Applied Thermal Engineering 116 (2017) 56–65

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Visualization and comparative investigations of pulsating ferro-fluid

heat pipe
A. Gandomkar a, M.H. Saidi b,⇑, M.B. Shafii a, M. Vandadi a, K. Kalan a
a
School of Mechanical Engineering, Sharif University of Technology, Iran
b
School of Mechanical Engineering, Sharif University of Technology, P.O. Box: 11155-9567, Tehran, Iran

h i g h l i g h t s

Nanofluid was unstable and also agglomerated and deposited in the copper pipe; whereas, it was stable in the Pyrex pipe.

For the Pyrex heat pipe, applying the magnetic field caused weaker performance.

For Pyrex PHP, increasing the concentration improved the performance.

For the copper PHP, applying the magnetic field had the best performance.

For copper PHP, concentration of the nanofluid in each case of the magnet and input power has an optimum value.

a r t i c l e i n f o a b s t r a c t

Article history: Pulsating heat pipes (PHPs) are among the best solutions for the electronics cooling due to their low cost,
Received 3 October 2016 effectiveness and being passive. Experiments to study the effective factors on heat transfer performance
Revised 29 December 2016 have been designed and as a result, improvement of ferrofluid PHP performance has been achieved. Two
Accepted 17 January 2017
different heat pipes made of copper and glass were prepared to investigate the behavior of magnetic
Available online 19 January 2017
nanofluids. In order to find the best condition for heat transfer performance, different concentrations
of nanofluid with a filling ratio of 50% were tested in 3 different cases of magnetic field. The results indi-
Keywords:
cated that the ferrofluid is more stable in the glass PHP. It also shows that the presence of magnetic field
PHP
Ferrofluid
in the copper PHP has the best outcome while in the glass PHP, the absence of magnetic field results
Visualization better. It was detected that using a more concentrated ferrofluid causes a better performance in the
Flow regime copper PHP only when the magnetic field is applied, while in the glass PHP, increasing the concentration
Thermal resistance of ferrofluid improves the performance of the PHP in all conditions of applying the magnetic field.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Durability is one of the many advantages of PHPs; they require
no valve, power supply, mechanical pumps or any other external
tools, which makes them essentially vibration and noise free. PHPs
are also popular due to their less dependence on orientation angle
and gravity compared to other types of heat pipes. They have
different applications in industry such as water treatment,
desalination, oil and gas pipelines, dryers, PVTs and so forth.
Pulsating heat pipes are made of three different parts namely
evaporator, adiabatic section and condenser. The way PHPs work
is that the liquid in the evaporator evaporates by absorbing heat
and after passing through the adiabatic path, its temperature drops
⇑ Corresponding author.
E-mail address: saman@sharif.edu (M.H. Saidi).
in the condenser and it turns back into liquid phase. Afterwards,
the condensed liquid goes back into the evaporator through the
pipe wall. This process is repeated in a way that the continuous
cycle of steam and liquid causes heat transmission.
Basically, PHPs, in addition to sensible heat transfer by the dis-
placement of the liquid phase, use latent heat transfer mechanism
by continuous phase change. This permanent phase change creates
a pressure difference between the different sections of the heat
pipe. The sudden pressure difference causes the movement of the
liquid inside the PHP. Several factors contribute to this transforma-
tion, including temperature, applied heat power, pipe diameter,
filling ratio, number of turns, thermo-physical characteristics of
liquids and some others [1–7].
To increase the transmitted heat capacity of the working fluid
and to improve heat pipe performance, fluids with higher conduc-
tivity coefficient are better choices due to their better heat transfer

http://dx.doi.org/10.1016/j.applthermaleng.2017.01.068
1359-4311/Ó 2017 Elsevier Ltd. All rights reserved.
 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65
properties. Among different fluids, nanofluids, especially metal
nanofluids are the common choices. Several nanofluids have been
previously examined for this task, among which diamond [8,9],
iron oxide [10], gold [11], copper [12], copper-nickel [13], titanium
oxide [14], silver [14–16], silicon oxide [17,18] and aluminum
oxide [18–20] can be mentioned.
Taslimifar et al. [21] studied on ferrofluid in a copper heat pipe
and examined the effect of input power, ferrofluid concentration,
location of the magnet and the orientation angle on its perfor-
mance. In their experiment, it is shown that ferrofluid’s thermal
performance is better than that of distilled water. It has been inter-
preted that startup and stability condition of ferrofluid is better
with magnet, in which thermo-magnetic shifting plays an impor-
tant role. Their results represent that a high concentration of fer-
rofluid worsens the thermal performance, and the best thermal
performance is at an angle of 67.5°.
Mohammadi et al. [22–25], studied the copper heat pipe perfor-
mance with ferrofluid and indicated that using ferrofluid leads to a
better efficiency compared to using distilled water and enhance-
ment of the magnetic field improves the PHP performance. Addi-
tionally their results show that the best thermal performance
happens when the magnet is placed in the evaporator section. They
have emphasized that in the presence of magnetic field in the
evaporator, thermal performance improves by increasing charge
field or ferrofluid concentration. On the other hand, a reduction
in concentration causes a reduction in viscosity. Therefore, by
decreasing the concentration of ferrofluid, the PHP has a better
performance in the absence of magnetic field; however, taking
everything into consideration, PHP performance is remarkably bet-
ter in the presence of magnetic field and a higher concentration of
ferrofluid.
Yang et al. [26] studied two-phase flow in a quartz-glass PHP
using distilled water and SiO2 nanofluid as working fluids. In the
case of using distilled water, vertical flow was detected.
Meanwhile for the nanofluid, by increasing heat power, column
flow, slug-plug flow and annular flow formed in the heat pipe.
This condition caused the thermal resistance in the PHP to
reduce notably while using nanofluid instead of distilled
water.
Karthikeyan et al. [27], have used copper nanofluid in a PHP
made of Pyrex with five turns and a diameter of 2.5 cm. Their
results demonstrated that most of the copper particles deposited
in the evaporator after 24 h and the rest of the particles remained
in the other parts like condenser. In addition, thermal efficiency of
copper nanofluid PHP was 33% better than water PHP, due to the
stuck copper particles in the evaporator which improved the boil-
ing phenomena and heat transfer.
Bhuwakietkumjohn and Rittidech [28] studied internal flow
effects on heat transfer characteristics in a PHP by using ethanol
and a mixture of nano silver-ethanol. They investigated the flow
regime in different lengths of evaporator at various temperatures
with ethanol and silver-ethanol nanofluid. Their results indicated
that by using silver-ethanol nanofluid instead of pure ethanol,
bubbly flow changes to slug-plug flow and annular flow
transforms into dispersed bubble flow. It was also stated that nano
silver-ethanol has a better thermal performance compared to
ethanol.
Effect of the material that PHP is made of on the thermal perfor-
mance of nanofluids in PHPs has not yet been investigated. Also
there are few studies about visualization of nanofluids in PHPs.
Hence, in this study we tested two different PHPs, one made of
copper and the other made of glass, charged with magnetic
nanofluids at a filling ratio of 50%. The effects of input heating
power, nanofluid concentration and presence of a magnetic field
are visualized and discussed to facilitate better design of future
electronic coolers.
57
2. Experimental setup and test procedure
2.1. Nanofluid specifications
Ferrofluid is made based on the method described by Gan-
domkar et al. [29]. According to this method, tetrahydrate metal
(II), chloride (FeCl24H2O) and metal (III) chloride hexahydrate,
(FeCl36H2O) with the ratio of respectively 2–1 from Fe3+ and
Fe2+ were dissolved in 240 ml of distilled water. Then the chemical
deposition was developed between 70 and 90 °C and vigorous stir-
ring after adding NH4OH. In this ferrofluid, Cetyltrimethylammo-
nium bromide (C19H42BrN) or CTAB with a weight ratio of 1–14
of Fe oxide acts as the surfactants.
Fig. 1 shows the analysis of the metal oxide nano particles mea-
surements by a Particle Size Analyzer (PSA) device. The size of the
particles is less than 10 nm as it is evident in Fig. 1. The stability of
the nanofluid was checked for five hours and during this interval,
no change was observed and nanofluid remained stable.
All nanofluids were developed with adding the surfactant and
placed in the ultrasonic mixer device and then injected into the
PHPs.
2.2. Experimental setup
An overview of the experimental setup is shown in Fig. 2a. Two
heat pipes, one made of copper and the other made of Pyrex were
prepared. Each one was one-turned with internal and external
diameters of 2 mm and 4 mm, respectively. The height of the cop-
per PHP was 45 cm, while the lengths of the evaporator, condenser
and the adiabatic section were 10 cm, 15 cm and 20 cm, respec-
tively. For the Pyrex PHP, lengths of the evaporator, condenser
and adiabatic section are 12 cm, 12 cm and 20 cm, respectively.
In order to observe the flow regime in the copper heat pipe, a
small piece of Pyrex was used in the adiabatic section and silicone
glue was used for joint sealing (Fig. 2b). A direct current power
supply and a nickel-chrome wire element were used to supply heat
power in the evaporator. To reduce contact resistance, tubular wire
was tightly wrapped around the tube. The element wrapped
around the pipe develops a magnetic field due to having electric
current. To neutralize this field, two coils were wrapped around
the pipe in opposite direction beside each other. Moreover, two
layers of electrical insulation cover were used to avoid electrical
current transfer to the heat pipe so that it did not affect the nano-
fluid. A 16 ⁄ 10 ⁄ 5 cm3 box was used to dissipate heat from the
condenser. Cold water at the temperature of (±18 °C) entered the
condenser reservoir and in order to fill or empty the heat pipe,
an input pipe was used in the adiabatic section. Fig. 2-c shows a
photo of the whole setup.
In order to measure the temperature, K-type thermocouples
with an accuracy of ±0.5 °C were used. 3 and 2 thermocouples were
used in the evaporator and condenser sections, respectively. The
thermocouple’s tip was covered by silicone paste to provide a bet-
Fig. 1. Particle Size Analysis (PSA) of ferro particles.
58 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65

Fig. 2. Experimental setup.

ter contact and to minimize sensitivity and measurement error. A
high speed and high resolution camera was used to observe the
flow regime. Fig. 2a shows a schematic of the device and the loca-
tion of the thermocouples.
After vacuuming, the PHP was filled by the magnetic nanofluid
as the working fluid with a filling ratio of 50%. The nanofluid was
tested in five different concentrations of 1, 1.5, 2, 2.5 and 3 g/L in
the copper heat pipe, while it was tested in only 2 concentrations
of 1.5 and 2 g/L in the Pyrex heat pipe. A 10 ⁄ 6 ⁄ 3 cm3 magnet
with a strength of 0.5 T at its center was used in the evaporator
(Fig. 2b). Three different cases for the magnet field were planned.
In first case, stable nanofluid without presence of the magnet
was examined. Then for second case, the magnet field was applied
to evaporator section. Finally for the last case, the magnet field was
removed. Thus the impact of the absence, presence and removal of
the magnetic field was examined in the PHP in vertical position, in
which it received heat from the bottom and the results were dis-
cussed. Input heat to the system was increased in steps of 5 W
(for the copper tube) and 10 W (for the Pyrex tube) until the dry-
out power was reached and for each input power, data were
recorded after the temperatures reached the quasi-equilibrium
value. Due to the existence of constraints and in order to prevent
damage to the power system, the evaporator temperature was con-
trolled not to exceed 140 °C.
Two layers of ceramic fiber wool (6 cm thickness) were used to
cover the evaporator and the condenser for the insulation purpose.
Although thermal power reaching to the fluid was different from
the overall power which power source provided, the total power
 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65
is the most important parameter in the heat pipe design. The heat
pipe performance is measured by the heat resistance measure-
ments and heat resistance is defined as follows:
Te  Tc
R¼
Q in
where T e , T c and Q in are evaporator average temperature, condenser
average temperature and total input power to the PHP, respectively.
The overall input power was measured by a digital power meter
(HIOKI, Model: 3286-20) with an accuracy of 0.001 W.
Inaccuracy of the thermometer is 0.5 °C. Error analysis was con-
ducted by the technique suggested by Holman [30]. The maximum
overall inaccuracy is approximately 12%.
3. Results and observations
The effects of magnetic field and concentration of Nano-fluids
as the two controlling parameters have been studied under the
influence of the material comprising the setup.
3.1. Copper and Pyrex PHP observations
The input power was increased from 5 W to 30 W (maximum
possible power to prevent dry-out) in the copper PHP and from
10 W to 60 W in the Pyrex pipe. The flow pattern was recorded
in the adiabatic section.
In the first case, magnetic field was not applied to the evapora-
tor section and temperature data and flow regime were recorded
for different powers. After that, the power source was turned off
for cooling purposes and then a magnetic field was placed in the
evaporator section and the same procedure as the previous case
was conducted. After this examination, the magnetic field was
eliminated and data were recorded for different powers again. All
the concentrations were tested for these 3 cases. In the following
sections, observations and results related to each case are
presented.
Fig. 3. Instability of ferrofluid in the copper PHP with concentrations of (a) 1.5 g/L and (b) 2.5 g/L.
59
3.1.1. The absence of magnetic field
In Fig. 3, ferrofluid with a concentration of 1.5 g/L (Fig. 3a) and
2.5 g/L (Fig. 3b) is shown after being injected to the system. After a
while Fe particles became agglomerated and the nanofluid got
unstable. As temperature varied and oscillation started, it was
ð1Þ
observed that the concentration of the dispersed particles
decreased, because most of the particles deposited over time and
momentum itself was not capable of moving the particles. During
this period, the oscillating fluid could be considered as deionized
water.
It was observed that the transition of the nanofluid to instable
conditions speeded up and larger agglomerated particles were seen
while the nanofluid concentration increased. At the concentration
of 1.5 g/L, it took about 5 min for the nanofluid to lose its stability,
while for the concentration of 2.5 g/L, it took about 3 min. In other
words, the more concentrated the nanofluid gets, the faster the
deposition occurs and the nanofluid turns unstable [31]. The point
that should be noted is that in all concentrations the nano-fluid
was unstable and nano-particles deposited and only deionized
water oscillation was observed. Particles deposition causes a pres-
sure drop in the evaporator. Therefore, the momentum is not cap-
able of moving the particles into the condenser.
Unlike the copper PHP, by applying various powers, the nano-
fluid remained stable in the Pyrex PHP. After a while, the concen-
tration decreased, which means some particles had deposited. As
shown in Fig. 4, the particles deposited at the bottom of the slug
column right after being injected into the PHP; but, after receiving
heat for some time, the nanofluid started to oscillate and got back
its stability and reached a certain concentration which was suit-
able for the tests (It took more than 3 h.) as it could be assumed
that the concentration remained stable (Fig. 4).
In nanoscale consideration, there are 3 main forces acting on
nano or micro particles. The interaction of these forces gives stabil-
ity to the nanofluid. They are: (1) Electrical repulsive force, created
by nano particles with the same electrical charge. (2) Attraction
force generated by the high surface-to-volume ratio energy: as
the surface energy per particles volume is considered high for nano
and micro scales, the particles tend to stick together to balance and
60 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65
Fig. 4. Stability of ferrofluid in the Pyrex PHP with a concentration of 1.5 g/L.
by increasing the surface, decrease the surface energy per volume.
(3) Gravity is always present and the higher the particle’s density
is, the bigger the gravity reacts.
The only force acting to support the nanofluid stability is the
repulsive force from nanofluid’s electrical charge which is also
called zeta potential charge. The reason that the nano particles
were unstable and agglomerated in the copper PHP is that the elec-
trical charge vanished after being in contact with the copper.
Therefore, there were no more repulsive force between the nano
particles to keep the fluid stable.
The high density of the particles and the effect of gravity as well
as the high surface-to-volume ratio energy of the nanoparticles
cause the formation of bigger particles and make a high density
deposit at the bottom of the pipe. These two factors have the high-
est effect on their instability and deposition of nanofluid in the
copper pipe. But in the glass PHP, due to nanoparticles’ charges,
an opposite charge is created on the glass’s outer surface and a
layer of nanoparticles cover the glass’s surface and establishes
electrostatic balance. Thus, the nanofluid charges are developed
on the inner wall which causes the repulsion of particles. As glass
is not an electrical conductor, electrical discharge does not take
place and only a small amount of the particles cover the glass sur-
face and cause a decrease in the concentration of the nanofluid and
its stability [31].
The pattern of flow regime is strongly dependent on the applied
heat flux and by an increase in the power and therefore increasing
the rate of bubble formation, generation of bubbly flow, slug-plug
flow, churn flow, semi-annular and annular flows are seen in the
PHP respectively [32]. It is assumed that most of the particles are
deposited due to the high density of the particles. Boiling rate, bub-
ble generation rate and thereby heat transmission increase due to
the metal particles deposition, but also viscosity of the fluid in the
evaporator increases [21–25].
In the experiments of ferrofluid in the copper and Pyrex PHP, by
increasing power from 5 W to 25 W (copper PHP) or 10 W to 60 W
(Pyrex PHP), slug-plug, churn, semi-annular and annular flow were
observed (Fig. 5).

Ferrofluid behavior by concentration changes
o Table 1 indicates the flow pattern in the absence of magnetic
field in the copper PHP. It can be seen that the increase in the
concentration and thereby increase in the viscosity of the
fluid is problematic at lower powers. According to the obser-
vations, at lower powers (5 W), oscillation is delayed when
the concentration is increased, which causes the fluid to
move slower; as a result, no flow regime is seen for concen-
trations of 2.5 g/L and 3 g/L and it seemed the increase in the
boiling points, due to the nanofluid concentration, cannot
overcome the increased viscosity of the fluid. But by increas-
ing heat to 10 W, the bubble generation rate reaches a level
that the fluid becomes capable of overcoming the surface
tension force and viscosity dominates; thus, oscillation
occurs at all concentration values.
o Another conclusion that could be made is that an increase in
concentration to a certain amount has positive effects on
flow regime, and negative effects are observed by increasing
it more than that certain amount. As seen in Table 1, increase
in concentration causes changes in flow regime. In general,
concentration effect can be evaluated at two power levels.
Increase in pressure drop is a result of increase in concentra-
tion due to the particles’ accumulation in the evaporator. At
lower powers, the fluid’s momentum is low and it is more
difficult for the fluid to deal with the pressure drop, so up
to a certain amount of concentration increase, the flow
regime improves and afterwards pressure drop has a more
negative effect. At higher powers, since the fluid’s momen-
tum is high, fluid has more power to deal with the pressure
drop, so it can move easier.
Table 2 shows the flow regime in the absence of magnetic field
based on input power and concentration in the Pyrex pipe. As seen
in the table, as the concentration goes up, the flow regime becomes
better. It is also notable that increase in concentration increases
the bubble generation rate and the Nano-fluid stability in the Pyrex
pipe leads to a lower pressure drop. Therefore, it’s not as noticeable
as it was in the copper PHP.
3.1.2. Magnetic field application in the evaporator
By applying magnetic field to the evaporator of the copper PHP,
the deposited metal particles move and stick to the pipe wall due
to the presence of the magnetic field and will have a better
arrangement compared to the previous condition. Therefore, the
boiling point increases and more steam will be produced compared
to the previous case. The conductivity of the particles increases due
to the presence of magnetic field, As a result, boiling and heat
transmission occur better. On the other hand, as the copper parti-
cles are stuck to the pipe wall there is more space for the fluid and
the pressure drop decreases, hence the sum of these two effects
helps the fluid move freely.
Flow patterns of 1.5 g/L ferrofluid in the copper and Pyrex PHP by
increasing power are shown in Fig. 6. In the copper PHP (Fig. 6a), it
can be seen that applying the magnetic field causes faster flows at
lower powers compared to the state where magnetic field is absent.
This could be due to the increase in boiling points, higher conductiv-
ity, and better arrangement of the particles, compared with the con-
dition where pressure is dropped due to the particles’ density.
In the Pyrex heat pipe (Fig. 6b), by applying the magnetic field,
all the particles remained around the magnetic field (evaporator)
 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65 61

Fig. 5. Flow pattern of 1.5 g/L ferrofluid in the absence of magnet with increasing power in (a) copper and (b) Pyrex PHP.

Table 1
Flow pattern for copper PHP, in the absence of magnet for various concentrations of ferrofluid (N: no regime; B: bubbly flow; S-P: slug-plug flow; C: churn flow; S-A: semi-annular
flow; A: annular flow).

Concentration (g/L) 5W 10 W 15 W 20 W 25 W
0 N B S-P C S-A
1 S-P S-P C S-A A
1.5 S-P C S-A S-A A
2 S-P C S-A S-A A
2.5 N C S-A A A
3 N C C S-A S-A

Table 2
Flow pattern for the Pyrex PHP, in the absence of magnet for various concentrations of ferrofluid.

Concentration (g/L) 10 W 20 W 30 W 40 W 50 W 60 W
1.5 S-P S-P S-P C C C
2.5 S-P S-P C C C S-A

Fig. 6. Flow pattern in the presence of the magnet, with increasing power for (a) copper and (b) Pyrex PHP.

and the fluid movement moved a few particles across the adiabatic
section. Whereas, this did not happen in the copper PHP.

Study of the behavior based on the concentration of nanofluid
In the copper PHP, after the magnetic field was applied, increas-
ing the concentration caused the flow pattern to accelerate, so
slug-plug flow could be observed for the concentrations of 2.5, 3 g/
L, as metal particles stuck to the pipe wall and the pressure drop
decreased. The situation leads the fluid to move more freely while
the particles conductivity increases in the presence of the magnet
and the boiling operation as well as heat pipe performance improves.
As shown in Table 3, increase in the concentration causes the devel-
opment of semi-annular and annular flows at smaller powers.
62 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65

Table 3
Flow pattern for copper PHP, in the presence of the magnet for various concentration of ferrofluid.

Concentration (g/L) 5W 10 W 15 W 20 W 25 W
1 S-P C S-A S-A A
1.5 S-P S-A S-A A A
2 C S-A S-A A A
2.5 C S-A S-A A A
3 C S-A A A A

Table 4
Flow pattern for the Pyrex PHP, in the presence of magnet for various concentrations of ferrofluid.

Concentration (g/L) 10 W 20 W 30 W 40 W 50 W 60 W
1.5 S-P C C C C S-A
2.5 S-P C C C S-A S-A

As shown in Table 4 for the Pyrex PHP, increase in the concen-
tration, improves the flow regime. In this system no change is seen
at the lower powers while a concentration of 2.5 g/L is applied and
only the size of the bubbles increases, which means the boiling rate
has increased. At higher powers such as 50 W and 60 W, semi-
annular and churn regimes are developed.
3.1.3. Elimination of the magnetic field
After the magnetic field is eliminated, in the copper PHP, the
momentum and the fluid movement cannot move the deposited
particles and the flow regime is similar to the previous conditions.
As seen in Table 5, in the condition where magnet field is
removed, increase in the concentration to a certain amount
improves the flow pattern and speeds up the flow at lower powers
and afterwards, flow drop occurs. As indicated, at both low and
high powers, just for a certain amount of concentration, increasing
the nanofluid’s concentration acts positively whereas after some
point has a negative effect.
As can be seen in Fig. 7, one of the differences between glass and
copper set-up could be related to the behavior of the particles after
the magnetic field is eliminated. In the copper PHP, after the mag-
netic field is eliminated, the metal particles remain deposited.
Whereas, in the Pyrex PHP, metal particles move with the water
in agglomerated form.
For the Pyrex heat pipe, the flow regime at the powers of 10–
30 W is slug-plug and by the increase of the power from 10 to
30 W, the speed and size of the bubbles increase. Compared with
the case where magnet is present, the speed and size of the bubbles
decreases due to the increase in fluid viscosity as particles are free
in the fluid the boiling points decrease due to the particles separa-
tion from the evaporator. At the range of 40–60 W, the prevailing
regime is churn flow which turns to semi-annular when the power
is increased to 60 W. Flow patterns in Pyrex PHP after removal of
the magnet for 2 concentrations are mentioned in Table 6.
3.2. Analysis of thermal performance of the PHP
3.2.1. Effect of the magnetic field
As mentioned in the preceding sections, the magnetic nanofluid
was examined in 3 conditions of magnetic field and 5 various con-
centrations in the copper PHP and in 3 conditions of magnetic field
and 2 different concentrations for Pyrex PHP. Thermal resistance of
the copper PHP in 3 different conditions of the magnetic field at the
concentration of 1.5 g/L based on input power was traced. As seen
in Fig. 8, the presence of magnetic field causes the best perfor-
mance for the heat pipe and the absence of magnetic field causes
the worst condition possible. This procedure is similar for all con-
centrations and magnetic field conditions.
As seen in Fig. 8, the least thermal resistance and thereby the
best PHP performance is due to the presence of magnetic field.
As the nanofluid became unstable and metal particles deposited
in the copper PHP, the presence of magnetic field caused the parti-
cles to stick to the pipe wall and as a result the boiling point
increased and also the conductivity of ferro particles increased
due to the presence of the magnet, therefore boiling and heat
transfer improved. Since the particles of the nanofluid stuck to
the wall of the pipe, the fluid moved in a smoother path; hence,
the best thermal resistance happens where the magnetic field is
present.
Considering Fig. 8, at lower powers, the performance after the
magnetic field was eliminated is similar to when the magnetic field
was present. It seems like the speed and the momentum of the
fluid did not have the power to separate all the stuck particles from
the evaporator wall, as movement of the particles from the adia-
batic section was not observed after the elimination of the mag-
netic field.
According to Fig. 8, by increasing the power and thereby the
speed of the fluid, the behavior of the PHP after the elimination
of the magnetic field was similar to the performance when the
magnetic field was absent. This behavior could be due to the sep-
aration of particles form the pipe wall and their movements in
the evaporator section, which causes the pressure drop to increase.
In both conditions of absence and elimination of the magnetic field,
there was weaker performance in developing bubbles compared to
the presence of magnetic field.
Since magnetic field has a remarkable influence on the arrange-
ment of the particles, it is predicted that after the elimination of
the magnetic field, based on momentum, the arrangement of par-
ticles changes; therefore, thermal performance deteriorates.
Another reason to this fact is the thermal resistance behavior for

Table 5
Flow pattern for the copper PHP, after the elimination of the magnet for various concentrations of ferrofluid.

Concentration (g/L) 5W 10 W 15 W 20 W 25 W
1 S-P C S-A S-A S-A
1.5 S-P C S-A S-A S-A
2 C S-A S-A S-A A
2.5 C S-A S-A S-A A
3 S-P C S-A S-A S-A
 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65 63

Fig. 7. Effect of elimination of the magnet on nano particles in (a) copper PHP and (b) Pyrex PHP.

Table 6
Flow pattern for the Pyrex PHP, after removal of the magnet for various concentrations of ferrofluid.

Concentration (g/L) 10 W 20 W 30 W 40 W 50 W 60 W
1.5 S-P S-P S-P C C S-A
2.5 S-P S-P C C C S-A

Fig. 9. Thermal Resistance of the Pyrex PHP vs. input power for various magnet
conditions and 1.5 g/L ferrofluid.
Fig. 8. Thermal Resistance of the copper PHP vs. input power for various magnet
conditions and 1.5 g/L ferrofluid.

the Pyrex PHP. Nanofluid was somehow stable in this system, in

the three conditions of magnetic field, which in the lower powers,
the behavior of the magnetic field after elimination is similar to the
state where magnetic field was present and at higher powers, the
behavior is similar to when magnetic field was not applied.
As seen in Fig. 9, for the Pyrex PHP, the absence of magnetic
field has the best performance among the other conditions. The
reason of this behavior can be related to the ferrofluid stability in
contrast to the copper PHP, in which the nanofluid was unstable.
The concentration of ferrofluid decreased compared to its initial
condition which means the particles stuck or deposited. Thus,
the fluid was more stable and the conductivity was higher com-
pared to the condition where the magnetic field was present or
eliminated.
After the magnetic field elimination, there was a better perfor-
mance compared with the state where magnetic field was present.
64 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65

This is due to the distribution of the agglomerated metal particles

in the fluid and their movements in the PHP. As seen earlier,
agglomerated particles were able to move in the fluid after the
magnetic field was eliminated, due to the fluid momentum. Since
some of the particles were distributed in the PHP and some were
still in the evaporator, the boiling point was less than the case that
the magnetic field was present; therefore, its conductivity coeffi-
cient was less. But, it was more than the case that the magnetic
field was absent.

3.2.2. Nanofluid concentration effect

In this part, the effect of the nanofluid concentration in the cop-
per PHP will be analyzed in detail in 3 different conditions of the
magnet. For the Pyrex PHP, only two concentrations were exam- Fig. 11. Thermal resistance of Pyrex PHP vs. input power, for two concentrations in
ined and under all magnet conditions, the effect of concentration the absence of magnet.
was similar. As a result, concentration will be analyzed in only
one magnetic condition.
6

Absence of magnetic field 1 g/lit
1.5 g/lit
5
2 g/lit
According to Fig. 10, in the absence of magnet and at the lower

Thermal resistance (K/W)

2.5 g/lit
powers such as 5 W, by increasing the concentration, thermal 4 3 g/lit
resistance increases accordingly and the heat pipe performance
worsens. Because in this power, the steam is weak; thus, by
3
increasing the concentration which leads to the deposition of the
particles, there is not enough power to move them. Deposited par-
2
ticles had no specific location and as a result the pressure drop
increased due to the presence of particles in the evaporator; there-
1
fore, the PHP did not perform well. As shown, in this condition, at
5 W, the best performance can be observed for the concentration of
1.5 g/L. By increasing the heat power and temperature of the evap- 0
0 5 10 15 20 25
orator, bubble generation rate increased and for the rest of the Input power (W)
powers, 2.5 g/L had the best performance.
Fig. 11 shows the effect of concentration in the Pyrex PHP in the Fig. 12. Thermal resistance of the copper PHP vs. input power, for various
concentrations in the presence of the magnet.
absence of magnet. By increasing concentration, thermal perfor-
mance of the Pyrex PHP improves, due to increased conductivity
and boiling points. Also in other magnet cases, increasing concen- 6
tration of the nanofluid, improved the performance of the PHP. 1 g/lit
1.5 g/lit
5

Applying the magnetic field 2 g/lit
Thermal resistance (K/W)

2.5 g/lit
4 3 g/lit
As the nanofluid was unstable in the copper PHP, the metal par-
ticles deposited, the magnetic field shifted the particles to the pipe
3
wall and the boiling points increased. On the one hand, it was
observed that the boiling phenomena improved by metal particles
2
improved as the concentration increased. On the other, since many
particles moved to the pipe wall, the fluid faced a rougher path to
1
move. Therefore, an efficient concentration was needed for this
condition.
0
0 5 10 15 20 25
10
Deionized water Input power (W)
9
1 g/lit
Fig. 13. Thermal resistance of copper PHP vs. input power, for various concentra-
Thermal resistance (K/W)

8 1.5 g/lit
tions after the removal of the magnet.
7 2 g/lit
6 2.5 g/lit

5 3 g/lit As can be seen in Fig. 12, the best thermal performance takes
4 place when the concentration is 2.5 g/L and in general, thermal
3 resistance decreased by increasing the concentration when the
2
magnetic field was applied, because the conductivity of magnetic
particles increases by applying the magnetic field, which leads to
1
a better performance in heat transmission and boiling.
0
0 5 10 15 20 25
Input power (W)
Elimination of the magnetic field

Fig. 10. Thermal resistance of copper PHP vs. Input power, for various concentra- In Fig. 13, the variation of the thermal resistance vs. heat power
tions in the absence of the magnet. for various concentrations can be seen.
 A. Gandomkar et al. / Applied Thermal Engineering 116 (2017) 56–65
This condition is very similar to the previous conditions at
lower powers for the following reasons:

Similarity of the flow regime

Metal particles staying stagnant or moving very slowly in the
evaporator
It is notable that the thermal resistance behavior in this condition
is similar to the previous one. The reason is that the low momentum
of the fluid was not able to separate the particles stuck to the pipe
wall. At higher powers, the flow regime was similar to the condition
where magnetic field was not applied. High fluid momentum pre-
vents the particles from sticking to the pipe wall and the regime is
not similar to when the magnetic field is applied anymore.
4. Conclusion
For a better understanding and evaluation of the performance of
PHPs using nanofluid, two different heat pipes from copper and
Pyrex were prepared. In order to approach the best condition of
thermal resistance at each input power, different concentrations
of nanifluid at a filling ratio of 50% were surveyed. The experiments
took place in three conditions which were absence, presence, and
elimination of the magnetic field. The results were compared
subsequently.
By analysis of the results, the following conclusions can be
made:

Nanofluid was unstable and also agglomerated and deposited in
the copper pipe; whereas, it was stable in the Pyrex pipe.

The influence of magnetic field and nanofluid concentration on
heat pipe performance strongly depended on the material of the
heat pipes and nanofluid’s stability. For the Pyrex heat pipe
while the condition of nanofluid was stable, applying the mag-
netic field caused weaker performance. On the other hand, for
the copper heat pipe, while nanofluid was unstable, applying
the magnetic field improved the performance.

The results of the copper PHP tests indicated that higher concen-
trations caused severe performance drop at lower powers and in
the case where magnetic field is not present. However, the per-
formance improved by applying the magnetic field at all concen-
trations and heat powers. The magnetic field application
speeded up the development of the regimes at lower powers.

In the copper PHP, the increase in the concentration showed a
different behavior which depended on heat power and magnet
field conditions; as increase in the concentration led to increase
in the boiling phenomena and pressure drop, which is why an
optimum concentration value was defined.

At lower powers, the performance of the copper PHP when the
magnetic field was eliminated, was similar to the condition that
the magnetic field was applied; additionally, at higher powers,
it was similar to when the magnetic field was absent. In the
Pyrex PHP, when the magnetic field was eliminated, the perfor-
mance quality was between the other two conditions.
Acknowledgement
Authors are grateful as this work has been supported by INSF.
References
[1] P. Charoensawan, S. Khandekar, M. Groll, P. Terdtoon, Closed loop pulsating
heat pipes: Part A: parametric experimental investigations, Appl. Therm. Eng.
23 (2003) 2009–2020.
[2] M.B. Shafii, A. Faghri, Y. Zhang, Thermal modeling of unlooped and looped
pulsating heat pipes, J. Heat Transfer 123 (2001) 1159–1172.
65
[3] M. Ebrahimi, M. Shafii, M. Bijarchi, Experimental investigation of the thermal
management of flat-plate closed-loop pulsating heat pipes with
interconnecting channels, Appl. Therm. Eng. 90 (2015) 838–847.
[4] S. Khandekar, N. Dollinger, M. Groll, Understanding operational regimes of
closed loop pulsating heat pipes: an experimental study, Appl. Therm. Eng. 23
(2003) 707–719.
[5] T. Sukchana, C. Jaiboonma, Effect of filling ratios and adiabatic length on
thermal efficiency of long heat pipe filled with R-134a, Energy Procedia 34
(2013) 298–306.
[6] H. Barua, M. Ali, M. Nuruzzaman, M.Q. Islam, C.M. Feroz, Effect of filling ratio
on heat transfer characteristics and performance of a closed loop pulsating
heat pipe, Procedia Eng. 56 (2013) 88–95.
[7] H. Han, X. Cui, Y. Zhu, S. Sun, A comparative study of the behavior of working
fluids and their properties on the performance of pulsating heat pipes (PHP),
Int. J. Therm. Sci. 82 (2014) 138–147.
[8] H. Ma, C. Wilson, Q. Yu, K. Park, U. Choi, M. Tirumala, An experimental
investigation of heat transport capability in a nanofluid oscillating heat pipe, J.
Heat Transfer 128 (2006) 1213–1216.
[9] C. Wilson, B. Borgmeyer, R. Winholtz, H. Ma, D. Jacobson, D. Hussey, M. Arif,
Visual observation of oscillating heat pipes using neutron radiography, J.
Thermophys. Heat Transfer 22 (2008) 366–372.
[10] H.R. Goshayeshi, I. Chaer, Comparison of copper and glass oscillating heat
pipes with Fe2O3 under magnetic field, Int. J. Low-Carbon Technol. (2015),
ctv021.
[11] P. Cheng, S. Thompson, J. Boswell, H. Ma, An investigation of flat-plate
oscillating heat pipes, J. Electron. Packag. 132 (2010) 041009.
[12] R.R. Riehl, N. dos Santos, Water-copper nanofluid application in an open loop
pulsating heat pipe, Appl. Therm. Eng. 42 (2012) 6–10.
[13] K. Park, H. Ma, Nanofluid effect on the heat transport capability in a well-
balanced oscillating heat pipe, J. Thermophys. Heat Transfer 21 (2007) 443–
445.
[14] H. Jamshidi, S. Arabnejad, M.B. Shafii, Y. Saboohi, Thermal characteristics of
closed loop pulsating heat pipe with nanofluids, J. Enhanc. Heat Transf. (2011)
18.
[15] S. Wannapakhe, S. Rittidech, B. Bubphachot, O. Watanabe, Heat transfer rate of
a closed-loop oscillating heat pipe with check valves using silver nanofluid as
working fluid, J. Mech. Sci. Technol. 23 (2009) 1576–1582.
[16] Y.-H. Lin, S.-W. Kang, H.-L. Chen, Effect of silver nano-fluid on pulsating heat
pipe thermal performance, Appl. Therm. Eng. 28 (2008) 1312–1317.
[17] Q.-M. Li, J. Zou, Z. Yang, Y.-Y. Duan, B.-X. Wang, Visualization of two-phase
flows in nanofluid oscillating heat pipes, J. Heat Transf. 133 (2011) 152211.
[18] J. Qu, H. Wu, Thermal performance comparison of oscillating heat pipes with
SiO2/water and Al2O3/water nanofluids, Int. J. Therm. Sci. 50 (2011) 1954–
1962.
[19] Y.-H. Hung, T.-P. Teng, B.-G. Lin, Evaluation of the thermal performance of a
heat pipe using alumina nanofluids, Exp. Thermal Fluid Sci. 44 (2013) 504–
511.
[20] H.R. Seyf, S. Kim, Y. Zhang, Thermal performance of an Al2O3–water nanofluid
pulsating heat pipe, J. Electron. Packag. 135 (2013) 031005.
[21] M. Taslimifar, M. Mohammadi, H. Afshin, M.H. Saidi, M.B. Shafii, Overall
thermal performance of ferrofluidic open loop pulsating heat pipes: an
experimental approach, Int. J. Therm. Sci. 65 (2013) 234–241.
[22] M. Mohammadi, M. Mohammadi, M. Shafii, Experimental investigation of a
pulsating heat pipe using ferrofluid (magnetic nanofluid), J. Heat Transfer 134
(2012) 014504.
[23] M. Mohammadi, M. Mohammadi, A.R. Ghahremani, M.B. Shafii, N.
Mohammadi, Experimental investigation of thermal resistance of a
ferrofluidic closed-loop pulsating heat pipe, Heat Transfer Eng. 35 (2014)
25–33.
[24] M Mohammadi, M Mohammadi, AR Ghahremani, M Shafii, Experimental study
of the effects of ferrofluid on thermal performance of a pulsating heat pipe, in:
9th International Conference on Nanochannels, Microchannels, and
Minichannels, American Society of Mechanical Engineers, 2011, pp. 435–440.
[25] M. Mohammadi, M. Mohammadi, N. Jahani, M. Shafii, Experimental
investigation of the effects of ferrofluid on the performance of the pulsating
heat, in: Pipe Proceedings of IMECE2011 International Mechanical Engineering
Congress and Exposition, November, 2011, pp. 11–17.
[26] H. Yang, S. Khandekar, M. Groll, Performance characteristics of pulsating heat
pipes as integral thermal spreaders, Int. J. Therm. Sci. 48 (2009) 815–824.
[27] V. Karthikeyan, K. Ramachandran, B. Pillai, A.B. Solomon, Effect of nanofluids
on thermal performance of closed loop pulsating heat pipe, Exp. Thermal Fluid
Sci. 54 (2014) 171–178.
[28] N. Bhuwakietkumjohn, S. Rittidech, Internal flow patterns on heat transfer
characteristics of a closed-loop oscillating heat-pipe with check valves using
ethanol and a silver nano-ethanol mixture, Exp. Thermal Fluid Sci. 34 (2010)
1000–1007.
[29] E. Gandomkar, S. Sabbaghi, B. hemmateenejad, Removal dye pollutions &
turbidity of inorganic & organic medium using nanostructures Master
graduate thesis, Shiraz University, 2014.
[30] J.P. Holman, Experimental Methods for Engineers, McGraw-Hill, New York,
2001.
[31] G. Cao, Synthesis, Properties and Applications, Imperial College Press, London,
2004.
[32] S. Khandekar, P. Charoensawan, M. Groll, P. Terdtoon, Closed loop pulsating
heat pipes Part B: visualization and semi-empirical modeling, Appl. Therm.
Eng. 23 (16) (2003) 2021–2033.