1. Introduction

The mechanical behavior of nanofluids concerns to attractive applications in tribology [1], microfluidics [2] and biological cell structures [3] among others. Potential optical techniques have been developed to measure fluid properties [4] based on optical resonator spectroscopy [5], interferometers [6] or diffraction [7]. Optical measurements are effective due to their non-invasive capacity, high speed response and small volume driving [8]. The surface tension of a liquid is a key parameter to be considered in its energy transfer potential because it is a direct measure of its wetting behavior. This parameter is the resultant of a strong cohesive force between the liquid molecules and a low adhesive force between the material and the boundary media [9]. Another crucial factor to consider in the stability mechanism of colloids is the dynamic viscosity, which is associated with the mechanical properties of a fluid. Knowing these nanofluid constants, enhancement in mass transfer, thermal conductivity or changes in absorption spectrum can be designed. A material with a remarkable acceptance to modify these properties in a fluid are the carbon nanotubes (CNTs), which are highly resorted in a wide range of fields of research due to their unique qualities, such as thermal conductivity, electrical conductivity, nonlinear optical effects, elasticity and density [10]. However, in order to facilitate the use of CNTs in engineering areas, it is essential their distribution in an easy continuous phase of separation [11]. In this direction, ethanol has been frequently employed as a carrier fluid of nanostructures due to their distinguished evaporation-induced processes. Particularly, multiwall carbon nanotubes (MWCNTs) suspended in ethanol have been intensely implemented in the fabrication of electronic devices in regards to the volatile characteristics of the ethyl alcohol. Outstanding applications of carbon/ethanol based nanofluids on nonlinear optics [12] or optical phase modulators have been presented [13]. By analyzing optomechanical oscillations in nanofluids, static and dynamic properties can be revealed [14]. With this motivation, in this work was analyzed the mechanical deformation induced by light interacting with MWCNTs based ethanol (EtOH-MWCNTs) samples.

2. Methods and materials

An aerosol pyrolysis process was employed for the preparation of the MWCNTs, from a solution containing hydrocarbon and an organometallic precursor. The aerosol was generated ultrasonically and it was supported by an argon flow of 2.5 l min⁻¹ and directed into a quartz tube at 800 °C [15]. The sample was obtained using 30 wt% concentration of MWCNTs.

In order to obtain the surface tension of the MWCNTs nanofluid, an optical measurement system was proposed. The method involves the immersion of a single mode optical fiber into the liquid sample to form a suspended drop at the tip of the fiber. The volume of the drop depends on the mechanical properties. This procedure was comparatively undertaken in three different fluids: distilled water, ethanol and EtOH-MWCNTs. A linearly polarized continuous wave (CW) laser system at 532 nm wavelength was propagated through the fiber. The drop of fluid originated two reflection points into the fiber. In order to approximate the reflection effects in the experiment, a mirror (M₃) was fixed near the end of the drop. The graphical abstract of these considerations is shown in Fig. 1(a). The feedback reflection R₃ was calculated by a three stage Fabry-Perot interferometer [16]:

 

Fig. 1 (a) Three mirrors setup of the suspended drop in the optical fiber, (b) Nonlinear mechano-optical scheme sensing by a Fabry-Perot interferometer.

Download Full Size | PDF

where r_i is the reflection coefficient of the mirror i and the phase length φ_i = 2πL_i/λ, λ is the optical wavelength and L_i is the distance from the mirror i to the mirror i + 1.

The processing route used for the preparation of the nanoparticles (NPs) influences their solubility in a nanofluid. As an example, a specimen of MWCNTs has a different length, diameter and purity in each nanotube [17]; although in similar samples equivalent average sizes may correspond, their collective nonlinear behavior can produce distinct effects. The experimental scheme proposed to explore the mechanical properties of a nanofluid by the assistance of mechano-optical effects integrating a Fabry-Perot interferometer is shown in Fig. 1(b). The procedure is to suspend a drop at the end of the fiber by a wetting process and then a 532 nm beam was propagated through the optical fiber to illuminate the drop sample. A beam splitter allows recording the optical feedback phenomena. The optical back-reflected (BR) signal was returned in the fiber to be acquired by a high speed PIN diode. Moreover, we used a Nd:YAG laser source at 532 nm wavelength, with y-linear polarization, 4 ns pulse duration and energy of 0.3 mJ per pulse to generate the mechanical disturbance. The optical fiber does not maintain the polarization of the beam. The use of polarized light could present limitation points for polarization-selectable conditions of birefringent nanofluids. The nanosecond pulses were focused into the drop sample perpendicular to the direction of the fiber.

3. Results

The MWCNTs were characterized by a transmission electron microscopy (TEM) analysis using a JEOL JEM-220FS. To obtain TEM images, a drop of CNTs dispersed in isopropanol was deposited on a carbon coated grid and allowed to dry in air. The multiwall nature of the nanotubes was clearly demonstrated in the amplification of an isolated tube showed in Fig. 2(a). In order to identify significant information of the optical properties of the samples, reflectance and transmittance spectra were measured by a USB2000 + XR1-ES spectrometer and a DH-2000 light source with 300-900 nm wavelength modulation. In Figs. 2(b) and 2(c) are shown the spectra of the H₂O, ethanol and EtOH-MWCNTs. By the other hand, we obtained the transmission spectrum of the BR signals related to ethanol and EtOH-MWCNTs; the differences in the spectra can be observed from the plots depicted in Fig. 2(d).

 

Fig. 2 (a) TEM micrograph of a CNTs bundle, an inset shows the multiwall nature of an isolated tube. (b) Reflectance and (c) Transmittance spectral of the samples. (d) Transmission spectra difference between EtOH and EtOH-MWCNTs. (e) Calibration measurements of the samples in study. (f) Dynamical evaporation behavior of EtOH and EtOH-MWCNTs.

Download Full Size | PDF

To calibrate the system of Fig. 1(a) a drop of H₂O was used. As an advantage in respect to other experiments in which the wavelength was modulated, we fixed the chromatic variable and controlled the distance between the end of the drop and the M₃ (L₂). The initial mirror distance was considered as 1.32 mm, in which the transmission was maximum. The final path of the mirror was 10 mm. The fitting of the experimental data was carried out by using Eqs. (1) and (2). The drop heights of the samples were recorder as 2 mm for the H₂O, and 1.6 mm for the EtOH and EtOH-MWCNTs. The experimental data and the fitting curves were plotted in Fig. 2(e). The numerical BR curves were shown with lines and the experimental data with markers. Distinction in the slopes of the data related to the BR signal of EtOH and EtOH-MWCNTs does not only concern to the dimensions of the drop but also to the absorbance of the samples. It is worth mentioning that this setup allows us to measure the evolution of the evaporation of the drop of a volatile substance, due to the MWCNTs were contained in ethyl alcohol. The experiment generates a fast and accurate estimation of the height drop in a single wavelength mode. The Fig. 2(f) shows the length of the ethylic drop as a function of time; it can be observed that the MWCNTs sample evaporates faster. The thermal conductivity of the base fluid is responsible for the acceleration process that occurs with the disperse phase of MWCNTs [18]. The incident wave generates an increment in the temperature of the pendant drop, facilitating the evaporation. The abrupt process originated after 50 seconds is due to a temperature increment derived from molecular hydrogen bonding of the air with the ethanol [19]. Taking into account these considerations, both fluids are stable for less than 50 seconds, opening the possibility to measure the mechano-optical response at low frequencies.

The calculation of the surface tension and viscosity was based on the induction of a mechanical disturbance resulting from a single high energy fast pulse of the Nd:YAG source. To compute the total elongation that happens in the pendant drop by the impulse perturbation, we analyzed the form of the transmitted optical signal by processing the image. The interferometer gives a circular pattern with constant dimensions after going through the drop and reflected into the M₃, which was been acquired by a digital camera of 720p resolution and 240 fps. The Fig. 3(a) shows the gradient image analysis of the calibration in which the width and the height was computed in 2.2 × 2.2 mm. The fluids presented a maximum deformation in the direction of the mechano-optical pulse that was assigned to the height, and a contraction is promoted in the width. These elongations and contractions were: H₂O, 3.7 mm and 2.08 mm; EtOH, 3.45 mm and 2.17 mm; EtOH-MWCNTs 3.1 mm and 2.2 mm. The Figs. 3(b)-3(d) illustrates the distorted gradient image of the H₂O, ethanol, and EtOH-MWCNTs, respectively when the optical pulse was induced. The right axis indicates the optical intensity of the CW beam through the disturbed fluids, and the color bar the magnitude of the gradient.

 

Fig. 3 Image related to the interferometric patterns perturbed by a nanosecond 532 nm pulse (a) calibration without fluid, (b) distilled water, (c) EtOH, (d) MWCNTs nanofluid.

Download Full Size | PDF

The results showed in Fig. 3 were used to calibrate the initial conditions of the mechanical behavior in each sample. We observed that the elongation occurs in the direction of the pulsed laser (y-axis) and it was conserved. The energy released affects the orthogonal axis as an inelastic behavior. Since this process is preserved until it is extinguished, a specific underdamping response is linked to each fluid. This principle was achieved because the mirror 2 in the Fabry-Perot interferometer was micro-modulated by the pump laser. The feedback signal was asserted by the Voigt model, with initial condition cero and an impulse input δ(t). The temporal response was obtained by the second-order ordinary differential equation [20]:

where m is the mass of the drop, η’ is a constant related to the dynamic viscosity, γ’ is a constant linked with the surface tension, x is the uniaxial strain (negative x-axis and positive y-axis), t is the time, n is the grade of the inelastic nonlinearity and f is the force exerted by the pump beam under the drop. Analogously to a mass-spring-damper system, the pending drop is joined by its molecules to the optical fiber, in which the force of deformation of the drop is proportional to the surface tension and the force velocity area to the dynamic viscosity: the height of the drop is proportional to the surface tension, while the change of the deformation in time is inversely proportional to the viscosity. We estimated the masses of the drops with the Young-Laplace equation [21] and the results were: m_H2O = 3.04 × 10⁻⁵kg, m_EtOH = 1.91 × 10⁻⁵kg, and m_MWCNTs = 5.1 × 10⁻⁵kg. The plots of the mechano-optical time response of the samples are shown in Fig. 4.

 

Fig. 4 Dynamical mechano-optical behaviors in the studied samples (a) H₂O, (b) EtOH, (c) EtOH-MWCNTs.

Download Full Size | PDF

Table 1 shows the mechanical properties of the three studied samples. The values of the H₂O and EtOH were taken from tables. By the other hand, the viscosity of the MWCNTs nanofluid was computed by the Brinkman equation [22] and the surface tension was taken from ref [9]. The values η’ and γ’ are the constants which fit the experimental data. Regarding that these variables had information of the fluid properties, they keep a close correlation ratio. This allowed us to estimate the properties of the EtOH-MWCNTs fluid with the next relations.

Table 1. Viscosity and surface tension of the studied samples.

View Table

where α is the fluid to compute and β is the fluid to compare. We acquired two values for each measure η_MWCNTs = 2.727 × 10⁻² P, γ_MWCNTs = 2.515 × 10⁻² N/m, η_MWCNTs = 2.7 × 10⁻² P, γ_MWCNTs = 2.492 × 10⁻² N/m, the first compared with H₂O and the second compared with EtOH. The error bar of the dynamic viscosity was around 6.4% and for the surface tension 1.4%. The low reflectance and transmittance of MWCNTs-EtOH around the 532 nm wavelength employed for the interferometric experiments is not an issue for the evaluations; this is because of the high sensitivity of the device that only takes a small sample in the drop tested. This method can be used for both single- and multi-wall CNTs; changes in droplets conditions to distinguish these samples could be derived from their characteristic viscosity, surface tension and thermal conductivity. These results pointed out potential applications for rheological instrumentation to detect variations in viscosity and surface tension in nanofluids without knowing the size or volume fraction of the dispersed NPs. Nanosecond interactions can be considered as an impulse input without exchange of matter in mechanical deformation.

4. Conclusions

Within this work were described surface tension and dynamic viscosity properties exhibited by MWCNTs incorporated to an ethanol suspension by mechano-optical measurements. The influence of the interferometric nature of light was analyzed to develop a non-contact technique with immediate applications for exploring mechanical properties in nanofluids. Modification in mechanical actions for nanophotonic sensors can be considered by governing the dynamic evolution of liquid samples with nanostructures.