1 Introduction

Heat pipes and vapor chambers perform excellently in efficient heat transport because of transferred energy via liquid–vapor phase change of the working fluid. Thus, the choice of the working fluid is vital in thermal performance of heat pipes.

Nanofluids are a new type of engineered heat transfer fluid developed by Argonne National Laboratory, USA. They are generated by suspending nanometer-sized metallic or nonmetallic particles in base fluids, such as water, engine oil, and ethylene glycol [1]. Many researchers have investigated thermal conductivity [2–9], convective heat transfer, and pool boiling heat transfer characteristics of nanofluids [10–24]. Nanoparticles, such as Ag, CuO, Al2O3, diamond, iron oxide, and Au, have been used to study the influence of nanofluids on thermal performance of micro-grooved heat pipes [25–28], mesh wicked heat pipes [29–31], sintered metal heat pipes [32], thermosyphons [33–35], and oscillating heat pipes [36–38].

Liu et al. [25] investigated the effects of aqueous Cu nanofluids with different nanoparticles on the thermal performance of a cylindrical micro-grooved heat pipe under different operating pressures. The addition of Cu and CuO nanoparticles to the base fluid can improve the thermal performance of the heat pipe. By contrast, the addition of SiO nanoparticles deteriorates heat transfer performance. The heat transfer mechanism was presented as follows: (1) Increased effective thermal conductivity of the nanofluid can enhance conductive heat transfer; (2) Decreased solid–liquid contact angle between the nanofluid and the heated surface can increase capillary force and stretch liquid film in the grooves; (3) Random collision of nanoparticles in the base fluid (Brownian motion), which is a nanoscale effect, can enhance convection; (4) The last aspect of the heat transfer mechanism is attributed to the coating layer formed on the heated surface during the heating process.

Tsai et al. [30] studied the thermal performance of a conventional circular meshed heat pipe by using an aqueous solution of gold nanoparticles. The thermal resistance of the heat pipe that used nanofluids is 37 % lower than that which used deionized water.

Chen et al. [31] studied the thermal performance of an axially flat mesh wicked heat pipe by using silver nanoparticle suspensions. The total resistance of the heat pipe that used nanoparticle suspensions is lower than that of the heat pipe that used DI water.

Kang et al. [32] used water/Ag nanofluids as the working fluid for a conventional sintered circular heat pipe with a 1 mm wick thickness. Nanoparticles with 10 and 30 nm diameters were used. Accordingly, the increase in the difference in wall temperature for heat pipe with nanofluid is smaller than that for a heat pipe filled with pure water under various heat loads.

Noie et al. [34] employed water/Al2O3 nanoparticle suspensions as the working fluid in a thermosyphon. For different input powers, the efficiency of the thermosyphon increased up to 14.7 % when Al2O3/water nanofluid was used instead of pure water.

Huminic et al. [35] also studied the heat transfer characteristic of a thermosyphon with iron oxide/water as the working fluid. Nanoparticles significantly influenced the enhancement of the heat transfer characteristic of the thermosyphon.

Qu et al. [36] experimented on the thermal performance of an oscillating heat pipe charged with Al2O3/water nanofluid. Alumina nanofluids significantly improved the thermal performance of the oscillating heat pipe. The change of surface condition at the evaporator due to the settlement of the nanoparticle was the major factor in the enhanced thermal performance of the oscillating heat pipe.

Ma et al. [38] investigated a nanofluid oscillating heat pipe to determine the effect of nanofluids on transport capability. The heat transport capability of the oscillating heat pipe increased significantly when it was charged with nanofluid. In addition, heat transport capability depends on the operating temperature.

Do et al. [39] studied whether the enhancement of heat transfer in nanofluid heat pipes can be attributed to changes in fluid thermophysical properties and to characteristics of a heat transfer surface. They constructed two mathematical models to investigate the heat transfer characteristics of a flat microheat pipe with a grooved wick by using water-based Al2O3. Model I represented the heat transfer enhancement caused by changes in fluid thermophysical properties under the assumption that nanofluids in the base fluid were uniformly dispersed and suspended without any accumulation or deposition on the heat transfer surface in the evaporation region. Model II represented the enhancement of heat transfer brought about by changes in the heat transfer surface caused by nanoparticle deposition. Model II was more applicable for predicting the thermal performance of nanofluid heat pipes.

Though many studies on the heat transfer enhancement for nanofluids use as the working medium in thermosyphons have been conducted, the heat transfer enhancement mechanism by using nanofluids has not yet been rigorously investigated.

In the previous report by Diao et al. [40], the micro-channel heat sink resulted in excellent thermal performance for a high-powered chip cooling application was investigated experimentally. In this study, nanofluid Cu/R141b was used as the working fluid for different volume concentrations to study the phase change heat transfer characteristic for a vertical plate with rectangular microchannels under a saturated vapor ambient at different operating pressures. A significant improvement in the heat transfer for the microchannel surface was achieved by using Cu/R141b nanofluids instead of pure R141b. Nanoparticle coverings (nano-coverings) used in enhancing heat transfer were also investigated. Nano-coverings on the evaporator surface can significantly affect heat transfer performance and are responsible for improved heat transfer of nanofluids on the plate with microchannels.

2 Nanofluid production

This study used Cu with an average diameter of 20 nm as nanoparticles. Cu/R141b nanofluids were produced via two-step method. Dried Cu nanoparticles were dispersed directly into R141b. Cu nanoparticles with an average diameter of 20 nm were purchased from a commercial manufacturer. To prevent formation of nanoparticle agglomerates in Cu–R141b mixture and to generate a stable nanofluid, an ample amount of dispersant SDBS was added to the base fluids, and solid–liquid mixtures were oscillated by using an ultrasonic cleaner for more than 10 h. The amount of the SDBS is twice that of nanoparticles for different nanoparticle volume concentrations. Thus, the volume concentrations of SDBS were twice that of nanoparticles. Cu/R141b nanofluids with three different volume concentrations of 0.001, 0.01, and 0.1 % were prepared for this experiment. Figure 1 shows SEM photographs of the nanofluids with different volume concentrations.

Fig. 1
figure 1

SEM images of Cu nanoparticles. a 0.001 % volume concentration, b 0.01 % volume concentration, c 0.1 % volume concentration

Full size image

3 Experimental setup

The experimental system is illustrated in Fig. 2. It mainly consists of a test vessel, condensing system, copper bar main heater and auxiliary heaters, power supply, and a data acquisition system. The internal dimensions of the test vessel are 70 mm × 60 mm × 125 mm. Four electrical heaters were inserted in the copper bar and used for evaporation/boiling tests. A water cooling condenser was mounted on the test vessel for vapor condensing. Cooling water was circulated through the condenser at a constant velocity to remove heat then placed in a water bath with a constant temperature. The cooling water temperature was controlled precisely to maintain a constant operating pressure and temperature in the vessel for different heat flux.

Fig. 2
figure 2

Schematic of experimental apparatus. 1 condenser, 2 pressure sensor, 3 Agilent D.A. system, 4 heating block, 5 primary heater, 6 voltage regulator, 7 auxiliary heater, 8 evaporation chamber, 9 heat transfer surface, 10 refrigerator

Full size image

Two P-100-type thermoresistances were used to measure liquid and vapor temperatures. Two auxiliary cartridge heaters were mounted at the bottom inside the test vessel to maintain the liquid and vapor at saturation temperature. A pressure transducer installed on the top plate was used to measure the operating pressure of the test vessel. A data acquisition system was used to monitor temperature readings throughout the copper block. The thermocouples readings were recorded by the multi-channel data acquisition system and the measurement readings were recorded at a frequency of 0.1 Hz. A computer was set up to record and process the experimental data.

The structure of the test heater is illustrated in Fig. 3. In this section, the locations for temperature measurement are shown. The heated surface is the top surface of a copper round-type block with a 20 mm diameter and heated by four electrical cartridge embedded heaters. The maximum power provided by all heaters could reach 1,000 W. The copper block was thermally insulated by a glass wool layer and fiber blankets.

Fig. 3
figure 3

Structure of heating block and arrangement of thermocouples

Full size image

The top surface of the copper was attached symmetrically to the plate with micro-channel. In the copper block, four holes, each 10 mm deep and 1 mm in diameter, were drilled into the center of the copper block at 10 mm intervals. Copper-constantan sheathed thermocouples with a diameter of 1.0 mm were embedded in each hole. With the temperature difference between thermocouples T 1 and T 2 as well as the thermal conductivity of the pure copper heating section, the steady-state 1-D axial average heat flux could be calculated. Hence, steady-state heat dissipation that results from evaporation/boiling from the micro-channel surface could be determined with high accuracy. The space between the contact surfaces was then filled with Ag thermal silicone, whose thermal conductivity is above 4.0 W m−1 K−1.

A schematic of the micro-channel surface is shown in Fig. 4. Twenty-five micro-channels were cut into the top surface of the copper blanket with a 44 × 60 mm2 area by using precision sawing. The test piece contains micro-channels that are 250 μm wide (W c ) and 500 μm high (d) with a 150 μm thick (p) intervening fin.

Fig. 4
figure 4

Micro-channel structure

Full size image

For each experimental iteration, the evaporator was filled with a pre-measured amount of working fluid. A filler tube was used after eliminating any non-condensable gases from the evaporator by using a vacuum pump. Figure 5 illustrates the arrangement of the heated area at the back of the copper plate with micro-channels. The initial liquid level in the reservoir is at the bottom edge of the heated region.

Fig. 5
figure 5

Schematic of copper plate with micro-channel and heater surface with respect to the liquid reservoir

Full size image

During the experiment, the working liquid was driven by capillary force into the micro-channels of the evaporator surface; phase change occurred because of input heat flux. At the surface of the micro-channels, the heat from the copper heater was absorbed by the working liquid. The vapor was condensed to release latent heat to the cooling water via convection. The condensed liquid was then drawn back to the evaporation chamber to complete a cycle. The heat transfer process mainly occurred at the liquid–vapor interface. Figure 6 shows a portion of the micro-channel that contains working liquid. Micro-channels stimulate thin liquid films and supply working liquid to the evaporating region. A concave vapor–liquid interface is formed. The curvature of the vapor–liquid interface changes as heat flux increases.

Fig. 6
figure 6

Schematic of meniscus in the micro-channels

Full size image

4 Experimental procedure

For each experiment, after the micro-channel surface was set, the vessel was emptied by using a vacuum pump; prepared nanofluid liquid was supplied into the vessel. Through the main heater, heat was supplied to the micro-channel surface. At the same time, heat input to the auxiliary heaters was adjusted to a proper value to maintain the working fluid at a saturated state. Tests were conducted after allowing the nanofluid temperature to level off at a constant temperature. During each run, power was supplied increasingly to the main heater and auxiliary heaters. In each run, the heating power increased monotonically. The heat flux q that dissipated through the evaporation plate under steady conditions could be determined by using Eq. (1)

$$q = \frac{{\lambda (T_{1} - T_{2} )}}{l}$$
(1)
$$T_{t} = T_{4} - q\frac{1}{\lambda }$$
(2)
$$R = \frac{{T_{t} - T_{s} }}{q}$$
(3)

where l is the distance between the thermocouples; T 1 , T 2 , T 3 , and T 4 are the temperature values of the four thermocouples in the copper heater; T t is the temperature for the back of the copper plate with micro-channels; T s is the saturation temperature of the working fluid at operating pressure; R is the thermal resistance.

Uncertainties in the temperature measurements and length (width) are ±0.1 K and 0.01 mm, respectively. The uncertainty of pressure transducer is ±0.5 %. The evaluation of uncertainty is based on the fact that a differential change of a function can be expressed in terms of differential changes of the dependent variables [12]. The uncertainties are calculated as follows: the heat flux was estimated to be less than ±7.22 %, and the thermal resistance is less than ±7.17 %.

5 Results and discussion

5.1 Results of nanofluid boiling experiments

Figure 7 shows the boiling curves of the average evaporation heat flux q versus the wall superheat for Cu nanofluids with volume concentrations of 0.001, 0.01, and 0.1 % in comparison with that of pure water under six steady operation pressures.

Fig. 7
figure 7

Relationship of heat flow density and the degree of superheat at different pressures and particle concentrations of nanofluids

Full size image

Heat flux increases as the wall superheat increases (Fig. 7). For the boiling values of tested nanofluids with volume concentrations of 0.001 and 0.01 %, the boiling curves of R141b-Cu nanofluids significantly shift to the left compared with that of pure R141b. This result indicates that for a given wall superheat, heat flux increases as nanoparticle concentrations escalate. Figure 7a–c show that at a lower wall superheat, heat flux for the nanofluid R141b-Cu with 0.1 % volume concentration is larger than that for the pure R141b. However, at a higher superheat degree, the curve for the nanofluids with 0.1 % concentration shifts to the right. Thus, heat transfer deterioration occurred at a higher wall superheat. These results indicate that nanoparticle concentration significantly influences evaporation/boiling heat transfer at every test pressure. Figure 7d, e show that as the wall superheat increases, the curve for the nanofluid with 0.1 % volume concentration shifts to the right. This phenomenon indicates that at higher heat flux, compared with base fluid R141b, heat transfer deterioration effect occurs. Figure 7f shows that heat transfer deterioration occurs in the operating range at 2.0 × 105 Pa.

Figure 8a–c show the evaporation/boiling thermal resistance of Cu/R141b nanofluid versus the wall superheat in the 0.86 × 105 Pa to 0.96 × 105 Pa pressure range. The thermal resistance decreases as the wall superheat increases. The thermal resistance of nanofluids with 0.001 and 0.01 % concentrations for the entire range is improved at different degrees because of nanoparticles. Reduction is highly noticeable at low operating pressure. Figure 8d–f indicate that thermal resistance first decreases as wall superheat increases and reaches the lowest point, and then begins to increase with further increase of wall superheat in the 1.0 × 105 Pa to 2.0 × 105 Pa pressure range. The thermal resistance reducing effect is not obvious for nanofluids with concentrations of 0.001 and 0.01 % compared with the base fluid R141b. Figure 8f shows a clear deterioration of heat transfer effect for the nanofluid with 0.1 % concentration compared with R141b at the operating pressure of 2.0 × 105 Pa. For the tested pressures, an optimal volume concentration exists, which corresponds to the lowest reduction effect of the thermal resistance of approximately 0.001 %. Thermal resistance for R141b is 50 % of the thermal resistance of nanofluid with 0.01 % volume concentration.

Fig. 8
figure 8

Relationship of thermal resistance and the degree of superheat at different pressures and particle concentrations of nanofluids

Full size image

The improvement in heat transfer for the nanofluids is likely due to the nanoparticle layer that formed on the microchannel surface. Compared with the clean microchannel surface, the nanoparticle layer that formed on the channel surface could extend the evaporation area to enhance evaporation heat transfer. Active nucleation site density on the nanoparticle layer may also enhance the boiling transfer process in the micro-channels. The increase in the thermal resistance with improved nanoparticle concentration of 0.1 % can be attributed to the corresponding thicker nanoparticle layer, which in turn offers improved thermal resistance. As the nanoparticle deposition increases, the additional layer of nanoparticle covering increases the covering thickness. Given this phenomenon, an optimal nanofluid concentration exists in which one observes minimum thermal resistance. Results from this study indicate that the optimal nanofluid concentration is approximately 0.01 %.

Figure 9 shows pictures of the micro-channel surface captured via 3D microscopy. The pictures indicate improved deposition with increased nanoparticle concentration. A significant deposit of nanoparticles was noted on the micro-channel surface after nanofluid evaporation/boiling experiments. The degree of coverage of Cu nanoparticle deposition increases as nanoparticle concentration improves. A higher nanofluid concentration indicates thicker nanoparticle covering on the heater surface, which increases thermal resistance.

Fig. 9
figure 9

Photographs of microchannel surface before and after experiment. a Structure of micro-channel before experiment, b structure of micro-channel after the experiment (0.001 % volume concentration), c structure of micro-channel after the experiment (0.01 % volume concentration), d structure of micro-channel after the experiment (0.1 % volume concentration)

Full size image

5.2 Source of nanoparticle deposition

Some researchers have studied the reason for the formation of a nanoparticle deposition layer on the heated surface. Kwark et al. [17] investigated mechanisms that may be responsible for depositing and adhering nanoparticles to the heater surface during nanofluid boiling. They reported that nanofluid boiling, specifically microlayer evaporation, is responsible for the deposition and adherence of nanoparticle coatings on the heater surface. This finding is consistent with previous studies by Kwark et al. [16] and Kim et al. [19], who hypothesized that the nanoparticle coatings, which formed during nanofluid boiling, are created as the vapor bubble’s microlayer evaporates. As vapor bubbles grow, the evaporating liquid leaves nanoparticles behind, which then concentrate at the base of the bubble, namely, the microlayer. As the microlayer evaporates, it leaves the nanoparticles behind, which then bond to the hot heater surface.

In the present study, two heat transfer modes can be considered: the evaporation heat transfer on the contact line (including meniscus) and the heat transfer by many micro-vapor bubbles generated in the micro-channels. Thus, the nanoparticle layer formed on the micro-channel surface is related to liquid evaporation on the contact line, and liquid evaporation beneath the bubble leaves nanoparticles behind, which then bond on the micro-channel surface.

To clarify the evaporation/boiling heat transfer mechanism for the nanofluids in the vapor chamber with rectangular micro-channels, we conducted a series of tests that used a nanoparticle-coated surface obtained by using a clean micro-channel surface in Cu nanofluids in pure R141b.

The nanoparticle layer accumulated on the micro-channel surface in the first test. In the second test, we investigated the effect of this layers on evaporation/boiling heat transfer performance in pure R141b. As previously mentioned, these nanoparticle layers were produced during nanofluid evaporation/boiling experiments at three nanofluid concentrations of 0.001, 0.01, and 0.1 %. Thus, the nanoparticle layer created during experiments that used 0.001, 0.01, and 0.1 % nanofluid concentrations will be defined as 0.001 % nanoparticle layer, 0.01 % nanoparticle layer, and 0.1 % nanoparticle layer, respectively.

Figure 10 plots the thermal resistance versus the wall superheat for nanoparticle layer covering surface in pure R141b under different operating pressures of 1.0 × 105 Pa to 2.0 × 105 Pa. For purposes of comparison, the thermal performance against the wall superheat for clean surface in nanofluids is presented in the same figure. Figure 10a–d show that the curve for the clean surface in nanofluids almost coincides with that for the nano-covering in pure R141b at a low operating pressure of 0.86 × 105 Pa to 0.96 × 105 Pa. Figure 10e, f show the discrepancy between the two curves at a higher operating pressure of 1.0 × 105 Pa to 2.0 × 105 Pa compared with that for the lower operating pressure. Figure 11 illustrates the thermal resistance versus the wall superheat for nanoparticle layer covering surface in pure R141b and for clean surface in nanofluids with the 0.01 % nanofluid concentration. Figure 11 shows the same trend as Fig. 10. The nanoparticle layer produced by using 0.001 and 0.01 % nanofluids produce almost the same heat transfer enhancement as nanofluids. Figure 12a–d show that for nanofluids with a higher concentration of 0.1 %, the result is a relatively large difference of thermal resistance values for the clean heater tested in nanofluids and the 0.1 % nanoparticle-coated heater tested in pure R141b for all test operating pressures. This finding means that the 0.1 % nano-covering is thicker. Thus, any nano-covering that is easily removed during evaporation/boiling has a distinct effect on heat transfer. When the nanoparticle -coated micro-channel surface was tested in pure R141b, evaporation/boiling on the micro-channel surface may detach some of the nanoparticle-covering from the surface. As a result, thermal resistance decreases for the 0.1 % nanoparticle-coated micro-channel surface compared with the thermal resistance produced with 0.1 % nanofluids.

Fig. 10
figure 10figure 10

Heat flow density and thermal resistance comparison between bare microchannel/0.001 vol % Cu nanofluid and 0.001 vol % coated microchannel surface/pure R141b

Full size image
Fig. 11
figure 11figure 11

Heat flow density and thermal resistance comparison between bare microchannel/0.01 vol % Cu nanofluid and 0.01 vol % coated microchannel surface/pure R141b

Full size image
Fig. 12
figure 12figure 12

Heat flow density and thermal resistance comparison between bare microchannel/0.1 vol % Cu nanofluid and 0.1 vol % coated microchannel surface/pure R141b

Full size image

6 Conclusions

An experimental investigation was conducted on the thermal performance of a micro-channel surface used in a vapor chamber by using Cu-R141b nanofluid as the working fluid. A series of tests was conducted to investigate the heat transfer characteristic of nanoparticle deposition by using the base fluid of R141b. The conclusions obtained in this study can be summarized as follows:

  1. 1.

    The use of R141b-based Cu nanofluids with volume concentrations of 0.001 and 0.01 % as the working fluid could enhance the thermal performance of the micro-channel surface, especially under lower operating pressures that range from 0.86 × 105 Pa to 0.96 × 105 Pa. The working fluid with nanoparticles could enhance the thermal performance of the micro-channel surface because of the following reason: The nanoparticles deposited at the heat transfer surface and formed a nanoparticle layer, thus extending the heat transfer area and increasing nucleation density to improve evaporation/boiling heat transfer. The extended heat transfer area and nucleation density increases simultaneously with increased nanoparticle concentration that could result in enhanced heat transfer effect. Optimal nanoparticle suspension was noted with 0.01 % volume concentration, which corresponds to minimum thermal resistance. Minimum thermal resistance is half that for R141b. Under some operating conditions, heat transfer deterioration was observed for nanofluid with 0.1 % volume concentration compared with R141b.

  2. 2.

    Operating pressure can significantly influence the heat transfer characteristic of micro-channel surface that use nanofluids. The maximum reduction effect of the thermal resistance was 50 %, which corresponds to a 0.01 % volume concentration of nanofluid at low operating pressure. The following quantitative reasons may explain such occurrence: (1) The properties of the working fluids can be influenced by pressure. Increased pressure causes variation in working fluid surface tension and latent heat of vaporization; (2) Nucleation density, bubble dynamics, and contact angle on the boiling surface can be affected by pressure. The conjunct role of the above enhancement factor determines the influence of pressure on heat transfer characteristics of nanofluids. The effect of thermal resistance of nanofluids for lower operating pressures is more apparent than that for higher operating pressures. A more evident deteriorating heat transfer effect is noted for the nanofluid with 0.1 % concentration compared with R141b at the operating pressure of 2.0 × 105 Pa.

  3. 3.

    Nanoparticle layers at the microchannel surface for nanofluids with different volume concentrations are observed. Experimental results for the thermal resistance of a nanoparticle deposition surface that used pure R141b showed that the formation of a nanoparticle layer at the microchannel surface is the main reason for the improved thermal performance of the microchannel surface that used nanofluids.