Design considerations for piezocomposite materials for electrical stimulation in medical implants

Abstract

Incidence of non-union following long bone fracture fixation and spinal fusion procedures is increasing, and very costly for patients and the medical system. Direct current (DC) electrical stimulation has shown success as an adjunct therapy to stimulate bone healing and increase surgery success rates, though drawbacks of current devices and implantable battery packs have limited widespread use. Energy harvesting utilising piezoelectric materials has been widely studied for powering devices without a battery, and a preclinical animal study has shown efficacy of a piezocomposite spinal fusion implant resulting in faster, more robust fusion. Most piezoelectric energy harvesters operate most effectively at high frequencies, limiting power generation from loads experienced by orthopaedic implants during human motion. This work characterises the efficient power generation capability of a novel composite piezoelectric material under simulated walking loads. Building on compliant layer adaptive composite stacks (CLACS), the power generation of mixed-mode CLACS (MMCLACS) is defined. Utilising poling direction to capitalise on in-plane strain generation due to compliant layer expansion, MMCLACS significantly increased power output compared to a standard piezo stack. The combination of radial and through-thickness poled piezoelectric elements within a stack to create MMCLACS significantly increases power generation under low-frequency dynamic loads. This technology can be adapted to a variety of architectures and assembled as a load-bearing energy harvester within current implants. MMCLACS integrated with implants would provide enough power to deliver bone healing electrical stimulation directly to the fusion site, decreasing non-union rates, and also could provide quantitative assessment of healing progression through load sensing.

Keywords: Bioelectric energy sources; electrical stimulation; fracture healing; sensing materials; smart materials; spinal fusion.

Figure 1.

Radial (a) and through-thickness (b) polarization of piezoelectric discs. Arrows represent poling direction.

Figure 2.

Experimental electromechanical setup and MMCLACS configurations. (a). MMCLACS were electromechanically tested to compare voltage and power produced at varying low frequency, sinusoidal compressive and torsion load using a biaxial MTS MiniBionix 858 with hydraulic grips. Load was applied and voltage output was measured across a shunt resistance sweep in series with MMCLACS. (b). Schematic showing R-CLACS layup. (c). Schematic showing RT-CLACS layup. (d). Schematic showing T-CLACS layup. In b-d. arrows represent poling direction and positive/negative electrodes are labeled on each PZT disc. CL represents the compliant layers interdigitated between each PZT disc.

Figure 3.

(a) Average power generated by each MMCLACS group as a function of applied compressive load at 2Hz. Power presented at resistance of maximum measured voltage - 2MΩ. (b) Average power generated by each MMCLACS group as a function of frequency at 1000N compressive load. Power presented at resistance of maximum measured voltage at the resistance corresponding to maximum power for each frequency: 5MΩ at 1Hz, 2MΩ at 2Hz, 1.5MΩ at 3Hz, 0.97MΩ at 5Hz. Note: Error bars represent one standard deviation and *represents a significant difference (p<.05).

Figure 4.

(a) Average power and voltage output for each MMCLACS group as a function of poling direction and resistance applied at 1000N and 2Hz. Note the resistance is plotted on a log scale for clarity. (b) Average power generation curves as a function of resistance for each MMCLACS group presented at each frequency tested at 1000N. This demonstrates the effect of frequency on power generated and resistance of maximum power generation, or optimal resistance. Error bars left off for clarity.

Figure 5.

(a). Average voltage generated from each MMCLACS group as a function of load applied (b). Average voltage generated from each MMCLACS group as a function of frequency applied. Note: voltage is the V_RMS equivalent calculated from the average amplitude of the AC voltage signal collected at each load, frequency and resistance. Error bars represent one standard deviation.

Figure 6.

(a) Average power density for each MMCLACS group as a function of load applied. Data presented for 5Hz frequency and resistance of maximum power, 0.97MΩ. (b). Average power density for each MMCLACS group as a function of frequency applied. Power data is presented at 1000N and the the resistance of maximum power for each frequency: 5MΩ at 1Hz, 2.5MΩ at 2Hz, 1.5MΩ at 3Hz, 0.97MΩ at 5Hz. Error bars left off for clarity.