicated that the APP flame retardant can produce better synergistic flame retardant effects with the castor-oil-based PUF than with PS-3152. The heat release rate (HRR) is the most important parameter for evaluating fire safety. Figs. 7 and 8 show that all PUFs exhibited a

100

250

2

HRR (kW/m )

PUF6 PUF7 PUF8

PUF6 PUF7 PUF8

200

Mass (%)

150 100

50

50 0 0

0

50

100 Time(s)

150

Fig. 7. Heat release rates of PUF6, PUF7 and PUF8.

200

0

50

100

150

Time (s) Fig. 9. Mass curves for PUF6, PUF7 and PUF8.

200

142

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143

Table 5 Cone calorimetry data for APP-filled PUFs. Samples

TTI (s)

Combustion time (s)

P-HRR (kW/m2 )

THR (MJ/m2 )

Char residue (%)

PUF6 PUF7 PUF8

3 3 3

210 155 125

226.63 188.57 176.51

12.81 10.60 9.08

13.50 18.63 19.11

sharp peak on the HRR curve approximately 20 s after ignition, with a heat release rate as high as 226.63 kW/m2 for PUF6, 188.57 kW/m2 for PUF7 and 176.51 kW/m2 for PUF8. Furthermore, it can be seen that the HRR curve changes from one peak for PUF6 to two peaks for PUF7, which is similar to that of PUF8. Thus, the mechanism of the thermo-oxidation degradation of PUF7 was similar to that of PUF8 and different from that of PUF6. The first peak is assigned to the development of the intumescent protective char. The HRR curve then plateaus in some cases in which the increase in HRR is suppressed because of the presence of the efficient protective char. The second peak is due to the gradual degradation of the protective layer as the sample is continuously exposed to heat. The formation of a new protective char in some formulations was observed as well. The total heat released (THR) by combustion is a function of time per unit area and is calculated by the integration of the heat release over a given time. From Fig. 8 and Table 5, it can be seen that the THR of PUF6, PUF7 and PUF8 are 12.81, 10.60 and 9.08 MJ/m2 , respectively. Overall, these values are relatively low, and the results show that APP can produce better synergistic flame-retardant effects with phthalic anhydride polyester polyols than with long-chain fatty polyols. When the temperature was above 300 ◦ C, the flame retardant ammonium polyphosphate (APP) decomposes into phosphoric acid and ammonia. Phosphoric acid, with a strong dehydration effect during heating, promoted the formation of char, which together with nitrogen and hard segments of PUFs formed a compact carbonaceous layer that acted as a physical protective barrier for heat transfer into the material, reducing the heat release. Ammonia, as a non-flammable gas, can delay the volatilisation loss of phosphorus compounds and accelerate the oxidation of phosphorus into char, preventing the oxygen from combusting and removing heat. Mass loss is another important parameter of polymer materials. The mass curves of PUF6, PUF7 and PUF8, shown in Fig. 9, reveal that the mass decreases from 15.8% to 21.0% when relative to that of pure RPUF. The char residues of PUF6, PUF7 and PUF8 increased by approximately 2.55%, 2.95% and 3.55%, respectively, compared to those of PUF2, PUF4 and PUF5 without the APP flame retardant, respectively. Therefore, APP can effectively promote the formation of a carbon layer by the benzene ring on PUF. The formation of denser char in greater quantities improves the flame retardancy. The cohesive, intumescent and impact char can suppress the flame and limit the heat and mass transfer from the polymer to the heat source, thereby preventing further decomposition and limiting the weight loss of the PUFs. 4. Conclusions CO-based polyols could be used as a raw material for preparing rigid polyurethane foams after alcoholysis and condensation. The foaming behaviours of the MCOs were similar to that of the commercial product PS-3152. The addition of a benzene ring structure was found to increase the reaction activity of polyol and the thermal stability and physical properties of the final PUFs. The reasonable distribution of soft segments and hard segments in the PUFs prepared from castor-oil-based polyester polyol had better mechanical properties than the PS-3152-based PUF. The MCO-modified PUFs also exhibited much higher thermal stability during the pyrolysis

process. The cone calorimetry results showed that adding APP into PUFs significantly decreases the heat release rate (HRR), total heat release (THR) and mass loss of PUF samples. These test results indicate that APP has a better synergistic effect with phthalic anhydride polyester polyols than long-chain fatty polyols. The results herein may lead to the development of a new type of polyurethane foam using castor oil as raw materials.

Acknowledgements This work was supported by National 12th Five-year Science and Technology Support Plan (Grant No. 2012BAD32B05), the Science Foundation of Chinese Academy of Forestry (Grant No. CAFINT2013K01) and the National Natural Science Foundation (Grant No. 31300490).

References Baser, S.A., Khakhar, D.V., 1993. Castor oil-glycerol blends as polyols for rigid polyurethane foams. Cell. Polym. 12 (5), 390–401. Belgacem, M.N., Gandini, A., 2008. Monomers, Polymers and Composites from Renewable Resources. Elsevier, Amsterdam. Campanella, A., Bonnaille, L., Wool, R., 2009. Polyurethane foams from soyoil-based polyols. J. Appl. Polym. Sci. 112, 2567–2578. Chian, K.S., Gan, L.H., 1998. Development of a rigid polyurethane foam from palm oil. J. Appl. Polym. Sci. 68 (3), 509–515. Corcuera, M.A., Rueda, L., Fernandez d’Arlas, B., Arbelaiz, A., Marieta, C., Mondragon, I., Eceiza, A., 2010. Microstructure and properties of polyurethanes derived from castor oil. Polym. Degrad. Stab. 95 (11), 2175–2184. Hablot, E., Zheng, D., Bouquey, M., Avérous, L., 2008. Polyurethane based on castor oil: kinetics, chemical, mechanical and thermal properties. Macromol. Mater. Eng. 293, 922–929. Hatakeyama, H., Hatakeyama, T., 2005. Environmentally compatible hybrid-type polyurethane foams containing saccharide and lignin components. Macromol. Symp. 224 (1), 219–226. Hatakeyama, H., Nakayachi, A., Hatakeyama, T., 2005. Thermal and mechanical properties of polyurethane-based geocomposites derived from lignin and molasses. Composites Part A 36 (5), 698–704. Hatakeyema, H., Tanamachi, N., Matsumura, H., Hirose, S., Hatakeyama, T., 2005. Bio-based polyurethane composite foams with inorganic fillers studied by thermogravimetry. Thermochim. Acta 431 (1-2), 155–160. Krämer, R.H., Zammarano, M., Linteris, G.T., Gedde, U.W., Gilman, J.W., 2010. Heat release and structural collapse of flexible polyurethane foam. Polym. Degrad. Stab. 95 (6), 1115–1122. Kurimoto, Y., Takeda, M., Koizumi, A., Yamauchi, S., Doi, S., Tamura, Y., 2000. Mechanical properties of polyurethane films prepared from liquefied wood with polymeric MDI. Bioresour. Technol. 74, 151–157. Kurimoto, Y., Takeda, M., Doi, S., Tamura, Y., 2001. Network structures and thermal properties of polyurethane films prepared from liquefied wood. Bioresour. Technol. 77, 33–40. Lefebvre, J., Bastin, B., Le Bras, M., Dusquesne, S., Ritter, C., Paleja, R., Poutch, F., 2004. Flame spread of flexible polyurethane foam: comprehensive study. Polym. Test. 23, 281–290. Maznee, T.N., Norin, Z.K.S., Ooi, T.L., Salmiah, A., Gan, L.H., 2001. Effect of additives on palm-based polyurethane foams. J. Oil Palm Res. 13 (2), 7–15. Mazo, P., Estenoz, D., Sponton, M., Rios, L., 2012. Kinetics of the transesterification of castor oil with maleic anhydride using conventional and microwave heating. J. Am. Oil Chem. Soc., http://dx.doi.org/10.1007/s11746-012-2020-3. McNeill, G.P., Berger, R.G., 1993. Enzymatic glycerolysis of palm oil fractions and a palm oil based model mixture: relationship between fatty acid composition and monoglyceride yield. Food Biotechnol. 7, 75–87. Niranjan, K., Sravendra, R., Jae, W.C., 2009. Synthesis and characterization of castor oil-modified hyperbranched polyurethanes. J. Appl. Polym. Sci. 112, 736–743. Noureddini, H., Medikonduru, V., 1997. Glycerolysis of fats and methyl esters. J. Am. Oil Chem. Soc. 74 (4), 419–425. Ogunniyi, D.S., 2006. Castor oil: a vital industrial raw material. Bioresour. Technol. 97 (9), 1086–1091. Ogunniyi, D.S., Fakayejo, W.R.O., Ola, A., 1996. Preparation and properties of polyurethanes from toluene diisoyanate and mixtures of castor oil and polyol. Iran. Polym. J. 5, 56–59. Oprea, S., 2010. Dependence of fungal biodegradation of PEG/castor oil-based polyurethane elastomers on the hard-segment structure. Polym. Degrad. Stab. 95, 2396–2404. Pan, M.J., Wang, X.H., Yao, C., 2008. Synthesis and applications of vegetable oil-based polyols. Mod. Chem. Ind. 36 (1), 35–36. Petrovic, Z.S., 2008. Polyurethanes from vegetable oils. Polym. Rev. 48, 109–155. Sharna, V., Kundu, P., 2008. Condensation polymers from natural oils. Prog. Polym. Sci. 33, 1199–1215.

M. Zhang et al. / Industrial Crops and Products 59 (2014) 135–143 Siwayanan, P., Ooi, T.L., Shaari, N.Z.K., Ahmad, S., Wiese, D., Chua, M.C., 1999. Recent developments in palm-based polyols. In: Proceedings of the 1999 PORIM International palm oil Congress (Olechemicals), pp. 59–64. Somani, K.P., Kansara, S.S., Patel, N.K., Rakshit, A.K., 2003. Castor oil based polyurethane adhesives for wood-to-wood bonding. Int. J. Adhes. Adhes. 23, 269–275. Trevino, A.S., Trumbo, D.L., 2002. Acetoacetylated castor oil in coatings applications. Prog. Org. Coat. 44, 49–54. Wang, H.J., Rong, M.Z., Zhang, M.Q., Chen, H.W., Czigany, T., 2008. Biodegradable foam plastics based on castor oil. Biomacromolecules 9, 615–623.

143

Williams, C.K., Hillmyer, M.A., 2008. Polymers from renewable resources: a perspective for a special issue of polymer reviews. Polym. Rev. 48, 1–10. Wu, M.Y., Yong, Z.Y., Jin, Y., 1988. The determination of the relative content of primary hydroxyl groups of polyether polyols by 19 F-NMR. Acta Polym. Sin. 2, 149–151. Xu, Y., Petrovic, Z., Das, S., Wilkes, G., 2008. Morphology and properties of thermoplastic polyurethanes with dangling chains in ricinoleate-based soft segment. Polymer 49, 4248–4258. Yeganeh, H., Hojati-Talemi, P., 2007. Preparation and properties of novel biodegradable polyurethane networks based on castor oil and poly(ethylene glycol). Polym. Degrad. Stab. 92 (3), 480–489.