Fig 1.
Illustration of a cyber-physically organized under-rubble biobotic swarm based mobile sensor network.
The illustration uses real images of insect biobots with neurostimulation backpacks and implanted stainless steel wire electrodes. Each of the ZigBee enabled system-on-chip based backpacks can be configured as a sensor node in the network.
Fig 2.
Stainless steel and EGaIn based electrodes.
(a) Electrode setup with commercially available 127 μm PFA insulated SS wire electrodes. One end of the electrodes are soldered onto a small connector for connection to a neurostimulation backpack or an external circuit using an FFC connector. Insulation from the other ends are removed by 3.5 cm using a scalpel; these are implanted into the mesothorax or flagellum of an antenna of a cockroach. (inset) tip of SS and EGaIn coated SS electrodes. (b) EGaIn in 1.5 cm long Tygon tubes (0.5 mm inner diameter, 1.5 mm outer diameter) for use as an electrode. The tube is filled with EGaIn to a length of 1 cm with EGaIn from one end, carefully preventing any trapped air bubbles inside. This end formed the electrolyte-electrode interface with saline while a metal wire was inserted 0.5 mm deep from the other end for connection to external circuit.
Fig 3.
Micro fabricated gold electrodes.
(a) Cross-sectional illustration (top) and SEM image (bottom) of a microfabricated gold electrode. The two thin horizontal lines running along the SEM image are the gold layers. (b) Microfabricated electrodes used in the experiments (left). The flat top end is inserted to an FFC connector for for connection to external circuit. (inset) Optical microscope and SEM images show an electrode pad at the electrode tip before and after electropolymerization using PEDOT:PSS conductive polymer.
Fig 4.
Equivalent circuit model of an electrolyte-electrode interface.
(a, b) Model of an interface between working and reference electrodes labeled as W.E. and R.E. respectively, shown in a 3-cell circuit (a) with a counter electrode C.E. and in a 2-cell circuit (b). The charge conducting interface is conceived of a parallel RC network with parameters Rct (the charge transfer resistance) and Cdl (the double layer capacitance). Rs is the electrolytic resistance. (c, d) Modified version of the equivalent circuit model in 4(a, b) for SS-EGaIn electrodes. An additional RC parallel network is considered for the intermediate layer between SS and EGaIn, where network parameters Rint and Cint function corresponding to Rct and Cdl respectively.
Fig 5.
EIS plots for all electrodes types.
Plots using 2-cell and 3-cell data for the same type of electrode have close match between impedances at corresponding frequencies. SS electrodes have high impedances, followed by Au electrodes comparable to SS at lower frequencies. EGaIn electrodes have low impedances with AuP the lowest among the tested electrodes. SS-EGaIn, a hybrid, formed by the addition of EGaIn to SS represents an enhancement of the SS electrodes.
Fig 6.
Mean values of interface circuit model parameters estimated from 2-cell and 3-cell measurements.
(a) Rct−charge transfer resistance, (b) Cdl−double layer capacitance, (c) Rs−electrolytic resistance, and (d) Rint and Cint−charge transfer parameters at the intermediate layer between SS and EGaIn in the SS-EGaIn electrodes. AuP and EGaIn electrodes have lower Rct and higher Cdl values designating them the most favorable among the tested electrodes. Model parameters estimated from 2-cell and 3-cell data of the same electrode types have approximately equal values.
Table 1.
Paired sample t-test of 2-cell and 3-cell measurement of electrodes.
Table 2.
Interface circuit model parameters–ideal interface.
Table 3.
Interface circuit model parameters–non-ideal interface.
Fig 7.
Charge injection capacity for all electrodes types in 2-electrode cells.
The charge injection capacity of all electrode types are determined from CV plots and were found to be correlated to the corresponding EIS impedance. AuP and EGaIn electrodes have the highest charge capacity, with AuP having a prominent redox reaction.
Fig 8.
EIS plots of electrodes obtained over a period of two weeks.
Results for 2-electrode cell data are shown at t = 0, 1 week, and 2 weeks. After two weeks, moderate changes in the impedance plots were observed, owing to the salt accumulation at the interfaces. EGaIn started to display some irregularity due to the changes on its oxide layer brought upon by interaction with water molecules. SS-EGaIn began behaving more like SS electrodes after two weeks due to disassociation of EGaIn from the surface of the SS electrodes.
Fig 9.
Mean change in 2-cell circuit model parameters with time.
(a) Rct−charge transfer resistance, (b) Cdl−double layer capacitance, (c) Rs−electrolytic resistance, and (d) Rint and Cint−charge transfer parameters at the intermediate layer between SS and EGaIn in the SS-EGaIn electrodes. Changes in the parameter values reflect the trends observed in the EIS plots. Rct changed for all electrodes except for EGaIn whose Cdl decreased by about five times over the period of two weeks. The Cdl of SS stayed fairly constant, while that of AuP changed, but remained considerably high. Parameters of the SS-EGaIn electrodes were found to gradually become similar to the initial parameters of the SS electrodes.
Fig 10.
Charge injection capacity of electrodes calculated from accelerated aging data.
The AuP electrodes were found to have the highest charge injection capacity which appeared consistent throughout the two weeks duration. It was followed by EGaIn electrodes which had the second highest charge capacity, and SS-EGaIn and SS electrodes which had relatively low capacity with SS electrodes having the lowest capacity. SS-EGaIn had increased charge injection capacity compared to SS, but it dropped and became more like that of SS with time as the EGaIn dislodged from the SS surface.