See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323237638

Portable thermoelectric power generation based on catalytic combustor for

low power electronic equipment

Article  in  Applied Energy · April 2018

DOI: 10.1016/j.apenergy.2018.02.012

CITATIONS READS

4 122

6 authors, including:

C. Fanciulli Hossein Abedi

Italian National Research Council Italian National Research Council
54 PUBLICATIONS   425 CITATIONS    4 PUBLICATIONS   5 CITATIONS   

SEE PROFILE SEE PROFILE

Laura Merotto Roberto Donde

Empa - Swiss Federal Laboratories for Materials Science and Technology Italian National Research Council
42 PUBLICATIONS   170 CITATIONS    39 PUBLICATIONS   178 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

LII measurements View project

shape memory alloy View project

All content following this page was uploaded by Hossein Abedi on 25 May 2018.

The user has requested enhancement of the downloaded file.

 Applied Energy 215 (2018) 300–308

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Portable thermoelectric power generation based on catalytic combustor for T

low power electronic equipment
⁎
C. Fanciullia, , H. Abedia, L. Merottob, R. Dondèb, S. De Iuliisb, F. Passarettia
a
CNR – Institute of Condensed Matter Chemistry and Technologies for Energy, Via Previati 1/E, 23900 Lecco, Italy
b
CNR – Institute of Condensed Matter Chemistry and Technologies for Energy, Via Cozzi 53, 20125 Milano, Italy

H I G H L I G H T S

• New TEG based on catalytic combustor is developed with a volume close to AA battery.
• The system is designed to be safe, accessible for portable use as electronic supplier.
• Our solution shows an alternative for low power supply open to further improvements.
• Characterization procedure is defined allowing deep understanding of the processes.

A R T I C L E I N F O A B S T R A C T

Keywords: In recent years, the portable technology is receiving a great interest and significant improvement due to the
Thermoelectric generator progresses in electronic technology development and energy storage solutions. The decrease in power require-
Catalytic combustor ments for working energy systems, due to the increased efficiency and to the reduction in components size, opens
Thermoelectric the access to new solutions for power supplying. In particular, alternative backup systems for battery charging or
Device
replacement could be designed taking advantage of unconventional technologies. It is the case of small pho-
tovoltaic portable panels or fuel cells technology: in these solutions different sources are used to produce limited
electrical powers required to keep devices on. In this paper, a thermoelectric solution for the power generation
has been considered: the generator has been designed and assembled starting from a catalytic combustor.
Catalytic combustion allows safe control of the processes, and the choice of a hydrocarbon fuel ensures the
power availability and a fast recharge. The size of the system is set to fit a volume close to the one of AA
batteries. The electrical power output obtained is close to 1 W with a cold side temperature below 40 °C. The
limited values of these physical parameters allow obtaining a portable and safe device. The generator has been
fully characterized in different ranges of fuel flow rates and the performances have been thoroughly analysed for
processes optimization and efficiency improvement.

1. Introduction
Thermoelectric (TE) technology allows the direct conversion of heat
into electrical power. The complete absence of moving mechanical
parts makes this technology highly reliable for long term waste heat
recovery. The technology takes advantage of intrinsic material prop-
erties, therefore the operating behavior of a thermoelectric device is
strictly and intrinsically related to the material used. Of course, some
boundary conditions have to be considered, such as the maximum
temperatures, the maximum heat flows and other parameters involved
in the conversion process. The dimensionless figure of merit, ZT, asso-
ciated to a thermoelectric material, describes the effective capability of
the material to convert a heat flow into electrical power [1]. ZT para-
meter is also related to the maximum efficiency that a thermoelectric
element can achieve operating at a certain ΔT. In Fig. 1 the efficiency of
a thermoelectric device working between hot (Th) and cold (Tc) tem-
peratures, ΔT, is plotted for several ZT values. Increasing ZT values, the
equation approaches the Carnot limit. At the same time the ZT slope
decreases with increasing temperature gradient. In the inset the same
data are reported in the range actually interested by available ther-
moelectric technologies.
As for the state of art of thermoelectric modules, chalcogenides
based on Bi, Sb, Te and Se compounds are the only materials currently
used for the fabrication of commercial modules. The materials

⁎
Corresponding author.
E-mail address: carlo.fanciulli@cnr.it (C. Fanciulli).

https://doi.org/10.1016/j.apenergy.2018.02.012
Received 6 July 2017; Received in revised form 2 February 2018; Accepted 3 February 2018
0306-2619/ © 2018 Published by Elsevier Ltd.
C. Fanciulli et al. Applied Energy 215 (2018) 300–308

Fig. 1. Thermoelectric conversion efficiency vs. ΔT for different ZT values. The solid line corresponds to the Carnot limit efficiency. ZT values greater than 1.5 are reported to suggest the
potential perspectives of the technology, even if they are, up to now, inaccessible.

characteristics in terms of chemical stability, mechanical properties and

thermoelectric behavior, coupled with the technological solutions re-
quired for the module assembly, impose the operating boundary con-
ditions for the devices. Usually, the best performances are obtained for
ΔT between 200 °C and 250 °C, with an upper limit for Th close to
300 °C. Higher working temperatures affect the reliability of the device,
leading to materials aging or device performances degradation and
failure. The ZT average value for this class of materials is close to 1
[2,3], therefore the theoretical efficiency achievable with actual com-
mercial system is below 12%.
Recently, thanks to a growing interest in energy recovery solutions,
several materials have been developed and studied [2–6] with the aim
to achieve higher ZT values, for improvement in conversion efficiency,
and higher working temperatures range. This second target, even not
involving a significant change in the efficiency of the system, is im-
portant to open new fields of application for the thermoelectric tech- Fig. 2. Thermoelectric and ‘engines’ efficiency crossover vs. Power Outputs. Two curves
nology. for TE devices corresponding to different ZT values are reported [10].
In recent years optimized thermoelectric generators (TEG) for large
scale applications based on low power requirements have been deeply
taken into account. In fact, each interface and thermal joint are re-
investigated. It is important to underline the high reliability of these
sponsible for losses in the heat transfer affecting the total efficiency of
systems and their compactness and noiseless operating conditions, as
the thermoelectric generator (TEG). This reduction in efficiency re-
well as their capability of taking advantage of any kind of heat, despite
presents the main limitation for TE technology. In all the applications
the low temperature. These aspects already made the TE solution suc-
based on TE modules, a significant drawback is due to the heat loss in
cessful for applications such as low temperature waste heat recovery or
the TE modules coupling [11,28]. Many interesting studies are oriented
power generation in deep space exploration [7–9]. In Fig. 2 the com-
to the coupling of the thermoelectric technology to other systems,
parison between the efficiencies of a TEG system and a standard engine
looking for an enhancement of the energy production performance of
is reported vs. the power output required [10]. It can be seen that below
the overall system.
few hundreds of Watts, the thermoelectric system is advantageous
The application of TE devices in micro power generation has gained
thanks to its constant behavior.
also interest, given the need for power requirement reduction to fit the
Despite of the practical absence of thermoelectric generators on the
battery outputs. This interest has been also triggered by the develop-
market, an increasing number of studies have been performed for the
ment of miniaturization of mechanical and electromechanical en-
evaluation of a thermoelectric solution in different cases of waste heat
gineering devices, especially in the areas of microelectronics, bio-
recovery or system efficiency improvement [11–27]. Here the attention
mechanics, microfabrication techniques [29]. The aim is the
moves on the capability of transferring the results produced on mate-
development of power-supply devices with high specific energy (small
rials to the modules: the presence of electrical and thermal contacts
size, low weight, long duration) [30,31]. The choice of the heat source
necessary to embed the active elements, strongly affects the efficiency
is strictly related to the achievement of different efficiency limits. Due
of the thermoelectric conversion. Looking at the results reported in
to the high energy content, hydrocarbon fuels can provide, even at low
literature and at the data for the commercial modules available, effi-
conversion efficiency, energy density higher than the most advanced
ciencies between 1% and 5% are reported for the modules. Going down
batteries [32–35]. Therefore, hydrocarbon-based devices as portable
with the chain, the losses related to the thermal chain needed to
power sources [34–39] could be considered a suitable solution as a heat
transfer the heat from the hot source and to the cold sink have to be
source. In this context, catalytic reactors [40–44] are of major interest:

301
C. Fanciulli et al.
Fig. 3. Schematic of the TEG system: the coupling of the combustor with two commercial thermoelectric modules. Water cooled heat sink is used as a cold side. The system is designed to
optimize symmetry.
they are used either for direct conversion of thermal energy released via
combustion to electrical energy using thermoelectric or thermo-pho-
tovoltaic devices [45–48]. Catalytic combustion seems to be the most
suitable solution to overcome the major drawbacks of conventional
homogeneous (gas-phase) combustion systems: the high (> 1500 °C)
operating temperatures and the size related combustion limitations
[45–50]. The high temperatures greatly limit material selection and
require extensive combustor insulation. The size reduction, involving
an increase in the surface area-to-volume ratio, increases the heat-loss
to heat-generation ratio. The strong loss in the small dimension induces
flame quenching [29,51,52] and is also responsible for an increase in
pollutants emission. Catalytic combustion allows full utilization of the
high energy densities of hydrocarbon fuels, but at notably lower oper-
ating temperatures than those typical of traditional combustion. Ad-
ditionally, catalytic systems are generally easier to start, more robust to
heat losses, and self-sustained at leaner fuel/air ratios [53,54].
There are many examples of micro- and meso-scale thermoelectric
power generators powered by catalytic combustion in literature
[55–59]. A thermoelectric generator made of silicon bonded to glass,
developed by Yoshida et al. [59], was able to produce 184 mW of
electrical power with an efficiency of 2.8% from the catalytic com-
bustion of hydrogen. Norton and co-workers [31] reported for their
integrated combustor-TE generator the production of 1 W maximum
power and a thermal-to-electrical conversion efficiency of 1.08% with
hydrogen as fuel. This team also reported the generation of 0.45 W
electrical power with propane as fuel at 0.42% conversion efficiency
[45]. In recent years, Marton and co-workers [60] developed a butane-
fuelled TEG, delivering 5.82 W maximum power with 2.53% conversion
efficiency. Several efforts have been made by different authors to in-
vestigate the hot gases recirculation in order to reduce heat losses and
thus increase the efficiency of this kind of systems [53,61,62]. Never-
theless, the demonstration of a fuel-based TEG suitable for portable
power with energy density comparable to that of a battery remains an
open challenge.
In this work a small scale thermoelectric generator powered by a
catalytic combustion chamber has been designed and fully character-
ized. The main target is to develop a portable device having a size
comparable with that of a standard AA battery commonly used in small
electronic systems. The target application is not commercial batteries
replacement, but developing a reliable, easily manageable system for
energy backups. Moreover, the use of hydrocarbon fuel ensures a
302
Applied Energy 215 (2018) 300–308
notable power availability and, at the same time, an easy, fast and safe
rechargeable solution, overcoming the problem of the time consump-
tion typical of battery charging. The system introduced in this work
aims for portability, trying to satisfy the mobile devices requirements.
Under this perspective, a set of boundary conditions have been imposed
for the characterizations of the device performances. In order to ensure
that the device external temperature does not exceed a temperature safe
for system handling, the thermal management needs to keep the outside
temperature below 40 °C. In order to do so, a limit of 50 °C has been set
to the cold side of the TEG.
The approach used in this work is therefore innovative: the focus is
not searching for the maximum ΔT corresponding to the highest elec-
trical power output, which is the usual approach for TE technology, but
the system’s safety and portability. The combined effects of the compact
design, the controlled and safe combustion, the electrical output able to
supply small electronic devices and the limited temperature involved,
open a perspective for a thermoelectric option as power backup next to
or alternative to the conventional batteries.
Another focus of this work is the fuel consumption: the system ef-
ficiency was thoroughly evaluated, focusing on the scale effect on the
processes involved. To allow the discussion on this subject, a new
measurement system for the characterization of the thermal flows in-
volved in the system has been developed. The low power involved in
the experiments, required an improvement in the characterization
process to obtain reliable and repeatable results. In this work, the data
analyses have been validated also by the use of an analytical model
helping the full understanding of the processes observed. The model
developed can be also used to identify possible improvements in waste
heats recoveries, distinguishing the different contributions to the heat
losses of the system.
2. Materials and methods
The sketch of the TEG developed in this work is reported in Fig. 3. It
has been built up coupling thermoelectric modules to a catalytic com-
bustor. Two 15 × 15 mm2 thermoelectric (TE) modules, based on
chalcogenides materials, are placed in a thermal chain consisting of a
catalytic combustor used as heat source and two 16 × 16 mm2 water-
cooled heat sinks. The resulting overall combustor size is
16 × 16 × 4 mm3.
The geometry used for the combustion chamber was chosen in order
C. Fanciulli et al.
to match the size of the TE devices. The combustor consists of a
stainless steel chamber closed on top and bottom by 0.5 mm thick
aluminium layers. The difference in thermal conductivities between the
two materials (36–54 W/m K for steel and 204 W/m K for Al), together
with the favourable geometry, allow obtaining a heat mainstream
strongly anisotropic, promoting the heat flow through the thermo-
electric elements and limiting the lateral losses. The materials and the
layer thicknesses are chosen in order to allow good homogenization of
the temperature at the surface between the combustor and the ther-
moelectric modules, ensuring a more stable and optimized operating
condition for the thermoelectric components.
Eight Pt-based catalytic pellets (Alumina cylinders with 1 wt% of Pt
on the surface, r = 1.6 mm, h = 3.2 mm) are placed into the combus-
tion chamber in ordered lines so that the height of the channel for gas
flowing is 0.8 mm. The heat sinks are made of copper and homo-
geneously cooled by a 13 °C water flowing in the circuit at 0.5 l/min
volume rate. The cooling system is designed to provide similar thermal
conditions on both the TEG branches.
In order to improve thermal contacts at the interfaces between the
TE modules, the combustor and the heat sinks, graphite sheets (100 μm
thick) are used. Graphite has been chosen for its capability of com-
pensating the effects of elements surface roughness and preferred to
other solutions, like thermal pastes, for its stability even at high tem-
perature.
To ensure the thermal chain, the system is held together by a spring
and a settable bolt. The compressive system allows obtaining a homo-
geneous pressure of 1 MPa on the surfaces, minimizing thermal losses,
compensating the increase of pressure on the TE modules due to the
thermal expansion, and preserving the correct alignment of the thermal
chain components. Two layers of mica are used to close the thermal
chain insulating the TEG from the mechanical support. Mica has low
thermal conductivity and good mechanical resistance, thus allowing the
best decoupling between the generator and the external pressing system
and, at the same time, supporting the pressure needed to obtain an
efficient thermal coupling of the chain elements. Fig. 4 shows the as-
sembled system with the spring for the pressure regulation.
The TE modules used are Adaptive ET-017-14-11, elements de-
signed for cooling applications. No specific datasheet for power gen-
eration are available. However, an operative limit below 200 °C at the
Fig. 4. Picture of system assembled (left) and rendering of the TEG. The pressure on the
system can be tuned using the calibrated spring on the top of the chain. The spring also
allows compensating the overpressure due to the thermal expansion of the different
elements of the chain, ensuring a constant operating pressure on the TE modules.
303
Applied Energy 215 (2018) 300–308
hot side has been considered, taking into account the limits reported in
the module datasheet related to the construction characteristics. An
identical module was characterized, obtaining a Seebeck response of
6.5 mV/°C and an internal resistance, useful for the evaluation of the
maximum power at the matching load, which is about 0.4 Ω.
The combustor is fuelled with an air/propane mixture close to the
stoichiometric conditions. Mass flow rates are measured and controlled
by thermal mass flow meters (Bronkhorst El-Flow F-201CV). The lower
limit for the gas flow control is 0.2 l/min. Gas flow rate values are
chosen in order to obtain stable ΔT values across the thermoelectric
modules. Moreover, in all the conditions a maximum temperature of
50 °C at the cold side has been set. This boundary condition results in
lower ΔT and thus reduced performance, but it has been chosen for
safety and portability of the generator device.
The flow rate values investigated range from the lower manageable
with the system preserving the precision required for the experiments
(0.3 Nl/min), up to the value producing a cold temperature close to
50 °C in regime conditions (1.3 Nl/min). Correspondingly, the gas re-
sidence time in the combustion chamber ranges from 16 to 3.7 ms and
the temperature gradient applied to the TE modules ranges between
50 °C and 130 °C.
Details of the TEG system assembled with the whole set of ther-
mocouples are shown in Fig. 5. Temperature measurements have been
carried out using K-type thermocouples placed at the hot side (in the
middle of 1 mm thickness copper layers between the combustion
chamber and TE modules), and at the cold side (in the middle of heat
sink’s wall at a distance of 1 mm from the TE modules). Furthermore,
four K-type thermocouples provide the measurement of the water
temperature at the inlet and the outlet of the heat sinks.
The recording system of the temperatures and electrical properties
characterization has been implemented by using three National
Instrument data acquisition (NI DAQ) modules (two NI 9211 for tem-
perature and a NI 9207 for voltage acquisition) and an Agilent 6060b
electronic load, interfaced with a LabVIEW program. The program
constantly records the temperatures of the system and performs a
routine for TEG performances characterization (applying decreasing
external loads), thus monitoring and recording the electrical outputs of
the system. The resistive loads applied range between 0.01 Ω and 1 KΩ.
All measurements are performed at steady-state conditions [63].
The temperature measurements accuracy is about ± 0.1 °C in the
Fig. 5. TEG system temperature acquisition: 8 thermocouples are used to acquire tem-
perature at hot and cold sides and at water inlet/outlet.
C. Fanciulli et al.
Fig. 6. TEG electrical output. Power and voltage versus current for upper TE module (upper plots), and for lower TE module (lower plots).
temperature range considered (from −20 °C to + 350 °C).
The overall chemical efficiency has been evaluated by means of
Fourier Transform Infrared Spectrometry (FT-IR) analysis of the ex-
haust gases using the experimental setup as reported in [64,65]. The
exhaust gases have been collected from the combustor outlet and sent
through a cut-off particulate filter and a water trap to a Thermo Sci-
entific Nicolet 6700 FT-IR spectrometer equipped with a variable-
pathlength heated gas-cell (Gemini Mars series 6.4 M, internal volume
of 0.75 l) positioned inside the instrument. The combustor exhaust
gases are sent to a distiller used to remove the water produced in the
combustion processes. The estimated error for FT-IR measurements is
lower than 1%. The exhaust pumping rate (0.3 Nl/min) is adjusted so as
to be slightly lower than the mass flow rate measured at the combustor
outlet. Quantitative analysis of gas concentration has been performed,
and the chemical efficiency, ηc, has been calculated as:
[CO2]
ηc =
[CO2] + [CO] + [UHC ]
where [UHC] is the concentration of the total unburnt hydrocarbons.
The chemical efficiency has been evaluated with TEG system working at
its matching load (maximum power production) status to be closest to
the system operating conditions.
The TEG outputs and the combustor performances have been si-
multaneously characterized at different operating conditions, i.e.
changing the ΔT values applied to the modules.
In this paper, the overall efficiency of the system has been evaluated
as the ratio of the electrical power produced by the thermoelectric
conversion system, given by the measured electrical outputs, and the
power Pin provided to the system, based on the lower heating value
(LHV) of propane [45]:
Pel
η=
Pin
Applied Energy 215 (2018) 300–308
Therefore, the overall system efficiency is related to the product of
the combustion efficiency and the thermoelectric conversion efficiency
of the modules:
η = ηc ηTE (3)
Moreover, an evaluation of the power dissipated at the heat sink has
been performed. The power transferred to the water is calculated as
[66]:
dm
Pwater = cp ΔTwaterin − out with cp = 41,82 J/kg K
dt (4)
where the water flow temperature is controlled by the chiller pump and
constantly recorded during the experiment.
The results of these analyses allow checking the power flows into
the thermal chain and the whole operating conditions. The thermal
chain optimization obtained decoupling the TEG system from the sup-
porting structure allowed minimizing the power losses at the heat sink
(1) throughout the structure. The final results of the new apparatus for the
system characterization allow an improved reliability on the output
data associated to the TEG.
3. Results and discussion
3.1. System characterization
Following the results described in previous works [64,65], the TEG
has been designed applying solutions for the highest performance
achievement. The choice of strategies has also been imposed by the
small size of the system, resulting in the need for a high optimization in
order to reach useful performances. Due to the small volume of the
combustion chamber, an adequate range of total flow (propane + air)
input needs to be defined. To do so, the gas behavior in the combustion
(2) chamber needs to be characterized as a function of the total feed flow
304
C. Fanciulli et al.
Fig. 7. TEG electrical output: Power and voltage versus current for both TE modules in connected mode.
rate. If the velocity of the gases entering the combustion chamber is too
high, the residence time may be too low to allow adequate completion
of combustion reactions, or even to allow the mixture ignition. On the
other hand, if the input speed is too low, the amount of gases in the
combustion chamber may not be sufficient for sustaining the combus-
tion reaction.
In Figs. 6 and 7 the results of the system electrical characterizations
are reported: the characteristic curves of voltage and power vs. the
current are plotted for increasing imposed temperature gradients (ΔT).
In Fig. 6 the single module outputs are considered, while in Fig. 7 the
results for the system with the two modules connected in series are
shown.
Looking at the behavior of the modules independently, it can be
noticed that the upper side module displays better electrical perfor-
mances. This slight asymmetry can be explained considering the heat
distribution inside the combustion chamber. As a consequence, the
combustor surfaces have different temperatures leading to different hot
side temperatures.
When the modules are electrically connected, the ΔT for the overall
system is considered as the average of the temperature differences
along the two branches. It can be observed that, despite the absence of
any signal conditioning, the connected configuration does not display
any reduction when compared to the sum of the single modules results.
The maximum power produced by the conversion system is 0.84 W,
corresponding to 1.12 A at 0.75 V. This performance was obtained with
a gas flow of 1.3 Nl/min and 5.6 g/h fuel consumption. The tempera-
ture at the hot side, equal to 160 °C, ensures safe and continuous long
term TE module operation.
3.2. Efficiency evaluation
In order to evaluate the system efficiency, a power balance analysis
has been performed considering the TEG in the matching load steady
state:
Pin = Pout
where Pin is the power provided to the system, calculated as the product
of propane mass flow rate entered in the combustor times the lower
calorific value of propane (LHV = 46.35 × 106 J/kg), and the Pout the
sum of the outputs. The power provided to the system is:
Pin = Pdissipated + Pel + LHVUHC
In Eq. (6), Pel is the electrical power delivered by the TEG system, at
the matching load state. The power dissipated by the system can be
calculated as the sum of the heat lost by the system through convection
and radiation, the heat transferred to the cooling water, and the heat
transferred at the exhaust gases, divided by the time interval of the test:
Applied Energy 215 (2018) 300–308
Qconv + Qrad + Q water + Qexhaust
Pdissipated =
texperiment (7)
The power transferred to the water have been calculated using Eq.
(4). For the calculation of the power dissipated by convection and ra-
diation, the system has been modelled divided into three different
stages, each one associated to a specific thermal gradient as respect to
the environment. The three stages are the combustor, the TE modules
and the heat sinks respectively. The contribution to the power dis-
sipated due to convection phenomena has been estimated using the
Newton's law for heat convection for each stage of the system:
Pconv = hc AΔT(stage,env) (8)
where A is the surface area exposed to the surrounding air. The con-
vection coefficients hc have been set equal to 15 W/m2 K for the ther-
moelectric modules, and to 5 W/m2 K for the lateral surfaces, exposed
to the air, of the other metallic components of the thermal chain.
The power losses related to radiation phenomena, have been cal-
culated for the stages by the Stefan-Boltzmann law:
4 4
Prad = σ ∊ A (Tstage −Tenv ) (9)
where σ is the Stefan-Boltzmann constant, A is the same surface area
considered for the convection case, ∊ is the emissivity for the present
case estimated equal to 0.6. The mean absolute temperature of each
single element has been used for the temperatures gradient between the
emitting system and the surroundings calculation.
The last power loss considered is the one associated to the exhaust
gasses. The mixture, mainly composed by air, exits from the system at
high temperature, so the power loss associated can be calculated as:
Pexhaust = ∑ ⎛ dm ⎞ c p ΔTgas
gas in − out
gas ⎝ dt ⎠ gas (10)
where the parameters for the different exhaust gas components at the
outlet temperature have been considered. The values corresponding to
(5) the different contributions are reported in Table 1. The main con-
tribution to the power dissipated is due to the water cooling. The
convection and radiation represent minor contribution to the power
losses. The relative error for the results ranges between 5% and 10%
depending on the gas flow rate. The errors associated to the different
components of power dissipation are mainly related to the thermo-
(6) couple precision ( ± 0.1 °C). The calculation has been done following
the propagation laws applied to Eqs. (4) and (8)–(10). In the case of
convection, the assumed coefficients can also be a source of error.
However, this contribution is a second order correction due to the small
size of the system. The main experimental error remains the one asso-
ciated to the power calculated by the water-cooled circuit, where the
error on the liquid flow determination is added to the thermocouple
305
C. Fanciulli et al. Applied Energy 215 (2018) 300–308

Table 1 Table 3
Power input/output for TEG in the matching load state assuming ideal combustion. Chemical efficiencies and wasted fuel for three different flow rates.

Flow rate Power in Power out (W) Total flow Theoretical chemical FT-IR chemical Wasted propane (g/h)
(W) rate (Nl/ efficiency (%) efficiency (%) (empirical – FTIR)
min)
10−5 kg/s Nl/min Pin Pel Pwater Pconv Prad Pexhaust LHVUHC
0.3 79 ± 8 70 ± 4 0.27–0.39
0.63 0.3 17.87 0.21 13.23 0.70 0.10 1.04 2.59 0.7 68 ± 6 63 ± 4 0.96–1.11
1.06 0.5 29.79 0.43 22.08 0.98 0.21 1.73 4.36 1.3 56 ± 4 57 ± 4 2.46–2.41
1.48 0.7 41.70 0.60 26.64 1.10 0.39 2.42 10.55
1.91 0.9 53.61 0.73 31.05 1.17 0.47 3.12 17.07
2.33 1.1 65.53 0.80 35.49 1.40 0.55 3.81 23.48 the chemical efficiency can be defined as the ratio between Pout and Pin,
2.76 1.3 77.44 0.84 39.92 1.79 0.64 4.51 29.74
where all the power loss contributions are considered. The results of
this calculation match those obtained with FT-IR measurements within
one. The relative errors resulted higher for low gas flowrates due to the the experimental uncertainty, as shown in Table 3. The higher dis-
low values of the temperature gradient between incoming and outgoing crepancy observed at low flows can be explained with the higher un-
water and the consequent greater weight of the absolute reading pre- certainty on the experimental data in both the FT-IR and the power
cision. measurements. This is due to the close-to-the-limit experimental con-
As reported in Table 1, the potential power of unburnt propane is a dition, affecting the performance of the characterization system.
part of overall power. Moreover, the mass of unburnt propane grows In Table 3 the amount of wasted propane in each case is also re-
significantly with increasing inlet gas flow rate. This suggests that a low ported in order to express the observed lack of efficiency in terms of fuel
residence time in combustion chamber might lead to a significantly consumption.
reduced propane combustion efficiency, thus resulting in an over-esti- Using the results obtained from the power analysis reported in this
mation of the power provided to the system. In order to consider the Section, the overall system efficiency can be estimated as the ratio
actual power provided to the system, a chemical efficiency needs to be between the power delivered by the system (i.e., the electrical output
introduced: measured) and the power Pin provided to the system. When the effi-
ciency is based on the theoretical power provided, the system efficiency
dmp ranges between 1 and 1.44%, as shown in Table 4.
Pin,actual = LHVηch
dt (11) However, if the actual power provided, as defined in Eq. (11) is used
The system chemical efficiency has been evaluated through FT-IR for the calculation, the efficiency is in the range 1.7–2.3%. These latest
measurements on the combustor exhaust gases at three different total values, also reported in Table 4, are closer to the results already ob-
flow rates (0.3 Nl/min, 0.7 Nl/min and 1.3 Nl/min). The exhaust gases served elsewhere in literature for thermoelectric generators based on
analysis performed through the acquisition of FT-IR spectra allows catalytic combustion [64]. Nevertheless, the theoretical values need to
determining the concentration of unburnt propane, carbon monoxide be taken into account for the use of the system in real TEG applications.
and carbon dioxide at the exhaust gas. Using the concentration mea- The results of this analysis lead to a better match of the system with
surements, it is possible to evaluate the chemical efficiency through Eq. the others proposed in literature. In particular, the thermoelectric ef-
(1), in which propane is the only detected unburnt hydrocarbon. The ficiency reflects the expected values for the modules operating in the
results are reported in Table 2. The errors reported correspond to the regime described. On the other hand, the combustion efficiency is
experimental standard deviations obtained by the FIIR measurements. shown to be strongly affected by the small size. Nevertheless, further
For the chemical efficiencies, the error reported has been determined device improvement can be obtained through combustion process op-
calculating the maximum ranges after repeated measurements. timization. Hence, two different strategies have been explored with this
A significant decrease in chemical efficiency of the combustion aim.
process with increasing gas flow rate is shown. This is consistent with The first approach involves residence time enhancement changing
the hypothesis of a residence time too short to allow complete fuel the free volume in the combustion chamber. Despite of the different
consumption. Furthermore, the small time characteristic of the catalytic attempts performed, this approach has been abandoned because the
action of the pellets and the reduced surface-to-volume ratio of the results were not effective in increasing the TEG performance.
catalyst in this small chamber are likely to affect the whole catalytic The second approach involves catalytic efficiency enhancement
process efficiency, thus contributing to the performance loss. through catalytic active surface increase. When 8 pellets are placed in
It is also worth noticing that, under the conditions of maximum the combustion chamber, a fraction of their total surface area covered
electrical power output (matching load applied), the chemical effi- in catalyst is in contact with the chamber walls and/or with adjacent
ciency decreases with respect to the open circuit case. This effect is pellets, so that the active catalyst surface is reduced. Using only 7
related to the increasing reversal Peltier effect due to the increasing pellets and changing their arrangement in the combustion chamber,
current flowing in the TE modules, that reduces the temperature of the attempts were made in order to increase the active catalyst surface area.
chamber surfaces and, in turn, the combustion performance. The active surface evaluation in different pellets arrangements was
The chemical efficiency estimated through FT-IR measurements can performed through image analyses, considering the contact surfaces as
be validated using the experimental data reported in Table 2. Namely, boundaries of circular sectors open or close to the catalytic action. The
two pellets configurations used to change the active catalytic surface
Table 2
FT-IR analyses results: exhaust gas species amounts and calculated chemical efficiencies.
Table 4
Flow C3H8 Std. CO2 (PPM) Std. CO (PPM) Std. Chemical FT-IR analysis results: exhaust gas species amounts and calculated chemical efficiencies.
rate (PPM) error error error efficiency
(Nl/ C3H8 CO2 CO (%) Flow rate Pin (W) Efficiency based Pin,actual (W) Efficiency based on
min) (PPM) (PPM) (PPM) (Nl/min) on Pin (%) Pin,actual (%)

0.3 5916 617 40,900 1801 10 344 70 ± 4 0.3 17.9 1.2 12.5 1.7
0.7 11,052 1229 59,879 1830 2004 71 63 ± 4 0.7 41.7 1.4 26.3 2.3
1.3 17,890 2088 75,269 1627 2830 178 57 ± 4 1.3 77.4 1.1 44.1 1.9

306
C. Fanciulli et al. Applied Energy 215 (2018) 300–308

Fig. 8. Test configurations with 7 pellets in the combustion chamber: at the left the condition called 7phs, having catalytic surface of 303 mm2; at the right the condition 7pls, with catalytic
surface 267 mm2.

Table 5 temperature lower than 50 °C for safe device handling. The perfor-
Test conditions with equal free volume in the combustion chamber and different catalyst mance obtained for the system is close to 1 W of electrical power pro-
surface. The temperature differences between the hot side and cold side of the TE and the
duced in the matching load configuration with a DC voltage of 0.75 V
corresponding measured chemical efficiencies are reported.
and a current of approximately 1.1 A. This work underlines the im-
Setup Gas flow rate Free volume Catalyst surface ΔT (°C) ηch (%) portance of size on catalytic combustion phenomena, proposing some
(Nl/min) (mm3) (mm2) easy solutions to achieve first order optimization of the system. Further
developments are in progress to exploit the system features to further
7phs 1.3 188 303 154 70 ± 3
7phs 0.7 188 303 116 77 ± 3
enhance its performance. The choice of different catalytic elements and
7pls 1.3 188 267 132 64 ± 3 the use of better performing thermoelectric modules are expected to
7pls 0.7 188 267 104 70 ± 3 strongly affect the final performance of such a device, making it more
suitable for large scale backup application as battery charger or similar.
The combined effects of the compact design, the controlled and safe
area, keeping the total free volume in the chamber constant, are shown combustion and the electrical output, able to supply small electronic
in Fig. 8. devices, can be effective for the development of sensors networks, an
Preliminary tests were performed at two different gas flow rates and enabling factor in the perspective of interactive environments and of
are shown in Table 5. Results show that a higher surface area available the Internet of Things. Such a solution can be mounted in small spaces,
for catalytic reaction results in higher combustion efficiency, as ex- on moving objects or face other critical conditions, providing an ef-
pected. It can be observed that an increase in the catalyst surface of fective backup to a battery powered system or to a communicating
about 11.5% (from 267 to 303 mm2), results in average increase of the device imposing no specific constraints. Moreover, the additional
temperature gradient across the modules of about 15 ± 3 °C, leading to choice of limiting the temperature involved, opens here a perspective
a chemical efficiency improved of about 10% for both the investigated for a thermoelectric option as power backup next to or alternative to
flows. The improvement in catalytic operation should help the final the conventional batteries in portable devices. Like for all the com-
power outputs of the system enabling an improvement of the TEG bustors reported in literature fuelled by propane, the fuel mixture
output between 5 and 10% depending on different flowrates. These doesn’t come from a self-breathing system, but from a forced stoichio-
projections have been made on the basis of the results obtained for the metric premixing phase. To further develop such systems to the real
system with 8 catalytic pellets. application the feeding strategy have still to be improved. However, the
TEG’s simple design, and the common materials and components used
in this work, represent a first step to envisage its use in large scale
4. Conclusions
applications.

In this work a thermoelectric generator based on a catalytic com-

Acknowledgements
bustor has been set up and characterized. The device has been designed
in a size comparable to the one of an AA class commercial battery
The authors would like to acknowledge the financial support for the
aiming for high portability. The system involves a liquid cooling circuit.
project I-ZEB in the framework of CNR-Regione Lombardia program,
This is nominally in contrast with the idea of a mobile solution, but the
the technical assistance of Mr Bassani, Mr. Fantin.
liquid volume required to support an efficient cooling is comparable to
the one needed to store the fuel for the combustor. Furthermore, the
References
system characterization has been performed keeping in mind its por-
table application. The temperature gradients chosen were not the
[1] Rowe DM. Thermoelecrtics handbook - macro to nano. Boca Raton (FL, USA): CRC
highest possible for the system operation, therefore the maximum Press, Taylor and Francis Group; 2006.
power produced by the device do not correspond to the highest [2] Snyder GJ, Toberer ES. Complex thermoelectric materials. Nat Mater
achievable performance. However, considering the limited cooling 2008;7:105–14.
[3] Medlin DL, Snyder GJ. Interfaces in bulk thermoelectric materials: a review for
capability due to the reduced size of the heat exchanger, the regimes current opinion in colloid and interface science. Curr Opin Colloid Interf Sci
chosen were the only accessible with the requirement of a cold side

307
C. Fanciulli et al.
2009;14:226–35.
[4] Pei Y, Wang H, Snyder GJ. Band engineering of thermoelectric materials. Adv Mater
2012;24:6125–35.
[5] Gaultois MW, Sparks TD, Borg CH, Seshadri R, Bonificio WD, Clarke DR. Data-
driven review of thermoelectric materials: performance and resource considera-
tions. Chem Mater 2013;25:2911–20.
[6] Alam H, Ramakrishna S. A review on the enhancement of figure of merit from bulk
to nano-thermoelectric materials. Nano Energy 2013;2:190–212.
[7] Hsu C-T, Huang G-Y, Chu H-S, Yu B, Yao D-J. Experiments and simulations on low-
temperature waste heat harvesting system by thermoelectric power generators.
Appl Energy 2011;88:1291–7.
[8] Walker ME, Theregowda RB, Safari I, Abbasian J, Arastoopour H, Dzombak DA,
et al. Utilization of municipal wastewater for cooling in thermoelectric power
plants: evaluation of the combined cost of makeup water treatment and increased
condenser fouling. Energy 2013;60:139–47.
[9] Schmidt GR, Sutliff TJ, Dudzinski LA. Radioisotope power: a key technology for
deep space exploration. Radioisotopes – applications in physical sciences, InTech;
2011 [chapter 20].
[10] Vining CB. An inconvenient truth about thermoelectrics. Nat Mater 2009;8:83–5.
[11] Liu W, Jie Q, Kim HS, Ren ZF. Current progress and future challenges in thermo-
electric power generation: from materials to devices. Acta Mater 2015;87:357–76.
[12] Vorobiev Y, Gonzalez-Hernandez J, Vorobiev P, Bulat L. Thermal-photovoltaic solar
hybrid system for efficient solar energy conversion. Sol Energy 2006;80:170–6.
[13] Hsu C-T, Huang G-Y, Chu H-S, Yu B, Yao DJ. Experiments and simulations on low-
temperature waste heat harvesting system by thermoelectric power generators.
Appl Energy 2011;88:1291–7.
[14] Hsiao YY, Chang WC, Chen SL. A mathematic model of thermoelectric module with
applications on waste heat recovery from automobile engine. Energy
2010;35:1447–54.
[15] Niu X, Yu J, Wang S. Experimental study on low-temperature waste heat thermo-
electric generator. J Power Sources 2009;188:621–6.
[16] Rowe DM, Min G. Evaluation of thermoelectric modules for power generation. J
Power Sources 1998;73:193–8.
[17] Riffet SB, Ma X. Thermoelectrics: a review of present and potential applications.
Appl Therm Eng 2003;23:913–35.
[18] Rowe DM, Min G. Design theory of thermoelectric modules for electrical power
generation. IEE Proc – Sci Meas Technol 1996;143:351–6.
[19] Bell LE. Cooling, heating, generating power and recovering waste heat with ther-
moelectric systems. Science 2008;321:1457–61.
[20] Ikoma K, Munekiyo M, Furuya K, Kobayashi M, Izumi T, Shinohara K.
Thermoelectric module and generator for gasoline engine vehicles. In: 17th
International conference on thermoelectrics proceedings; 1998. p. 464–7.
[21] Kishi M, Nemoto H, Hamao T, Yamamoto M, Sudou S, Mandai M, et al. Micro-
thermoelctric modules and their application to wrist watches as an energy source.
In: 18th International conference on thermoelectrics proceedings; 1999. p. 301–7.
[22] Yang J. Potential applications of thermoelectric waste heat recovery in the auto-
motive industry. In: 24th International conference on thermoelectrics proceedings;
2005. p. 155–9.
[23] Min G, Rowe DM. Conversion efficiency of thermoelectric combustion systems.
IEEE Trans Energy Convers 2007;22:528–34.
[24] Yang D, Yin H. Energy conversion efficiency of a novel hybrid solar system for
photovoltaic, thermoelectric and heat utilization. IEEE Trans Energy Convers
2011;26:662–70.
[25] Wang N, Han L, He H, Park N-H, Koumoto K. A novel high-performance photo-
voltaic-thermoelectric hybrid device. Energy Environ Sci 2011;4:3676–9.
[26] Salvador JR, Cho JY, Ye Z, Moczygemba JE, Thompson AJ, Sharp JW, et al.
Conversion efficiency of skutterudite-based thermoelectric modules. PCCP
2014;16:12510–20.
[27] Yazawa K, Shakouri A. Cost-efficiency trade-off and the design of thermoelectric
power generators. Environ Sci Technol 2011;45:7548–53.
[28] Baranovski LL, Snyder GJ, Toberer ES. Effective thermal conductivity in thermo-
electric materials. J Appl Phys 2013;113:24904.
[29] Vican JGBF, Gajdeczko BF, Dryer FL, Milius DL, Aksay IA, Yetter RA. Development
of a microreactor as a thermal source for micro-electromechanical systems power
generation. Proc Combust Inst 2002;29:909–16.
[30] Fernandez-Pello AC. Micropower generation using combustion: issues and ap-
proaches. Proc Combust Inst 2002;29:883–99.
[31] Norton DG, Voit KW, Brüggemann T, Vlachos DG, Wetzel ED. Portable power
generation via integrated catalytic microcombustion-thermoelectric devices. In:
24th Army science conference, Orlando, FL; 2004.
[32] Hua J, Wu M, Kumar K. Numerical simulation of the combustion of hydrogen-air
mixture in micro-scaled chambers Part II: CFD analysis for a micro-combustor.
Chem Eng Sci 2005;60:3507–15.
[33] Patil AS, Dubois TG, Sifer N, Bostic E, Gardner K, Quah M, et al. Portable fuel cell
systems for America’s army: technology transition to the field. J Power Sources
2004;136:220–5.
[34] Yadav S, Sharma P, Yamasani P, Minaev S, Kumar S. A prototype micro-thermo-
electric power generator for micro-electromechanical systems. Appl Phys Lett
2014;104:123903.
[35] Dunn-Rankin D, Leal EM, Walther DC. Personal power systems. Prog Energy
Combust Sci 2005;31:422–65.
308
View publication stats
Applied Energy 215 (2018) 300–308
[36] Choi W, Kwon S, Dong Shin H. Combustion characteristics of hydrogen-air pre-
mixed gas in a sub-millimeter scale catalytic combustor. Int J Hydrogen Energy
2008;33:2400–8.
[37] Aravind B, Raghuramb GKS, Kishore VR, Kumar K. Compact design of planar
stepped micro combustor for portable thermoelectric power generation. Energy
Convers Manage 2018;156:224–34.
[38] Yadav S, Yamasani P, Kumar S. Experimental studies on a micro power generator
using thermo-electric modules mounted on a micro-combustor. Energy Convers
Manage 2015;99:1–7.
[39] Lee SI, Um DH, Kwon OC. Performance of a micro-thermophotovoltaic power
system using an ammonia-hydrogen blend-fueled micro-emitter. Int J Hydrogen
Energy 2013;38:9330–42.
[40] Norton DG, Wetzel ED, Vlachos DG. Thermal management in catalytic micro-
reactors. Ind Eng Chem Res 2006;45:76–84.
[41] Ouyang X, Bednarova L, Besser RS, Ho P. Preferential oxidation (PrOx) in a thin-film
catalytic microreactor: advantages and limitations. AIChE J 2005;51:1758–72.
[42] Rebrov EV, De Croon MHJM, Schouten JC. A kinetic study of ammonia oxidation on
a Pt catalyst in the explosive region in a microstructured reactor/heat-exchanger.
Chem Eng Res Des 2003;81:744–52.
[43] Deshmukh SR, Vlachos DG. A reduced mechanism for one-step rate expressions for
catalytic combustion of small alkanes on noble metals. Combust Flame
2007;149:366–83.
[44] Kaisare NS, Deshmukh SR, Vlachos DG. Stability and performance of catalytic mi-
croreactors: simulation of propane catalytic combustion on Pt. Chem Eng Sci
2008;63:1098–116.
[45] Federici JA, Norton DG, Bruggeman T, Voit KW. Catalytic microcombustor with
integrated thermoelectric elements for portable production. J Power Sources
2006;161:1469–78.
[46] Ahn J, Eastwood C, Sitzki L, Ronney PD. Gas-phase and catalytic combustion in
heat-recirculating burners. Proc Combust Inst 2005;30:2463–72.
[47] Zhang S, Yuan XZ, Hin JNC, Wang H, Friedrich KA, Schulze M. A review of pla-
tinum-based catalyst layer degradation in proton exchange membrane fuel cells. J
Power Sources 2009;194:588–600.
[48] Wu M, Yetter RA, Yang V. Combustion in meso-scale vortex combustors: experi-
mental characterization. In: 42nd AIAA aerospace sciences meeting and exhibit, 5–8
Jan. 2004, Reno, Nevada. AIAA 2004-980; 2004.
[49] Wierzbicki TA, Lee IC, Gupta AK. Combustion of propane with Pt and Rh catalysts
in a meso-scale heat recirculating combustor. Appl Energy 2014;130:350–6.
[50] Neyestanaki AK, Kumar N, Lindfors L-E. Catalytic combustion of propane and
natural gas over Cu and Pd modified ZSM zeolite catalysts. Appl Catal B
1995;7:95–111.
[51] Maruta K. Micro and mesoscale combustion. Proc Combust Inst 2011;33:125–50.
[52] Yang WM, Chou SK, Shu C, Xue H, Li ZW, Li DT, et al. Microscale combustion
research for application to micro thermo-photovoltaic systems. Energy Convers
Manage 2003;44:2625–34.
[53] Vijayan V, Gupta AK. Combustion and heat transfer at meso-scale with thermal
recuperation. Appl Energy 2010;87:2628–39.
[54] Vijayan V, Gupta AK. Flame dynamics of a meso-scale heat recirculating combustor.
Appl Energy 2010;87:3718–28.
[55] Aranguren P, Astrain D, Rodríguez A, Martínez A. Experimental investigation of the
applicability of a thermoelectric generator to recover waste heat from a combustion
chamber. Appl Energy 2015;152:121–30.
[56] Chou SK, Yang WM, Chua KJ, Li J, Zhang KL. Development of micro power gen-
erators – a review. Appl Energy 2011;88:1–16.
[57] Kinsella CE, O’Shaughnessy SM, Deasy MJ, Duffy M, Robinson AJ. Battery charging
considerations in small scale electricity generation from a thermoelectric module.
Appl Energy 2014;114:80–90.
[58] Xiao H, Qiu K, Gou X, Ou Q. A flameless catalytic combustion-based thermoelectric
generator for powering electronic instruments on gas pipelines. Appl Energy
2013;112:1161–5.
[59] Yoshida K, Tanaka S, Tomonari S, Satoh D, Esashi M. High-energy density miniature
thermoelectric generator using catalytic combustion. J Microelectromech Syst
2006;15:195–203.
[60] Marton CH, Haldeman GS, Jensen KF. Portable thermoelectric power generator
based on a microfabricated silicon combustor with low resistance to flow. Ind Eng
Chem Res 2011;50:8468–75.
[61] Shirsat V, Gupta AK. A review of progress in heat recirculating meso-scale com-
bustors. Appl Energy 2011;88:4294–309.
[62] Tolmachoff ED, Allmon W, Waits CM. Analysis of a high throughput n-dodecane
fuelled heterogeneous/homogeneous parallel plate microreactor for portable power
conversion. Appl Energy 2014;128:111–8.
[63] Pierce RD, Stevens RJ. Experimental comparison of thermoelectric module char-
acterization methods. J Electron Mater 2015;44:1796–802.
[64] Merotto L, Fanciulli C, Dondè R, De Iuliis S. Study of a thermoelectric generator
based on a catalytic premixed meso-scale combustor. Appl Energy
2016;162:346–53.
[65] Abedi H, Merotto L, Fanciulli C, Dondè R, De Iuliis S, Passaretti F. Study of the
performances of a thermoelectric generator based on a catalytic meso-scale H2/
C3H8 fuelled combustor. J Nanosci Nanotechnol 2017;17:1592–600.
[66] Incropera FP, DeWitt DP, Bergman TL, Lavine AS. Fundamentals of heat and mass
transfer. 6th ed. Hoboken (NJ, USA): John Wiley & Sons; 2007.