Received: 4 May 2020

| Revised: 20 June 2020

| Accepted: 29 June 2020

DOI: 10.1002/ese3.784

RESEARCH ARTICLE

Multi-source ambient energy harvester based on RF and thermal

energy: Design, testing, and IoT application

Azamat Bakytbekov1 | Thang Q. Nguyen1 | Weiwei Li1 | Anton Lee Cottrill2 |

Ge Zhang2 | Michael S. Strano2 | Khaled N. Salama1 | Atif Shamim1

1
Computer, Electrical and Mathematical
Science and Engineering Department,
Abstract
King Abdullah University of Science and Billions of wireless sensing devices must be powered for IoT applications. Collecting
Technology, Thuwal, Saudi Arabia energy from the ambient environment to power sensor nodes is a promising solu-
2
Chemical Engineering Department,
tion. Solar energy has been one of the main sources of ambient energy due to its
Massachusetts Institute of Technology,
Cambridge, MA, USA availability, higher power density, and the maturity of the solar photovoltaic indus-
try. However, there are many scenarios (indoor environment, outdoor environment
Correspondence
during nighttime, poor weather conditions, underground, etc) where ambient solar
Azamat Bakytbekov, King Abdullah
University of Science and Technology, energy is either not available or not sufficient for practical applications. For such
Bldg. 3, Seaside, Fl. 2, 2247-WS07, Thuwal scenarios, other renewable sources of energy must be sought. Typically, not enough
23955, Saudi Arabia.
Email: [email protected]
power is collected from one ambient source to charge sensor nodes for continuous
operation. In this work, we present a multi-source energy harvester that collects RF
Funding information and thermal energy (both available 24 hours) from the ambient environment simulta-
King Abdullah University of Science and
Technology, Sensor Initiative, Grant/Award neously. The RF energy harvester is multi-band and collects power from GSM (900,
Number: OSR-2015-Sensors-2700 1800 MHz) and 3G (2100 MHz). The thermal harvester converts diurnal temperature
fluctuations to electrical energy using high thermal effusivity phase change material.
Extensive field testing has been performed in three different conditions—outdoors,
indoors, and buried underground—to highlight the usefulness of the multi-source
energy harvester in all these environments. When one source is disabled, the har-
vester still generates energy from the remaining active source and can enable con-
tinuous operation of futuristic IoT sensors. As a proof of concept, a real-world IoT
application is demonstrated, where temperature and humidity sensors are powered
by the multi-source energy harvester. Continuous robust operation of the sensors and
wireless data transmission after each 3.7 seconds are expected when both harvesters
operate in full mode. Scenarios, where only single thermal energy harvester or only
single RF energy harvester operates, are also demonstrated and data transmission
with average time intervals of 30 seconds and 9 minutes is achieved, respectively.

KEYWORDS
ambient energy sources, multi-source energy harvester, RF energy harvesting, self-powered IoT
wireless sensors, thermal energy harvesting

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original
work is properly cited.
© 2020 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.

Energy Sci Eng. 2020;8:3883–3897.  wileyonlinelibrary.com/journal/ese3 | 3883

3884
|    BAKYTBEKOV et al.

1 | IN TRO D U C T ION

It is estimated that the number of wireless devices will

reach 75 billion in 2025.1 This is possible because of the
emerging Internet of things (IoT) paradigm, which will
connect a huge number of devices and sensors into one
network to optimize and facilitate people's daily activi-
ties. It is clear that the power consumption of these de-
vices is the main limitation with regard to large-scale
implementation of the IoT network. Ideally, the sensor FIGURE 1 Multi-source ambient energy harvester
nodes must be autonomous, self-powered, and low-cost.
Collecting energy from the ambient environment and and microbial fuel cells are combined to achieve a reliable
using it to power IoT sensor nodes might solve the issue operation of an underwater sensor network. Furthermore,
of power consumption. In general, IoT devices can ben- hybrid harvesters containing different combinations of
efit from solar energy, as light is available in most out- single renewable energy harvesters are presented in.6,7 All
door locations and the power density of solar energy is the hybrid harvesters mentioned above are dependent on a
1000 times higher than the average of other ambient en- specific location and source, such as wind, water flow, and
ergy sources.2 However, in some scenarios, such as night vibration, among others. Thus, the limitation of these hy-
time, cloudy days, and indoor environments, or situations brid harvesters is that they cannot be utilized universally in
in which solar cells are covered by dust or shadowed by any environment, as their sources are location specific. In
nearby objects, the solar cells do not operate at all or other words, they cannot be used in the circumstances de-
their efficiency significantly decreases. Under such cir- scribed above (indoors, night time outdoors, poor weather
cumstances, other ambient energy sources must be uti- conditions, underground, etc).
lized. However, there are limitations, such as low power Harvesting ambient RF energy and thermal energy has
level availability in the ambient environment and imper- great potential, as both sources are available 24 hours a day
fect efficiencies of different harvesters. In addition, the and are not location specific. RF energy harvesting can be
typical energy collected from stand-alone ambient energy done through a receiving antenna that can capture elec-
harvesters is low to drive wireless sensors. Stand-alone tromagnetic signals from available wireless bands and a
harvesters may also not be able to provide continuous op- rectifying circuit that transforms these signals into useful
eration due to functional constraints as well as external electrical energy. On the other hand, the thermal energy
factors, such as the unpredictable availability of ambient harvester can generate energy from the diurnal cycle (tem-
energy sources. Therefore, it is advantageous to combine perature fluctuations in an environment, particularly be-
several ambient energy harvesters into one module to tween day and night time). The available power density of
achieve (a) sufficient collected energy to power sensor RF and thermal sources can be combined to generate more
nodes and (b) continuous operation of sensor nodes so power to be suitable for IoT applications. An RF energy
that they can compensate for each other's limitations. harvester can be improved by harvesting from multiple
Several ambient energy sources are available, such as available bands at the same time so that output power can
wind, water flow, vibration, waste heat, temperature dif- be increased.8 In turn, a thermal energy harvester can be
ference, and electromagnetic radiation, among others. The optimized and tuned specifically to the temperature fluctu-
choice of energy harvesters that are integrated into one ation of the given environment.9
module depends on the type of application, location, en- In this paper, we present a multi-source ambient energy
ergy source availability, surrounding environment, and so harvester that collects energy from RF signals and diur-
forth. A considerable amount of research has been done nal temperature fluctuations to power sensor nodes of the
in designing hybrid multi-source energy harvesters to in- IoT (Figure 1). We perform smart system integration by
crease the amount of collected energy and address the lim- combining RF and thermal energy harvesters into one mod-
itations of single-source harvesters. In,3 a harvester that ule without them degrading each other's performance. The
collects energy from wind, light, and water flow is used to multi-source harvester is characterized in many locations,
power a data acquisition unit. In,4 a power conversion effi- such as indoor, outdoor, and underground environments.
ciency (PCE) boost is demonstrated by using a combination As a proof of concept, we show an IoT application where
of electromagnetic energy (RF) and kinetic energy (vibra- continuous operation (with certain time intervals) of tem-
tion) harvesters. Another example of a hybrid harvesting perature and humidity sensors is achieved when powered
system is presented in,5 where piezoelectric harvesting by the multi-source ambient energy harvester.
BAKYTBEKOV et al.    
| 3885

2 | M U LTI-S OU RC E E N E RGY shown both experimentally and theoretically to enhance the

H A RV ESTE R D E SIGN A N D power density of the thermal harvester.9,10
FAB R ICAT IO N We first develop an analytical model (Figure 2A) with no
flux at the internal boundaries (xj = 0) and no heat transfer
2.1 | Thermal energy harvester design and resistance at the external boundaries (xj = Lj). From our pre-
fabrication vious analytical model in,9 the time-averaged power density
(pavg) for a thermal harvester subjected to sinusoidal input
Transient temperature variations are pervasive in our sur- temperature fluctuations is described as:
roundings, coming from the diurnal cycle, human activity,
and exhaust heat from vehicles and computers, among others. (1)
√
pavg ≈ 𝜂e1 TA 𝜔Q
However, temperature fluctuation has not yet been exploited
as a ubiquitous energy source. A novel device called a thermal where η is the efficiency of the thermoelectric heat engine,
resonator for transient thermal energy harvesting has recently as shown in Equation (2), e1 is the thermal effusivity of
been developed.10 The key is to convert temporal fluctuations mass 1, as given by Equation (5), TA is the amplitude of
into a spatial gradient, by having thermal masses with dis- temperature fluctuations, 𝜔 is the angular frequency of
tinct properties on two sides of a thermoelectric generator. temperature fluctuations, and Q is a tuning factor that is
Two characteristics distinguish a thermal resonator from pre- given in Equation (3).
vious works: (a) the incorporation of high thermal effusiv-
ity (e = k𝜌Cp ) materials and (b) the capability of targeting
√ √
� �
T 1 + ZT − 1
any input frequencies. This leads to a device that generates 𝜂 = 1− C (2)
TH
√
electricity from broadband temperature variations, requiring 1 + ZT + TC ∕TH
neither a static gradient nor a high frequency oscillation. In
a thermal resonator (Figure 2A), thermoelectric modules are where TC is the cold reservoir temperature, TH is the hot reser-
sandwiched between an aluminum heat fin and a metal foam voir temperature, and ZT is the thermoelectric figure of merit,
impregnated with the phase change material (PCM). The which is assumed to be 1 for the commercial Bi2Te3 thermo-
PCM is a material that undergoes a phase transition (for ex- electrics used in this work.
ample from liquid to solid), which is usually a constant-tem-
perature process, and this temperature is the phase transition 1
� � �√ �� �2
Q= Re sech iv − 1 (3)
temperature. The PCM has much higher effective heat capac- 8
ity than common materials due to its latent heat (released or
absorbed energy during the constant-temperature process). where ν is the dimensionless frequency:
When the thermal energy harvester is operated near the phase
transition temperature of PCM, a great amount of heat can L21
v=𝜔 (4)
be stored or released with very little temperature change. In 𝛼1
contrast to the PCM, a heat fin is made of a material with low
heat capacity, which will follow the changing environmental L1 is the length of thermal mass 1 and α1 is the thermal
temperature closely. This will create a time-varying tempera- diffusivity of thermal mass 1. For PCM, the apparent thermal
ture difference across the thermoelectric generator (white effusivity due to latent heat is:
pads in Figure 2A), which can harvest this difference to pro-
duce electricity. The high heat capacity of PCM is combined
√
k𝜌h
with the high thermal conductivity of a metal foam to cre- eapp = (5)
2TA
ate the high thermal effusivity. This characteristic has been

F I G U R E 2 (A) Schematic diagram of

the analytical model. (B) Analytical model
prediction of the average power density
as a function of thermal capacitance and
conductance
3886
|    BAKYTBEKOV et al.

where k is the thermal conductivity of the material (assume no 2.2 | RF energy harvester design and
change in thermal conductivity with phase), ρ is the density, fabrication
and h is the latent heat per mass.
Thermal effusivity (e1) measures a material's ability to ex- With the rapid development of wireless technologies, there
change heat with its surroundings. The comparison between a are more electromagnetic waves around us that carry sig-
pure PCM (eicosane, E) and a metal/PCM composite (Ni/E) nals from different standards, such as GSM, 3G, WiFi, and
in Figure 2B exhibits the enhancement of thermal effusiv- so forth. Intense usage of wireless devices and services di-
ity through the embedded metal matrix. We also construct rectly leads to more electromagnetic energy around us, as cell
a numerical model accounting for the heat flux through the phone base stations radiate electromagnetic signals in an om-
thermoelectric modules, and the convective heat transfer at nidirectional fashion. This ambient electromagnetic energy
the external boundaries. The numerical model agrees better can be collected and converted into useful electrical energy.
with the experimental value than with the analytical one. The The RF energy harvester consists of an antenna that can
thermal resonator is optimized to work persistently in a harsh capture wireless signals from the ambient environment and
environment, such as the desert of Saudi Arabia.9 circuits (impedance matching circuit, rectifier circuit) and
Ultra-high-molecular-weight (UHMW) polyethylene with that converts AC signals into DC (Figure 4A). First, an am-
the dimensions of 10 × 10 × 15 cm3 is used as a package for bient RF power measurement is performed to identify the
the thermal harvester (Figure 3A). A high thermal effusiv- most powerful RF bands. Measurements are conducted eight
ity PCM composed of eicosane (T* = 36°C; h = 250J g−1) times a day in five different locations to obtain information
vacuum impregnated within a highly porous and highly about the time and location dependence. To increase the ac-
thermally conducting nickel matrix (Ni/E) is fabricated spe- curacy of the measurements, cable and reference antenna
cifically for application in the desert environment of Saudi losses are de-embedded, and the power averaging option
Arabia. Eicosane is chosen because its phase transition tem- on a spectrum analyzer is selected.11 The result of the mea-
perature matches the desired temperature, and because of its surement shows that the most powerful bands are GSM900,
high heat storage capabilities. In the first step, the UHMW 3G, and GSM1800 in a descending order, compared with
container is filled with the phase change material (nickel other available bands, such as WiFi. This result is consid-
foam and eicosane), as shown in Figure 3B. Four heat en- ered valid, as most mobile phone operators in Saudi Arabia
gines are sandwiched between two copper plates and placed utilize GSM900 and GSM1800 for calls and 3G at 2.1 GHz
on top of the UHMW box (Figure 3C,D). The fabricated ther- for Internet connections. In addition, the maximum output
mal harvester is displayed in Figure 3E. power of GSM900 is more than the GSM1800, which is also

Nickel mesh Copper

& Eicosane Plate

(A) (B) (C)

Heat
Engines
F I G U R E 3 (A) UHMW container of
the thermal harvester. (B) UHMW container
filled with phase change material: nickel
mesh and eicosane. (C - D) Four heat
engines sandwiched between two copper
plates. (E) Fabricated thermal energy
(D) (E) harvester
BAKYTBEKOV et al.    
| 3887

F I G U R E 4 (A) Block diagram of the

RF energy harvester. (B) Cantor fractal. (C) Matching Rectifier Load
Antenna
Body of the thermal harvester. (D) Pieces Network
of UHMW and inner filling prepared for
characterization. Cantor fractal antenna with (A)
(B)
(E) one iteration (F) two iterations (G) three
iterations (H) Reflection coefficient of the
UHMW Copper
antennas with different iterations Plates
Box

Inner
Filling (D)

(C)

(E) (F) (G)

(H)

consistent with the collected data. The results of the ambient fractal pattern used for this project is called a Cantor fractal
RF power measurement can be seen in Figure S1. (Figure 4B). Its pattern is as follows: the first line segment is
copied and split up into three parts, and the middle section of
those parts is deleted. The same process is reiterated several
2.2.1 | Antenna design and fabrication times to achieve the required multi-band performance.
The antenna is designed using Ansys High Frequency
To harvest from three bands at the same time, a straightfor- Structure Simulator (HFSS) software by simulating the whole
ward approach would be to design three antennas and three body and inner filling of the thermal harvester to achieve ac-
rectifier circuits with three matching networks tuned to curate results (Figure 4C). Small pieces of the UHMW poly-
each of these frequency bands (900 MHz, 1800 MHz, and ethylene and inner filling (mix of eicosane and nickel foam),
2100 MHz). However, this solution is neither cost-effective as shown in Figure 4D, are prepared for the characterization
nor space-effective. Instead, an efficient solution is to de- in terms of electrical properties. The UHMW is characterized
sign a single rectenna (rectifying antenna) that can harvest on an impedance analyzer, and a dielectric constant of 2 and
from multiple bands simultaneously. For that, a multi-band loss tangent of 0.02 are measured at 1 GHz. As the UHMW
antenna and a multi-band matching network are required. is found to be quite lossy at higher frequencies, we fabricate
A fractal approach can be used to achieve multi-band per- an antenna on a different low loss material and attach it on
formance, as fractals can provide several resonances due to the side faces of the UHMW box, instead of printing the an-
the combination of different lengths.12 Fractals are structures tenna directly on the UHMW box. In this way, the effect of
with a self-repetitive pattern across different scales. The the lossiness of the UHMW can be minimized. The chosen
3888
|    BAKYTBEKOV et al.

F I G U R E 5 Current distribution at (A)

900 MHz. (B) 1800 MHz. (C) 2100 MHz.
(D) Reflection coefficient for different
segment widths. (E) Fabricated Cantor
fractal antenna. (F) Reflection coefficient
of the antenna. (G) Fabricated multi-source
energy harvester

substrate of the antenna is 3D printed Preperm ABS300 one resonance. So, three iterations are required to achieve
material with a dielectric constant of 3 and a loss tangent the triple-band antenna performance. Figure 4E-G shows
of 0.004 at 2.4 GHz. Metallization is done through screen the evolution of the antenna design with different number
printing of silver paste with a conductivity of 107 S/m.13 All of iterations and Figure 4H shows the reflection coeffi-
these real values are inserted into the HFSS software for ac- cients of these antennas. As it can be seen in Figure 4H,
curate results. The challenging part is to characterize a mix the first iteration's segment length is optimized to have a
of nickel mesh with eicosane that becomes a homogenous resonance at 2.1 GHz (black dotted curve). As at 2.1 GHz
mass and changes its phase from solid to liquid at different the wavelength is 82.5 mm, the length of the first Cantor
temperatures. The measurement of the inner filling through segment is 82.5 mm. Once the second iteration is added
the impedance analyzer does not lead to any result, as the mix following the Cantor fractal structure (Figure 4F), another
has characteristics of both metal and a dielectric. To facilitate resonance appears at around 1.7 GHz (blue curve). The
the building of the model of the thermal harvester on HFSS, length and width of the small separations between the main
it is assumed that (a) the filling of the box only contains bulk segments are important variables to adjust the resonance
nickel as it is difficult to simulate a metallic mesh and (b) position so the second resonance can be moved to 1.8 GHz.
the effect of nickel mesh is dominant compared to eicosane It is worth to notice that the resonance at around 1.3 GHz is
particles affecting the RF performance. the harmonic of the 2.1 GHz resonance which is also pres-
Two Cantor fractals are designed to be the two arms of ent in the previous black dotted curve of the first iteration.
an antenna. The idea behind the Cantor fractal design of the The third iteration can be added as shown in Figure 4G to
triple-band antenna is that each iteration is responsible for achieve the resonance at 900 MHz frequency. Once the
BAKYTBEKOV et al.    
| 3889

required resonances are in place, the impedance matching 2.2.2 | Rectifier and matching circuit
must be improved. It is considered an industry standard for design and fabrication
the impedance matching when the required band is lower
than −10 dB, which implies 90% power transmission to the To convert the harvested RF power from the antenna to us-
antenna (see the green line in Figure 4H). able DC power for subsequent circuits (sensors, transceiv-
Figure 5A-C shows the current distribution of the an- ers), a rectifier circuit must be employed. This circuit can be
tenna at 0.9 GHz, 1.8 GHz, and 2.1 GHz frequencies, implemented using either CMOS technology18 or the pack-
respectively. It can be noticed that the high currents are aged diode.19 In this work, a single series diode is employed,
placed in the corresponding iteration segments with respect as this is the best candidate for high sensitivity and minimum
to frequencies as expected from the design steps. Areas diode loss (one diode only). Conventionally, the harvested
with the high currents are the most sensitive locations for RF energy generates a small voltage at the output of the an-
impedance matching and thus, the width of the segments is tenna. The magnitude of this voltage is in terms of uV, which
a valuable parameter to optimize. Figure 5D displays the is unable to turn on a normal diode. Therefore, a special type
reflection coefficient for different widths of the segments. of diode with high sensitivity must be selected. A Schottky
As it can be seen from the graph, at 31 mm the resonance diode SMS 7630-079 from Skyworks20 is chosen since it
at 900 MHz is not quite matched (levels at −10 dB) and at has low junction capacitance and resistance, low series re-
27 mm, two higher resonances (1.8 and 2.1 GHz) are moved sistance, and, most importantly, low turn-on voltage. The
to the right so that the antenna does not work at 1.8 GHz. rectifier circuit is simulated in Keysight Advanced Design
Therefore, the optimized value of the width must be found System (ADS) software to obtain the input impedance of the
where the resonances are at the right frequencies with the diode. The challenge here is that the input impedance of the
decent level of matching. As it can be seen from the graph, rectifier varies with the input power level, load condition, and
the width of the segment must be 29 mm to achieve the best frequency (Figure S2). For example, the input impedance of
performance. the rectifier circuit at −20 dB m for 900 MHz, 1800 MHz,
The antenna of the RF energy harvester is fabricated on and 2100 MHz is 35 - j553, 15 - j274, and 12 - j252, re-
2 mm thick ABS300 substrate with a dielectric constant of spectively. However, the antenna described in Section 2.2.1
3 and a loss tangent of 0.004.14 The substrates are printed is designed to have an impedance of 50 Ω. To transmit the
on a Makerbot Replicator desktop 3D printer.15 The fractal maximum power from the antenna to the rectifier circuit, the
pattern of the antenna is metallized through screen printing input impedance of the rectifier circuit must be matched to
using conductive silver paste from DuPont.16 The fabri- 50 Ω. Therefore, we design a triple-band impedance match-
cated antenna is shown in Figure 5E. Two parts of the an- ing network consisting of four transmission line stubs and
tenna are placed on two adjacent side faces of the thermal two radial stubs. This impedance matching network is dif-
harvester box. As one antenna covers only two side faces ferent from the triple-band rectifier presented in8 due to its
of this box, we integrate two RF harvesters on the thermal different substrate and impedance conditions. The layout of
harvester box so that (a) the RF harvester has 360̊ azimuthal the circuit is analyzed and optimized by ADS EM simulation.
coverage around the thermal harvester and (b) the power The circuit board is fabricated using the same combination
collected by the RF harvester is doubled. The final version of 3D printing and screen printing as for the antenna fabrica-
of the multi-source ambient energy harvester can be seen tion. Schematic and fabricated versions of the circuit board
in Figure 5G. Figure 5F shows the reflection coefficient are shown in Figure 6A,B. The reflection coefficient and rec-
results of the simulation and measurement in one graph. It tification results are presented in Figure S3 in the supporting
can be noticed that the resonances at 1.8 GHz and 2.1 GHz information section.
merged with each other and produce one wide bandwidth
in measurement. However, green line indicates that the an-
tenna covers all the required bands as the curves are below 3 | FIELD TESTING AND IOT
−10 dB (equivalent of 0.1, power ratio of reflected power APPLICATION
to input power) industry standard line which implies 90%
power transmission from the antenna to the subsequent Field testing is conducted in three different scenarios to as-
circuit.17 So, it can be concluded that the antenna works sess the performance of the multi-source ambient energy har-
well at 900 MHz, 1800 MHz, and 2.1 GHz frequencies and vester where one source can compensate for the absence of
deviation from the simulation results might be caused by the other one. The multi-source energy harvester is tested in
the fabrication tolerances. After comparing the simulated an outdoor environment, indoor environment, and buried un-
and measured results, the previously made assumptions are derground. Outputs of each energy harvester are connected to
proven true, as the measured reflection coefficient of the DrDAQ data loggers from Picotech,21 where output voltages
antenna is similar to the simulation results. of the harvesters can be recorded with a certain sampling rate.
3890
|    BAKYTBEKOV et al.

(A) (B)

FIGURE 6 (A) Schematic diagram of the triple-band matching network and rectifier circuit. (B) Fabricated matching network and rectifier
circuit

(A) F I G U R E 7 (A) Multi-source energy

Multi-source harvester in the outdoor environment. (B)
energy Output voltage from the RF harvester. (C)
harvester Output voltage from the thermal harvester
Laptop

Data
Loggers

(B) (C)

In turn, the data loggers are connected to a laptop, where the voltage of the RF harvester is collected across a 5 kΩ load, as
above-mentioned parameters can be monitored in real time. it is chosen to achieve maximum PCE. Output voltages of the
Measurement is conducted for 24 hours for each of the loca- thermal harvester are collected across a 100 Ω load resistance
tions, starting from 9 am in the morning. The sampling rate is to achieve maximum power transfer, as the internal resist-
set to 10 seconds to be small enough to capture the changes ance of the thermal harvester is around 100 Ω. This loading
of fluctuating RF power and large enough for the thermal condition is used for all the following measurements in the
harvester, as the temperature changes gradually. The output three different environments.
BAKYTBEKOV et al.    
| 3891

3.1 | Outdoor environment 3.2 | Indoor environment

Figure 7A shows the measurement setup used to test the multi- Figure 8A shows the measurement setup used in the indoor
source energy harvester in the outdoor environment. The am- environment. As can be seen in Figure 8B, the collected out-
bient available RF and thermal energy have some patterns, as put voltage from the RF source is no less than 100 mV at any
observed during the pre-design assessment of available power time, as in the previous case, and in total it generates 96 uWh
levels. RF energy patterns depend on users' activity on their of energy in 24 hours. As displayed in Figure 8C, the output
mobile phones and 3G internet surfing. In the evaluation stage voltage of the thermal harvester is around 0 mV and it is al-
of the available ambient RF power, we find that users' activ- most constant throughout the 24 hours. This result is consid-
ity is highest from morning to afternoon and increases again ered valid, as the temperature of the indoor environment does
in the evenings. Early morning hours have the lowest avail- not change much, remaining constant at 22-24°C. Once the
able RF power. In Figure 7B, the same pattern can be observed temperature of the eicosane reaches equilibrium with the am-
in terms of output voltage from the RF harvester. Minimum bient temperature, the thermal harvester stops generating any
voltage levels are at 100 mV during the whole measurement output voltage. This test is conducted in Saudi Arabia, where
period, and there are considerable peaks of up to 300 mV from almost all indoor environments (offices, laboratories, etc)
the morning to the afternoon. This happens when active users are controlled environments through AC. However, there are
are close to the multi-source energy harvester or passing by it. many places with natural ventilation systems that will have
Overall, over 24 hours, 105 uWh of energy is collected from indoor environments with similar temperature trends to the
the RF energy harvester. Thermal energy depends on the tem- outdoor environment. Thus, in these environments the ther-
perature fluctuations during different parts of the day. The pat- mal harvester will operate as in Section 3.1 and will collect
tern is quite consistent (with some exceptions) in Saudi Arabia energy.
for a particular time of the year, as can be seen from the histori-
cal data.22 The thermal energy harvester shows an output volt-
age pattern consistent with this pattern. Figure 7C shows that 3.3 | Underground environment
the thermal energy harvester generates the maximum output
voltage of 180 mV at around 12 pm because this is the warm- In this experiment, the multi-source energy harvester is bur-
est time of day. Output voltage decreases gradually with the ied underground (Figure 9A). In this case, the RF energy
temperature drop in the afternoon, and at around 6 pm it starts harvester does not generate any energy (Figure 9B). This
showing negative output voltage (the ambient temperature is is because the soil is lossy and the ambient power level is
lower than the melting point of the eicosane). Overall, the ther- low, so not enough RF power reaches the RF harvester when
mal harvester generates 1200 uWh of energy in 24 hours. it is buried underground. On the other hand, the thermal

(A)

Data
Laptop Loggers

Multi-source
energy
harvester
(B) (C)

F I G U R E 8 (A) Multisource energy

harvester in the indoor environment. (B)
Output voltage from the RF harvester. (C)
Output voltage from the thermal harvester
3892
|    BAKYTBEKOV et al.

harvester operates properly, as can be seen in Figure 9C. The the multi-source harvester, we must assess whether it can
trend of the output voltage is really similar to the one in the power up a wireless sensor node for continuous operation.
Section 3.1 (outdoor environment); however, the scale of For this experiment, we use a CYALKIT-E02 sensor bea-
the output voltage is different. This can be explained by the con.23 This board includes an energy harvesting power man-
fact that the soil is better thermal conductor compared to air agement IC (PMIC) S6AE103A, Bluetooth BLE transceiver
and thus the energy collected in 24 hours is expected to be CYBLE 022, and sensors (temperature and humidity), as il-
higher in this case as compared to the outdoor environment. lustrated in Figure 10A. On the receiver side, a smartphone
At around 12 pm, the thermal harvester shows highest output with the installed application is used to visualize the readings
voltage of 250 mV. Overall, it produces 2300 uWh energy in of the temperature and humidity sensors in a graph.
24 hours. The challenge here is that the output voltages of the thermal
and RF energy harvesters are quite small (a few hundred mV),
whereas the sensor node requires 2 V to operate. Therefore,
3.4 | IoT application of sensors powered by a step-up converter is used to provide the appropriate voltage
ambient multi-source energy harvester for the node. TI BQ2550424 is selected due to its low input
voltage sensitivity of operation (~100 mV). An output volt-
Based on the measured data from the previous sections, the age of 3 V can be achieved using this step-up converter, with
harvested energy is sufficient to power a node in a typical an efficiency of 40%. As stated in the datasheet of the BLE
wireless sensor for the IoT network. To verify the utility of transceiver, the energy needed for one transmission is 200 uJ,

(A)
Laptop
Multi-source
energy
harvester

*Harvester is fully
buried during
experiment. It is
shown for illustration
purpose only

(B) (C)

F I G U R E 9 (A) Multisource energy harvester buried underground. (B) Output voltage from the RF harvester. (C) Output voltage from the
thermal harvester
BAKYTBEKOV et al.    
| 3893

(A) PMOS
Step-up (Switch)
Mul-source
converter BLE Transceiver
energy harvester 1200uF
(BQ25504) + MCU
10M Power (CYBLE022)
BLE Receiver
managerment
(Cell Phone)
PMIC
(S6AE103A)
BAT_OK Temperature +
(Acve high when NMOS Humidity sensors
2.8V> Vout>2.4V) (Inverter)

(B)

(C) (D)

(E)

F I G U R E 1 0 (A) Block diagram of the measurement setup. (B) Photo of the measurement setup. (C) Readings from the temperature sensor.
(D) Readings from the humidity sensor (E) Data transmission time interval
3894
|    BAKYTBEKOV et al.

FIGURE 11 (A) Readings from the temperature sensor. (B) Readings from the humidity sensor. (C) Data transmission time interval

but the peak current needed can be as high as 10 mA, since and 2.4 V. Below 2.4 V, the switch automatically breaks the
the transmission time is short (a few milliseconds). To pro- connection.
vide such a high current, the harvested energy must be stored Based on the collection from both RF and thermal har-
in a capacitor in advance. A 1200 uF capacitor is placed at the vesters, total 1300 uWh in 24 hours, (Section 3.1), we can
output of the step-up converter to accumulate the energy. The calculate the energy generated each second which equals
voltage of this capacitor is charged from 0 V, and the charging 54.2 uJ. Energy consumption of the BLE for one transmis-
time depends on the harvested power, which ultimately dic- sion, as described previously, is 200 uJ. By dividing this
tates the time interval between data transmissions. value to the energy generated per second from both harvest-
Another challenge appears when the voltage on the ca- ers, we can achieve data transmission from the sensors after
pacitor reaches 2 V, which is the required voltage of the sen- each 3.7 seconds throughout the whole day continuously.
sor node operation. The sensor node will attempt to draw the Two assumptions have been made during this calculation
required current (10 mA), but it cannot do this because the which are, (a) power generated from the energy harvesters are
capacitor at this instance is not fully charged. When the sen- evenly distributed throughout the whole day and (b) outputs
sor node draws the current, the voltage on the capacitor will of the harvesters are combined without any losses (step-up
drop below 2 V and the power management IC (PMIC) of the converter losses and etc). Clearly, the distribution of the col-
sensor node will shut down the operation. This loop contin- lected energy throughout the day varies depending on the
ues, and thus enough energy cannot be collected for one data external factors, so the contribution of the thermal harvester
transmission. To prevent this situation, we need to charge the is greater during this particular test (Section 3.1), as more
storage capacitor to a higher voltage (ie, higher than 2 V) and thermal than RF power is collected. However, this situation
maintain it above 2 V until enough energy is stored to guaran- can be reversed based on another location, where more RF
tee one data transmission. This is done by using a switch be- power can be collected as compared to thermal.
tween the sensor node and the step-up converter. In brief, the As demonstrated in Sections 3.2 and 3.3, there can be sit-
switch and the step-up converter are installed in a way that uations in which one of the sources is not available, so the
the connection between the sensor node and the harvester is IoT sensor node has to operate while powered by a single
maintained as long as the collected voltage is between 2.8 V source. Therefore, in the following we consider two scenarios
BAKYTBEKOV et al.    
| 3895

in which only the single thermal harvester or the single RF the outdoor environment fluctuate considerably and change
harvester operates. gradually over time, depending on the time of day.
Figure 10B displays the setup of the system for the first
test, where only the thermal harvester operates. The results of
the sensor readings can be seen in Figure 10C,D. As shown 4 | CONCLUSION
in Figure 7C, the power collected from the thermal harvester
depends on the temperature fluctuations, so different time This paper presented a multi-source ambient energy har-
intervals achieved between data transmissions, ranging from vester. The advantage of this harvester is that it can work
8 seconds to 90 seconds, throughout the 6 hours of measure- 24 hours a day in environments where solar energy is not
ments from 9 am to 3 pm (Figure 10E). The trend is also very available. It collects energy from thermal and RF energy
clear that the time interval is small (less than 30 seconds) be- sources simultaneously. In an experiment, harvesters were
tween 9 am and 12 pm which can be correlated with Figure 7C smartly integrated into one module, and additive manu-
where the thermal harvester generates maximum power at the facturing techniques were utilized during fabrication. The
warmest period of day in outdoor environment. In the after- multi-source energy harvester was tested in three different
noon, output voltage of the thermal harvester drops gradually conditions: indoors, outdoors, and buried underground. The
which corresponds to the increase in time interval between paper also demonstrated a real-world IoT application with
data transmissions (60 seconds to 90 seconds). In other words, temperature and humidity sensors transmitting the data to a
time interval indicates how fast the thermal harvester can phone via Bluetooth.
charge the capacitor that powers up the temperature and hu-
midity sensors. Overall, average of 29 seconds time interval ACKNOWLEDGMENTS
between data transmissions has been achieved when the IoT The research reported in this publication was supported by
sensors are powered by the thermal harvester only. This value the King Abdullah University of Science and Technology
is considered as a reasonable time interval as the change in (KAUST). The authors acknowledge the KAUST Sensor
temperature and humidity in such a small time period is neg- Initiative, under award OSR-2015-Sensors-2700, for the
ligible. Video demonstration of this experiment can be found financial support. Special thanks go to Esraa Fakeih, who
in multimedia section of Supporting Information. helped in collecting power measurement data on the KAUST
From Figures 7B and 8B, it can be seen that the output campus.
voltages of the RF energy harvesters vary from 100 mV to
peaks of 300 mV. For this experiment, the indoor environ- ORCID
ment is chosen to test the IoT wireless sensor node when only Azamat Bakytbekov https://orcid.
RF power is available. The results of the temperature and hu- org/0000-0002-7247-4876
midity readings from the indoor environment (in this case,
an office) are shown in Figure 11A,B. An average time in- R E F E R E NC E S
terval of 8 minutes 55 seconds is achieved for the duration of 1. Statista Research Department. Internet of Things - Number of
6 hours (Figure 11C). Depending on the availability of ambi- Connected Devices Worldwide 2015–2025; 2019. https://www.stati​
ent RF power levels, the time interval ranges from 8 minutes sta.com/stati​stics/​47126​4/iot-numbe​r-of-conne​cted-devic​es-world​
wide/. Accessed 27 October 2019
15 seconds to 11 minutes 30 seconds. As the collected power
2. Ahmed S, Kakkar V. An electret-based angular electrostatic en-
from the RF harvester is lower than that from the thermal har-
ergy harvester for battery-less cardiac and neural implants. IEEE
vester (shown in the previous sub-section), the time interval Access. 2017;5:19631-19643.
is longer. Even if the output voltages of the thermal and the 3. Morais R, Matos SG, Fernandes MA, et al. Sun, wind and water
RF energy harvesters are converted to up to 3 V of required flow as energy supply for small stationary data acquisition plat-
voltage, currents are very different, and the charging of the forms. Comput Electron Agric. 2008;64(2):120-132.
capacitor occurs more slowly in the RF energy harvester 4. Gu X, Hemour S, Guo L, Wu K. Integrated cooperative ambient
case. However, the conclusion based on this experiment is power harvester collecting ubiquitous radio frequency and kinetic
energy. IEEE Trans Microw Theory Tech. 2018;66(9):4178-4190.
that when the thermal harvester is off and the weakest energy
5. Sai Srujana B, Neha PM, Harigovindan VP. Multi-source energy
source is operating, temperature and humidity readings can
harvesting system for underwater wireless sensor networks. Proc
be achieved every 9 minutes, which still seems to be a decent Comput Sci. 2015;46:1041-1048.
rate with which to monitor the temperature and humidity lev- 6. Carli D, Brunelli D, Benini L, Ruggeri M. An effective multi-
els in an indoor environment. As displayed in Figure 11A,B, source energy harvester for low power applications. In: 2011
the temperature and humidity values are constant over time Design, Automation & Test in Europe, Grenobl; 2011:1-6.
with minimum fluctuations. This is expected for the indoor 7. Hu Z, Qiu J, Wang X, et al. An integrated multi-source energy
environment, as it is controlled by an AC system. In contrast, harvester based on vibration and magnetic field energy. AIP Adv.
2018;8(5):56623.
Figure 10C,D shows that the temperature and humidity of
3896
|    BAKYTBEKOV et al.

8. Bakytbekov A, Nguyen TQ, Huynh C, Salama KN, Shamim A. AUTHOR BIOGRAPHIES

Fully printed 3D cube-shaped multiband fractal rectenna for ambi-
ent RF energy harvesting. Nano Energy. 2018;53:587-595.
Azamat Bakytbekov received
9. Cottrill AL, Zhang GE, Liu AT, et al. Persistent energy harvesting
in the harsh desert environment using a thermal resonance device:
his bachelor's degree from the
design, testing, and analysis. Appl Energy. 2019;235:1514-1523. Electrical Engineering and
10. Cottrill AL, Liu AT, Kunai Y, et al. Ultra-high thermal effusivity Electronics Department at
materials for resonant ambient thermal energy harvesting. Nature Nazarbayev University in
Commun. 2018;9:664. Kazakhstan in 2015, and his
11. Keysight. A-Series FieldFox Analyzers. http://liter​ature.cdn.keysi​ master's degree from the
ght.com/litwe​b/pdf/N9927​-90001.pdf. Accessed June 2017
Electrical Engineering
12. Bakytbekov A, Maza AR, Nafe M, Shamim A. Fully inkjet printed
Department at King Abdullah
wide band cantor fractal antenna for RF energy harvesting applica-
tion. In: 11th European conference on antennas and propagation University of Science and Technology (KAUST) in 2017.
(EUCAP), Paris; 2017:489-491. He is currently PhD student at KAUST under supervision
13. Bakytbekov A, Shamim A. Additively manufactured triple-band of Dr Atif Shamim. His research interests include multi-
fractal antenna-on-package for ambient RF energy harvesting. band and wideband printed antennas, printed electronics,
In: 13th European Conference on Antennas and Propagation RF energy harvesters and multi-source energy harvesters.
(EuCAP), Krakow, Poland; 2019:1-3.
14. Preperm Webshop. PREPERM 3D ABS300 Filament. https://www.
Thang Q. Nguyen received the
prepe​r m.com/websh​op/produ​ct/prepe​r m-3d-abs-%c9%9br-3-0-
filam​ent/. Accessed June 2018
BE degree in telecommunica-
15. Makerbot. Makerbot Replicator 2X Experimental 3D Printer. tion engineering from Ho Chi
https://pages.maker​bot.com/ap-repli​cator​-2x.html. Accessed Minh University of Technology
March 2017 (HCMUT), Ho Chi Minh City,
16. Dupont Inc. Printed Electronics Materials. https://www.dupont. Vietnam in 2017. He joined the
com/elect​ronic​-mater​ials/print​ed-elect​ronics.html. Accessed June Sensors Lab at KAUST, Saudi
2018 Arabia as a visiting student in
17. Balanis C. Antenna Theory. Hoboken, NJ: Wiley; 2016.
summer 2017 where he con-
18. Kotani K, Sasaki A, Ito T. High-efficiency differential-drive
CMOS rectifier for UHF RFIDs. IEEE J Solid-State Circ.
ducted research on multi-band RF energy harvesting sys-
2009;44(11):3011-3018. tems. He is currently a master student at KAUST. His re-
19. Song C, Huang Y, Carter P, et al. A novel six-band dual CP search focus on RF energy harvesting systems and
rectenna using improved impedance matching technique for am- Wireless sensor systems.
bient RF energy harvesting. IEEE Trans Antennas Propagat.
2016;64(7):3160-3171. Weiwei Li received the BS de-
20. Skyworks. Surface-Mount Mixer and Detector Schottky Diodes.
gree in packaging engineering,
https://www.skywo​rksinc.com/-/media/​SkyWo​rks/Docum​ents/
and MS degree in material
Produ​cts/201-300/Surfa​ce_Mount_Schot​tky_Diodes_20004​1AE.
pdf. Accessed March 2017 physics and chemistry from
21. Pico Technology. Data Logger, Oscilloscope, Signal Generator Beijing Institute of Graphic and
and More. https://www.picot​ech.com/data-logge​r/drdaq/​overview. Communication, Beijing,
Accessed August 2017 China, in 2010 and 2013, re-
22. World Weather Online. Jeddah Historical Weather. https://www. spectively, and the PhD degree
world​weath​eronl​ine.com/jedda​h-weath​er-histo​ry/makka​h/sa.aspx. in nanoscience and technology
Accessed October 2019
from National Center for
23. Cypress Semiconductor. CYALKIT-E02 Solar-Powered BLE
Sensor Beacon Reference Design Kit. https://www.cypre​ss.com/
Nanoscience and Technology,
docum​entat​ion/devel​opmen​t-kitsb​oards/​cyalk​it-e02-solar​-power​ Chinese Academy of Sciences, Beijing, China, in 2017.
ed-ble-senso​r-beaco​n-refer​ence-design. Accessed September Currently, he is working as a post-doctoral fellow at King
2019 Abdullah University of Science and Technology, Thuwal,
24. Texas Instruments. Ultra Low Power Boost Converter with Saudi Arabia. His research interests include synthesis of
Battery Management for Energy Harvester. Nano-Power functional materials, printed electronics, and radio fre-
Management. http://www.ti.com/produ​ ct/BQ25504. Accessed quency devices.
September 2019
BAKYTBEKOV et al.    
| 3897

Anton Lee Cottrill received Khaled N. Salama received

his bechalor degree in his masters and doctorate de-
Chemistry and Chemical grees from the Electrical
Engineering from Vanderbilt Engineering Department at
University in 2013, and his Stanford University. His work
doctorate degree from the on CMOS sensors for molecu-
Department of Chemical lar detection was awarded the
Engineering at Massachusetts Stanford-Berkeley Innovators
Institute of Technology in Challenge Award in biological
2018. During his PhD, Anton conducted a series of re- sciences and was acquired by Lumina Inc in 2008. He was
search on novel thermal materials and devices, including an assistant professor at RPI between 2005 and 2008. He
thermal diodes, thermal resonators, and high-effusivity is a Professor and founding chair member in the Electrical
materials and published 5 papers on these topics. His engineering department at King Abduallah University of
work on harvesting temperature fluctuation from envi- Science and Technology. He is currently the director of
ronment with thermal resonators was published on the sensor initiative at KAUST. He is the co-author of 200
Nature Communications. papers and 14 patents on sensors.

Ge Zhang received his bache- Atif Shamim received MS

lor's degree in chemical engi- (2004) and PhD degrees (2009)
neering in 2017 from Tsinghua in electrical engineering from
University. He is currently a Carleton University, Canada. In
PhD student in Prof. Michael 2010, he joined KAUST and is
Strano's laboratory at currently an Associate Professor.
Massachusetts Institute of His research work has won sev-
Technology. His research inter- eral best paper awards in presti-
ests include novel energy har- gious conferences and has been
vesting systems such as thermal resonator, and building highlighted in international media. He has won numerous
power sources to enable the independent operation of awards, OCRI Researcher of the Year (2008), OCRI
cell-sized micro-electronic systems dispersed in Entrepreneur of the year (2010), 1st prize in CMC
environment. Competition (2007) and 1st prize in Canada's national busi-
ness plan competition (2009). He is an author of 200 inter-
Michael S. Strano is currently national publications, inventor on 20 patents and has given
the Carbon P. Dubbs Professor 40 invited talks at international forums.
in the Chemical Engineering
Department at the SUPPORTING INFORMATION
Massachusetts Institute of Additional supporting information may be found online in
Technology. His research fo- the Supporting Information section.
cuses on biomolecule/nanopar-
ticle interactions and the sur-
face chemistry of low How to cite this article: Bakytbekov A, Nguyen TQ, Li
dimensional systems, nano-electronics, nanoparticle sepa- W, et al. Multi-source ambient energy harvester based
rations, and applications of vibrational spectroscopy to on RF and thermal energy: Design, testing, and IoT
nanotechnology. application. Energy Sci Eng. 2020;8:3883–3897. https://
doi.org/10.1002/ese3.784