Rula Alrawashdeh, Yi. Huang and Qian Xu, University of Liverpool, UK.
Experimental measurements of implantable antennas in the actual human bodies are often impossible but it is critical that they comply with SAR legislation. Simulation with anatomical body models provides a very helpful and precise tool for evaluating implantable antennas. These models provide accurate data about the antenna radiation characteristics and matching behavior.
Implantable antennas play a major role in bio-telemetric communications. Although implantable antennas are very beneficial, their design is very challenging. In addition to the challenges of small antennas (small gain and narrow bandwidth), more challenges are introduced because of the complex and lossy structures of the human bodies in which these antennas radiate. Simplified body models introduce many uncertainties into the design of implantable antennas because of their uniform structures, small sizes and inaccurate equivalent dielectric material. To validate performance of implantable antennas accurately, evaluations have to be conducted using anatomical body models....
In this article, the importance of carrying out evaluation in the anatomical human bodies is demonstrated by comparing the performance of a flexible implantable antenna in a simplified body model of an elliptic cylindrical shape, dimensions of 180, 100 and 50 mm and muscle equivalent material with the performance in the CST® Laura human voxel model.
Antenna performance in a simplified body model
The selected antenna is a meandered flexible antenna which can be bent around implants of 3.2 mm in radius and 10 mm in length. The antenna is initially designed in the simplified body model. The antenna has a loop structure because of the preferred performance of magnetic antennas inside the non-magnetic human body. The flat and bent structures of the proposed antenna are shown in Figure 1. The width of each horizontal radiator (W1 = W2 = W3) is 0.5 mm whereas the width of the top radiator (W4) and of each vertical meander (Wv) is 1 mm, the length of the horizontal meanders (Lh and Lh2) is 20 mm and 9 mm, respectively. The spacing between each pair of the horizontal meanders (Sv) is 2.5 mm and the spacing (S) between the symmetrical halves of the antenna is 2 mm. These dimensions are selected to cover the MedRadio and 433 MHz ISM bands for S1,1 < -10 dB as shown in Figure 2.
It is expected that the simplified body model underestimates the gain and radiation efficiency because of its smaller size in comparison with the full anatomical body model. However, these values have been computed for the purposes of comparison. A gain value of -28.4 dBi and a radiation efficiency of 0.057% have been computed inside this model.
Implantable antennas have to comply with the specific absorption rate (SAR, 1.6 W/kg per 1 g average, 2 W/kg per 10 g average) in order to prevent hazardous heating of the biological tissues. The 1 g SAR is computed because it is more restrictive than the 10 g SAR. A 1 g averaged SAR value of 624 W/kg has been computed for an input power of 1 W - based on this SAR value, the antenna can be provided with a power of 2.6 mW (4 dBm) and still satisfy the 1 g averaged SAR limitations. This is much larger than the input power that is normally provided to implantable devices, 0 dBm. The 3D radiation pattern is also checked in this body model, and an omnidirectional radiation pattern is obtained as shown in Figure 3. This symmetric radiation pattern is obtained because of the symmetric structure around the implantable antenna where all the antenna parts are surrounded by muscle.
The antenna was also simulated in a human-shaped model made of the same material. In this model, the resonant frequency was shifted down to 338 MHz and the overall gain and radiation efficiency are degraded (-48.8 dBi and 0.000283%). This is due to the larger equivalent permittivity and conductivity of muscle in comparison with a multilayer anatomical human body model, which includes layers of fat and skin which have a smaller permittivity and conductivity.
Antenna performance in an anatomical body model
To validate the antenna performance accurately, simulations using the anatomical body models have to be conducted. The performance of implantable antennas may differ from one person to another, and so multiple voxel body models of different age, gender and structure are available in the CST voxel family as shown in Figure 4.
To show the importance of implantable antennas evaluation in the anatomical human body models and how the performance differs from the simplified body model, this article uses the CST Laura human voxel body model. This model represents a 43-year-old female with height of 163 cm and weight of 51 kg. The antenna is placed in the arm of this model as shown in Figure 5 with the same orientation in the simplified body model. Simulations are conducted with the full body using the transient time solver, with discrete ports used to feed the antenna. The reflection coefficient is shown in Figure 6.
It can be seen from the figure that this antenna has an ultra-wide bandwidth of at least 203 MHz (327 MHz - 530 MHz) for S1,1 < -10 dB which covers the MedRadio and 433-434 MHz ISM bands. The resonant frequency has been shifted up from 392 to 413 MHz because the antenna is directly implanted beneath the fat layer in the body model. The gain value is reduced to -45 dBi in the full body model and this is because of the larger losses in the larger body model which has more lossy tissues. The 3D radiation pattern becomes directive and asymmetric in this body model as shown in Figure 7 because of its asymmetric structure. A 1-g averaged SAR value of 580 W/kg is obtained for this orientation. By comparing the SAR in both models, it can be concluded that the homogeneous body model overestimates the SAR value.
The antenna orientation is a very important parameter to consider, and can only be examined in the anatomical body model. As explained above, in simplified and symmetric body models all parts of the antenna are surrounded by the same material – muscle. However, this is not the case in the anatomical body models, where each part of the antenna are surrounded by different materials. This is especially true for flexible and conformal antennas.
The antenna is simulated rotated around four different angles around each axis. The antenna still covers both bands of interest no matter what its orientation. Some orientations increased the gain and radiation efficiency are increased while decreasing the 1 g SAR, since some of the antenna parts are close to the fat layer, which has smaller conductivity than muscle (s = 0.04 and 0.79 S/m, respectively). However, this is accompanied with an upshift of the resonant frequency and thus reduces the band margin below the MedRadio bands to be 20 MHz for some orientations because the total effective permittivity and length around/of the antenna become smaller. The radiation pattern also differs from one orientation to another as expected. The maximum radiation is obtained in a direction that is away from the body most of the time.
This data has practical applications. The implant around which the antenna is bent can be placed during the surgery in different orientations and the orientation which provides optimum performance has to be selected. In addition, capsule antennas rotate while passing through the digestive tract; their performance has to be checked at different orientations in order to guarantee its desired performance.
Anatomical computer-based body models such as the CST voxel body models play an important role in the evaluation and validation of the performance of implantable antennas. Although the design of implantable antennas can be accelerated by the use of simplified body models, it doesn't provide accurate data about their radiation characteristic. The use of the anatomical body models is necessary as a source of reliable data and a substitute for the impossible experiment validation in the full human bodies. Some of the effective parameters on the antenna performance such as the antenna orientation can only be evaluated using anatomical body models.
R. Alrawashdeh, Y. Huang and P. Cao, “Flexible meandered loop antenna for implants in MedRadio and ISM bands,” Electronics Letters, vol. 49, 1515-1517, 2013.