Multiband Dual-Meander Line Antenna for Body-Centric Networks’ Biomedical Applications by Using UMC 180 nm

Abstract

A new, compact, on-chip antenna architecture for 5G body-centric networks’ (BCNs) applications is presented in this paper. The integrated antenna combines two turns of dual-meander lines (DML) on two stacked layers and a metal ground layer. The proposed DML antenna structure operated at resonant bands 22 GHz, 34 GHz, 44 GHz, and 58 GHz with an operating bandwidth up to 2 GHz at impedance bandwidth ≤−7.5 dB (VSWR—Voltage Standing Wave Ratio ≤ 2.5) and antenna gain about −20 dBi, −15 dBi, −10 dBi, and −1 dBi, respectively. Then it was compared with conventional single-meander line antenna. The proposed structure decreased the resonant frequency by 22%, increased number of tuning bands, and broadened the operating bandwidth by 25%, 15%, 10%, and 20% for the tuning bands to be a suitable choice for high-data -ate biomedical applications. Furthermore, the proposed antenna was simulated and studied for its performance on and inside the human body to test the integration effect in wearable equipment. The results showed that the antenna had acceptable performance in both locations. All simulations of the proposed antenna were done were done by using Ansys HFSS (high-frequency structure simulator) v.15 (Ansys, Canonsburg, PA, USA). The DML (Digital Microwave Links) antenna was fabricated by using UMC (United Microelectronics Corporation) 180 nm CMOS (Complementary Metal–Oxidesemi–Conductor) technology with a total area of 1150 µm × 200 µm and the results showed a good agreement between measured and simulated results.

1. Introduction

The 5G technologies have brought about massive growth of millimeter-wave (mm-wave) body-centric networks (BCNs) communications, shown in Figure 1 [1]. BCNs are a subject area combining Wireless Body-Area Networks (WBANs), Wireless Sensor Networks (WSNs), and Wireless Personal Area Networks (WPANs), and refers to human-self and human-to-human networking with the use of wearable and implantable wireless sensors [2,3,4,5,6,7,8,9]. It has numerous applications in healthcare, smart homes, personal entertainment, identification systems, space exploration, and military. Healthcare applications require the antenna and associated electronics to support multi-channels to be able to deal with a group of patients with each channel. The channel should be supported by large bandwidth to communicate with many sensors at the same time. Multiband antenna is a potential solution for wearable/implantable sensors as it can support both data transfer and integrated wireless powering techniques, as energy harvesting/wireless power transfer module, to reduce the need for regular battery replacement [10,11]. Attenuation on millimeter-wave link is observed to be very weak and would not result in communication failure at 28–38 GHz [12] and 60 GHz [13]. However, for wireless power transfer applications, there are proper choices of frequency to ensure high RF-DC power conversion. Some studies have been done at 24 GHz [14,15], 28 GHz [16], 35 GHz [17], and 60 GHz [18]. Recently, the multiband, mm-wave, on-chip antennas have been also considered and discussed in literature [19,20,21].

In this paper, a multiband on-chip antenna for mm-wave BCNs is proposed with three frequency bands for data transfer and a fourth band for wireless powering module. Taking into consideration that, as the frequency increases, the DC to RF/RF to DC conversion efficiency and the power decrease. Thus, the resonant band at 22 GHz was selected for power transfer and the two resonant bands at 34 GHz and at 58 GHz with low attenuation were selected for data transfer. The selected resonant frequency at 44 GHz band is usually used in some cancer detection applications [22,23]. The previous results, in Gabriel et al. and IT IS website [24,25], showed the electric properties of human tissues are varying with frequency and could affect the performance of the antenna. In order to accurately model and design a wearable/implantable antenna, the dispersion characteristics of the different tissues are taken into consideration.

The proposed multiband, on-chip antenna is designed and fabricated by UMC 180-nm CMOS technology. The proposed antenna structure, shown in Figure 2, consists of three layers. Each layer is composed of a separate metal layer, an upper meander line (UML), a lower mender line (LML), and a metal ground plane. Compared to normal meander line antenna, the proposed structure doubles tuning frequencies and increases the bandwidth for each tuning frequency. The operation of the proposed antenna resonates at four bands: 22 GHz, 34 GHz, 44 GHz, and 58 GHz. The enhanced bandwidth for each tuning frequency makes it a good candidate for Ultra WideBand (UWB) transceivers that need many resonant frequencies for data and power transmission.

The organization of the paper is set as follows: Section 2 presents a detailed explanation for the dual-meander line (DML) antenna design. The simulation results for the proposed antenna and comparison with conventional single-layer meander line antennas are presented in Section 3. In Section 4, the effect of the human body for both implanted and wearable applications on the antenna performance are demonstrated. Measurements of the proposed antenna and other antenna parameters are shown in Section 5. Finally, a conclusion of the paper is shown in Section 6.

2. Antenna Configuration and Design

In order to provide good wireless communication from outside the body to inside, factors such as high tissue conductivity, biocompatibility, and small antenna size must be taken into consideration. Furthermore, simulation models and physical models are also important to predict the behavior of the antenna in the presence of the human body. The most commonly used models are one-layer models and three-layer models, presented and compared in [23].

The results showed that no significant differences were observed in the measured input parameters unless the antenna is to be implanted in the fat layer [23]. The dual-meander line (DML) antenna is fabricated using UMC 180-nm CMOS process. The technology layers consist of a low resistivity silicon substrate, six metal layers embedded in inter-dielectric layers, with the upper metal layer, M₆, being covered with a dielectric passivation layer. As shown in Figure 3, the proposed DML configuration was implemented at two stacked layers, the upper two turns of meander line (UML) at metal layer M₆ and two turns of the lower-meander line (LML) at metal layer M₄. The presence of the LML increased the total antenna length, which reduced the fundamental resonant frequency by 22% compared to the single-meander line (SML).

A sheet of metal was added as a ground plane to increase bandwidth with a size of 1525 µm × 250 µm. The ground plane was implemented at the bottom metal layer M₁ and connected to four ground PAD_S. Both meander line layers had identical structure. Line finger had a line width of 15 µm, line length of 1150 µm, and line spacing of 20 µm between each two fingers in the same layer. The UML had five fingers while the LML had only four fingers. The LML had the same alignment as the UML but shifted by 17.5 µm, which increased the number of tuning bands. A signal PAD and four ground PADs, each 80 µm × 60 µm, were set and separated by 150 µm. The ground layer was connected to the four ground PADs through four vias. The UML had a 50-Ω line connection with the signal PAD and both meander layers were stacked through Via1 (ChemoMetec, Allerod, Denmark).

3. Antenna Performance Results

The design and performance of the DML antenna were simulated using a high-frequency structure simulator (HFSS) ver. 15, which is a three-dimensional electromagnetic field simulator. Simulation results showed that the proposed structure had four bands of operations. The first band at 22 GHz with bandwidth extended from 20 GHz to 26 GHz, the second band at 34 GHz with bandwidth extended from 31 GHz to 37 GHz, the third band at 44 GHz with bandwidth extended from 42 GHz to 46 GHz, and the fourth band at 58 GHz extended from 56 GHz to 65 GHz at −7.5 dB reflection coefficient. To investigate the effect of the proposed dual-meander structure, simulation for a single-meander line (SML) antenna located at the metal layer M₆ with same previous dimensions was carried out. The results showed the resonant frequency of SML at 28 GHz compared to 22 GHz for DML. The two higher resonant frequencies of SML at 43 GHz and 56 GHz with narrow bandwidth reduced to 34 GHz and 44 GHz with broadened bandwidth for the DML antenna, respectively. Simulated reflection coefficients for both DML and SML antenna are shown in Figure 4a, while the real and imaginary parts for the input impedance are shown in Figure 4b. Simulations showed that the existence of the lower meander line increased the number of operational bands and improved the bandwidth for each band. The current distribution at different resonant frequencies on the two meander lines’ surfaces of the proposed antenna is shown in Figure 5.

4. Analysis of Human-Body Effect

Body-centric networks (BCNs) are mainly used in personal healthcare in 5G and IoT (Internet of Things). The major field in this area is to monitor a group of patients and control a large number of sensors with high data rates. The 5G band from 57 to 64 GHz has received much interest for BCN. To test the operation of the proposed antenna for wearable WBAN applications, two scenarios were conducted. In the first scenario, the proposed antenna under the test (AUT) was implanted inside a human body, as shown in Figure 6a. In the second scenario, the AUT was placed outside a human body with different separation, as shown in Figure 6. The dielectric properties of the human body are characterized up to 65 GHz based on the extrapolation of the data obtained through measurements up to 20 GHz [24,25]. Single relaxation time of Debye model was considered for modeling the permittivity data from 15 GHz to 65 GHz frequency range [26]. The effect of the human body was added to the proposed DML antenna as a single, effective layer with thickness 300 µm, relative dielectric constant ε_r = 11, and effective conductivity δ = 2 S/m. By setting frequency dependency using Debye model in the HFSS simulator, the reflection coefficient magnitude and phase were calculated, as shown in Figure 7.

The results showed that when the human body effect was included in simulations, the antenna impedance matching was more significant for the antenna implanted inside the human body. The resonant frequency shifted down and the operating bandwidth increased. Moreover, the phase of the reflection coefficient was slightly changed when the antenna was placed outside a human body, and it was abruptly changed when the antenna was implanted in the human body. The changed results of the antenna performance are due to the conductivity of the human body, which adds extra load on the antenna surface and changes in the electrical properties of the substrate. However, when the antenna was implanted in the human body, it became more dispersive and the operating bandwidths at −7.5 dB (VSWR ≤ 2.5) were extended from 20 GHz to 45 GHz and from 57 GHz up to 65 GHz. However, in both simulation scenarios, the antenna bandwidth still operated in the selected bands of operations.

5. Measurement of the Proposed Antenna

The proposed DML antenna was fabricated by using UMC 180 nm and the reflection coefficient was measured in the microstrip lab at the Electronics Research Institute. The reflection coefficient was achieved by using on-wafer probing and the measurement setup is shown in Figure 8a. The measurement setup was composed of one coplanar RF PicoProbe GSG (GGB industries, Naples, FL, USA 67 GHz connected through a coaxial cable to ZVA67 (Rohde & Schwarz, Columbia, MD, USA) Rohde and Schwarz vector network analyzer (VNA) from 10 MHz to 67 GHz. The fabricated UMC 180-nm die (mini@sic) 1525 µm × 1525 µm was fixed at the holder of PM5 KurlSuss (SUSS Microtec, Germany) manual-probe station with the three tips of the PicoProbe (GSG) are fixed to the Die PADS (signal PAD and its two surrounding Ground PADS). The photo of the fabricated antenna is shown in Figure 8b.

Figure 9a shows the comparison between the simulated and measured reflection coefficients of the proposed antenna. The |S11| result showed that there was a good agreement between the lower and the upper of the operating frequencies’ antenna bands. There was about −3 dB shift in the measured reflection coefficient results, which could be attributed to many factors as the thickness of the layer, which introduced some tolerance on the thickness, metal holder was not taken into account. In addition, uncertainty in the dielectric material properties were specified up to 15 GHz while the simulated dielectric material properties of the layer were identified up to 65 GHz so that the permittivity of the dielectric layer was somewhat higher than the permittivity that was identified at lower frequency. Another antenna performance of the proposed DML antenna was the realized antenna gain and radiation efficiency versus frequency, as shown in Figure 9b. The peak gain of the proposed antenna was −1 dBi at the end of the operating band at 62 GHz, while the radiation efficiency was about 35%.

The simulated 3D polar radiation patterns of the realized antenna gain at four resonant frequencies, 22 GHz, 34 GHz, 44 GHz and 58 GHz, are shown in Figure 10a–d, respectively. The normalized 2D radiation pattern of the proposed antenna with and without the presence of the human body at two main coordinates, φ = 90° and Ɵ = 90°, were compared at different resonant frequencies, as shown in

Table 1. The 2D polar radiation pattern at different resonant frequencies.

Plane	22 GHz	34 GHz	45 GHz	58 GHz
φ = 0°
Ɵ = 90°

without human body, with human body.

.

In references Huang et al. and Cao et al. [19,27], antennas resonated at 24 GHz but with narrow bandwidth and a larger area than the proposed antenna. The proposed antenna has low antenna gain at lower frequency; this can be explained due to the small area which reduces the aperture area of the antenna. The proposed antenna was compact when compared to other structures and it had a large number of operational bands. The multiband of operation with wideband bandwidth makes the proposed antenna a good choice for BCN to cover a large number of sensors for many patients. Essentially, the antenna provides a wireless power transfer system that can be integrated for battery-less applications.

Table 1 shows the effect of the human body on the antenna performance by reducing the antenna directivity and increasing the radiation pattern beam width.

Table 2. Comparison of the proposed antenna with other on-chip antennas.

	Reference
	[27]	[19]	[20]	[28]	[29]	[30]	This Work
Freq. GHz	24	24/60	40–50	60	94	67	22/34/44/58
Gain dBi	−8	−9/−1	−0.8 to 3.3	−3.2	Rehman-2.71	−8	−20/−15/−10/−1
BW GHz	-	0.18/0.7	5.3	5.5	15	4	5/6/4/8
Area mm2	3 mm length	0.794	1.87	2.25	0.22	0.875	0.23 mm2
Tech.	0.13 μm CMOS	0.13 μm CMOS	0.18 μm CMOS	0.18 μm CMOS	IHP 0.13 μm Bi-CMOS	0.18 μm CMOS	0.18 μm CMOS

shows the performance comparison between the proposed antenna and other previously reported in literature within the same frequency range. In the 60 GHz band, the proposed DML antenna indicated a comparable gain with Huang et al. and Wang and Sun [19,28] with larger bandwidth and achieved multiband of operation. While Khan et al. and Song et al. [29,30] had lower gain and single band of operation at frequency higher than the proposed antenna, both antennas had the same efficiency (35%) at 65 GHz range.

6. Conclusions

A compact dual-meander line (DML)-integrated antenna for mm-wave BCNs’ biomedical applications was proposed in this paper. The proposed antenna consisted of two stacked layers of meander lines and a ground metal layer. The dual-meander line structure increased the number of tuning bands and improved the bandwidth. The proposed antenna resonated at four bands, 22 GHz, 34 GHz, 44 GHz, and 58 GHz, with a maximum gain of −1 dB at 62 GHz and radiation efficiency of 35%. The proposed DML resonated at 22 GHz with reduction in resonant by 6 GHz with 22% area reduction, large number of tuning bands, and a simplified integration with implantable/wearable systems when it is compared with a traditional single-meander line antenna. The proposed antenna had acceptable performance when it was implanted or worn on the human body.