With the gradually increasing development of dense electrical circuit electronics, several aspects were considered for the typical improvements on a heat pipe, referred to as a heat superconductor. Nanofluids with excellent thermal conductivity 1 were applied as the working fluids (WFs) of traditional heat pipes to enhance the thermal performance 23. However, either high-cost metal nanoparticles solved in water were used as nanofluids or these heat pipes were operated only in a horizontal or limited tilt arrangement, rather than the vertical arrangement, in which their thermal performance deteriorates because nanoparticles always accumulate on the evaporator of heat pipes with evaporation converting water to vapor.

An oscillating heat pipe 456 with an additional heat transfer mechanism of conventional force through the interval liquid slug and vapor bubble beyond that of the phase change can overcome the deposition problem. Its closed loop suppresses the entrainment limit, that is, the shear force between condensing and evaporating flow in high-input power 7. However, its multi-turn architecture limits its integration with electronics.

Among low-cost metal-oxide nanofluids 8910, magnetic nanofluids (MNFs) have the magnetically enhanced thermal conductivity 89 to compete with metal nanofluids. However, the common disadvantage of developed MNFHPs in overcoming the deposition problems is the extra-induced problem of magnetic manipulation resulting from unsuitable designs of either active or complex magnetic fields 1112. To avoid the extra power consumption from actively controlling the MNF flowing, a novel MNFHP in a closed-loop architecture was proposed for several main objects: the first was the flowing pattern of the interval vapor moving the MNF slug; the second was the superior suppression of the entrainment limit; and the third was the expansible ability for the integration with some energy storage 13.

Figure 1 shows a closed-loop type of MNFHP. Its main difference from traditional heat pipes is the inner structure, that is, the additional iron porous nozzle in both evaporator and condenser, and magnets on its evaporator and condenser. The iron porous nozzle suppresses magnetic attraction for MNF flowing and forms the flowing pattern of the interval flowing of MNF slug and vapor bubble. Similarly, copper meshes sintered on whole inner surface were also used to transport MNFs by the capillary force against the magnetic attraction.

The used WFs included water as the reference and different concentrations of MNFs composed of a water solvent and magnetic nanoparticles of Fe₃O₄ (Taiwan Advanced Nanotech Corp., Taoyuan,Taiwan). The crystallines of magnetic particles were analyzed, as shown in Figure 2a, using a powder X-ray diffractometer (D-500, Siemens Corp., Belgium, Germany) with Cu-Kα radiation at a wavelength of 0.15418 nm. Figure 2a shows that the phases of Fe₃O₄ agreed with the standard diffraction spectrum (JCPDS no. 65–3107).

The size distribution of the magnetic particles was investigated, as shown in Figure 2b, using dynamic laser scattering (Nanotrac 150, Microtrac Corp., PA, USA). The average hydrodynamic diameter was 10 nm ± 14 nm. The diameter that was detected using dynamic laser scattering is a hydrodynamic diameter because detecting the Brownian motion of the particles was conducted by probing the Doppler frequency shift of the scattered light with respect to the incident light.

The magnetization properties of MNFs were examined using a vibration sample magnetometer (Model 4500, EG&G Corp, California, USA). The concentration of MNFs was diluted with deionized (DI) water. For example, the magnetization of 4 emu/g, equal to 0.8% in volumetric fraction, was obtained (Figure 2c).

Regarding the analysis of the thermal properties of an evaporator or a condenser, Figure 3a,b shows that T_WF increased or decreased exponentially over time for WFs of water or MNFs at different fields. For MNFs, both T WF , f − T tube and τ decreased with Hs to optimal values at H of 200 Oe, where T_WF,f was the final temperature of WF. Furthermore, both T WF , f − T tube and τ of MNFs were smaller than those of water. Based on the one-dimensional model, T WF , f − T tube decreased with the increase of k_WF under the same final heat transfer, Q_sensible,f, because Q_sensible,f is equal to the system of heat loss. In addition, τ could be evaluated according to the dominator of C_WF / k_WF. Therefore, the variations of T WF , f − T tube and τ were related to the thermal parameters of k_WF and C_WF.

Furthermore, considering the variation of Hs and MNF concentrations in both the evaporator and condenser, Figure 3c indicates that all k_MNFC_W / k_WC_MNF increased with Hs and have an optimal value at a specific H. The optimal phenomenon was not significant for low concentrations of MNFs. Moreover, the k_MNFC_W / k_WC_MNF of the evaporator was larger than that of the condenser. The reasons were explained in k_WF and C_WF at different Hs and temperatures.

For magnetic nanoparticles of Fe₃O₄k_MNF was larger than k_W in proportion to the concentration 89. k_MNF increased with Hs and had an optimal value at a specific H89. The specific field increased, whereas the concentration was lower 8. Because of the formation of magnetic clusters, C_MNF decreased with Hs, becoming smaller than 300 Oe 14. Because the origin of mechanism is the same as that of k_MNF, there was possibly an optimal value for the variation of C_MNF with Hs. This was valid for other H-dependent optical properties of MNFs based on this mechanism 15. Regardless, k_MNF / C_MNF had an optimal value at a specific H, but k_W / C_W maintained the same value for all Hs.

In WF temperature, k_MNF decreased as the temperature rose 16, but k_W almost remained the same 17. Similarly, C_MNF decreased slightly as the temperature rose, but C_W almost remained the same 14. Effectively, k_MNFC_W / k_WC_MNF is smaller at high temperatures than at low temperatures if the effect of C_MNF is smaller than that of k_MNF.

Therefore, for the enhanced time of heat transfer in an evaporator or a condenser, k_MNFC_W / k_WC_MNF, as depicted in Figure 3c, achieved up to 3.5 and 1.6, separately, under the optimal condition of MNFs of 0.8% in volumetric fraction at 200 Oe.

Furthermore, by filling MNFs with 0.8% in volumetric fraction into an entire MNFHP, thermal performance of R_MNF / R_W at different Hs was analyzed in Figure 4a, indicating that R_MNF / R_W was less than 1 at all Hs except zero field. The worst R_MNF at zero fields was unexpected according to k_MNFC_W / k_WC_MNF of a single evaporator or a condenser in the Q_sensible test, larger than 1. Possibly, MNF with the high concentration easily resulted in the deposition of magnetic nanoparticles on the tube wall of the evaporator. Conversely, for the smaller R_MNF / R_W at the nonzero field, the deposition phenomenon was possibly avoided. The reason was that the more violent scour on the tube walls of a condenser and an evaporator caused by the faster MNF flowing under higher k_MNFC_W / k_WC_MNF might decrease the MNP deposition. The minimum of R_MNF / R_W occurred at the same as 200 Oe with k_MNFC_W / k_WC_MNF of a single evaporator and condenser in Figure 3c. Furthermore, R_MNF / R_W of the entire MNFHP could be viewed as the ratio of the effective thermal conductivity (k_W,MNFHP / k_MNF,MNFHP) with a simple conduction model in that the thermal resistance R was equal to k_WF,MNFHPA_tube / d_tube, where the effective thermal conductivity of the entire MNFHP filled with some WF (k_WF,MNFHP), the cross area (A_tube) and diameter (d_tube) of the tube are in form, derived from the one-dimensional conduction model. Therefore, the minimal value of R_MNF / R_W of approximately 0.6 was equal to this maximal value of k_MNF,MNFHP / k_W,MNFHP of 1.6. This value is similar with the smaller optimal value of k_MNFC_W / k_WC_MNF in a single condenser rather than that of the larger single evaporator. It could be concluded that the thermal performance of an entire MNFHP was limited by the smaller one because of the series assembly of an evaporator and a condenser, as well as a minor thermal resistance of the flowing mechanism.

Moreover, under the optimal condition of MNFs of 0.8% in volumetric fraction at 200 Oe, the total thermal performance of the entire MNFHP under the optimal condition achieved the maximal P of 60 W and the flowing pattern of the MNF slug and vapor bubble, as shown in the photos at the time interval of 1 sec in Figure 4b, was observed in the acrylic tube between the evaporator and the condenser, thus confirming that the flowing pattern was the design.

This study developed an MNFHP in a closed-loop scheme with magnetically enhanced thermal properties, the flowing pattern of the interval liquid slug and vapor bubble, and the separated paths of evaporated fluids and condensed fluids. The feasibility of an MNFHP with magnetically enhanced thermal performance is valid. The enhanced thermal conductivity ratio of an MNFHP showed that the entire MNFHP was limited by the k_MNFC_W / k_WC_MNF of a condenser subunit. This was useful in designing an MNFHP.

DI, deionized; MNFs, magnetic nanofluids; MNFHP, magnetic-nanofluid heat pipe; WFs, working fluids.

The authors declare that they have no competing interests.

YCC conducted the experiment. JJC provided the idea and proofread the manuscript. CCH provided technical support for the manufacturing equipment. All authors read and approved the final manuscript.

YCC has been an assistant professor in the Department of Mechanical Engineering, Chinese Culture University. His research interests include thermodynamics, heat transfer in electronics, and energy engineering. JJC has been an associate professor in the Institute of Electro-Optical Science and Technology, National Taiwan Normal University. His research interests include photonics, energy, and biomagnetism. CCH was a PhD degree candidate in the Department of Mechanical Engineering, National Central University. He had rich practical experiences of related manufacturing equipment in addition to his major in micro- and nano-scale manufacturing.