SEARCH

SEARCH BY CITATION

Keywords:

  • thiol-ene;
  • intracortical electrode;
  • smart polymer;
  • neural interface;
  • plasticization

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

Neural interfaces have traditionally been fabricated on rigid and planar substrates, including silicon and engineering thermoplastics. However, the neural tissue with which these devices interact is both 3D and highly compliant. The mechanical mismatch at the biotic–abiotic interface is expected to contribute to the tissue response that limits chronic signal recording and stimulation. In this work, novel ternary thiol-ene/acrylate polymer networks are used to create softening substrates for neural recording electrodes. Thermomechanical properties of the substrates are studied through differential scanning calorimetry and dynamic mechanical analysis both before and after exposure physiological conditions. This substrate system softens from more than 1 GPa to 18 MPa on exposure to physiological conditions: reaching body temperature and taking up less than 3% fluid. The impedance of 177 µm2 gold electrodes electroplated with platinum black fabricated on these substrates is measured to be 206 kΩ at 1 kHz. Specifically, intracortical electrodes are fabricated, implanted, and used to record driven neural activity. This work describes the first substrate system that can use the full capabilities of photolithography, respond to physiological conditions by softening markedly after insertion, and record driven neural activity for 4 weeks. © 2013 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater, 102B: 1–11, 2014.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

Silicon-based neural interfaces have been shown to have poor stability during chronic implantation.[1-3] This general failure of devices, usually within a year of implantation, has been in part attributed to the extreme mechanical mismatch between the silicon substrate and neural tissue.[4, 5] To address this problem, advances in the field of flexible electronics, using polymers as substrates, have been incorporated in the design of neural interfaces.[6] Neural interfaces designed to be inserted in the tissue, such as intracortical or intrafascicular electrodes, have generally used various polyimides or Parylene-C (modulus ∼3 GPa) as flexible replacements for silicon (modulus ∼140 GPa). Devices have been fabricated from these materials in forms that maintain sufficient stiffness to allow implantation into tissue.[7, 8] These and other similar materials have also been used in flexible electronics research to successfully demonstrate a host of passive and active electronic devices.[9, 10] Silicone-based elastomeric neural interfaces (modulus ∼2 MPa) have also shown promise in the literature.[11, 12] These silicone-based materials are an obvious choice for substrates for in vivo electronics because of their low modulus and widely recognized biocompatibility.[13, 14] However, these materials lack the requisite stiffness to penetrate even soft tissue.[15] Stiff secondary insertion aids for soft elastomeric devices have been proposed, but necessitate larger insertion footprints that increase trauma at the implant site. The benefits of smart polymers, rigid during insertion and elastomeric during use, present a property space of interest in neural interfaces for polymers capable of softening from glassy to rubbery in response to a stimulus. Previous devices demonstrated in this space were limited by unnecessarily large fluid uptake and strict temperature and processing limits during photolithography.[16] A schematic listing the moduli of common neural interface materials, biological tissue, and the proposed softening substrate is shown in Figure 1.

image

Figure 1. Schematic demonstrating the (1) transfer-by-polymerization process in which the monomers comprising the responsive polymer system described are polymerized directly onto a sheet of gold. The solution is comprised of tri-thiol (i), tri-ene (ii), and diacrylate (iii) monomers. After polymerization, the bottom part of the mold is removed exposing the gold film. (2) This film is then patterned and an insulating Parylene-C film is subsequently deposited and patterned. Devices are then micromachined from the surrounding substrate. (3) Schematic indicating the modulus of materials used in neural interfaces relative to biological tissues. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Synthesizing polymers capable of softening might be envisioned with several different chemistries. For instance, the so-called thiol-ene” reaction is a general name for a well-understood reaction between an alkene (“ene”) and thiol functionality through radical or Michael-type addition.[17] This reaction has been described as a “click” reaction, a set of conditions first proposed by Kolb et al. that identifies high-yield, facile reactions.[18] Unlike the radical mechanism through which (meth)acrylates polymerize, the thiol-ene reaction proceeds through a step-growth mechanism. The result is a stoichiometric reaction between thiols and enes, when using enes that do not readily homopolymerize. When both monomers have exactly two reactive functionalities, a linear polymer is formed, and when either monomer has three or more functionalities, a cross-linked network is formed. The reaction proceeds in a slightly different manner when the alkene can homopolymerize and participate in the thiol-ene reaction, such as with (meth)acrylates. The reaction that occurs has been described as mixed-mode, where (meth)acrylate homopolymerization and thiol-ene reactions proceed simultaneously.[19]

The thiol-ene reaction is not inhibited by oxygen allowing for simple reaction setups. Low cure stresses are present in the final polymer because there is less volumetric shrinkage and delayed gelation. This leads to highly uniform, dimensionally stable polymer networks. The presence of flexible thioether linkages in the final polymer significantly decreases the glass transition temperature (Tg) and provides a key challenge in developing high Tg thiol-ene systems.[20] Ternary thiol-ene/acrylate systems create more highly cross-linked networks because of the homopolymerization of the multifunctional acrylate, which subsequently increases the Tg.[19] These resulting copolymers share characteristics with both acrylics and thiol-ene polymers. The thiol, ene, and diacrylate monomers used in this work are shown in Figure 1.

Responsive neural interfaces made with these types of substrates can behave as rigid polymers, such as polyimides, during insertion, but subsequently soften to more closely resemble elastomers. There are relatively few examples of this type of interface in the literature. Hess et al., Harris et al., and Ware et al. have demonstrated softening intracortical electrodes based on the significant swelling of thermal- and water-sensitive polymer substrates.[16, 21, 22] However, these substrates present difficulties in photolithography because of poor dimensional stability in processing, leading to poor registration for devices consisting of multiple patterned layers or cracking of the thin-film conductors used for the electrodes.

In this work, a dimensionally stable, smart polymer capable of softening in response to physiological conditions, with low fluid uptake, is demonstrated. This novel thiol-ene/acrylate system is compatible with the so-called “transfer-by-polymerization” method, described previously and briefly shown schematically in Figure 1, and is also compatible with more traditional photolithographic techniques.[16] In fact, precise control over the thermomechanical properties of the thiol-ene/acrylate substrate is demonstrated. Intracortical microelectrode arrays are fabricated, implanted, and used in acute in vivo recordings. In the opinion of the authors, these substrates are the first published example of substrates that soften considerably in physiological conditions, but maintain compatibility with standard high-yield, high-resolution photolithographic processes. In the future, devices synthesized with these materials and techniques may enable new paradigms in neuroscience research and hold promise for the treatment of epilepsy, stroke, tinnitus, Parkinson's, and other neurological disorders.

EXPERIMENTAL METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

Materials

Tricyclo[5.2.1.02,6]decanedimethanol diacrylate (TCMDA), 1,3,5-triallyl-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (TATATO), and 2,2-dimethoxy-2-phenylacetophenone (DMPA) were purchased from Sigma Aldrich. Tris[2-(3-mercaptopropionyloxy)ethyl] isocyanurate (TMICN) was purchased from Wako Chemicals. These monomers are depicted in Figure 1. All chemicals were used as received without further purification. All processing steps, except polymer synthesis, were performed in a Class 10,000 clean room.

Dynamic mechanical analysis

Dynamic mechanical analysis (DMA) was performed on a Mettler Toledo DMA 861e/SDTA. Samples were cut into cylinders approximately 1.2 mm thick and ∼3 mm in diameter. The mode of deformation was shear, and strain was limited to a maximum of 0.3%. Samples were tested at a heating rate of 2°C/min. The frequency of deformation shown is 1 Hz. Tests were conducted in a nitrogen atmosphere. DMA on explanted samples was performed by immediately removing the sample and testing in shear parallel-plate configuration. It should be noted that all samples that were implanted or swollen were chosen to be appropriately sized for mechanical testing (∼1 mm thick) and are not representative of the dimensions of the fabricated neural interfaces (∼35 µm thick). Tg by DMA is denoted as the peak of tanδ. Each composition was tested at least twice. All modulus data presented here is the shear modulus. The shear modulus (G) is related to the Young's modulus (E) through Poisson's ratio (ν) as given by inline image. Poisson's ratio was not measured for these polymers; however, ν can be generally taken as approximately 0.35 for glassy polymers and 0.5 for polymers above the glass transition. Detailed explanation of the testing of swollen or explanted samples can be found in Supporting Information.

Differential scanning calorimetry

Differential scanning calorimetry (DSC) was performed on a Mettler Toledo DSC 1 with an intracooler option. Samples were heated from room temperature to 100°C, cooled to −50°C, and subsequently heated to 200°C. Data shown are of only the second heating ramp. All heating and cooling rates were fixed at 10°C/min. Tests were conducted in a nitrogen atmosphere. Tg by DSC is denoted as the midpoint of the transition.

Swelling measurements

Swelling was measured by weight change after immersion in phosphate-buffered saline at 37°C at several time points over 1 month. Samples consisted of ∼10 mg laser machined cylinders 3 mm in diameter and 1.2 mm in height. The dry mass of each sample was measured and recorded with a balance with 0.01 mg precision. At each desired time point, each sample was removed from the PBS, and the surface of the polymer was gently dried using an absorbent wipe. The swollen mass was then recorded. Swelling is calculated as the mass change from dry to swollen normalized to the dry mass. Data reported are the average of five samples.

Polymer synthesis and electrode transfer-by-polymerization

The appropriate quantities of TCMDA and TATATO were mixed. A total of 0.1 wt % DMPA of total monomer concentration was dissolved into the solution. The vial was covered in aluminum foil to prevent incident light from contacting the monomer solution. In the covered vial, the appropriate amount of TMICN was added. Without exposing the solution to light, the vial was mixed through vortexing and sonication until the solution was visually homogenous and without air bubbles. The monomer solution was cast between two glass slides (75 × 50 mm) separated by a glass spacer, 1.2 mm or 35 µm thick. Polymerization was performed using a cross-linking chamber with five overhead 365 nm UV bulbs (UVP via Cole-Parmer) for 15 min.

Device fabrication

Devices were fabricated using standard photolithography to pattern both gold and Parylene-C. The electrodes were then electroplated with platinum. Further detailed fabrication can be found in the Supporting Information.

Neuronal cell culturing

Both glass and polymer surfaces were coated with poly-l-lysine overnight at room temperature. Cortical neurons from E16 mice were dissociated using trypsin and were cultured in neurobasal media containing B27. Cells were fixed using 4% paraformaldehyde after 3, 7, and 14 days in vitro culture and were stained with β-tubulin. The coverage of neuronal cell bodies, axons, and dendrites was quantified using Image J, and statistical analysis was performed using one-way analysis of variance. Data are presented as the average of six samples for each substrate at each time point. Error bars represent standard deviation.

Impedance spectroscopy

Impedance spectroscopy was performed using a CH Instruments (Austin, TX) potentiostat. A three-electrode configuration was used, and the tests were performed in phosphate-buffered saline with a Ag/AgCl reference electrode. Frequencies between 100 and 10,000 Hz were tested using a 5-mV sinusoidal potential.

Neural recording

Cortical probes were implanted 800 µm deep in the primary auditory cortex of rats. Detailed surgical procedures and recording parameters can be found in Supporting Information. Auditory stimuli (broadband clicks and tones) were presented at 60 dB in a double-walled acoustic chamber under light ketamine anesthesia. Error bars are the 95% confidence intervals (CIs) for the iso-intensity tuning curves.

Statement of IRB protocol

All animal work regarding the culturing of neurons was performed in agreement with University of Texas at Arlington IACUC protocol. All animal work in the cortex was conducted in accordance with the IACUC procedures at the University of Texas at Dallas.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

The goal of this work is the design of a substrate system for flexible neural interfaces that: (1) responds to physiological conditions triggering a tunable drop in modulus to reduce mechanical mismatch across the biotic–abiotic interface and (2) demonstrates compatibility with standard techniques used for the patterning of multilayer devices through photolithography. Substrates were produced through the photopolymerization of stoichiometric quantities of TATATO and TMICN with a varied amount of TCMDA. TCMDA was varied from 0 to 31 mol %. Figure 2(a,b) shows the shear dynamic mechanical response of the synthesized system of polymers. Tg, as denoted by the peak of tangent delta, increases with increasing TCMDA from 49 to 58°C. Shear modulus below the glass transition remains relatively unaffected by the addition of the diacrylate, but rubbery modulus, Gr, increases from 3.0 to 3.6 MPa. It should be noted that each of the synthesized compositions remains rigid through physiological temperatures (and thus during surgery), denoted by a vertical line in Figure 2(a), while dry. This is a requisite for implantation of a polymer film, 35 µm thick, into cortical tissue. Figure 3(a–c) shows the response of the synthesized substrate system to simulated physiological conditions. Swelling of each composition was monitored during immersion in saline solution over 4 weeks. Each sample swells less than 3% over 4 weeks. Samples of each of these compositions were also implanted onto the surface of the brain. The effect of this exposure to physiological conditions on thermal properties is demonstrated by DSC thermograms in Figure 3(b). The appearance of a slight endothermic peak near 0°C for each swollen sample and a decrease of the Tg, denoted by a step in the heat capacity of the sample of approximately 15°C, is observed for each of the three compositions. The peak near 0°C indicates the presence of small amount of free or weakly bound water capable of melting. Samples that were implanted subcutaneously or swollen in vitro showed similar behavior, and DSC of these materials can be seen in Supporting Information Figure S1. Figure 3(c) demonstrates the change in mechanical properties after this implantation. Although the body temperature shear modulus for each of the dry compositions is between 360 and 460 MPa, after 1 week of implantation, the shear modulus of each composition drops significantly to between 23 and 4.7 MPa. The modulus after exposure to physiological conditions increases with diacrylate concentration in the network. DMA of in vitro swollen materials can be found in Supporting Information Figure S2.

image

Figure 2. Comparison of the dynamic mechanical response of the thiol-ene-acrylate system. Storage modulus as a function of temperature indicates that each sample remains largely glassy at 37°C and that rubbery modulus increases with diacrylate content (a). The glass transition temperature, denoted by the peak of tanδ, also increases with diacrylate content (b).

Download figure to PowerPoint

image

Figure 3. Swelling analysis indicates that samples take up less than 3% water after immersion for up to 4 weeks in simulated physiological conditions (a). Thermal analysis shows significant plasticization after immersion for 1 week by both (b) DSC and (c) DMA. Physiological temperature is indicated with a vertical bar.

Download figure to PowerPoint

During some implants, both the 0 and 23 mol % exhibited occasional probe buckling. This was not the case for the 31 mol % TCMDA sample; the onset of the glass transition is far enough above physiological temperature to ensure softening will not occur during implantation. As such, the 31 mol % TCMDA sample was selected as the substrate for the fabrication of intracortical electrode arrays despite being the stiffest of the three materials in the rubbery regime. To begin to understand the biological response to this novel material system, embryonic mouse cortical neurons were cultured on poly-d-lysine–coated samples of the 31 mol % TCMDA containing sample. Cultures exhibited good viability over the entirety of the test, 14 days, as indicated in Figure 4(a). Figure 4(b) provides quantification of the density of neuronal cell bodies, axons, and dendrites in the cultures on both glass and the thiol-ene/acrylate. The two sample sets (n = 6 at each time point) were not statistically different at any of the observed time points.

image

Figure 4. β-Tubulin labeling of E 16 mouse cortical neurons after 3, 7, and 14 days in vitro (DIV) culture (20×) (a). Quantification of coverage of neuronal cell bodies, axons, and dendrites on glass and polymer surfaces over the culturing period (n = 6) (b). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.].

Download figure to PowerPoint

Intracortical microelectrode arrays were fabricated using a modified version of a transfer-by-polymerization technique described previously. The device is designed to be percutaneous, and electrical contact is made to the device by bonding with a conductive adhesive or by attachment to a zero insertion force connector. Figure 5(a,b) shows a single penetrating shank and a photograph of a 16-channel intracortical array fabricated on the 31 mol % TCMDA containing sample. The minimum feature size for the device is 15 µm. Across 16 measured electrodes, average impedance spectroscopy of the 177 µm2 gold electrodes and the 177 µm2 electrodes coated in electroplated platinum black is shown in Figure 5(c). The impedance of the gold electrodes at the physiologically relevant frequency of 1 kHz is 1.84 ± 0.62 MΩ (95% CI). After coating with platinum black, the impedance dropped an order of magnitude to 206 ± 116 kΩ (95% CI).

image

Figure 5. Photograph of a 16-channel intracortical electrode array next to an American penny (a), with an optical micrograph of a single penetrating shank (scale bar 60 µm) (b). Theoretical buckling analysis of this probe based on moduli measured before and after implantation. Gray area represents the force necessary to penetrate the pia of a rat given the dimensions of the penetrating shaft (c). Impedance spectroscopy of electrodes (n = 16) before and after electroplating platinum (error bars represent 95% CI) (d). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.].

Download figure to PowerPoint

The ability to consistently insert these electrodes without buckling was evaluated. Previous studies by Najafi and Hetke and Harris et al. have studied the insertion force of intracortical electrodes.[15, 23] In Figure 5(d), the theoretical critical buckling force for a softening electrode, 35 µm thick and 150 µm wide, is shown as a function of length. It is shown that the dry substrate has the prerequisite stiffness required to penetrate the pia of a rat and reach tissue of interest within 1.5 mm of the cortical surface. With low moisture absorption at body temperature (37°C), the softening substrate will no longer have the ability to penetrate this soft neural tissue. During insertion of these intracortical electrodes (800 µm long), buckling was not observed.

Three rats were implanted with intracortical softening probes. Of those, two animals had two shank arrays with eight recording sites and one animal had a four-shank array with 16 recording sites. The implanted devices were placed in the primary auditory cortex (layer IV/V), and driven neural activity was recorded in anesthetized subjects. Neural recordings consisted of multiunit clusters and local-field potentials (LFPs) on each electrode site in response to broadband click trains and iso-intensity tuning curves.[24] Figure 6(a) shows the fraction of recording sites with driven action potentials and driven local field potentials over 4 weeks postimplantation. Twenty-four hours after implantation, 100% of the channels displayed driven LFP, and 65% of channels recorded driven multiunit activity. Although more than 90% of the channels displayed some form of driven local field potential activity out to 4 weeks, there were also 15% of the channels with driven multiunit activity that could be sorted from background noise at 4 weeks.

image

Figure 6. Recording performance of softening intracortical electrode arrays in the auditory cortex of rats. Percentage of channels with driven local field potential and spiking activity over 1 month implantation (a). The extent to which this driven activity exceeded the measured spontaneous response (b). Absolute value of the average maximum evoked potential over time (c). The noise floor and spike SNR ratio over time for driven channels (d). All error bars represent 95% CI.

Download figure to PowerPoint

For the channels that displayed driven activity, the average maximum response for LFPs and peristimulus time histograms (PSTH) were normalized to the maximum activity before the stimulus was applied. This value is the average drivenness and is shown in Figure 6(b) for both local field potentials and spiking activity with a 95% CI. LFPs exhibited a maximum average evoked response more than 23.1 ± 3.08 times greater than the maximum of the average spontaneous response immediately after implantation. This drivenness of the LFPs reduced to 12.5 ± 1.27 over 4 weeks. The average maximum amplitude of the LFPs remained fairly consistent between 80 and 120 μV over the course of 4 weeks, Figure 6(c). Immediately after surgery, the PSTH drivenness was 2.63 ± 1.59. At 4 weeks, the drivenness was 4.89 ± 3.39 for channels that displayed driven neural activity.

The multiunit activity on driven channels is characterized by the average noise of the neural activity and the driven spike signal-to-noise (SNR) ratio as shown in Figure 6(d). By using a bone-screw as a ground, we obtained an average noise floor between 9 and 14 μV. The driven spike SNR was calculated as the absolute value of the maximum amplitude divided by the standard deviation of the noise floor. The initial driven spike SNR was 2.58 ± 0.51 and increased to 3.83 ± 1.00 after 4 weeks, although this increase was not statistically significant.

Figure 7 shows LFP waveforms, PSTHs, multiunit cluster waveform shapes, and iso-intensity tuning curves over 4 weeks from a single animal. For the local field potentials, four channels, one from each shaft, were chosen to demonstrate the recording over time. For the PSTHs, multiunit waveforms and tuning curves, a characteristic channel represents driven neural activity observed using a softening implant. The amplitude and SNR of the multiunit clusters are typical of planar electrodes despite the small electrode area. The LFPs have different waveform shapes across shanks but similar shapes on individual shanks, as expected due to the proximity of the electrodes. The iso-intensity tuning curves show frequency selectivity across all time points.

image

Figure 7. From a single animal, select recording data of driven neural activity are shown. The gray bar indicates the onset of the stimulus. Local field potentials from four different channels are shown over the 1 month of implantation (left). Peristimulus time histograms from the same channel of spiking activity to broadband clicks with the associated waveforms and mean waveform (black) (y scale bar is 140 µV and x scale bar is 2.5 ms) (middle). Iso-intensity tuning curves obtained from single frequency tones from a single channel indicate frequency selectivity at the chosen electrode (error bars represent 95% CI) (right).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

Development of substrate materials for chronic neural interfaces is driven by one basic tenet: the device must (under demanding physiological conditions) maintain a reliable electrical connection between the portion of the nervous system of interest and a computer. This reliable chronic connection has proved elusive and numerous biological and materials science strategies have been used. This article describes a substrate system for photolithographically defined neural interfaces capable of softening after implantation to reduce the mechanical mismatch across the biotic–abiotic interface. This substrate demonstrates this marked change in properties without a large uptake of conductive bodily fluids; thus providing a nonconductive substrate that insulates the electrode materials. The thiol-ene/acrylate substrates designed in this work adhere to gold, a common electrode and conductor material, and are compatible with standard photolithography processing with demonstrated feature sizes as small as 5 µm. This system is the first that provides the combination of each of the following: (1) responsiveness to physiological conditions to more closely match the modulus of neural tissue; (2) compatibility for the patterning of high-density electrodes by using the full capabilities of photolithography; and (3) low water uptake during implantation such that implant dimensions and insulating nature are unaffected by physiological conditions. The authors believe that this parameter space forms a potential basis from which to design chronic, invasive neural interface substrates to enable a new generation of clinically relevant biomedical devices.

The thiol-ene system was selected through a screening process of commercially available hydrophobic thiol and alkene monomers: chosen monomers are shown in Figure 1 and can be polymerized without solvent, with Tgs above 37°C and yield uniform polymer networks. It was hypothesized that a highly uniform polymer network with a glass transition just above physiological conditions could be sufficiently high modulus to allow insertion into soft tissues (based on the dimensions of predicate devices) and subsequently soften as a result of plasticization by a small amount of fluid. As such, this device could mimic current successful polyimide and Parylene-C devices allowing penetration into soft tissue and approach the modulus of silicone devices after implantation. Control over the Tg was established by adding a rigid diacrylate comonomer capable of increasing crosslink density of the network without significantly sacrificing network uniformity. Three compositions are evaluated here for their ability to meet these criteria. The materials described herein contain a stoichiometric ratio of thiol and ene while diacrylate concentration is varied. Figure 2(a) demonstrates the glassy behavior of each of these compositions at 37°C. The narrow width of the tanδ peak for each composition indicates a highly uniform polymer network is present for each sample. Another key advantage of thiol-ene polymers is their favorable interaction with gold, which adheres poorly to most materials. Poor adhesion necessitates use of a rigid adhesion layer or either titanium or chromium. Free thiols and thioether linkages found throughout the chosen polymer system adhere well to gold allowing processing without any adhesion layer. The use of transfer-by-polymerization, where the monomer solution is polymerized directly on the gold surface improves this adhesion further, as has been reported on distinct polymer systems.

The mechanical properties of each of the polymer samples are markedly changed through exposure to simulated physiological conditions. DMA and DSC, shown in Figure 3(b,c), 1 week after implantation on the surface of the brain indicate significant plasticization and subsequently a drop in modulus compared with the dry samples. The result was confirmed with samples implanted subcutaneously and immersed in PBS. Modeling of a hypothetical implanted soft neural interface suggested that a device with a modulus of 6 MPa could reduce tissue deformation around the device by two orders of magnitude.[4] Minimizing this tissue deformation has the potential to contribute to the goal of reducing the formation of the glial scar around the implant, which isolates the brain's electrical activity from the electrodes. Although this drop is significant, the resulting modulus of the polymer substrate is still three to four orders of magnitude stiffer than cortical tissue. This marked change in properties occurs despite low water uptake, less than 3% over 1 month for all three samples. Although minimizing the mechanical mismatch across the biotic–abiotic interface is important, other parameters such as device toughness are also of importance. This is specifically the case while interfacing with the peripheral nervous system, an environment in which related peripheral softening electronic devices may be subject to large deformations. Further development of the neural interface described and also toward future softening electronic devices will require significant iteration, but the authors believe the tunability of the presented polymer substrate will aid in continued optimization and the fabrication of highly functional devices.

A similar thiol-ene system was demonstrated compatible with osteocytes in the literature, but limited investigation has been presented on this system.[25] Evaluating the full biological response to this novel material is underway, but efforts were made to begin to understand this response. Embryonic neurons were cultured on the 31 mol % TCMDA sample. Cultures were possible only after coating with poly-d-lysine because of the hydrophobic nature of the polymer, but the viability of the cells indicates no inherent toxicity of the substrate to neurons.

A softening, intracortical electrode array has been fabricated using full photolithography. During photolithography, the maximum temperature reached was 85°C. In a more extreme example of this temperature stability of these novel polymers, devices were able to be soldered to provide electrical connection. This violates one of the presumed basic processing guidelines used in flexible electronics of not surpassing Tg during processing. Surpassing the glass transition during processing normally leads to warping and cracking of thin-film patterned layers, rendering the device unusable. These reported failures are attributed to the large coefficient of thermal expansion of polymers above their glass transition and to the release of cure or processing stresses.

The devices in this work do not fail when heated above Tg. This advancement is attributed to two key virtues of thiol-ene/acrylate as a substrate system. The first is dimensional stability over a large temperature range. This is imparted through the low cure stress nature of the thiol-ene reaction and the polymer's thermoset nature. The second is the polymerization of the substrate directly onto the carrier wafer using the transfer-by-polymerization process. This leads to excellent compatibility between the substrate and carrier over wide temperature ranges. Other key advantages of this system are chemical resistance to materials and solvents, including alcohols, water, acids, bases, and limited resistance to acetone and high optical transparency. Understanding the true limits of the substrate is the subject of ongoing investigation.

Impedance spectroscopy for the fabricated intracortical electrode array (electrode area of 177 µm2) is presented in Figure 5(c). Impedance of approximately 2 MΩ indicates consistency with previously published thin-film gold electrodes. Coating with platinum black was performed to decrease impedance, and as such noise, without increasing the geometric surface area of the device. Over the sub-chronic time period tested here, these softening intracortical electrode arrays were successfully used to record driven local field potential and spiking information on multiple electrodes implanted in the auditory cortex of the rat. These data are roughly comparable in quality to the results of other groups using commercially available planar silicon electrodes.[26] It should be noted however that the larger implant and distinct electrode positioning and surgical technique used compared with commercially available, silicon-based arrays makes direct comparison unreliable. Driven neural activity was recorded using electrode arrays coated in platinum black over 1 month postimplantation in three animals. Driven neural activity exhibited a latency of 10 ms, which is highly typical of activity in the primary auditory cortex. Nearly every channel demonstrated the ability to record driven LFPs, but the number of channels with driven spiking activity decreases, similar to the effect seen in other arrays. For channels with driven activity, no significant change in recorded LFP amplitude, spike SNR, noise floor or drivenness rating was observed. Importantly, this demonstrates that the electrode array continues to function in a stable manner despite the plasticization and softening of the substrate. Figure 7 provides examples of this activity for a representative channel, which was among the best at each time point. Strongly driven multiunit activity, which is tuned to certain frequencies, can be easily identified. Although this electrode design was sufficient to observe driven neural activity, the array design has not been optimized for reliable high SNR single-unit recording. The data presented here suggest that these electrodes can be used in the collection of useful neural activity for at least sub-chronic experiments. Direct comparisons of these electrodes to silicon and Parylene-C electrode arrays are underway and in detail immunohistochemical and recording longevity studies will be reported at later times. Although these are important metrics for chronic devices, the authors believe that the new design paradigm enabled by softening, processable, low swelling polymers, should stand on its own as a useful tool for the neural engineering community.

This work describes the development of a novel substrate platform for responsive neural interfaces. Using a facile “click” reaction substrates that are rigid for insertion and subsequently soften one to two orders of magnitude are developed. It should also be noted that this softening, although significant, still leaves the substrate between three and four orders of magnitude stiffer than cortical tissue. Polymers that soften even further are also an area of active exploration. These novel substrates are compatible with traditional lithography techniques generally reserved for unresponsive substrates. Using the stability of the described substrate and process, electrode materials that more effectively interact with the nervous system than gold for both stimulation and recording will also be investigated.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

A ternary thiol-ene/acrylate substrate developed for softening, neural interfaces has been synthesized and characterized. Control of the glass transition temperature and modulus in physiological conditions was achieved through varying diacrylate content from 0 to 31 mol %. Each sample softened considerably in physiological conditions. The shear modulus measured after explant from physiological conditions increased with diacrylate content from 4.7 to 23 MPa. Several neural interfaces were fabricated using a transfer-by-polymerization process. Excellent dimensional stability of the synthesized film enables processes that exceed the glass transition by more than 80°C. The limit of this temperature stability has not yet been found. The physiological response of these polymers, combined with excellent tolerance to photolithographic processes, makes this a promising system of substrates for neural interfaces that decrease the mechanical mismatch at the biotic–abiotic interface while offering increasingly advanced device and design capabilities.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

T. W. and W. V. have a significant financial interest in Syzygy Memory Plastics, Inc. This financial interest has been disclosed to UT Dallas, and a conflict of interest management plan is in place to manage the potential conflict of interest associated with this research program.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information
  • 1
    Zhong Y, Bellamkonda RV. Dexamethasone-coated neural probes elicit attenuated inflammatory response and neuronal loss compared to uncoated neural probes. Brain Res 2007;1148:1527.
  • 2
    Stice P, Muthuswamy J. Assessment of gliosis around moveable implants in the brain. J Neural Eng 2009;6:046004.
  • 3
    Polikov V, Tresco P, Reichert W. Response of brain tissue to chronically implanted neural electrodes. J Neurosci Methods 2005;148:118.
  • 4
    Subbaroyan J, Martin DC, Kipke DR. A finite-element model of the mechanical effects of implantable microelectrodes in the cerebral cortex. J Neural Eng 2005;2:103.
  • 5
    Thelin J, Jörntell H, Psouni E, Garwicz M, Schouenborg J, Danielsen N, Linsmeier CE. Implant size and fixation mode strongly influence tissue reactions in the CNS. PLoS One 2011;6:e16267.
  • 6
    Hassler C, Boretius T, Stieglitz T. Polymers for neural implants. J Polym Sci Part B: Polym Phys 2011;49:1833.
  • 7
    Seymour J, Langhals N, Anderson D, Kipke D. Novel multi-sided, microelectrode arrays for implantable neural applications. Biomed Microdevices 2011;13:441451.
  • 8
    Rousche PJ, Pellinen DS, Pivin DP Jr, Williams JC, Vetter RJ, Kirke DR. Flexible polyimide-based intracortical electrode arrays with bioactive capability. IEEE Trans Biomed Eng 2001;48:361371.
  • 9
    Gowrisanker S, Quevedo-Lopez MA, Alshareef HN, Gnade BE, Venugopal S, Krishna R, Kaftanoglu K, Allee D. A novel low temperature integration of hybrid CMOS devices on flexible substrates. Org Electron 2009;10:12171222.
  • 10
    Sekitani T, Kato Y, Iba S, Shinaoka H, Someya T, Sakurai T, Takagi S. Bending experiment on pentacene field-effect transistors on plastic films. Appl Phys Lett 2005;86:073511.
  • 11
    Kim D-H, Viventi J, Amsden JJ, Xiao J, Vigeland L, Kim Y-S, Blanco JA, Panilaitis B, Frechette ES, Contreras D. Dissolvable films of silk fibroin for ultrathin conformal bio-integrated electronics. Nat Mater 2010;9:511517.
  • 12
    Schuettler M, Stiess S, King BV, Suaning GJ. Fabrication of implantable microelectrode arrays by laser cutting of silicone rubber and platinum foil. J Neural Eng 2005;2:S121.
  • 13
    Achyuta AKH, Polikov VS, White AJ, Lewis HGP, Murthy SK. Biocompatibility assessment of insulating silicone polymer coatings using an in vitro glial scar assay. Macromol Biosci 2010;10:872880.
  • 14
    Lacour S, Benmerah S, Tarte E, FitzGerald J, Serra J, McMahon S, Fawcett J, Graudejus O, Yu Z, Morrison B. Flexible and stretchable micro-electrodes for in vitro and in vivo neural interfaces. Med Biol Eng Comput 2010;48:945954.
  • 15
    Harris J, Hess A, Rowan S, Weder C, Zorman C, Tyler D, Capadona J. In vivo deployment of mechanically adaptive nanocomposites for intracortical microelectrodes. J Neural Eng 2011;8:046010.
  • 16
    Ware T, Simon D, Arreaga-Salas DE, Reeder J, Rennaker R, Keefer EW, Voit W. Fabrication of responsive, softening neural interfaces. Adv Funct Mater. Forthcoming.
  • 17
    Hoyle CE, Lowe AB, Bowman CN. Thiol-click chemistry: A multifaceted toolbox for small molecule and polymer synthesis. Chem Soc Rev 2010;39:13551387.
  • 18
    Kolb HC, Finn M, Sharpless KB. Click chemistry: Diverse chemical function from a few good reactions. Angew Chem Int Ed 2001;40:20042021.
  • 19
    Senyurt AF, Wei H, Hoyle CE, Piland SG, Gould TE. Ternary thiol−ene/acrylate photopolymers: Effect of acrylate structure on mechanical properties. Macromolecules 2007;40:49014909.
  • 20
    Carioscia JA, Schneidewind L, O'Brien C, Ely R, Feeser C, Cramer N, Bowman CN. Thiol–norbornene materials: Approaches to develop high Tg thiol–ene polymers. J Polym Sci Part A: Polym Chem 2007;45:56865696.
  • 21
    Hess A, Dunning J, Harris J, Capadona JR, Shanmuganathan K, Rowan SJ, Weder C, Tyler D, Zorman C. A bio-inspired, chemo-responsive polymer nanocomposite for mechanically dynamic microsystems. In: Solid-State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, June 21-25, 2009.
  • 22
    Harris JP, Hess AE, Rowan SJ, Weder C, Zorman CA, Tyler DJ, Capadona JR. In vivo deployment of mechanically adaptive nanocomposites for intracortical microelectrodes. J Neural Eng 2011;8:046010.
  • 23
    Najafi K, Hetke JF. Strength characterization of silicon microprobes in neurophysiological tissues. IEEE Trans Biomed Eng 1990;37:474481.
  • 24
    Rennaker RL, Street S, Ruyle AM, Sloan AM. A comparison of chronic multi-channel cortical implantation techniques: Manual versus mechanical insertion. J Neurosci Methods 2005;142:169176.
  • 25
    Nordberg A, Antoni P, Montaez MI, Hult A, Von Holst H, Malkoch M. Highly adhesive phenolic compounds as interfacial primers for bone fracture fixations. ACS Appl Mater Interfaces 2010;2:654657
  • 26
    Kozai TDY, Langhals NB, Patel PR, Deng X, Zhang H, Smith KL, Lahann J, Kotov NA, Kipke DR. Ultrasmall implantable composite microelectrodes with bioactive surfaces for chronic neural interfaces. Nat Mater 2012;11:10651073.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmb32946-sup-0001-suppinfo.doc436KFigure S1: DSC of each of the tested samples in the dry condition, after 1 week swollen and 1 week implanted on the surface of the brain. Both in vitro and in vivo methods result in similar plasticization of the polymer networks.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.