nanoparticles

nanoparticles

Polymer 53 (2012) 4787e4799 Contents lists available at SciVerse ScienceDirect Polymer journal homepage: www.elsevier.com/locate/polymer Engineerin...

2MB Sizes 4 Downloads 60 Views

Polymer 53 (2012) 4787e4799

Contents lists available at SciVerse ScienceDirect

Polymer journal homepage: www.elsevier.com/locate/polymer

Engineering biodegradable polymer blends containing flame retardant-coated starch/nanoparticles Seongchan Pack a, *, Ezra Bobo b, Neil Muir c, Kai Yang a, Sufal Swaraj d, Harald Ade d, Changhong Cao e, Chad S. Korach e, Takashi Kashiwagi f, Miriam H. Rafailovich a, * a

Department of Materials Science and Engineering, State University of New York at Stony Brook, Stony Brook, NY 11794-2275, USA University of Pennsylvania, Philadelphia, PA, USA Uniondale High School, Uniondale, NY 11553, USA d Department of Physics, North Carolina State University, Raleigh, NC 27695, USA e Department of Mechanical Engineering, State University of New York at Stony Brook, Stony Brook, NY 11794, USA f Fire Research Division, National Institute of Standards and Technology, Gaithersburg, MD 20899-8665, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 February 2012 Received in revised form 27 July 2012 Accepted 4 August 2012 Available online 21 August 2012

We have shown that the addition of resorcinol di(phenyl phosphate) (RDP)-coated starch can improve the compatibility to either Ecoflex or poly(lactic acid) (PLA). The increased compatibilization enhanced the tensile properties such as yield strength and impact toughness. In particular, we examined the effect of addition of RDP-coated starch on thermal responses of a blend of Ecoflex/PLA. We found that the combination of RDP-coated starches with nanoclays could render the blends self-extinguishing since they are formed as a shell-like chars on the exposure surface against heat, which can prevent the melt polymers against dripping. With an examination on the scanning transmission X-ray microscopy (STXM) images of the blends, the Ecoflex domains were well dispersed in the PLA matrix, while the domains became smaller when the RDP-coated starch was added. Moreover, we demonstrated that the introduction of either flat-like or tube-like clays could provide an increase of interfacial area on the RDPcoated starch surfaces, where each polymer chain preferentially segregates to either the starch or the clay surface. Thus, large complex in-situ grafts with polymers can be formed at the interfaces. Additionally, the complex in-situ grafts could influence flammability of the blends. We have shown that the addition of RDP-coated clays can decrease the mass loss rate of Ecoflex/Starch blends, while a lot of nanofiber are formed on the chars surface, which are entangled each other with the clay platelets. The mechanical properties of the chars structures were examined by nano-indentation, where a good elastic chars formation could keep the internal pressures built up with decomposed gases from melt polymers as well as ductility of the chars could play an important role on releasing the internal gases through small vents on its surface, steadily where a good elastic and ductile chars formation could require keeping the internal pressures built up with decomposed gases from melt polymers. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Biodegradable polymer blends Nanotube Flame retardants

1. Introduction Environmental concerns are rapidly growing regarding the large amounts of waste which is accumulating. As a result, interest in biodegradable polymers has been also risen sharply both in academia and industry [1e8]. The replacement of conventional polymers with biodegradable alternatives though, poses a significant challenge since the mechanical, thermal, and electronic properties of the original material have to be reproduced. For * Corresponding authors. Tel.: þ1 631 632 2843; fax: þ1 631 632 575. E-mail addresses: [email protected] (S. Pack), [email protected] notes.cc.sunysb.edu (M.H. Rafailovich). 0032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.polymer.2012.08.007

example, biodegradable polymers tend to absorb water more readily, are not easily compatibilized with other polymers, and the mechanical properties may be more difficult to control. To overcome those problems, biodegradable polymers have to be modified by either chemical/physical treatments on the polymers or the addition of compatibilizer. For instance, polylactic acid (PLA) is a biodegradable polymer, which can be produced by synthetic methods [3], has good mechanical properties and many potential applications as packing materials and automotives parts [4]. Although the biodegradable polymer can be applied to the commody products, the PLA should need to be rendered flame retardant because it is highly flammable. Reti et al. demonstrated that addition of 40% flame retardant formulation (FRs) containing 30%

4788

S. Pack et al. / Polymer 53 (2012) 4787e4799

Polymer

MW or grade

Manufacturer

samples, and determined relationships of the structure property of chars which resulted in blends that could have a good mechanical properties.

Starch Ecoflex Polylactic acid (PLA)

Melojel F BX 7011 2002D

National Starch BASF NatureWorks Co. Ltd.

2. Experimental section

Table 1 Polymers were used in this study.

2.1. Materials ammonium polyphosphate (APP) and either 10% starch or 10% lignin in the PLA matrix could achieve the vertical flammable designation [5]. Since PLA is non-charring polymer the addition of intumescing FRs have been used to pass the stringent flammable test. Zhan et al. showed that 25% intumescing FRs containing bisphosphorate diphosphoryl melamine was added in a PLA matrix [6]. Another method to achieve the self-extinguishing designation with PLA is to mix it with another biodegradable polymer, which can produce chars when heated. Starch is a good example of a thermally stable, charring polymer, where the carbonization process is known [5]. However, it is also well known that starch crystallizes easily and is difficult to blend with other materials [7,8]. Recently, researchers in NIST have studied that the addition of nanoclays resulted in the synergistic effects in producing more char formations when the Cloisite 20A was blended with the catalyst containing ammonium phosphomolybdate hydrate in poly(styrene-co-acrylonitrile) [9]. However, since the di-tallow surfactants on the clays have been considered as a toxic substance in the environmental community, an alternative surfactants have seek out in nanocomposite Pack et al. showed that montmorillonite (MMT) clays coated with resorcinol di(phenyl phosphate) (RDP) were very effective in the compatibilization of highly immiscible polymer blends [10]. Since RDP is a biodegradable compound with good thermal stability, we describe in this manuscript how it can play a role in improving the compatibility of starch with biodegradable polymers. In Ref. [11] Pack et al. also pointed out that in order to impart a polymer blend with multiple properties, such as high impact strength and good thermal response, a single nanoparticle was not sufficient to achieve the proper response. They showed that combination of nanoparticles with different surface interactions or morphologies was needed to achieve the synergy, which is necessary to simultaneously optimize multiple properties. Here we describe the use of clay tubes or platelets, coupled with starch nanoparticles, all coated with RDP surfactant, in optimizing the polymer blends of Ecoflex and PLA for both mechanical as well as flame retardant properties. To characterize the degree of the compatibility and flame retardancy, we used complementary techniques such as dynamic mechanical analysis (DMA), electron microscopy, a scanning transmission X-ray microscopy (STXM), and cone calorimetry on 1.5 mm molded

Halloysite nanotubes (HNTs), tubular-like clays were provided from NaturalNano, Inc. The length of the HNTs was 0.5e1 mm and the diameter was w500 nm. Cloisite Naþ was provided by Southern Clay Products, INC. Three biodegradable polymers were used in this study: a corn-based starch, known as Melojel, was supplied by National Starch. A poly(lactic acid) (PLA) was a commercial grade provided by Amco Plastic Inc. Ecoflex (grade F BX 7011) A commercial grades of biodegradable polymers was an aliphatic-aromatic co-polyester, (Ecoflex, F BX 7011), provided from BASF. The properties of the biodegradable polymers using in this study are tabulated in Table 1. Resorcinol di(phenyl phosphate) (RDP) was used as an absorbent, which was provided by ICLSupresta Inc. 2.2. Preparation of the RDP-coated starch, MMT clay and Halloysite nanotube To obtain the FR-coated particles, 20 wt% of RDP was first poured into a 200 ml beaker and then the beaker was placed on a hotplate stirrer at a temperature of 80  C. Once the viscosity of the RDP in the beaker becomes decreased to the point where it can easily be stirred manually or with a magnetic stirring bar, 80 wt% of starch, MMT clays, or Halloysite tubes were poured into the beaker while stirring mildly. The mixture was continuously stirred for an additional 15e20 min at 80  C in order to ensure the complete absorption of the RDP into the respective particles. The beaker was inserted in a vacuum oven at 100  C for 24 h to remove any unabsorbed liquid. The final RDP-coated particles were placed in a storage room before blending with the biodegradable polymers. The RDP coated MMT clays were characterized with atomic force microscopy and X-ray diffraction, as described in Ref. [10], which showed a uniform coating of 2.23 nm thick forming around individual platelets. The Halloysite tubes were also thoroughly coated with RDP as shown in Fig. 1, where we show SEM images of the neat and coated nanoclay tubes. Here, as well we see a thick uniform layers forming with no exposed clay surfaces. EDAX spectroscopy inserted in the figure confirms that the composition of the coating contains phosphorous.

Fig. 1. SEM images and EDAX spectra: (a) Neat HNTs (b) RDP-coated HNTs ( The scale bar is 1 µm).

S. Pack et al. / Polymer 53 (2012) 4787e4799

4789

2.3. Starch/polymer nanocomposites A C.W. Brabender with two screw roller blades, Type EPL-V501, and equipped with a direct current drive (type GP100), heated chamber was used to mix the polymers and the fillers. The polymeric pellets were first inserted to the chamber at a rotation speed of 20 rpm at 170  C. The additives, such as RDP-coated starch, were gradually added into the chamber and mixed at the same rpm for 2 min. The entire melt mixture was further blended at 100 rpm for 15 min under nitrogen gas flow, which prevented degradation of the mixture from heat-induced oxidation. The mixture was allowed to cool to room temperature in the chamber, and then using a hot press, small pieces of the mixture were molded into different shapes required for various experiments. Scanning transmission X-ray microscopy (STXM) was performed at beam line 5.3.2 at Advanced Light Source at Lawrence Berkeley National Laboratory to identify the chemical composition of polymer blends. The resolution of a forced X-ray beam was about 40 nm. This beam enables to scan cross-sections with 70e80 nm thickness. The cross-sections were cut from the bulk compounds using a ReichertJung Ultracut E microtome and then they were lift on coated copper mesh TEM grids without a supporting film on DI water surfaces. Nearedge X-ray Absorption Fine Structures (NEXAFS) spectra of starch, RDP and Ecoflex were measured beforehand to obtain the appropriate energies for images of polymer blends. The copper grids were mounted in the sample chamber which was evacuated to 0.3 mbar and subsequently refilled with 1/3 atm of helium. The intensity of the focused x-ray beam transmitted through the cross-sectioned films was recorded using a scintillator and photomultiplier tube and measured as a function of energy and position. The STXM images at

Scheme 1. A diagram of a specimen for tensile testing.

specific photon energies were representative to strong absorption of one of phases, where dark phases in the images were corresponded to high energy absorption and light phases were obtained at low energy absorption. Detail explanations and research studies of this scanning instrument can be found elsewhere [12,13]. Underwriters Laboratories (UL) 94 V0 assay was conducted in a vertical burning chamber to measure flammability of samples. The samples of dimensions 125 mm  13 mm  1.5 mm were molded and tested using the protocol established by ASTM D3801/ ISO 1210 UL 94 V0. Further details of the experimental methods, including the test diagram and figures are described in Ref. [14]. Transmission and Scanning Electron Microscopy (TEM/SEM) was conducted on cross sectional samples, which were placed on a coated copper mesh grid. The sections were directly viewed from a FEI Tecnai12 BioTwinG2 TEM at 80 kV and digital images were obtained from an AMT XR-60 CCD digital camera system. The RDP agents in the polymer matrixes were verified by an energy dispersive X-ray spectroscopy (EDXS) in the SEM (LEO-1550) with

Fig. 2. (a) The NEXAFS spectra of Starch, Ecoflex, and RDP and STXM images: (b) Starch/Ecoflex (50/50 wt%) (c) RDP Starch/Ecoflex (50/50 wt%), (d) RDP Starch/Ecoflex/Halloysites (60/40/5 wt%) (the scale bar is 3 mm and the image was taken at 284.7 eV).

4790

S. Pack et al. / Polymer 53 (2012) 4787e4799

Fig. 3. STXM images and NEXAF spectra: (a) PLA/Ecoflex/Starch (76/19/5 wt%), and (b) PLA/Ecoflex/RDP Starch (76/19/5 wt%). (The STXM images were taken at 284.7 eV and the scale bar is 5 mm).

a Schottky field-emission gun where the surfaces of RDP coated fillers were coated with a few micrometers of gold in order to make the specimens conduct. 2.4. Cone calorimetry The mass loss rate (MLR) measurements of samples were performed at the National Institute of Standards and Technology (NIST). The samples were made by molding 24  2 g of the composite into a 75 mm  75 mm  5 mm square. The samples were then wrapped on three sides with thin aluminum foil, exposing one side in the direction of the thermal radiation. The samples were exposed, in ambient atmosphere, to an external radiant flux of 50 kW/m2, normal to the sample surface and the MLR were measured as a function of exposure time. The residues in the thin aluminum foiled containers were collected and their morphology was taken by an optical microscopy. The standard uncertainty of the measured MLR was 10%. The gasification test was conducted by ASME/ISO 5660. Instrumented indentation was performed using a nanoindenter (Micro Materials NanoTest) with a 5 mm radius of curvature diamond cone indenter. Depth controlled indentations were performed with the maximum depth set to 2 mm. Char samples were mounted with an epoxy adhesive to aluminum sample holders for testing. Due to the large variation in char topology, regions were identified using a light microscope that were flat and free of pores. Loaddisplacement data was obtained using a constant loading and

unloading rate of 0.1 mN/s. Typical maximum loads reached were on the order of 1e12 mN. A 5 s hold segment was taken at the maximum displacement before unloading to take into account creep effects. On each sample, three identical indentations were performed with a minimum spacing interval of 60 mm between successive indents to ensure no interaction between residual plastic deformations. To determine elastic modulus of the char material, the slope (S) of the initial unloading data in conjunction with the spherical contact area is used in the conventional Oliver and Pharr method to determine a reduced modulus [10]. The contact radius as a function of the plastic depth (hp) of the char is given as:

a ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2Rhp  h2p ;

where R is the radius of the indenter (5 mm) and hp is calculated using the Oliver and Pharr method from the unloading curve. The reduced modulus (Er ) can be written as:

Er ¼

3 S 1=2 4 2Rhp  h2p

where S is the slope of the unloading curve, andEr is represented by:

1 1  n2s 1  n2i ¼ þ Es Ei Er

S. Pack et al. / Polymer 53 (2012) 4787e4799

4791

Fig. 4. STXM images at different energies at 284.7 eV and 288.5 eV: (aeb) PLA/Ecoflex/Starch (76/19/5 wt%), and (ced) PLA/Ecoflex/RDP Starch (76/19/5 wt%).

With knowledge of the elastic modulus and Poisson’s ratio of the indenter tip (Ei ¼ 1141 GPa, and ni ¼ 0.07) it is possible to determine the elastic modulus of the chars (Es) from the reduced modulus, assuming a char Poisson ratio of ns ¼ 0.5, typical for incompressible materials. Hardness (H) of the chars was calculated by the ratio of the maximum applied load (Pmax) divided by the contact area (A ¼ pa2). The reported material properties of the char surfaces were averaged over three discrete indentations. Tensile testing was conducted on an Instron tensile tester (model: 5840 Series Single Column Systems) equipped with a 5 kN load cell and grips. The measurements were performed at a rate of 0.5 mm/min and the data was analyzed in order to extract the mechanical properties; tensile strength, yield strength, and impact strength. The dimensions of the specimen for the tensile test were Table 2 UL-94-V0 results of RDP Starch/Ecoflex polymer blends. Sample

UL-94

Dripping

RDP Starch/Ecoflex (20/80 wt%) RDP Starch/Ecoflex (40/60 wt%) RDP Starch/Ecoflex (50/50 wt%) RDP Starch/Ecoflex (60/40 wt%) RDP Starch/Ecoflex (70/30 wt%) RDP Starch/Ecoflex (80/20 wt%) Starch/Ecoflex (50/50 wt%)

V2 V2 V2 V2 V0 V0 No grade

Yes Yes Yes Yes No No No

molded in accordance with ASTM D638, or the shape of the specimen is described in Scheme 1. 3. Results and discussion 3.1. STXM The morphology of a 50/50 wt% Starch/Ecoflex blend was imaged using STXM and the results are described in Fig. 2. The absorption spectra of RDP, starch, and Eecoflex are shown in Fig. 2a. From spectra, RDP has two absorption peaks at 285.1 and 287.2 eV, while those of Ecoflex are at 284.7 eV and those of starch are at Table 3 The results of tensile experiments with a series of RDP Starch/Ecoflex blends. Sample

Yield strength (Mpa)

Tensile strength (MPa)

Impact strength (MPa)

RDP Starch/Ecoflex (20/80 wt%) RDP Starch/Ecoflex (40/60 wt%) RDP Starch/Ecoflex (50/50wt%) RDP Starch/Ecoflex (60/40 wt%) RDP Starch/Ecoflex (70/30 wt%) RDP Starch/Ecoflex (80/20 wt%) Starch/Ecoflex (50/50 wt%)

5.75 3.83 3.56 3.32 0.61 0.62 5.37

12.32 5.73 3.26 2.72 1.19 0.73 5.95

9.18 6.67 4.02 4.9 0.26 0.25 0.6

4792

S. Pack et al. / Polymer 53 (2012) 4787e4799

Table 4 The results of UL-94-V0 flammable test of RDP Starch/Ecoflex blends with Halloysites nanotube (HNTs). Sample

Weight %

UL-94V0

Dripping

Neat Starch/EcoflexÒ/HNT Starch RDP/EcoflexÒ/HNT

60/40/5 60/40/5

NGa V0

No No

a

Burnt to the clamp.

286.6 eV. Hence in order to map the surface of the sample, we can place a window on the blend and obtain the spatial distribution of all the components, which was taken at 284.7 eV. The results are shown in Fig. 2bed. From the figure, dark regions corresponded to the Ecoflex phase while white regions are starch phases, where they are phase separated in the matrix. From the Fig. 2c we can see that the RDP map overlaps completely with the starch map, indicating that it is indeed localized into the starch particles. However, RDP coated starch and Ecoflex are still phase separated. As a result, the addition of RDP on the starch phase may not obtain a good dispersion of Ecoflex phase in the matrix, which results from phase segregation of Ecoflex. However, when 5% HNTs clays are added in the blend system, the phase segregations severely occur with smaller Ecoflex domains, where they appear in the shear direction in the Fig. 2d. This segregation could be contributed by an increase of interfacial areas, which results from being attached the HNTs on the RDPcoated starch. As we have shown in Fig. 1, RDP can be strongly absorbed on the clay surface. An excess of RDP oligomers may be favorable onto the clay surfaces at a given high shear. Hence this strong interaction between the clay surface and the RDP-coated starch could increase in compatibility of the blend. Recently, Pack et al. demonstrated that the RDP absorbed on flat-like (e.g. Montmorillonite) clays could be easily exfoliated in homopolymers since there was no energy penalty for the exfoliation, which was confirmed by obtaining lower contact angles for polymers/RDP clays [10]. It could indicate that the lower surface energy might reduce the interfacial tension between the polymers/RDP clays. Since HNTs clays tend to segregate onto the RDP-coated starch, the RDP may be more distributed to its starch surfaces. Therefore, the starch phases containing the RDP adsorbed to HNTs can be more compatible to Ecoflex domains. The RDP coated starch can be mixed as filler in a polymer blend system.

Fig. 5. The results of impact strength of RDP Starch/Ecoflex blends with HNTs.

A polymer blend of Ecoflex/PLA with RDP-coated starch: Another biodegradable polymer, 76% PLA was melt-blended with 19% Ecoflex either with 5% RDP-coated starch or with uncoated starch at 170  C. The cross-sections from the two polymer blends were exposed by X-ray beam to examine their chemical compositions in aid of NEXAFS spectra. In Fig. 3a we have shown that PLA and Eecoflex have very sharp interfaces at each other, and that different NEXAFS spectra are obtained from each phase. It can indicate that they are phase separated to each other even if the uncoated starch is added in the blend. Moreover, Ecoflex domains (dark regions) are seen as aggregated phases in the PLA matrix. However, the addition of the RDP-coated starch could improve the compatibility of the blend, which is shown in Fig. 3b. From the figure we can see that the domains decrease in size as well as increase the degree of dispersion. Since the RDP can act as a surfactant on the starch surface, the RDP-coated starch segregate onto the interfaces between Ecoflex and PLA thereby may increase the interfacial width. As a result, the sharp interfacial area is opaque. Different STXM images of the two blends taken at 284.7 eV and 288.5 eV are shown in Fig. 4. From the figure we can see that the Ecoflex domains are well dispersed in the PLA matrix when the RDP-coated starch is added. Therefore, the RDP-coated starch can improve the degree of the compatibility of the two biodegradable polymers. 3.2. Thermal resistance and mechanical properties of biodegradable blends A polymer blend of RDP Starch/Ecoflex: Improving the compatibility of the RDP-coated Starch/Ecoflex blend has an influence on flame retardant(FR) and mechanical properties. We had previously shown that the addition of 5% Cloisite 20A to either PC/SAN or PS/ PMMA blend with FR particles could increase the degree of the dispersion of FR particles into their preferential phase, which resulted in the enhanced flame retardant properties such as passing the UL-94-V0 test [14]. We have also showed that the addition of RDP-coated clays could not only decrease the amount of FR formulations (normally 19 wt%) but also increase the compatibility of the polymer blends [10]. Therefore, we propose that the absorption of RDP to either particles or starch may enhance compatibilization for biodegradable polymer blends thereby improve its mechanical and flame retardant properties. In this section we will show that the RDP coated starch can be a good candidate for a flame retardant agent as well as compatibilizer. Flame retardant and tensile properties of RDP Starch/Ecoflex blends are shown in Tables 2 and 3, respectively. From the tables we can find that the addition of RDP allows the blend of Starch/Ecoflex (50/50 wt%) to pass a designation of UL-94 V2, while in the absence of RDP, the blend is very flammable and t1 is about 180 s as time to burn. On the other hand, impact property of the blends is also different. From Table 3 we can see that the impact strength of the blend is increased by nearly an order of magnitude when the starches are coated with RDP. These improvements can be attributed to an increase of the compatibility of Starch/Ecoflex blend in aid of the absorption of RDP. Moreover, we also can see that increasing the ratio of RDP-coated starch to Ecoflex increase in flame retardant properties, where the UL-94 grading is upgraded to V0 from V2 when 70% RDP-coated starch is blended with 30% Ecoflex. However, obtaining this enhanced flame retardancy results in an expense of the decreased impact properties, which is shown in Table 3. However, the blends of RDP Starch/Ecoflex are still not selfextinguishing, where the ratio of RDP starch is less than 70% in the blend with Ecoflex. In order for the blends to make it selfextinguishing, we add HNTs in a blend of 60% RDP starch/40% Ecoflex. As we have shown STXM images in the previous section, the addition of the HNTs can increase the degree of dispersion of

S. Pack et al. / Polymer 53 (2012) 4787e4799

4793

Fig. 6. The results of impact strength and stress-strain curves of PLA/Ecoflex blends with different nanofillers.

the Ecoflex domains. This enhancement in the dispersion of Ecoflex results in rendering the blend with the HNTs self-extinguishing. From Table 4 we can see that when the 60% RDP-coated starch is mixed with 40% Ecoflex and 5% HNTs clays, the blend is passed UL94-V0 compared to the blend of neat Starch/Ecoflex/HNTs in the same composition, which is ranked as V2 with dripping in the vertical flame test. Furthermore, we have also mentioned that the increase of interfacial areas by the attachment of HNTs clays on the surfaces of RDP starch could improve a distribution of RDP molecules along the interfaces between the RDP Starch/Ecoflex. As a result of that, from Fig. 5 we can see that the impact strength of the blend with RDP coated starch is increased by a factor of 2 compared to that of the blend with uncoated starch.

A polymer blend of PLA/Ecoflex with RDP-coated particles: As we have shown that Ecoflex polymers are not compatible to the PLA matrix, the addition of RDP coated starch can enhance the compatibility of the blend. This improved compatibility can affect on mechanical properties of the blend. Furthermore, PLA and Ecoflex are very flammable so that it is very difficult to pass a stringent flame test such as UL94-V0. Therefore, it is worthy to obtain flame retardant biodegradable polymers. In this section we have a blend of 20% Ecoflex/80% PLA with different particles to investigate the materials properties of the blend. From Fig. 6 we can see that the impact strength of the Ecoflex/PLA blend with 2.5% RDP-coated starch and 2.5% Cloisite Naþ clay is highest among other polymer blends. This increased toughness could be attributed

4794

S. Pack et al. / Polymer 53 (2012) 4787e4799

somewhat decrease but still increases in a factor of 4 compared to that of the neat polymer blend. It may indicate that unbounded RDP either the starch or the clays can improve the compatibility of the blend. Furthermore, the blends containing the two types of particles are obtained UL-94 V0 grade. This synergy from a combination of two different particles in size and shape was previously reported in the published paper [11]. 3.3. Cone calorimetry

Fig. 7. The TEM image of Ecoflex/PLA/RDP-coated clay (20/80/5 wt%).

to the increased interfacial area in which the excess of RDP on the RDP starch may be absorbed on the clays surface. Since RDP clays tended to be good compatible to any polymers [10] the Cloisite Naþ with the RDP may be segregated into one of components in the blend. As a result of that, the exfoliated Clositie Naþ in Ecoflex domains is observed through a TEM image of the PLA/Ecoflex blend, which is shown in Fig. 8. From the figure we can see that we can see that the RDP-coated clays are well dispersed and exfoliated in the Ecoflex domain as well as some of the clays are seen at the interfaces between Ecoflex/PLA. Since most RDP-coated clays segregate to the Ecoflex domain, which is relatively ductile. When 5% RDP clays are added, it would expect that the impact strength of the blend may increase because of the improved compatibility. However, the impact strength is almost same as that of the PLA/ Ecoflex blend, which is shown in Fig. 7. It may be explained by the fact that Ecoflex becomes a brittle phase. Therefore, the blend with the RDP clays is more brittle even if the clays are exfoliated in the Ecoflex. On the other hand, when the RDP-coated starch is added in the blend with RDP clays, the impact strength of the blend

Fig. 8. The MLR of Starch/Ecoflex polymer blends with different nanoparticles.

Thermal responses of polymer blends are more complicated than that of homopolymers since the blend systems tend to phase segregate once heat is applied to the blend air interface. Furthermore, if one of polymers in the blend produces chars during decomposition, the combustion behaviors of the blend are much more difficult to understand and/or predict. We have shown in previous publications that the addition of Cloisite 20A can stabilize the blend against phase segregation. Moreover, the clays can be very effective to protect melt polymers from the heat fronts, by which leads to produce a lot of solid residues containing clay platelets. However, if a charring polymer exists in a polymer blend, the clays could not be as the same effective as they did at noncharring systems since expanding char surface with gases could be controlled by the surface elasticity and hardness [10]. Here we examined a blend of 60% starch and 40% Ecoflex with different nanoparticles in flammability, which was conducted in gasification with N2. The results of mass loss rate (MLR) are shown in Fig. 8. From the figure we can see that the MLR trace of the blend without any nanofillers constantly increases until a broad peak MLR appear at about 28 g/m2s. The peak is decreased to a steady plateau at about 15 g/m2s once 5% HNTs clays are added. Moreover, when the 5% RDP-coated clays are mixed in the blend of Starch/ Ecoflex, the steady plateau is more decreased to nearby 13 g/m2s. However, the plateau is rebounded to the broad peak when 2% RDP clays are replaced with 2% HNTs in the blend of Starch/Ecoflex with 5% RDP-clays, which could deteriorate somewhat the improved flame retardancy of the blend. This may result from a poor network-like char structures formed during decomposition, where the addition of HNTs could interrupt forming the expandable char layers on the exposed surface [10,15]. On the other hand, when 3% Cloisite 20A (C20A) is added in the blend with 2% HNTs. A steady plateau appears at about 11 g/m2s and then the MLR trace slightly rises up to 15 g/m2s until about 350 s. After that, it is similar to the MLR trace of the blend with 5% RDP clays, where both the MLR traces decrease gradually at the end of the combustion. Therefore, the addition of either 5% RDP clays or the combination of C20A with HNTs could show the better performance in flammability because the rate of mass loss of both the blends is lower than that of the blend with either the RDP clays/HNTs or the HNTs alone. These results from the cone calorimetry can be confirmed by examining optical and electron images of their residues, which is shown in Figs. 9 and 10. Previously, we reported and postulated that the degree of the segregation of clays to either exposed surfaces or polymers interfaces depends on the surface energy and/or interactions between polymer/FR agents [10]. We also reported in the previous paper that the FR absorbed clay platelets could stabilize to form intumescent chars when the melt polymers decomposed because of the relatively higher surface energy of the FR coated clays. Since starch is a good charring polymer but very brittle, it is difficult for the starch to form an expandable char layers. As a result of that, from the Fig. 9a we can see that in the case of unfilled blend the residue surface looks a hard shell-like char layers covered. However, in the cases of the blends with either 5% HNTs or 5% RDP clays, the residues surface relatively looks a soft shell-like char layers covered, which is shown in Fig. 9bec. A closer examination on the optical

S. Pack et al. / Polymer 53 (2012) 4787e4799

4795

Fig. 9. The residues of Starch/Ecoflex blends with the nanoparticles: (a) Starch/Ecoflex (60/40 wt%), (b) Starch/Ecoflex/HNTs (60/40/5 wt%), (c) Starch/Ecoflex/RDP-coated clays (60/ 40/5 wt%), (d) Starch/Ecoflex/RDP-coated clays/HNTs (60/40/3/2 wt%), and (e) Starch/Ecoflex/C20A/HNTs (60/40/3/2 wt%).

images of both the HNTs and the RDP clays residue shows that the char surface of the HNTs clays is relatively rougher than that of the RDP clays. This difference might lead to the fact that the MLR of the RDP clays was lower than that of the HNTs since the addition of RDP coated clays could keep forming intumescent chars from the Starch/ Ecoflex blend. Moreover, the optical images of the residues from either RDP clays/HNTs or C20A clay/HNTs blend show that even though it is similar in quantity of the two residues covered in the container, the chars surface of the C20A/HNTs is much smooth than

that of the RDP clays/HNTs, which may result in keeping forming intumescent chars during combustion, Thus, the difference of the two residues in char morphology could lead to obtaining the different results in the MLR measurement. 3.4. Chars morphology An examination of the SEM images of chars in microstructure can explain the different results from the cone calorimetry, which is

4796

S. Pack et al. / Polymer 53 (2012) 4787e4799

Fig. 10. SEM images/EDAX spectra of the residues: (a) Starch/Ecoflex (60/40 wt%), (bec) Starch/Ecoflex/HNTs (60/40/5 wt%), (dee) Starch/Ecoflex/RDP-coated clays (60/40/5 wt%), (feg) Starch/Ecoflex/RDP-coated clays/HNTs (60/40/3/2 wt%), and (hei) Starch/Ecoflex/C20A/HNTs (60/40/3/2 wt%).

S. Pack et al. / Polymer 53 (2012) 4787e4799

4797

Fig. 10. (continued).

shown in Fig. 10. From the figure we can see that, in the case of unfilled blend, the char surfaces look strand-like bundles, and a higher resolution of the SEM image shows that a lot of small holes, about a few micro sizes in diameter, exist at the bundles. However, when either the HNTs or the RDP clays are added, the bundles look coated with the nanoparticles, which can be confirmed by the EDAX spectrum of the chars, where the peak of Si/Al is relatively higher than that of C/O. This segregation of the nanoparticles to the bundles is consistent with the results that we have reported in the previous paper [10,11]. Furthermore, in the case of RDP clays chars, many nanofibers are observed onto the clays, where they look entangled with each other in Fig. 10e. This kind of network-like structures is more clearly shown at a lower resolution in the SEM images, which is shown in Fig. 10def, where the fibers are formed in a crystalline-like structure on the bundles surface. Discovery of the nanofibers may result from a byproduct, where a chemical reaction could occur at the interfaces between the RDP clay and the Ecoflex. Since we have shown that the RDP coated clays can be well

Fig. 11. (a) The elastic modulus and (b) hardness of Starch/Ecoflex blends with different nanoparticles.

Scheme 2. A model of compatibilization, where the RDP-coated starch is added in Ecoflex/PLA/Cloisite Naþ blend.

4798

S. Pack et al. / Polymer 53 (2012) 4787e4799

exfoliated and/or intercalated in Ecoflex at the section of STXM, the fibers could be made out of the polymers/RDP, which can be confirmed by the EDAX spectrum. It shows that the C peak is much higher than that in the case of HNTs chars. Additionally, in the case of RDP clays/HNTs, from Fig. 10feg we can see that there are no fibers observed because the two nanoparticles may interact with either other, rather than either the polymers/RDP clays or polymers/HNTs. This existence of competition at the interfaces could cause an agglomeration of the two particles on the char surface, which is shown in Fig. 10g. However, in the case of C20A/HNTs, a kind of the agglomeration does not obtain at the char surfaces, which is shown in Fig. 10hei. Therefore, since C20A/HNTs does not have any interaction, the C20A first may segregate to the air surface and then the HNTs clays are segregated underneath the C20A clays, which could obtain the spider net-like char formation at the exposed surfaces. 3.5. Nanoindentation In the previous paper we have reported that in order to sustain and/or reduce internal pressure built up with decomposed gases from polymers, the expanding char surface may need to have good elasticity. Otherwise any crack can occur at the exposed surface from heat fronts. Therefore, to measure elastic modulus and hardness of chars surface is an important procedure for understanding a mechanism of intumescing chars to elucidate a correlation between char structures and flame retardant properties, such as MLR and HRR. We here used nanoindentation to obtain both the elastic modulus and the hardness of chars. The results are shown in Fig. 11aeb. From the figure we can see that the elastic modulii and hardness of the chars from either the unfilled blend or the blend with RDP clay/HNTs are almost same, E0 ¼ 0.2 GPa, H ¼ 0.03e0.05. This may indicate that their chars are very fragile and weak so that intumescing chars could be collapsed before increased internal pressure is slowly released throughout the small vent holes, as shown in the SEM images. However, when either HNTs or RDP clays are added, the elastic modulii are E0 ¼ 5.3 GPa and E0 ¼ 2.5 GPa, respectively. These results may be consistent with the char morphologies in the SEM images, where the HNTs are a uniformly covered on the char surface, whereas the RDP clays are also well segregated on the surface, as well as the soft parts, nanofibers, are dominated on the surface. The soft segments may result in the lower hardness obtained from the measurement. Therefore, the chars formed by the HNTs are more elastic behavior, which may indicate that a maximum expansible char surface formed by gases is dependent on the elastic modulus. This can be contributed to the results of the MLR, where the mass loss rate of the HNTs case is highest among the other cases. Furthermore, the hardness property may explain that once gases release through the vent holes, internal pressures are controlled by a relative hardness on the expanded char surface. If hardness is lower, internal pressures are relatively steady during combustion. As a result of that, we can elucidate the result that the time of mass loss in both the RDP clays and C20A/ HNTs is longer than that of the HNTs case alone. Therefore, in order to construct a superior flame retardant material, it may be considered with the mechanical properties such as elastic modulus and hardness. 3.6. Model From the above experiments we can find that the RDP itself is not imparting the improved flame retardant and mechanical properties of the biodegradable polymers. In the case of the blend of Ecoflex/Starch, the charring starch phases must be properly distributed in the Ecoflex matrix in order to obtain the better

materials property. Since RDP acts as a surfactant, the RDP coated starch is able to be well dispersed in the matrix thereby the impact property of the blend with 50% RDP coated starch is much higher than that of the blend with 50% uncoated starch (see Fig. 5). Moreover, the addition of organophilic clays was shown in increasing the compatibility of immiscible polymer blends when they exfoliated and/or intercalated favorably at interfaces in which the segregation of clays could be also achieved by the adsorption of RDP onto the clay surfaces [10-1114]. Therefore, a good dispersion of the RDP-coated starch may occur by the fact that clay platelets segregate onto the starch surface. As we have shown in Fig. 6, in the case of PLA/Ecoflex/RDP starch blend, the further increased impact strength can be obtained when the Cloisite Naþ clays are added in the blend, where a synergy may exist at the interfaces between the RDP-coated starch and the clays surface. We therefore propose the following model, based on the TEM micrographs and STXM images shown above. From Fig. 7 we found that the RDP coated clays segregated preferentially to the Ecoflex phase, while the RDP coated starch segregated preferentially to the PLA phase. When both types of particles are present, the common RDP coating allows both types of particles to preferentially segregate together, as shown in Scheme 2. Since the clays have Ecoflex chains adsorbed, while the starch particles have PLA chains adsorbed, the combined starch/clay particles form the basis of a very large in-situ graft complex, which can greatly enhance the compatibility, as well as the mechanical and thermal properties of the nanocomposite.

4. Conclusion We have confirmed that the resorcinol di(phenyl phosphate) (RDP)-coating method can apply to either starch itself or the combination of the starch with nanopartilces in a blend of Eecflex/ Poly(lactic acid) (PLA) in order to improve the compatibility of the blend. We demonstrated that the introduction of either flat-like or tube-like clays could provide an increase of interfacial area on the RDP-coated starch surfaces, where each polymer chain preferentially segregates to either the starch or the clay surface. Thus, large complex in-situ grafts with polymers can be formed at the interfaces. Therefore, this formation of in-situ grafts at interfaces enhanced the tensile properties such as yield strength and impact toughness. The complex in-situ grafts could also influence the flammability of the blends. We found that the combination of RDP-coated starch with nanoclays could also render the blends selfextinguishing. We found that the clay segregated to the polymer interfaces, thereby acting as compatiblizers, while the RDP/starch particles migrated to the surface. When exposed to heat, the clay prevented phase segregation, while the RDP starch migrated to the surface where it reacted and formed a hard shell that prevented dripping. The synergy between the clay and starch particles was driven by their common surface energy obtained from the RDP coating. We also observed that the mechanical properties of the chars were another factor which correlated with enhanced flame retardance. As the composite burns, hot gases form in the interior of the sample, exerting pressure on the char enclosure. If the char is brittle, cracks will form, exposing the interior to the hot front, thereby increasing the MLR and HLR. When the chars are ductile, they expand, thereby releasing the internal pressure without cracking and protecting the internal material from the heat front. In this case, we found that the char of the blend consisting of either RDP-coated clays or C20A/HNTs was the most ductile, and testing also showed that it was the most efficient at flame retardance.

S. Pack et al. / Polymer 53 (2012) 4787e4799

Acknowledgment The authors thank Ms. Susan C. Van Horn from Central Microscopy Imaging Center (C-MIC) at Stony Brook University for taking TEM images. This work was supported by the NSF-MRSEC program. References [1] [2] [3] [4] [5]

Kim Meera, Lee Sun-Ja. Carbohydr Polym 2002;50:331e7. Ma XF, Yu JG, Wan JJ. Carbohyrate Polym 2006;64:267e73. Tokiwa Y, Calabia BP. Appl Mirobiol Biotechnol 2006;72:244e51. Kimura K, Horikoshi Y. Fujitsu Sci Tech J 2005;41:173e1437. Reti C, Casetta M, Duquesne S, Bourbigot S, Delobel R. Polym Adv Technol 2008;19:628e35. [6] Zhan J, Song L, Nie S, Hu Y. Polym Degrad Stabil 2009;94:291e6.

4799

[7] Walenta E, Fink H, Weigel P, Ganster J, Schaaf E. Macromol Mater Eng 2001; 286:462e71. [8] Mani R, Bhattacharya M. Eur Polym J 2001;37:515e26. [9] Liu M, Zhang M, Zammarano M, Gilman JW, Kashiwagi T. Polym Degrad Stabil 2011;96:1000e8. [10] Pack S, Kashiwagi T, Cao C, Korach C, Rafailovich M. Macromolecules 2010;43: 5338e51. [11] Pack S, Kashiwagi T, Stemp D, Koo J, Si M, Sokolov JC, et al. Macromolecules 2009;42:6698e707. [12] Ade H, Winesett DA, Smith AP, Qu S, Ge S, Sokolov J, et al. Europhys Lett 1999; 45:526e32. [13] Kilcoyne ALD, Tyliszczak T, Steele WF, Fakra S, Hitchcock P, Franck K, et al. J Synchrotron Radiat 2003;10:125e36. [14] Pack S, Si M, Koo J, Sokolov JC, Koga T, Kashiwagi T, et al. Polym Degrad Stabil 2009;93:306e26. [15] Kashiwagi T, Mu M, Winey KI, Cipriano B, Raghavan S, Pack S, et al. Polymer 2008;49:4358e68.