The study of mechanical behavior and flame retardancy of castor oil phosphate-based rigid polyurethane foam composites containing expanded graphite and triethyl phosphate

Polymer Degradation and Stability 98 (2013) 2784e2794

Contents lists available at ScienceDirect

Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab

The study of mechanical behavior and flame retardancy of castor oil phosphate-based rigid polyurethane foam composites containing expanded graphite and triethyl phosphate Liqiang Zhang, Meng Zhang, Yonghong Zhou*, Lihong Hu Institute of Chemical Industry of Forestry Products, Research Institute of New Technology, CAF, Nanjing 210042, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 August 2013 Received in revised form 10 October 2013 Accepted 20 October 2013 Available online 30 October 2013

The goal of this work was the synthesis of novel flame-retarded polyurethane rigid foam with a high percentage of castor oil phosphate flame-retarded polyol (COFPL) derived from renewable castor oil. Rigid flame-retarded polyurethane foams (PUFs) filled with expandable graphite (EG) and diethyl phosphate (TEP) were fabricated by cast molding. Castor oil phosphate flame-retarded polyol was derived by glycerolysis castor oil (GCO), H2O2, diethyl phosphate and catalyst via a three-step synthesis. Mechanical property, morphological characterization, limiting oxygen index (LOI) and thermostability analysis of PUFs were assessed by universal tester, scanning electron microscopy (SEM), oxygen index testing apparatus, cone calorimeter and thermogravimetric analysis (TGA). It has been shown that although the content of P element is only about 3%, the fire retardant incorporated in the castor oil molecule chain increased thermal stability and LOI value of polyurethane foam can reach to 24.3% without any other flame retardant. An increase in flame retardant was accompanied by an increase in EG, TEP and the cooperation of the two. Polyurethane foams synthesized from castor oil phosphate flameretarded polyol showed higher flame retardancy than that synthesized from GCO. The EG, in addition to the castor oil phosphate, provided excellent flame retardancy. This castor oil phosphate flame-retarded polyol with diethyl phosphate as plasticizer avoided foam destroy by EG, thus improving the mechanical properties. The flame retardancy determined with two different flame-retarded systems COFPL/EG and EG/COFPL/TEP flame-retarded systems revealed increased flame retardancy in polyurethane foams, indicating EG/COFPL or EG/COFPL/TEP systems have a synergistic effect as a common flame retardant in castor oil-based PUFs. This EG/COFPL PUF exhibited a large reduction of peak of heat release rate (PHRR) compared to EG/GCO PUF. The SEM results showed that the incorporation of COFPL and EG allowed the formation of a cohesive and dense char layer, which inhibited the transfer of heat and combustible gas and thus increased the thermal stability of PUF. The enhancement in flame retardancy will expand the application range of COFPL-based polyurethane foam materials. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Castor oil Expanded graphite Triethyl phosphate Flame retardant Polyurethane foam

1. Introduction Rigid polyurethane foams (RPUF) that are an interesting family of polymers, with excellent thermal insulation, electrical insulation, chemical resistance and toughness combined with good lowtemperature flexibility, are extensively used in insulation in refrigerators, construction materials, chemical pipelines [1,2], thermal insulation, space filling and other applications due to its excellent properties such as closed-cell structure, low thermal conductivity, low moisture permeability and high compressive

* Corresponding author. Tel.: þ86 25 85482520; fax: þ86 25 85482777. E-mail addresses: [email protected], [email protected] (Y. Zhou). 0141-3910/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.polymdegradstab.2013.10.015

strength. However, RPUF are highly flammable and susceptible to degradation upon exposure to elevated temperature during fire accident, which constitutes a serious concern and restricts its application. Numerous studies have aimed to improve the fire behavior and thermal stability of RPUF. One of the problems facing polyurethanes nowadays is their dependence on petroleum derivative products. In particular, polyurethanes, offering a broad variety of properties that are useful in different areas of applications, are very interesting materials that can be prepared from reactants obtained from renewable resources of wide availability. Due to oil crisis, governmental politics, global warming effects, legislation, economical factors and the growth of awareness towards environmental preservation, have led corporations and researchers to search for new processes, products and

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

alternative raw materials in the polymer industry that minimize damage to the planet. New materials should have the same or better physical and chemical properties than those produced from petroleum; at the same time, the cost performance should also be compatible with commercial materials in the market. In the last years, the increasing interest in searching for a sustainable chemistry is promoting the replacing of petroleum derived raw materials with renewable raw materials in the production of polymers [3e5]. In this way, bio-based materials obtained from renewable resources are receiving considerable attention for an increasing amount of applications [6e9] from a social, environmental and energy standpoint, with the increasing emphasis on issues concerning waste disposal and depletion of non-renewable resources. Substitution of agricultural oils for petrochemically derived feedstocks in polyurethane synthesis has been an area of intense research and development for several decades. PU based on renewable resources has generated worldwide interest, especially development of vegetable oil based polyurethanes [10e12]. Among the natural oils, castor oil is an important renewable resource and widely used as a starting material for many industrial products [13]. In recent years, castor oil has attracted a lot of attention because of its wide possible applications. Castor oil is a renewable raw material that has been attracting research efforts due to its use in coatings, adhesives, paints, sealants, encapsulating compounds [14,15]. Researchers have been developing new types of polyurethane, using castor oil as a precursor for bio-polyurethane and polyurethane foams. Castor oil is an abundant and renewable natural resource available in large quantities from castor-oil plant seed; it is a kind of relatively low cost material which offers a priori possibility of biodegradation. Bio-based materials derived from castor oil are used to synthesize natural polyols, which are used as raw materials in the preparation of bio-based elastomeric polyurethanes. The unique feature of castor oil is that it contains considerable amount of hydroxylated triacylglycerols, which are important ingredients for polyurethane polymer. PU obtained from castor oil has certain disadvantages, including: low hydroxyl number leading to low modulus materials, a slower rate of curing of secondary hydroxyl groups [16] and low flame retardancy. The use of castor oil (CO) as polyol replacement in polyurethane formulations has been reported by several authors. However, the original concentration of hydroxyl reactive groups is not enough to obtain rigid PU. To offset these disadvantages, the chemical modification of castor oil is considered. Castor oil is transesterified or alcoholyzed with polyhydroxy alcohols, most commonly by glycerol, pentaerythritol, triethanolamine and trimethylol propane. Transesterification leads to an increase in hydroxyl value of the system thereby rendering hardness to the product, while long chain fatty acids induce flexibility. So transesterification is a possible route of modification to increase the hydroxyl groups in the oil structure [17]. The inflammability of PUF is due to the absence of information about the real structure of castor oil, which is a complex 12hydroxy-9-cis-octadecenoic glyceride structure. The existence of ester group and alkyl makes it flammable. By considering these issues, it is necessary to improve the flame retardancy of PUFs that is realized through react with diethyl phosphate in order to evaluate their flame retardancy. In order to promote further utilization of castor oil in new fields of industrial products, it is crucial to control flame retardancy and hydroxyl value of products derived from castor oil. However, there is little study about castor oil-based flame-retarded polyols. Since polyurethane foams are, in general, flammable materials, many efforts have been directed to improve their flame retardancy. However, the additive flame retardants are usually easy to separate out from PUF, especially for micromolecular liquid flame retardants, which decreases the permanent flame retardancy. Therefore, a

2785

great deal of effort has been devoted to exploiting flame-retarded polyols for PUF, making the flame retardant element link to the PUF matrix by covalent bond. Polyurethane foam can be specifically modified in order to increase its flame retardancy by dispersing flame-retarded fillers and using modified polyols with flame retardant element such as halogen, phosphorus, nitrogen and silicon. Usually, the halogenated additives release corrosive, obscuring and toxic smoke, which pollutes environment, erode instruments, and even damages people’s health. As a result, the flame retardant additives which have good flame retardant efficiency and hardly pollute environment are particularly needed. Expandable graphite prepared from natural graphite by chemical treatment is a type of graphite intercalation compound. The excellent properties of expandable graphite kept most excellent characteristics of natural graphite and of its own, such as low price, overcame hard, abundance, permeability, electrical conductivity, increases of mechanical properties of polymers, high porosity and exchange surface, make it and its derivatives very useful as functional carbon materials that can be applied in various fields, such as conductivity polymer [18], airtight materials, oil absorbents, highpower batteries, electrodes, military materials [19,20], sealing, catalyzing mechanism, space flight military affairs, environmental protection etc. [21]. Another important property for expandable graphite is the flame retardancy. Some studies implied that EG could produce good fire-retardant properties for some polymers, such as polyolefins [22], polyurethane foam [23], coating [24], etc. In addition, EG after expansion can be used as biomedical materials due to its pore structure and absorptive capacity [25]. EG or modified EG is also another kind of typical flame retardant for a wide range of polymers, giving satisfactory fire retardancy in polyurethane. A number of studies have been conducted on expanded graphitereinforced flame retardant polyurethane foams. This is because expandable graphite obtained by partial exfoliation, while being cheap, also has a high aspect ratio, and it forms a “wormlike” structure at high flame retardancy. EG is a structural modification of graphite obtained from intercalated [26] or oxidized [27] graphite via thermal reduction. This treatment of the graphite results in a light worm-like structure. When exposed to a heat source, EG, occupies hundred times its initial volume and generates a voluminous structure, thus providing fire-retardant performance for the polymeric matrix [28]. The special layer structure of EG is treated with sulfuric acid, nitric acid or acetic acid, which are intercalated into the graphite crystal structure. As to the origin of the flame retardancy of EG, it has been established that a “worm-like” structure layer can be formed on the surface of the materials due to the expansion of EG during burning, and such a layer of graphite can prevent heat and oxygen entering the bulk [29], thus providing fire resistance to the polymeric matrix. Lower size EG particles led to lower volume expansion ratio and less flame retardant efficiency [30]. Because the boiling point of graphite is above 3000  C, EG can maintain its integrity in the flame zone and provide better fire protection than many other flame retardants. EG acts mainly in the condensed phase as a smoke suppressant and an insulator. If the expanded carbon layers are too unstable, EG needs to be combined with other flame retardants such as TEP to form stable intumescent layers. In the present work, castor oil phosphate flame-retarded polyols were synthesized and we chose commercial expandable graphite and TEP in order to ensure flame retardancy of polyurethane foam. Mechanical property, morphological characterization, limiting oxygen index and thermostability analysis of PUFs were assessed by universal tester, scanning electron microscopy, oxygen index testing apparatus, cone calorimeter and thermogravimetric analysis. It reveals that polyurethane foam prepared from COFPL, EG and TEP exhibits high LOI, excellent thermostability and mechanical property.

2786

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

2. Experimental 2.1. Materials Castor oil (industrial grade) was purchased from Sinopharm Chemical Reagent. Glycerol, hydrogen peroxide, phosphoric acid, triethyl phosphate and triethylamine were obtained from Nanjing Chemical Reagent Co., Ltd. Triphenylphosphine and sodium methoxide were from Shanghai Zhanyun Chemical Co., Ltd. The raw material employed in polyurethane foam was polymeric MDI (methane diphenyl diisocyanate), was obtained from Guangzhou Meixing Chemical Co., Ltd. Catalyst employed is: N,N-dimethyl cyclohexylamine (DMCHA) kindly supplied by Aladdin Chemical Co., Ltd.; surface active agent: polysiloxaneepolyether copolymer (AK8804) purchased from Jiangsu Maysta Chemical Co., Ltd.; blowing agent: blend of 1,1-dichloro-1-fluoroethane (HCFC-141B) and water. Castor oil phosphate was synthesized following the detailed procedure described elsewhere in our previous studies. Their chemical structure is given in Fig. 1. Triethyl phosphate (TEP), reagent grades, reagent grades, were purchased from Nanjing Chemical Reagent Co., Ltd. Expanded graphite (EG) giving an expansion rate of 200 mL/g, was supplied by Qingdao Laixi electronics factory. The largest particle size of EG is less than 100 mesh and its pH value is 7.2. 2.2. Synthesis of castor oil phosphate flame-retarded polyol A reaction kettle equipped with a mechanical stirrer, condenser pipe, thermometer and provision for nitrogen flushing was charged with dry castor oil 500 g (0.54 mol) and catalysts (sodium methoxide 0.18 g and triethylamine 3.57 g). The temperature was raised quickly with continued stirring and maintained at 200  C after continuous nitrogen for 30 min. Then 123.3 g (1.34 mol) of glycerol was taken in the reaction kettle. The temperature was maintained at 180  C for 3 h. GCO (100 g), formic acid (7.75 g) and phosphoric acid (0.25 g) were mixed in a four-necked round-bottom flask equipped with a tetrafluoroethylene stirrer, a thermometer and a condenser pipe. This was heated to 40  C in a water bath. After that, hydrogen peroxide was dropped to the reaction flask in 30 min. The mixture

was then heated to 60  C and stirred for 6 h. And then the reaction mixture was cooled to room temperature and washed to pH ¼ 7 with sodium hydroxide solution. The resulting product was dried over anhydrous sodium sulfate. 40 g of diethyl phosphate and triphenylphosphine (0.5 g) were taken in a 500 ml four-necked round bottom flask provided with a mechanical stirrer, thermometer and a water condenser. Then the reaction proceeded with continuous stirring at 70e75  C with persistent addition of epoxidized GCO (EGCO) until the acid value was below 1.0 mgKOH/g. The chemical reaction between EGCO and diethyl phosphate producing COFPL is shown in Fig. 1. 2.3. Preparation of PU foams Due to the plasticizing effect of high TEP content, their amount cannot exceed certain limits in foam formulations. Therefore, TEP content was fixed at under 15 g/100 g of polyols. Also PU rigid foam without flame retardant has been prepared as a reference material. Castor-based polyols were premixed at room temperature with a small amount of silicone oil AK8804 (surfactant), N,N-dimethyl cyclohexylamine (catalyst), HCFC-141B and water. The mixture was stirred with a propeller stirrer for 2 min at approximately 1500 rpm to ensure a homogeneous mix and mixed with MDI using the a propeller stirrer (1800 rpm) for 10 s at constant NCO/OH ratio 1.2:1 to guarantee that each OH group would be reacted, and poured into a steel box (15  15  15 cm), closed during the foaming. The obtained PU foams were put in an oven for complete cure for 24 h at 80  C in order to complete the polymerization reaction, before carrying out characterization. The compositions of investigated castor oil-based polyurethane foams are shown in Table 1. 2.4. Measurements Compression test: The compressive properties were tested with a CMT4000 universal testing machine (Sheng Zhen, China) according to GB/T8813-2008. At least three samples were tested to obtain average values. Size of the specimens was 50  50  50 mm (width  length  thickness).

Fig. 1. The synthesis of COFPL.

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

2787

Table 1 Composition of investigated polyurethane foams. Sample

PUF-1#

PUF-2#

PUF-3#

PUF-4#

PUF-5#

PUF-6#

PUF-7#

PUF-8#

PUF-9#

PUF-10#

COFPL (g) GCO (g) EG (g) TEP (g) AK8804 (g) DMCHA (g) Water (g) HCFC-141B (g) MDI (g)

0 100 0 0 2 1 0.5 25 120

0 100 10 0 2 1 0.5 25 120

0 100 20 0 2 1 0.5 25 120

0 100 30 0 2 1 0.5 25 120

100 0 0 0 2 1 0.5 25 120

100 0 30 0 2 1 0.5 25 120

100 0 10 15 2 1 0.5 25 120

100 0 15 10 2 1 0.5 25 120

100 0 20 10 2 1 0.5 25 120

100 0 30 5 2 1 0.5 25 120

Scanning electron microscopy study: PU foams were investigated with a scanning electron microscope (SEM) 3400N Hitachi Co. The specimens were mounted on an aluminum stub and sputter coated with a thin layer of gold to avoid electrostatic charging during examination. Fire retardancy test: Limiting oxygen index test was carried out according to the GB/T2406-1993 to determine the relative flammability of foam. Test specimen dimensions were 100  10  10 mm (length  width  thickness). Thermogravimetry: Thermogravimetry analysis was carried out using STA 409 PC/PG. Small amount of sample was placed in the platinum pan before it was put in the furnace. Then, the sample was heated from 30 to 800  C with a heating rate of 10  C/min. Cone Calorimetry Test: The cone calorimetry test was carried out by using a cone calorimeter FTT2000 according to ISO5 660-1. Each specimen, with the dimensions of 20 mm  100 mm  100 mm, was wrapped in aluminum foil and exposed horizontally to an external heat flux of 35 kW/m2. All samples were run in duplicate and the average value was reported.

384.2  C, indicating that COFPL is more thermal stability than GCO with the linkage of PeOeC. May be the first decomposition stage is contributed to the diethyl phosphate linked with PeC. When it was reached to above 210  C, the diethyl phosphate degradated and phosphinic acid was generated. Phosphinic acid makes COFPL dehydrate to produce carbonization char with abundant phosphorus carbonization zone and thus O2 and heat are prevented, making COFPL more stable. Carbon residue rate of GCO and COFPL is 0% and 4.47%, illustrating COFPL is more stable and has flameretardant function. 3.2. Effect of COFPL content on cell morphology

The TGA and DTG analysis of GCO and COFPL are illustrated in Figs. 2 and 3, respectively, and the parameters are summarized in Table 2. It can be noted that pristine GCO shows only one main step of weight loss. The onset decomposition temperature of 335.1  C is observed while COFPL has onset decomposition temperature near 340  C in the second stage. The maximum decomposition temperature of COFPL has similar result 406.3  C, while that of GCO is

The cell morphology is an important factor that affects the physicalemechanical properties of the polyurethane foam. Fig. 4 represents the morphologies of EG filled PUF with the same magnifications. Comparing with Fig. 4 (a, c, d, e), it can be seen that the cell shapes are approximately spherical, polyhedral and symmetrical as a whole in both the pristine foam and the EG filled foams. Regular cell shapes are because EG particles have dimensions similar to that of the cell and are located inside the cell, thus increasing the average cell size of the foam. Compared with the pristine foam, the EG filled foams containing 10, 20 and 30 wt% EG exhibit obviously enlarged cell size and more cells are opened varying the EG concentration from 10 to 30 wt% and there are some holes in the thin cell wall in PUF-4# formed due to local internal stress or unbalanced foam growth. The cell size becomes larger compared to the pristine PUF, which is a reasonable result because the addition of EG means the reduction of the polyurethane resin content. Larger cell size is because the increase in EG content causes the combination of small and large cell structures in the PUF. When

Fig. 2. TGA curves of GCO and COFPL.

Fig. 3. DTG curves of GCO and COFPL.

3. Results and discussion 3.1. Thermogravimetric analysis (TGA and DTG) of GCO and COFPL

2788

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

Table 2 The parameters of TGA and DTG of GCO and COFPL. Stage one

GCO COFPL

Stage two

Stage three

Onset ( C)

Tdmax ( C)

Rmax (%/min)

Onset ( C)

Tdmax ( C)

Rmax (%/min)

Onset ( C)

Tdmax ( C)

Rmax (%/min)

211.1

258.2

3.74

335.1 340.3

384.2 406.3

22.98 20.18

440.5

459.5

6.54

Fig. 4. SEM micrographs for PUF-1# (a), PUF-5# (b), PUF-2# (c), PUF-3# (d), PUF-4# (e), PUF-6# (f), PUF-4# (e00 ) magnified from (e) and PUF-6# (f00 ) magnified from (f).

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

2789

Fig. 5. SEM micrographs for PUF-2# (c0 ), PUF-3# (d0 ), PUF-4# (e0 ) and PUF-6# (f0 ).

the EG content reaches as high as 30 wt% in PUF-4#, the cell size becomes obviously heterogeneous and shows bubble collapse, as compared with PUF-1#, PUF-2# and PUF-3# that have the small level of EG. This result indicates that adding too much EG leads to ruptures of the cell walls. Sample PUF-6# with EG content 30 wt% is lesser to holes in the thin cell wall compared to sample PUF-4#. The SEM micrographs in Fig. 4(e,f) show that the shape and size of the cells are approximately spherical and symmetrical as a whole, but there are some evident signs of collapse or collisions in the cell system of EG/GCO/ PUF as the arrows shown in Fig. 4(e00 ) magnified from Fig. 4(e), which indicates incompatibility between the EG particles and the PUF matrix. Unlike the EG/GCO/PUF system, the cells of EG/COFPL/ RPUF (Fig. 4(f)) are more compact and there is less cell collapse and collision. When COFPL is blended into the PUF, no evident collapse or collisions happen in the cell system (Fig. 4(f)). There are many obvious gaps around the EG particles, indicating the incompatibility between the inorganic EG particles and the organic foam matrix. In contrast, hardly any gaps or cracks are found around the EG/COFPL/PUF matrix, indicating strong adhesion between the EG particles and the foam matrix. As EG particles are blended into RPUF in Fig. 4(f00 ) magnified from Fig. 4 (f), EG is dispersed in the matrix without any cracks and the matrix’s struts of the cells are nearly intact, thus the high interfacial interaction between EG and PUF matrix is originated from the high interfacial adhesion between EG particles and PUF. This is because of the existence of diethyl phosphate grafting in the castor oil phosphate chain, which as one type of surfactant is more sensitive to surface effects due to the mixed surface/bulk conduction mode. This surfactant effect makes connect more compact in the surface of expandable graphite and PU matrix. To further investigate the effect of expandable graphite on the char formation of PUFs during combustion, the morphologies of the char residues left after limiting oxygen index test are characterized

by SEM. Fig. 5 shows SEM of PUFs residues morphologies perpendicular to the heating direction after combustion in limiting oxygen index test. Seen from Fig. 5, the residual char contains many compact graphite layers which overlap each other. Due to the combustion effect, the graphite layers expand dramatically perpendicular to the heating direction and the distance of the graphite layers enlarges to produce worm-like structure, which is about several hundred times of that of expandable graphite, directly showing the expansion effect of EG on PUFs. Comparing with the images of c0 , d0 , e0 and f0 , the foam structure of PUF-2# is completely destroyed, following the generation of worm-like residual char layer, which implies the complete decomposition of PUF-2#. PUF-4# has more worm-like residual char layer than PUF-3# after combustion and the border shape of PUF-4# is smoother than PUF-3#, indicating flame retardancy is more excellent with the increase of EG content. From image f0 , it reveals that the residual char layer is more compact than other samples. This is because of the existence of diethyl phosphate grafting in the castor oil phosphate molecular chain, when it was burned, diethyl phosphate degradated and phosphinic acid was generated. Phosphinic acid makes PUF dehydrate to produce compact carbonization zone and this compact carbonization layer enters into worm-like residual char layer as a connected bridge between the gaps of worm-like residual char layer, making the worm-like residual char layer more compact. Not only char quantity but also char quality formed during combustion are crucial to the flame retardancy ability of PUFs. 3.3. Compression strength Given their significant application as foam materials, the compressive strength of PUFs is a critical property, the apparent density is a very important physical property which has great influence on the compression strength. The results of compression

2790

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

Table 3 Compression strength of polyurethane foams. Sample

PUF-1#

PUF-2#

PUF-3#

PUF-4#

PUF-5#

PUF-6#

PUF-7#

PUF-8#

PUF-9#

PUF-10#

Density (kg/m2) Compression strength (MPa)

36.8 0.2014

37.0 0.1781

37.1 0.1801

37.2 0.1980

36.7 0.2180

37.3 0.2212

37.1 0.1358

37.1 0.1397

37.2 0.1746

37.2 0.1792

strength and apparent density are summarized in Table 3. As shown in Table 3, the results show that the flame retardants significantly influence these properties. Higher EG loading results in higher density of EG PUFs. This is because the addition of flame retardants increases the PUF’s weight and so affects the degree of foaming. Compared with PUF-1#ePUF-4#, the compressive strength decreases drastically from 0.2014 MPa to 0.1980 MPa with the increase of EG loading from 0% to 30%, which is due to the incompatibility between EG and the polyurethane matrix. With the inclusion of expandable graphite, the compression strength of PUFs is sacrificed. The primary causes of the deterioration of the compression strength of the PUFs are the collapse of PUF’s cell because of the large size of EG particles and the poor interfacial adhesion between EG particles and polyurethane matrix. Additionally, PUF-5# displays more effective reinforcement of compression strength than PUF-1#. The reasons may be as follows. This significant improvement may be attributed to the increased hydroxyl value from the opened epoxy bond, which makes the hydroxyl value and degree of functionality increased. The enlargement of degree of functionality makes crosslink density to increase. It was confirmed that crosslinking system is formed and the degree of crosslinking is increased. The dominant effect is the extremely high crosslink density (caused by the reaction between increasing hydroxyl and isocyanate groups of polyurethane). These results indicate that stiffness of the foams has been improved through the chemical modification. However, it reveals an opposite conclusion compared with PUF5# and PUF-6#, the results show that the introduction of EG causes small increase of compression strength, thanks to the plasticizing effect of diethyl phosphate grafting in the castor oil phosphate molecular chain, reduces the increase in cell size and the surfactant effect makes the connect more compact in the surface of expandable graphite and PU matrix. 3.4. Flammability tests The results of these fire tests are summarized in Table 4. In case of non-FR treated PUF prepared from GCO, a LOI value of 20.1% was observed. It can be seen that the LOI increases linearly with increasing EG amount in GCO-based PUF systems, while in the inexistence of TEP. Higher content of EG in the PUFs decreases their burning rates and thus can be considered to offer better flame protection for the foams. A similar trend is observed in case of LOI test results of PUFs prepared from COFPL, where the addition of 30 wt% EG significantly raises the LOI value of the PUFs. The use of EG is very effective in enhancing LOI, thus showing effective flame retardancy action of GCO- or COFPL-based polyurethane systems. Additionally, PUF-5# displays more effective reinforcement of LOI than PUF-1#. This is because the existence of diethyl phosphate grafted in the polyol performs a good effect on flame retardation behavior of this self-cured PU system. Therefore, COFPL functions as

a phosphorus-containing reactive flame-retarded polyol for this self-curing PU system without any other flame retardant. In the same way, compared PUF-4# with PUF-6#, the oxygen index test shows that the LOI increases for PUF-6#. The presence of diethyl phosphate leads to an increase of about 28.3% from 24.5%, using COFPL as the polyol. The results obtained have clearly shown that the effectiveness of diethyl phosphate substituent group formed in the castor oil phosphate is good. Analyzing more deeply the char morphology by means of SEM (Fig. 5), it is possible to note a more compact char layer with the presence of “worms”, deriving from the expansion of EG and diethyl phosphate. We suppose that a synergistic effect could take place between diethyl phosphate and EG. In particular, foam prepared from COFPL containing 5 wt% TEP and 30 wt% EG exhibits the highest LOI value 29.7%, whereas the analogous EG (30 wt%) containing foam exhibits relatively lower LOI value 28.3%. Even if the improvements on LOI values for EG filled foams are satisfactory, for TEP/EG filled foams, the relative LOI is much greater than for EG filled foams. In general, the use of both fire retardants together leads to a synergistic effect: in the presence of a constant amount of TEP (10 wt%) and an increasing quantity of EG the LOI rises more than the case of an additive effect. We suppose that a synergistic effect could take place because of complementary mechanisms of fire retardancy of TEP and EG. In the same way, compared PUF-4# with PUF-6#, the oxygen index test shows that the LOI increases for PUF-6#. The presence of diethyl phosphate leads to an increase of about 28.3% from 24.5%, using COFPL as the polyol. The results obtained have clearly shown that the effectiveness of diethyl phosphate substituent group formed in the castor oil phosphate is the same. Analyzing more deeply the char morphology by means of SEM (Fig. 5), it is possible to note a more compact char layer with the presence of “worms”, deriving from the expansion of EG and diethyl phosphate. We suppose that a synergistic effect could take place between diethyl phosphate and EG. Based on the analysis above, one can conclude that EG/TEP/ diethyl phosphate system has a synergistic effect as a common flame retardant, which is in accordance to the results of the aforementioned flammability parameters. In addition, not only char quantity but also char quality formed during combustion are crucial to the flame retardancy ability of PUF. And EG is an excellent flame retardant adapting to COFPL- and GCO-based PUFs. To further evaluate the flame retardancy of PUF, the original GCO-based PUF, EG filled PUFs and EG filled COFPL-based PUF are respectively examined by cone calorimetric measurement and thermogravimetric analysis. 3.5. Cone calorimetric measurement Cone calorimetry test based on the oxygen consumption principle has been extensively employed for investigating the effect of

Table 4 The LOI values of polyurethane foams. Sample

PUF-1#

PUF-2#

PUF-3#

PUF-4#

PUF-5#

PUF-6#

PUF-7#

PUF-8#

PUF-9#

PUF-10#

LOI (%)

20.1

21.3

23.0

24.5

24.3

28.3

27.1

28.1

29.0

29.7

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

the proposed flame retardant systems on the behavior 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 behavior of materials in real fires. Table 5 shows the cone results of GCO/EG, COFPL/EG based PUFs. Because of the real fire condition, it is often the basis of escape time designing for one in a real fire. The longer the time to ignition (TTI) and the lower the PHRR, it is the better chance to reduce the loss and casualty in a real fire. From Table 5, TTI of PUF-6# increases to 6 s while that of PUF-4# is 5 s. PHRR of PUF-6# is 127.78 kW/m2, which is lower than that of PUF-4#, showing the better flame retardancy of the EG especially for COFPL-based PUF. As a whole, the above results indicate that the flame retardancy is improved for PUF by the addition of COFPL, and which is very important for the practical usage as halogen-free flame retardant PUFs. The curves of HRR (heat release rate) versus time for both samples which are recognized to quantify the size of the fire [31], are shown in Fig. 6. From Fig. 6 and Table 5, it can be clearly seen that EG filled PUF burns very slowly after ignition and a sharp single HRR peak appears with a peak heat release rate as high as 164.69 kW/m2 for GCO-based PUF. With the addition of COFPL, the flammability of PUF is obviously restrained, it could be found that there are two peaks or a combustion process after an HRR peak for the COFPL-based PUF, and PHRR value of COFPL/EG sample (PUF6#) was decreased to about 127.78 kW/m2. The one HRR peak is because the sample is gradually burnt. In the second case, the first peak is assigned to the development of the intumescent protective char. After the first peak, the HRR curve forms a plateau 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 degradation of the protective layer gradually as the sample is continuously exposed to the heat and the formation of a new protective char in some formulations [32]. When it is burnt, diethyl phosphate makes PUF dehydrate to produce compact carbonization zone and this compact carbonization layer enters into worm-like residual char as a connected bridge between the gaps of wormlike residual char layer, making the worm-like residual char layer more compact. This char layer is more compact than that of GCO/EG filled PUF, which making the PHRR lower. Total heat released by combustion (THR), is a function of time per unit area and it is calculated by the integration of heat release for given time. Compared with PUF-4# displays in Fig. 7, the THR of PUF-6# is decreased to 21.80 MJ/m2. The data above indicates that parts of the PUF are protected without completely combusted. From Table 5, it can be seen that the total heat evolved per total mass loss (THE/TML) decreases significantly using COFPL as the polyol. This is because EG thermally expanded and absorbed heat from the system at 180e300  C, forming a worm-like protective layer on PUF surface. This layer acts as a physical protective barrier for heat transfer into the material, resulting in decreased heat release. These

Fig. 6. HRR curves for GCO filled PU foam, COFPL filled PU foam.

indicate a cumulative effect between EG and COFPL on flame inhibition. The synergistic effect may be attributed to the decomposition of diethyl phosphate to yield phosphinic acid with strong dehydration during heating, and the phosphinic acid promotes the formation of char, which together with EG forms a compact carbonaceous layer with stabilized structure. Generally, the emission of smoke along with HRR also plays a critical role in fire conditions. The smoke generation is an important hazard parameter in many flame situations. The effect of flame retardants on smoke formation was also measured. Total smoke release (TSR) is shown in Table 5. TSR of PUG-6# is about 470.63 m2, which is lower than PUF-4# (497.60 m2) in the combustion. The results of low TSR of both PUFs meant that EG performs an important function in smoke suppression. It can be noticed that the addition of EG in PUFs leads to a remarkable low TSR. This is because when exposed to fire, EG creates a carbon char with some small porous holes on the surface, which improves the smoke suppressing performance. Furthermore, when COFPL is used together with EG, the TSR of the PUF is lower than that of GCO foam containing the same EG content, due to the formation of compact carbon layer. For COFPL-based foam, more residues are observed as shown in Fig. 5. This indicates that the incorporation of EG and COFPL can promote the formation of carbonaceous materials in the

Table 5 The cone calorimeter data for GCO filled PU foam, COFPL filled PU foam.

2

PHRR (kW/m ) THR (MJ/m2) TSR (M2 m2) THE/TML (MJ/m2 g) CO/CO2 ratio TTI EHC SEA CO CO2

PUF-6#

PUF-4#

127.78 21.80 470.63 1.82 0.0302 6 20.21 534.4 0.0692 2.29

164.69 27.50 497.60 2.29 0.0243 5 24.15 345.49 0.0563 2.32

2791

Fig. 7. Total heat evolved (THE) for GCO and COFPL filled PU foam.

2792

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

condensed phase, showing the better flame retardancy of the expandable graphite especially for COFPL-based PUF. The CO/CO2 ratio of PUF with COFPL is 0.0302 while that is 0.0243 with GCO. It shows that the incomplete combustion rate increases due to COFPL addition, which is determined that if COFPL is introduced, the combustion control function is carrying out. 3.6. Thermogravimetric analysis (TGA and DTG) of polyurethane foams The TGA and DTG curves in nitrogen of flame-retarded foams are reported in Figs. 8 and 9, respectively. Figs. 8 and 9 show the TGA and DTG curves of polyurethane foams with various contents of EG at the linear heating rate of 10  C min1 under N2 atmosphere, and the corresponding values are listed in Table 6. The initial decomposition temperature can be considered as the temperature at which the weight loss is about 5%. Saunders and Frisch [33] summarized the four possible types of reactions that may take place in the thermal decomposition of urethanes. The tendency for a particular mechanism depends on the chemical nature of the groups, adjacent to the urethane linkage and the environmental conditions. Polyurethane degradation usually starts with dissociation of the urethane bonds, carbon dioxide and isocyanate evaporation. From Fig. 9 it can be noted that polyurethane foams show two main steps of weight loss (in the range of 250e400  C and 400e600  C). PUF-5# and PUF-6# foams show also an additional weight loss step at lower temperature (around 200  C). Considering that the boiling temperature of diethyl phosphate is about 203  C, those additional steps can be likely attributed to the loss of diethyl phosphate from polyurethane and this result is well in accordance with the thermogravimetric analysis of COFPL, which has a degradation step at about 210  C. It is clearly seen that the pristine foam (PUF-1#) begins to decompose around 260  C and its thermal degradation is composed of two main steps with a maximum weight loss temperature of 463.6  C. Compared with pristine polyurethane foam prepared from GCO, the modified polyurethane foam (PUF-5#) prepared from COFPL has slightly lower onset degradation temperature and the temperature of the maximum DTG peak is also slightly shifted to lower temperature. Firstly, the weight loss between 217 and 336  C is attributed to the breaking of urethane bond [34,35]. The degradation of diethyl phosphinate can occur by means of two competing processes: decomposition or vaporization. The scission

Fig. 9. DTG curves of polyurethane foams.

to phosphinic acid is energetically preferred; the first step is most probably related to the decomposition of urethane bond while the slightly lower onset degradation temperature of polyurethane foam prepared from COFPL, related to the diethyl phosphite, already shown for decomposition of COFPL. As same to COFPL, when it was reached to above 210  C, the diethyl phosphate degradated and phosphinic acid was generated. Phosphinic acid makes PUF dehydrate to produce carbonization char layer with abundant phosphorus and benzene ring and thus O2 and heat are prevented, making PUF more stable. The decomposition stage of polyurethane lies in the range from 336 to 480  C resulted from a faster weight loss of castor oil molecules, which is accorded with the literature [36]. The last stage (>475  C) corresponds to further oxidation of the crosslinked network and gradual oxidation of the char residue. From Fig. 9 it can be noted that residue char yield of EG filled GCO-based PUFs increases with the increasing EG content, especially for 30 wt% EG loading, residue char yield at 700  C increases to 17.07%, about 2.22 times of that of pristine GCO-based PUF (7.69%). Residue char yield of 10 wt% EG loading and 20 wt% EG loading is 11.12% and 13.86%, respectively, all higher than that of pristine PUF but lower than that of 30 wt% EG loading, showing the better flame retardancy of expandable graphite especially for GCObased PUFs. The results indicate the thermal stability of composites is enhanced during high temperature period. Those data provide positive evidence that EG can promote the formation of carbonaceous materials in the condensed phase. It can be also noted that the onset decomposition temperatures increase to 260.3  C (10 wt% EG), 268.9  C (20 wt% EG) and 269.4  C (30 wt% EG) comparing with pristine PUF(260  C), respectively. The results indicate that EG delays the decomposition because of the formation of efficient char layer due to the reaction of carbon and sulfuric acid which can be

Table 6 The corresponding values of thermogravimetric curves. Stage one

Fig. 8. TGA curves of polyurethane foams.

PUF-1# PUF-2# PUF-3# PUF-4# PUF-5# PUF-6#

Stage two

Onset ( C)

Tdmax ( C)

Rmax (%/min)

Onset ( C)

Tdmax ( C)

Rmax (%/min)

260 260.3 268.9 269.4 248 277.5

272.2 284.5 285.2 287.6 326.8 327.9

5.25 3.55 3.99 3.51 2.69 2.54

380 340 345 340 380 380

463.6 463.8 464.5 467.3 457.6 458.8

6.60 5.22 6.23 5.90 4.58 4.35

Residue char yield (%) 7.69 11.12 13.86 17.07 15.97 23.78

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

explained by Fig. 5. The phenomenon can be explained by the following description. When exposed to a heat source at above 160  C, the blowing effect which is from the intercalation compounds (H2SO4) decomposed to gaseous products (CO2, SO2 and H2O) causes an increase of the volume of PUF by about 100 times. The fused resin and carbonaceous compound can stick a large amount of expandable graphite in the expanding process. A large amount of EG is embedded in the char structure in a fibered way, so EG can enhance char structure and improve its resistance to crack. It is useful to form residual char under the protection of the char layers, and the acid in the EG may be helpful to form residual char too. The “worm-like” structure developed by graphite expansion suffocates the flame and the compact char layer formed limits the heat and mass transfer from polymer to the heat source, preventing further decomposition of the material. Compared with the TGA and DTG curves of EG filled GCO-based PUFs, the EG/COFPL-based PUF shows more char residue, due to the pyrolysis of EG and COFPL. The maximum decomposition temperature of EG/COFPL-based PUFs in the onset decomposition step is higher than that of corresponding EG/GCO-based PUFs, which indicates that COFPL and EG enhance the thermal stability of PUFs, which indicates that COFPL and EG enhance the thermal stability of PUFs. Moreover, COFPL and EG slow down the maximum thermal degradation rate. Compared PUF-6# with PUF-5# prepared from COFPL, the maximum decomposition temperature in the onset decomposition step increases and more residues are observed while the maximum thermal degradation rate reduces significantly. It can be concluded that there is a synergistic effect of EG and COFPL. EG acts as a flame retardant in the condensed phase and COFPL acts as a flame retardant in the gas and condensed phase. When it is exposed to a heat source, diethyl phosphate substituent group grafting in COFPL molecular chain degrades to form PO, PO2 and EG creates a great deal of insulative layer on the surface of the polyurethane foam, which collectively provide fire retardant performance. 4. Conclusions The results show that the flame retardant efficiency gets better with COFPL in the foam at the fixed EG weight percent or with increase in the EG weight percent in the foam. The presence of diethyl phosphate within the molecular structure of COFPL further improves the flame retardant efficacy of the PUFs significantly. The flame retardancy determined with COFPL/EG or EG/COFPL/TEP systems reveals increased resistance in polyurethane foams, indicating EG/COFPL or EG/COFPL/TEP systems have a synergistic effect as a common flame retardant in castor oil based PUFs. Mechanical strength of EG/COFPL/TEP polyurethane foam also shows as high as that of virgin polyurethane foam. The thermogravimetric analysis results for the PUFs indicate that the most effective EG/COFPL/PUF exhibits lower maximum thermal degradation rate and higher and more compact residue char yield. COFPL shows higher thermal stability over GCO. This EG/COFPL PUF exhibits a large reduction of peak of heat release rate compared to EG/GCO PUF. The SEM results show that the incorporation of COFPL and EG allows the formation of a cohesive and dense char layer, which inhibits the transfer of heat and combustible gas and increases the thermal stability of PUF. EG content of 30 wt% in COFPL turns to be excellent for its foamability than GCO with the isocyanate component, as verified by LOI, thermogravimetric analyzer and cone calorimetry test. These results additionally indicate that EG/TEP/COFPL system works more effectively than EG, TEP or COFPL in mechanical property and heat resistance enhancement of the PUFs. On the contrary, the enhancement in flame retardancy will expand the application range of COFPL based polyurethane foam materials.

2793

References [1] Cao X, Lee LJ, Widya T, Macosko C. Polyurethane/clay nanocomposites foams: processing, structure and properties. Polymer 2005;46(3):775e83. [2] Silva MC, Takahashi JA, Chaussy D, Belgacem MN, Silva GG. Composites of rigid polyurethane foam and cellulose fiber residue. J Appl Polym Sci 2010;117: 3665e72. [3] Williams CK, Hillmyer MA. Polymers from renewable resources: a perspective for a special issue of polymer reviews. Polym Rev 2008;48:1e10. [4] Gandini A. Polymers from renewable resources: a challenge for the future of macromolecular materials. Macromolecules 2008;41(24):9491e504. [5] Van Haveren J, Scott EL, Sanders J. Bulk chemicals from biomass. Biofuels Bioprod Bioref 2008;2:41e57. [6] Hatti-Kaul R, Törnvall U, Gustafsson L, Börjesson P. Industrial biotechnology for the production of bio-based chemicals e a cradle-to-grave perspective. Trends Biotechnol 2007;25:119e24. [7] Campanella A, Bonnaille LM, Wool RP. Polyurethane foams from soyoil-based polyols. J Appl Polym Sci 2009;112:2567e78. [8] Deka H, Karak N. Bio-based hyperbranched polyurethanes for surface coating applications. Progr Org Coat 2009;66:192e8. [9] Fritzen-Garcia MB, Oliveira IRWZ, Zanetti-Ramos BG, Fatibello-Filho O, Soldi V, Pasa AA. Carbon paste electrode modified with pine kernel peroxidase immobilized on pegylated polyurethane nanoparticles. Sens Actuators B Chem 2009;139:570e5. [10] Javni I, Petrovic ZS, Guo A, Fuller R. Thermal stability of polyurethanes based on vegetable oils. J Appl Polym Sci 2000;77:1723e34. [11] Javni I, Zhang W, Petrovi ZS. Effect of different isocyanates on the properties of soy-based polyurethanes. J Appl Polym Sci 2003;88:2912e6. [12] John J, Bhattacharya M, Turner RB. Characterization of polyurethane foams from soybean oil. J Appl Polym Sci 2002;86:3097e107. [13] Ogunniyi DS. Review paper castor oil: a vital industrial raw material. Bioresour Technol 2006;97:1086e91. [14] Yeganeh H, Mehdizadeh MR. Synthesis and properties of isocyanate curable millable polyurethane elastomers based on castor oil as a renewable resource polyol. Eur Polym Mater 2004;40:1233e8. [15] Liu TM, Bui VT. Instrumented impact testing of castor-oil-based polyurethanes. J Appl Polym Sci 1995;56:345e54. [16] Knaub P, Camberlin Y. Castor oil as a way to fast-cured polyurethane ureas. Eur Polym J 1986;22(8):633e5. [17] Khot SN, Lascala JJ, Can E, Morye SS, Williams GI, Palmese GR, et al. Development and application of triglyceride-based polymers and composites. J Appl Polym Sci 2001;82:703e23. [18] He F, Fan J, Lau S. Thermal, mechanical, and dielectric properties of graphite reinforced poly(vinylidene fluoride) composites. Polym Test 2008;27:964e70. [19] Uhl FM, Yao Q, Wilkie CA. Formation of nanocomposites of styrene and its copolymers using graphite as the nanomaterial. Polym Adv Technol 2005;16: 533e40. [20] Stankovich S, Dikin DA, Dommett GHB, Kohlhaas KM, Zimney EJ, Stach EA, et al. Graphene-based composite materials. Nature 2006;442:282e6. [21] Yu RG, Qiao XJ. Synthesis and application of expandable graphite as a nanometer compound material. Chin J Mater Rev 2003;17:125e6. [22] Xie RC, Qu BJ. Expandable graphite systems for halogen-free flame retardant of polyolefins I. Flammability characterization and synergistic effect. J Appl Polym Sci 2001;80:1181e9. [23] Duquesne S, Delobel R, Michel LB, Camino G. A comparative study of the mechanism of action of ammonium polyphosphate and expandable graphite in polyurethane. Polym Degrad Stab 2002;77:333e44. [24] Modesti M, Lorenzettia A, Simioni F, Caminom G. Expandable graphite as an intumescent flame retardant in polyisocyanurate-polyurethane foams. Polym Degrad Stab 2002;77:195e202. [25] Shen WC, Wen SZ, Cao NZ, Zheng L, Zhou W, Liu YJ, et al. Expanded graphite e a new kind of biomedical material. Carbon 1999;37:351e8. [26] Chen GH, Weng WG, Wu DJ, Wu CL, Lu JR, Wang PP, et al. Preparation and characterization of graphite nanosheets from ultrasonic powdering technique. Carbon 2004;42(4):753e9. [27] Zhan D, Ni ZH, Chen W, Sun L, Luo ZQ, Lai LF, et al. Electronic structure of graphite oxide and thermally reduced graphite oxide. Carbon 2011;49(4): 1362e6. [28] Shi L, Li ZM, Yang MB, Yin B, Zhou QM, Wang JH. Expandable graphite for halogen-free flame-retardant of high-density rigid polyurethane foams. Polym Plast Technol Eng 2005;44:1323e37. [29] Modesti M, Lorenzetti A. Flame retardancy of polyisocyanurateepolyurethane foams: use of different charring agents. Polym Degrad Stab 2002;78:341e7. [30] Shi L, Li ZM, Xie BH, Wang JH, Tian CR, Yang MB. Flame retardancy of different-sized expandable graphite particles for high-density rigid polyurethane foams. Polym Int 2006;55:862e71. [31] Ren WT, Peng ZL, Zhang Y, Zhang YX. Water-swelling elastomer prepared by in situ formed lithium acrylate in chlorinated polyethylene. J Appl Polym Sci 2004;92:1804e12. [32] Wu XF, Wang LH, Wu C, Yu JH, Xie LY, Wang GL, et al. Influence of char residues on flammability of EVA/EG, EVA/NG and EVA/GO composites. Polym Degrad Stab 2012;97:54e63. [33] Saunders JH, Frisch KC. Polyurethane: chemistry and technology, vol. 1. New York: Wiley-Interscience; 1962. p. 106e21.

2794

L. Zhang et al. / Polymer Degradation and Stability 98 (2013) 2784e2794

[34] Semenzato S, Lorenzetti A, Modesti M, Ugel E, Hrelja D, Besco S. A novel phosphorus polyurethane FOAM/montmorillonite nanocomposite: preparation, characterization and thermal behavior. Appl Clay Sci 2009;44:35e42. [35] Jasin’ska L, Haponiuk JT, Balas A. Dynamic mechanical properties and thermal degradation process of the compositions obtained from unsaturated poly

(ester urethanes) cross-linked with styrene. J Therm Anal Calorim 2008;93: 777e81. [36] Zhang L, Huang J. Effects of nitrolignin on mechanical properties of polyurethaneenitrolignin films. J Appl Polym Sci 2001;80:1213e9.