estudiantes

Carbon 40 (2002) 271–276

Rheological examination of C 60 in low density solutions a, b c Karen Lozano *, Abel Gaspar-Rosas , Enrique V. Barrera a

The University of Texas Pan American, Department of Engineering, 1201 West University Drive, Edinburg, TX 78539, USA b Paar Physica USA, Driftwood, TX 78619, USA c Rice University Department of Mechanical Engineering and Materials Science, Houston, TX 77005 -1892, USA Received 28 June 2000; accepted 28 February 2001

Abstract Fullerene (C 60 ) in solutions of decahydronaphthalene (decalin) and a petroleum solvent viscous standard (PSVS) were studied to understand the rheological properties of fullerene-laden solutions. Although fullerene solubility limits have been published for a variety of solvents, little has been reported on the effect that fullerenes have on flow properties of fluids. In this study, the solvents were studied up to the point of saturation, whereby measurements of solubility, density, viscosity and elasticity were conducted varying the concentration level of C 60 . Rheological measurements based on molecular interactions and on distortion of the flow were studied. No significant elastic contribution from the fullerenes resulted for the solutions below saturation. A pseudoplastic behavior with a lubrication effect imparted by the C 60 molecules was observed in the decalin solutions at concentrations below the saturation level. The PSVS solutions remain Newtonian for all C 60 concentrations while leading to an increase in viscosity.  2002 Elsevier Science Ltd. All rights reserved. Keywords: A. Fullerene; B. Mixing; D. Viscoelasticity

1. Introduction With the recent interest in developing new material systems out of fullerenes and nanotubes, their processability into material systems is becoming a focus for scientific developments [1,2]. Recent research in the processing development of fullerene-containing materials (polymers, metals and ceramics) [3–5] has led to a need for a better understanding of the properties of fullerenes in solutions. Their rheological behavior in solutions will provide for further development of materials systems that rely on ‘wet’ processing. Wet processing involves delivering the fullerenes to preferred sites in a material system by a solution or slurry method [6]. Balta-Calleja et al. processed fullerenes in polypropylene yet the rheological behavior was not reported [7]. While deposition methods have proven to be quite promising, powder metallurgical methods, polymer mixing, and sol–gel processing involving wet fullerenes are also of interest. Powder processing using dry fullerenes has relied on thermal methods of separating fullerenes from each other via sublimation [3,4]. It has been observed that fullerenes can easily be dispersed in powder mixtures *Corresponding author.

using slurry methods where the fullerenes are in solution. Sheng et al. [8] mixed fullerenes in methylnaphthalene with copper powder for composite processing and observed that the slurry method was effective at producing less grain growth than the dry mixing procedure. As the influence of fullerenes on the rheological behavior of fluids is better understood, the effects of fullerene derivatives (functionalized fullerenes and nanotubes) can also be determined more fully. Currently single-walled nanotubes can be purchased in solution yet the rheological properties of these solutions have not been studied. Perhaps as these studies continue it will be seen that the higher order fullerenes (C 120 , C 240 and nanotubes) can be used to study basic rheology of fluids and that more complete understanding will come as they are functionalized and form well understood networks. This paper represents a study that evaluates the rheological behavior of C 60 solutions where decalin and a petroleum solvent for viscous standard (PSVS) were selected. Decalin was selected for its ease of handling compared to other solvents like CS 2 , benzene, etc. and the PSVS was chosen since it is a Newtonian fluid with well-defined stable properties. The various solvents that can be used for fullerene separation (toluene, xylene, decalin, etc.) typically can only hold less than 2 g / l of C 60 in solution before

0008-6223 / 02 / $ – see front matter  2002 Elsevier Science Ltd. All rights reserved. PII: S0008-6223( 01 )00093-8

K. Lozano et al. / Carbon 40 (2002) 271 – 276

272

reaching saturation, therefore, the amount of fullerenes in any given solution may be too low to see changes in the elastic properties of the fluid even though they may influence the viscosity of the fluid. Currently, fullerene separation involves relatively static methods of separating fullerenes but it is likely that future methods may rely on processing requiring significant flow and shear conditions. Therefore, this research will be useful for materials processing, fullerene separation and the basic understanding of the rheological properties of fullerenes in solutions.

2. Experimental section Fullerenes (C 60 ) of 99.95% purity were obtained from Nanotechnologies of Texas, Inc. and used in the as-received condition. The solvents used were laboratory reagent grade decalin from Aldrich which was a mixture of cis and trans and PSVS from Cannon Instrument Company composed of 83% C7–C10 saturated hydrocarbons, 14% C81 aromatics, 0.8% xylene, 0.4% toluene, and 0.2% ethylbenzene, according to the fluid materials datasheet. The solubility level of C 60 in the PSVS was expected to be associated with the xylene, toluene, and ethylbenzene levels as reported by Ruoff et al. [9]. Decalin and PSVS were selected because of low toxicity, rate of evaporation, and aggressiveness of the solvent to the measuring instruments. From Ruoff et al. [9], decalin was also found to provide relatively high fullerene solubility. Fullerenes are soluble in organic solvents, which tend to be flammable and highly toxic therefore, the safe use of the solvents was an important aspect in the selection process [10]. The PSVS was selected because it is used as a viscosity standard for calibration of rheological instruments, therefore, its rheological properties are well known. Table 1 shows the labeled physical properties of the solvents noting that both are selected where their density and viscosity values are close to water. Fullerene-laden solutions were prepared similarly to procedures reported by Ruoff et al. [9] and Sivaraman et al. [11]. The samples were weighed before and after mixing to record possible loss of sample due to vaporization of the solvent, where no evidence of evaporation was found. An excess amount of fullerenes was confirmed by the residue of C 60 left on the bottles after stirring. The saturated solutions were filtered through a 0.45-mm polytetrafluoroethylene filter since fullerene agglomerates tend to be micron in size. Table 1 Labeled values of the physical properties of the selected solvents Solvent

Boiling point (8C)

Density (g / ml 208C)

Viscosity (Pa s 258C)

Decalin PSVS Water

191 153 100

0.8865 0.7912 0.99823

2.415310 23 9.847310 24 8.904310 24

Absorbance measurements were carried out using a 9430 Ultraviolet / Visible (UV/ Vis) spectrophotometer manufactured by IBM Instruments Inc. Absorbance measurements of decalin and PSVS were conducted without fullerenes to assure that the solvents do not absorb strongly in the spectral region of the fullerene, which is reported to be at 328 nm in hexane [9,11,12]. Density measurements were conducted according to the Paar Physica oscillating U-tube technique with a standard deviation of 65310 4 g / cm 3 . The temperature of the solutions was thermostatically controlled within 60.018C. The sample volume in the U-tube was |0.7 cm 3 and the experiments were repeated for each concentration measured to assure accuracy. Rheological examinations of the structural behavior of the samples were conducted in a Paar Physica dynamic capillary rheometer (DCR). Measurements were conducted by relatively low strain controlled oscillation tests [13]. Strain was set up to oscillate between 1 and 10 mm / mm, the frequency was selected at 2 Hz, the capillary used consisted of a 100-mm length and 1 mm diameter. The standard diaphragm (connected to the capillary) made of neoprene was exchanged for a commercial latex type balloon to ensure better resistance to the solvents. The sample volume was |1.5 ml and tests were repeated to evaluate the stability of the testing method, which proved to be of high accuracy. The distortion of the flow was studied in a Paar Physica viscometer AMV200. The selected measuring system was the 1.6 with a capillary diameter of 1.6 mm and a ball diameter of 1.5 mm. The viscometer inclination angles varied between 20 and 758.

3. Results and discussion The UV/ Vis spectra for C 60 in solution with decalin and PSVS are shown in Fig. 1. The UV/ Vis spectra for the decalin and PSVS with zero percent of C 60 and at their saturation level are shown. It can be seen that the C 60 absorbance wavelength is localized at 328 nm for decalin and 330 nm for the PSVS since decalin and PSVS do not have absorbance peaks in this region. As expected from work by Ruoff et al. [9], solubility of fullerene in PSVS was associated with the fullerene–toluene, ethylbenzene and xylene interactions. These solutions tend to hold C 60 to a limited solubility level. The absorbance value for C 70 can be seen at 377 nm [14] and it was corroborated by a solution made of fullerene extract and decalin, which is also shown in Fig. 1. The extract is a mixture of |85% C 60 , and 15% C 70 , with trace amounts of higher order fullerenes. Hexane and the extract in solution with decalin are shown as references for the wavelength peaks. Besides the change in color of the solvents [15], which was observed, absorbance spectra are a clear measure of the presence of fullerenes. The confirmation of absorbance values and their linear increase with increasing C 60 con-

K. Lozano et al. / Carbon 40 (2002) 271 – 276

273

Fig. 1. UV/ Vis spectra for C 60 in solution with decalin and PSVS. Hexane and fullerene extract in solution with decalin are shown as references for the wavelength peaks.

centration (below saturation) provided for selection of sample conditions ranging from no C 60 to saturated states. Fig. 2 showed the linear change in absorbance with fullerene concentration where the maximum solubility values for decalin and PSVS are 1.9 mg / ml and 0.15 mg / ml, respectively. Ruoff et al. [9] reported a solubility value of 2.2 mg / ml for cis-decalin and 1.3 mg / ml for trans-decalin. Fig. 3 shows that the change in density as a function of fullerene concentration was linear up to saturation for both solutions. The density as a function of temperature for the two solvents is shown in Fig. 4. These temperature-dependent results are important in providing detailed information about the solubility of fullerenes in these solutions for slurry processing since saturation conditions should be chosen based on processing temperature to reduce agglomerates from occurring. Fullerenes add to an increase in density as they disperse in solution. Viscosity and elasticity measurements were conducted as a function of strain sweep to identify the linear viscoelastic strain range of the solutions at a constant frequency. Figs. 5 and 6 show the viscous component in the form of the loss modulus vs. strain and the elastic component as the storage modulus vs. strain for decalin and PSVS solutions, respectively. The maximum standard deviation for the measurements of each viscous component was 0.019 for decalin and 0.0128 for PSVS. For the elastic component, the maximum standard deviation was found to be 0.0533 for decalin and 0.0526 for PSVS. In general, the curves describe the solutions as linear viscoelastic fluids. From this behavior it can be inferred that the fullerene molecules readily move with their fluid medium and the

Fig. 2. Absorbance vs. C 60 concentration for (A) decalin and (B) PSVS. Equations show the linear fitting of the results.

interaction of the fullerene with the solvent molecules increases viscosity with no apparent contribution to the moduli as a function of the oscillatory strain measured [13]. The state of dispersion is considered homogeneous where the fullerenes acted as individual molecules and not as aggregates. If the particle distribution was non-uniform then different kinds of interactions would be involved such as a significant elastic contribution [16]. The oscillatory strain sweep showed a dominant behavior of the viscous part, differing from the elastic part by more than one order of magnitude. The structural behavior of the solutions can be considered as lacking in molecular interaction between the fullerenes and the solvents, the fullerenes contributing no inelastic component to the modulus. The weak elastic component at the measured conditions for these solutions can be the result of a low concentration of fullerene in the solvent and proof that the fullerenes are well dispersed in the solvent and behave as individual molecules with a uniform particle distribution.

274

K. Lozano et al. / Carbon 40 (2002) 271 – 276

Fig. 3. Density vs. C 60 percentage of saturation for (A) decalin and (B) PSVS. Equations show the linear fitting of the results.

For powder processing by wet methods, this may be a good result because the lack of elasticity will allow the flow of the solution to cover the metal powders smoothly. The ability of the solvents to wet the powders will provide for a better dispersion of the fullerenes on the powders. The distortion of the flow for decalin solutions can be observed from Fig. 7. The deformations produced by the AMV200 are higher than those analyzed based on molecular interactions (DCR). Therefore it is observed that at higher deformations the viscosity behavior of decalin can resemble a pseudoplastic behavior, which at low deformations was observed to follow a Newtonian behavior. It can also be observed that at concentrations higher than 60% of fullerene saturation a lubrication effect is observed. In the case of the PSVS solutions, the increase in deformation values did not affect the Newtonian behavior. The PSVS solution showed an increase in viscosity with increasing fullerene concentration. Fluids with fullerene concentrations greater than the saturation levels need to be studied, it is expected that the

Fig. 4. Density vs. temperature for (A) decalin and (B) PSVS in which the different C 60 concentrations appear superimposed due to the scale.

pseudo-particle behavior of fullerene agglomerates will dominate the fluid rheological properties. Note that this study investigated the rheological behavior of fullereneladen solutions below saturation and up to a strain condition of 10. It is likely that by going to more extended strain values additional rheological features as those observed in the viscometer for the decalin solutions will be seen. Molecular measurements at higher strain values might show contributions to the elastic modulus given by fullerenes. The C 60 being a monosize particle with 1 nm in size, a dispersion of C 60 into the solvent will give a constant, measurable molecular volume and molecular surface area in the solvent. These contributions will assist in the development of viscoelastic standards for rheological flow studies.

K. Lozano et al. / Carbon 40 (2002) 271 – 276

275

Fig. 6. (A) Loss modulus vs. strain and (B) storage modulus vs. strain for PSVS samples with different concentration levels of C 60 . Fig. 5. (A) Loss modulus vs. strain and (B) storage modulus vs. strain for decalin samples with different concentration levels of C 60 .

4. Conclusions The rheological behavior and density of decalin and PSVS at different concentrations of C 60 was analyzed. Density analyses showed increases in density proportional to fullerene concentration. Rheological studies at relatively low deformations showed that fullerenes in solution increased viscosity, but did not contribute to behavioral alterations in the storage and loss moduli of the fluid. A negligible elastic effect from the fullerenes was obtained over the measured frequency for both solvents. The measurements at relatively high deformations showed a pseudoplastic behavior for the decalin solutions with a fullerene lubricating effect. In the case of the PSVS, a Newtonian behavior with increasing viscosity with fullerene content was observed.

Fig. 7. Dynamic viscosity vs. time for decalin as a function of fullerene concentration below saturation levels.

276

K. Lozano et al. / Carbon 40 (2002) 271 – 276

Acknowledgements This research was funded by the National Science Foundation under grant number DMR-9357505 and by the Texas Higher Education Coordinating Board under grant number 003604-056.

References [1] Files BS, Mayeaux BM. Adv Mater Processes 1999;156:47. ¨ [2] Ajayan PM, Redlick P, Ruhle M. J Microsc 1997;185:275. [3] Barrera EV, Sims J, Callahan DL, Provenzano V, Milliken J, Holtz RL. J Mater Res 1994;9:2662. [4] Barrera EV, Sims J, Callahan DL. J Mater Res 1995;10:366. [5] Holtz RL, Barrera EV, Milliken J, Provenzano V. Mater Res Soc Symp Proc 1994;351:375. [6] Lozano K, Sheng X, Gaspar-Rosas A, Chibante F, Barrera EV. In: Physical examination and handling of wet and dry C60, TMS Publications, 1997, p. 205.

[7] Balta Calleja FJ, Giri L, Asano T, Mieno T, Sakurai A, Ohnuma M et al. J Mater Sci 1996;31:5153. [8] Sheng XY. Processing and characterization of copper–fullerene systems, Rice University, 1998, Ph.D. dissertation. [9] Ruoff RS, Tse DS, Malhotra R, Lorents DC. J Phys Chem 1993;97:3379. [10] Browning E. Toxicity of industrial organic solvents, USA: Chemical Publishing Co, 1953. [11] Sivaraman N, Dhamodaran R, Kaliappan I, Srinivasan TG, Vasudeva Rao PR, Mathews CK. J Org Chem 1992;57:6077. [12] Cataldo F. Spectrochim Acta 1995;3:405. [13] Mezger T. A little course in rheology, Germany: Physica, 1991. [14] Sivaraman N, Dhamodaran R, Kaliappan I, Srinivasan TG, Vasudeva Rao PR, Mathews CK. Fullerene Sci Technol 1994;2:233. [15] Catalan J, Saiz JL, Laynez JL, Jagerovic N, Elguero J. Angew Chem Int Ed Engl 1995;34:105. [16] Jansma JB, Qutubuddin S. J Rheol 1995;39:161.