- Email: [email protected]
polypropylene nanocomposites
Composites Science and Technology 105 (2014) 166–173
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Laser-induced thermo-oxidative degradation of carbon nanotube/polypropylene nanocomposites Stephen F. Bartolucci a,⇑, Karen E. Supan a,b, Jeffrey M. Warrender a, Christopher E. Davis a,c, Lawrence La Beaud d, Kyler Knowles d, Jeffrey S. Wiggins d a
U.S. Army Benét Laboratories, Armaments Research Development and Engineering Center, 1 Buffington St., Watervliet, NY 12189, USA Norwich University, Department of Mechanical Engineering, Northfield, VT 05663, USA Massachusetts Institute of Technology, Department of Chemical Engineering, Cambridge, MA 02139, USA d University of Southern Mississippi, School of Polymers and High Performance Materials, Hattiesburg, MS 39406, USA b c
a r t i c l e
i n f o
Article history: Received 5 May 2014 Received in revised form 2 September 2014 Accepted 24 September 2014 Available online 5 October 2014 Keywords: A. Nanocomposites A. Carbon nanotubes A. Nanoclays B. Thermal properties B. High-temperature properties
a b s t r a c t Nanocomposites of multi-walled carbon nanotubes and polypropylene were fabricated using twin screw extrusion and subjected to heating rates on the order of 100,000 degrees per second using laser pulse heating. Raman and infrared spectroscopy, as well as electron microscopy observations, showed that the nanotubes retain their structure and that a nanotube-rich layer formed at the surface of the composites during laser heating. By varying the weight fraction of filler in the composites, a 54% reduction in mass loss per laser pulse was achieved, compared to the pure polymer. The nanotube composites also displayed lower mass loss than nanoclay composites under the same conditions, despite having a significantly lower weight fraction of additive. Further reduction in mass loss per laser pulse was achieved in solvent-processed samples, which appear to have better dispersion than the extrusion-processed composites. We attribute these observations to the formation of a nanotube network formed on the surface, which acts as a protective barrier, slowing the degradation of the underlying polymer. Several observations made in this study are consistent with those reported in previously published slow heating rate studies, showing that while some phenomena are unique to transient heating, there are some commonalities between the two regimes. Published by Elsevier Ltd.
1. Introduction Organic-based polymers are increasingly being used to replace inorganic materials in some applications where the material may be subjected to elevated temperatures. Since most polymers have a relatively low melting and thermal stability temperature compared to materials like metals and ceramics, the addition of particles to the polymer matrix is often needed to produce a composite with improved thermal stability. There has been a substantial amount of research on the use of nanoparticles as additives to polymers to improve both the mechanical and thermal properties. Fillers such as nanoclay and carbon nanotubes have been among the most widely investigated for many years [1–3]. Carbon nanotubes are high aspect ratio carbon nanostructures that may be thought of qualitatively as graphene layers that have been rolled into a tubular shape. Carbon nanotubes exhibit extremely high strength, modulus, and thermal and electrical conductivity, and
⇑ Corresponding author. http://dx.doi.org/10.1016/j.compscitech.2014.09.018 0266-3538/Published by Elsevier Ltd.
when dispersed well in a polymer matrix, can form a network structure throughout the composite. The thermal stability or flammability of carbon nanotube nanocomposites has been studied previously [4,5]. In traditional thermal stability studies, the nanocomposite is heated at rates on the order of 10 °C/min. Using thermogravimetric analysis (TGA), the mass change of the sample is recorded and the thermal stability is determined as a measure of the mass loss versus temperature. The thermal degradation of the nanocomposite can be determined under both oxidative and non-oxidative conditions. Kashiwagi et al. [6,7] showed that 1 and 2 vol% multi-walled carbon nanotubes (MWCNT) in polypropylene increased the thermal stability of the polymer during TGA tests in air and nitrogen, and decreased the heat release rate in a cone calorimeter and gasification tests. The high aspect ratio carbon nanotubes formed a networked structure covering the sample surface, acting as a thermal shield and slowing the degradation of the underlying polymer. Incorporation of MWCNTs into other polymer materials has shown similar behavior [8–10]. It has been shown that effective dispersion of the nanotubes in the polymer matrix is critical for the reduction
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
in mass loss rate during flammability tests [11]. Well-dispersed nanotubes form a homogeneous network layer, whereas poorly dispersed nanotubes form a discontinuous layer consisting of fragmented islands of networked nanotubes on the composite surface. Well-dispersed nanotubes also lead to a higher viscosity of the polymer melt, making bubble formation and coalescence more difficult. Decreased bubbling and bursting leads to a nanotube network layer that is more intact. This continuous protective network layer leads to improved fire retardance. Carbon nanotubes have also been shown to provide superior enhancement to fire retardance compared to other carbon nanoadditives, such as carbon nanofibers and carbon black [5]. Certain applications, such as ballistic environments, have heating rates that are many orders of magnitude larger than those typically achievable in thermogravimetric analysis or flammability studies, necessitating the use of alternative methodologies to study such regimes. Laser pulse heating (LPH) has been used to achieve heating rates on the order of 107 K/min [12–14]. In order to achieve photothermal heating and to avoid photochemical effects (as seen with the use of UV wavelength lasers), an infrared laser with variable pulse duration on the order of milliseconds is used to heat the sample surface. By adding the same weight percentage of carbon black to each of the composites, uniform optical absorption of the laser light among samples of various compositions can be achieved [15,16]. Studying the thermal degradation of polymer nanocomposites with this method combines phenomena seen in both polymer ablation and polymer thermal stability/flammability research. We can observe ablation of material, melting and bubbling of the heated polymer, and degradation of the polymer chains. This group has previously studied highly transient heating of polymer/nanoclay nanocomposites [17] and observed decreased mass loss with increasing amounts of montmorillonite nanoclay in a polypropylene matrix. Spectroscopic characterization showed that a silicate-rich layer forms on the surface of the LPH region of the composite, providing a mechanism of protection for the underlying polymer. While carbon nanotube composites have been studied at slow heating rates, they have not been studied at high heating rates. The objective of this study is to determine how carbon nanotube/polymer composites behave under highly transient heating and whether correlations can be made to the behaviors observed during low-heating rate thermal stability and flammability testing.
167
2. Experimental 2.1. Composite preparation Isotactic Polypropylene (PP) pellets (Total Petrochemical Type 3371, MFR = 2.8 g/10 min at 230 °C, 2.16 kg) were dried for 16 h under vacuum at room temperature. A master batch (Nanocyl PP2001CNT/Thermoplastic) at 20 wt% MWCNTs (with average 9.5 nm diameter and 1.5 lm length) were compounded with appropriate amounts of the neat PP resin to achieve desired final loadings using a co-rotating twin screw extruder and a controlled feeder. Zone 1 was held at 170 °C and zones 2–4 and the die were held at 220 °C and screw speed at 200 RPM. Pellets were extruded through a 1/1600 circular die and passed through an ice bath prior to being pelletized. They were then passed through the extruder for a second time under identical conditions. Using electrical conduction as a measurement for MWCNT dispersion, it was found that two sequential passes through the extruder (under identical conditions) improved dispersion. Additional passes through the extruder yielded increased conductivity up to five passes, though physical properties did not show a corresponding increase beyond two passes. For this reason (as well as additional processing time), the compounding step was limited to two passes. Composites were made with 0.5, 1.0, 2.0, 3.0 and 5.0 wt% MWCNT. Compounded PP/MWCNT pellets were melt-pressed at 200 °C in a steel die with PTFE release film. No pressure was applied for 5 min while the melt press equilibrated and the PP began to melt/flow. Pressure was then applied in 13.7 MPa increments to 68.9 MPa. The pellets were left under 68.9 MPa for 300 s, at which time they were removed and cooled to room temperature. The newly-formed discs were removed from the die and any flash material was trimmed from the edges. All materials were compounded with 1 wt% carbon black (Cabot, Monarch 120) to give the materials uniform optical absorption during laser irradiation. For comparison with the 1 and 2 wt% MWCNT composites fabricated by extrusion, samples of the same weight loading of nanotubes were fabricated by a solvent processing method. MWCNTs (NanoLab PD15L5-20) were horn sonicated (Cole-Parmer VCX500) at 20% amplitude for 45 min in p-xylene (ACROS CAS #10642-3) with 1% Triton X-100 (ACROS CAS #9002-93-1) surfactant. In this case, the nanotubes are nominally 15 nm in diameter and 5–20 lm in length. The mixture was heated on a hot plate to
Fig. 1. The dispersion of MWCNTs in the polypropylene matrix as seen in the cross sections of cryo-fractured samples for (a) 0.5, (b) 1.0, (c) 2.0, (d) 3.0, and (e) 5.0 wt% MWCNT composites.
168
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 2. Mass change for a single laser pulse at 3.1 J, 5 ms. MWCNT and MMT nanoclay composites comparison (left) and a closer view of the MWCNT composites (right).
180 °C and isotactic polypropylene (Scientific Polymer Products Inc– CAS #9003-07-0) was mixed in until it melted in the solution. The solution of MWCNT and PP was added to an ethanol anti-solvent (Acros CAS #64-17-5), upon which the PP-MWCNT composite precipitated out of solution. The mixture was magnetically stirred for 120 s to ensure completion of reaction. The PP-MWCNT composite was separated and allowed to dry for 16 h at 110 °C under vacuum. Samples for laser testing (discs of 10 mm diameter and 2.5 mm thickness) were compression molded on a Carver press (Carver 3851-0) at 190 °C under about 50 MPa of pressure and then allowed to cool slowly. 2.2. Laser pulse heating The Laser Pulse Heating (LPH) system is composed of an Nd:YAG laser (MegaWatt Lasers, Inc) that sends 1064 nm light into an optical fiber to improve the beam uniformity. The output from the fiber is passed through a 50 mm focal length collimating lens, followed by a 250 mm focal length objective lens, which produces a laser spot with a 3.0 mm diameter. The incident laser energy was measured using a beam splitter (Newport) and an energy meter (Ophir Nova). The laser pulse energy and pulse duration can be adjusted, but all samples in this paper were irradiated at 3.1 J pulses lasting 5 ms, corresponding to a fluence of 44 J/cm2 and an intensity of 8.8 kW/cm2. The samples were irradiated in air to give thermo-oxidative conditions. The reflectance of the samples was measured in a UV–VIS spectrophotometer (Perkin–Elmer Lambda 40) at 1064 nm and was found to be between 7.6% and 8.0% reflective for all MWCNT loadings; however, because the ablation plume from the sample has been observed to occlude the incident beam late in the pulse, and because this effect is more pronounced with increased nanotube loading, we do not claim that the energy absorbed by the sample is equal to the incident flux, nor that it is the same for all samples. 2.3. Characterization The mass of the samples was measured before and after irradiation on a microbalance with a readability of ±0.001 milligram
(Sartorius ME-36 S). Five samples for each condition were run for statistics. An FEI Helios Nanolab 600i Field Emission Electron Microscope (FE-SEM) operating at 5 kV was used for electron microscopy imaging. Samples were coated with a thin layer of gold to reduce charging effects. Transmission Electron Microscopy (TEM) was conducted on a JEOL 2010F operating at 200 kV. Material from the LPH regions was carefully removed with fine tweezers and placed on a carbon coated copper grid for viewing. Fourier Transform Infrared Spectroscopy (FTIR) was performed on a Perkin– Elmer Spectrum One Spectrometer operating in Attenuated Total Reflectance mode (ATR) between 4000 and 515 cm1 at 4 cm1 resolution. Raman spectroscopy of the composites was performed before and after LPH. A Renishaw Ramascope 2000 was used with a 785 nm wavelength laser at 1% power, 1200 line grating and a 20 objective. The so-called G-line and D-line, at 1585 cm1 and 1345 cm1, respectively, are characteristic features of the graphitic layers; the former corresponds to the tangential vibration of the carbon atoms, and the latter corresponds to defective graphitic structures. Comparison of the ratio of these two peaks’ intensities gives a measure of the quality of the bulk samples. Thermal diffusivity (a) was measured using the Laser Flash technique (ASTM E1461) at 25 and 100 °C. Bulk density (d) values were calculated from the sample’s geometry and mass and specific heat (Cp) was measured using differential scanning calorimetry. Thermal conductivity (k) values were calculated as a product of these quantities, i.e. k = aCpd.
3. Results and discussion 3.1. Dispersion of MWCNT The dispersion quality of MWCNT in the PP matrix was studied in the cryo-fractured cross-sections of composite material with varying MWCNT wt% and is seen in Fig. 1. Each micrograph is a representative view of each sample. A design of experiment was performed with varying twin screw extrusion processing parameters to determine the parameters that gave composites with the
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
169
seen in Fig. 2. For the same laser conditions, as the concentration of nanotubes increase from 0 to 5 wt%, the mass loss for a single laser pulse decreases by over 54%, from 532 lg to 243 lg. This is a significant improvement in thermal stability with low loading (5 wt%) of MWCNT. The mass loss from tests with the same LPH parameters on montmorillonite (MMT) nanoclay composites is shown for comparison. The details of the MMT composites study are described elsewhere [17]. Clearly, MWCNT composites show superior behavior to the nanoclay composites at far smaller filler loadings. This could be from factors such as higher thermal conductivity and an enhanced network structure in the MWCNT composites [19]. An increase in beam occlusion during irradiation is observed with increasing MWCNT content and this could account for some of the observed mass change differences with the pure PP, however, because this effect is observed later in the pulse, the general trend of decreasing mass loss with increasing MMT/ MWCNT content would still be observed even if beam occlusion could be prevented. The structure of the composite after LPH and the mechanisms by which a reduced mass loss is realized is explored in the following sections. 3.3. FTIR analysis
Fig. 3. Infrared spectroscopy (a) before and after LPH for the 5 wt% MWCNT material (inset shows magnified region of the after LPH spectrum) and (b) for all the composites after LPH.
best-achievable dispersion. Viscoelastic and electrical conductivity tests were conducted on the variously processed materials to determine which composites likely had the best MWCNT dispersion. The composites with the highest electrical conductivity and viscoelastic properties had likely formed a percolated network. While the dispersion of the nanotubes is generally good in this best-achieved material, there was still an occasional large agglomerate of nanotubes in the matrix that measured tens of microns in length. These agglomerates are likely masterbatch that has not been thoroughly distributed and dispersed in the pure PP [18]. Perfect dispersion of individual nanotubes throughout the matrix is desirable, as this would give the most networked structure.
Fig. 3 shows the FTIR spectra that were collected from the LPH regions of the nanotube composites. Fig. 3a shows representative spectra before and after LPH for the 5 wt% MWCNT composite with an inset showing a magnified and baselined portion of the spectrum, while Fig. 3b shows the spectra for all the composites after LPH. A pure PP material has strong characteristic absorption peaks for CAH bending at 1450 and 1375 cm1 and CAH stretching at 2800–3000 cm1 from the CAH bonds along the aliphatic polymer chains, as well as the CH3 side groups along these chains. The absorption bands associated with the polymer decrease significantly in the LPH samples as the nanotube content increases. For the 5 wt% MWCNT composite, only very weak absorption peaks remain after LPH. It is likely that the surface structure consists primarily of a nanotube network, with little polymer remaining. As the nanotubes are heated rapidly by the laser heating, it is possible that defects in the nanotube structure could react with oxygen in the air to form CAO and C@O bonds. The FTIR spectra do not show strong absorption bands for these moieties, although a faint peak for CAO is seen in the 5 wt% LPH composite at 1190 cm1. It is most visible in the 5 wt% LPH sample and fades away as the amount of MWCNT in the composite decreases. A baselined and higher magnification inset spectrum in Fig. 3a shows a small peak for CAO at 1190 cm1. The thermal shock may actually remove or limit oxygen containing groups from forming on the nanotubes [20], as this is a common practice to remove oxygen groups from the surface of graphite oxide to form graphene platelets. The heating rates in these current experiments are orders of magnitude higher than those used in thermal shock of graphene to remove the oxygen functionalities, therefore, it may not be surprising to find little evidence for oxygen functionalities on the MWCNTs. FTIR has showed that the polymer becomes less prevalent on the surface after LPH. Raman spectroscopy was then used to study the nanotube structure that remains on the LPH surface.
3.2. Mass loss data
3.4. Raman spectroscopy
In thermal stability testing of polymers and polymer composites, the mass loss of the material is monitored as a function of temperature as the temperature increases linearly, or is held isothermally at an elevated temperature. Analogously, we record the mass loss of the PP/MWCNT composite after LPH as a measure of thermal stability and resistance to high temperature degradation. The results of the mass loss for the MWCNT composites are
The Raman data for the composite materials was recorded before and after LPH. The characteristic peaks at the so-called Gline and D-lines (1585 cm1 and 1345 cm1) were analyzed, and the average D/G ratio for all samples before and after LPH was unchanged at 1.69. This ratio represents a nanotube that has a significant amount of defects in the structure. Since the D/G ratio data is essentially the same before and after LPH, the defects have likely
170
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 4. The surface of the nanotube composites after LPH with (a) 0, (b) 0.5, (c) 1.0, (d) 2.0, (e) 3.0, (f) 5.0 wt% MWCNT. Islands of high concentration of nanotubes are clearly seen in each composite, with a more contiguous networked layer being formed in the higher weight % composites.
3.5. LPH microstructures
Fig. 5. Cryo-fractured cross-section of 5 wt% MWCNT composite through the LPH region. The top surface of the LPH spot (A) and the composite below the surface (B) of the LPH spot.
arisen during nanotube and composite processing. It is known that melt-mixing processing methods, such as those used in this work, can create defects in nanotubes even resulting in the shortening of their length [21]. There does not appear to be a higher defect structure after LPH according to the Raman data. This suggests that the nanotube structure remains largely intact during LPH, while the polymer structure mostly degrades as evidenced by the disappearance of the Raman peaks [22] at 809 and 840 cm1. These peaks represent the crystalline order and the amorphous polypropylene structure, respectively.
Figs. 4 and 5 show the surface and cross section of the nanotube composites after LPH. In Fig. 4 the LPH surface changes drastically as the amount of MWCNTs is increased from 0 to 5 wt%. The surface of the pure PP is smooth and represents a surface that has experienced clean ablation of the polymer into the gaseous phase. Any melting that occurred was likely followed by rapid solidification of the polymer. As the amount of MWCNTs increase in the composite, the surface begins to show islands of high concentration of nanotubes. Eventually, the composites with higher weight loading of MWCNTs show a relatively contiguous layer of networked nanotubes across the surface, in addition to some high concentration islands of MWCNTs. Islands of MWCNTs could be agglomerates from the original composite that have been exposed as islands as the polymer around them is degraded by the laser heating, or could have formed by the localized accumulation of poorly dispersed nanotubes; the latter has been reported previously. [5,11]. The proposed formation mechanism for the latter explanation entails the result of fluid convection accompanying bubble formation and rise of bubbles to the surface through the molten sample layer and bursting of the bubbles at the surface. A representative cross-section of the 5 wt% LPH surface in Fig. 5 shows the networked nanotube structure on the surface, which we denote ‘‘region A’’ and the subsurface composite, denoted ‘‘region B’’. There is no unambiguous evidence of vigorous bubble formation in Fig. 5. Examination of the 0.5, 1.0, 2.0 and 3.0 wt% MWCNT subsurface structures also did not show evidence of vigorous bubble formation. The agglomerates seen were most likely due to poorly dispersed starting material, exposed during LPH, with bubbling having little or no effect on island formation during
171
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
their structure. The predominant features were the MWCNT, which can be seen in Fig. 6, in (a) and (b). The image in (b) clearly shows the graphitic walls that comprise the nanotubes. The tubes are not perfectly straight, and show significant waviness in the graphitic structure, which could be due to lattice bending and atomic vacancies. This is consistent with the description of Raman data presented in Section 3.4, which suggests the nanotubes have a significant density of defects in their structure. Another feature seen in the TEM is shown in (c), which is a carbonaceous particle with a concentric turbostratic nanostructure. These particles are nearly identical to previously published images of carbonaceous soot [23], both in structure and size. In this study, it is believed that these particles are formed from degraded polymer. While the soot particles and carbon black had similar structures, the soot particles were observed after LPH in composites that did not initially contain any carbon black. The TEM observations confirm the spectroscopic data showing that nanotubes with a defective structure remain intact on the LPH surface, along with small amounts of degraded polymer. 3.7. Thermal conductivity data It is known that MWCNTs can increase the thermal conductivity of a polymer [24,25]. MWCNTs have a high intrinsic thermal conductivity that is several orders of magnitude higher than a typical polymer. When the nanotubes are added to a polymer, the interfaces created, dispersion, and composite structure all play significant roles in the resulting thermal conductivity of the composite. The thermal conductivity of our nanotube composites was measured and was shown to have approximately a 25% increase at 25 °C when 5 wt% MWCNT was added. This increase is consistent with what others have reported [7]. The data are shown in Table 1. In theory, as the laser light begins to heat the surface of the composite, the higher thermal conductivity composites should be able to conduct heat away from the surface faster than the pure PP resulting in a slower rise to the degradation and ablation temperature and hence, a lower mass loss. This could play a small role in the improved performance of the higher weight % nanotube composites with respect to the pure PP. 3.8. Networked structure
Fig. 6. TEM images of the nanotube network layer on the LPH surface with (a) entangled individual nanotubes, (b) an individual nanotube with observable graphitic wall structure (c) carbonaceous soot particles (entire field of view) showing a wavy, concentric turbostratic nanostructure. The soot microstructure is similar to that of the carbon black particles.
LPH. Because the time scales used during LPH are orders of magnitude smaller than those used in flammability testing, some of the mechanisms observed in flammability tests may not be observable in LPH testing, such as island growth due to bubbling processes. 3.6. Transmission electron microscopy TEM of the networked layer on the surface of the LPH region was conducted to more closely examine the particles present and
The networked structure layer acts as a thermal shield to reduce the exposure of the underlying polymer composite and reduce degradation. As seen in the cross-section of Fig. 5, the layer is on the order of several microns thick. The layer is generally contiguous, but can have regions where exposed polymer can be seen from the plan view of the LPH region. The network of individual nanotubes can also take on a bundle-like secondary structure as seen before in flammability testing [8]. The nanotube network must be responsible for the reduced mass loss of the composites when subjected to the high temperatures caused by photothermal heating by the laser. When the polymer composite absorbs the laser light, the temperature will increase as energy is transferred to
Table 1 Thermal conductivity measurements. Material
Thermal conductivity (W/m-K) @ 25 °C
Thermal conductivity (W/m-K) @ 100 °C
Pure PP 0.5 wt% MWCNT 1.0 wt% MWCNT 2.0 wt% MWCNT 3.0 wt% MWCNT 5.0 wt% MWCNT
0.218 0.228 0.224 0.242 0.252 0.273
0.220 0.216 0.215 0.238 0.248 0.267
172
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 7. (a) Cryo-fractured cross-section of a solvent processed PP/MWCNT composite with 2 wt% MWCNT. Individual nanotubes are seen protruding from the surface. The surfaces after LPH for the (b) 1 wt% MWCNT and (c) 2 wt% MWCNT composites (d) the fracture surface through the LPH region for 2 wt% MWCNT solvent-processed composite with the top surface of the LPH spot (A) and the composite below the surface (B) of the LPH spot.
the polymer chains, nanotubes, and carbon black. In the presence of air, the polypropylene chains undergo oxidative dehydrogenation accompanied by hydrogen abstraction [26]. Extensive chain scission and volatile fragment formation results in the mass loss and degradation of the polymer. The networked nanotube structure must either reduce the exposure of the polymer to heat or reduce the mass loss in the form of volatile formation and escape from the surface, or both of these. Qualitatively, we observe a similar networked layer providing a protective barrier to the incident radiative flux as reported previously at slow heating rates [6,7]. To further establish that this is the dominant mechanism providing the thermal stability at high heating rates, we are examining the behavior of composites exposed to multiple laser pulses, variable pulse durations, and pulse energies and this will be the subject of a future publication. It is evident that the nanotubes remain after the exposure to elevated temperatures at high heating rate during LPH. The temperatures achieved are not known at this time and is the topic of ongoing work in our laboratory. It has been reported that MWCNTs are stable in air to at least 420 °C under slow heating rates of 2 °C/ min in thermogravimetric experiments [27]. That same work also showed that an increase to 20 °C/min heating rate increased the thermal stability of MWCNTs in air by over 100 °C at the 50% weight loss mark. Others have shown the oxidative degradation of graphitized MWCNTs took place between 640 and 780 °C [28] with the thermal stability in oxidative conditions increasing in MWCNTs that have fewer defects. The defects likely lead to oxidative instability by providing edges and dangling bonds by which each nanotube can be ‘‘unzipped’’ and peeled away. While the Raman data suggests we have a MWCNT structure with defects before and after LPH, we see an intact networked nanotube structure after LPH. We can infer that either we did not achieve temperatures above 420 °C, or that the increased thermal stability of the MWCNT is a result of the very high heating rates encountered during LPH. The rapid cooling rate after LPH also limits the time spent at elevated temperatures. Some preliminary temperature analysis in our laboratory suggests that the temperature exceeds 420 °C, and we therefore tentatively identify the latter interpretation as the more plausible. 3.9. Solvent-processed composites Composites with 1 and 2 wt% MWCNT were fabricated with a solvent-based method in order to compare the mass loss behavior during LPH with those fabricated by twin screw extrusion. Melt blending is generally less effective at dispersing nanotubes in polymers compared to solvent blending methods. It was expected that solvent processed samples could have better dispersion [29]. The cryo-fractured cross-section of the 2 wt% MWCNT composite fabricated by solvent-based method is shown in Fig. 7a and it shows
very good dispersion quality. During LPH under the conditions previously described, the composites fabricated with solvent-based methods had significantly less mass loss than the composites fabricated by extrusion. The 1.0 and 2.0 wt% MWCNT solvent-processed composites had a mass loss of 293 ± 19 and 258 ± 2 lg, respectively. Compare this with the extrusion-processed composites, which had a mass loss of 467 ± 6 and 391 ± 9 lg, respectively. This represents a 36% (average) reduction in mass loss for the two composites studied. One noteworthy contributing factor could be the MWCNT aspect ratio, as mass loss rates of composites have been shown to decrease with MWCNT of increasing aspect ratio. [30]. The nanotubes used in our solvent-based samples had a larger initial aspect ratio (nominal A.R. = 833) than those used in the extrusion samples (nominal A.R. = 157). It is possible that differences in thermal conductivity of the solvent-based material could account for differences in mass change compared to the extruded samples; the thermal properties of the solvent-based material have not yet been measured. However, we consider the more likely explanation to be the improved dispersion quality of the solventprocessed samples compared to the extruded samples. The solvent-processed samples appear to have good dispersion with only small areas of low nanotube content on the cryo-fractured surface. The LPH surfaces in Fig. 7b–d appear to be a generally contiguous network of nanotubes without any areas of sparse or agglomerated concentration as seen in the extrusion processed LPH surfaces. The seemingly better dispersion quality of the MWCNTs and the more contiguous networked layer on the surface of the composite after LPH resulted in the greater thermal stability of the solvent-processed material compared to the extruded material, as measured by the decreased mass loss. Additional characterization of the solvent-processed material will appear in a future publication.
4. Conclusions Nanocomposites of PP/MWCNT were fabricated using twin screw extrusion, subjected to LPH, and subsequently characterized using mass loss data, spectroscopy, and microscopy. The Raman and FTIR spectroscopy showed that the nanotubes retain their structure during LPH and that a nanotube-rich layer formed at the surface of the composites during laser heating. The presence of this layer was confirmed by electron microscopy observations. If the nanotubes are well-dispersed and form a contiguous networked layer, it can act as a protective shield for the underlying polymer, preventing degradation. The presence of MWCNTs in the polymer resulted in a lower mass loss of the composite when subjected to highly transient heating during LPH. The PP-5 wt% MWCNT composite showed a 54% reduction in mass loss per laser pulse compared to pure PP. The nanotube composites were also shown to outperform nanoclay composites under the same
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
conditions, despite having a significantly lower weight loading of additive. Solvent-processed samples, which appear to be well-dispersed, display improved thermal stability compared to the extruded materials, as indicated by a lower mass loss during LPH. While this processing method may show better thermal stability results due to better MWCNT dispersion, the processability of bulk nanotube composites by extrusion is higher than solventbased methods. In conclusion, the MWCNTs can reduce degradation of the composite during LPH by forming a protective surface layer that acts as a shield for the underlying polymer and by dissipating heat more quickly from increased thermal conductivity. These observations are consistent with those seen during previously published slow heating rate studies and this shows that while some phenomena are unique to transient heating and laser irradiation of polymer nanocomposites, there are some commonalities between the two regimes.
[12] [13]
[14]
[15] [16]
[17]
[18]
[19]
References [1] Thostenson ET, Ren Z, Chou TW. Advances in the science and technology of carbon nanotubes and their composites: a review. Compos Sci Technol 2001;61(13):1899–912. [2] Giannelis EP. Polymer layered silicate nanocomposites. Adv Mater 1996;8(1):29–35. [3] Paul DR, Robeson LM. Polymer technology: nanocomposites. Polymer 2008;49(15):3187–204. [4] Gilman JW, Jackson CL, Morgan AB, Harris RH, Manias E, Giannelis EP, et al. Flammability properties of polymer-layered-silicate nanocomposites. Polypropylene and polystyrene nanocomposites. Chem Mater 2000;12:1866–73. [5] Kashiwagi T, Du F, Douglas JF, Winey KI, Harris RH. Nanoparticle networks reduce the flammability of polymer nanocomposites. Nat Mater 2005;4(12):928–33. [6] Kashiwagi T, Grulke E, Hilding J, Harris R, Awad W, Douglas J. Thermal degradation and flammability properties of poly(propylene)/carbon nanotube composites. Macromol Rapid Comm 2002;23:761–5. [7] Kashiwagi T, Grulke E, Hilding J, Groth K, Harris R, Butler K, et al. Thermal and flammability properties of polypropylene/carbon nanotube nanocomposites. Polymer 2004;45:4227–39. [8] Bocchini S, Frache A, Camino G, Claes M. Polyethylene thermal oxidative stabilisation in carbon nanotubes based nanocomposites. Eur Polym J 2007;43:3222–35. [9] Schartel B, Potschke P, Knoll U, Abdel-Goad M. Fire behaviour of polyamide 6/ multiwall carbon nanotube. Eur Polym J 2005;41:1061–70. [10] Boccini S, Annibale E, Frache A, Camino G. MWNT surface self-assembling in fire retardant polyethylene-carbon nanotubes composites. E-Polymers 2008;20:1–11. [11] Kashiwagi T, Du F, Winey KI, Groth KM, Shields JR, Bellayer SP, et al. Flammability properties of polymer nanocomposites with single-walled
[20]
[21]
[22] [23]
[24] [25] [26]
[27]
[28]
[29] [30]
173
carbon nanotubes: effects of nanotube dispersion and concentration. Polymer 2005;46:471–81. Cote PJ, Kendall G, Todaro ME. Laser pulse heating of gun bore coatings. Surf Coat Tech 2001;146–147:65–9. Warrender JM, Mulligan CP, Underwood JH. Analysis of thermo-mechanical cracking in refractory coatings using variable pulse-duration laser pulse heating. Wear 2007;263:1540–4. Bartolucci SF, Zhang W, Warrender J, Lee S. Structure and morphology of carbon nanofiber polycarbosilane precursor derived nanocomposites. Nanosci Nanotechnol Lett 2010;2:1–5. Price D, Milnes GJ, Gao F. Laser pyrolysis as a model for fire behaviour behind the flame front. Polym Degrad Stabil 1996;54:235–40. Gao F, Price D, Milnes GJ, Eling B, Lindsay CI, McGrail PT. Laser pyrolysis of polymers and its relation to polymer fire behavior. J Anal Appl Pyrol 1997;40– 41:217–31. Bartolucci SF, Supan KE, Wiggins JS, LaBeaud L, Warrender JM. Thermal stability of polypropylene-clay nanocomposites subjected to laser pulse heating. Polym Degrad Stabil 2013;98:2497–502. Villmow T, Potschke P, Pegel S, Haussler L, Kretzschmar B. Influence of twinscrew extrusion conditions on the dispersion of multi-walled carbon nanotubes in a poly(lactic acid) matrix. Polymer 2008;49:3500–9. Rahatekar SS, Zammarano M, Matko S, Koziol KK, Windle AH, Nyden M, et al. Effect of carbon nanotubes and montmorillonite on the flammability of epoxy nanocomposites. Polym Degrad Stabil 2010;95:870–9. Schniepp HC, Li J-L, McAllister MJ, Sai H, Herrera-Alonso M, Adamson DH, et al. Functionalized single graphene sheets derived from splitting graphite oxide. J Phys Chem B 2006;110:8535–9. Andrews R, Jacques D, Minot M, Rantell T. Fabrication of carbon multiwall nanotube/polymer composites by shear mixing. Macromol Mater Eng 2002;287(6):395–403. Minogianni C, Gatos K, Galiotis. Estimation of crystallinity in isotropic polypropylene with raman spectroscopy. Appl Spectrosc 2005;59(9):1141–7. Posfai M, Gelencser A, Simonics R, Arato K, Li J, Hobbs PV, et al. Atmospheric tar balls: particles from biomass and biofuel burning. J Geophys Res 2004;109:D06213. Yang Y, Grulke EA, Zhang ZG, Wu G. Thermal and rheological properties of carbon nanotube-in-oil dispersions. J Appl Phys 2006;99:114307. Han Z, Fina A. Thermal conductivity of carbon nanotubes and their polymer nanocomposites: a review. Prog Polym Sci 2011;36:914–44. Zanetti M, Camino G, Reichert P, Mulhaupt R. Thermal behaviour of poly(propylene) layered silicate nanocomposites. Macrom Rapid Comm 2001;22:176–80. Mahajan A, Kingon A, Kukovecz A, Konya Z, Vilarinho P. Studies on the thermal decomposition of multiwall carbon nanotubes under different atmospheres. Mater Lett 2013;90:165–8. Bom D, Andrews R, Jacques D, Anthony J, Chen B, Meier MS, et al. Thermogravimetric analysis of the oxidation of multiwalled carbon nanotubes: evidence for the role of defect sites in carbon nanotube chemistry. Nano Lett 2002;2:615–9. Moniruzzaman M, Winey KI. Polymer nanocomposites containing carbon nanotubes. Macromolecules 2006;39:5194–205. Cipiriano BH, Kashiwagi T, Raghavan SR, Yang Y, Grulke EA, Yamamoto K, et al. Effects of aspect ratio of MWNT on the flammability properties of polymer nanocomposites. Polymer 2007;48:6086–96.
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Laser-induced thermo-oxidative degradation of carbon nanotube/polypropylene nanocomposites Stephen F. Bartolucci a,⇑, Karen E. Supan a,b, Jeffrey M. Warrender a, Christopher E. Davis a,c, Lawrence La Beaud d, Kyler Knowles d, Jeffrey S. Wiggins d a
U.S. Army Benét Laboratories, Armaments Research Development and Engineering Center, 1 Buffington St., Watervliet, NY 12189, USA Norwich University, Department of Mechanical Engineering, Northfield, VT 05663, USA Massachusetts Institute of Technology, Department of Chemical Engineering, Cambridge, MA 02139, USA d University of Southern Mississippi, School of Polymers and High Performance Materials, Hattiesburg, MS 39406, USA b c
a r t i c l e
i n f o
Article history: Received 5 May 2014 Received in revised form 2 September 2014 Accepted 24 September 2014 Available online 5 October 2014 Keywords: A. Nanocomposites A. Carbon nanotubes A. Nanoclays B. Thermal properties B. High-temperature properties
a b s t r a c t Nanocomposites of multi-walled carbon nanotubes and polypropylene were fabricated using twin screw extrusion and subjected to heating rates on the order of 100,000 degrees per second using laser pulse heating. Raman and infrared spectroscopy, as well as electron microscopy observations, showed that the nanotubes retain their structure and that a nanotube-rich layer formed at the surface of the composites during laser heating. By varying the weight fraction of filler in the composites, a 54% reduction in mass loss per laser pulse was achieved, compared to the pure polymer. The nanotube composites also displayed lower mass loss than nanoclay composites under the same conditions, despite having a significantly lower weight fraction of additive. Further reduction in mass loss per laser pulse was achieved in solvent-processed samples, which appear to have better dispersion than the extrusion-processed composites. We attribute these observations to the formation of a nanotube network formed on the surface, which acts as a protective barrier, slowing the degradation of the underlying polymer. Several observations made in this study are consistent with those reported in previously published slow heating rate studies, showing that while some phenomena are unique to transient heating, there are some commonalities between the two regimes. Published by Elsevier Ltd.
1. Introduction Organic-based polymers are increasingly being used to replace inorganic materials in some applications where the material may be subjected to elevated temperatures. Since most polymers have a relatively low melting and thermal stability temperature compared to materials like metals and ceramics, the addition of particles to the polymer matrix is often needed to produce a composite with improved thermal stability. There has been a substantial amount of research on the use of nanoparticles as additives to polymers to improve both the mechanical and thermal properties. Fillers such as nanoclay and carbon nanotubes have been among the most widely investigated for many years [1–3]. Carbon nanotubes are high aspect ratio carbon nanostructures that may be thought of qualitatively as graphene layers that have been rolled into a tubular shape. Carbon nanotubes exhibit extremely high strength, modulus, and thermal and electrical conductivity, and
⇑ Corresponding author. http://dx.doi.org/10.1016/j.compscitech.2014.09.018 0266-3538/Published by Elsevier Ltd.
when dispersed well in a polymer matrix, can form a network structure throughout the composite. The thermal stability or flammability of carbon nanotube nanocomposites has been studied previously [4,5]. In traditional thermal stability studies, the nanocomposite is heated at rates on the order of 10 °C/min. Using thermogravimetric analysis (TGA), the mass change of the sample is recorded and the thermal stability is determined as a measure of the mass loss versus temperature. The thermal degradation of the nanocomposite can be determined under both oxidative and non-oxidative conditions. Kashiwagi et al. [6,7] showed that 1 and 2 vol% multi-walled carbon nanotubes (MWCNT) in polypropylene increased the thermal stability of the polymer during TGA tests in air and nitrogen, and decreased the heat release rate in a cone calorimeter and gasification tests. The high aspect ratio carbon nanotubes formed a networked structure covering the sample surface, acting as a thermal shield and slowing the degradation of the underlying polymer. Incorporation of MWCNTs into other polymer materials has shown similar behavior [8–10]. It has been shown that effective dispersion of the nanotubes in the polymer matrix is critical for the reduction
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
in mass loss rate during flammability tests [11]. Well-dispersed nanotubes form a homogeneous network layer, whereas poorly dispersed nanotubes form a discontinuous layer consisting of fragmented islands of networked nanotubes on the composite surface. Well-dispersed nanotubes also lead to a higher viscosity of the polymer melt, making bubble formation and coalescence more difficult. Decreased bubbling and bursting leads to a nanotube network layer that is more intact. This continuous protective network layer leads to improved fire retardance. Carbon nanotubes have also been shown to provide superior enhancement to fire retardance compared to other carbon nanoadditives, such as carbon nanofibers and carbon black [5]. Certain applications, such as ballistic environments, have heating rates that are many orders of magnitude larger than those typically achievable in thermogravimetric analysis or flammability studies, necessitating the use of alternative methodologies to study such regimes. Laser pulse heating (LPH) has been used to achieve heating rates on the order of 107 K/min [12–14]. In order to achieve photothermal heating and to avoid photochemical effects (as seen with the use of UV wavelength lasers), an infrared laser with variable pulse duration on the order of milliseconds is used to heat the sample surface. By adding the same weight percentage of carbon black to each of the composites, uniform optical absorption of the laser light among samples of various compositions can be achieved [15,16]. Studying the thermal degradation of polymer nanocomposites with this method combines phenomena seen in both polymer ablation and polymer thermal stability/flammability research. We can observe ablation of material, melting and bubbling of the heated polymer, and degradation of the polymer chains. This group has previously studied highly transient heating of polymer/nanoclay nanocomposites [17] and observed decreased mass loss with increasing amounts of montmorillonite nanoclay in a polypropylene matrix. Spectroscopic characterization showed that a silicate-rich layer forms on the surface of the LPH region of the composite, providing a mechanism of protection for the underlying polymer. While carbon nanotube composites have been studied at slow heating rates, they have not been studied at high heating rates. The objective of this study is to determine how carbon nanotube/polymer composites behave under highly transient heating and whether correlations can be made to the behaviors observed during low-heating rate thermal stability and flammability testing.
167
2. Experimental 2.1. Composite preparation Isotactic Polypropylene (PP) pellets (Total Petrochemical Type 3371, MFR = 2.8 g/10 min at 230 °C, 2.16 kg) were dried for 16 h under vacuum at room temperature. A master batch (Nanocyl PP2001CNT/Thermoplastic) at 20 wt% MWCNTs (with average 9.5 nm diameter and 1.5 lm length) were compounded with appropriate amounts of the neat PP resin to achieve desired final loadings using a co-rotating twin screw extruder and a controlled feeder. Zone 1 was held at 170 °C and zones 2–4 and the die were held at 220 °C and screw speed at 200 RPM. Pellets were extruded through a 1/1600 circular die and passed through an ice bath prior to being pelletized. They were then passed through the extruder for a second time under identical conditions. Using electrical conduction as a measurement for MWCNT dispersion, it was found that two sequential passes through the extruder (under identical conditions) improved dispersion. Additional passes through the extruder yielded increased conductivity up to five passes, though physical properties did not show a corresponding increase beyond two passes. For this reason (as well as additional processing time), the compounding step was limited to two passes. Composites were made with 0.5, 1.0, 2.0, 3.0 and 5.0 wt% MWCNT. Compounded PP/MWCNT pellets were melt-pressed at 200 °C in a steel die with PTFE release film. No pressure was applied for 5 min while the melt press equilibrated and the PP began to melt/flow. Pressure was then applied in 13.7 MPa increments to 68.9 MPa. The pellets were left under 68.9 MPa for 300 s, at which time they were removed and cooled to room temperature. The newly-formed discs were removed from the die and any flash material was trimmed from the edges. All materials were compounded with 1 wt% carbon black (Cabot, Monarch 120) to give the materials uniform optical absorption during laser irradiation. For comparison with the 1 and 2 wt% MWCNT composites fabricated by extrusion, samples of the same weight loading of nanotubes were fabricated by a solvent processing method. MWCNTs (NanoLab PD15L5-20) were horn sonicated (Cole-Parmer VCX500) at 20% amplitude for 45 min in p-xylene (ACROS CAS #10642-3) with 1% Triton X-100 (ACROS CAS #9002-93-1) surfactant. In this case, the nanotubes are nominally 15 nm in diameter and 5–20 lm in length. The mixture was heated on a hot plate to
Fig. 1. The dispersion of MWCNTs in the polypropylene matrix as seen in the cross sections of cryo-fractured samples for (a) 0.5, (b) 1.0, (c) 2.0, (d) 3.0, and (e) 5.0 wt% MWCNT composites.
168
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 2. Mass change for a single laser pulse at 3.1 J, 5 ms. MWCNT and MMT nanoclay composites comparison (left) and a closer view of the MWCNT composites (right).
180 °C and isotactic polypropylene (Scientific Polymer Products Inc– CAS #9003-07-0) was mixed in until it melted in the solution. The solution of MWCNT and PP was added to an ethanol anti-solvent (Acros CAS #64-17-5), upon which the PP-MWCNT composite precipitated out of solution. The mixture was magnetically stirred for 120 s to ensure completion of reaction. The PP-MWCNT composite was separated and allowed to dry for 16 h at 110 °C under vacuum. Samples for laser testing (discs of 10 mm diameter and 2.5 mm thickness) were compression molded on a Carver press (Carver 3851-0) at 190 °C under about 50 MPa of pressure and then allowed to cool slowly. 2.2. Laser pulse heating The Laser Pulse Heating (LPH) system is composed of an Nd:YAG laser (MegaWatt Lasers, Inc) that sends 1064 nm light into an optical fiber to improve the beam uniformity. The output from the fiber is passed through a 50 mm focal length collimating lens, followed by a 250 mm focal length objective lens, which produces a laser spot with a 3.0 mm diameter. The incident laser energy was measured using a beam splitter (Newport) and an energy meter (Ophir Nova). The laser pulse energy and pulse duration can be adjusted, but all samples in this paper were irradiated at 3.1 J pulses lasting 5 ms, corresponding to a fluence of 44 J/cm2 and an intensity of 8.8 kW/cm2. The samples were irradiated in air to give thermo-oxidative conditions. The reflectance of the samples was measured in a UV–VIS spectrophotometer (Perkin–Elmer Lambda 40) at 1064 nm and was found to be between 7.6% and 8.0% reflective for all MWCNT loadings; however, because the ablation plume from the sample has been observed to occlude the incident beam late in the pulse, and because this effect is more pronounced with increased nanotube loading, we do not claim that the energy absorbed by the sample is equal to the incident flux, nor that it is the same for all samples. 2.3. Characterization The mass of the samples was measured before and after irradiation on a microbalance with a readability of ±0.001 milligram
(Sartorius ME-36 S). Five samples for each condition were run for statistics. An FEI Helios Nanolab 600i Field Emission Electron Microscope (FE-SEM) operating at 5 kV was used for electron microscopy imaging. Samples were coated with a thin layer of gold to reduce charging effects. Transmission Electron Microscopy (TEM) was conducted on a JEOL 2010F operating at 200 kV. Material from the LPH regions was carefully removed with fine tweezers and placed on a carbon coated copper grid for viewing. Fourier Transform Infrared Spectroscopy (FTIR) was performed on a Perkin– Elmer Spectrum One Spectrometer operating in Attenuated Total Reflectance mode (ATR) between 4000 and 515 cm1 at 4 cm1 resolution. Raman spectroscopy of the composites was performed before and after LPH. A Renishaw Ramascope 2000 was used with a 785 nm wavelength laser at 1% power, 1200 line grating and a 20 objective. The so-called G-line and D-line, at 1585 cm1 and 1345 cm1, respectively, are characteristic features of the graphitic layers; the former corresponds to the tangential vibration of the carbon atoms, and the latter corresponds to defective graphitic structures. Comparison of the ratio of these two peaks’ intensities gives a measure of the quality of the bulk samples. Thermal diffusivity (a) was measured using the Laser Flash technique (ASTM E1461) at 25 and 100 °C. Bulk density (d) values were calculated from the sample’s geometry and mass and specific heat (Cp) was measured using differential scanning calorimetry. Thermal conductivity (k) values were calculated as a product of these quantities, i.e. k = aCpd.
3. Results and discussion 3.1. Dispersion of MWCNT The dispersion quality of MWCNT in the PP matrix was studied in the cryo-fractured cross-sections of composite material with varying MWCNT wt% and is seen in Fig. 1. Each micrograph is a representative view of each sample. A design of experiment was performed with varying twin screw extrusion processing parameters to determine the parameters that gave composites with the
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
169
seen in Fig. 2. For the same laser conditions, as the concentration of nanotubes increase from 0 to 5 wt%, the mass loss for a single laser pulse decreases by over 54%, from 532 lg to 243 lg. This is a significant improvement in thermal stability with low loading (5 wt%) of MWCNT. The mass loss from tests with the same LPH parameters on montmorillonite (MMT) nanoclay composites is shown for comparison. The details of the MMT composites study are described elsewhere [17]. Clearly, MWCNT composites show superior behavior to the nanoclay composites at far smaller filler loadings. This could be from factors such as higher thermal conductivity and an enhanced network structure in the MWCNT composites [19]. An increase in beam occlusion during irradiation is observed with increasing MWCNT content and this could account for some of the observed mass change differences with the pure PP, however, because this effect is observed later in the pulse, the general trend of decreasing mass loss with increasing MMT/ MWCNT content would still be observed even if beam occlusion could be prevented. The structure of the composite after LPH and the mechanisms by which a reduced mass loss is realized is explored in the following sections. 3.3. FTIR analysis
Fig. 3. Infrared spectroscopy (a) before and after LPH for the 5 wt% MWCNT material (inset shows magnified region of the after LPH spectrum) and (b) for all the composites after LPH.
best-achievable dispersion. Viscoelastic and electrical conductivity tests were conducted on the variously processed materials to determine which composites likely had the best MWCNT dispersion. The composites with the highest electrical conductivity and viscoelastic properties had likely formed a percolated network. While the dispersion of the nanotubes is generally good in this best-achieved material, there was still an occasional large agglomerate of nanotubes in the matrix that measured tens of microns in length. These agglomerates are likely masterbatch that has not been thoroughly distributed and dispersed in the pure PP [18]. Perfect dispersion of individual nanotubes throughout the matrix is desirable, as this would give the most networked structure.
Fig. 3 shows the FTIR spectra that were collected from the LPH regions of the nanotube composites. Fig. 3a shows representative spectra before and after LPH for the 5 wt% MWCNT composite with an inset showing a magnified and baselined portion of the spectrum, while Fig. 3b shows the spectra for all the composites after LPH. A pure PP material has strong characteristic absorption peaks for CAH bending at 1450 and 1375 cm1 and CAH stretching at 2800–3000 cm1 from the CAH bonds along the aliphatic polymer chains, as well as the CH3 side groups along these chains. The absorption bands associated with the polymer decrease significantly in the LPH samples as the nanotube content increases. For the 5 wt% MWCNT composite, only very weak absorption peaks remain after LPH. It is likely that the surface structure consists primarily of a nanotube network, with little polymer remaining. As the nanotubes are heated rapidly by the laser heating, it is possible that defects in the nanotube structure could react with oxygen in the air to form CAO and C@O bonds. The FTIR spectra do not show strong absorption bands for these moieties, although a faint peak for CAO is seen in the 5 wt% LPH composite at 1190 cm1. It is most visible in the 5 wt% LPH sample and fades away as the amount of MWCNT in the composite decreases. A baselined and higher magnification inset spectrum in Fig. 3a shows a small peak for CAO at 1190 cm1. The thermal shock may actually remove or limit oxygen containing groups from forming on the nanotubes [20], as this is a common practice to remove oxygen groups from the surface of graphite oxide to form graphene platelets. The heating rates in these current experiments are orders of magnitude higher than those used in thermal shock of graphene to remove the oxygen functionalities, therefore, it may not be surprising to find little evidence for oxygen functionalities on the MWCNTs. FTIR has showed that the polymer becomes less prevalent on the surface after LPH. Raman spectroscopy was then used to study the nanotube structure that remains on the LPH surface.
3.2. Mass loss data
3.4. Raman spectroscopy
In thermal stability testing of polymers and polymer composites, the mass loss of the material is monitored as a function of temperature as the temperature increases linearly, or is held isothermally at an elevated temperature. Analogously, we record the mass loss of the PP/MWCNT composite after LPH as a measure of thermal stability and resistance to high temperature degradation. The results of the mass loss for the MWCNT composites are
The Raman data for the composite materials was recorded before and after LPH. The characteristic peaks at the so-called Gline and D-lines (1585 cm1 and 1345 cm1) were analyzed, and the average D/G ratio for all samples before and after LPH was unchanged at 1.69. This ratio represents a nanotube that has a significant amount of defects in the structure. Since the D/G ratio data is essentially the same before and after LPH, the defects have likely
170
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 4. The surface of the nanotube composites after LPH with (a) 0, (b) 0.5, (c) 1.0, (d) 2.0, (e) 3.0, (f) 5.0 wt% MWCNT. Islands of high concentration of nanotubes are clearly seen in each composite, with a more contiguous networked layer being formed in the higher weight % composites.
3.5. LPH microstructures
Fig. 5. Cryo-fractured cross-section of 5 wt% MWCNT composite through the LPH region. The top surface of the LPH spot (A) and the composite below the surface (B) of the LPH spot.
arisen during nanotube and composite processing. It is known that melt-mixing processing methods, such as those used in this work, can create defects in nanotubes even resulting in the shortening of their length [21]. There does not appear to be a higher defect structure after LPH according to the Raman data. This suggests that the nanotube structure remains largely intact during LPH, while the polymer structure mostly degrades as evidenced by the disappearance of the Raman peaks [22] at 809 and 840 cm1. These peaks represent the crystalline order and the amorphous polypropylene structure, respectively.
Figs. 4 and 5 show the surface and cross section of the nanotube composites after LPH. In Fig. 4 the LPH surface changes drastically as the amount of MWCNTs is increased from 0 to 5 wt%. The surface of the pure PP is smooth and represents a surface that has experienced clean ablation of the polymer into the gaseous phase. Any melting that occurred was likely followed by rapid solidification of the polymer. As the amount of MWCNTs increase in the composite, the surface begins to show islands of high concentration of nanotubes. Eventually, the composites with higher weight loading of MWCNTs show a relatively contiguous layer of networked nanotubes across the surface, in addition to some high concentration islands of MWCNTs. Islands of MWCNTs could be agglomerates from the original composite that have been exposed as islands as the polymer around them is degraded by the laser heating, or could have formed by the localized accumulation of poorly dispersed nanotubes; the latter has been reported previously. [5,11]. The proposed formation mechanism for the latter explanation entails the result of fluid convection accompanying bubble formation and rise of bubbles to the surface through the molten sample layer and bursting of the bubbles at the surface. A representative cross-section of the 5 wt% LPH surface in Fig. 5 shows the networked nanotube structure on the surface, which we denote ‘‘region A’’ and the subsurface composite, denoted ‘‘region B’’. There is no unambiguous evidence of vigorous bubble formation in Fig. 5. Examination of the 0.5, 1.0, 2.0 and 3.0 wt% MWCNT subsurface structures also did not show evidence of vigorous bubble formation. The agglomerates seen were most likely due to poorly dispersed starting material, exposed during LPH, with bubbling having little or no effect on island formation during
171
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
their structure. The predominant features were the MWCNT, which can be seen in Fig. 6, in (a) and (b). The image in (b) clearly shows the graphitic walls that comprise the nanotubes. The tubes are not perfectly straight, and show significant waviness in the graphitic structure, which could be due to lattice bending and atomic vacancies. This is consistent with the description of Raman data presented in Section 3.4, which suggests the nanotubes have a significant density of defects in their structure. Another feature seen in the TEM is shown in (c), which is a carbonaceous particle with a concentric turbostratic nanostructure. These particles are nearly identical to previously published images of carbonaceous soot [23], both in structure and size. In this study, it is believed that these particles are formed from degraded polymer. While the soot particles and carbon black had similar structures, the soot particles were observed after LPH in composites that did not initially contain any carbon black. The TEM observations confirm the spectroscopic data showing that nanotubes with a defective structure remain intact on the LPH surface, along with small amounts of degraded polymer. 3.7. Thermal conductivity data It is known that MWCNTs can increase the thermal conductivity of a polymer [24,25]. MWCNTs have a high intrinsic thermal conductivity that is several orders of magnitude higher than a typical polymer. When the nanotubes are added to a polymer, the interfaces created, dispersion, and composite structure all play significant roles in the resulting thermal conductivity of the composite. The thermal conductivity of our nanotube composites was measured and was shown to have approximately a 25% increase at 25 °C when 5 wt% MWCNT was added. This increase is consistent with what others have reported [7]. The data are shown in Table 1. In theory, as the laser light begins to heat the surface of the composite, the higher thermal conductivity composites should be able to conduct heat away from the surface faster than the pure PP resulting in a slower rise to the degradation and ablation temperature and hence, a lower mass loss. This could play a small role in the improved performance of the higher weight % nanotube composites with respect to the pure PP. 3.8. Networked structure
Fig. 6. TEM images of the nanotube network layer on the LPH surface with (a) entangled individual nanotubes, (b) an individual nanotube with observable graphitic wall structure (c) carbonaceous soot particles (entire field of view) showing a wavy, concentric turbostratic nanostructure. The soot microstructure is similar to that of the carbon black particles.
LPH. Because the time scales used during LPH are orders of magnitude smaller than those used in flammability testing, some of the mechanisms observed in flammability tests may not be observable in LPH testing, such as island growth due to bubbling processes. 3.6. Transmission electron microscopy TEM of the networked layer on the surface of the LPH region was conducted to more closely examine the particles present and
The networked structure layer acts as a thermal shield to reduce the exposure of the underlying polymer composite and reduce degradation. As seen in the cross-section of Fig. 5, the layer is on the order of several microns thick. The layer is generally contiguous, but can have regions where exposed polymer can be seen from the plan view of the LPH region. The network of individual nanotubes can also take on a bundle-like secondary structure as seen before in flammability testing [8]. The nanotube network must be responsible for the reduced mass loss of the composites when subjected to the high temperatures caused by photothermal heating by the laser. When the polymer composite absorbs the laser light, the temperature will increase as energy is transferred to
Table 1 Thermal conductivity measurements. Material
Thermal conductivity (W/m-K) @ 25 °C
Thermal conductivity (W/m-K) @ 100 °C
Pure PP 0.5 wt% MWCNT 1.0 wt% MWCNT 2.0 wt% MWCNT 3.0 wt% MWCNT 5.0 wt% MWCNT
0.218 0.228 0.224 0.242 0.252 0.273
0.220 0.216 0.215 0.238 0.248 0.267
172
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
Fig. 7. (a) Cryo-fractured cross-section of a solvent processed PP/MWCNT composite with 2 wt% MWCNT. Individual nanotubes are seen protruding from the surface. The surfaces after LPH for the (b) 1 wt% MWCNT and (c) 2 wt% MWCNT composites (d) the fracture surface through the LPH region for 2 wt% MWCNT solvent-processed composite with the top surface of the LPH spot (A) and the composite below the surface (B) of the LPH spot.
the polymer chains, nanotubes, and carbon black. In the presence of air, the polypropylene chains undergo oxidative dehydrogenation accompanied by hydrogen abstraction [26]. Extensive chain scission and volatile fragment formation results in the mass loss and degradation of the polymer. The networked nanotube structure must either reduce the exposure of the polymer to heat or reduce the mass loss in the form of volatile formation and escape from the surface, or both of these. Qualitatively, we observe a similar networked layer providing a protective barrier to the incident radiative flux as reported previously at slow heating rates [6,7]. To further establish that this is the dominant mechanism providing the thermal stability at high heating rates, we are examining the behavior of composites exposed to multiple laser pulses, variable pulse durations, and pulse energies and this will be the subject of a future publication. It is evident that the nanotubes remain after the exposure to elevated temperatures at high heating rate during LPH. The temperatures achieved are not known at this time and is the topic of ongoing work in our laboratory. It has been reported that MWCNTs are stable in air to at least 420 °C under slow heating rates of 2 °C/ min in thermogravimetric experiments [27]. That same work also showed that an increase to 20 °C/min heating rate increased the thermal stability of MWCNTs in air by over 100 °C at the 50% weight loss mark. Others have shown the oxidative degradation of graphitized MWCNTs took place between 640 and 780 °C [28] with the thermal stability in oxidative conditions increasing in MWCNTs that have fewer defects. The defects likely lead to oxidative instability by providing edges and dangling bonds by which each nanotube can be ‘‘unzipped’’ and peeled away. While the Raman data suggests we have a MWCNT structure with defects before and after LPH, we see an intact networked nanotube structure after LPH. We can infer that either we did not achieve temperatures above 420 °C, or that the increased thermal stability of the MWCNT is a result of the very high heating rates encountered during LPH. The rapid cooling rate after LPH also limits the time spent at elevated temperatures. Some preliminary temperature analysis in our laboratory suggests that the temperature exceeds 420 °C, and we therefore tentatively identify the latter interpretation as the more plausible. 3.9. Solvent-processed composites Composites with 1 and 2 wt% MWCNT were fabricated with a solvent-based method in order to compare the mass loss behavior during LPH with those fabricated by twin screw extrusion. Melt blending is generally less effective at dispersing nanotubes in polymers compared to solvent blending methods. It was expected that solvent processed samples could have better dispersion [29]. The cryo-fractured cross-section of the 2 wt% MWCNT composite fabricated by solvent-based method is shown in Fig. 7a and it shows
very good dispersion quality. During LPH under the conditions previously described, the composites fabricated with solvent-based methods had significantly less mass loss than the composites fabricated by extrusion. The 1.0 and 2.0 wt% MWCNT solvent-processed composites had a mass loss of 293 ± 19 and 258 ± 2 lg, respectively. Compare this with the extrusion-processed composites, which had a mass loss of 467 ± 6 and 391 ± 9 lg, respectively. This represents a 36% (average) reduction in mass loss for the two composites studied. One noteworthy contributing factor could be the MWCNT aspect ratio, as mass loss rates of composites have been shown to decrease with MWCNT of increasing aspect ratio. [30]. The nanotubes used in our solvent-based samples had a larger initial aspect ratio (nominal A.R. = 833) than those used in the extrusion samples (nominal A.R. = 157). It is possible that differences in thermal conductivity of the solvent-based material could account for differences in mass change compared to the extruded samples; the thermal properties of the solvent-based material have not yet been measured. However, we consider the more likely explanation to be the improved dispersion quality of the solventprocessed samples compared to the extruded samples. The solvent-processed samples appear to have good dispersion with only small areas of low nanotube content on the cryo-fractured surface. The LPH surfaces in Fig. 7b–d appear to be a generally contiguous network of nanotubes without any areas of sparse or agglomerated concentration as seen in the extrusion processed LPH surfaces. The seemingly better dispersion quality of the MWCNTs and the more contiguous networked layer on the surface of the composite after LPH resulted in the greater thermal stability of the solvent-processed material compared to the extruded material, as measured by the decreased mass loss. Additional characterization of the solvent-processed material will appear in a future publication.
4. Conclusions Nanocomposites of PP/MWCNT were fabricated using twin screw extrusion, subjected to LPH, and subsequently characterized using mass loss data, spectroscopy, and microscopy. The Raman and FTIR spectroscopy showed that the nanotubes retain their structure during LPH and that a nanotube-rich layer formed at the surface of the composites during laser heating. The presence of this layer was confirmed by electron microscopy observations. If the nanotubes are well-dispersed and form a contiguous networked layer, it can act as a protective shield for the underlying polymer, preventing degradation. The presence of MWCNTs in the polymer resulted in a lower mass loss of the composite when subjected to highly transient heating during LPH. The PP-5 wt% MWCNT composite showed a 54% reduction in mass loss per laser pulse compared to pure PP. The nanotube composites were also shown to outperform nanoclay composites under the same
S.F. Bartolucci et al. / Composites Science and Technology 105 (2014) 166–173
conditions, despite having a significantly lower weight loading of additive. Solvent-processed samples, which appear to be well-dispersed, display improved thermal stability compared to the extruded materials, as indicated by a lower mass loss during LPH. While this processing method may show better thermal stability results due to better MWCNT dispersion, the processability of bulk nanotube composites by extrusion is higher than solventbased methods. In conclusion, the MWCNTs can reduce degradation of the composite during LPH by forming a protective surface layer that acts as a shield for the underlying polymer and by dissipating heat more quickly from increased thermal conductivity. These observations are consistent with those seen during previously published slow heating rate studies and this shows that while some phenomena are unique to transient heating and laser irradiation of polymer nanocomposites, there are some commonalities between the two regimes.
[12] [13]
[14]
[15] [16]
[17]
[18]
[19]
References [1] Thostenson ET, Ren Z, Chou TW. Advances in the science and technology of carbon nanotubes and their composites: a review. Compos Sci Technol 2001;61(13):1899–912. [2] Giannelis EP. Polymer layered silicate nanocomposites. Adv Mater 1996;8(1):29–35. [3] Paul DR, Robeson LM. Polymer technology: nanocomposites. Polymer 2008;49(15):3187–204. [4] Gilman JW, Jackson CL, Morgan AB, Harris RH, Manias E, Giannelis EP, et al. Flammability properties of polymer-layered-silicate nanocomposites. Polypropylene and polystyrene nanocomposites. Chem Mater 2000;12:1866–73. [5] Kashiwagi T, Du F, Douglas JF, Winey KI, Harris RH. Nanoparticle networks reduce the flammability of polymer nanocomposites. Nat Mater 2005;4(12):928–33. [6] Kashiwagi T, Grulke E, Hilding J, Harris R, Awad W, Douglas J. Thermal degradation and flammability properties of poly(propylene)/carbon nanotube composites. Macromol Rapid Comm 2002;23:761–5. [7] Kashiwagi T, Grulke E, Hilding J, Groth K, Harris R, Butler K, et al. Thermal and flammability properties of polypropylene/carbon nanotube nanocomposites. Polymer 2004;45:4227–39. [8] Bocchini S, Frache A, Camino G, Claes M. Polyethylene thermal oxidative stabilisation in carbon nanotubes based nanocomposites. Eur Polym J 2007;43:3222–35. [9] Schartel B, Potschke P, Knoll U, Abdel-Goad M. Fire behaviour of polyamide 6/ multiwall carbon nanotube. Eur Polym J 2005;41:1061–70. [10] Boccini S, Annibale E, Frache A, Camino G. MWNT surface self-assembling in fire retardant polyethylene-carbon nanotubes composites. E-Polymers 2008;20:1–11. [11] Kashiwagi T, Du F, Winey KI, Groth KM, Shields JR, Bellayer SP, et al. Flammability properties of polymer nanocomposites with single-walled
[20]
[21]
[22] [23]
[24] [25] [26]
[27]
[28]
[29] [30]
173
carbon nanotubes: effects of nanotube dispersion and concentration. Polymer 2005;46:471–81. Cote PJ, Kendall G, Todaro ME. Laser pulse heating of gun bore coatings. Surf Coat Tech 2001;146–147:65–9. Warrender JM, Mulligan CP, Underwood JH. Analysis of thermo-mechanical cracking in refractory coatings using variable pulse-duration laser pulse heating. Wear 2007;263:1540–4. Bartolucci SF, Zhang W, Warrender J, Lee S. Structure and morphology of carbon nanofiber polycarbosilane precursor derived nanocomposites. Nanosci Nanotechnol Lett 2010;2:1–5. Price D, Milnes GJ, Gao F. Laser pyrolysis as a model for fire behaviour behind the flame front. Polym Degrad Stabil 1996;54:235–40. Gao F, Price D, Milnes GJ, Eling B, Lindsay CI, McGrail PT. Laser pyrolysis of polymers and its relation to polymer fire behavior. J Anal Appl Pyrol 1997;40– 41:217–31. Bartolucci SF, Supan KE, Wiggins JS, LaBeaud L, Warrender JM. Thermal stability of polypropylene-clay nanocomposites subjected to laser pulse heating. Polym Degrad Stabil 2013;98:2497–502. Villmow T, Potschke P, Pegel S, Haussler L, Kretzschmar B. Influence of twinscrew extrusion conditions on the dispersion of multi-walled carbon nanotubes in a poly(lactic acid) matrix. Polymer 2008;49:3500–9. Rahatekar SS, Zammarano M, Matko S, Koziol KK, Windle AH, Nyden M, et al. Effect of carbon nanotubes and montmorillonite on the flammability of epoxy nanocomposites. Polym Degrad Stabil 2010;95:870–9. Schniepp HC, Li J-L, McAllister MJ, Sai H, Herrera-Alonso M, Adamson DH, et al. Functionalized single graphene sheets derived from splitting graphite oxide. J Phys Chem B 2006;110:8535–9. Andrews R, Jacques D, Minot M, Rantell T. Fabrication of carbon multiwall nanotube/polymer composites by shear mixing. Macromol Mater Eng 2002;287(6):395–403. Minogianni C, Gatos K, Galiotis. Estimation of crystallinity in isotropic polypropylene with raman spectroscopy. Appl Spectrosc 2005;59(9):1141–7. Posfai M, Gelencser A, Simonics R, Arato K, Li J, Hobbs PV, et al. Atmospheric tar balls: particles from biomass and biofuel burning. J Geophys Res 2004;109:D06213. Yang Y, Grulke EA, Zhang ZG, Wu G. Thermal and rheological properties of carbon nanotube-in-oil dispersions. J Appl Phys 2006;99:114307. Han Z, Fina A. Thermal conductivity of carbon nanotubes and their polymer nanocomposites: a review. Prog Polym Sci 2011;36:914–44. Zanetti M, Camino G, Reichert P, Mulhaupt R. Thermal behaviour of poly(propylene) layered silicate nanocomposites. Macrom Rapid Comm 2001;22:176–80. Mahajan A, Kingon A, Kukovecz A, Konya Z, Vilarinho P. Studies on the thermal decomposition of multiwall carbon nanotubes under different atmospheres. Mater Lett 2013;90:165–8. Bom D, Andrews R, Jacques D, Anthony J, Chen B, Meier MS, et al. Thermogravimetric analysis of the oxidation of multiwalled carbon nanotubes: evidence for the role of defect sites in carbon nanotube chemistry. Nano Lett 2002;2:615–9. Moniruzzaman M, Winey KI. Polymer nanocomposites containing carbon nanotubes. Macromolecules 2006;39:5194–205. Cipiriano BH, Kashiwagi T, Raghavan SR, Yang Y, Grulke EA, Yamamoto K, et al. Effects of aspect ratio of MWNT on the flammability properties of polymer nanocomposites. Polymer 2007;48:6086–96.