- Email: [email protected]
Investigation of exfoliated graphite nanoplatelets (xGnP) in improving thermal conductivity of paraffin wax-based phase change material
Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
Contents lists available at ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Investigation of exfoliated graphite nanoplatelets (xGnP) in improving thermal conductivity of paraffin wax-based phase change material Jinglei Xiang, Lawrence T. Drzal n Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824-1226, USA
a r t i c l e i n f o
abstract
Article history: Received 17 December 2009 Received in revised form 15 October 2010 Accepted 31 January 2011 Available online 23 February 2011
Composite phase change materials (PCM) for latent heat thermal energy storage were made by mixing two different kinds of exfoliated graphite nanoplatelets (xGnP-1 and xGnP-15) into paraffin wax. Direct casting and two roll milling were used to prepare samples. The investigation on the thermal and electrical conductivity of nanocomposites with these two nanoplatelets was performed. Higher thermal conductivity of composite PCM can be achieved with nanofillers of larger aspect ratio, better orientation and lower interface density. The thermal physical properties of the nanocomposites were investigated by differential scanning calorimetry and thermal gravimetric analysis. It was found that the latent heat of the nanocomposites was not adversely affected by the presence of xGnP nanoplatelets and the thermal stability improved. & 2011 Elsevier B.V. All rights reserved.
Keywords: Exfoliated graphite nanoplatelets Paraffin wax Latent heat storage Thermal conductivity Two roll mill
1. Introduction Latent heat storage materials have been reported in the literature as useful for thermal energy storage applications because of their high energy density [1–3]. Heat is stored through the phase transition within the material, such as solid–solid or solid–liquid transitions. The most popular form of phase change selected for thermal energy storage is the solid–liquid transition, since there is a large melting enthalpy associated with the phase change while volumetric and temperature variation during the transition is small. Some common phase change materials are the inorganic compounds which include salt hydrates, salts, metals and alloys. Organic compounds include paraffin, non-paraffin organics and poly-alcohols depending on specific applications [1]. In particular; paraffin wax has attracted numerous attentions for its low cost, moderate energy densities, low vapor pressure, negligible supercooling and chemical inertness [4]. However, one of the intrinsic disadvantages associated with paraffin wax and other organic PCM are their low thermal conductivity, which severely limited the rate of absorbing and releasing heat from and to the environment. Various high thermally conductive fillers (e.g. metallic fins [5], ceramic powder fillers, graphitic carbon fibers [6], carbon nanofibers [7], graphite particles [8] and exfoliated graphite [9]) have been reported to improve the effective thermal conductivity of phase change materials. However, one
n
Corresponding author. E-mail address: [email protected] (L.T. Drzal).
0927-0248/$ - see front matter & 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2011.01.048
usually needs a very high concentration of these traditional conductive fillers to achieve a noticeable improvement in thermal conductivity, resulting in increased density of the composites, processing difficulties and increased cost. In addition, the filler material will replace the active matrix, reducing its thermal energy storage capacity. Recently, single and multi-walled carbon nanotubes, carbon blacks, exfoliated graphite nano platelets with intrinsically high conductivities have been used in making nanocomposites which become electrically and thermally conductive at much lower loading levels compared with traditional fillers [10]. Carbon nanotubes nanocomposites have been reported to have an electrical percolation threshold on the order of 0.01% by volume due to their high aspect ratios [11]. The thermal conductivity depends largely on aspect ratio, dispersion, and interfacial area with the polymer matrix [12]. Homogenous dispersions of carbon nanotubes in polymers have long been a major challenge and the high cost of producing nanotubes offset some of their advantages. Carbon black particles, with its extremely high surface area and easily aggregated state, pose great difficulty in achieving uniform dispersion. Nonetheless, carbon blacks are usually mixed with other conductive particles to make a hybrid structure in which carbon blacks serve as bridging channels, therefore enhancing heat transfer among the major heat carrying particles such as carbon nanotubes and carbon fibers [13]. Exfoliated graphite nano platelets, usually produced from graphite intercalated compounds, are particles consisting of several layers of graphene sheets. They have very high aspect ratio comparable to that of carbon nanotubes. Drzal et al. successfully developed a microwave exfoliation and
1812
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
ultrasonic grinding process to prepare exfoliated graphite nanoplatelets (xGnP) of different sizes and surface areas [14]. These particles have been incorporated into different thermoplastic and thermoset materials to improve the electrical, thermal and mechanical properties of nanocomposites [15,16]. In this work, xGnP of two sizes and aspect ratios were mixed with paraffin wax. The viscosity of paraffin wax is low and is thus conducive to dispersing conductive particles uniformly in the polymer matrix. The thermal and electrical conductivities of nanocomposites with nanofillers of different sizes were investigated. Both isotropic and anisotropic samples were prepared to study the effect of particle orientation on the conductivities of nanocomposites. Furthermore, the experimental data were compared with effective medium theory first proposed by Nan et al. [17] on carbon nanotubes and Foygel et al’s percolation theory [18] to study the conductivities of nanocomposites below and above percolation threshold. The models imply the mechanisms governing the marginal improvement in thermal conductivity as opposed to dramatic improvement in electrical conductivity observed around percolation threshold in nanocomposites. In addition, the effect of adding different xGnP particles on melting/crystallization enthalpies, latent heat and thermal stability of paraffin wax will also be discussed.
2. Experimental 2.1. Materials Exfoliated graphite nanoplatelets (xGnP) are prepared from exfoliating the sulfuric acid intercalated natural graphite (A3772) purchased from Asbury Graphite Mills, Inc, NJ. The microwave thermal exfoliation process [14] provides a rapid thermal treatment to the graphite intercalated compounds (GIC), which quickly vaporizes the acids trapped in between the layers of the graphite resulting in a rapid expansion of graphite gallery. The microwave treatment proves superior to other traditional thermal treatments in that the heating process is quick and the energy input density is high. The volume of expanded graphite produced with microwave heating is over 500 times that of the initial volume occupied by GIC. Pulverization methods such as sonication and ball milling are used to reduce the expanded graphite to a desired size. Graphite nanoplatelets with a lateral size of about 15 mm and a thickness of around 10 nm are produced by sonication for 2h (xGnP-15). Nanoplatelets of 1 mm in diameter (xGnP-1) are produced by further size reduction of the xGnP-15 particles with ball milling for 72 h. The surface area of the exfoliated graphite nanoplatelets were measured by Brunauer–Emmet– Teller (BET) using N2 adsorption at 77 K. For xGnP-15, the surface area is around 20–40 m2/g. For xGnP-1, the surface area is around 100–130 m2/g. Paraffin wax (n-docosane) with melting temperature of 53–57 1C, xylenes (reagent grade) were purchased from Sigma-Aldrich and used directly. 2.2. Sample preparation 2.2.1. Casting Paraffin wax was melted at 80 1C. xGnP were then added to the liquid paraffin under constant stirring at 1, 2, 4, 6, 8, and 10 wt%. At higher xGnP loadings (i.e. 46 wt%), xylenes were used as a solvent to dissolve wax first, reducing the melt viscosity and facilitate the uniform dispersion of xGnP. The liquid composite was dip sonicated for 1 h prior to solvent evaporation. Samples were poured into a customized aluminum mold (shown in Fig. 1a) and solidified at room temperature. Disk samples of 1 in in
Fig. 1. (a) Photograph of the customized aluminum mold; (b) photograph of Bolling two roll mill and (c) Scheme for the sample preparation routes: casting and two roll milling.
diameter and 5 mm in average thickness were used for subsequent experiments.
2.2.2. Two roll milling In order to study the effect of particle orientation on thermal conductivity, composite samples from cast molding was further processed by two roll mill (shown in Fig. 1b) in which extensive
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
shear force and compression at the nip point of the rolls allow particles to align within the plane. In two roll mill processing, the quality of the composite film depends on the temperature of the rolls, a higher temperature yet still below the melting point of paraffin reduces the chance of film shrinkage by giving the wax molecules enough thermal energy to undergo stress relaxation while the film is still wrapped around the rolls. The speed of the roll is adjusted so that the front roll rotates faster than the rear roll. It is worth noting that the speed difference between the rolls effectively creates enough shear force at the gap of the two rolls, facilitating the dispersion of xGnP particles and their alignment in the matrix. The composite materials were processed in the rolls for 10 min before they were removed. The final product composite sheet is around 0.4–0.5 mm in thickness. Both the through plane and in plane thermal conductivity was measured. Fig. 1c shows the scheme of sample preparation routes.
3. Characterization 3.1. Scanning electron microscopy (SEM) To investigate the dispersion of xGnP particles in paraffin matrix at different loadings, disk samples prepared by casting and two roll milling were liquid nitrogen fractured. The fractured surface was oxygen plasma treated for 5 min with a Radio Frequency power at 550 W. The treated sample was later gold coated with the Denton Vacuum sputter coater to ensure that the sample is electrically
1813
conductive. A JEOL (JSM 6400) scanning electron microscope with an accelerating voltage of 15 kV was used to examine the fractured composite surface at a working distance of 15 mm. 3.2. Thermal conductivity The thermal conductivity of paraffin/xGnP composites was measured with a Unitherm (TM) Model 2022 (Anter Corporation, Pittsburgh, PA). The test was performed according to ASTM E1530. For samples prepared by cast molding, the average thickness of the sample is 5 mm, and it can be directly used for through-plane thermal conductivity measurement. However, for the samples prepared by two roll mill, the through plane thermal resistance is too small to be measured due to the very small thickness of the composite thin film (0.4–0.5 mm). 4 or 5 sheets were stacked in parallel so that the resistance falls in the range of the instrument. As for the in-plane measurement, the sample was made by compressing the thin sheets released from the roll mills in a closed rectangular mold to a desired thickness, and used a circular iron punch to produce a 1 in disk sample from the cross section of the block (shown in Fig. 2a and b). All the samples were tested at 20 1C under an applied load of 20 psi. 3.3. Electrical conductivity measurement The resistivity of paraffin/xGnP nanocomposites was measured with a Gamry instrument under FAS2TM Femtostat plug system
Fig. 2. (a) Schematics for in plane thermal conductivity measurement of the two roll milled sample and (b) photograph of the sample prepared from compressing the two roll milled sheets to a desired thickness (1 in) and a circular cross section of that was produced by a punch;
1814
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
and potentiostatic mode. A rectangular portion was fabricated by TM a Buehler Isomet low speed saw from the cast molded disk samples for the measurements. The resistance of the sample was measured in the frequency range from 0.6 Hz to 100 kHz and resistivity of samples were calculated by taking the geometry of the sample into account using the following equation:
atmosphere. The temperature range scanned was from 25 to 500 1C with a 25 1C/min ramp rate and a 4 1C/min resolution upon thermal decomposition events.
r ¼ RA=L
4.1. Morphology
ð1Þ
where R is the resistance of the sample at 1 Hz, A is the cross section area of the sample and L is the length of the sample. 3.4. Differential scanning calorimetry (DSC) A sample size of around 5–10 mg was loaded to the DSC sample cell and the data was collected for the 2nd run at a scanning rate of 5 1C/min. The solid–solid transformation in paraffin wax, the melting temperature, the crystallization temperature, and the latent heat of paraffin were analyzed by Universal Analysis 2000 Version 4.5A. 3.5. Thermogravimetric analysis (TGA) A sample of 10–15 mg of composite samples was loaded to the Pt pan and the experiments were conducted in nitrogen
4. Results and discussions
Fig. 3(a) shows the fractured cross section of the 4 wt% xGnP-1/ paraffin nanocomposites prepared by casting. The xGnP-1 particles are either embedded in the paraffin matrix evidenced by the white lines protruding from the background or lying on the surface suggested by their irregular but sharp edges different from the smooth and soft paraffin wax. At the same nanofiller loading, the xGnP-15 particles in the nanocomposites are more easily recognized due to their larger size shown in Fig. 3(b). It is also found that unlike the morphology reported by Kalaitzidou et al. [16] in polypropylene/xGnP nanocomposites that most xGnP particles tend to roll up or fold on itself during sample preparation (extrusion and injection molding), the xGnP particles in paraffin nanocomposites maintained their platelet-like shape as a result of the low viscosity of the polymer matrix and thus the high aspect ratio of xGnP is not negatively affected in sample fabrication. In the samples prepared by two roll mill with xGnP-15 loading at 4,
Fig. 3. (a) 4 wt% xGnP-1/paraffin by casting; (b) 4 wt% xGnP-15/paraffin by casting; (c) 4 wt% xGnP-15/paraffin by two roll mill; (d) 10 wt% xGnP-15/paraffin by two roll mill; (e) enlarged edge view of the nanoplatelets (scale bar 100 nm) and (f) enlarged edge view of one platelet (scale bar: 10 nm).
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
In order to find out the effect of dispersion of xGnP particles of different sizes on the electrical conductivity of the resulting nanocomposites, the resistivity of paraffin/xGnP-1 and paraffin/ xGnP-15 samples were measured and compared in Fig. 4. For nanocomposites with xGnP-15 as fillers, the resistivity undergoes a sharp decrease from 1010 (O cm) to around 103 (O cm) at 1 vol% which suggests that conductive paths were formed in paraffin. However, in xGnP-1 composites, the system does not percolate even at 2 vol%. Despite the fact that the absolute number of nanofillers in xGnP-1 composites is almost 250 (15 15) times more than for xGnP-15 composites at the same loading level, the aspect ratio of xGnP-1 is much smaller than xGnP-15 (100 vs. 1000). It can also be seen from SEM micrographs that the plateletlike morphology for xGnP-15 particles are maintained better than xGnP-1, which shed light on the importance of filler morphology and aspect ratio on the electrical conductivity of nanocomposites. 4.3. Thermal conductivity paraffin/xGnP nanocomposites The thermal conductivity of paraffin/xGnP nanocomposites prepared by casting at different filler concentrations was measured and shown in Fig. 5(a). The thermal conductivity increases linearly for xGnP-15 samples up to 2 vol% beyond which the data exhibit super linear behavior. In xGnP-1 composites, thermal conductivity also increases linearly with nano platelet loading, but the improvement is not as effective as xGnP-15 although the number density of nanofiller particles is much higher at the same loading level. Compared with larger filler particles, smaller particles introduce more phonon scattering interfaces at the boundaries. Heat transport in graphite nanoplatelets occurs by phonons of varying frequencies. In the xGnP polymer nanocomposites, phonons are
12 Resistivity (ohm cm) Log scale
xGnP-15 composite PCM
10 xGnP-1 composite PCM
8 6 4 2 0 0
1
2 3 xGnP loadings (vol%)
4
Fig. 4. Electrical resistivity of xGnP/paraffin composite PCM.
5
xGnP-15 nanocomposites
Thermal conductivity (W/mK)
4.2. Electrical conductivity of paraffin/xGnP nanocomposites
3.5 3
xGnP-1 nanocomposites Curve fit beyond percolation
2.5 2 1.5 1 0.5 0 0
0.02
0.04
0.06
xGnP volume fraction 3.5 two roll mill through plane
3 Thermal conductivity (W/mK)
and 10 wt% shown in Fig. 3(c and d, respectively). It is very clear that particles assume a preferential alignment as indicated by the white lines in the micrograph. Fig. 3(e and f) shows the enlarged edge views of the nanoplatelets, it is also evident that the individual nanoplatelet is thin with thickness less than 10 nm and thus they have a very high aspect ratio. The distance between particles was increased due to the expansion of material at the rolling nip point and it is the same compressive and shearing force that contributed to the alignment of the particles.
1815
two roll mill in plane cast through plane 2.414
2.5 2
1.71
1.5 1
0.737
0.701
0.496
0.5
0.31
0 0.02
0.05 xGnP-15 volume fraction
Fig. 5. (a) Through plane thermal conductivity of xGnP/paraffin prepared by casting and curve fitting with Foygel’s model above percolation and (b) thermal conductivity of xGnP-15/paraffin prepared by two roll mill.
transported from one xGnP platelet to another via the polymer in between. Since there is a difference in acoustic properties between the polymer and graphite nanoplatelets and in this case, the paraffin wax and xGnP interacts with each other through weak Van der Waals forces, only low frequency phonon vibration modes are available to carry a small amount of heat energy. The high frequency phonons, which are the major heat energy carriers, interact with other phonons before they could be transferred to some low energy vibration states to couple with the polymer matrix. As a result, a very high thermal interface resistance exists between the boundaries and gives rise to a reduction in heat transfer in nanocomposites. This process would be more prevalent in nanocomposites containing finer filler particles due to a higher density of interfaces. Given the fact that the xGnP-15 is superior to xGnP-1 particles in improving the thermal conductivity of paraffin nanocomposites, two samples prepared by two roll mill containing 2 and 5 vol% xGnP-15 in paraffin matrix were used for the measurements. Both samples show higher thermal conductivity in the in-plane direction rather than in the through-plane direction shown in Fig. 5(b). Aligning the particles in its basal plane direction effectively utilizes the high in-plane thermal conductivity of the graphite
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
nanoplatelets. However, the thermal conductivity of the two roll milled sample is not as high as the cast mold samples at the same loading level. On the one hand, random orientations of xGnP in paraffin for the cast molded samples leads to overlapping among particles which facilitates phonon transport due to the larger thermal contact area. On the other hand, higher xGnP loadings also lead to possible aggregation in the wax which contributes to a large scattering in the data. The two roll mill process, by comparison, not only aligns the particles in the matrix but also breaks up the aggregates and separates them, destroying the percolating networks and introduces more phonon scattering at the interfaces. 4.4. Effect of thermal interface resistance
bx ¼
g 2ðK11 Km Þ ; g K11 þ Km
bz ¼
g K33 1 Km
ð2Þ
where f is the volume fraction of graphite platelets, Assume the graphite platelet is coated with a very thin interfacial thermal barrier g g layer, K11 and K33 are the equivalent thermal conductivities along transverse and longitudinal axes of a graphite platelet, respectively, and can be expressed as: g K11 ¼
Kg ; 1 þð2ak Kg =tKm Þ
g K33 ¼
Kg 1 þð2ak Kg =dKm Þ
ð3Þ
where t and d are the thickness and diameter of xGnP; and ak is the Kapitza radius defined by: ak ¼ Rk Km
Rk = 7E-8
Rk = 1E-8
Rk = 3E-8
Rk = 5E-8
Rk = 9E-8
10
1 0.1
As already mentioned in the previous section, the thermal interface resistance presents a significant barrier for phonons to transfer from one conductive filler particle to another surrounded by a polymer matrix, it is interesting to predict the thermal interface resistance from experimental data. Many models have been proposed to predict the thermal conductivity of nanocomposites, but few have taken the interfacial resistance into account. Nan et al. [17] proposed a model based on carbon nanotubes and assumed an interfacial barrier layers between the nanotubes and the polymer, a Maxwell–Garnett type effective medium approach was used to predict the thermal interface resistance. For the case of isotropically oriented ellipsoidal particles including platelets, the thermal conductivity ratio Ke/Km (defined as effective thermal conductivity vs. matrix thermal conductivity) can be expressed by : Ke 3 þ f ðbx þ bz Þ ¼ ; 3f bx Km
Experimental
Thermal conductivity ratio (ke/km) Log scale
1816
ð4Þ
where Rk is the thermal interface resistance. The model described above assumes that the nanofiller particles are completely surrounded by the polymer matrix and they do not touch each other. Since the percolation threshold for xGnP-15/paraffin composites is around 1 vol% as suggested by electrical resistivity measurement, the experimental data used for curve fitting is restricted to those below 1 vol%. Assume the diameter and thickness of the xGnP-15 particles are 15 mm and 10 nm and the in-plane and through-plane thermal conductivities are 3000 and 10 W/m K, respectively. The thermal interface resistance at the boundary of graphite and paraffin is estimated to be 7 9 108 m2 K=W, this Rk value is very close to the interfacial resistance across the carbon nanotube matrix reported by Huxtable et al. [19] which is about 8:3 108 m2 K=W. Fig. 6 shows the curve fitting result using this model. As the xGnP concentration increases to beyond the percolation threshold (1 vol%), a non-linear model presented by Foygel et al. [18] is used to model the nanocomposite system. This approach is based on using Monte Carlo’s simulation to model percolating clusters consisting of nanotubes or any randomly oriented conductive fillers of large aspect ratio. The key parameters such as the conductivity exponent of the percolation network and the contact resistance can be derived from the model. The thermal conductivity of the paraffin/xGnP-15 in the
1 volume fraction (%) Log scale
Fig. 6. Thermal conductivity of paraffin/xGnP-15 below percolation with Nan’s model (unit of Rk: m2K/W).
vicinity of the percolation threshold is described by the following equation:
sðf; aÞ ¼ s0 ½ffc ðaÞtðaÞ
ð5Þ
where s0 is the pre-exponential factor that depends on the conductivity of individual particle or the contact between them, ^ is their volume fraction, ^c is the critical volume fraction at the percolation threshold, t(a) is the conductivity exponent that depends on the aspect ratio a. From the result of electrical conductivity measurements, the percolation threshold for the paraffin/xGnP-15 nanocomposites is around 0.8% by volume. The least square fit to the equation with s0 and t(a) as the fitting parameters gives s0 ¼120 W/mK and t(a)¼1.26 for paraffin/ xGnP-15 composites. The resistance of individual particle or their contact resistance (whichever is larger) R0 could be estimated by equation: R0 ¼ ðs0 Lfc
tðaÞ 1
Þ
ð6Þ
where L is the lateral dimension of xGnP particles. The obtained value R0 ¼2.4 105 K/W is in good agreement with the thermal resistance of an individual xGnP-15 particle: RxGnP ¼2 105 K/W, estimated using an experimental room temperature value of sxGnP ¼3000 W/mK as the thermal conductivity, length L¼15 mm and thickness t ¼10 nm. Thus it is suggested that for a nanocomposite system above percolation threshold, the effective thermal conductivity is dominated by the conductive particles. 4.5. The effect of xGnP particles on thermal properties of nanocomposites Differential scanning calorimetry was used to investigate the effect of xGnP particles on thermal physical properties of paraffin nanocomposites and its capacity to store thermal energy. Both the melting and cooling behaviors of paraffin/xGnP were recorded, two peaks were observed from the DSC analyses which correspond to solid–solid transformation and melting/crystallization of paraffin wax shown in Fig. 7(a–d). The relevant enthalpies were calculated by integration the peaks above the base line given by the software itself. The temperatures that characterize the melting and crystallization peak are not affected significantly by addition of xGnP particles. The melting and crystallization enthalpies for paraffin nanocomposites at low xGnP concentration do not deviate from neat paraffin very much as opposed to the samples at high loading levels when the composite’s ability to absorb and release heat is degraded because of the increasing replacement of paraffin wax with xGnP particles. However, the presence of more xGnP particles in the paraffin matrix does not negatively affect the thermal
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
Neat Paraffin
2wt% xGNP15/Paraffin 8wt% xGNP15/Paraffin
0.5
2.5
0
2
Heat flow (W/g)
Heat flow (W/g)
4wt% xGNP15/Paraffin
-0.5 -1 -1.5 -2 -20
0
20
40
60
80
Neat Paraffin
2wt% xGNP15/Paraffin
4wt% xGNP15/Paraffin
8wt% xGNP15/Paraffin
1.5 1 0.5 0 -20
-2.5 100
0
20
40
60
80
100
Temperature (C)
Neat Paraffin
2wt% xGNP1/Paraffin
Neat paraffin
2wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
0.5
2.5
0
2
Heat flow (W/g)
Heat flow (W/g)
Temperature (C)
-0.5 -1 -1.5 -2 -2.5 -20
1817
0
20
40 60 Temperature (C)
80
100
1.5 1 0.5 0 -20
0
20
40 60 Temperature (C)
80
100
Fig. 7. (a) Heating curves of xGnP-15/paraffin; (b) cooling curves of xGnP-15/paraffin; (c) heating curves of xGnP-1/paraffin and (d) cooling curves of xGnP-1/paraffin. Table 1 Heating and cooling characteristics of xGnP-15 composite PCM (average of three samples).
Table 2 Heating and cooling characteristics of xGnP-1 composite PCM (average of three samples).
Sample names
Melt peak (1C)
Melting enthalpy of composite (J/g)
Melting enthalpy of wax (J/g)
Cryst. peak (1C)
Cryst. enthalpy of composite (J/g)
Cryst. enthalpy of wax (J/g)
Sample names
Melt peak (1C)
Melting enthalpy of composite (J/g)
Melting enthalpy of wax (J/g)
Cryst. peak (1C)
Cryst. enthalpy of composite (J/g)
Cryst. enthalpy of wax (J/g)
Neat 2 wt% 4 wt% 8 wt%
54.54 54.82 55.13 54.9
185.9 188.6 183.5 171.9
185.9 192.4 191.1 186.8
51.73 51.27 51.87 51.63
186.8 189.9 185 172.9
186.8 193.7 192.7 187.9
Neat 2 wt% 4 wt% 8 wt%
54.54 54.33 53.97 54.35
185.9 186.05 180.4 176.65
185.9 189.85 187.9 192
51.73 52.09 52.22 51.94
186.8 187.15 180.2 178.1
186.8 191 187.7 193.5
physical properties of wax itself. This can be further supported by compensating the presence of xGnP in the composites and recalculating the melting and crystallization enthalpies for paraffin wax alone in the nanocomposites using Equation (7). It is found that for both xGnP-1/xGnP-15 composites, the heat absorbing and releasing capacity during heating and cooling for paraffin did not degrade but improve slightly with xGnP loadings.
DHwax ¼
DHcomp 1xG
ð7Þ
where DHcomp is latent heat of composite PCM with xGnP. DHwax is latent heat of wax in PCM. xG is the weight fraction of xGnP. Tables 1 and 2 collected data from paraffin nanocomposites containing 0, 2, 4 and 8 wt% xGnP-15 or xGnP-1. 4.6. Thermal gravimetric analysis Since graphite is known for its chemical inertness and resistance to thermal degradation, thermal decomposition of paraffin
nanocomposites upon addition of xGnP particles was analyzed by TGA in nitrogen atmosphere shown in Fig. 8. It is shown in the analysis that thermal decomposition of paraffin nanocomposites gradually shifted to a higher temperature with increasing amount of xGnP particles. Although the interaction between graphite platelets and paraffin is weak, xGnP particles still interact with the surrounding matrix and improve the overall stability of composite PCM. The remaining weight at 500 1C, at which temperature paraffin wax was completely thermally decomposed but the graphite is unaffected, was close to the initial loadings 0, 2, 4 and 8 wt%.
5. Conclusions Exfoliated graphite nanoplatelets of two sizes (xGnP-1 and xGnP-15) were successfully mixed with paraffin wax to make a composite PCM of high latent heat and high thermal and electrical conductivity. Nanocomposites prepared by casting with xGnP-15
Weight retained
1818
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Neat Paraffin
2wt% xGNP15/Paraffin
4wt% xGNP15/Paraffin
8wt% xGNP15/Paraffin
Acknowledgements The author appreciated the generous help from Brian Rook from Composite Materials and Structures Center on operating the two roll mill. Financial support for this project was provided by the Michigan Economic Development Corporation through the 21st Century Jobs Fund. References
0
100
200
300
400
500
600
Weight retained
Temperature (C)
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
100
Neat Paraffin
2wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
200
300
400
500
600
Temperature (C) Fig. 8. TGA analysis of xGnP-1/paraffin and xGnP-15/paraffin nanocomposites in N2 atmosphere.
show lower electrical percolation and higher thermal conductivity than xGnP-1 counterparts due to a higher interface density. Two roll milling enables the platelet particles to orient uniformly in the matrix enhancing the in-plane thermal conductivity of the nanocomposites. However, the milling process also separates the particles disrupting percolated network and increasing polymer graphite interfaces. The interfacial thermal resistance estimated by Nan’s model to be 7–9 10 8 m2K/W, was found to be comparable to that of carbon nanotubes. Experimental data also fit Foygel’s model suggesting that thermal conductivity above the percolation point for xGnP-15 PCM was more affected by the intrinsic conductivity of fillers instead of the interface. DSC results show the latent heat of paraffin nanocomposites were not degraded by adding xGnP. On the contrary, latent heat of paraffin itself and the thermal stability of the composite PCM improved slightly in the presence of xGnP.
[1] S.M. Hasnain, Review on sustainable thermal energy storage technologies, part I: heat storage materials and techniques, Energy Convers. Manage 39 (1998) 1127–1138. [2] M. Telkes, Thermal storage for solar heating and cooling, in: Proceedings of the Workshop on Solar Energy Storage Subsystems for the Heating and Cooling of Buildings, Charlottesville, Virginia, USA, 1975. [3] K.K. Pillai, B.J. Brinkworth, The storage of low grade thermal energy using phase change materials, Appl. Energy 2 (1976) 205–216. [4] A.S. Luyt, Role of polymers in developing phase change materials for energy storage, eXpress Polymer Lett 2 (2008) 147. [5] R. Velraj, R.V. Seeniraj, B. Hafner, C. Faber, K. Schwarzer, Experimental analysis and numerical modeling of inward solidification on a finned vertical tube for a latent heat storage unit, Sol. Energy 60 (1997) 281–290. [6] J. Fukai, M. Kanou, Y. Kodama, O. Miyatake, Thermal conductivity enhancement of energy storage media using carbon fibers, Energy Convers. Manage 41 (2000) 1543–1556. [7] A. Elgafy, K. Lafdi, Effect of carbon nanofiber additives on thermal behavior of phase change materials, Carbon 43 (2005) 3067–3074. [8] S. Pincemin, R. Olives, X. Py, M. Christ, Highly conductive composites made of phase change materials and graphite for thermal storage, Sol. Energy. Mater. Sol. C. 92 (2008) 603–613. [9] S. Kim, L.T. Drzal, High latent heat storage and high thermal conductive phase change materials using exfoliated graphite nanoplatelets, Sol. Energy. Mater. Sol. C. 93 (2009) 136–142. [10] G.-W. Lee, J.T. Lee, S.-S. Lee, M. Park, J. Kim, Comparison of thermal properties between inorganic fillers and acid-treated multiwall nanotube/polymer composites, J Mater. Sci. 40 (2005) 1259–1263. [11] C. Park, Z. Ounaies, K.A. Watson, R.E. Crooks, J. Smith, S.E. Lowther, J.W. Connell, E.I. Siochi, J.S. Harrison, T.L.S. Clair, Dispersion of single wall carbon nanotube by in-situ polymerization under sonication, Chem. Phys. Lett. 364 (2002) 303–308. [12] F.H. Gojny, M.H.G. Wichmann, B. Fiedler, I.A. Kinloch, W. Bauhofer, A.H. Windle, K. Schulte, Evaluation and identification of electrical and thermal conduction mechanisms in carbon nanotube/epoxy composites, Polymer 47 (2006) 2036–2045. [13] J. Heiser, J.A. King, Thermally conductive carbon filled nylon 6,6, Polym. Compos. 25 (2004) 186–193. [14] H. Fukushima, Graphite nanoreinforcements in polymer nanocomposites. Ph.D Thesis, Michigan State University, East Lansing, MI, USA, 2003. [15] K. Kalaitzidou, Exfoliated graphite nanoplatelets as reinforcement for multifunctional polypropylene nanocomposites. Ph.D Thesis, Michigan State University, East Lansing, MI, USA, 2006. [16] K. Kalaitzidou, H. Fukushima, L.T. Drzal, Multifunctional polypropylene composites produced by incorporation of exfoliated graphite nanoplatelets, Carbon 45 (2007) 1446–1452. [17] C.W. Nan, G. Liu, Y. Lin, M. Li, Interface effect on thermal conductivity of carbon nanotube composites, Appl. Phys. Lett. 85 (2004) 3549–3551. [18] M. Foygel, R.D. Morris, D. Anez, S. French, V.L. Sobolev, Theoretical and computational studies of carbon nanotube composites and suspensions, Phys. Rev. B 71 (2005) 104201–104208. [19] S. Huxtable, D.G. Cahill, S. Shenogin, L. Xue, R. Ozisik, P. Barone, M. Usrey, M.S. Strano, G. Siddons, M. Shim, P. Keblinski, Interfacial heat flow in carbon nanotube suspensions, Nat. Mater. 2 (2003) 731–734.
Contents lists available at ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Investigation of exfoliated graphite nanoplatelets (xGnP) in improving thermal conductivity of paraffin wax-based phase change material Jinglei Xiang, Lawrence T. Drzal n Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824-1226, USA
a r t i c l e i n f o
abstract
Article history: Received 17 December 2009 Received in revised form 15 October 2010 Accepted 31 January 2011 Available online 23 February 2011
Composite phase change materials (PCM) for latent heat thermal energy storage were made by mixing two different kinds of exfoliated graphite nanoplatelets (xGnP-1 and xGnP-15) into paraffin wax. Direct casting and two roll milling were used to prepare samples. The investigation on the thermal and electrical conductivity of nanocomposites with these two nanoplatelets was performed. Higher thermal conductivity of composite PCM can be achieved with nanofillers of larger aspect ratio, better orientation and lower interface density. The thermal physical properties of the nanocomposites were investigated by differential scanning calorimetry and thermal gravimetric analysis. It was found that the latent heat of the nanocomposites was not adversely affected by the presence of xGnP nanoplatelets and the thermal stability improved. & 2011 Elsevier B.V. All rights reserved.
Keywords: Exfoliated graphite nanoplatelets Paraffin wax Latent heat storage Thermal conductivity Two roll mill
1. Introduction Latent heat storage materials have been reported in the literature as useful for thermal energy storage applications because of their high energy density [1–3]. Heat is stored through the phase transition within the material, such as solid–solid or solid–liquid transitions. The most popular form of phase change selected for thermal energy storage is the solid–liquid transition, since there is a large melting enthalpy associated with the phase change while volumetric and temperature variation during the transition is small. Some common phase change materials are the inorganic compounds which include salt hydrates, salts, metals and alloys. Organic compounds include paraffin, non-paraffin organics and poly-alcohols depending on specific applications [1]. In particular; paraffin wax has attracted numerous attentions for its low cost, moderate energy densities, low vapor pressure, negligible supercooling and chemical inertness [4]. However, one of the intrinsic disadvantages associated with paraffin wax and other organic PCM are their low thermal conductivity, which severely limited the rate of absorbing and releasing heat from and to the environment. Various high thermally conductive fillers (e.g. metallic fins [5], ceramic powder fillers, graphitic carbon fibers [6], carbon nanofibers [7], graphite particles [8] and exfoliated graphite [9]) have been reported to improve the effective thermal conductivity of phase change materials. However, one
n
Corresponding author. E-mail address: [email protected] (L.T. Drzal).
0927-0248/$ - see front matter & 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2011.01.048
usually needs a very high concentration of these traditional conductive fillers to achieve a noticeable improvement in thermal conductivity, resulting in increased density of the composites, processing difficulties and increased cost. In addition, the filler material will replace the active matrix, reducing its thermal energy storage capacity. Recently, single and multi-walled carbon nanotubes, carbon blacks, exfoliated graphite nano platelets with intrinsically high conductivities have been used in making nanocomposites which become electrically and thermally conductive at much lower loading levels compared with traditional fillers [10]. Carbon nanotubes nanocomposites have been reported to have an electrical percolation threshold on the order of 0.01% by volume due to their high aspect ratios [11]. The thermal conductivity depends largely on aspect ratio, dispersion, and interfacial area with the polymer matrix [12]. Homogenous dispersions of carbon nanotubes in polymers have long been a major challenge and the high cost of producing nanotubes offset some of their advantages. Carbon black particles, with its extremely high surface area and easily aggregated state, pose great difficulty in achieving uniform dispersion. Nonetheless, carbon blacks are usually mixed with other conductive particles to make a hybrid structure in which carbon blacks serve as bridging channels, therefore enhancing heat transfer among the major heat carrying particles such as carbon nanotubes and carbon fibers [13]. Exfoliated graphite nano platelets, usually produced from graphite intercalated compounds, are particles consisting of several layers of graphene sheets. They have very high aspect ratio comparable to that of carbon nanotubes. Drzal et al. successfully developed a microwave exfoliation and
1812
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
ultrasonic grinding process to prepare exfoliated graphite nanoplatelets (xGnP) of different sizes and surface areas [14]. These particles have been incorporated into different thermoplastic and thermoset materials to improve the electrical, thermal and mechanical properties of nanocomposites [15,16]. In this work, xGnP of two sizes and aspect ratios were mixed with paraffin wax. The viscosity of paraffin wax is low and is thus conducive to dispersing conductive particles uniformly in the polymer matrix. The thermal and electrical conductivities of nanocomposites with nanofillers of different sizes were investigated. Both isotropic and anisotropic samples were prepared to study the effect of particle orientation on the conductivities of nanocomposites. Furthermore, the experimental data were compared with effective medium theory first proposed by Nan et al. [17] on carbon nanotubes and Foygel et al’s percolation theory [18] to study the conductivities of nanocomposites below and above percolation threshold. The models imply the mechanisms governing the marginal improvement in thermal conductivity as opposed to dramatic improvement in electrical conductivity observed around percolation threshold in nanocomposites. In addition, the effect of adding different xGnP particles on melting/crystallization enthalpies, latent heat and thermal stability of paraffin wax will also be discussed.
2. Experimental 2.1. Materials Exfoliated graphite nanoplatelets (xGnP) are prepared from exfoliating the sulfuric acid intercalated natural graphite (A3772) purchased from Asbury Graphite Mills, Inc, NJ. The microwave thermal exfoliation process [14] provides a rapid thermal treatment to the graphite intercalated compounds (GIC), which quickly vaporizes the acids trapped in between the layers of the graphite resulting in a rapid expansion of graphite gallery. The microwave treatment proves superior to other traditional thermal treatments in that the heating process is quick and the energy input density is high. The volume of expanded graphite produced with microwave heating is over 500 times that of the initial volume occupied by GIC. Pulverization methods such as sonication and ball milling are used to reduce the expanded graphite to a desired size. Graphite nanoplatelets with a lateral size of about 15 mm and a thickness of around 10 nm are produced by sonication for 2h (xGnP-15). Nanoplatelets of 1 mm in diameter (xGnP-1) are produced by further size reduction of the xGnP-15 particles with ball milling for 72 h. The surface area of the exfoliated graphite nanoplatelets were measured by Brunauer–Emmet– Teller (BET) using N2 adsorption at 77 K. For xGnP-15, the surface area is around 20–40 m2/g. For xGnP-1, the surface area is around 100–130 m2/g. Paraffin wax (n-docosane) with melting temperature of 53–57 1C, xylenes (reagent grade) were purchased from Sigma-Aldrich and used directly. 2.2. Sample preparation 2.2.1. Casting Paraffin wax was melted at 80 1C. xGnP were then added to the liquid paraffin under constant stirring at 1, 2, 4, 6, 8, and 10 wt%. At higher xGnP loadings (i.e. 46 wt%), xylenes were used as a solvent to dissolve wax first, reducing the melt viscosity and facilitate the uniform dispersion of xGnP. The liquid composite was dip sonicated for 1 h prior to solvent evaporation. Samples were poured into a customized aluminum mold (shown in Fig. 1a) and solidified at room temperature. Disk samples of 1 in in
Fig. 1. (a) Photograph of the customized aluminum mold; (b) photograph of Bolling two roll mill and (c) Scheme for the sample preparation routes: casting and two roll milling.
diameter and 5 mm in average thickness were used for subsequent experiments.
2.2.2. Two roll milling In order to study the effect of particle orientation on thermal conductivity, composite samples from cast molding was further processed by two roll mill (shown in Fig. 1b) in which extensive
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
shear force and compression at the nip point of the rolls allow particles to align within the plane. In two roll mill processing, the quality of the composite film depends on the temperature of the rolls, a higher temperature yet still below the melting point of paraffin reduces the chance of film shrinkage by giving the wax molecules enough thermal energy to undergo stress relaxation while the film is still wrapped around the rolls. The speed of the roll is adjusted so that the front roll rotates faster than the rear roll. It is worth noting that the speed difference between the rolls effectively creates enough shear force at the gap of the two rolls, facilitating the dispersion of xGnP particles and their alignment in the matrix. The composite materials were processed in the rolls for 10 min before they were removed. The final product composite sheet is around 0.4–0.5 mm in thickness. Both the through plane and in plane thermal conductivity was measured. Fig. 1c shows the scheme of sample preparation routes.
3. Characterization 3.1. Scanning electron microscopy (SEM) To investigate the dispersion of xGnP particles in paraffin matrix at different loadings, disk samples prepared by casting and two roll milling were liquid nitrogen fractured. The fractured surface was oxygen plasma treated for 5 min with a Radio Frequency power at 550 W. The treated sample was later gold coated with the Denton Vacuum sputter coater to ensure that the sample is electrically
1813
conductive. A JEOL (JSM 6400) scanning electron microscope with an accelerating voltage of 15 kV was used to examine the fractured composite surface at a working distance of 15 mm. 3.2. Thermal conductivity The thermal conductivity of paraffin/xGnP composites was measured with a Unitherm (TM) Model 2022 (Anter Corporation, Pittsburgh, PA). The test was performed according to ASTM E1530. For samples prepared by cast molding, the average thickness of the sample is 5 mm, and it can be directly used for through-plane thermal conductivity measurement. However, for the samples prepared by two roll mill, the through plane thermal resistance is too small to be measured due to the very small thickness of the composite thin film (0.4–0.5 mm). 4 or 5 sheets were stacked in parallel so that the resistance falls in the range of the instrument. As for the in-plane measurement, the sample was made by compressing the thin sheets released from the roll mills in a closed rectangular mold to a desired thickness, and used a circular iron punch to produce a 1 in disk sample from the cross section of the block (shown in Fig. 2a and b). All the samples were tested at 20 1C under an applied load of 20 psi. 3.3. Electrical conductivity measurement The resistivity of paraffin/xGnP nanocomposites was measured with a Gamry instrument under FAS2TM Femtostat plug system
Fig. 2. (a) Schematics for in plane thermal conductivity measurement of the two roll milled sample and (b) photograph of the sample prepared from compressing the two roll milled sheets to a desired thickness (1 in) and a circular cross section of that was produced by a punch;
1814
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
and potentiostatic mode. A rectangular portion was fabricated by TM a Buehler Isomet low speed saw from the cast molded disk samples for the measurements. The resistance of the sample was measured in the frequency range from 0.6 Hz to 100 kHz and resistivity of samples were calculated by taking the geometry of the sample into account using the following equation:
atmosphere. The temperature range scanned was from 25 to 500 1C with a 25 1C/min ramp rate and a 4 1C/min resolution upon thermal decomposition events.
r ¼ RA=L
4.1. Morphology
ð1Þ
where R is the resistance of the sample at 1 Hz, A is the cross section area of the sample and L is the length of the sample. 3.4. Differential scanning calorimetry (DSC) A sample size of around 5–10 mg was loaded to the DSC sample cell and the data was collected for the 2nd run at a scanning rate of 5 1C/min. The solid–solid transformation in paraffin wax, the melting temperature, the crystallization temperature, and the latent heat of paraffin were analyzed by Universal Analysis 2000 Version 4.5A. 3.5. Thermogravimetric analysis (TGA) A sample of 10–15 mg of composite samples was loaded to the Pt pan and the experiments were conducted in nitrogen
4. Results and discussions
Fig. 3(a) shows the fractured cross section of the 4 wt% xGnP-1/ paraffin nanocomposites prepared by casting. The xGnP-1 particles are either embedded in the paraffin matrix evidenced by the white lines protruding from the background or lying on the surface suggested by their irregular but sharp edges different from the smooth and soft paraffin wax. At the same nanofiller loading, the xGnP-15 particles in the nanocomposites are more easily recognized due to their larger size shown in Fig. 3(b). It is also found that unlike the morphology reported by Kalaitzidou et al. [16] in polypropylene/xGnP nanocomposites that most xGnP particles tend to roll up or fold on itself during sample preparation (extrusion and injection molding), the xGnP particles in paraffin nanocomposites maintained their platelet-like shape as a result of the low viscosity of the polymer matrix and thus the high aspect ratio of xGnP is not negatively affected in sample fabrication. In the samples prepared by two roll mill with xGnP-15 loading at 4,
Fig. 3. (a) 4 wt% xGnP-1/paraffin by casting; (b) 4 wt% xGnP-15/paraffin by casting; (c) 4 wt% xGnP-15/paraffin by two roll mill; (d) 10 wt% xGnP-15/paraffin by two roll mill; (e) enlarged edge view of the nanoplatelets (scale bar 100 nm) and (f) enlarged edge view of one platelet (scale bar: 10 nm).
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
In order to find out the effect of dispersion of xGnP particles of different sizes on the electrical conductivity of the resulting nanocomposites, the resistivity of paraffin/xGnP-1 and paraffin/ xGnP-15 samples were measured and compared in Fig. 4. For nanocomposites with xGnP-15 as fillers, the resistivity undergoes a sharp decrease from 1010 (O cm) to around 103 (O cm) at 1 vol% which suggests that conductive paths were formed in paraffin. However, in xGnP-1 composites, the system does not percolate even at 2 vol%. Despite the fact that the absolute number of nanofillers in xGnP-1 composites is almost 250 (15 15) times more than for xGnP-15 composites at the same loading level, the aspect ratio of xGnP-1 is much smaller than xGnP-15 (100 vs. 1000). It can also be seen from SEM micrographs that the plateletlike morphology for xGnP-15 particles are maintained better than xGnP-1, which shed light on the importance of filler morphology and aspect ratio on the electrical conductivity of nanocomposites. 4.3. Thermal conductivity paraffin/xGnP nanocomposites The thermal conductivity of paraffin/xGnP nanocomposites prepared by casting at different filler concentrations was measured and shown in Fig. 5(a). The thermal conductivity increases linearly for xGnP-15 samples up to 2 vol% beyond which the data exhibit super linear behavior. In xGnP-1 composites, thermal conductivity also increases linearly with nano platelet loading, but the improvement is not as effective as xGnP-15 although the number density of nanofiller particles is much higher at the same loading level. Compared with larger filler particles, smaller particles introduce more phonon scattering interfaces at the boundaries. Heat transport in graphite nanoplatelets occurs by phonons of varying frequencies. In the xGnP polymer nanocomposites, phonons are
12 Resistivity (ohm cm) Log scale
xGnP-15 composite PCM
10 xGnP-1 composite PCM
8 6 4 2 0 0
1
2 3 xGnP loadings (vol%)
4
Fig. 4. Electrical resistivity of xGnP/paraffin composite PCM.
5
xGnP-15 nanocomposites
Thermal conductivity (W/mK)
4.2. Electrical conductivity of paraffin/xGnP nanocomposites
3.5 3
xGnP-1 nanocomposites Curve fit beyond percolation
2.5 2 1.5 1 0.5 0 0
0.02
0.04
0.06
xGnP volume fraction 3.5 two roll mill through plane
3 Thermal conductivity (W/mK)
and 10 wt% shown in Fig. 3(c and d, respectively). It is very clear that particles assume a preferential alignment as indicated by the white lines in the micrograph. Fig. 3(e and f) shows the enlarged edge views of the nanoplatelets, it is also evident that the individual nanoplatelet is thin with thickness less than 10 nm and thus they have a very high aspect ratio. The distance between particles was increased due to the expansion of material at the rolling nip point and it is the same compressive and shearing force that contributed to the alignment of the particles.
1815
two roll mill in plane cast through plane 2.414
2.5 2
1.71
1.5 1
0.737
0.701
0.496
0.5
0.31
0 0.02
0.05 xGnP-15 volume fraction
Fig. 5. (a) Through plane thermal conductivity of xGnP/paraffin prepared by casting and curve fitting with Foygel’s model above percolation and (b) thermal conductivity of xGnP-15/paraffin prepared by two roll mill.
transported from one xGnP platelet to another via the polymer in between. Since there is a difference in acoustic properties between the polymer and graphite nanoplatelets and in this case, the paraffin wax and xGnP interacts with each other through weak Van der Waals forces, only low frequency phonon vibration modes are available to carry a small amount of heat energy. The high frequency phonons, which are the major heat energy carriers, interact with other phonons before they could be transferred to some low energy vibration states to couple with the polymer matrix. As a result, a very high thermal interface resistance exists between the boundaries and gives rise to a reduction in heat transfer in nanocomposites. This process would be more prevalent in nanocomposites containing finer filler particles due to a higher density of interfaces. Given the fact that the xGnP-15 is superior to xGnP-1 particles in improving the thermal conductivity of paraffin nanocomposites, two samples prepared by two roll mill containing 2 and 5 vol% xGnP-15 in paraffin matrix were used for the measurements. Both samples show higher thermal conductivity in the in-plane direction rather than in the through-plane direction shown in Fig. 5(b). Aligning the particles in its basal plane direction effectively utilizes the high in-plane thermal conductivity of the graphite
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
nanoplatelets. However, the thermal conductivity of the two roll milled sample is not as high as the cast mold samples at the same loading level. On the one hand, random orientations of xGnP in paraffin for the cast molded samples leads to overlapping among particles which facilitates phonon transport due to the larger thermal contact area. On the other hand, higher xGnP loadings also lead to possible aggregation in the wax which contributes to a large scattering in the data. The two roll mill process, by comparison, not only aligns the particles in the matrix but also breaks up the aggregates and separates them, destroying the percolating networks and introduces more phonon scattering at the interfaces. 4.4. Effect of thermal interface resistance
bx ¼
g 2ðK11 Km Þ ; g K11 þ Km
bz ¼
g K33 1 Km
ð2Þ
where f is the volume fraction of graphite platelets, Assume the graphite platelet is coated with a very thin interfacial thermal barrier g g layer, K11 and K33 are the equivalent thermal conductivities along transverse and longitudinal axes of a graphite platelet, respectively, and can be expressed as: g K11 ¼
Kg ; 1 þð2ak Kg =tKm Þ
g K33 ¼
Kg 1 þð2ak Kg =dKm Þ
ð3Þ
where t and d are the thickness and diameter of xGnP; and ak is the Kapitza radius defined by: ak ¼ Rk Km
Rk = 7E-8
Rk = 1E-8
Rk = 3E-8
Rk = 5E-8
Rk = 9E-8
10
1 0.1
As already mentioned in the previous section, the thermal interface resistance presents a significant barrier for phonons to transfer from one conductive filler particle to another surrounded by a polymer matrix, it is interesting to predict the thermal interface resistance from experimental data. Many models have been proposed to predict the thermal conductivity of nanocomposites, but few have taken the interfacial resistance into account. Nan et al. [17] proposed a model based on carbon nanotubes and assumed an interfacial barrier layers between the nanotubes and the polymer, a Maxwell–Garnett type effective medium approach was used to predict the thermal interface resistance. For the case of isotropically oriented ellipsoidal particles including platelets, the thermal conductivity ratio Ke/Km (defined as effective thermal conductivity vs. matrix thermal conductivity) can be expressed by : Ke 3 þ f ðbx þ bz Þ ¼ ; 3f bx Km
Experimental
Thermal conductivity ratio (ke/km) Log scale
1816
ð4Þ
where Rk is the thermal interface resistance. The model described above assumes that the nanofiller particles are completely surrounded by the polymer matrix and they do not touch each other. Since the percolation threshold for xGnP-15/paraffin composites is around 1 vol% as suggested by electrical resistivity measurement, the experimental data used for curve fitting is restricted to those below 1 vol%. Assume the diameter and thickness of the xGnP-15 particles are 15 mm and 10 nm and the in-plane and through-plane thermal conductivities are 3000 and 10 W/m K, respectively. The thermal interface resistance at the boundary of graphite and paraffin is estimated to be 7 9 108 m2 K=W, this Rk value is very close to the interfacial resistance across the carbon nanotube matrix reported by Huxtable et al. [19] which is about 8:3 108 m2 K=W. Fig. 6 shows the curve fitting result using this model. As the xGnP concentration increases to beyond the percolation threshold (1 vol%), a non-linear model presented by Foygel et al. [18] is used to model the nanocomposite system. This approach is based on using Monte Carlo’s simulation to model percolating clusters consisting of nanotubes or any randomly oriented conductive fillers of large aspect ratio. The key parameters such as the conductivity exponent of the percolation network and the contact resistance can be derived from the model. The thermal conductivity of the paraffin/xGnP-15 in the
1 volume fraction (%) Log scale
Fig. 6. Thermal conductivity of paraffin/xGnP-15 below percolation with Nan’s model (unit of Rk: m2K/W).
vicinity of the percolation threshold is described by the following equation:
sðf; aÞ ¼ s0 ½ffc ðaÞtðaÞ
ð5Þ
where s0 is the pre-exponential factor that depends on the conductivity of individual particle or the contact between them, ^ is their volume fraction, ^c is the critical volume fraction at the percolation threshold, t(a) is the conductivity exponent that depends on the aspect ratio a. From the result of electrical conductivity measurements, the percolation threshold for the paraffin/xGnP-15 nanocomposites is around 0.8% by volume. The least square fit to the equation with s0 and t(a) as the fitting parameters gives s0 ¼120 W/mK and t(a)¼1.26 for paraffin/ xGnP-15 composites. The resistance of individual particle or their contact resistance (whichever is larger) R0 could be estimated by equation: R0 ¼ ðs0 Lfc
tðaÞ 1
Þ
ð6Þ
where L is the lateral dimension of xGnP particles. The obtained value R0 ¼2.4 105 K/W is in good agreement with the thermal resistance of an individual xGnP-15 particle: RxGnP ¼2 105 K/W, estimated using an experimental room temperature value of sxGnP ¼3000 W/mK as the thermal conductivity, length L¼15 mm and thickness t ¼10 nm. Thus it is suggested that for a nanocomposite system above percolation threshold, the effective thermal conductivity is dominated by the conductive particles. 4.5. The effect of xGnP particles on thermal properties of nanocomposites Differential scanning calorimetry was used to investigate the effect of xGnP particles on thermal physical properties of paraffin nanocomposites and its capacity to store thermal energy. Both the melting and cooling behaviors of paraffin/xGnP were recorded, two peaks were observed from the DSC analyses which correspond to solid–solid transformation and melting/crystallization of paraffin wax shown in Fig. 7(a–d). The relevant enthalpies were calculated by integration the peaks above the base line given by the software itself. The temperatures that characterize the melting and crystallization peak are not affected significantly by addition of xGnP particles. The melting and crystallization enthalpies for paraffin nanocomposites at low xGnP concentration do not deviate from neat paraffin very much as opposed to the samples at high loading levels when the composite’s ability to absorb and release heat is degraded because of the increasing replacement of paraffin wax with xGnP particles. However, the presence of more xGnP particles in the paraffin matrix does not negatively affect the thermal
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
Neat Paraffin
2wt% xGNP15/Paraffin 8wt% xGNP15/Paraffin
0.5
2.5
0
2
Heat flow (W/g)
Heat flow (W/g)
4wt% xGNP15/Paraffin
-0.5 -1 -1.5 -2 -20
0
20
40
60
80
Neat Paraffin
2wt% xGNP15/Paraffin
4wt% xGNP15/Paraffin
8wt% xGNP15/Paraffin
1.5 1 0.5 0 -20
-2.5 100
0
20
40
60
80
100
Temperature (C)
Neat Paraffin
2wt% xGNP1/Paraffin
Neat paraffin
2wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
0.5
2.5
0
2
Heat flow (W/g)
Heat flow (W/g)
Temperature (C)
-0.5 -1 -1.5 -2 -2.5 -20
1817
0
20
40 60 Temperature (C)
80
100
1.5 1 0.5 0 -20
0
20
40 60 Temperature (C)
80
100
Fig. 7. (a) Heating curves of xGnP-15/paraffin; (b) cooling curves of xGnP-15/paraffin; (c) heating curves of xGnP-1/paraffin and (d) cooling curves of xGnP-1/paraffin. Table 1 Heating and cooling characteristics of xGnP-15 composite PCM (average of three samples).
Table 2 Heating and cooling characteristics of xGnP-1 composite PCM (average of three samples).
Sample names
Melt peak (1C)
Melting enthalpy of composite (J/g)
Melting enthalpy of wax (J/g)
Cryst. peak (1C)
Cryst. enthalpy of composite (J/g)
Cryst. enthalpy of wax (J/g)
Sample names
Melt peak (1C)
Melting enthalpy of composite (J/g)
Melting enthalpy of wax (J/g)
Cryst. peak (1C)
Cryst. enthalpy of composite (J/g)
Cryst. enthalpy of wax (J/g)
Neat 2 wt% 4 wt% 8 wt%
54.54 54.82 55.13 54.9
185.9 188.6 183.5 171.9
185.9 192.4 191.1 186.8
51.73 51.27 51.87 51.63
186.8 189.9 185 172.9
186.8 193.7 192.7 187.9
Neat 2 wt% 4 wt% 8 wt%
54.54 54.33 53.97 54.35
185.9 186.05 180.4 176.65
185.9 189.85 187.9 192
51.73 52.09 52.22 51.94
186.8 187.15 180.2 178.1
186.8 191 187.7 193.5
physical properties of wax itself. This can be further supported by compensating the presence of xGnP in the composites and recalculating the melting and crystallization enthalpies for paraffin wax alone in the nanocomposites using Equation (7). It is found that for both xGnP-1/xGnP-15 composites, the heat absorbing and releasing capacity during heating and cooling for paraffin did not degrade but improve slightly with xGnP loadings.
DHwax ¼
DHcomp 1xG
ð7Þ
where DHcomp is latent heat of composite PCM with xGnP. DHwax is latent heat of wax in PCM. xG is the weight fraction of xGnP. Tables 1 and 2 collected data from paraffin nanocomposites containing 0, 2, 4 and 8 wt% xGnP-15 or xGnP-1. 4.6. Thermal gravimetric analysis Since graphite is known for its chemical inertness and resistance to thermal degradation, thermal decomposition of paraffin
nanocomposites upon addition of xGnP particles was analyzed by TGA in nitrogen atmosphere shown in Fig. 8. It is shown in the analysis that thermal decomposition of paraffin nanocomposites gradually shifted to a higher temperature with increasing amount of xGnP particles. Although the interaction between graphite platelets and paraffin is weak, xGnP particles still interact with the surrounding matrix and improve the overall stability of composite PCM. The remaining weight at 500 1C, at which temperature paraffin wax was completely thermally decomposed but the graphite is unaffected, was close to the initial loadings 0, 2, 4 and 8 wt%.
5. Conclusions Exfoliated graphite nanoplatelets of two sizes (xGnP-1 and xGnP-15) were successfully mixed with paraffin wax to make a composite PCM of high latent heat and high thermal and electrical conductivity. Nanocomposites prepared by casting with xGnP-15
Weight retained
1818
J. Xiang, L.T. Drzal / Solar Energy Materials & Solar Cells 95 (2011) 1811–1818
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Neat Paraffin
2wt% xGNP15/Paraffin
4wt% xGNP15/Paraffin
8wt% xGNP15/Paraffin
Acknowledgements The author appreciated the generous help from Brian Rook from Composite Materials and Structures Center on operating the two roll mill. Financial support for this project was provided by the Michigan Economic Development Corporation through the 21st Century Jobs Fund. References
0
100
200
300
400
500
600
Weight retained
Temperature (C)
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
100
Neat Paraffin
2wt% xGNP1/Paraffin
4wt% xGNP1/Paraffin
8wt% xGNP1/Paraffin
200
300
400
500
600
Temperature (C) Fig. 8. TGA analysis of xGnP-1/paraffin and xGnP-15/paraffin nanocomposites in N2 atmosphere.
show lower electrical percolation and higher thermal conductivity than xGnP-1 counterparts due to a higher interface density. Two roll milling enables the platelet particles to orient uniformly in the matrix enhancing the in-plane thermal conductivity of the nanocomposites. However, the milling process also separates the particles disrupting percolated network and increasing polymer graphite interfaces. The interfacial thermal resistance estimated by Nan’s model to be 7–9 10 8 m2K/W, was found to be comparable to that of carbon nanotubes. Experimental data also fit Foygel’s model suggesting that thermal conductivity above the percolation point for xGnP-15 PCM was more affected by the intrinsic conductivity of fillers instead of the interface. DSC results show the latent heat of paraffin nanocomposites were not degraded by adding xGnP. On the contrary, latent heat of paraffin itself and the thermal stability of the composite PCM improved slightly in the presence of xGnP.
[1] S.M. Hasnain, Review on sustainable thermal energy storage technologies, part I: heat storage materials and techniques, Energy Convers. Manage 39 (1998) 1127–1138. [2] M. Telkes, Thermal storage for solar heating and cooling, in: Proceedings of the Workshop on Solar Energy Storage Subsystems for the Heating and Cooling of Buildings, Charlottesville, Virginia, USA, 1975. [3] K.K. Pillai, B.J. Brinkworth, The storage of low grade thermal energy using phase change materials, Appl. Energy 2 (1976) 205–216. [4] A.S. Luyt, Role of polymers in developing phase change materials for energy storage, eXpress Polymer Lett 2 (2008) 147. [5] R. Velraj, R.V. Seeniraj, B. Hafner, C. Faber, K. Schwarzer, Experimental analysis and numerical modeling of inward solidification on a finned vertical tube for a latent heat storage unit, Sol. Energy 60 (1997) 281–290. [6] J. Fukai, M. Kanou, Y. Kodama, O. Miyatake, Thermal conductivity enhancement of energy storage media using carbon fibers, Energy Convers. Manage 41 (2000) 1543–1556. [7] A. Elgafy, K. Lafdi, Effect of carbon nanofiber additives on thermal behavior of phase change materials, Carbon 43 (2005) 3067–3074. [8] S. Pincemin, R. Olives, X. Py, M. Christ, Highly conductive composites made of phase change materials and graphite for thermal storage, Sol. Energy. Mater. Sol. C. 92 (2008) 603–613. [9] S. Kim, L.T. Drzal, High latent heat storage and high thermal conductive phase change materials using exfoliated graphite nanoplatelets, Sol. Energy. Mater. Sol. C. 93 (2009) 136–142. [10] G.-W. Lee, J.T. Lee, S.-S. Lee, M. Park, J. Kim, Comparison of thermal properties between inorganic fillers and acid-treated multiwall nanotube/polymer composites, J Mater. Sci. 40 (2005) 1259–1263. [11] C. Park, Z. Ounaies, K.A. Watson, R.E. Crooks, J. Smith, S.E. Lowther, J.W. Connell, E.I. Siochi, J.S. Harrison, T.L.S. Clair, Dispersion of single wall carbon nanotube by in-situ polymerization under sonication, Chem. Phys. Lett. 364 (2002) 303–308. [12] F.H. Gojny, M.H.G. Wichmann, B. Fiedler, I.A. Kinloch, W. Bauhofer, A.H. Windle, K. Schulte, Evaluation and identification of electrical and thermal conduction mechanisms in carbon nanotube/epoxy composites, Polymer 47 (2006) 2036–2045. [13] J. Heiser, J.A. King, Thermally conductive carbon filled nylon 6,6, Polym. Compos. 25 (2004) 186–193. [14] H. Fukushima, Graphite nanoreinforcements in polymer nanocomposites. Ph.D Thesis, Michigan State University, East Lansing, MI, USA, 2003. [15] K. Kalaitzidou, Exfoliated graphite nanoplatelets as reinforcement for multifunctional polypropylene nanocomposites. Ph.D Thesis, Michigan State University, East Lansing, MI, USA, 2006. [16] K. Kalaitzidou, H. Fukushima, L.T. Drzal, Multifunctional polypropylene composites produced by incorporation of exfoliated graphite nanoplatelets, Carbon 45 (2007) 1446–1452. [17] C.W. Nan, G. Liu, Y. Lin, M. Li, Interface effect on thermal conductivity of carbon nanotube composites, Appl. Phys. Lett. 85 (2004) 3549–3551. [18] M. Foygel, R.D. Morris, D. Anez, S. French, V.L. Sobolev, Theoretical and computational studies of carbon nanotube composites and suspensions, Phys. Rev. B 71 (2005) 104201–104208. [19] S. Huxtable, D.G. Cahill, S. Shenogin, L. Xue, R. Ozisik, P. Barone, M. Usrey, M.S. Strano, G. Siddons, M. Shim, P. Keblinski, Interfacial heat flow in carbon nanotube suspensions, Nat. Mater. 2 (2003) 731–734.