In order to obtain a stable nanofluid, several water-based nanofluids were analysed and various parameters were investigated: different preparation methods, various kinds of dispersants varying both the concentration of the nanoparticles and of the dispersants. As already described, for each nanofluid, the mean size value was obtained, repeating the measurements almost every day for 30 days, both for the nanofluid stored in static mode and for the same nanofluid after mechanical shaking. Moreover, the Zeta potential measurements and the suspensions visual observation were used for analysing the nanofluid's stability.
Comparison between different dispersion techniques
Initially, some tests on 0.1 wt.% solutions of TiO2, CuO and SWCNHs in water were performed, comparing the three different dispersion techniques without dispersants.
Ball milling method
Table shows nanoparticles' mean diameters at different days from their dispersions by different methods. Only 2 days are presented for the ball milling because, only after 4 days for TiO2 nanofluids, the nanoparticles got completely precipitated.
Nanoparticles mean diameters at three different days from their preparation by three different methods, by DLS measurements on static sample
The mean particle size obtained by ball milling was over the nanometric range (day 1). These nanofluids turned out to be unstable. In fact, from the first to the last day of measurement, the mean diameter decreased since at the constant height from the base of the cell, where the average diameter was measured, only the smaller particles remained in suspension and therefore could be detected, while the bigger ones got precipitated at the bottom of the cell. After 14 and 4 days, respectively, for CuO and TiO2 nanofluids, the nanoparticles, as highlighted by visual inspection, got completely precipitated and the concentration of the particles in suspension was too low to allow the measurements using the nanosizer.
Moreover, the Zeta potential was around +10 mV for CuO-water nanofluid and around 0 mV for TiO2-water nanofluid. These low values are typical of unstable solutions.
Considering the poor results obtained for the suspensions prepared by the ball milling process, this method was no longer tested, and other techniques were preferred.
The mean diameter of CuO, TiO2 and SWCNH nanoparticles dispersed in water by sonication method are presented in Table , at days 1, 4 and 15. This method proved to be more effective than the ball milling method in reducing aggregates. However, in terms of stability, for CuO nanoparticles, the results are similar to those obtained by ball milling method, since they could not be measured after 15 days, because of particle precipitation, as highlighted by visual observation. Also in TiO2-water nanofluid, a precipitation occurred, even if being slower than with ball milling, as shown in Figure which presents the nanoparticles' size distributions for water containing TiO2 at days 1, 4 and 15.
Nanoparticles size distribution for water containing 0.1 wt.% TiO2 dispersed by means of the sonication method. At (thick line) day 1, (dashed line) day 4 and (dashed-dotted line) day 15.
In SWCNHs-water nanofluid, a stable population with a 100-nm average diameter was observed, although with the presence of larger particles, with a mean diameter of approximately 4 μm, according to DLS measurements, which disappeared after 24 days, probably because of settling down.
The measured Zeta potentials were approximately +10, +50 and +35 mV for CuO, TiO2 and SWCNHs water-based nanofluids, respectively. Owing to the strong opacity of the SWCNHs nanofluid, it was necessary to dilute that suspension to perform the Zeta potential measurements. Considering the strong instability of TiO2 nanoparticles, the value obtained is in disagreement with the empirical limit of |30| mV, over which a nanofluid should remain stable.
The mean diameters of CuO, TiO2 and SWCNH nanoparticles in water, dispersed by the homogenization method, are presented in Table , which shows the differences in them at days 1, 5 and 15 after preparation. The CuO-based fluid shows aggregates having mean diameters of 1 μm or more and precipitation in 8 days, as highlighted also by the visual inspection (Figure ).
The CuO-water nanofluid, showing precipitation just after 8 days.
In TiO2-water nanofluid, all the aggregates observed on the first day precipitated after 21 days, as measured by DLS, while the other nanoparticles tended to settle down.
The SWCNHs nanofluid turned out to be quite stable. In fact, the mean size measured by DLS the first day was almost constant for 33 days, as shown by Figure . However, from day 5, a micrometric aggregate was found, indicating a partial instability of the solution. Moreover, the mean particle size in water was slightly higher than the size measured in the powder.
Nanoparticles size distribution for water containing 0.1 wt% SWCNH, dispersed by means of the homogenization method.without dispersant. At (thick line) day 1 and (dashed line) day 33.
The Zeta potentials for the CuO and TiO2 nanofluids were approximately +10 and +35 mV, respectively, while for the SWCNHs-water nanofluid, it was not possible to obtain a stable value, even after diluting the suspension.
Therefore, the homogenization process proved to be the most effective method for preparing nanofluids. However, these preliminary results pointed out that the precipitation of the CuO nanoparticles was evident even after a few days with any of the three analysed methods. For this reason, this nanofluid was no longer investigated.
At this point, in order to improve the stability of TiO2 and SWCNHs nanofluids, different dispersants were tested.
Use of dispersants and acidification of the solutions
All the fluids discussed in this section were prepared with the high-pressure homogenization method, considering its superiority over the other methods. Table shows nanoparticles' mean diameters and standard deviations at different days from their dispersion.
Nanoparticles mean diameters and standard deviations at different days from their preparation by means of the homogenization method
Initially, two acidic solutions having pH 4-5 prepared with citric acid or hydrochloric acid were tested for the titanium dioxide-water nanofluid. In view of the potential use of these nanofluids in, e.g. hydraulic circuits, lower pH values were not considered. However, these acids were ineffective in producing stable suspensions at these pH values, since the particle precipitation was visually evident.
Therefore, a non-ionic dispersant, PEG 600, was investigated, based on [13
]. Various concentrations of PEG and TiO2
were measured. The variation along time of TiO2
-PEG nanoparticle mean diameters, with TiO2
at 0.01, 0.1 and 1 wt.% and PEG at 0.02, 0.2 and 2 wt.%, respectively, are shown in Figure .
Figure 5 Nanoparticles mean diameter. Diameter in relation to the time elapsed from the day of preparation, for water containing (a) 0.01 wt.% TiO2 + 0.02 wt.% PEG: (filled square) static, (open square) shaken; (b) 0.1 wt.% TiO2 + 0.2 wt.% PEG: (filled triangle) (more ...)
The first nanofluid (at TiO2 concentration of about 0.01 wt.%) became unstable, i.e. just after 5 days, an aggregation occurred, and after 18 days, all the nanoparticles settled down (as gathered by visual observation). The irregular trend shown in the figure is probably due to the instability of the suspension.
On the contrary, the other samples were quite stable. In the case of static solutions, the mean size slightly decreased to around 70 nm after a few days and then it remained stable, indicating only a partial precipitation. However, after a simple mechanical shaking a mean particle size of approximately 130 nm was repeatedly recovered, suggesting the absence of further aggregation phenomena. This result is of interest because it suggests a possible application in devices where the fluids are frequently or continuously stirred, e.g. in plants with forced circulation. All the measurements provided average diameters higher than the 21 nm of the base powder, but the aggregates grew just after preparation, keeping nanometric and constant dimensions even after 30 days. In order to highlight this behaviour, Figure represents the nanoparticle size distribution for water-TiO2 at 0.1 and 0.2 wt.% PEG. After 30 days, while the static sample shows a smaller average diameter than at the first day, the shaken nanofluid gives the same value, i.e. no further aggregation was detected.
Nanoparticles' size distribution for water containing 0.1 wt.% TiO2 + 0.2 wt.% PEG. At (thick line) day 1, (dashed line) day 30 for static and day 30 for shaken (dashed-dotted line).
The measured Zeta potential was +40 mV for the nanofluids containing 1 and 0.1 wt.% TiO2, supporting their non-aggregating tendency, while the values obtained for the 0.01 wt.% TiO2 fluid were not stable. The PEG:TiO2 = 2:1 ratio turned out to be effective, but further research is needed to optimize nanoparticle and dispersant concentration as a function of their application.
SWCNHs-water nanofluids with SDS as dispersant were tested in several concentrations. An anionic dispersant was chosen based on [14
]. The investigated fluids were
• water +0.01, 0.1 and 1 wt.% SDS at 0.01, 0.1 and 1 wt.% SWCNHs, respectively;
• water +0.01 wt.% SWCNHs +0.03 wt.% SDS.
Figure represents the mean particle diameters as a function of time for the nanofluid in static mode and for the same nanofluid after mechanical shaking.
Figure 7 Nanoparticles' mean diameter. Diameter in relation to the time elapsed from the day of preparation, for water containing (a) 0.01 wt.% SWCNHs + 0.01 wt.% SDS: (filled circle) static, (open circle) shaken; 0.01 wt.% SWCNHs + 0.03 wt.% SDS: (filled triangle) (more ...)
Water-SWCNHs containing 0.01 wt.% SDS formed aggregates, which are visible in Figure in the upper curve relative to the shaken nanofluid. In order to improve the stability of this suspension, a higher SDS:SWCNHs ratio was tested. The result is shown in the same figure with triangle, where the suspension containing 0.01 wt.% of SWCNHs and 0.03 wt.% of SDS showed a very stable behaviour for 39 days, keeping a mean diameter of about 120 nm.
Water-SWCNHs containing 0.1 wt.% SDS (Figure ) shows a constant diameter around 100 nm, i.e. a value very similar to the one measured for the powder, for both the static and stirred sample even after 25 days, suggesting a good stability of the fluid.
Analogous behaviour was shown by water-SWCNHs containing 1 wt.% SDS (Figure ), though the mean diameter of nanoparticles was about 180 nm.
The measured Zeta potential was around -40 mV, negative as expected in the case of anionic dispersant [6
], for all the studied SWCNHs-nanofluids, supporting their non-aggregating tendency. Owing to the strong opacity of the solutions at 0.1 and 1 wt.%, they were diluted to perform the Zeta potential measurements.
In conclusion, the water-based nanofluids containing SWCNHs and SDS proved to be very stable and further investigation on their properties is underway.