Applied Thermal Engineering 194 (2021) 117072

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Effects of filling ratio of a long cryogenic Pulsating Heat Pipe

Maria Barba ∗, Romain Bruce, Florent Bouchet, Antoine Bonelli, Bertrand Baudouy
Université Paris-Saclay, CEA, Département des Accélérateurs, de la Cryogénie et du Magnétisme, 91191, Gif-sur-Yvette, France

ARTICLE INFO ABSTRACT

Keywords: In the development of cryogenic pulsating heat pipes for cooling superconducting devices, numerous progres-
Cryogenic pulsating heat pipes sive heat load tests have been performed at different filling ratios with neon, argon and nitrogen as working
Heat and mass transfer fluids. The tests have been carried on with a pulsating heat pipe system connected to an external buffer volume
Two-phase flow
during the entire test (open configuration) and with the pulsating heat pipe system isolated from any external
Nitrogen
volume (closed configuration). From this two group of tests, it has been possible to determine the minimum
Neon
Argon
filling ratio required to operate in stable conditions and also to determine the optimum filling ratio giving the
Filling ratio highest thermal performance of the system for each working fluid. This filling ratios are always below 50%,
revealing the key role of the fluid flow circulation in pulsating heat pipes, mainly ensured by the expansions
and contractions of the vapor parts, in the global heat transfer. In addition, the analysis of the temperatures
and pressure evolution of a progressive heat load test with a high filling ratio reveals that an excess of liquid
impedes the flow circulation and therefore reduces the heat transfer from the evaporator to the condenser of
the pulsating heat pipe.

1. Introduction the amount of liquid inside the PHP, see detailed description in Sec-
tion 3.2) is fixed, leaving this parameter less documented. Nevertheless,
Pulsating Heat Pipes, also known as PHP, are thermal links com- it is possible to find some studies related to the filling ratio in the field
posed of a single capillary tube with a serpentine shape. In the most of cryogenics.
common design, the tube connects an evaporator zone to a condenser For example, T. Mito et al. [3] have tested a vertical 16 cm long PHP
zone, separated in the middle by an adiabatic section, as illustrated in with 10 turns with an inner diameter of 0.78 mm using neon as working
Fig. 1. The maximum inner diameter of the capillary tube is defined fluid, at different filling ratios: 15.9% and 31.7% with the condenser
with the help of the Eötvös dimensionless number (Eö) [1], which fixed at 27 K, and 50.1%, 79.8% and 95.4% with the condenser fixed
considers the equilibrium between gravitational and capillary forces. at 26 K. During each test, the heat load at the evaporator section was
The working fluid in the tube operates at phase change conditions, gradually increased from 0 to 1.5 W. It appears that with a filling ratio
resulting in a two-phase flow composed of liquid slugs and vapor of 31.7%, the PHP provides the highest values of the equivalent (or
plugs surrounded by a thin liquid film which facilitates vapor plugs effective) thermal conductivity (defined below in Section 3.1), reaching
circulation. Permanent thermal instabilities create an oscillating flow
8000 W/m K at 1.5 W. Liang et al. have also studied this parameter in
allowing the heat to transfer from the evaporator to the condenser. The
different cryogenic pulsating heat pipes operating with neon. Firstly,
heat is transferred through the liquid parts, which have a higher density
they tested a vertical 16.5 cm long PHP with an inner diameter of
and heat capacity (sensible heat), and due to all the phase-changes
0.9 mm and 8 turns [6]. Increasing gradually the heat load at the
occurring simultaneously (latent heat) [2].
evaporator section from 0.51 to 4.93 W, different filling ratios were
During the last decade the interest in pulsating heat pipes has
tested: 13.6%, 24.5%, 33 %, 40.4% and 49.5%. In this case, the highest
increased due to their lightness and low fluid consumption. Due to this,
they have been studied in the field of low temperatures as a thermal values of equivalent thermal conductivity were reached with a filling
link to be integrated in a cryogenic scheme for cooling superconducting ratio of 24.5%, going from 6000 W/m K at 0.51 W to 22 000 W/m K
magnets [3–5]. The majority of the studies focuses on the heat transfer at 4.93 W. Secondly, they tested a longer PHP (48 cm), also in vertical
performance or the maximum heat transfer capability of the system position, with 10 turns and an inner diameter of 1 mm [7]. Increasing
according to geometric parameters (length, diameter, number of turns the heat load at the evaporator section gradually from 2 to 26 W,
of the serpentine, etc.) and, generally, the filling ratio (which indicates the filling ratios tested were: 15.3%, 22.1%, 28.9%, 35.9%, 43.1%

∗ Corresponding author.
E-mail address: [email protected] (M. Barba).

https://doi.org/10.1016/j.applthermaleng.2021.117072
Received 15 January 2021; Received in revised form 6 April 2021; Accepted 4 May 2021
Available online 19 May 2021
1359-4311/© 2021 Elsevier Ltd. All rights reserved.
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 1. Schematic illustration of a Pulsating Heat Pipe.

and 51.5%. With 28.9%, they reach a maximum equivalent thermal thermally linked to a cryocooler. More details about the experimental
conductivity of 33 750 W/m K at 12 W but the system dries out at facility, the thermal shields and the support and protecting structures
16 W. Nevertheless, with a filling ratio of 35.9%, even if the maximum are indicated in [10, Ch. 2]. The pulsating heat pipe is instrumented
equivalent thermal conductivity was slightly lower, around 32 500 with pressure and temperature sensors distributed throughout the evap-
W/m K, the system was able to transfer up to 26 W. Actually, with orator, condenser and adiabatic sections, as indicated in Fig. 2. More
a filling ratio of 35.9%, between 2 and 16 W the equivalent thermal details about the integration of the sensors in the experimental facility
conductivity increases at each power step, going from 23 000 to 32 500 are indicated in [10, Ch. 2]. Globally, the temperature sensors are
W/m K and then decreases after 16 W reaching 31 750 W/m K at 26 W. thermalized at the surface of the tubes (adiabatic section) and inserted
They also tested a similar PHP, with a bended adiabatic section, and in the copper plates (evaporator and condenser). The pressure sensors
a total length of 70 cm [8]. This latter also was operated in vertical are inserted in specific insert pieces so that the tip of the sensor is
position. With 10 turns and an inner diameter of 1 mm, this system was adjusted to be flush with the inside diameter of the tube. In that way,
tested at the following filling ratios: 15.2%, 23.3%, 30.7% and 38.5%. no part obstructs the flow and no extra dead volumes is added for
Only when operating with a filling ratio of 30.7%, the PHP was able this measurement. The typical measurement error, including the data
to transfer the maximum applied heat load at the evaporator: 36 W, acquisition electronic chain, is ± 0.25 K for the Pt100 sensors, ±25 mK
reaching an equivalent thermal conductivity 30 000 W/m K. Finally, for the Cernox sensors and ±0.05% of the measurement for the Kulite
Fonseca et al. [9] tested a cylindrical PHP with 40 turns, an inner pressure sensors. For the error of the power is estimated to be 2.5% of
diameter of 0.5 mm and operating in vertical position. After testing the measurement and this also includes the error propagation by the
various filling ratios between 10 and 40% with nitrogen as working electronic chain.
fluid and a heat load at the evaporator going from 1 to 4.5 W, they
noticed that they reached the highest equivalent thermal conductivity 2.2. Filling procedure and test description
of 70 000 W/m K at a filling ratios around 20%.
From these studies, it can be noticed that high equivalent thermal As mentioned in the introduction, the PHP has been tested with
conductivities are given by relatively low filling ratios. To confirm two-phase neon, two-phase argon and two-phase nitrogen. To fill the
these tendency with a higher number of experimental data, numerous PHP with the different working fluids and empty it after each test,
progressive heat load tests have been performed in a meter-scale pul- a gas supply circuit has been conceived (as is presented in Fig. 3).
sating heat pipe operating with neon, argon and nitrogen as working This latter is composed of a pressure transducer, three gas tanks with
fluids, and are presented in this manuscript. The evolution of the the different working fluids and their corresponding buffer volumes, a
equivalent thermal conductivity during these progressive heat load pumping system and several valves and pressure regulators.
tests with different filling ratios is analyzed in order to determine the At the beginning of every test, the PHP and gas supply circuit are
minimum filling ratio required to operate in stable conditions, and the vacuumed down in order to remove impurities such as air or other
optimum filling ratio, giving the highest thermal performance or the unwanted gases. Then, after closing all the valves, the temperature of
maximum heat load transfer for each cryogenic fluid. We propose in the condenser is fixed at 27 K when using neon, 89 K when using
this manuscript an extensive study of the effect of the filling ratio on argon and 75 K when using nitrogen, with a temperature controller.
the heat performance of a 1 m long pulsating heat pipe in a horizontal The evaporator is heated at 5 W until a temperature difference of
orientation. 5 K between the evaporator and the condenser is reached, promoting
fluid oscillations inside the PHP during the filling process [11]. At this
2. Experimental conditions point, the filling procedure starts: firstly, the buffer volume is filled to
a desired pressure using an absolute pressure transducer (APT) that
2.1. Experimental facility will determine the amount of fluid and the filling ratio in the PHP.
The gas tank’s pressure regulator (PR) can be closed once the buffer
The experimental facility is composed of a one-meter long hori- volume and gas supply circuit is filled to a desired pressure. Then the
zontal pulsating heat pipe (see Fig. 2) composed of a stainless steel PHP can be filled by opening valve Va7. The pressure difference in
capillary tube having an inner diameter of 1.5 mm and bent in 36 the buffer volume between the beginning and the end of the filling
turns. The condenser, adiabatic and evaporator sections have a length process determines the quantity of fluid injected into the PHP. The
of 330 mm each and the fluid is introduced into the PHP through the evaporator heater delivers heat during this filling process. After a few
condenser section. A heater is fixed at the surface of the evaporator minutes, the temperature of the evaporator stabilizes meaning that the
section to control the amount of heat transfer. The condenser section is PHP is sufficiently filled with working fluid and able to transfer the heat

2
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 2. Schematic illustration of the experimental facility and locations of the sensors. Types of sensors: Pt100 temperature sensor, 1070 CX-SD Cernox temperature sensor
and Kulite pressure sensor.

Fig. 3. Schematic illustration of the gas supply circuit connected to the PHP with the corresponding lengths and diameters.

from the evaporator to the condenser, i.e. operating in a stable PHP been calculated at each power step. In this section, a detailed descrip-
mode. During this stable PHP mode, the temperature of the evaporator tion of these two parameters is given.
keeps stable while a certain amount of heat load is applied at its
surface, as detailed in [5], meaning that this heat is being transferred 3.1. Equivalent thermal conductivity
to the condenser. At this point, two options are possible to operate
the PHP: “open configuration” and “closed configuration”. The “open
The equivalent thermal conductivity 𝜆𝑒𝑞 is determined by:
configuration” is when the PHP valve (Va7) and the buffer valve (Va3,
Va4 or Va5, depending on the operating fluid used) are open during the 𝑄⋅𝐿
𝜆𝑒𝑞 = , (1)
entire PHP operation. The “closed configuration” is when valve Va7 is 𝐴 ⋅ 𝛥𝑇
closed during the entire PHP operation. In both cases, the heaters will where 𝑄 (W) is the input power at the surface of the evaporator, 𝐿
be operated with same heat load increments and for all three working (m) is the length of the tubing part (990 mm and 10 mm for the
fluids. The heat load will be ramped up at increments of 5 W every 30 capillary sleeves connecting the different tubing parts), 𝐴 (𝑚2 ) is the
to 60 min. cross-sectional area of the tubes (only the inner diameter is considered
due to the low thermal conductivity of stainless steel at the operating
3. Parameters studied temperatures) and 𝛥𝑇 (K) is the temperature difference between the
evaporator and the condenser. Since the maximum difference between
To determine the relation between the amount of fluid inside the the measurements of all the sensors of the evaporator is lower than
PHP and its global thermal performance during progressive heat load 0.1 K (and the same for the sensors of the condenser), the temperature
tests, the equivalent thermal conductivity and the filling ratio have difference 𝛥𝑇 is given by measurements from one sensor located at

3
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 4. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in open configuration using neon as working fluid at different
initial filling ratios.

the center of the top surface of the evaporator and one sensor located The liquid filling ratio 𝐹 𝑅(%), also known as the filling ratio, is
at the center of the top surface of the condenser. These temperatures determined by:
are averaged in time at each power step when they reach a stable 𝑉𝑙
value after the transient regime. The equivalent thermal conductivity 𝐹𝑅 = ⋅ 100, (2)
𝑉𝑝ℎ𝑝
is always calculated when a steady-state regime is reached that the
temperature of the evaporator reaches a stable value at each power where 𝑉𝑙 is the volume of the liquid in the PHP and 𝑉𝑝ℎ𝑝 is the volume of
step. the PHP. It is obtained from the following mass conservation equation:
We have computed the error of the determination of the equivalent
thermal conductivity using the propagation error equation with inde- 𝜌1 𝑉2 − 𝜌2 𝑉2 = 𝜌4 𝑉4 + 𝜌5 𝑉5 + 𝜌𝑙 𝑉𝑙 + 𝜌𝑣 𝑉𝑣 , (3)
pendent variables, i.e., no covariant terms [12]. We have considered where 𝑉2 = 𝑉1 + 𝑉3 , 𝜌1 𝑉2 is the mass of the volume 𝑉2 before the filling
the error on the length of the tubing part 𝐿 to be 2 mm and the one process, 𝜌2 𝑉2 is the mass left in the volume 𝑉2 after the filling process,
on the diameter about 0.05 mm. These values are conservative since 𝜌4 𝑉4 is the mass of the intermediate volume 𝑉4 inside the cryostat, 𝜌5 𝑉5
the PHP length is known at less than 2 mm and that the tube of our
is the mass of the volume 𝑉5 of the tube thermalized at the surface of
PHP has been processed by cold drawing technique. The relative error
the thermal link between the evaporator and the cryocooler, 𝜌𝑙 𝑉𝑙 is
is dominated by the error on the diameter and is almost independent of
the liquid mass in the PHP and 𝜌𝑣 𝑉𝑣 is the vapor mass in the PHP,
the temperature increase. It spans from 7.15% for a 𝛥T of 5 K to 7.12%
being 𝑉𝑙 + 𝑉𝑣 = 𝑉𝑝ℎ𝑝 the total volume of the PHP. The inlet tube is
for a 𝛥T of 30 K.
included in the total volume of the PHP. All the mentioned volumes are
indicated in Fig. 3 (see the color version of this paper). The different
3.2. Filling ratio
densities are obtained using the NIST Refprop Database [13] knowing
As mentioned above, the quantity of fluid injected into the PHP the corresponding pressure and temperature at each section thanks to
is given by the pressure difference in the buffer volume between the the numerous sensors.
beginning and the end of the filling process. The method to determine Then, the amount of liquid inside the PHP can be determined by
the filling ratio and the insertion pressure is explained as follows (more 𝜌1 𝑉2 − 𝜌2 𝑉2 − 𝜌4 𝑉4 − 𝜌5 𝑉5 − 𝜌𝑣 𝑉𝑝ℎ𝑝
details are given in [10, Ch. 2]): 𝑉𝑙 = , (4)
𝜌𝑙 − 𝜌𝑣

4
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 5. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in open configuration using argon as working fluid at different
initial filling ratios.

Conversely, fixing the initial filling ratio (𝐹 𝑅), the insertion pres- example is given here for a test with an initial filling ratio of 50%
sure is obtained with NIST Refprop Database [13] knowing the tem- with nitrogen in the closed configuration. More details are given in [10,
perature and the density given by Ch. 3]. The maximum difference between both values goes from ±0.34%
at the beginning of the test to ±1.46% at the end of the test. For other
𝜌4 𝑉4 + 𝜌5 𝑉5 + 𝜌𝑙 ( 𝐹100
𝑅
𝑉𝑝ℎ𝑝 ) + 𝜌𝑣 (1 − 𝐹𝑅
)𝑉
100 𝑝ℎ𝑝
+ 𝜌2 𝑉2
𝜌1 = . (5) fluids and tests, the difference is in the same order and quite small. For
𝑉2 simplicity, the filling ratio at saturation conditions is considered as the
It is important to note that the calculation is performed assuming real filling ratio, as mentioned above.
that the entire PHP is at the temperature of the condenser (27 K when
using neon, 89 K when using argon and 75 K when using nitrogen) 4. Experimental results
and in saturation conditions [14]. Nevertheless, the temperature of the
adiabatic and the evaporator sections are higher than the condenser’s As mentioned in Section 2.2, during each progressive heat load test,
one, so the real filling ratio remains slightly different (lower) than the temperature of the condenser is fixed while the input power at the
the initial one. The real filling ratio is later calculated by using the surface of the evaporator is increased in 5-watts increments every 30 to
experimental data obtained during each test. 60 min. Consequently, at each power step, the pressure of the system, as
To estimate the uncertainty of the filling ratio, we have considered well as the temperature of the evaporator and the temperatures of the
two ways of computing it. The first way is to compute the filling ratio in adiabatic section increase before reaching stable values. After several
considering all the vapor and liquid parts at saturation conditions at the power steps, when the PHP reaches its heat transfer capacity limit, the
measured pressure. Secondly, the filling ratio is calculated considering temperature of the evaporator increases constantly without reaching a
all the vapor and liquid parts at the temperature of the evaporator and stable value (total dry-out of the evaporator described in [15]) and the
the condenser respectively, and always using the measured pressure. test is stopped. For example, experimental results of a single progressive
However, in reality in the PHP, the temperature of the vapor parts is heat load test using nitrogen as working fluid are given in [5], using
necessarily between the evaporator’s temperature and the saturation neon as working fluid are given in [16] and using argon as working
temperature and the temperature of the liquid parts is necessarily fluid are given in [10, Ch. 4].
between the condenser’s temperature and the saturation temperature. For the present study, the evolution of the real filling ratio (so-called
This gives the two possible limit values of the real filling ratio. An filling ratio) as well as the thermal conductivity have been plotted as

5
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 6. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in open configuration using nitrogen as working fluid at different
initial filling ratios.

conductivity as a function of the heat load applied at the evaporator

during progressive heat load tests performed at different initial filling
ratios in open configuration with neon, argon and nitrogen, respec-
tively. Both parameters have been calculated at each power step when
the PHP is operating in stable conditions. In all the figures, the initial
filling ratios are indicated in the legend of each figure. As mentioned in
Section 3.2, the initial filling ratio is estimated assuming that the entire
PHP is at the temperature of the condenser (27 K with neon, 89 K with
argon and 75 K with nitrogen) at the beginning of each test. The filling
ratio is calculated after each test with the experimental data obtained
and it is always lower than the initial one because the PHP is not at
the temperature of the condenser during the entire test. At each power
step, the temperatures of the adiabatic section increase, being higher
Fig. 7. Minimum filling ratios necessary to start to operate in stable conditions at
than the condenser’s one.
different heat loads with neon ( ), argon ( ) and nitrogen ( ) as working fluids. It can be seen in Fig. 4 that the filling ratio decreases when the
heat load increases with neon as working fluid. The last point of each
curve corresponds to the maximum heat load value giving a stable
PHP operation, meaning that at higher heat loads the evaporator dries
a function of the heat load applied at the evaporator for the three
out. From this data, it becomes possible to determine the minimum
working fluids in open and closed configuration and are presented
filling ratio necessary to transfer a certain amount of heat load in stable
hereunder.
conditions: a minimum filling ratio of 11% is necessary to transfer
15 W, a minimum filling ratio of 15.5% is necessary to transfer 20 W,
4.1. Open configuration a minimum filling ratio of 13.5% is necessary to transfer 25 W, a
minimum filling ratio of 19% is necessary to transfer 30 W and a
Figs. 4–6 presents the evolution of the filling ratio (obtained from minimum filling ratio of 20% is necessary to transfer 35 W. Concern-
experimental results) as well as the evolution of the equivalent thermal ing the evolution of the equivalent thermal conductivity, it can be

6
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 8. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in closed configuration using neon as working fluid at different
initial filling ratios. Rmk: the initial filling ratio of 20% is marked with a single point because dry-out occurred just after increasing the heat load to 10 W.

noticed that with an initial filling ratios between 20 and 50%, the mass is liquid and given by 𝐹 𝑅 × 𝜌𝑙 × 𝑉𝑝ℎ𝑝 . Considering that 𝑉𝑝ℎ𝑝 is
equivalent thermal conductivity increases, reaches a maximum peak constant and that the liquid density 𝜌𝑙 almost does not change with
and decreases again. Above an initial filling ratio of 60%, this latter pressure, mainly the filling ratio (𝐹 𝑅), so the proportion of liquid,
increases constantly until the end of the test. The highest equivalent diminishes due to this mass transfer to the external components. The
thermal conductivity achieved was 64 000 W/m K at a heat load of fact that the system continues to operate in stable conditions even if
35 W and a filling ratio of 23.5%. Fig. 5 shows that, when operating the amount of liquid in the PHP (filling ratio) diminishes more than the
with argon, the filling ratio also decreases when the heat load increases. amount of vapor, shows the adaptability of the PHP. This reveals the
In this case, the minimum filling ratios are: 20% for a heat load of key role of the fluid flow circulation, mainly ensured by the expansions
10 W, 24.5% for a heat load of 15 W, 30.5% for a heat load of 20 W and contractions of the vapor parts, in the global heat transfer of the
and 33.5% for a heat load of 25 W. With respect to the equivalent PHP [2], with small amounts of liquid slugs “carrying” the heat (by
thermal conductivities, the highest values achieved are 78000 and sensible heat and thermal advection) [10, Ch. 4]. This becomes more
79 000 W/m K, at 15 and 20 W and filling ratios of 24.5% and 30.5%, evident at higher heat loads and explains filling ratio values below 50%
respectively. When operating with nitrogen as working fluid, we find to transfer the maximum heat load and to reach the highest thermal
a similar behavior, with a filling ratio decreasing when increasing the conductivity, as observed by [3,6–8], for the three operating fluids (at
heat load, as illustrated in Fig. 6. In this case, a minimum filling ratio 23.5% with neon, at 30.5% with argon and at 18% with nitrogen).
of 16.5% is necessary to transfer 10 W of heat load, 22.5% to transfer In fact, if there is too much liquid, the pressure drop becomes too
20 W and 42% to transfer 30 W. Concerning the equivalent thermal important and the system needs a higher pressure difference between
conductivities, the highest values are reached at lower filling ratio of the evaporator and the condenser for the fluid movement. Nevertheless,
18% and heat loads as low as 5 W. But with such a low filling ratio, the in some cases, when the filling ratio is too low according to the
PHP system dries-out early during the test, just after applying 10 W. applied heat load, the amount of liquid is not enough (necessity of a
During open configuration, the filling ratios of all the three working minimum filling ratio) and the PHP dries out. Concerning the minimum
fluids decrease as the heat loads increase. In fact, when increasing the filling ratio, it can be seen in Fig. 7 that these minimum values are
heat load, the temperatures and the operating pressure inside the PHP higher when using argon, followed by nitrogen and neon. For example,
increase, leading a certain mass transfer from the PHP to the gas supply at 20 W, the minimum filling ratio necessary to operate in stable
circuit and the buffer volume. In the PHP, the major part of the total conditions is 30.5%, 22.5% and 15.5% for argon, nitrogen and neon,

7
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 9. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in closed configuration using argon as working fluid at different
initial filling ratios. Rmk: the initial filling ratio of 70% is marked with a single point because dry-out occurred just after increasing the heat load to 10 W.

respectively. With neon, the thermophysical properties that promote comparing also to the open configuration, for the same amount of heat
the fluid circulation (low surface tension and dynamic viscosity, high load, filling ratios are slightly higher for the three fluids. This is due to
coefficient (𝑑𝑝∕𝑑𝑇 )𝑠𝑎𝑡 [2,17] and low latent heat), as well as its high the fact that the system is closed when a heat load of 5 W is applied to
coefficient 𝜌𝑙 × 𝐶𝑝𝑙 ([10, App.𝐹 ]), indicate that the heat is mainly carry the evaporator section during the beginning of the test. The operating
by the liquid (sensible heat), needing less liquid (lower filling ratio) temperatures and pressure are lower than at higher heat loads, so there
to start to carry the same amount of heat. On the contrary, argon has is almost any mass transfer to the outside before closing the PHP.
a very low 𝜌𝑙 × 𝐶𝑝𝑙 coefficient (followed by nitrogen) which indicates Focusing on the experimental results with neon as working fluid,
that the heat transfer through the liquid parts is very poor, and higher it can be seen in Fig. 8 that the PHP reaches the maximum equivalent
amounts of liquid (filling ratio) are necessary to start to transfer the thermal conductivity between 5 and 25 W with a filling ratio of 19.5%,
heat. going from 40 000 to 58 000 W/m K, and between 25 and 55 W with a
filling ratio of 25%, going from 59 000 to 71 000 W/m K. The maximum
4.2. Closed configuration heat transferred was 50 W, which resulted in a thermal conductivity of
71 000 W/m K and at an “optimum” filling ratio of 25%. It is important
Figs. 8–10 presents the evolution of the filling ratio as well as the to remark that, during this test with a filling ratio of 25% and at
evolution of the equivalent thermal conductivity as a function of the 50 W, the temperature of the condenser increased from 27 to 28.2 K
heat load applied at the evaporator during progressive heat load tests because the limits of stable working conditions of the cold head of the
performed at different initial filling ratios (indicated in the legend of cryocooler were reached and the test was stopped at 55 W for security
the figures) in closed configuration with neon, argon and nitrogen as reasons (high operating pressure) but the limits of the PHP system itself
working fluids. Also in this case both parameters have been calculated were not reached, meaning that the evaporator part did not dry-out.
at each power step when the PHP is operating in stable conditions. When operating with argon as working fluid, similar observations can
First of all, as it can be noticed from the three figures that, in be made. Fig. 9 reveals that between 5 and 10 W, the PHP reaches the
closed configuration, the evolution of the filling ratio is very stable. In maximum equivalent thermal conductivity with a filling ratio of 20.5%
this case, contrary to the open configuration, when increasing the heat going from 57 000 to 75 000 W/m K. Between 10 and 30 W the PHP
load, the temperatures and the operating pressure increase but there is reaches the maximum equivalent thermal conductivity with a filling
no mass transfer from the PHP to the exterior gas lines. Furthermore, ratio of 29.5% with a peak at 20 W of 84 000 W/m K. At a filling ratio

8
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Fig. 10. Evolution of the filling ratio and the equivalent thermal conductivity during progressive heat load tests in closed configuration using nitrogen as working fluid at different
initial filling ratios.

of 29.5%, the maximum heat is transferred and the maximum thermal The “optimum” filling ratios are all below 50%, which confirms
conductivity is reached, being this value the “optimum” filling ratio. what it has been observed from tests in open configuration: the im-
Fig. 10 illustrates the evolution of the filling ratio and the equivalent portant role of the vapor parts in the fluid flow circulation to ensure a
thermal conductivity when using nitrogen as working fluid. The PHP stable heat transfer. Furthermore, comparing these “optimum” filling
reaches the maximum equivalent thermal conductivity with a filling ratio values between the three fluids, even if nitrogen and argon
ratio between 17.5 and 18%, but the systems dries out at low heat present similar ranges of “optimum” filling ratios (29.5% for argon and
loads close to 10 W. Nevertheless, increasing the filling ratio, the between 27 and 42% for nitrogen), it is interesting to remark that, when
system is able to transfer higher amounts of heat load. For example, using neon, the “optimum” filling ratio is lower (25%). This can also
with a filling ratio of 27%, the PHP reaches an equivalent thermal be noticed when analyzing experimental results in open configuration
conductivity of 120 000 W/m K at 5 W to 95 000 W/m K at 20 W, and (see previous Section 4.1), where for the same heat load, the operating
the maximum heat load (25 W) is transferred in stable conditions with filling ratios are considerably lower than with the two other fluids.
an average filling ratio of 42% and an equivalent thermal conductivity Once again, since neon displayed the thermophysical properties that
of 60 000 W/m K. In this case, it is less evident to identify an “optimum” promote mass flow (low surface tension and dynamic viscosity, high
value, but it becomes clear that with filling ratios between 27 and coefficient (𝑑𝑝∕𝑑𝑇 )𝑠𝑎𝑡 [2,17] and low latent heat, making the fluid easy
42%, a reasonable equilibrium between a high thermal conductivity to change its phase), indicating that the heat is mainly carried by the
and a maximum heat load transfer can be reached. With argon as liquid (sensible heat), therefore it needs less liquid to carry the same
working fluid, as well as with nitrogen, at higher filling ratios, the heat or more.
system is able to transfer higher amounts of heat load but with lower It has already been mentioned that when the filling ratio is too
thermal conductivity, meaning that the maximum thermal performance low, the system dries out because the amount of liquid is not enough
of the PHP does not correspond to the maximum applied heat load. to transfer the heat. But it would be complementary to analyze the
This a direct consequence of the calculation of the equivalent thermal behavior of the PHP with an excessive amount of liquid (high filling
conductivity itself, where at some point the value of the temperature ratios). It can be seen in Fig. 11 the evolution of the PHP pressure,
difference between the evaporator and the condenser overcomes the the saturation temperature, the average temperatures of the evaporator
value of the applied heat load. This is not the case with neon, but, and the condenser and the temperatures of the adiabatic section of a
as mentioned above, the heat transfer capacity limit of the PHP is not progressive heat load experiment in closed configuration using argon
reached. as working fluid with a filling ratio of 43.8% (60% of initial filling

9
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

Nomenclature

Roman letters
A cross-sectional area m2
𝐶𝑝 specific heat J/(kg K)
𝐿 length m
𝑝 pressure Pa
𝑄 input power W
𝑇 temperature K
𝑉 volume m2

Greek letters
𝜆𝑒𝑞 equivalent thermal conductivity W/(m K)
𝜌 density kg/m3

Dimensionless numbers
Fig. 11. Evolution of the PHP pressure ( ), the saturation temperature ( ),
𝐸 𝑜̈ Eötvös number; 𝐸 𝑜̈ = 𝐷2 𝑔(𝜌𝑙 − 𝜌𝑣 )∕𝜎
the average temperatures of the evaporator ( ) and the condenser ( ) and the
temperatures of the adiabatic section ( ) of a progressive heat load experiment in Subscripts
closed configuration using argon as working fluid with a filling ratio of 43.8% (60%
initially).
𝑙 related to the liquid phase
𝑝ℎ𝑝 related to the pulsating heat pipe
𝑠𝑎𝑡 related to saturation conditions

ratio). This test is presented as an example of what it has been observed Acronyms
in experimental results from progressive heat load tests with high FR Filling Ratio
filling ratios with the three fluids. In this test, the temperature of the PHP Pulsating Heat Pipe
condenser oscillates around 89 K during the entire test and the heat
Declaration of competing interest
load applied at the evaporator goes from 5 to 10 W. It can be observed
that the temperature of the evaporator never stabilizes, increasing by
The authors declare that they have no known competing finan-
installments, indicating that the fluid stops to flow and the heat is
cial interests or personal relationships that could have appeared to
not transferred. This stop in the fluid circulation can be confirmed
influence the work reported in this paper.
by the stop of pressure oscillations (horizontal slope) and the stop of
temperature oscillations in the adiabatic section, suggesting an excess References
of fluid creating a pressure drop high enough to stop the flow.
Remark: At high filling ratios, usually higher than 50%, the temper- [1] E.T. White, R.H. Beardmore, The velocity of rise of single cylindrical air bubbles
ature of the evaporator hardly stabilizes, as showed in Fig. 11. Due to through liquids contained in vertical tubes, Chem. Eng. Sci. 17 (5) (1962)
this, it is important to mention that the equivalent thermal conductivity 351–361.
[2] Y. Zhang, A. Faghri, Advances and unsolved issues in pulsating heat pipes, Heat
must be considered merely as indicative since it is calculated from an
Transf. Eng. 29 (1) (2008) 20–44.
average temperature of the condenser.
[3] T. Mito, K. Natsume, N. Yanagi, H. Tamura, T. Tamada, K. Shikimachi, N.
Hirano, S. Nagaya, Development of highly effective cooling technology for a
superconducting magnet using cryogenic OHP, IEEE Trans. Appl. Supercond. 20
5. Conclusions (3) (2010) 2023–2026.
[4] L.D. Fonseca, F. Miller, J. Pfotenhauer, A helium based pulsating heat pipe for
superconducting magnets, in: Advances in Cryogenic Engineering, 1573, 2014,
The numerous progressive heat load tests performed at different pp. 28–35.
filling ratios with neon, argon, and nitrogen in open and closed config- [5] R. Bruce, M. Barba, A. Bonelli, B. Baudouy, Thermal performance of a meter-scale
uration reveal that a minimum filling ratio is needed to start operating horizontal nitrogen pulsating heat pipe, Cryogenics 93 (2018) 66–74.
the pulsating heat pipe in stable conditions and ensure a minimum [6] Q. Liang, Y. Li, Q. Wang, Experimental investigation on the performance of a
heat transfer. It also shows that the optimum filling ratio giving the neon cryogenic oscillating heat pipe, Cryogenics 84 (2017) 7–12.
highest thermal performance is always below 50%, revealing the key [7] Q. Liang, Y. Li, Q. Wang, Effects of filling ratio and condenser temperature on
the thermal performance of a neon cryogenic oscillating heat pipe, Cryogenics
role of the fluid flow circulation, mainly ensured by the expansions
89 (2018) 102–106.
and contractions of the vapor parts in the global heat transfer of the
[8] Q. Liang, Y. Li, Q. Wang, Study on a neon cryogenic oscillating heat pipe with
pulsating heat pipe. Furthermore, it can be noticed that these filling
long heat transport distance, Heat Mass Transfer 54 (2018) 1721–1727.
ratios values are considerably lower with neon than with the two other [9] L.D. Fonseca, F. Miller, J. Pfotenhauer, Experimental heat transfer analysis of a
fluids, due to its low flow resistance and its high heat transfer through cryogenic nitrogen pulsating heat pipe at various liquid fill ratios, Appl. Therm.
the liquid parts by sensible heat. Moreover, it is well known that when Eng. 130 (2018) 343–353.
the filling ratio is very low, the pulsating heat pipe dries out and [10] M.A. Barba Higueras, Study of Meter-scale Horizontal Cryogenic Pulsating Heat
the efficient heat transfer is not guaranteed, but the analysis of the Pipes (Ph.D. thesis), University of Paris-Saclay, 2019, http://www.theses.fr/

temperatures and pressure evolution during a progressive heat load 2019SACLS224.

[11] Y. Li, Q. Wang, S. Chen, B. Zhao, Y. Dai, Experimental investigation of the
test with a high filling ratio reveals that excessive amounts of liquid
characteristics of cryogenic oscillating heat pipe, Int. J. Heat Mass Transfer 79
cause the stop of the flow circulation interrupting the heat transfer.
(2014) 713–719.
From all of this, we can conclude that a suitable filling ratio is given [12] P.R. Bevington, D.K. Robinson, Data Reduction and Error Analysis for the
by liquid/vapor proportion that avoids to have stop flows and to dry Physical Sciences, second ed., WCB McGraw-Hill, 1992.
out the pulsating heat pipe system, and this suitable value is highly [13] NIST Reference Fluid Thermodynamic and Transport Properties Database
dependent on the physical properties of the working fluid. (REFPROP): Version 9.1.

10
M. Barba et al. Applied Thermal Engineering 194 (2021) 117072

[14] L.D. Fonseca, F. Miller, J. Pfotenhauer, Design and operation of a cryogenic [16] M. Barba, R. Bruce, F. Bouchet, A. Bonelli, B. Baudouy, Thermal Study of a
nitrogen pulsating heat pipe, in: Advances in Cryogenic Engineering, Vol. 101, one-meter long Neon Cryogenic Pulsating Heat Pipe, in: IOP Conference Series:
2015. Materials Science and Engineering, 27th International Cryogenic Engineering
[15] M. Barba, R. Bruce, A. Bonelli, B. Baudouy, Experimental study of large-scale Conference and International Cryogenic Materials Conference.
cryogenic pulsating heat pipe, in: 21st Cryogenic Engineering Conference and [17] R. Bertossi, V. Ayel, B. Mehta, C. Romestant, Y. Bertin, S. Khandekar, Motion
International Cryogenic Materials Conference, IOP Conf. Ser.: Mater. Sci. Eng. of liquid plugs between vapor bubbles in capillary tubes: a comparison between
278 (2017) 012156. fluids, Heat Mass Transfer 53 (11, SI) (2017) 3315–3327.

11