Download Performance evaluation of a natural convective

Performance evaluation of a natural convective-cooled concentration solar
thermoelectric generator coupled with a spectrally selective high temperature
absorber coating
Aaditya Anand*, V. Praveen Kumar, Harish C. Barshilia1
Nanomaterials Research Laboratory, Surface Engineering Division
CSIR – National Aerospace Laboratories, Bangalore 560017, India
Abstract
Solar energy can be directly harnessed for power generation by using Solar Thermoelectric
Generator (STEG) technology, which comprises of solar absorbers integrated with
thermoelectric materials. STEGs behave as solid state heat engines which can utilize the heat
energy of the sun to produce a temperature gradient across a thermoelectric device, which is in
turn converted to electrical energy. In this paper, we focus on investigating the performance of
the solar absorber subsystem that employs a high temperature spectrally selective coating on a
stainless steel substrate. We have performed temperature measurements on the absorber coating
exposed to solar irradiation flux at different optical concentration ratios (10-100) and validated
the experimental data using a numerical heat transfer model in COMSOL Multiphysics. This has
been combined with the high temperature emittance measurements of the coating to develop a
predictive efficiency model for the STEG system as a function of the thermoelectric figure of
merit at a hot side temperature range of 100 to 500°C. Further, we have experimentally examined
the performance of a natural convective cooled STEG consisting of a series combination of three
commercial Bi2Te thermoelectric modules coupled to the selective absorber coating. The
1
Author to whom correspondence should be addressed
E-mail: [email protected]
Fax # +91-80-2521 0113; Tel. # +91-80-2508 6494
*
Also at Department of Mechanical Engineering, BITS Pilani – Pilani, India
1
maximum power generated from the STEG has been measured at different concentration ratios
and the peak efficiency of the system has been calculated in the feasible temperature range of the
thermoelectric module.
Keywords: Solar Thermoelectric Generator, Spectrally Selective absorber coating, Optical
concentration, Efficiency
1
Introduction
Growing energy demand across the world has necessitated the need to explore new
technologies for power generation. The sun’s energy presents great potential to serve as a viable
alternative source of power as it is a renewable and sustainable source of energy [1].Solar
photovoltaic systems and large scale solar thermal power systems are the most common
technologies currently being used to harness solar power [2-5].In addition to these technologies,
the area of solar thermoelectrics represents another option that can convert the sun’s heat to
useful electrical energy [6], especially for small scale applications like micro-power systems.
Thermoelectric generators have historically been used primarily in waste heat recovery
systems but they can also be an effective means of small scale power generation when integrated
with solar power. Such generators are called Solar Thermoelectric Generators (STEGs) and use
thermal energy to create a temperature gradient across the hot and cold junctions of a
thermoelectric material which is used to generate voltage as per the principle of Seebeck effect
[7].STEGs typically work as solid-state heat engines and therefore act as a portable power
generating system for standalone rooftop arrangements. They do not have any requirement of
moving parts or working fluids for operation and therefore have captured interest in recent times
[8,9].
2
The performance of an STEG system is largely dependent on two of its components: the
solar absorber and the thermoelectric material. The need to maximize the performance of each of
these components leads us to realize the important role played by the materials employed in the
system. Spectrally selective absorbers are used so as to absorb maximum amount of radiation
incident on their surface (i.e., high absorptance, α), while at the same time emitting less radiation
(i.e., low thermal emittance, ε) to ensure a high efficiency of the absorber [10,11].Therefore, they
must be stable and retain their selective properties at high temperatures to minimize radiation
losses. Similarly, high figure of merit (ZT) materials need to be employed since the
thermoelectric efficiency of the system is directly dependent on the value of ZT in the
temperature range of operation [12].Moreover, to achieve large temperature gradients,
techniques like optical and thermal concentration have to be used so as to concentrate the solar
beam to produce a large value of irradiation flux. Therefore, the system efficiency of an STEG is
governed by interplay of various parameters, which need to be optimized to achieve a low cost
and effective design for power generation.
The pioneering research on STEGs for power generation was carried out by Maria Telkes
in 1954. Her experiments demonstrated 3.35% system efficiency under optical concentration
with a lens employing the best Zn-Sb and Bi alloy combination of that period [13].This reported
efficiency was not bettered until 2011 when Daniel Kraemer et al. [14]reported an efficiency of
4.6% for nanostructured bismuth telluride (Bi2Te3) thermoelectric materials using solar selective
absorbers and thermal concentration by employing a large absorber. With the arrival of high ZT
materials and the need for further exploring the performance potential of STEG technology, there
have been various recent works in this field. Studies featuring the use of nanostructured materials
and thermal concentration, like the work by Kraemer et al. and innovative designs employing
3
segmented and cascaded thermoelectric materials to optimize material performance as carried
out by McEnaney et al. [15]have shed light on various experimental improvements in design and
approach. Further, various works like the studies performed by Ram and Amatya [16] and A.
Pereira et al. [17] have also been dedicated to modeling and simulation of STEG performance
using thermodynamic models and finite element techniques to optimize experimental methods
and validate experimental observations.
In our present study, we have evaluated the performance of an STEG that uses a
commercial Bi2Te3 thermoelectric module coupled with a high temperature spectrally selective
absorber coating. In the first part of the paper, we have determined the thermal performance of
our coating under different optical concentration ratios using a Fresnel lens to determine the
steady state temperatures that can be attained and sustained by it. In the subsequent sections, we
have made theoretical predictions of the re-radiation losses, absorber efficiency and the net
STEG system efficiency using the temperature-concentration ratio correlation of our coating. We
have also carried out high temperature emittance measurements of the coating and incorporated
this emittance variation while predicting the system efficiencies as a variation with ZT in the
temperature range of 100 to 500°C. Finally, we have experimentally evaluated the performance
of an STEG that uses this absorber coating as a collector, integrated with three thermoelectric
modules in a stacked arrangement. A finned aluminum heat sink has been used to facilitate
natural convective cooling to produce the desired temperature gradient, and voltage, power
output, and efficiency measurements have been carried out.
2
Experimental details
The absorber coatings were deposited on stainless steel (SS) substrates (dimensions 35
mm × 35 mm  2 mm) by a Four-Cathode Reactive Unbalanced Direct Current (DC) Magnetron
4
Sputtering System with 7.25˝ diameter targets. Absorptance and emittance of the coatings were
measured using Solar Spectrum Reflectometer (model SSR) and Emissometer (Model AE) of
M/s. Devices and Services. The emittance was measured at 82°C. The optical properties of the
absorber coatings were also measured using UV-Vis-NIR (Cary 500i, Varian) and FTIR (Nicolet
6700, Thermo Scientific). As will be discussed later, the emissivity at high temperatures was also
measured for the absorber coating.
To understand the usefulness of such a coating, initial theoretical calculations were
performed in which the absorber efficiency of a system using a selective coating was compared
to that of a system using a polished stainless steel surface as an absorber. According to StefanBoltzmann’s law, any surface radiates energy to its surroundings at a rate proportional to its
emittance and the fourth power of its temperature (T4), which is a thermal loss for any energy
system. To establish the suitability of a spectrally selective coating in such an application,
analysis of these re-radiation losses was also performed by including the high temperature
emittance measurements carried out.
In the experimental study, the steady state temperatures attainable using the selective
absorber coating were determined under concentrated solar irradiation flux using a Fresnel lens
at various optical concentration ratios. The experiments were performed in Bengaluru, India
when the Direct Normal Irradiation (DNI) as calculated from Ref [18]was about 865𝑊/𝑚2 . To
avoid optical tracking issues, the measurements were taken during noon time. A suitable fixture
was fabricated for the outdoor rooftop setup for carrying out the temperature measurements. A
Fresnel lens, 49 x 65 𝑐𝑚2 was used for concentrating the incident solar irradiation so as to
produce a large value of flux (up to 100 suns) at the surface of the absorber. The solar selective
substrate was mounted on a ceramic brick which was used as it is a very good thermal insulator
5
so as to prevent heat loss due to conduction to obtain maximum possible surface steady state
temperature.
To study the variation of the absorber temperature with optical concentration ratio, the
setup was designed so as to enable changing of distance between the sample and the lens. This
was done to enable change of focus area, and therefore change in the optical concentration. A Pt100 temperature sensor, having an approximate error of ± 20 𝐶, connected to a temperature
display indicator was clamped to the substrate. During the experiment, the entire frame at any
point of time was adjusted such that the direct radiation from the sun fell normally on the
absorber. This was to ensure that peak solar flux was achieved at all times. The sample was
allowed to attain steady state and the temperature readings were taken. These values were
validated using a numerical heat transfer simulation performed in COMSOL Multi-physics.
In another set of experiments, the solar selective substrate was used in an STEG system
to directly convert solar energy to electrical energy for micro-power applications. A schematic
diagram of the experimental setup has been shown in Fig. 1(a). A commercial bismuth telluride
HZ-2 thermoelectric module manufactured by Hi-Z Technology Inc. having 97 p-n thermocouples was fixed below the substrate. A finned aluminum heat sink was used below the module
to enhance the natural air convective cooling at its bottom surface due to increased surface area.
For the initial open circuit voltage tests, the output terminals of the module were connected to a
digital voltmeter and the voltage measurements were recorded at different hot side temperatures
by varying the concentration ratio. However, as no external cooling mechanism was employed,
this arrangement with a single module could not yield a sufficient temperature gradient to
generate desired output voltages. Therefore, three thermoelectric modules were connected
electrically in series and placed in a thermally parallel stacked arrangement below the solar
6
absorber as shown in Fig. 1(b). Thermal grease was applied on each surface of the modules and
the bottom of the absorber so as to ensure electrical insulation and near perfect thermal contact.
By varying the optical concentration ratio of the incident solar radiation, the different voltage
readings obtained were recorded. Fig. 1(c) shows the STEG system with aluminum heat sink,
Fresnel lens and voltage display indicator. The concentrated focal spot of the solar irradiation on
the spectrally selective coating can be seen in Fig. 1(d).
In the subsequent maximum power measurement tests, the electrical circuit was designed
so as to operate at maximum power output by connecting a load resistance equal to the internal
resistance of the circuit [19,20].This was done by connecting a 100 Ω potentiometer as the
variable load and its resistance was varied starting from 1 Ω to match the internal resistance of
the thermoelectric modules. The internal resistance as specified by Hi-Z Technology Inc. was 4
Ω for the single module setup and 12 Ω for the series setup with three modules. However, it is
expected that maximum power will be obtained at a slightly higher value of resistance because of
the internal resistance of other components in the circuit like ammeter, connecting wires, etc.
Again, the optical concentration was varied and the power output was calculated by measuring
current and voltage. The temperatures of the surface of the solar absorber and the aluminum heat
sink were also monitored during the experiments.
3
3.1
Results and discussion
Spectrally selective coating on W coated SS substrates
A spectrally selective coating was used to harness the solar energy for generating
temperatures in the range of 100-512°C (for short durations) depending up on the solar
concentration. The absorber coating consisted of a tandem multilayer stack of tungsten (W),
titanium aluminum nitride (TiAlN), titanium aluminum silicon nitride (TiAlSiN), titanium
7
aluminum silicon oxy-nitride (TiAlSiON) and titanium aluminum silicon oxide (TiAlSiO) coated
on stainless steel 304 substrate. The W layer was deposited using a balanced magnetron
sputtering system and all the other constituent layers of the tandem stack were deposited using a
four-cathode reactive unbalanced pulsed direct current magnetron sputtering technique. The
tungsten layer deposited on the substrate decreased the thermal emittance of stainless steel 304 to
0.03-0.04 from its intrinsic emittance of 0.12-0.13. Fig. 2 shows the typical reflectance data of
the absorber coating in both visible and IR regions, depicting very high reflectance in the IR
region. This high reflectance of the absorber coating is caused because of W interlayer, which
acts as a good IR reflector. The W layer also helps to improve the thermal stability of the
spectrally selective coating by acting as a diffusion barrier between the SS substrate and tandem
absorber. This tandem stack exhibits absorptance of 0.956 and emittance of 0.07 at 82°C.
Detailed thermal performance analysis shows that the tandem absorber was stable up to 350°C in
air for 1050 h under cyclic heating conditions with  = -0.031, ε = +0.01 and up to 600°C in
vacuum for 1000 h under cyclic heating conditions with  = -0.011, ε = +0.03. The details of
the absorber coatings are provided elsewhere [21]. The radiative properties of the solar absorber
coating was measured in the temperature range of 100-500°C using a direct method wherein the
sample was raised to elevated temperature and the power input required to maintain the elevated
temperature was measured. More details about the high temperature emissivity measurements
can be found elsewhere [22].These measurements were also confirmed using FTIR and UV-VisNIR spectrometer data. The emissivity values were 10.1, 11.2, 13.8, 17.7 and 22.3% at 100, 200,
300, 400 and 500°C, respectively. The uncertainties of the direct calorimetric total hemispherical
emittance measurements are approximately 5%. As will be discussed in Section 3.3, these high
temperature emittance measurements were used to predict the re-radiative losses and the STEG
8
system efficiency in an evacuated environment. To our knowledge, this presents a more
comprehensive analysis of STEG efficiency as we have accounted for increase in emittance with
temperature, thus predicting realistic values of attainable efficiencies, especially at higher
temperatures.
3.2
Steady state temperature variation of spectrally selective coating with optical
concentration ratio
Optical concentration using Fresnel lens focuses the solar irradiation onto the absorber
coating which heats up the sample. The temperature recorded at steady state when the absorber is
not coupled to the STEG is attained as a result of energy balance between the incoming solar
flux on the surface and the thermal losses due to re-radiation, natural air convection and heat
conduction to the contact surfaces. This temperature varies as the optical concentration is varied
because of change in the heat flux incident on the absorber.
The high temperature spectrally selective coating absorbs nearly 96% of the incident flux,
and emits only 7% (at 82°C) of energy in the form of re-radiation which results in higher steady
state temperatures as compared to normal surfaces like stainless steel or copper substrates. From
the thermal performance tests, it was observed that a maximum temperature of 512°C was
obtained on this coating at a concentration ratio of 100. At the same concentration ratio, the
temperature of polished stainless steel surface was found to be only 340°C, which demonstrates
the effectiveness of the selective coating. The coating was found to be thermally stable in air up
to about 480°C, beyond which it suffered some degradation. As the concentration ratio was
increased from 10 to 100, the temperature of the selective coating was found to increase from
120 to 512°C. This shows the effect of optical concentration on the steady state temperatures that
can be achieved using the spectrally selective absorber.
9
The numerical simulations of the above experimental setup were performed in COMSOL
Multi-physics. A heat transfer model with surface to surface radiation was created for the
experimental setup with appropriate boundary conditions to include conductive, convective and
radiative losses. The absorptance in the solar radiation spectrum region and emittance in the IR
region for the materials used were also given as inputs to the model. The physics controlled mesh
feature of COMSOL was used for meshing (Fig. 3(a)) and the computational results obtained
showed a good match with the experimental observations. Fig. 3(b) shows the simulated
temperature contour of the experiment geometry at a concentration ratio of 75 at which an
absorber temperature of 484°C was predicted by the COMSOL model. Fig. 4 shows a
comparison of the experimental and simulated results for the spectrally selective coating and
polished stainless steel surface at different concentration ratios.
3.3
Heat transfer and efficiency analysis of an STEG using a spectrally selective coating
The basic principle of a concentration photothermal solar energy converter is based on a
simple energy balance at the surface of the absorber. A concentration system is employed so that
the absorber receives more number of photons per unit area per unit time and therefore a higher
flux as compared to the direct normal radiation. Some part of the radiation is reflected, some part
is lost as re-radiated energy and the remaining is absorbed, which can be used to extract valuable
energy. This extracted energy is used to create a temperature gradient to harvest power when
coupled to a thermoelectric device.
The heat balance at the surface of the absorber yields the following equation assuming no
conduction losses due to contact [23,24]:
4
4
𝑄𝑢𝑠𝑒𝑓𝑢𝑙 = 𝛼𝑋𝑞 − 𝜎𝜖(𝑇𝑎𝑏𝑠
−𝑇𝑠𝑢𝑟𝑟
)−ℎ(𝑇𝑎𝑏𝑠 − 𝑇𝑠𝑢𝑟𝑟 ),
10
1
here, α is the absorptance, X is the concentration ratio, q is the heat flux per unit area, σ is the
Stefan-Boltzmann constant, ε is the emittance, h is the convective heat transfer coefficient, and
Tabs and Tsurr are the temperatures of the absorber and surroundings respectively.It is seen from
Eq. 1 that the useful heat available for power generation increases as the absorptance of the
surface increases and the emittance of the surface decreases. Therefore, both absorptance and
emittance play an important role in this energy conversion technique and this is where the
effectiveness of a spectrally selective coating can be utilized. The high absorptance of such a
coating for the lower wavelength solar irradiation and its low emittance in the thermal IR reradiation range make it an ideal absorber for this application.
The absorber efficiency is defined as the ratio of the useful heat energy that can be passed
on to power generation stage to the energy that is incident on the absorber surface. It is
dependent on the absorptance, the incident heat flux and the re-radiation losses. As the
temperature of the absorber increases, this efficiency decreases because the re-radiation losses
are directly proportional to T4. An increase in the losses leads to lesser amount of useful heat
being transferred which results in lesser efficiency of the absorber coating. Eq. 2 shows this
dependence of efficiency on temperature [25].
𝜂𝑎𝑏𝑠𝑜𝑟𝑏𝑒𝑟 =
4
𝑄𝑎𝑏𝑠 − 𝑄𝑙𝑜𝑠𝑠
𝜎𝜖𝑇𝑎𝑏𝑠
⇒ 𝜂𝑎𝑏𝑠𝑜𝑟𝑏𝑒𝑟 = 𝛼 −
𝑄𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡
𝑋𝑞
2
Fig. 5 shows a plot of the absorber efficiency and re-radiation losses of the spectrally
selective coating used in the experiments. This analysis has been done assumingvariation of
emittance with temperature as per recorded measurements and considering only thermal reradiation losses so as to illustrate the efficiencies that can be achieved in vacuum. The
concentration ratio required for a particular temperature has been taken from the experimental
results shown in Fig. 4. Convective losses to surroundings and conduction losses from absorber
11
to other surfaces in thermal contact have not been considered. The values of the heat flux,
concentration ratio and absorptance have been taken according to the experimental conditions for
the selective coating used.
When the absorber coating is coupled to a thermoelectric device, the useful heat in the
form of thermal energy is utilized to generate electrical power. The thermoelectric device used in
such an STEG has its own efficiency of energy conversion which depends on the temperature
gradient across the device and the temperature dependent figure of merit (ZT) of the
thermoelectric materials employed:
𝑆 2𝜎
𝑍𝑇 =
𝑇 ,
𝑘
3
where S is the Seebeck coefficient, σ is the electrical conductivity and k is the thermal
conductivity of the thermoelectric material. The maximum efficiency of a thermoelectric device
[26] is given by:
𝜂𝑇𝐸 = (1 −
𝑇𝐶 √1 + 𝑍𝑇𝑀 − 1
)
,
𝑇𝐻 √1 + 𝑍𝑇𝑀 + 𝑇𝐶
4
𝑇𝐻
here, TCand TH are the temperatures (in K) of the hot and cold side of the thermoelectric device,
Z is the effective figure of merit in the operating temperature range and TM is the mean of the hot
and cold side temperatures. From the above equation, it is clear that the thermoelectric device
efficiency increases as the hot side temperature increases and as ZT of the thermoelectric
material increases.
The net system efficiency of solar to electrical energy conversion can be expressed as a
product of the absorber subsystem efficiency and the thermoelectric device subsystem efficiency.
Fig. 6 shows the variation of the system efficiency with ZT for the spectrally selective coating
used at different values of temperature. The concentration ratio for each temperature has been
12
chosen from the experimental observations of the previous section (Fig. 3) and emittance
variation with temperature has been accounted for from the high temperature emittance data. The
value of ZT depends on the thermoelectric materials used in the module. The module used in the
present experiment has an effective ZT of close to 0.4 in the temperature range of interest [27].
Values of ZT more than 3 have been reported at temperatures close to 600 K for quantum-dot
super-lattices [28]but generally a benchmark value of ZT = 1 is taken for predicting the
performance of a thermoelectric device [12,14].
It is seen that, theoretically, an efficiency of about 4.67% can be achieved with the present
STEG setup when operated at a hot side temperature of 300°Cwith the cold side at 30°C. It
should be noted that under similar experimental conditions, a system efficiency of 9.28% with an
effective ZT of 1 and 16% with an effective ZT of 3, can be achieved in the operating
temperature range. It is also observed that, for a particular value of ZT, the increase in efficiency
from one temperature to the next decreases at higher temperatures. Therefore, the different
temperature lines are closer to each other at higher temperatures. This is because the rate of
increase of re-radiative losses becomes very high at higher temperatures as also apparent from
Fig. 5. This affects the absorber efficiency and therefore the net system efficiency.
3.4
Performance characterization of STEG
According to Seebeck’s relation, the open circuit voltage generated by the thermoelectric
module is directly proportional to the temperature difference between the hot and cold surfaces.
The constant of proportionality is the difference between the Seebeck coefficients of the p-type
and n-type materials used in the module. This relationship is expressed as [17,19]:
𝑉𝑂𝐶 = (𝑆𝑝 − 𝑆𝑛 )(𝑇𝐻 − 𝑇𝐶 )
.
5
13
The thermoelectric generator used in the experiments consisting of three modules
coupled to the solar selective coating, can be modeled using the electrical circuit shown in Fig. 7.
The circuit consists of the open circuit voltage source, VOCgenerated by virtue of the temperature
difference across the three modules, the internal resistances of each of the modules in series
causing a voltage drop, and the load resistance, RL. The current flowing in the circuit through the
load resistance is given by:
𝐼=
𝑉𝑂𝐶
,
𝑅𝐿 + 𝑅𝑖𝑛𝑡
6
where, Rint is the total internal resistance of the three modules. The power output across the load
is then calculated as:
𝑃𝑜𝑢𝑡 = 𝐼 2 𝑅𝐿 = (
𝑉𝑂𝐶
)2 𝑅𝐿 .
𝑅𝐿 + 𝑅𝑖𝑛𝑡
7
Therefore, knowing the open circuit voltage generated for a given temperature of the hot
side of the thermoelectric module, the theoretical power output can be calculated for different
loads using the above equation. In order to obtain maximum power output, the load resistance
should be set equal to the internal resistance of the circuit. In this matched-load situation, the
electrical power generated is given by:
𝑃𝑜𝑢𝑡 𝑡 =
𝑉𝑂𝐶 2 𝑉𝑂𝐶 2
=
.
4𝑅𝐿
4𝑅𝑖𝑛𝑡
8
The open circuit voltage measurements performed at different temperatures have been
plotted in Fig. 8(a). A maximum open-circuit voltage of 5.4 V has been obtained at a temperature
of 280°C of the module hot side. This plot has been used to calculate the theoretical power
output expected at the different temperatures under matched load condition using Eq. 8. The
actual power output at matched load for a particular temperature has been calculated using the
output voltage and current measurements shown by the ammeter and voltmeter, respectively.
14
𝑃𝑜𝑢𝑡 𝑚 = 𝑉𝑜𝑢𝑡 × 𝐼
9
The incident power is calculated using the average solar flux irradiation. This value was
around 845 W/m2 on the days when the measurements were carried out. Therefore, the incident
power for a particular concentration ratio X can be evaluated as:
𝑃𝑖𝑛𝑐 = 845 × 𝐴 × 𝑋 ,
10
where, A is the area of the selective absorber. The incident and measured electrical power are
used to calculate the actual system efficiency of the STEG system. The net system yield is
estimated from:
𝜂𝑆𝑇𝐸𝐺 =
𝑃𝑜𝑢𝑡 𝑚
.
𝑃𝑖𝑛𝑐
11
Fig. 8(b) shows the theoretical and measured electrical power outputs (in W) at matched
load condition for the STEG setup of the present work plotted against the incident power on the
absorber coating. It is seen that a maximum measured power of 0.66 W is observed for an
incident power of 57 W at a concentration ration of 62. The graph in Fig. 8(c) shows the
measured power output (in mW) as a variation of hot surface temperature and the corresponding
values of net system efficiency as calculated from Eq. 11. As the temperature increases from 170
to 220°C, there is a small increase in efficiency from 0.83 to 1.18% and a maximum of 1.25%
system efficiency is measured at 260°C.
At a hot side temperature of 280°C (concentration ratio of 62 suns), the performance of
two STEG arrangements has been compared in Fig. 9. The plot in Fig. 9(a) shows the values of
the voltage and power outputs for a single thermoelectric module as the external load resistance
of the circuit is varied. Under the same experimental conditions, a three module stacked
arrangement setup has also been tested by varying the load. As expected, the voltage output
obtained at a particular load is greater for the three module setup, whereas the current through
15
the load is greater for the single module setup. This is due to the higher internal resistance of
three modules in series. The peak power values obtained at the individual matched load
conditions of 4 Ω and 12 Ω have also been compared to illustrate the higher power output
potential of the stacked arrangement in this experiment. The peak output power of the stacked
arrangement at this temperature is observed to be 0.66 W as compared to 0.57 W for the single
thermoelectric module. The temperature near the cold side of the thermoelectric module was
monitored at the Al heat sink and was found to be around 65±2°C for the stacked arrangement
and about 145±2°C for the single module arrangement. Therefore, because of the much larger
temperature gradient, a higher power output was expected. The reason for not obtaining a
significant power increase in the stacked arrangement can be attributed to loss in power due to
the temperature mismatch when thermoelectric generators are electrically connected in series
[29].
3.5
Discussion
The system performance achieved in this study compares well with related experiments
done in the field of power generation using solar thermoelectrics.
Daniel Kraemer et al.
[14]have developed flat panel thermoelectric generators that work on thermal concentration
using large absorbers to achieve a peak efficiency of 4.6% by imposing a 200°C temperature
gradient with the cold side at 20°C. They have achieved the highest reported efficiency for flatpanel STEG using a new approach of thermal concentration employing nanostructured
thermoelectric materials having an effective ZT of 1.03 at optimal operating conditions.
However, to achieve this performance, they have used very large solar absorbers having an area
of about 250 times that of the thermoelectric element, and have performed the efficiency
measurements in vacuum by controlling the cold side temperature. The power required for
16
vacuum generation and cost of large area deposition of selective coatings would result in a net
system input much more than that of the present study.
R. J Ram and R. Amatya [16]have measured an efficiency of 3% for commercial Bi2Te3
modules at an optical concentration of 66 suns. This is the highest system efficiency that has
been achieved using a commercial thermoelectric module in an STEG. The thermodynamic
analysis presented in their work predicts an efficiency of 5.6% for an STEG at an optical
concentration of 120 suns. This agrees well with our theoretical analysis presented in Fig. 6 from
which it is seen that at a value of ZT = 0.4 and a hot side temperature between 400 and 500°C
(which corresponds to an optical concentration of 120 suns in our experimental setup), a system
efficiency between 5.6% and 6.3% can be achieved by maintaining the cold side at 30°C. The
higher value of the predicted efficiency in our study may be due to the high selective properties
of the coating that has been considered for this analysis.
In another work, Emmanuel Ogbonnaya et al. [30]have developed STEGs for micropower applications that generate 9.15 mW using spectrally selective coatings by employing a
collector cooled by a copper water block. They have demonstrated the superiority of a selective
coating by comparing the performance of the STEG using a non-selective absorber in which only
2.01 mW was generated. However, the thermoelectric efficiency of 2.87% predicted in their
study for a temperature difference of only 7°C is highly overestimated as the unit of temperature
considered for calculation is incorrect (°C instead of K). Though, they have not employed optical
concentration, more power can be harnessed by doing so without the requirement of external
cooling on the cold side. Similarly, A. Pereira et al. [17]have performed high optical
concentration experiments using SiGe thermoelectric materials to achieve a system efficiency of
1.6% at a temperature gradient of 400°C. They have performed predictive efficiency
17
computations for high working temperatures (700-900°C) at high optical concentrations (>100
suns). One of their important conclusions was that SiGe based thermoelectric materials require a
hot side temperature greater than 700°C to outperform Bi2Te3 based thermoelectric materials at
250°C. Therefore, it is seen form their simulations that Bi2Te3 materials are the best contenders
for solar thermoelectric power generation in the temperature range of operation of our present
study.
In yet another work, LaurynBaranowski et al. [25] have performed a theoretical modeling
analysis of the STEG efficiency and have predicted an efficiency of 15.9% at a hot side
temperature of 1000°C for today’s thermoelectric materials (ZT = 1) and an efficiency of 30.6%
at 1500°C and ZT = 2. However, it should be noted that the predictive model presented in this
work assumes the effective module ZT of the thermoelectric material to be constant in the
temperature range between the hot and cold sides. But, in practice, the figure of merit varies with
temperature and there will be a significant decrease in the predicted efficiencies of their model if
this variation is considered because of the very large temperature gradient considered in their
analysis.
The efficiency achieved in the present study is slightly on the lower side compared to the
above mentioned works but the experimental method used does not require any source of power
input to the system for cooling of the cold side or for the creation of an evacuated environment.
Our objective was to determine the performance of the spectrally selective coating when coupled
to an STEG with minimum input power and system cost. Additionally, the experimental data of
temperature variation of the coating with optical concentration and its emittance variation with
temperature have enabled the development of a physical and realistic theoretical efficiency
model for the STEG coupled to the selective absorber. Further research to improve efficiency
18
using this setup can be done by reducing the convective losses using a glass cover and employing
a larger absorber to utilize the concept of thermal concentration. Also, by employing a vacuum
chamber and higher ZT materials, much higher system efficiencies can be achieved but with
significant increase in total cost of system design.
4
Conclusion
In this study, we have demonstrated the effectiveness of a high temperature spectrally
selective absorber coating (α = 0.956, ε = 0.07) and investigated its performance in
thermoelectric power generation. We have experimentally established that the spectrally
selective coating is thermally stable up to a temperature of approximately 512°C for short
durations under high optical concentration ratio (100) which is about 150°C greater than that of
uncoated stainless steel substrate surface. We have carried out emittance measurements for the
coating at different temperatures and observed that the emittance increases by about three times
(from 0.07 to 0.22) as the temperature is increased from 100 to 500°C.
We have also presented a comprehensive analysis of the variation of STEG system
efficiency with the figure of merit of the thermoelectric materials at different temperatures for
the selective coating that has been used in the present study. This has been achieved by
performing temperature measurements for the STEG setup at different optical concentration
ratios and integrating these experimental observations with the variable emittance measurement
data of the coating. This model predicts an efficiency of 4.67% for our optical concentration
STEG system at a hot side temperature of 300°C and cold side temperature of 30°C for a
commercial thermoelectric module having ZT~0.4 in vacuum conditions. With today’s
thermoelectric materials (ZT = 1), we can achieve up to 9.3% system efficiency using our
coating under the same operating conditions, assuming an ideal optical system.
19
In the last part of the paper, we have presented experimental results for an STEG system
consisting of the high selectivity coating coupled to a series stacked combination of three
commercial Bi2Te3 thermoelectric modules. A natural convective cooled heat sink has been
incorporated with the STEG which has enabled the design of an effective low cost virtual power
source for micro-power applications. A solar to electrical conversion efficiency of 1.2% has been
achieved at a hot side temperature of 280°C under an optical concentration ratio of 62, at an
input solar flux of ~845 W/m2 under matched load condition.
Acknowledgements
The authors would like to thank Director CSIR-NAL for giving permission to publish this
work. Daniel Kraemer and Lee Adragon Weinstein of Massachusetts Institute of Technology,
USA are thanked for carrying out the high temperature emittance measurements for the
spectrally selective coating. Dr. Manju Nanda, ALD, NAL is thanked for making the electrical
circuit. This work is partially supported by CSIR under TAPSUN Program (NWP 0054).
20
List of figures
Fig. 1 (a): Schematic of STEG experimental setup showing- (1) Fresnel lens (650 mm X 490
mm); (2) Solar concentration radiation ; (3) Aluminum cover plate; (4) High temperature solar
selective coating (35 mm X 35 mm); (5) Stainless steel substrate with IR reflector; (6) Variable
resistor; (7) Thermoelectric modules ; (8) Aluminum heat sink.
(b): Thermally parallel stacked arrangement of commercial Bi2Te3 thermoelectric modules
electrically connected in series.
(c): Rooftop experimental setup of STEG with adjustable Fresnel lens frame to ensure that
absorber receives direct normal irradiation of the sun.
(d): Concentrated solar focal spot on the spectrally selective absorber surface at a concentration
ratio of 65. Aluminum heat sink is visible below.
Fig. 2: Reflectance data of spectrally selective coating recorded in UV-Vis-NIR and IR regions,
displaying very low reflectance in the solar spectrum region and very high reflectance in the IR
region.
Fig. 3: (a) Physics controlled mesh in COMSOL Multiphysics of the temperature measurement
setup - The central region of the geometry consisting of solar absorber, ceramic brick insulator
and aluminum base plate is divided into 3456 tetrahedral finite elements for computation.
(b): Temperature contour of the simulated results in COMSOL at an optical concentration ratio
of 75 - A maximum temperature is obtained at the center of the selective coating as the ceramic
brick insulator ensures that the thermal energy is contained in the region of the coating to obtain
maximum temperature.
Fig. 4: Comparison of experimental and numerical temperatures for SS substrate and solar
selective coated substrate at different concentration ratios - A selective absorber obtains higher
temperatures compared to a normal surface by virtue of its high absorptance and low emittance.
Fig. 5: Re-radiation losses and corresponding absorber efficiencies for the selective coating as a
function of surface temperature at a heat flux of 865 W/m2. Inset graph shows variation of
emittance with temperature for the coating that has been incorporated into the efficiency relation
to predict realistic efficiencies that can be achieved in evacuated conditions.
Fig. 6: Variation of ideal system efficiency for the STEG system with ZT of the thermoelectric
material at different hot side temperatures with cold side at 30°C - Region of interest shows the
feasible operation region for the scope of the present experimental study that uses a commercial
Bi2Te3 thermoelectric module.
21
Fig. 7: Electrical circuit model of the STEG consisting of 3 thermoelectric modules in series
each having internal resistance of ~ 4Ω. Variable load is connected to measure maximum output
power at matched load condition.
Fig. 8: (a) Variation of open-circuit voltage with hot side temperature of the thermoelectric
module setup obtained experimentally.
(b) Theoretical and measured power output as a function of the incident power on the absorber
surface – A maximum power output of 0.66 W is recorded at an incident power of 57 W which
corresponds to an optical concentration of 62 suns in the experimental setup used.
(c) Measured power output plotted against the hot side temperature and corresponding STEG
system efficiencies which show utilization of the input solar flux – Maximum efficiency of 1.25
% recorded at 260°C for the STEG system of the present experimental study.
Fig. 9: Voltage and power outputs of the STEG as a function of the load resistance R L for (a)
Single module arrangement and (b) Stacked arrangement consisting of three modules in series –
Peak power in each case (0.57 and 0.66 W, respectively) correspond to the load matching
situation. Higher power output in (b) demonstrates optimized arrangement for the present
experimental setup.
22
References
1. B.Parida, S. Iniyan and R.Goic, A review of solar photovoltaic technologies, Renewable
and Sustainable Energy Reviews 15(2011) 1625-1636.
2. M. Ortega, P. delRío and E. A. Montero, Assessing the benefits and costs of renewable
electricity, Renewable and Sustainable Energy Reviews 27(2013) 294–304.
3. J.A Duffie and W.A Beckman, Solar Engineering of Thermal processes, 2nd edition,
Wiley-Interscience, New York, 1991.
4. G. Boyle, Renewable Energy: Power for a Sustainable Future, 3rd edition, 2012.
5. P.Sundarraj, D. Maity, S.Sinha Roy and Robert A.Taylor, Recent advances in
thermoelectric materials and solar thermoelectric generators, RSC Adv. 4 (2014) 4686046874.
6. H.S. Lloyd, Solar thermoelectric generators, 1961, Patent No. US 2,984,696.
7. P. Lee, L. Cai, P. Zhai, X, Tang, Q. Zhang and M. Niino, Design of a concentration solar
thermoelectric generator, Journal of Electronic Materials 39(2010) 1522-1530.
8. J. Xiao, T. Yang, P. Li, P. Zhai and Q. Zhang, Thermal design and management for
performance optimization, Applied Energy 93(2012) 33-38.
9. Jincan Chen, Thermodynamic analysis of a solar driven thermoelectric generator, Journal
of Applied Physics 79 (1996) 2717-2721.
10. B. O. Seraphin, Solar Energy Conversion, Springer-Verlag, Berlin, 1979.
11. N. Selvakumar and H.C. Barshilia, Review of physical vapor deposited spectrally
selective coatings for mid- and high-temperature solar thermal applications, Solar Energy
Materials and Solar Cells 98 (2012) 1-23.
23
12. H. J. Goldsmith, J. E. Giutronich and M. M Kaila, Solar thermoelectric generation using
Bismuth Telluride alloys, Solar Energy 24 (1980) 435-440.
13. M. Telkes, Solar Thermoelectric Generators, Journal of Applied Physics 265 (1954) 765777.
14. D. Kraemer, B. Poudel, H. Feng, J.C. Caylor, Bo Yu, X. Yan, Yi Ma, X. Wang, D. Wang,
A. Muto, K.McEnaney, M.Chiesa, Z.Ren and G. Chen, High-Performance flat-panel
solar thermoelectric generators with high thermal concentration, Nature materials 10
(2011) 532-538.
15. K.McEnaney, D. Kraemer, Z.Ren and G. Chen, Modeling of concentrating solar
thermoelectric generators, Journal of Applied Physics 110 (2011) 074502 1-6.
16. R.
Amatya
and
R.
J
Ram,
Solar
Thermoelectric
Generator
for
micro-
powerapplications,Journal of Electronic Materials 39(2010) 1735-1740.
17. A. Pereira, T. Caroff, G. Lorin, T. Baffie, K. Romanjek, S. Vesin, K. Kusiaku, H.
Duchemin, V. Salvador, N. Miloud-Ali, L. Axila, J. Simon, High temperature solar
thermoelectric generator – Indoor characterization method and modeling, Energy 84
(2015) 485-492.
18. G.N Tiwari and S. Dubey, Fundamentals of Photovoltaic Modules and Their
Applications, RSC Energy Series No.2, 2010.
19. Sarah Watzman, Design of a Solar Thermoelectric Generator - Thesis, Ohio State
University, 2013.
20. D.K Paul and P. Gardner, Maximum power transfer theorem: A simplified approach, Int.
Journal of Electrical Engineering Education 35 (1998) 271-273.
21. H. C. Barshilia,Indian Patent Application No. 2226DEL2014.
24
22. D. Kraemer, K. McEnaney, F. Cao, Z.Ren, G. Chen, Solar Energy Materials and Solar
Cells 132 (2015) 640-649.
23. Frank. P Incropera, Fundamentals of Heat and Mass Transfer, 6th edition, 2006.
24. G. Chen, Theoretical efficiency of solar thermoelectric energy generators, Journal of
Applied Physics 109 (2011) 104908 1-8.
25. L. Baranowski, G. Snyder and S. Toberer, Concentrated Solar Thermoelectric
Generators, Energy and Environmental Science 5 (2012) 9055-9067.
26. M.L Olsen et al., A high temperature high efficiency solar thermoelectric generator
prototype, Energy Procedia 49(2014) 1460-1469.
27. Hi-Z Technology Inc.,www. hi-z.com.
28. J. Li, W. Liu and Li-Dong Zhao, High performance nanostructured thermoelectric
materials, NPG Asia Materials 2 (2010) 152-158.
29. A. Montecucco, J. Siviter and A. Knox, The effect of thermoelectric generators
electrically connected in series or parallel, Applied Energy 124 (2014) 47-54.
30. E. Ogbonnaya and L. Weiss, Micro Solar Thermal power harvesting using thermoelectric
generator,National Society of Black Engineers,Technical Proceedings of 38th Annual
Convention, PA, 2012.
25
26
27
28
29
30
31
32
33
34