Solar Energy Materials & Solar Cells 96 (2012) 124–130

Contents lists available at SciVerse ScienceDirect

Solar Energy Materials & Solar Cells

0927-02

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/solmat

Effective dispersion of multi-wall carbon nano-tubes in hexadecane throughphysiochemical modification and decrease of supercooling

Shuo Zhang a, Jian-Yong Wu b, Chi-Tat Tse b, Jianlei Niu a,n

a Department of Building Service Engineering, The Hong Kong Polytechnic University, Kowloon, Hong Kong, PR Chinab Department of Applied Biology & Chemical Technology, The Hong Kong Polytechnic University, Kowloon, Hong Kong, PR China

a r t i c l e i n f o

Article history:

Received 8 April 2011

Accepted 14 September 2011Available online 2 October 2011

Keywords:

Hexadecane

Multi-wall carbon nano-tube

Dispersion

Nucleating agent

Decrease of supercooling

48/$ - see front matter & 2011 Elsevier B.V. A

016/j.solmat.2011.09.032

esponding author. Tel.: þ852 27667781; fax

ail address: bejlniu@polyu.edu.hk (J. Niu).

a b s t r a c t

Prevention of supercooling is essential for phase change material (PCM) utilization. In this study, multi-

wall carbon nano-tube (MWCNT) particles were dispersed in an organic liquid n-hexadecane used to

decrease supercooling. Various surfactants were tested as additives to overcome the rapid aggregation

and sedimentation of the nanoparticles in the organic liquid. Stable and homogenous dispersion was

attained through surface modification of the MWCNT particles with strong acids H2SO4 and HNO3, plus

the addition of 1-decanol as a surfactant to the organic liquid. Thermal analysis of the n-hexadecane

with well dispersed MWCNT particles at concentrations ranging from 0.1% to 10% w/w by differential

scanning calorimeter (DSC) indicated that the supercooling of n-hexadecane was significantly

decreased with the concentration of 0.1% and 0.5% but only slightly with the concentrations over

1.0%. It appears that well dispersed nanoparticles provided stable foreign nuclei of proper size to

promote the heterogeneous nucleation process and accelerate crystallization process, thus the super-

cooling was significantly reduced. The obvious effects of MWCNT particles on the decrease of

supercooling of n-hexadecane provide promising way of improving the performance of system energy

efficiency in building cooling and heating applications.

& 2011 Elsevier B.V. All rights reserved.

1. Introduction

In recent years, microencapsulated phase change material(MPCM) slurry and PCM emulsion have been investigated forbuilding heating and cooling applications in view of their potentialthermal performance and utilization flexibilities [1–3]. However,one of the major problems in using the PCM as thermal storagematerial is supercooling, i.e., when a PCM liquid is cooled, freezingusually occurs at a lower temperature than the melting point. Asthe latent heat is only released below the supercooled temperature,large temperature difference between charging and discharging isneeded to fully utilize the latent heat, which is undesirable for theenergy efficiencies of energy storage applications [4].

As suggested by some researchers [5], an effective approachfor decreasing supercooling is the addition of liquid or solidnucleating agents to the PCM liquids as the seeds and catalystsfor nucleation and crystal growth. A liquid nucleating agent has ahigher melting point than that of the main heat storage material,and is first solidified upon cooling to act as a nucleus of crystalformation. Several studies have been conducted on liquid nucle-ating agents in various liquids, such as 1-Tetradecanol (2 wt%) for

ll rights reserved.

: þ852 27746146.

microencapsulated n-Tetradecane [6], 1-octadecanol (10 wt%) formicrocapsulated n-octadecane [4], and paraffin wax (0.8–10 wt%)for tetradecane and hexadecane paraffin-in-water emulsion [7].Solid nucleating agents, such as nanoparticles and impurityparticles, acting as nucleation centers to enhance the nucleationprogress have shown promising application potentials. As pro-posed by Oliver and Calvert [8], the crystallization processes ofmost PCM liquids are controlled by the heterogeneous nucleationmechanism. The phase-transition behavior of the PCM liquidsis complicated and very sensitive to small amounts of impurities.He et al. [9] showed that the addition of TiO2 nanoparticles intopure water effectively reduced the supercooling of water. Zhanget al. [10] reported that the effects of nanoparticles on super-cooling of pure water are strongly dependent on their surfacewettability. Among the three additive candidates, a-Al2O3 hadmore effect than g-Al2O3 and SiO2, and the effect was morenotable at a lower concentration of 0.3% than at 0.5%. However,up to now, there is no systematic method to select additive forreducing the supercooling. This is because the essential factorsaffecting the nucleation have not been clarified.

A major problem with the use of nanoparticles as nucleatingagents is the poor dispersibility of nanoparticles in the liquid.Because of their high aspect ratios, large specific surface area, andsubstantial van der Waals attractions, carbon nano-tubes tend toself-aggregate into bundles spontaneously. In addition, the high

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130 125

flexibilities increase the possibility of nano-tubes entanglementand close packing [11,12]. The structure of MWCNT particle playsan important role on its poor solubility and dispersivity in eitherwater or organic solvents [11]. The carbons on the tubular partare joined with three neighbor carbons (1201) to form thehexagonal structure hence the hybridization state is sp2, whilethe carbons on the end caps are pentagons and heptagons, whichare sp3 hybridized. The sp2 hybridized carbons stay in the pzorbital and are responsible for p–p interactions causing aggrega-tion between MWCNT particles. But these p-electrons also pro-mote adsorption of various chemicals on the MWCNT particlesurface via p–p stacking interactions [11,13]. Two commonapproaches are available for dispersion of MWCNT: the use ofexternal mechanical forces and physicochemical modification ofthe particle surface properties [14]. Mechanical approach can onlytemporarily break the interactions between MWCNT particles andthe particles will aggregate again after the force is removed.Physicochemical approaches can be classified into physical andchemical methods. Physical methods mostly involve the adsorp-tion of chemical surfactants onto the MWCNT surface, which donot change the chemical properties of the MWCNT surface.Chemical methods involve the surface modification or functiona-lization of the particles with various chemical reagents toimprove their dispersibility in the liquid and to increase theresistance to aggregation.

Our previous study [15] has investigated the effects of super-cooling on the effective MPCM thermal storage capacity and theimpact on building cooling utilization. This study is to evaluate theuse of MWCNT nanoparticles as nucleating agent to decreasesupercooling in the organic liquid PCM n-hexadecane. Addition ofvarious chemical surfactants and chemical modification of theparticle surface were attempted to attain a stable and well dispersedsuspension of nanoparticles in the organic liquid. The effectivenessof well dispersed MWCNT particles as nucleating agent for decreas-ing of supercooling was evaluated at various concentrations.

2. Experimental

2.1. Materials

In this study, n-hexadecane C16H34 (99%) was chosen as thePCM liquid (purchased from International Laboratory USA), andmulti-walled carbon nano-tube (MWCNT) as the nucleating agent,which had an outer diameter 10–20 nm, length 0.5–2 mm and495% purity (purchased from Chengdu Organic Chemicals Co. Ltd.,Chinese Academy of Sciences, China). Several surfactants weretested as additives for dispersing the MWCNT particles in hexade-cane, including sodium dodecyl sulfate (SDS), cetyl trimethylam-monium bromide (CTAB), polyvinylalcohol (PVA), polyethyleneglycol (PEG), tetramethylethylenediamine (TEMED), Triethylamine(TEA), glacial acetic acid (AcCOOH), 1-decanol (decan-1-ol), sal-icylic acid (SA), Tween-80 (polysorbate 80), and Triton X-100(C14H22O(C2H4O)n), which were all of analytical grade.

2.2. Dispersion of MWCNT nanoparticles in hexadecane

Two methods were attempted to improve the dispersion ofMWCNT particles in the PCM liquid, the addition of a chemicalsurfactant into the PCM liquid and the surface-modification of thenanoparticles plus the addition of a surfactant. All dispersionexperiments were performed in glass test tubes using the ultra-sonication technique. For the dispersion with surfactants, theMWCNT particles were added at 30–50 mg/L to hexadecane ineach test tube, and one of the surfactants mentioned above wasadded to the tube at 1% (w/v). The tubes were ultrasonicated by

an ultrasound probe (Sonics Vibra-CellTM, model VCX130) for5 min at 30% amplitude of power.

Surface modification of the MWCNT particles was performedin a mixture of two strong acids, concentrated H2SO4 (98%) andHNO3 (70%) at 3:1 volume ratio. MWCNT particles was added toH2SO4–HNO3 in a test tube and sonicated in an ultrasonic bath for6 h and then heated with reflux at 65 1C for 4 h. After coolingdown to room temperature, the acid liquid was diluted withdeionized water and the MWCNT particles were spun down at18,000 rpm for 2 h. After removal of the liquid, the solid particleswere dried at 80 1C in an oven for about 24 h.

The surface-modified MWCNT particles were re-dispersed inhexadecane by ultrasonication using an ultrasound probe with 30%amplitude of power for 5 min. Each of the surfactants was added tothe dispersion and the dispersion was sonicated in an ultrasoundbath for 20 min. Based on the experimental results as shown laterin Section 3.1, surface-modification plus the addition of 1-decanolto the PCM liquid was the most effective method for the stabledispersion of the nanoparticles. Therefore, the surface-modifiedMWCNT particles were dispersed in 1-decanol by ultrasonicationfor 10 min to attain an excess amount of 1-decanol being coated onthe MWCNT particles. This yield a stock of 1-decanol-coated,surface-modified MWCNT particles, which was applied to generatethe final MWCNT-hexadecane slurry at various concentrationsfrom 0.1 wt% to 10 wt% for the following studies.

2.3. Characterization and analysis of nanoparticles

Surface properties of the modified MWCNT particles wereanalyzed by Fourier transform infrared (FTIR) spectrometry on aNicolet Avatar 360 FTIR instrument using the KBr pellet method.

The size distribution of nanoparticles dispersed in hexadecanewas measured by dynamic light scattering analysis using aMalven Zetasizer (model 3000 HSA) at 901 scattering angle and25 1C. Each sample was scanned 100 times and the averageparticle size (nm) and the polydispersity index were computedby the Zetasizer 3000HSA-Advanced Software 15.

The morphology of original and modified MWCNT particleswas examined with transmission electron microscopy (TEM)using a JEOL 2011 instrument at a high voltage of 200 kV andpoint resolution 0.23 nm. The TEM was operated by the Jeol FasTEM software and the images were processed by Gatan DigitalMicrograph.

2.4. Thermal analysis of PCM

Thermal analysis of the PCM with the modified MWCNT as thenucleating agent was performed on a differential scanning calori-meter (DSC) (METTLER TOLEDO DSC-822e) equipped with a thermalanalysis data station. Samples (10 mg each) were placed intohermetically sealed aluminum pans and treated with the followingtemperature program: the sample was first cooled to the initialtemperature of �5 1C for at least 15 min for stabilization, and thenheated from �5 1C to 30 1C at a rate of 5 1C/min, held for 5 min at30 1C, and finally cooled from 30 1C to �5 1C at a rate of 5 1C/min.STARe software was used to analyze and plot the thermal data.

3. Results and discussion

3.1. Dispersion of MWCNT in hexadecane

3.1.1. Dispersion of original MWCNT by surfactants

Table 1 shows the suspension time, i.e. the period for completesedimentation of the original MWCNT particles in hexadecanesupplemented with the eleven surfactant chemicals. A longer time

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130126

period means a greater effectiveness of a surfactant in enhancing thesuspensions, which can be described in this descending order:Tween-80oTriton X-100ocontrol�AcCOOH�SDS�CTABoPVA�PEGoSAoTEAo1-decanoloTEMED.

Compared with that of the control (30 min), the suspensionperiod of MWCNT in hexadecane was shorter or decreased withthe addition of several of the surfactants, most significantly thetwo nonionic surfactants, Tween-80 and Triton X-100. It appearsthat these surfactants accelerated aggregation and sedimentationof MWCNT particles in the organic liquid. The ionic surfactantsincluding SDS and CTAB had only slight effect on the dispersion ofMWCNT, probably because the hydrophobic tails of these surfac-tants had a stronger interaction with the hexadecane than withMWCNT. The polymeric surfactants including PVA and PEGgreatly increased the suspension period, which was probablyachieved by their attachment or adsorption onto the MWCNTparticles, creating steric resistance to aggregation. However, theirhigher molecular weight and low solubility in the hexadecanemay be unfavorable for the dispersion. In comparison, the aminesincluding TEA and TEMED were more favorable. Their aminogroups perhaps initially attached to the MWCNT surface by theelectron donation of NH group and the steric tail, increasing theresistance to aggregation, though not sufficient to overcome thestrong p–p interaction among MWCNT particles. As for the polarmolecules, 1-decanol had a remarkable effect to enhance thedispersion, though AcCOOH and SA only had slight effect.1-decanol may interact with the nanoparticles through electro-static force by its negatively charged OH group, as the NH groupwith the amines. Among all surfactants tested, TEMED and1-decanol were the most effective to enhance the dispersion.

3.1.2. Surface modification of MWCNT

Surface modification of the MWCNT particles was performedin a mixture of two strong acids, concentrated H2SO4 and HNO3.The results of surface properties of the modified MWCNT particleswere analyzed by FTIR spectroscopy, shown in Fig. 1. There is

Table 1Effects of surfactants on dispersion of original MWCNT in hexadecane.

Surfactant Suspension

time (min)

Surfactant Suspension

time (min)

Control 30 TEA 60

SDS 25 AcCOOH 30

CTAB 25 SA 40

PVA 35 1-decanol 70

PEG 35 Tween-80 15

TEMED 85 Triton X-100 25

Fig. 1. FTIR spectrum of the m

a peak appearing in the FTIR spectrum of the acid treated MWCNTparticles around 1764 cm�1, which is assigned to the CQO bondstretching. The peak around 2922 cm�1 is the anhydrous O–Hbond stretching. The peaks round 1341 cm�1 and 1050 cm�1areO–H bond bending. The peak around 1195 cm�1 is the C–Ostretching. These peaks indicate that COOH groups were intro-duced to the end or the side of the MWCNT particles, therebyhaving functionalized or modified surfaces of MWCNT particles.The peak around 1107 cm�1 is the C–O stretching in acidanhydride, which may be caused by polymerizing the acid groupduring the drying process in oven.

It took longer time to disperse the modified MWCNT particlesthan the original MWCNT particles 110 min versus 45 min. Thesteric effect of carboxylic acid group as well as the irregularity ofthe modified MWCNT particle surface presented in lower effec-tiveness of packing and slower rate of aggregation. The aggregationof modified MWCNT particles eventually appeared, probably dueto low solubility of COOH in hexadecane, the hindrance providedby small size of the acid group, and also the hydrogen bondinginteraction among acid groups in modified MWCNT particles.

3.1.3. Dispersion of modified MWCNT by surfactants

The effects of the tested surfactants on the dispersion of themodified MWCNT in hexadecane are shown in Table 2. Theeffectiveness can be presented in the descending order: Tween-80�Triton X-100oAcCOOHoTEMED�PEGoPVAoTEAoSDS�controloCTABoSAo1-decanol.

It appears that the surfactants Tween-80 and Triton X-100accelerated the aggregation and sedimentation of the modifiedand original MWCNT particles in the hexadecane, probably becausethe two surfactants were not soluble well in hexadecane and thedensity was higher, so that the MWCNT particles sank to the bottomdue to the density difference. The ionic surfactants including SDSand CTAB had better effects on the dispersion of the modifiedMWCNT particles than that of the original particles, but they werestill not as good as expected since they could not be ionized in

odified MWCNT particles.

Table 2Effects of surfactants on dispersion of modified MWCNT in hexadecane.

Surfactant Suspension

time (min)

Surfactant Suspension

time (min)

Control 70 TEA 65

SDS 70 AcCOOH 20

CTAB 105 SA 140

PVA 40 1-decanol 290

PEG 25 Tween-80 17

TEMED 30 Triton X-100 15

0

10

20

30

40

50

60

70

80

0 1000 2000 3000 4000 5000

Inte

nsity

(%)

Particle Diameter (nm)

Fig. 3. Size distribution of MWCNT-1-decanol in hexadecane.

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130 127

hexadecane and are only dependent on dipole–dipole interaction.The polymer dispersion using PVA and PEG had similar results withthe original ones since they interacted on MWCNT by physicallywrapping, and thus only twisted the particles with little chemicalinteraction. Although TEMED had a remarkable effect on thedispersion of the original MWCNT particles, its effects on modifiedMWCNT particles was greatly weakened. It was because the electrondonating effect of amines on the original MWCNT no longer existedin the modified MWCNT, and the new interaction between themwas hydrogen bonding. The dispersion effect of TEMED might havebeen hindered by the di-amine forming a bridge-liked structure thatlinked MWCNT particles together. Since using long straight carbonchain mono-amine might improve the situation, long straightcarbon chain alcohol (1-decanol) and carboxylic acid (SA) were usedas substitutes. The hydrogen bonding in these two molecules isbelieved stronger than that in amines since the electronegative ofOH is stronger than NH. Polar molecule SA and 1-decanol showedhuge improvements on the dispersion of the modified MWCNT,though AcCOOH had little effect on both modified and originalMWCNT. It was because AcCOOH had little solubility in hexadecane,although the modified MWCNT had a very good solubility inAcCOOH due to the hydrogen bonding. In contrast, the polar groupof SA and 1-decanol could form hydrogen bonding in MWCNT, andthe long carbon chain tail rendered them well soluble in hexade-cane, which leads to a stable dispersion.

Among all chemicals used, 1-decanol showed the best perfor-mance of dispersion for both original and modified MWCNT. Henceit was chosen for the further investigation with increased concen-tration in modified MWCNT. The dispersion effect was greatlyincreased since 1-decanol could completely coat the MWCNTparticles. The schematic diagram of the interaction between1-decanol and modified MWCNT particle is shown in Fig. 2. Thevisible aggregation was negligible even after seven days.

Based on the dispersion analysis and comparison above, mod-ified MWCNT assisted by 1-decanol (MWCNT-1-decanol) werewell dispersed in liquid n-hexadecane to form a stable n-hexade-cane/MWCNT-1-decanol slurry. A series of such slurry of differentMWCNT-1-decanol concentrations (0.1–10% w/w) were preparedfor the further investigation of the effect on supercooling.

3.2. The particle size of MWCNT-1-decanol

The average hydrodynamic diameter of MWCNT-1-decanol inhexadecane was 2761.3 nm or 2.76 mm (Fig. 3) as measured bydynamic light scattering analysis (Malven Zetasizer) with the

CNT

C

O OH

OH

C

O

HO

C

O

OHO

H

C

OHO

OH

O

H

Decan-1-ol

Hydrogen bondings

Fig. 2. Schematic diagram of the interaction between 1-decanol and

modified MWCNT.

count rate at 200.3þ6.6 kilo count per second (kcps) at 25 1C. Therange difference within the repeated measurements wasþ36.1 nm, therefore the MWCNT particles were well dispersed.The polydisperse index was 0.585, which means the range ofparticle size distribution was medium.

3.3. Morphology of MWCNT

TEM micrographs of the original MWCNT and MWCNT-1-decanol are presented in Figs. 4 and 5, respectively. The sizeand shape of MWCNT particles can be observed clearly and areconsistent with the parameters provided by manufacturer, withan outer diameter of 10–20 nm and a length of 0.5–2 mm.Compared with the morphology of original MWCNT, the surfaceof MWCNT-1-decanol was rougher. There were some masses onthe surface of the nano-tubes, and the mass may be caused bysurfactant modification and the addition of decan-l-ol. Thesecan be observed more clearly in the TEM micrographs (Fig. 6),showing the crystal structure of nano-tubes.

3.4. Supercooling reduction

In the nucleation process, the formation of the nuclei isassociated with a change in the free energy of the system. At agiven supersaturation and temperature, there is a critical value ofthe free energy when stable nuclei of critical size are formed. Thisbehavior of the free energy change, DG, associated with theformation of the nucleus was shown in Fig. 7 as a function ofnucleus radius r. When the radius r of the nuclei is smaller than rn,the nuclei dissolve. However, when r4rn, the nuclei are stableand grow. DGn is a barrier for nucleation, and only when r4rn

there is a reduction in the free energy of the nucleus with anincrease in its size. The higher the activation barrier DGn, the moredifficult it is to attain stable nuclei. Nucleation process may behomogeneous or heterogeneous. Homogeneous nucleation ispossible when there is no external source, in ideally pure liquids,whereas heterogeneous nucleation occurs when the system con-tains nanoparticles and/or impurity particles [5].

According to the classical nucleation theory, well dispersedMWCNT particles are expected to act as seeds during the processof crystallization, thus to be used as nucleating agents to reducethe supercooling of hexadecane. DSC technique was used toanalyze the thermal property of the hexadecane/MWCNT-1-decanol slurry. Fig. 8 showed the DSC melting and freezing curvesof the pure hexadecane at 5 1C/min scanning rate and thecharacteristic temperatures including melting temperature Tm,melting peak temperature Tm.peak, freezing temperature Tf, andfreezing peak temperature Tf.peak. The difference between Tm.peak

and Tf.peak is the supercooling DT. It can be observed that the

Fig. 4. TEM micrographs of original MWCNT on different scales showing the morphology: (a) 200 nm and (b) 50 nm.

Fig. 5. TEM micrographs of well dispersed MWCNT-1-decanol on different scales showing the morphology: (a) 200 nm and (b) 50 nm.

Fig. 6. TEM micrographs of different MWCNT showing the crystal structure of MWCNT: (a) original MWCNT and (b) MWCNT-1-decanol.

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130128

melting/freezing peak temperature is 22.67 1C and 14.08 1C,respectively, so the supercooling is 8.59 1C.

The DSC heating/cooling curves of the hexadecane/MWCNT-1-decanol slurry obtained at a heating and cooling rate of 5 1C/minare presented in Fig. 9, and the melting and crystallizationproperties calculated from the DSC tests are presented inTable 3. It can be observed that the freezing temperature of thehexadecane/MWCNT-1-decanol slurry varied with the MWCNT

concentrations while the melting temperature remained funda-mentally unchanged. With the addition of the nucleating agent of0.1 wt%, the freezing temperature clearly increased from 14.08 1Cto 17.42 1C, and the supercooling of hexadecane decreased by43%, from 8.59 1C to 4.91 1C. With the addition of the nucleatingagent of 0.5 wt%, the supercooling reduction effect was alsosignificant but weaker than at the lower concentration. Whenthe concentrations of the nucleating agent exceeded 1.0 wt%, the

Fig. 7. Change in Gibbs free energy DG as a function of radius r of nucleus formed

in a supersaturated medium [5].

Freezing

Melting

Tf

Tf.peak

Tm

Tm.peak

Supercooling

Fig. 8. DSC melting and freezing curves of the pure hexadecane at 5 1C/min

scanning rate and the characteristic temperatures Tm, Tm.peak, Tf and Tf.peak.

Hea

t F

low

(W

/g)

Temperature (°C)

Exo

Endo

NA(10.0%)

NA(2.0%)

NA(1.0%)

NA(0.5%)

NA(0.1%)

NA(0.0%)

-10 0 10 20 30 40

-10 0 10 20 30 40

Hea

t F

low

(W

/g)

Temperature (°°C)

NA(0.0%)

NA(0.1%)

NA(0.5%)

NA(1.0%)

NA(2.0%)

NA(10.0%)

Exo

Endo

Fig. 9. DSC curves of the hexadecane with MWCNT of different fractions as the

nucleating agent (NA): (a) heating process and (b) cooling process.

Table 3Melting and crystallization properties of hexadecane with MWCNT of different

fractions as the nucleating agent (NA).

DHm (J/g) DHf (J/g) Tm.peak (1C) Tf.peak (1C) DT (1C)

NA (0.0 wt%) 227.13 225.81 22.67 14.08 8.59

NA (0.1 wt%) 196.58 196.33 22.33 17.42 4.91

NA (0.5 wt%) 256.28 250.59 22.25 15.42 6.83

NA (1.0 wt%) 237.90 239.77 23.83 13.67 10.16

NA (2.0 wt%) 244.81 244.22 22.75 13.83 8.92

NA (10.0 wt%) 219.60 223.03 22.70 14.08 8.67

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130 129

effect of reducing supercooling was not seen. Judging from theseresults, it appears that there existed an optimal or a most effectiveconcentration range of the nucleating agent and this agrees withexperimental research reported by other researchers [16,17]. Thefact that the mass concentration required is small has positivepractical implications.

3.5. Discussion

The crystal structure of n-hexadecane (even-carbon n-alkaneswith no26) is triclinic structure [5], and that of MWCNT particleswith a diameter of 10–20 nm (more than 25 A) is honeycombstructure [18]. The similar crystal structure of MWCNT andn-hexadecane is the basis for MWCNT to be used as the nucleat-ing agent. The experimental results help to confirm this hypoth-esis. Without any foreign additives, formation of stable nucleirelying on homogeneous nucleation is a ‘‘sluggish’’ process, whichis the main cause of large supercooling. In this experimentalinvestigation, the well dispersed nanoparticles provided stableforeign nuclei of proper size to promote the heterogeneousnucleation process and accelerate crystallization process, thusthe crystallization temperature was raised and the supercoolingwas significantly reduced.

The small concentration of additives required has a number ofpositive implications for the potential engineering application ofthe technology. Firstly, it will have negligible impact on thevolumetric latent heat of the PCM; secondly, the cost of the nanomaterials can be much reduced.

4. Conclusions

It was difficult to disperse the original MWCNT particles in ahexadecane liquid. The surfactant used to disperse MWCNT mustbe soluble to the hexadecane, and the results showed that thefunctional group in the surfactants that are supposed to reactwith the acid groups in MWCNT should be as exposed as possible,otherwise the branch near the functional group would producehindrance and lower the dispersion effect. Modified MWCNTparticles plus the addition of 1-decanol as surfactant weresuccessfully dispersed in hexadecane. It can be expected thatusing 1-decanol to disperse MWCNT in other higher molecularalkanes that are similar to hexadecane such as pentadecane andheptadecane is possible; the results may also imply that usingother fatty alcohols, such as nonadecan-1-ol and undecan-1-ol,which have similar structures to 1-decanol to assist to disperseMWCNT in hexadecane is also possible.

It is encouraging that well dispersed MWCNT of relatively lowconcentration had significant effect in reducing supercooling ofhexadecane. With the addition of 0.1 wt% MWCNT, the super-cooling of hexadecane can decrease by 43%, which produced themost significant effect among the test samples. It is also interest-ing to note that there was an effective concentration range ofnanoparticles for supercooling reduction, and better results

S. Zhang et al. / Solar Energy Materials & Solar Cells 96 (2012) 124–130130

cannot be obtained by continuously increasing nanoparticleconcentration.

For thermal cooling storage in air conditioning system andother building applications, reducing supercooling as much aspossible is essential. Large supercooling will decrease the COP ofthe cooling storage process in electric-driven cooling storagesystems, and reduce the utilization hours or the per-volumestorage capacities of natural cooling systems such as evaporativecooling system.

References

[1] M.M. Farid, A.M. Khudhair, S.A.K. Razack, S. Al-Hallaj, A review on phasechange energy storage: materials and applications, Energy Conversion andManagement 45 (2004) 1597–1615.

[2] N. Zhu, Z. Ma, S. Wang, Dynamic characteristics and energy performance ofbuildings using phase change materials: a review, Energy Conversion andManagement 50 (2009) 3169–3181.

[3] L. Huang, M. Petermann, C. Doetsch, Evaluation of paraffin/water emulsion asa phase change slurry for cooling applications, Energy 34 (2009) 1145–1155.

[4] X.-x. Zhang, Y.-f. Fan, X.-m. Tao, K.-l. Yick, Crystallization and prevention ofsupercooling of microencapsulated n-alkanes, Journal of Colloid and InterfaceScience 281 (2005) 299–306.

[5] K. Sangwal, Additives and Crystallization Processes: From Fundamentals toApplications, John Wiley & Sons Ltd., West Sussex, 2007.

[6] Y. Yamagishi, T. Sugeno, T. Ishige, H. Takeuchi, A.T. Pyatenko, An evaluationof microencapsulated PCM for use in cold energy transportation medium,in: Proceedings of the 31st Intersociety Energy Conversion EngineeringConference, IECEC 96, Washington, DC, USA, 1996, pp. 2077–2083.

[7] L. Huang, E. Gunther, C. Doetsch, H. Mehling, Subcooling in PCMemulsions—part 1: experimental, Thermochimica Acta 509 (2010) 93–99.

[8] M.J. Oliver, P.D. Calvert, Homogeneous nucleation of n-alkanes measured bydifferential scanning calorimetry, Journal of Crystal Growth 30 (1975)343–351.

[9] Q. He, W. Tong, Y. Liu, Experimental study on super-cooling degree ofnanofluids for cryogenic cool storage, Journal of Refrigeration 28 (2007)33–36 (in Chinese).

[10] X.J. Zhang, P. Wu, L.M. Qiu, X.B. Zhang, X.J. Tian, Analysis of the nucleation ofnanofluids in the ice formation process, Energy Conversion and Management51 (2010) 130–134.

[11] L. Vaisman, H.D. Wagner, G. Marom, The role of surfactants in dispersion ofcarbon nanotubes, Advances in Colloid and Interface Science 128–130 (2006)37–46.

[12] G. Sun, G. Chen, J. Liu, J. Yang, J. Xie, Z. Liu, R. Li, X. Li, A facile geminisurfactant-improved dispersion of carbon nanotubes in polystyrene, Polymer50 (2009) 5787–5793.

[13] J. Zhang, H. Zou, Q. Qing, Y. Yang, Q. Li, Z. Liu, X. Guo, Z. Du, Effect of chemicaloxidation on the structure of single-walled carbon nanotubes, The Journal ofPhysical Chemistry B 107 (2003) 3712–3718.

[14] R. Rastogi, R. Kaushal, S.K. Tripathi, A.L. Sharma, I. Kaur, L.M. Bharadwaj,Comparative study of carbon nanotube dispersion using surfactants, Journalof Colloid and Interface Science 328 (2008) 421–428.

[15] S. Zhang, J. Niu, Experimental investigation of effects of supercooling onmicroencapsulated phase-change material (MPCM) slurry thermal storagecapacities, Solar Energy Materials and Solar Cells 94 (2010) 1038–1048.

[16] Y. Liu, Study on Preparation and Thermal Properties of Phase ChangeNanocomposites for Cool Storage, Ph.D. Thesis, College of Power Engineering,Chongqing University, Chongqing, 2005 (in Chinese).

[17] K. Tomura, J. Kawasaki-shi, Heat storage substance, heat storage agent, heatreservoir, heat transport medium, cold insulation agent, cold insulator, meltingpoint control agent for heat storage agent, supercooling preventive agent forheat storage agent and method for producing main agent of heat storage agentheat transport medium, or cold insulation agent, United States, 2010.

[18] J. Tersoff, R.S. Ruoff, Structural properties of a carbon-nanotube crystal,Physical Review Letters 73 (1994) 676.