Young-Sik Seo, Zachariah Hughes, Deena Isom, Minh Quoc Nguyen,. Sanchali Deb ... stimulation efficacy and need for surgical placement of the permanent ...
2012 European Microwave Conference, Amsterdam RAI, The Netherlands, Oct. 28 – Nov. 2 2012
Wireless Power Transfer for a Miniature Gastrostimulator Young-Sik Seo, Zachariah Hughes, Deena Isom, Minh Quoc Nguyen, Sanchali Deb, Smitha Rao* and J.-C. Chiao Electrical Engineering Department, The University of Texas at Arlington, *Med-Worx, TX Abstract — In this work, wireless power transfer issues were investigated for a gastroparesis management system consisting of an endoscopically-implantable miniature batteryless/rechargeable wireless gastrostimulator and a wearable battery-operated transmitter module. Output power for stimulation and input power consumption in the transmitter as well as the wireless power transfer efficiency of two coil antennas with significantly different sizes were examined. With an implant antenna coil of 1×3.5 cm2 and 5-cm radius transmitter antenna coils of varying coil turns, power transfers and efficiencies at the individuallytuned resonance of 1.3MHz were experimentally obtained at various coil separation distances from 4 to 10 cm between two coils. Taking power consumption in the wearable module and tolerance of parasitic capacitance in the environment into consideration, a 17-turn transmitter coil was chosen. RF losses through tissues were also examined with saline mimicking the abdomen tissues. At a distance of 10 cm as the worst scenario, a load current 3.02 mA was achieved which was still sufficient for gastrostimulation. Measurements of output voltage and efficiency distribution were mapped in 3-D to examine the effects of antenna misalignment due to body and stomach motion. This work provided insights into a complete system design for an endoscopically-implantable gastrostimulator that could relieve the suffering of many gastroparesis patients. Keywords: wireless power; inductive coupling; implant; gastrostimulator; mapping
I.
INTRODUCTION
Gastroparesis is a motility disorder associated with symptoms such as nausea, vomiting, early satiety, weight loss, abdominal pain, and stomach spasms. Chronic gastroparesis patients often suffer from the delayed gastric emptying of solid meals [1]. The World Health Organization reported in 2000 that 171 million people globally suffer from gastroparesis [2]. Idiopathic and post-surgical gastroparesis have been estimated to rise to 366 million by the year 2030. The number is expected to increase significantly with the increase in the number of diabetic patients globally and the considerably high risk of gastroparesis (50%) as a co-morbidity of diabetes mellitus, whether type-I or II Drug therapies with prokinetics have been used for symptomatic control, but none is proven effective. About 40% of gastroparesis patients could not tolerate the chronic use of prokinetics due to strong side effects [3]. Relief of symptoms has been reported with gastric electrical stimulation (GES) in clinical trials [4, 5]. At present, a neurostimulator (EnterraTM, Medtronic) approved by the Food and Drug Administration has been employed to deliver energy to stomach tissues through two wires. It delivers an electrical pulse train at 14 Hz and a current of 1‒5 mA. It can be wirelessly programmed before implantation to deliver different doses of current. The reader
antenna has to be physically touching or very close to the cross section surface of the neurostimulator, since it was designed for transmitting radio signals through skin. To determine the stimulation efficacy and need for surgical placement of the permanent stimulator, a temporary endoscopic stimulation (EndoStim) method is used. Electrodes are attached inside the stomach placed via endoscope with connection wires coming out through esophagus, throat and nose, to a stimulator held in front of the patient’s chest [6]. The patient has to wear the device with the wires taped to his/her face continuously from 2 weeks to 3 months. The transnasal wires adversely affect the lifestyle and normal functioning of the patient. Once the doctor determines that the stimulation could reduce symptoms, the patient will then receive permanent implantation. The batterybased EnterraTM device is large (60mm×55mm×10mm) and hence can only be placed surgically. The procedure lasts 1−3 hours, requires general anesthesia and results in hospitalization for 1 to 5 days. The large dimensions of the device are due to the battery. In addition, the limited battery lifetime necessitates repeated surgeries to replace the implant every 3‒5 years. This brings additional pain and costs to the gastroparesis patients who are already weak and often have multiple complications. II.
ENDOSCOPICALLY IMPLANTABLE GASTROSTIMULATOR
The aforementioned issues call for a dire need of a miniature gastrostimulator that is small enough to be implanted inside stomach via out-patient endoscopic procedures, does not require battery or requires a small rechargeable battery to avoid future repeated replacement, and is wirelessly controllable to eliminate all wires. Fig. 1 shows the concept. We have previously developed such a gastrostimulator prototype [7]. The wireless stimulator measures 35×10×8 mm3 and weighs 5.2 g, which could be either a batteryless one or one with a coin rechargeable battery, powered or recharged by inductive coupling. In our previous work [7], the implants were tested in live pigs. The electrogastrogram (EGG) recordings
Figure 1. The concept of wireless power transfer for a gastrostimulaotr. A wearable battery-operated power transmitter sends RF energy to operate the implan attached on the mucosal tissue of the stomach. Page 1
Figure 2. The equivalent circuit for the wireless power transfer via inductive coupling. L1 and L2 represent the antenna coils.
demonstrated that the gastric slow waves became more regular and of constant amplitudes when stomach tissues were stimulated. The frequency-to-amplitude ratios of EGG signals also changed significantly with stimulation. The experiments were conducted using a gastrostimulator with a 3.0-V, 11-mAh, rechargeable lithium-ion battery [7]. The battery needs to be recharged after 19 hours of use, with approximately 30 minutes required for each charging cycle. This miniature stimulator was designed for the human stomach, where mucosal impedance varies between 200 and 800 Ω [6], and should deliver currents in the range of 1‒5 mA. The delivered current measured in the pig model was at 1.9 mA, lower than the anticipated value of 5 mA value for humans. The reason for the lower current was due to the higher impedance of 1751 Ω that was measured in the pig stomach. Nonetheless, the feasibility of an endoscopically-implantable gastrostimulator was demonstrated and the function of modulating stomach slow waves was also confirmed even with a lower current. Based on the preliminary results in animals, in this work we focused on improvement in wireless power transmission between the implant device and the wearable transmitter module which will be used to directly power up or recharge the implant. This serves two purposes. First, it is preferred to further reduce the size of the implant. While the circuitry space could be reduced significantly with an ASIC (applicationspecific integrated circuit) chip, the battery size remains to dominate the size. Thus, it is our aim to develop a batteryless stimulator that could be powered by an external wearable module. The wearable module needs to be operated by its own batteries, which demands high efficiency in energy transfer in order to prolong battery life and keep battery small/light. For the recharging purpose, even though the patient could wear a transmitter module that plugs into an AC power source, it is better to minimize the magnetic field strength by increasing the coupling efficiency in order to avoid local tissue heating due to intensive continuous RF power transmission. III.
maximal exposure [8]. A transistor IRF510 (Fairchild) was used in the class-E amplifier switched by a 6-Vpp square wave at a 50% duty cycle from a source. The coil coupling could be modeled as the near-field mutual inductance M in a transformer. The load resistance, RL, was considered as the tissue impedance, although usually time-, location- and patientdependent, was specified in the range of 500 to 1500 Ω that was measured in human stomach tissues. From our previous work [9], the lowest power transfer efficiency was measured at the low impedance (500 Ω) and increased when load impedance increased. Thus we fixed the load resistance at 500 Ω to examine the worst scenario cases. B. Antenna Design In our animal experiments [8], the implant cross section determined the coil antenna dimensions as 1×3.5cm2 since the metal wire was wrapped around the rectangular-shape printed circuit board to maximize its aperture size. The coil wire connected to a tuning capacitor, followed by a voltage multiplier and voltage regulator. It should be noted that the turn number of the coil antenna and the metal wire gauge were limited in order to limit the thickness of the package so that the implant could be delivered through an endoscope. We chose AGW-26 wire with a tradeoff between unit-length resistance and wire thickness. 13 turns of wire were wrapped around the circuit board making it 2 mm in thickness and producing an inductance of 10 µH. The transmitter antenna was made of AWG-26 copper wires with various radii from 4 cm to 6 cm and various coil turn numbers. Previous experiments [9] showed that the 5-cm radius antenna, in general consumed a lower input current, had a smaller physical size so it would be more comfortable to wear, and had a lower inductance so the matching capacitance would be sufficiently large to tolerate parasitic capacitance in the environment preventing the resonance frequency from shifting. The antennas were made with a consistent tensile strength applied to the wires when wrapping the coil. C. Mesurement Setup An apparatus was made to perform accurate 3-D measurements for various distances (z) and alignment (x and y) between the two antennas. From z=4 to 10 cm, slots were made
SYSTEM DESIGN
A. Circuit Configuration The equivalent circuit of the system for wireless power experiments is illustrated in Fig. 2. L1 and L2 represent the transmitter and implant coil antennas which have a significant difference in physical size. C1 and C2 are matching capacitances for resonance. R1 and R2 are equivalent series resistance (ESR) contributed by the metal coil wires and components. V1 is the source which generated an amplified sinusoidal waveform at 1.3 MHz through a class-E amplifier. The 1.3 MHz frequency was chosen because of the low electromagnetic energy attenuation in tissues and permitted
Figure 3. Input and output powers and efficiencies at RL=500 Ω. The tag antenna was fixed at a distance of 4cm from a 5-cm radius transmitter antenna. Page 2
1 cm apart along a wooden rail to hold the transmitter antenna vertically, facing the implant antenna which was mounted on a large acrylic plate with alignment grids. The centers of the transmitter and implant coils were aligned as (x=0, y=0) and the implant coil was allowed to move ±8 cm in both x and y directions. To avoid interferences, metal objects and equipment were kept at a sufficient distance from the measurement setup. The rectified output voltage across the load resistor was measured to find the output current/power while the input current to the class-E amplifier was measured to compute the input power. The wireless power transfer efficiency is the ratio between input and output powers. IV.
EXPERIMENTS AND RESULTS
A. Power and Efficiency Measurments Input and output powers as well as the wireless power transfer efficiency were measured at 4-cm distance for the 5cm radius antennas with various coil turns. Once the transmitter antenna coil was made, the matching capacitances C1 and C2 were adjusted to achieve the resonance resulting the output voltage to reach the maximum. The result is shown in Fig. 3. The highest efficiency obtained with the 18-turn (N=18) coil was 6.61%. However, the efficiencies between 14 and 18 turns were not very different varying within 0.17%. In contrast, the input power drawn from the system decreased as the number of coil turns increased. The 17- and 18-turn coils had at least 9% less power consumption than others. Practically, for our battery-operated portable transmitter module, 17- or 18turn coil antennas would be more suitable since they consumed less current. In comparison among the higher turn numbers of coils, another practical parameter we considered was the parasitic effect. As the turn number of coil increased, the self-inductance increased resulting in a smaller matching capacitance (226 pF and 121 pF for 14 and 18 turns), as shown in Fig. 4, which makes the circuit more susceptible to parasitic effects from environmental changes. For example, the effective dielectric constant near the coil wires varied when the transmitter coil antenna was firmly conformed on the pig abdomen skin indicated by the resonant frequency shift and the consequential
Figure 5. Input and output powers for different media, air and 0.4% NaCl solution, at various distances between two antennas. The 17turn coil antenna was used.
change of efficiency. Therefore, since the 17 and 18-turn coils generated similar efficiency and transferred power, a coil with fewer turns was preferred. B. Tissue Effects In our previous work during the in vivo animal experiments [10], the pig abdominal wall was estimated to be 3 cm in thickness. By comparing the received current at a distance of 3 cm in air, the attenuation of RF energy in the 3-cm thick tissue reduced the amount of stimulation current to 77%, from 2.93 mA to 2.26 mA, and to 76%, from 6.56 mA to 5 mA, for two different input power settings. In this work, we used saline to mimic the tissue attenuation effect. The 17-turn coil was placed at a 4-cm distance from the implant coil. A bag of 0.4% solution of NaCl was placed between the coils. The transferred power and efficiency were compared between air and saline as shown in Fig. 5. At 4 cm, the loss of transferred power in the saline solution was about 30% of the power transferred through the free space. The differences became smaller as the distance increased as expected since the transferred power decayed with distance in magnetic field coupling. It should be noted that the input power gradually increased with the distance. This is due to the effective dielectric constant variation from the water that changed the parasitic capacitance in the coil. Nonetheless, the output currents obtained at the 10 cm distance were 3.05 mA through the free space, and 3.02 mA through the saline solution, which was above the required minimal electrical current of 1mA for gastrostimulation. C. Voltage and Efficiency Distribution Maps To map the voltage and efficiency distribution at different distances, the implant antenna was moved in x-, y- and zdirection. The results are shown in Figs. 6 and 7. Both voltage and efficiency distributions were illustrated in absolute and normalized values. The normalized values were the data normalized to the maximum value at that antenna separation distance which occurred at the centers.
Figure 4. Output current at RL=500 Ω. The matching capacitance decreases due to the increasing self-inductace as the number of turns increases. Less matching capacitance meaning the system is more vulnerable to parasitic effects from the surrounding environment.
In both figures, it was clear that the output power and efficiency reached the maximum when two coil antennas were centered. However, these maps were to provide us information about the effects of antenna misalignment due to body and stomach motions. The average 3-dB (50%) power radii in each Page 3
(a)
(b)
Figure 6. The output voltage distribution (at x-y plane) at various distances (z) in (a) absolute and (b) normalized values.
x-y plane were 3.7, 3.6, 3.75, 3.83, 4.15, 4.4 and 4.56 cm at the distances of 4, 5, 6, 7, 8, 9, 10 cm, respectively, for a transmitter antenna size of 5 cm and 17 turns. Because the 3dB points did not follow a perfectly circular shape in the x-y plane, the radius was averaged over the distances from all 3-dB points to the center. Similarly, the efficiency maps also provided the average 50% efficiency drop points at 3.46, 3.42, 3.67, 3.85, 4.09, 4.43 and 4.57 cm from the centers at the distances of 4, 5, 6, 7, 8, 9, 10 cm, respectively. The 50% efficiency contours in each plane are shown in Fig. 8. At a 4cm distance, the produced current at the center was 14.78 mA while it became 9.32 mA at a radius of 4cm. At a 10-cm distance, the produced current at the center was 3.34 mA while it became 2.52 mA at a radius of 4 cm. Although the measurements were conducted in air, the estimated load current and efficiency values could be extrapolated from the saline experiment. V.
CONCLUSIONS
In this work, the wireless power transfer issues in terms of output power and input power consumption as well as the transfer efficiency were investigated for a gastroparesis management system consisting of an endoscopicallyimplantable miniature batteryless/rechargeable wireless gastrostimulator and a wearable battery-operated transmitter module. The transfer powers and efficiencies were examined at various distances between the antennas and with the effects of tissue attenuation mimicked by that of saline. The output voltage and efficiency maps were generated to examine the effects of antenna misalignment. The obtained information will be used to guide the final system design for a miniature batteryless wireless gastrostimulator that could ease the suffering of many gastroparesis patients. ACKNOWLEDGMENTS We sincerely thank Intel, Texas Instruments, Texas Health Resources, and Med-Worx for their support.
(a)
(b)
Figure 7. The wireless power transfer efficiency distribution (at x-y plane) at various distances (z) in (a) absolute and (b) normalized values.
Figure 8. The contours of 50% efficiency at different antenna separation distances.
REFERENCES [1]
Gastroparesis Clinical Research Consortium, “Psychological dysfunction is associated with symptom severity but not disease etiology or degree of gastric retention in patients with gastroparesis,” Am J Gastroenterol. Vol. 105, No. 11, pp. 2357-67, 2010. [2] S. Wild, et. al., “Global prevalence of diabetes: estimates for the year 2000 and projections for 2030.” Diabetes Care, Vol. 27, No. 5, pp.10471053. 2004. [3] F. E. Eckhauser, et. al., "Safety and Long-Term Durability of Completion Gastrectomy in 81 Patients with Postsurgical Gastroparesis Syndrome," Am.Surg., Vol. 64, No. 8, pp.711-717, 1998. [4] C. Anand, et. al., “Gastric electrical stimulation is safe and effective: a long-term study in patients with drug-refractory gastroparesis in three regional centers,” Digestion. Vol. 75, pp. 83-89. 2007. [5] J. Forster, et. al., “Further Experience with Gastric Stimulation to Treat Drug Refractory Gastroparesis,” Am.J.Surg. Vo. 186, pp. 690-695, 2003. [6] T. Abell, et. al., "Gastric Electrical Stimulation for Medically Refractory Gastroparesis," Gastroenterology, Vol. 125, No. 2, pp.421-428, 2003. [7] S. Deb, et. al., “An Endoscopic Wireless Gastrostimulator (with video),” Gastrointestinal Endoscopy, Vol. 75, No. 2, pp 411- 415, 2012. [8] IEEE Standard, "IEEE Standard for Safety Levels With Respect to Human Exposure to Radio Frequency Electromagnetic Fields, 3 kHz to 300 GHz," IEEE Std C95.1-2005. 2006. [9] Y.S. Seo, et. al., “Wireless Power Transfer by inductive Coupling for Implantable Batteryless Stimulators,” IEEE International Microwave Symposium, 2012, accepted. [10] S. Deb, et. al., “Miniature Wireless Gastro-Stimulators for Gastric ysmotility,” Medical Engineering & Physics, 2012, in review.
Page 4
