Thermal Properties of TiO₂ Nanoparticle-Treated Transformer Oil and Coconut Oil

Abstract

This study investigates the impact of incorporating TiO₂ nanoparticles into two types of oils at different temperatures and with varying volume fractions: transformer oil (NYTRO LIBRA) and virgin coconut oil (manufactured by Govi Aruna Pvt. Ltd., Gampaha, Sri Lanka). The nanofluids were prepared using a two-step method by adding CTAB (cetyltrimethylammonium bromide) surfactant. To minimize nanoparticle agglomeration, this study employed relatively low-volume fractions. Thermal properties by means of thermal conductivity, thermal diffusivity, and volumetric heat capacity were measured in accordance with ASTM (American Society for Testing and Materials) standard methods using a multifunctional thermal conductivity meter (LAMBDA thermal conductivity meter). The measured thermal conductivity values were compared with theoretical models and previous research findings. It was confirmed that the modification of thermal properties was enhanced by doping TiO₂ nanoparticles with different volume fractions.

1. Introduction

‘Nanofluid’ has been a developing topic in the engineering and scientific community since the beginning of the 21st century for various applications in engineering fields ever since its initial introduction by Choi in 1995 [1,2]. The term nanofluid is used to describe a solid–liquid mixture containing nanoscale particles with an average size of less than 100 nm with any kind of a base fluid.

Numerous productive research efforts have been carried out using nanoparticles, particularly in the context of thermal conductivity across various fields. These nanoparticle-based fluids, termed nanofluids, have various applications in industrial areas. A rapid increase in the amount of research efforts related to nanofluids over the past two decades can be observed. This literature has investigated several factors that influence the thermal properties of nanofluids, such as nanoparticle concentration, particle size, particle shape, the thermal resistivity of interfacial layers, and Brownian motion [3,4,5].

It has been reported that nanofluids can be applied in various industrial cooling applications, such as electrical power systems, electronic cooling applications, biomedical systems, and the automobile industry [6,7,8]. Apart from experimental works, several studies have been carried out to model the thermal conductivity of nanofluids, considering both macroscopic and microscopic properties. However, widening the scope of empirical studies to apply nanofluids in real-time applications and build theoretical models is essential for advancing future heat transfer studies.

Researchers have explained numerous mechanisms related to nanofluids that explain the enhancement in the thermal properties of nanofluids, such as Brownian motion, nanoparticle clustering, and thermal liquid layering. Brownian motion is the random movement of nanoparticles within the fluid. This movement helps to improve the thermal conductivity of the nanofluid through thermal contact between the base fluid and the particles. Nanoparticle clustering is the aggregation of the nanoparticles that form thermal bridges within the fluid to enhance thermal conductivity. However, even though this clustering helps to improve thermal conductivity, it may reduce the stability of some nanofluids while lowering the positive effect that can be created by the Brownian motion. Thermal liquid layering is the formation of liquid molecules around the nanoparticles. This layer helps to transfer the heat between the solid and liquid phases of the nanofluid. However, the properties of this layer are different between nanofluid types. According to researchers, these mechanisms are yet to be studied further, and in the future, there will be other mechanisms to explain the behavior of the nanofluids. There are several thermal conductivity models that have been proposed, including the effect of these mechanisms. However, these models are not often used in research since these microscopic properties are complicated to analyze compared to macroscopic properties.

Nanofluids are generally developed by doping nanoparticles such as metal nanoparticles (Au, Ag, Cu), metal oxides (TiO₂, Al₂O₃, CuO, Fe₂O₃), and carbon-based nanoparticles (Graphene, Fullerene, Carbon Nano Tubes) into liquids such as water, mineral oils, vegetable oils, and synthetic and natural esters. Surfactants are used to improve the nanoparticle suspension in liquids where necessary. Consequently, these efforts have enhanced the dielectric and thermal properties of base fluids. Some of the related productive research works carried out to enhance the thermal properties of fluids are mentioned below.

Sohel et al. [9] discovered the thermal properties (such as thermal conductivity, thermal diffusivity, and specific heat capacity) of Al, Al₂O₃, and TiO₂ nanoparticles suspended in ethyl glycol and engine oils and compared their results with several theoretical implementations concerning different volume fractions. Another study validated the advantage of using Al₂O₃ and Fe₂O₃ nanoparticles in engine oils for different mass fractions, considering temperature as another variable. Their primary measurement was thermal conductivity; the measurement method was the transient hot-wire method [10]. Eric et al. [11] investigated the thermal performance of Al₂O₃ and Al₂O₃–Fe hybrid-based nanofluids by comparing theoretical models of dynamic viscosity and thermal conductivity. The CeO₂–water nanofluid also enhanced thermal conductivity and dynamic viscosity compared to pure water [12].

Sunil et al [13]. dispersed CaO nanoparticles in rice bran oil, both of which are rarely used nanoparticle and base oil types compared to other research studies. Both materials also have biodegradable quality. This effort also enhanced the thermal properties of base oils by a recognizable level. Using exfoliated hexagonal boron nitride (h-BN) nanoparticles with vegetable oil is another successful effort to develop a biodegradable nanofluid [14]. Zbigniew et al. [15] used TiO₂ and C₆₀ nanoparticles with natural ester as an electro-insulative liquid in power transformers. Replacing petroleum-based products with bio-based solutions is becoming a promising research area in the effort to move towards green technology.

The thermophysical properties of carbon-based nanofluids such as graphene, amorphous graphene, graphene quantum dots, controlled reduced graphene oxide, carbon nanotubes, and fullerene are critically discussed in several research papers that consider their unique characteristics compared to other metallics and metal oxide nanoparticles [16]. Almeida et al. [17] dispersed graphene nanoparticle transformer oils and investigated thermophysical properties such as kinematic viscosity, surface tension, density, and specific resistance. It has been reported that amorphous graphene is also a good candidate to enhance the electrical and thermophysical properties of transformer oils due to their interaction between nanosheets and the clustering of nanostructures [18]. Zhang et al. [19] dispersed controlled reduced graphene oxide in water for different concentrations, which also represented the advantage of applying them in industrial cooling applications. Primarily, this study demonstrated that controlled reduced graphene oxide can enhance the thermal conductivity of water more than other graphene nanoparticles. Amiri et al. [20] showed that amine-treated graphene dots can produce enormously stable nanofluids with superior thermal conductivity and higher breakdown voltage, confirming their potential.

Researchers have developed several theoretical models to predict the thermal conductivity of nanofluids throughout the past few decades. Maxwell’s model was the first thermal conductivity empirical model [21]. Most researchers have followed Maxwell’s model and developed some advanced models, such as the Hamilton crosser model and Bruggeman model. However, recent research articles show that developing a common thermal conductivity model for all nanofluids is challenging, considering microscopic parameters such as nanolayer thermal conductivity, nanolayer thickness, particle size, particle shape, and particle interaction between different materials. As a result, researchers started to develop specific models for specific nanoparticles and base fluids.

In this research work, TiO₂ (titanium dioxide) was selected as the nanoparticle, and the base fluids used were transformer oil and coconut oil. Several research studies have verified that TiO₂ can enhance the electro-insulation properties of transformer oils [22,23]. Even though some previous studies combined TiO₂ and transformer oils, there are very few research articles for TiO₂-based coconut oils to the best of our knowledge [24,25]. This study compares the thermal conductivity, thermal diffusivity, and volumetric heat capacity of TiO₂-based nanofluids. In addition, the experimental values of thermal conductivity are compared with three theoretical models, namely the Maxwell model [21], the Maxwell–Garnett model [21], and the Pak and Cho model [26], as well as alongside several other experimental studies.

2. Materials and Methods

2.1. Materials

TiO₂ anatase-type nanopowder was purchased from Ningbo Jiweina New Material Technology Co., Ltd., Beijing, China. Two base fluids were used to prepare nanofluids. The transformer oil was NYTRO LIBRA [27,28], and the organic virgin coconut oil was manufactured by Govi Aruna (Pvt.) Ltd. [29,30]. The properties of the TiO₂ nanoparticles are provided in

Table 1. Properties and characteristics of nanoparticles.

Nanoparticle Type	TiO2 Anatase
Purity	99.5%
Color	White powder
Particle size	3–5 nm
BET value	150–200 m2/g
Density	3.89 g/cm3
Phase	Anatase

, and the SEM image is shown in Figure 1.

Table 2. Properties and characteristics of base fluids.

Base Oil Type	Transformer Oil	Coconut Oil
Color	Transparent	Transparent
Density	843.74 kg/m3	880.40 kg/m3
Thermal Conductivity at 30 °C	108.12 mW/K	156.21 mW/K

represents the details of the two base fluids. Figure 2 represents the appearances of the nanoparticle and the two base fluids.

2.2. Preparation of Nanofluids

A two-step preparation method was used to utilize the base fluids with the nanoparticles. It was decided that the TiO₂ nanoparticles be dispersed in the base fluid with different volume fractions from 0.002 vol.% to 0.012 vol.%.

$$Volume Percentage=\frac{V_{np}}{V_{np}+V_{bf}} \times 100\%=\frac{\raisebox{1ex}{$m_{np}$}\!\left/ \!\raisebox{-1ex}{${\rho }_{np}$}\right.}{\raisebox{1ex}{$m_{np}$}\!\left/ \!\raisebox{-1ex}{${\rho }_{np}$}\right.+\raisebox{1ex}{$m_{np}$}\!\left/ \!\raisebox{-1ex}{${\rho }_{np}$}\right.} \times 100$$

(1)

The volume percentage of the nanofluid was calculated from Equation (1) using the data from Table 1 and Table 2. The weights were measured with an accuracy of 0.1 mg using a BSA224S-CW precision balance, Sartorius AG, Goettingen, Germany. After measuring the weights, the base oil beaker was placed on a magnetic stirrer and the temperature increased to 60 °C. The nanopowder was then added to the oil with CTAB (cetyl trimethyl ammonium bromide) at a 1:0.5 ratio. Magnetic stirring was carried out for 1 h at a temperature of 60 °C at 500 rpm. The nanofluid samples were then sonicated for 4 h in the bath-type ultrasonicator at 60 °C to eliminate particle agglomeration. The physical appearance of base oils, nanoparticles, surfactants, and nanofluids is shown in Figure 3. Each prepared sample was left undistributed and observed by visual inspection for five days.

2.3. Thermal Conductivity Measurement

A LAMBDA Thermal Conductivity Meter is shown in Figure 4, was used to measure thermal conductivity, thermal diffusivity, and volumetric heat capacity according to ASTM D7896-19 standardization [31]. This instrument uses a stationary hot-wire resistance method to measure the liquids’ thermal conductivity and diffusivity. The required liquid volume was 40 mL, and the selected temperature range was 40 to 120 °C. The selected range is significantly broader compared to previous studies. This instrument is capable of measuring values from 10 to 2000 mWm⁻¹ K⁻¹. A PT-100 temperature sensor was included in the sample cup, which had ±0.1 °C accuracy. A dry block calibrator from the same manufacturer was used to reduce the vibrations and maintain a constant temperature during measurement.

Table 3. Parameters of the LAMBDA Thermal Conductivity Meter.

Parameter	Value
Measuring Range	10 to 2000 mW/(mK)
Reproducibility	±1%
Temperature Range	−50 to 300 °C
Temperature Accuracy	±0.1 °C

represents the parameters of the LAMBDA Thermal Conductivity Meter.

The functional core of the stationary hot-wire method is a long cylindrical heat source. The theory of hot-wire instruments produces a constant heat stream input, q, into the surrounding liquid. This will cause a time-dependent temperature field in the liquid, as represented. The temperature rise is influenced by the changing electrical resistance of the hot wire. The following equation can be used to measure the thermal conductivity of the liquid.

$$\lambda =\frac{q}{4\pi ({\vartheta }_1-{\vartheta }_{2)}}ln\left(\frac{t_2}{t_1}\right)$$

(2)

where λ is thermal conductivity (Wm⁻¹ K⁻¹), q is a constant heat stream (Wm⁻¹), and ϑ is temperature at time t.

When a constant voltage is supplied to the circuit, the platinum wire’s electric resistance rises with the wire’s temperature. Measured data for rising temperature are linear against logarithmic time intervals. The thermal conductivity is calculated from the slope of the rise in the wire’s temperature against the logarithmic time interval obtained by Equation (1).

Temperature regulation of the instruments is ensured by a standard PI (proportional and integral) controller with pre-calibrated controller constant values.

2.4. Thermal Diffusivity Measurement

Thermal diffusivity is a measurable thermal property according to the ASTM D7896-19 standard method and a critical factor in heat transfer fluids [31]. Thermal diffusivity measures how quickly the temperature will change when it is cooled and heated. It is the thermal inertia of a material. Thermal diffusivity and kinematic viscosity both play similar roles in heat transfer.

$$\mathrm{Thermal} \mathrm{Diffusivity}, {\mathrm{a}}_{\mathrm{T}}=\frac{K}{\rho C_p}$$

(3)

Here, K is the thermal conductivity, $\rho$ is the density of the liquid, and C_p is heat capacity at a constant pressure.

2.5. Volumetric Heat Capacity Calculation

Volumetric heat capacity can be calculated using thermal conductivity and thermal diffusivity.

$$C_{vol}=\frac{\lambda }{a_T}$$

(4)

Here, λ is the thermal conductivity and a_T is the thermal diffusivity. This study calculated volumetric heat capacity using the measured values of thermal conductivity and thermal diffusivity.

3. Results and Discussion

3.1. Stability of Nanofluid with Time

The stability of the transformer and coconut oil with TiO₂ nanoparticles was analyzed over five days. Both oil samples with 0.002 vol.% concentration displayed the highest stability compared to other higher volume concentrations. On day 4, a thick layer of TiO₂ in the 0.012 vol.% samples became observable, with other samples exhibiting the same kind of layer on day 5 as it is shown in Figure 5.

Highly concentrated nanofluids have more nanoparticles that can increase the chance of colliding with each other than lower-concentrated nanofluids. Since these higher-concentrated nanofluids tend to create larger aggregates within the fluids than lower-concentrated nanofluids, this process is called particle agglomeration. This happens due to particle–particle interaction that tends to create larger aggregates, making higher-concentrated nanofluids less stable than lower-concentrated nanofluids. This phenomenon can be easily observed in oil-based nanofluids rather than water or ethylene-glycol-based nanofluids because water and ethylene glycol base fluids are polarized materials, such as TiO₂ nanoparticles, that can create strong bonds with the solid particles.

3.2. Effect of Temperature on Thermal Conductivity

Variation in the thermal conductivity of nanofluid as a function of temperature at different volumetric fractions is shown in Figure 6. The thermal conductivity of base oils and nanofluids was evaluated over the temperature range of 40 to 120 °C. The findings indicate that coconut oil has significantly higher thermal conductivity than transformer oil, making it a potential alternative for future heat transfer applications. Both oils exhibit a decrease in thermal conductivity with increasing temperature, which is consistent with most oil types [32].

Even though there is some previous research related to TiO₂-based transformer oils, further research is needed on TiO₂-based coconut oils. It can be observed that the impact of nanoparticles on increasing thermal conductivity has a more significant effect on transformer oils rather than coconut oils. Furthermore, TiO₂ nanoparticles have a more substantial effect on enhancing the thermal conductivity of water and ethylene glycol fluids [33,34]. This observation can be justified because water or ethylene glycol base fluids can allow nanoparticles to perform Brownian motion throughout the nanofluid, which is a critical factor in increasing the thermal conductivity of nanofluids.

3.3. Effect of NanoparticleVolume Fraction on Thermal Conductivity

Variation in the thermal conductivity of nanofluid as a function of different volume fractions at different temperatures is shown in Figure 7. As the results indicate, thermal conductivity decreases with increasing temperature. However, increments in thermal conductivity alongside nanoparticle volume fractions can be discovered. Adding nanoparticles to the oil also alters the behavior of the base oil in terms of thermal conductivity, which leads to an increasing trend for thermal conductivity at all concentrations under any temperature conditions.

3.4. Comparison of Thermal Conductivity Enhancement with Experimental Values from Theoretical Model Values at 40 °C

According to the experimental results as given in

Table 4. Comparison of measured values of thermal conductivity ratios with the values predicted by different thermal conductivity ratio models for transformer oil-based nanofluids at 40 °C.

Nanoparticle Volume Fraction	0.002	0.004	0.008	0.012
Experimental Values	1.009	1.029	1.034	1.042
Maxwell Model (% diff)	1.003 (0.59%)	1.007 (2.14%)	1.014 (1.93%)	1.021 (2.01%)
Maxwell–Garnett Model (% diff)	1.004 (0.50%)	1.007 (2.14%)	1.014 (1.93%)	1.022 (2.01%)
Pak and Cho Model (% diff)	1.015 (0.59%)	1.030 (0.10%)	1.060 (2.51%)	1.090 (4.61%)

and

Table 5. Comparison of measured values of thermal conductivity ratios with the values predicted by different thermal conductivity ratio models for coconut oil-based nanofluids at 40 °C.

Nanoparticle Volume Fraction	0.002	0.004	0.008	0.012
Experimental Values	1.007	1.007	1.010	1.014
Maxwell Model (% diff)	1.003 (0.40%)	1.006 (0.10%)	1.012 (0.20%)	1.018 (0.39%)
Maxwell–Garnett Model (% diff)	1.003 (0.40%)	1.007 (0.00%)	1.009 (0.01%)	1.018 (0.39%)
Pak and Cho Model (% diff)	1.015 (0.79%)	1.030 (2.28%)	1.060 (4.85%)	1.090 (7.50%)

, TiO₂ nanoparticles considerably enhance the thermal conductivity of transformer oils more than coconut oils, which may be attributed to the higher viscosity of coconut oils.

When comparing the different thermal conductivity models, it can be observed that the Maxwell–Garnett model is more suitable for predicting TiO₂–oil-based nanofluids, especially in the case of coconut oil-based nanofluids.

3.5. Comparison of Thermal Conductivity Results with Previous Research on TiO₂-Nanofluids

When comparing previous research results with oil-based nanofluids, it can be observed that the contribution of TiO₂ nanoparticles is much less compared to water or ethylene glycol-based nanofluids. This phenomenon might be caused by nanoparticles’ difficulty in performing Brownian motion in oils rather than fluids like water and ethylene-glycol according to

Table 6. Thermal conductivity results of previous TiO₂-based nanofluids.

Nanoparticle Volume Fraction (%)	Base Fluid	Nanoparticle	Concentration	Maximum Enhancement
[35]	Diathermal oil	TiO2—Anatase	0.2–1.0 vol.%	7.1%
[15]	Synthetic ester oil	TiO2	0.82 g·L−1	3.2%
[36]	Synthetic ester oil	TiO2—Rutile	0.005–0.05 vol.%	8.3%
[37]	SAE 15W40 engine oil	TiO2	0.1–1.0 wt. %	4.54%
[38]	Ethylene glycol	TiO2	1.0–7.0 vol.%	19.52%
[39]	BioGlycol:water (20:80) mixture	TiO2	0.5–2.0 vol.%	12.6%
[40]	Water:ethylene glycol (60:40) mixture	TiO2	0.5–1.5 vol.%	15.35%
Present Study	Transformer oil	TiO2—Anatase	0.002–0.012 vol.%	4.2%
	Coconut oil	TiO2—Anatase	0.002–0.012 vol.%	1.4%

.

The TiO₂ nanoparticles are more likely to be dispersed in liquids like water and ethylene glycol because of the common polarized properties of the particles and the liquids. This strong dispersion within these types of fluids does not create aggregates. Since there is no agglomeration of particles, individual nanoparticles move freely and randomly within these liquids (Brownian motion) more so than they do in oils. This phenomenon causes more collisions with the liquid molecules. This behavior of the particles helps the migration of nanoparticles from warmer to colder regions and makes for more effective heat transfer within the fluid.

3.6. Effect of Temperature and Volume Fraction on Thermal Diffusivity

According to Figure 8, it can be identified that the dispersed nanoparticles have no significant impact on both base oils. This may be partly due to the slight variation in thermal conductivity since the specific heat capacity has a negligible variation due to there being no phase change in the material. However, it can be identified that the thermal diffusivity of coconut oil is slightly higher than that of transformer oil, which again confirms that coconut oil can be identified as a potential candidate for industrial cooling applications. Also, the thermal diffusivity of both oil types decreases with temperature.

3.7. Effect of Temperature and Volume Fraction on Volumetric Heat Capacity

The volumetric heat capacity of a nanofluid can be identified as follows:

$$\mathrm{Volumetric} \mathrm{Heat} \mathrm{Capacity}={\rho }_{nf}{c}_{nf}={\varphi }_{np}{\rho }_{np}{c}_{np}+{\varphi }_{bf}{\rho }_{bf}{c}_{bf}$$

(5)

Here, ϕ is the volume ratio of nanoparticles and the base fluid, ρ is the density, c is the specific heat capacity, and a substance’s specific volumetric heat capacity is given by ρc.

Volumetric heat capacity is the amount of heat required to change the temperature of a unit volume of the substance by one Kelvin, and ρ is the density or mass per unit volume. Volumetric heat capacity also describes the ability of a unit volume of a substance to store internal energy while undergoing a given temperature change but without a phase change of the material. Comparing the two base fluids implies the advantage of using coconut oil according to Figure 9. However, the data for volumetric heat capacity are calculated according to Equation (4).

4. Conclusions

In this study, two nanofluids, TiO₂/transformer oil and TiO₂/coconut oil, were prepared following the methods described to observe the behavior of thermal properties with temperature. The following conclusions can be made based on the observations of each analysis conducted.

• Observation of thermal conductivity at different temperatures of TiO₂/transformer oil and TiO₂/coconut oil nanofluids for different volume fractions.

The thermal conductivity of the nanofluid decreased with the temperature and increased with the volume fraction of TiO₂ nanoparticles in both TiO₂/transformer oil and TiO₂/coconut oil nanofluids. The highest thermal conductivity enhancements at 0.012 vol.% at 40 °C were 4.2% and 1.4% for transformer and coconut oil, respectively. Better thermal conductivity was observed in the TiO₂/coconut oil nanofluid throughout the 40–120 °C temperature range. However, the contribution of TiO₂ to increasing thermal conductivity is less in coconut oil than in transformer oil.

b.

Observation of thermal diffusivity at different temperatures of TiO₂/transformer oil and TiO₂/coconut oil nanofluids for different volume fractions

The thermal diffusivity of both nanofluids decreased with the temperature increase. However, there was no significant impact on thermal diffusivity by volume fractions of TiO₂ nanoparticles in both TiO₂/transformer oil and TiO₂/coconut oil nanofluids throughout the temperature range of 40–120 °C.

c.

Observation of volumetric heat capacity at different temperatures of TiO₂/transformer oil and TiO₂/coconut oil nanofluids for different volume fractions.

The volumetric heat capacity of both the nanofluids was increased with the temperature increase. Comparatively higher volumetric heat capacity values were observed in TiO₂/coconut oil throughout the 40–120 °C temperature range.

d.

Comparison of practical results obtained with the theoretical models for thermal conductivity enhancement.

The theoretical values of the thermal conductivity of TiO₂/transformer oil and TiO₂/coconut nanofluid were close to the practical values, with less than a 5% difference for each measurement. The TiO₂/coconut oil nanofluid’s practical results were very similar to the theoretical results of the Maxwell–Garnett model. Considering the percentage difference between experimental and theoretical model values of thermal conductivity, the Maxwell–Garnett model generates much more accurate predictions than other theoretical models.

e.

Stability of the nanofluid samples.

It can be observed that both types of nanofluids show higher stability at the lower concentrations. This is an important observation since even though higher concentrations enhance the thermal properties of the fluid samples, researchers have to maintain the nanoparticle concentration at some level to avoid nanoparticle agglomeration. Due to this behavior, even though higher-concentrated nanofluid samples exhibit higher thermal properties, their thermal properties will deteriorate with time.

Regarding the base oil properties, coconut oil exhibits superior thermal properties compared to transformer oil. However, the addition of nanoparticles enhances the thermal properties of transformer oil rather than coconut oil.

The findings of this study provide a significant dataset for future experiments on nanofluids, which can be highly applicable in designing heat transfer and electrical insulation systems for various industrial applications.