Dielectric and thermal behavior investigation of Mn-Zn nano ferrite-fluid for transformer oil applications

2.1 Synthesis of Mn-Zn nano ferrite

To synthesize Mn_0.5Zn_0.5Fe₂O₄ nanoparticles (NPs), the sol–gel technique was employed. The starting materials are manganese (II) nitrate hexahydrate (Mn(NO₃)₂·6H₂O, CAS number 157–66-4), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O, CAS number 10196–18-6) and iron (III) nitrate nonahydrate (Fe(NO₃)₂·9H₂O, CAS number 7782–61-8). All have a minimum purity of 99 % and were provided by Merck Co. The appropriate amounts of them were dissolved in the right amount of distilled water and well blended by using a mechanical stirrer with constant velocity at 80 °C for 20 min. The obtained nitrates solution was added to citric acid (C₆H₈)₇ H₂O) with a 1:1 M ratio between citric and nitrates. To adjust and maintain the pH value of the mixture at 7, an ammonia solution (NH₄OH) was added. By continuous heating at 200 °C under the same stirring conditions, the deep brown gel is created. Gel drying was obtained by burring it in an auto-combustion reaction. Finally, the dried powder was annealed at 550 °C for 5 h.

2.2 Characterization of synthesized Mn-Zn nano ferrite

The phase and crystallinity identification of as-synthesized Mn-Zn nano ferrite were performed by XRD using a Rigaku D/max 2550 V X-ray diffractometer with CuK_α irradiation (λ = 1.54056 Å). The morphology of these samples was investigated by a high-resolution transmission electron microscope (HRTEM) of Joel type JEM-2100. In addition, FTIR absorption spectra of the synthesized powder were recorded in the range of wavelengths (200–1800 cm⁻¹) using a Perkin–Elmer model 1430 spectrometer of ferrite nanopowder that was mixed with potassium bromide powder in tablet form. Moreover, the magnetic measurements were done at room temperature (RT) via a vibrating sample magnetometer (VSM) of type 9600–1 LDJ, with a maximum applied field of 2000 Oe.

2.3 Fabrication of nano ferrite-fluid samples

Nano ferrite-fluid samples were prepared by the addition of Mn-Zn nano ferrite to a commercial mineral oil. An enhancement in the dielectric properties was obtained at a low concentration of NPs as report in previous work (Elsad et al., 2020). The currently used concentrations of Mn-Zn nano ferrite as mineral oil nonfillers were chosen as, 0.0, 0.03, 0.06, 0.1, 0.2, and 0.4 g/L. Here it is worth mentioning that dispersion of Mn-Zn nano ferrite into oil was obtained without using any chemical surfactant to avoid the probability of the surfactant chains coiling up in the oil. Many researchers reported that at low concentrations, metal oxides such as titania, which has a density value close to that of Mn-Zn ferrite, are suspended in the oil without any sedimentation for a short period during which the required experimental test can be carried out. Note that the recorded densities of Mn-Zn ferrite and titania were 4.23 and 4.8 g/cm³, respectively (Mansour et al., 2019; Elsad et al., 2020). Moreover, Mn-Zn ferrite was suspended in distilled water with a PH value similar to that of miner oil to measure zeta potential. The importance of zeta potential is to determine the intensity of the EDL repulsive forces among particles as well as to characterize the stability of a colloidal system. Fig. 1 shows the Zeta potential data of Mn-Zn nano ferrite. This figure reflects a moderate zeta potential value for Mn-Zn nano ferrite to be 8.73 mV. The positive value of the zeta potential indicates a presentation of the positive charge on the Mn-Zn nano ferrite. Such a moderate Zeta potential value that is obtained during the circulation of nanofluid between the external heat exchanger and transformer tank inhibits the sedimentation of Mn-Zn NPs. To obtain a homogeneous distribution of the particles through the oil, each sample was subjected to the ultrasonic wave's sonication for about 25 min using a model UP 400 s sonicator at a power of 400 W and a frequency of 24 kHz. Indeed, the moisture content in nanofluid samples greatly affects the dielectric parameters. As a result, the nanofluid and pure oil samples were stored in a vacuum oven overnight at 40 °C to get rid of any air bubbles and/or moisture that may have formed during the sonication process. The under-investigation nanofluid samples were kept out for cooling in the vacuum for 20 min before starting the measurements.

Close

Fig. 1

Zeta potential data of as-synthesized Mn-Zn ferrite nanoparticles.

Export to PPT

2.4 Breakdown voltage of nano ferrite-fluid samples

The breakdown voltage of the dielectric oil, used in the power transformer has a great influence on the transformer s performance. So the pure oil and the oil that is loaded with various concentrations of Mn-Zn nano ferrite were exposed to the breakdown test. This test was performed according to the ASTM D1816 standard at (RT), in which the applied AC voltage ramp rate is adjusted at 500 V/s, and the mushroom electrode shape with 2 mm gap spacing is used (Elsad et al., 2020; Habashy et al., 2019). For each sample, fifteen breakdown voltage tests were measured and then the average value was estimated for each sample. The electrodes and sides of the cell were cleaned between each of the consecutive measurements.

2.5 Dielectric spectroscopy of nano ferrite-fluid samples

The dielectric constant (ε′) as well as the dissipation factor (tan δ) measurements of the under-investigated nano ferrite-fluid samples were measured in the frequency range of 30 Hz to 1 MHz, according to IEC 60247–2004 protocol, using an Agilent LCR meter of type E4980A. The examined fluid sample was put inside the test cell, which comprises a parallel cylindrical electrode with a 19-mm radius, and a spacing gap of 0.3 mm between them. The parallel mode of capacitance (C_p) and resistance (R_p) is considered the equivalent circuit of the tested sample. Both ε′ and tan δ were estimated based on such modes using the following equations:

Where ${\epsilon }^{\prime }$ is a real part of permittivity, ${\epsilon }^{″}$ is an imaginary part of permittivity, ${\epsilon }_o$ is free space permittivity, A is the cross-section area of the electrode, d is the gap spacing between two electrodes, and $f$ is the frequency of applied AC voltage. The readings of both C_p and R_p were acquired using LabVIEW-based software as the average values of 128 readings for each at each applied voltage.

2.6 The thermal characteristic measurements

Cooling of the transformer winding is one of the two main functions of the transformer oil. So, the thermal characteristic measurements of nano ferrite-fluid samples were done using the heating–cooling experiment according to the recommended test reported in (Du et al., 2015). The schematic diagram for thermal conductivity evaluation is illustrated in Fig. 2. It principally consists of a cylindrical testing section, a heating coil, the AC power supply, and a thermocouple that is fixed above the heating coil by about 5 mm. In order to simulate the rise in temperature of the power transformer oil during its operation due to power loss through winding, the heating coil inside the nanofluid sample was connected to an AC power supply with an output power of 10 W, and the temperature was recorded with time. The coil was used for the heating process for each one of the nanofluid samples, and then the electric power was disconnected when the temperature reached 75 °C. Thereafter, the cooling process was launched by recording the temperature of the nanofluid samples with time during the cooling process. The variations in temperatures during the heating and cooling processes for the nano ferrite-fluid samples were recorded at a time-sequence interval of 1 min.

Close

Fig. 2

The schematic diagram for the experimental thermal conductivity setup for the heating–cooling test.

Export to PPT

3.1 Structural of as-synthesized Mn_0.5Zn_0.5Fe₂O₄ NPs

Fig. 3 shows the XRD pattern of the as-synthesized Mn-Zn nano ferrite NPs. The XRD of the synthesized sample perfectly matched that of the Mn–Zn ferrite spinel cubic structure. XRD peaks at (1 1 1), (2 2 0), (3 1 1), (2 2 2), (4 0 0), (4 2 2), (5 1 1), (4 4 0), (6 2 0) and (5 3 3) indicate that the obtained structure is perfectly matched with the data of spinel structure ICDD card No. (74–2402) (El-Dek et al., 2017). Notably, the characteristic peak of hte Mn–Zn ferrite spinel cubic structure at (3 1 1) is obtained. Also, minor impurities peaks that are linked to α-Fe₂O₃ and denoted by (*) on Fig. 3 were observed. The XRD result also confirms that the synthesized sample has a single phase with no appearance of any extra lines that may refer to the absence of secondary phases. The accurate lattice constant, a of as-synthesized Mn_0.5Zn_0.5Fe₂O₄ nanoparticles was obtained by using the Nelson-Riley (N-R) extrapolation method. The calculated values of the lattice parameter a_cal, of the as-synthesized Mn_0.5Zn_0.5Fe₂O₄ at different (hkl) orientations plotted against the N-R function F(θ) and the accurate lattice constant is determined from the extrapolation of fitting sightline for that relation. The values of lattice parameter a_cal, and the N-R function, F(θ) were calculated from the following equations (Katz, 1964):

Close

Fig. 3

XRD pattern of as-synthesized Mn0.5Zn0.5Fe2O4 ferrite nanoparticles. inset is showing the Nelson–Riley (N-R) function F(θ) versus calculated lattice constant to evaluate the accurate lattice constants.

Export to PPT

where (hkl) are the Miller indices, λ is the wavelength of XRD Cu target = 1.5406 Å, and θ is the diffraction angle corresponding to the (hkl) plane.

The accurate value of the lattice constant was obtained from extrapolation of the inset of Fig. 3 straight line and was found to be around 8.3952 Å. Such obtained value is quite close to the standard value in card no. (74–2402) (El-Dek et al., 2017). The average crystallite size (D) of the synthesized Mn-Zn ferrite has been calculated using Scherrer’s formula (Scherrer; Langford and Wilson, 1978):

where β is the full width at the half maximum of the peak in the XRD pattern. The estimated D values for all peaks have reflected that the investigated Mn_0.5Zn_0.5Fe₂O₄ NPs are of the nano-size range between 19 and 34.9 nm with an average value of about 26.7 nm, whereas the D value for the most preferred orientation plane, (3 1 1), was found to be 27 nm.

The HRTEM micrograph of the synthesized Mn_0.5Zn_0.5Fe₂O₄ NPs is demonstrated in Fig. 4. This figure shows that the obtained shape of the synthesized ferrite sample is semi-spherical with a diameter ranging from 21 to 33 nm. Such values are in good agreement with the estimated size range from XRD. The obtained agglomeration of the NPs as seen in Fig. 4 could be attributed to the commonly high magnetic moment of ferrite particles as well as to the incorporation of Zn in the NPs (Manova et al., 2004). The FTIR absorption peaks of interstitial sites of both tetrahedral and octahedral groups of the spinel lattice are the main characteristic tools for the structure description. The FTIR spectrum of Mn_0.5Zn_0.5Fe₂O₄ NPs recorded in the atmosphere at RT is given in Fig. 5. Two main strong peaks at υ₁ = 565 and υ₂ = 386 cm⁻¹ correspond to spinel cubic structure (Guo et al., 2010). The high-frequency band (υ₁) refers to tetrahedral complexes, and the low-frequency band (υ₂) refers to octahedral complexes, which can be attributed to the stretching vibration frequency of the metal–oxygen at the tetrahedral and octahedral sites, respectively (Sivakumar et al., 2011). The band at 873 cm⁻¹ related to the C-N stretching vibration and the C–H bending vibration. The band at 1036 cm⁻¹ could be assigned to the nitrate traces. Whereas, the band at 1450 cm⁻¹ is linked to the carboxylate group symmetric vibration (C = O) bonded to the nanoparticle surface. This result confirms the formation of Mn_0.5Zn_0.5Fe₂O₄ NPs with a spinel structure, agreeing with the XRD diffraction results. Moreover, the band obtained at 1608 cm⁻¹ connected to O–H stretching vibrations interacting across H bonds (Gunjakar et al., 2008; Foroughi et al., 2015; Sangmanee and Maensiri, 2009; El-Okr et al., 2016).

Close

Fig. 4

HRTEM micrographs of as-synthesized Mn_0.5Zn_0.5Fe₂O₄ nanoparticles: (a) 50,000 × magnification with 50 nm scale and, (b) 25,000 × magnification with a 100 nm scale.

Export to PPT

Close

Fig. 5

FTIR spectrum of as-synthesized Mn_0.5Zn_0.5Fe₂O₄ nanoparticles.

Export to PPT

3.2 Magnetic characteristics of the synthesized Mn_0.5Zn_0.5Fe₂O₄ NPs

The magnetization-magnetic field (M−H) hysteresis curve of the under-investigation Mn_0.5Zn_0.5Fe₂O₄ NPs is shown in Fig. 6. Such a figure shows the existence of an ordered magnetic structure localized in the spinel structure that refers to the ferromagnetic behavior of the investigated sample. The obtained values of the coercive field and the remanence magnetization are found to be 206.43 Oe and 16.98 emu/g, respectively. The magnetic moment (η_B) could be estimated using the measured saturation magnetization, M_s, using the following formula (Gabal, 2009).

Close

Fig. 6

M−H hysteresis loop curve of Mn_0.5Zn_0.5Fe₂O₄ NPs at 300 K. The inset represents the specific magnification of the curve between −200 and 200 Oe.

Export to PPT

where M_s and M_w are the saturation magnetization and the molecular weight of the specimen, respectively. In general, the magnetic characteristics of nano ferrites depend on the particle size and the kind of cation‘s structure. Table 1 illustrates the obtained ferromagnetic characteristic parameters in comparison with other works that were done on magnetite and Ni-Zn nano ferrite prepared by the sol–gel route (Lemine et al., 2012; Arais et al., 2019; Hemida et al., 2018; Ounacer et al., 2021; Ahilandeswari et al., 2020; Abdulhamid et al., 2019). The measured coercivity has a higher value as shown in Table 1 (Segal et al., 2000). The sample size agrees with a magnetic single domain and its rotation is mainly responsible for coercitivity, which is linked to the effective anisotropy of particles as reported to the Stonere-Wohlfarh model (Stoner and Wohlfarth, 1991). The sum of the contributions of the shape anisotropy and the magneto-crystalline anisotropy is the effective anisotropy of each particle (Ido et al., 1986). So, the semi-spherical shape with different diameters as obtained from HRTEM raises effective anisotropy due to its higher aspect ratio, which causes the high coercivity. It can be noticed that Mn-Zn ferrite is a noticeable candidate for insertion as a filler into the fluid as compared with other magnetic ferrite composites such as magnetite, Ni-Zn, Co-Mg, Ba-Nd, and Ni-Co-ferrites due to its high saturation magnetization (M_s), magnetic moment (η_B), and moderate particle size (Nalbandian et al., 2008; Goldman, 2006).

3.3 Electrical breakdown voltage of the investigated nano ferrite-fluid

The variation of the AC breakdown voltage of oil as a function of Mn-Zn ferrite NPs is illustrated in Fig. 7. As it is seen, there is an enhancement in breakdown voltage with increasing the content of Mn-Zn ferrite NPs to 0.2 g/L. However, the breakdown voltage tends to reduce at a concentration of more than 0.2 g/L. Indeed, the initiation of streamers through transformer oil is related to space charge density, distortion of the electric field through oil, and the content of dissolved water in transformer oil (Lv et al., 2017). The probability of creating a large space charge for pure oil could be high near the top electrode, therefore, the distributions of an electric field through the sample are non-homogeneous. In contrast, the remarkable interfacial zones for nanofluid, formed at the contact region between NPs and base oil, play a crucial role in improving the breakdown voltage. More specifically, the high density of the interfacial zones, obtained at low concentrations of NPs, enhances the trapping and de-trapping processes of free electrons in electronic trap sites that are causing a reduction in the speed and energy of free electrons (Lv et al., 2017). Therefore, the probability of further electron production during the ionization impact becomes limited, and further development of the streamer is obstructed. Furthermore, the capture of free electrons in the interface region around NPs, which are distributed through the transformer oil, reduces the space charge density. Hence, the non-homogeneity of the electric field through the magnetic nanofluid samples could be limited. On the other hand, some of the dissolved water in the transformer oil may be attached to the surface of NPs (Elsad et al., 2020). The reduction in space charge density, the distortion electric field, and the adsorption of the dissolved water could be the probable reasons for the enhancement in breakdown voltage for the investigated nano ferrite-fluid up to 0.2 g/L. The trap creation is linked to the suspension degree of NPs through transformer oil. So, the reduction in breakdown voltage for concentrations grater than 0.2 g/L may be assigned to the probability of Mn-Zn ferrite NP agglomerations. Such agglomerations decrease the density of traps for free-electron and thin channel paths are easily offered.

Close

Fig. 7

The variation of AC breakdown voltage measured after 1 day from preparation of transformer oil as a function of Mn-Zn ferrite NPs concentration.

Export to PPT

3.4 Dielectric spectroscopy of the investigated nano ferrite-fluid

Fig. 8 shows the dependency of ${\epsilon }^{\prime }$ in the investigated nano ferrite-fluid samples on the applied frequency at different concentrations of Mn-Zn ferrite NPs. This figure shows a slight reduction approaching to constant ${\epsilon }^{\prime }$ with raising the applied field frequency. The same dependency of ${\epsilon }^{\prime }$ on frequency has been reported by Liu et al. and Mansour et al. (Mansour et al., 2016; Liu et al., 2011). On the other hand, Fig. 8 indicates that the ${\epsilon }^{\prime }$ value is decreasing with increasing the concentration of Mn-Zn ferrite NPs until 0.06 g/L and then increases. The decrease percentage in ${\epsilon }^{\prime }$ value obtained at 0.06 g/L is approximately 4 % compared to that obtained for pure oil. The inset of Fig. 8 shows the variation of ${\epsilon }^{\prime }$ as-synthesized Mn-Zn ferrite NPs at RT. In this figure, the reduction in ${\epsilon }^{\prime }$ from 300 at low frequency to around 10 at high frequency can be noticed. According to superior results, there is a clear difference between the values of ${\epsilon }^{\prime }$ both Mn-Zn ferrite NPs and mineral oil. The decrease in the ${\epsilon }^{\prime }$ values at 0.06 g/L concentration could be explained because such a concentration of NPs offered a suitable high density of the interfacial zones between Mn-Zn ferrite NPs and the oil. Indeed, the increase in the interfacial zones leads to limitations in the electric dipole motion (Elsad et al., 2020; Habashy et al., 2019). In addition to the high density of the interfacial zones, the dielectric feature of the host oil plays an essential role in the change of ${\epsilon }^{\prime }$ values of especially at low concentrations of NPs. For concentrations higher than 0.06 g/L, the ${\epsilon }^{\prime }$ value of Mn-Zn ferrite NPs has a predominant effect on ${\epsilon }^{\prime }$ value of magnetic nanofluid, which is reflected in the higher value obtained ${\epsilon }^{\prime }$ in comparison with that found for pure oil.

Close

Fig. 8

Variation of ${\epsilon }^{\prime }$ for the studied nano ferrite-fluid as a function of the applied frequency at RT. The inset shows the variation of ${\epsilon }^{\prime }$ as-synthesized Mn-Zn ferrite NPs at RT.

Export to PPT

The variation of the tan δ for the investigated nano ferrite-fluid samples as a function of the applied frequency is represented in Fig. 9. This figure reflects that the tan δ value of each investigated nano ferrite-fluid dramatically declined with increasing the applied frequency up to 700 Hz. The small hump that appeared in all investigated samples at 100 Hz can be assigned to the moisture problem. Beyond this interval of frequency, the variations of tan δ values exhibit an almost stable plateau. The significant reduction in tan δ at low-frequency intervals can be explained using an equivalent circuit that consists of a parallel connection between resistance, denoting the dissipative portion of the nanofluid response, and reactance, denoting the storage element (Aziz et al., 2018). The inrush current through each branch of the equivalent circuit relies on the values of resistance and reactance, which are related to the applied frequency. The increase in the applied field frequency causes a reduction in the current flowing through the resistance due to the decrease in the value of reactance. The inset of Fig. 9 shows the variation of tan δ for transformer oil with Mn-Zn ferrite NPs concentration at 300 Hz. As it is seen, the tan δ value is significantly reduced with the increase in Mn-Zn ferrite NPs concentration to 0.1 g/L. Beyond this concentration, the tan δ value is slightly enhanced. The insertion of Mn-Zn ferrite NPs into power transformer oil has two contrary influences on the AC conductivity: enhancing the density of the charge carriers due to the existence of inorganic filler and capturing the charge carriers in traps formed at the surface of NPs. The reduction in the recorded tan δ values of the nano ferrite-fluid samples compared to pure power transformer oil refers to the predominant influence of charge-carrying trapping due to the formed interfacial zones between Mn-Zn ferrite NPs and oil.

Close

Fig. 9

The variation of tan δ of the investigated nano ferrite-fluid as a function of applied frequency. The inset represents the variation of tan δ with Mn-Zn ferrite NPs concentration at an applied frequency of 300 Hz.

Export to PPT

3.5 Thermal behavior of the investigated nano ferrite-fluid during heating and cooling processes

The evaluation of thermal properties for all the investigated nano ferrite-fluid samples was investigated considering two procedures, heating and cooling procedures. Fig. 10 illustrates the temperature variation of the investigated Mn-Zn nano ferrite-fluid samples with lapse time during heating and cooling processes. This figure indicates that the rate of increase in temperature of the nano ferrite-fluid samples during heating is slower than that recorded for the pure transformer oil.

Close

Fig. 10

Temperature variation of nano ferrite-fluid during the lapse time of heating and cooling processes.

Export to PPT

However, the rate of cooling was enhanced by introducing Mn-Zn ferrite NPs. The thermal time constant (τ) of cooling is a significant tool to reflect the variation rates of temperature. The variation rate of temperature during the cooling process is represented by the following formula (Bressin and Willmer, 2000):

where T, T_s, T_a, and t are the instantaneous temperature of the sample during cooling, the initial temperature at the beginning of the cooling process, the ambient temperature, and time, respectively. The thermal time constant, τ was estimated for each nanofluid sample from the slope of the straight line that obtained from the relation between Ln (T-T_a) versus time (Bressin and Willmer, 2000). Fig. 11 shows the variation of Ln (T-Ta) as a function of time during the cooling process for all investigated nanofluid samples. The estimated thermal time constant during cooling is illustrated in Table 2. The recorded results reflect that the thermal time constant of transformer oil during cooling is decreased by introducing the Mn-Zn ferrite NPs.

Close

Fig. 11

The variation of Ln (T-Ta) as a function of time during the cooling for all nanofluid samples.

Export to PPT

The reduction in the thermal time constant refers to an enhancement in the rate of heat dissipation. Accordingly, the presence of Mn-Zn ferrite NPs in mineral oil plays an essential role in heat dissipation from the heating coil to the surrounding medium across the oil during the heating and cooling procedures. Indeed, the thermal behavior of magnetic nanofluid has been considerably influenced by the thermal conductivity of both NPs and base oil and the surface-to-volume ratio of NPs (Saidur et al., 2011). The enhancement in heat dissipation of nanofluid cannot be only attributed to the higher thermal conductivity of NPs than that of the base fluid but also to the interparticle interaction in the electric double layer (EDL), that is created around NPs through a fluid as reported, by Saidur et al. (Saidur et al., 2011). Here it is worth mentioning that the real usage of the investigated Mn-Zn nano ferrite-fluid at a high external magnetic field generated from winding in-service operation as well as temperature gradient causes a probable quiet enhancement in heat convective through the ferromagnetic feature of the investigated NPs. The enhancement in heat dissipation by introducing Mn-Zn ferrite NPs with high magnetic properties of the investigated Mn-Zn ferrite NPs may inhibit the formation of local hot spots, which cause the breaking of the oil’s molecular composition and insulation fallout.