In the era of the rapidly developing Internet of Things (IoT), wireless sensor networks have become a key enabler for a wide range of applications, including environmental monitoring, smart agriculture, intelligent transportation, and infrastructure management. However, the vast majority of sensor nodes are still powered by conventional batteries, which have limited capacity and finite lifetime. When the battery is exhausted, the node may fail, and replacing batteries in remote or hostile environments is often impractical and expensive. Therefore, continuous and sustainable power supply has become an urgent issue restricting the further development of IoT systems. Because the average power consumption of typical IoT sensor nodes is less than 10 mW, environmental energy harvesting offers a promising solution to achieve self-sustaining operation.
Among various ambient energy sources, solar energy has the highest power density. However, solar harvesting alone cannot provide stable output due to its dependence on weather conditions and time of day. On the other hand, radio frequency (RF) energy from cellular base stations, Wi-Fi access points, and broadcast towers is ubiquitously present in the environment, albeit at much lower power densities. The combination of solar and RF energy collection—referred to as hybrid energy harvesting—can exploit the complementary characteristics of these two sources: solar energy provides high power when available, while RF energy can be collected continuously regardless of lighting conditions. This thesis focuses on the design of photovoltaic (PV) panel integrated antennas and low-power rectifying circuits for such hybrid systems.
The core contribution of my work is the integration of solar panels into antenna structures so that the aperture of the solar panel can face the sun while the antenna radiation pattern points toward RF sources. This is a nontrivial requirement because conventional antennas with broadside radiation have their main beam orthogonal to the panel plane. If the panel is oriented toward the sky, the beam would point upward, away from typical terrestrial RF sources. Therefore, I have designed several antennas with end-fire or tilted beams that allow simultaneous solar energy collection and RF energy harvesting. Additionally, I designed a low-power broadband rectifier circuit that operates in the same frequency band as the antennas, enabling a complete energy harvesting system.
Antenna Fundamentals Used in the Design
In order to design antennas for hybrid energy harvesting, it is essential to understand the basic performance parameters. The input impedance of an antenna is defined as
$$Z_{in} = \frac{U_{in}}{I_{in}} = R_{in} + jX_{in}$$
where \(R_{in}\) is the resistance consisting of radiation resistance and loss resistance, and \(X_{in}\) is the reactance. The reflection coefficient between the transmission line and the antenna is
$$\Gamma = \frac{Z_{in} – Z_0}{Z_{in} + Z_0}$$
and the voltage standing wave ratio is
$$VSWR = \frac{1 + |\Gamma|}{1 – |\Gamma|}$$
The impedance bandwidth is generally defined as the frequency range over which \(VSWR < 2\) or \(|S_{11}| < -10\,\text{dB}\). The relative bandwidth is expressed as
$$BW = \frac{f_H – f_L}{f_0} \times 100\%$$
where \(f_H\) and \(f_L\) are the upper and lower frequencies of the operating band, and \(f_0\) is the center frequency.
The radiation pattern and directivity are also important. The directivity \(D\) of an antenna is defined as
$$D = \frac{U_{max}}{U_0}$$
where \(U_{max}\) is the maximum radiation intensity and \(U_0\) is the radiation intensity of an isotropic radiator with the same total radiated power. The gain is related to directivity by
$$G = \eta D$$
where \(\eta\) is the antenna efficiency. Polarization is another critical property; for linear polarization the electric field vector traces a straight line, while for circular polarization the axial ratio should be less than 3 dB.
Among the feeding methods for microstrip patch antennas, direct feeding methods such as microstrip line feeding and coaxial probe feeding are simple but suffer from narrow bandwidth. In contrast, indirect methods such as aperture coupling and proximity coupling can significantly broaden the impedance bandwidth. In my designs, I utilize aperture coupling and coplanar waveguide feeding to achieve wideband operation and to separate the DC path of the solar panels from the RF path of the antenna.
Photovoltaic Panel Integrated Reconfigurable Antenna
The first antenna I designed is a photovoltaic panel integrated mechanically reconfigurable antenna. The motivation is to realize a broadside radiation pattern that can be tilted so that the main beam points toward terrestrial RF sources while the solar panel aperture remains upward. The antenna structure is shown in the following schematic representation: it consists of a 1×4 photovoltaic panel periodic superstrate placed on the top layer, a ground plane with a rectangular slot, and a microstrip feed line on the bottom layer. The photovoltaic panels are used as the periodic superstrate because their back electrode is made of aluminum, which is conductive and can serve as part of the radiating structure. The use of the PV panel superstrate not only allows the panel to be fully exposed to sunlight but also enhances the impedance bandwidth and gain of the antenna.
The antenna operates around 2.5 GHz. The overall size is 233 mm × 200 mm × 3 mm. The photovoltaic panel superstrate consists of four solar cells, each with dimensions \(W_1 \times W_2\), and the gap between adjacent cells is \(W_3\). The ground plane contains a rectangular slot of width \(W_4\) and length \(L_4\). The microstrip feed line has dimensions \(W_5 \times L_5\) and is oriented perpendicular to the slot. By mechanically shifting the top substrate relative to the lower substrate, the main beam can be steered between \(-30^\circ\) and \(+30^\circ\). The aperture coupling feeding method ensures that the RF energy harvesting circuit and the solar energy harvesting circuit are electrically isolated, eliminating the need for additional decoupling structures.
I performed a current distribution analysis to understand the operating principle. In the first mode, where the slot is located beneath the third and fourth solar cells, the surface current is mainly concentrated on these two cells, while the current on the first and second cells is weaker and opposite in phase. This results in a radiation pattern tilted to \(+30^\circ\). In the second mode, the slot is located beneath the first and second cells, producing a beam tilt to \(-30^\circ\). The tapered amplitude and phase distribution across the periodic superstrate are the key mechanisms for beam tilting.
I also carried out parametric studies to optimize the antenna. The following table summarizes the optimized parameters:
| Parameter | Value (mm) | Description |
|---|---|---|
| \(L_1\) | 200 | Length of substrate 1 |
| \(L_2\) | 170 | Width of substrate 1 |
| \(L_3\) | 233 | Length of substrate 2 |
| \(W_1\) | 140 | Length of each PV cell |
| \(W_2\) | 31 | Width of each PV cell |
| \(W_3\) | 0.5 | Gap between PV cells |
| \(L_4\) | 60 | Length of slot |
| \(W_4\) | 3.0 | Width of slot |
| \(L_5\) | 135.5 | Length of microstrip feed |
| \(W_5\) | 3.1 | Width of microstrip feed |
| \(H_1\) | 1 | Thickness of substrate 1 |
| \(H_2\) | 2 | Thickness of substrate 2 |
The simulated \(|S_{11}| < -10\,\text{dB}\) frequency ranges for the two modes are 2.26–2.92 GHz and 2.26–2.91 GHz, while the measured ranges are 2.28–3.05 GHz and 2.30–3.02 GHz. The measured results are slightly shifted to higher frequencies, which is attributed to tolerances in the dielectric constant and fabrication. The realized gain in the operating band is 5.2–11 dBi for mode 1 and 5–10.8 dBi for mode 2. At 2.5 GHz, the measured maximum gains are 8.7 dBi and 10.6 dBi for the two modes, respectively.
The radiation patterns measured at 2.3 GHz, 2.5 GHz, and 2.7 GHz agree well with simulations. The main beam directions are \(-30^\circ\) for mode 1 and \(+30^\circ\) for mode 2. This reconfigurable behavior allows the antenna to adapt to different locations of RF sources while the solar panel always faces upward. In outdoor testing under sunlight, one PV cell produced a voltage of 614.3 mV and a current of 227.7 mA, confirming that the solar panel works properly despite being integrated into the antenna structure.
Photovoltaic Panel Integrated Dual-Beam Antenna
To further increase the angular coverage of RF energy harvesting, I designed a second antenna that produces two simultaneous beams pointing at \(+30^\circ\) and \(-30^\circ\). This dual-beam characteristic is achieved by extending the periodic superstrate to 1×6 PV cells. The operating principle is based on the superposition of the two configurations of the reconfigurable antenna described above. When the slot is centrally located beneath the array, the current distribution on the left three cells and the right three cells has opposite phases, creating two tilted beams.
The antenna geometry is shown in the following description. The overall dimensions are 228 mm × 190 mm × 2 mm, with a center frequency of 2.5 GHz. The structure consists of, from top to bottom, a 1×6 PV panel superstrate, a dielectric substrate with relative permittivity 2.65 and thickness \(H\), a coplanar waveguide (CPW) feed structure on the bottom side, and a metallic reflector plate placed at a distance \(H_1\) below the dielectric substrate. The CPW feed includes two inverted-L shaped slots etched in the ground plane. The horizontal slot dimensions are \(W_3 \times L_3\), and the vertical slot dimensions are \(W_4 \times L_4\). The gap between the two slots is \(W_5\). The PV cells are serially connected using metal traces and inductors to provide DC extraction while preventing RF leakage.
The presence of the metallic reflector significantly improves the front-to-back ratio and increases the forward gain. In the simulation, the back lobe level is reduced from 4.5 dBi to about \(-8\) dBi after adding the reflector, and the peak forward gain is increased by approximately 3 dB.
I performed parametric analyses for the dual-beam antenna. The key parameters are summarized in the following table:
| Parameter | Value (mm) | Description |
|---|---|---|
| \(L_1\) | 228 | Length of substrate |
| \(W_1\) | 190 | Width of substrate |
| \(L_2\) | 24 | Length of each PV cell |
| \(W_2\) | 144 | Width of each PV cell |
| \(G_1\) | 3.5 | Gap between cells |
| \(L_3\) | 4.8 | Horizontal slot length |
| \(W_3\) | 44 | Horizontal slot width |
| \(L_4\) | 116.4 | Vertical slot length |
| \(W_4\) | 0.3 | Vertical slot width |
| \(W_5\) | 3.8 | Distance between slots |
| \(H\) | 2 | Substrate thickness |
| \(H_1\) | 20 | Reflector distance |
The effect of the PV superstrate length \(W_2\) on \(|S_{11}|\) was investigated. For \(W_2 = 140\) mm, the high-frequency matching is poor; for \(W_2 = 148\) mm, the matching is good but the bandwidth is narrower. The optimum is 144 mm. The slot width \(W_3\) strongly affects the resonant frequency and matching: 44 mm yields the best overall performance. The substrate width \(W_1\) mainly affects low-frequency matching; 190 mm is the best trade-off. The reflector distance \(H_1\) was varied from 18 to 22 mm, and 20 mm provides good matching at both band edges.
The simulated and measured reflection coefficients are in good agreement. The simulated bandwidth (\(|S_{11}| < -10\,\text{dB}\)) is 2.13–2.74 GHz, while the measured bandwidth is 2.11–2.81 GHz. The measured gain in the operating band is 6.6–9.7 dBi, with a peak gain of 9.5 dBi at 2.5 GHz. The normalized radiation patterns at 2.4 GHz, 2.5 GHz, and 2.6 GHz show two equal beams at \(+30^\circ\) and \(-30^\circ\). This dual-beam antenna increases the angular coverage of the RF energy harvesting system, allowing it to capture power from multiple sources in different directions.
For the solar harvesting part, the six PV cells connected in series produced an open-circuit voltage of 3.833 V and a short-circuit current of 468.6 mA under strong sunlight, proving the effectiveness of the solar panel integration.
Photovoltaic Panel Integrated Yagi Antenna
For scenarios where the RF source is located at a long distance and the antenna must have high directivity along the horizontal direction, I designed a photovoltaic panel integrated Yagi antenna. The Yagi antenna is known for its end-fire radiation, which naturally satisfies the requirement that the antenna aperture (the plane of the solar panel) is vertical rather than horizontal. In this design, the solar panels are used as parasitic director elements of the Yagi antenna, while the driven element and reflector are conventional metallic structures.
The antenna dimensions are 142 mm × 66 mm × 1.5 mm, with a center frequency of 2.5 GHz. The top surface of the substrate contains a coplanar strip ground plane and the PV parasitic structures. The driven element has dimensions \(W_1 \times L_1\), and the reflector has dimensions \(R_w \times Y_1\). The directors are arranged in a row, with the first director spaced \(d_0\) from the driven element and subsequent directors spaced \(d_1\) apart. To increase the total area of the solar panels, I added four additional parasitic patches on the upper and lower sides of the director array. These extra patches have the same dimensions \(W_2 \times L_2\) as the directors and are positioned symmetrically. The bottom layer of the substrate contains a microstrip feed line with two sections of lengths \(L_3\) and \(L_4\) and width \(W_3\). A shorting via with diameter \(W_4\) connects the end of the microstrip line to the ground plane, forming a half-wavelength driven element. The SMA connector feeds the antenna from the side: the inner conductor is soldered to the microstrip line, and the outer conductor is connected to the coplanar ground.
Loading the four parasitic patches increases the total solar panel area, which is beneficial for solar energy harvesting. The simulated \(|S_{11}|\) comparison shows that adding these parasitic elements shifts the operating band slightly lower but does not significantly change the relative bandwidth of about 34%. The gain variation is within 1 dB.
The optimized design parameters are listed in the following table:
| Parameter | Value (mm) | Description |
|---|---|---|
| \(L_1\) | 24.5 | Length of driven element |
| \(W_1\) | 8 | Width of driven element |
| \(L_2\) | 28 | Length of directors / parasitic patches |
| \(W_2\) | 13 | Width of directors / parasitic patches |
| \(R_w\) | 18 | Width of reflector |
| \(d_0\) | 15 | Spacing from driven element to first director |
| \(d_1\) | 8 | Spacing between directors |
| \(d_2\) | 7 | Distance from parasitic patch to substrate edge |
| \(d_3\) | 18 | Vertical spacing between parasitic patches |
| \(d_r\) | 25 | Spacing from reflector to driven element |
| \(X_1\) | 142 | Substrate length |
| \(Y_1\) | 66 | Substrate width |
| \(L_3\) | 39 | Length of first microstrip section |
| \(L_4\) | 5 | Length of second microstrip section |
| \(W_3\) | 2 | Width of microstrip feed |
| \(W_4\) | 1 | Diameter of shorting via |
| \(L_5\) | 9 | Distance from microstrip line to coplanar strip |
| \(H\) | 1.5 | Substrate thickness |
The driven element length \(L_1\) has a strong influence on the operating frequency. Increasing \(L_1\) shifts the band downward and degrades the matching. The optimum \(L_1\) is 24.5 mm. The director spacing \(d_1\) affects the coupling between directors: 8 mm gives the widest bandwidth. The reflector width \(R_w\) affects the low-frequency matching; 18 mm yields the best performance. The spacing \(d_0\) determines the separation between the two resonance frequencies; 15 mm creates a good match. The parasitic patch spacing \(d_3\) and edge distance \(d_2\) have minor effects, and the chosen values (18 mm and 7 mm, respectively) provide good high-frequency matching.
The measured reflection coefficient shows \(|S_{11}| < -10\,\text{dB}\) over 2.03–3.28 GHz, while the simulated bandwidth is 2.16–3.02 GHz. The measured gain within the band ranges from 6.9 to 9 dBi, with a peak of 9.5 dBi. The end-fire radiation pattern is well maintained at 2.2 GHz, 2.5 GHz, and 2.8 GHz. The solar panel array, consisting of the directors and four parasitic patches, produced a voltage of 4.73 V and a current of 111.9 mA under sunlight. This confirms that the integrated PV Yagi antenna can simultaneously harvest solar energy and RF energy.
Low-Power Broadband Rectifier Circuit
The RF energy collected by the antennas must be converted into DC power. I designed a low-power broadband rectifier circuit operating at a center frequency of 2.5 GHz. The rectifier topology is based on a full-wave voltage doubler, which is suitable for low input power levels because it utilizes both half-cycles of the RF signal and provides a higher DC output voltage compared to a half-wave rectifier.
First, I selected the appropriate Schottky diode. A comparison among SMS-7630, HSMS-285C, and HSMS-286C was carried out using a simple half-wave rectifier circuit. The following table summarizes the simulated rectification efficiency at 2.5 GHz for different input power levels:
| Input Power (dBm) | HSMS-285C | SMS-7630 | HSMS-286C |
|---|---|---|---|
| -20 | 12% | 9% | 4% |
| -10 | 38% | 31% | 18% |
| 0 | 55% | 50% | 42% |
| 10 | 35% | 43% | 55% |
Since the target is low ambient RF power, HSMS-285C is the best choice because it provides the highest efficiency below 0 dBm input power.
The rectifier circuit consists of a five-stage low-pass filter as the input matching and harmonic suppression network, followed by the voltage doubler rectifier, and an output DC filter. The initial design without explicit matching yielded a narrow bandwidth and low efficiency. Therefore, I designed a single-stub matching network consisting of a shunt open-circuit stub and a series transmission line. The matching network transforms the impedance of the rectifier to 50 Ω over a wide frequency range.
The load impedance was optimized as well. The efficiency versus load resistance curve shows a maximum at \(R_L = 3.2\,k\Omega\). Therefore, the rectifier was designed for a 3.2 kΩ load.
The simulated efficiency versus input power at 2.5 GHz is shown in the following table:
| Input Power (dBm) | Rectification Efficiency |
|---|---|
| -15 | 27% |
| -10 | 42% |
| -5 | 55% |
| 0 | 67% |
| 5 | 61% |
| 10 | 50% |
The efficiency increases with input power up to 0 dBm and then decreases because the diode impedance changes at high power levels, causing impedance mismatch. The frequency response at different input power levels shows that for -10 dBm input, the efficiency exceeds 20% from 2.3 GHz to 2.7 GHz, giving a broadband performance of 400 MHz. At 0 dBm input, the efficiency exceeds 40% over 2.35–2.80 GHz.
I fabricated the rectifier circuit and measured its performance using a signal generator and a digital multimeter. The measured maximum efficiencies at 0 dBm, -5 dBm, -10 dBm, and -15 dBm are 63%, 53%, 40%, and 25%, respectively. The measured efficiency curves are slightly lower than simulations due to diode parasitics and fabrication tolerances, but the broadband characteristic is preserved.
System-Level Testing
To validate the complete hybrid energy harvesting system, I connected each of the three designed antennas to the rectifier and performed end-to-end RF-to-DC conversion tests. A standard gain horn antenna was used as the transmitter, and the photovoltaic panel integrated antennas were used as receivers. The output DC voltage across the 3.2 kΩ load was measured at different distances and input power levels.
The following table summarizes the measured DC output voltages:
| RF Input Power | Distance | Reconfigurable Antenna | Dual-Beam Antenna | Yagi Antenna |
|---|---|---|---|---|
| 20 dBm | 0.5 m | 760 mV | 550 mV | 560 mV |
| 20 dBm | 1 m | 660 mV | 310 mV | 380 mV |
| 15 dBm | 0.5 m | 406 mV | 280 mV | 320 mV |
| 15 dBm | 1 m | 350 mV | 145 mV | 206 mV |
| 10 dBm | 0.5 m | 205 mV | 122 mV | 144 mV |
| 10 dBm | 1 m | 158 mV | 57 mV | 84 mV |
As expected, the output voltage decreases with increasing distance due to the higher path loss and with decreasing input power. The reconfigurable antenna with higher maximum gain provides larger output voltage than the other two antennas. These measurements confirm that the proposed photovoltaic panel integrated antennas together with the low-power broadband rectifier can convert ambient RF energy into usable DC power. At the same time, the solar panels integrated into the antennas produce voltages and currents sufficient to power low-duty-cycle IoT sensors.
Conclusion and Future Work
In this thesis, I have investigated the design of photovoltaic panel integrated antennas and rectifier circuits for hybrid energy harvesting systems. The main challenge is that the solar panel aperture must face the sun, while the antenna beam must point toward terrestrial RF sources. I addressed this challenge by designing antennas with tilted or end-fire radiation patterns: a mechanically reconfigurable antenna with beam steering, a dual-beam antenna with simultaneous \(+30^\circ\) and \(-30^\circ\) beams, and a Yagi antenna with end-fire radiation. All three antennas integrate solar panels as part of the radiating structure without blocking sunlight. The aperture coupling or coplanar waveguide feeding methods ensure that the solar DC extraction circuit and the RF path are isolated.
The measured results show that the reconfigurable antenna has an impedance bandwidth of more than 25% and a maximum gain of 10.5 dBi; the dual-beam antenna achieves 25% bandwidth with two beams of gain above 9 dBi; the Yagi antenna exhibits 33% bandwidth and a peak gain of 9.5 dBi. The rectifier circuit achieves 40% efficiency at -10 dBm input power and maintains efficiency above 20% over a 400 MHz band. System tests confirm that the integrated antennas and rectifier can harvest RF energy from a nearby transmitter, while the solar panels continue to operate normally.
Future work will focus on extending the photovoltaic panel integrated antenna concept to phased arrays and on designing lower-profile dual-beam antennas. The current dual-beam antenna has a relatively high profile due to the reflector plate; reducing the height while maintaining gain is a potential improvement. Additionally, I plan to investigate the integration of transparent conductive materials and mesh structures to increase the optical transparency of the antennas and further improve solar collection efficiency. The ultimate goal is to develop a self-sustaining power supply module for IoT sensor nodes that can operate reliably in diverse environmental conditions.