Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Apr 21.
Published in final edited form as: Annu Int Conf IEEE Eng Med Biol Soc. 2019 Jul;2019:3766–3769. doi: 10.1109/EMBC.2019.8856692

In Vitro Reactive-Accelerated-Aging Assessment of Anisotropic Conductive Adhesive and Back-End Packaging for Electronic Neural Interfaces

Cary A Kuliasha 1, Jack W Judy 1
PMCID: PMC8059779  NIHMSID: NIHMS1680629  PMID: 31946694

Abstract

Bioelectronic neural interfaces can fail in vivo due to water penetration and corrosion of the packaging technology used to protect sensitive portions of the device. Although anisotropic conductive adhesive (ACA) is gaining popularity in the neural interface community to connect fabricated electrode arrays with back-end packages, the durability of ACA in chronic implants is largely unknown. We have designed a platform that uses an aggressive reactive-accelerated aging (RAA) environment to rapidly assess the ability of ACA and silicone-rubber encapsulation to maintain electrical integrity in vitro. All RAA experiments were performed at 77°C, for 24 days, and with 10 to 20 mM H2O2, which approximates a 1 year implantation. Results from these experiments showed that ACA rapidly fails (i.e., 2 to 4 days RAA) due to water absorption through the silicone encapsulant. Electrical impedance spectroscopy (EIS) confirmed water penetration through the package and the resulting corrosion of the sensitive metallic components.

I. INTRODUCTION

Implantable bioelectronic neural interfaces are currently being developed to address a host of neural diseases and conditions by sensing (recording) and activating (stimulating) the body’s nervous system [1]. A main challenge in the development of next-generation neural interfaces is engineering devices that provide robust, stable, and chronic performance in vivo for the full duration of the implant. Excellent progress has been made in electrode design and material selection with devices showing stable performance in vivo for months [2], [3]; however, longer studies over years typically reveal decreases in performance that indicate the onset of failure of the device (i.e., abiotic) or its interface with the body (i.e., biotic) [4], [5]. The factors behind these failures are potentially numerous, complicated, and often interrelated making it difficult to investigate and design around failure. Determining biotic failure mechanisms often requires long-term in vivo studies; however, many abiotic failures can be investigated using rapid in vitro accelerated aging (AA) tests that consist of soaking devices in simulated body fluids (e.g., phosphate buffered saline: PBS) at elevated temperatures (>37°C) to simulate chemical degradation and accelerate device failure [6].

Abiotic device failure typically involves material delamination, degradation, or corrosion due to water penetration and chemical attack, and it can be further separated into failure at the electrode site, failure of the bulk device, or failure of the back-end packaging. Since many non-wireless neural interface technologies are not long enough to enable percutaneous electrical connections, back-end packaging is often needed to connect the interface to longer leads. Many back-end connection technologies are used including soldering and “microflex” ball-bonding [7] and newer methods such as microspring interconnects [8] and anisotropic conductive adhesive (ACA) / film (ACF) [9]. ACA was first pioneered by the consumer electronics industry and uses conductive particles suspended in a dielectric adhesive to electrically connect components. One advantage of ACA over other techniques is the capability of forming high-density electrical connections. As bioelectronic devices are miniaturized and channel counts increase, this capability is becoming more valuable. However, ACA’s durability in long-term (chronic) implants is largely unknown.

Using any of the available connection technologies, the sensitive back-end package must be designed to mitigate the damage of moisture intrusion and corrosion that can result in loss of device performance and failure. Hermetic ceramic/metal packages provide the best longevity [10]; however, they have limited applications, especially in a research environment, due to their size and complexity. Non-hermetic polymeric materials (e.g., silicone rubber, epoxy, polyimide, or parylene-C) are often used to encapsulate back-end packaging [11]; however, due to the non-hermiticity of the materials, additional care must be taken to mitigate the impact of moisture penetration. AA has been widely used to assess packaging technology [12], [13]. Unfortunately, AA tests fail to adequately replicate the in vivo environment due to a decreased concentration of reactive oxygen species (ROS) commonly present during chronic inflammation [14], and this can result in not identifying possible failures during in vitro AA testing. In response to this drawback, researchers at the FDA have developed a reactive-accelerated aging (RAA) platform that incorporates H2O2 as a source of ROS that has been shown to replicate some in vivo failure modes encountered in a variety of cortical devices [15].

Herein we report on the use of an aggressive RAA platform to validate the ability of ACA to maintain electrical connection within a package encapsulated in silicone rubber. Thin-film polyimide test structures with shorted contact pairs were fabricated and connected to printed circuit boards (PCBs) using ACA, and all exposed metal sites and structures were encapsulated in silicone. Electrical impedance spectroscopy (EIS) and DC resistance measurements were used to longitudinally assess the durability of the silicone encapsulant and the ability of ACA to maintain electrical connections with the PCB.

II. Materials And Methods

A. Device Fabrication, Bonding, and Packaging

ACA test-structures were fabricated in a class 100-1000 clean room from laminated thin film polyimide (U-Varnish S, UBE Ind.) with hydrogenated amorphous silicon carbide (a-SiC:H) and 400 nm thick sputtered Ti-Pt-Au-Pt-Au conductive traces/contact sites according to Kuliasha et al. [16] (Fig. 1). Test-structures contained 20 individual contact pads (10 pairs) with 6 adjacent pairs shorted via metal traces (12 pads) and 8 isolated pads. Custom FR4 PCBs (Imaganeering, Inc.) with dimensions of 7×5×0.25 mm3 containing 20 isolated 200 μm diameter, 650 μm center-to-center pitch contact pads connected to vias were electroless nickel and immersion gold (ENIG) plated so the metal sites were positively “bumped” above the surrounding solder mask by 15 to 20 μm, determined by optical profilometry. Bonding sites need to be positively projected from the surrounding surface for ACA to be effective. The PCB was also designed with a continuity structure consisting of 2 shorted vias that do not connect to an adjacent contact pad as an internal control.

Figure 1.

Figure 1.

Open in a new tab

Device design, ACA bonding, and validation. (A1) Schematic cross-section of both the flexible test-structure (top) and PCB (bottom) highlighting the electrical contact pads and the application of ACA. Inset images show top down views of the (A2) flexible test-structure and the (A3) PCB bonding site and (A3) optical profilometry of the ENIG plated PCB indicating the “bumped” nature of the contact sites. (B1) Schematic cross-section of an ACA connected test-strucuture highlighting the different electrial pathways possible including the low resistance “Rc” path along the conductive traces and the high resistance “Rd” path through the dielectric material. Inset images show (B2) nanoCT cross-section of two PCB contact pads (red) connected to the flexible test-structure via ACA particles and (B3) top down image of 2 contact pad pairs of the ACA bonded test-structure.

ACA liquid precursor (124-20, Creative Materials, Inc.) was applied to ENIG plated PCB contact pads using a syringe needle and B-staged at 80 °C until tacky. Next, the contact pads of the thin-film test-structure were optically aligned and bonded to the PCB contact pads using a flip-chip placement system (Model 950, Semiconductor Equipment Corp.). A silicone gasket was placed over the bond area, to ensure even application of pressure, and the assembly was cured at 150 °C at 30 psi. NanoCT was used to image the bonding joints (N = 1, Phoenix V-Tome-X M240, General Electric). Insulated 30 AWG stainless steel wires (R30B, Jonard Tools) were then soldered to PCB vias of bonded devices, and the ACA-PCB-wire bundle was encapsulated in medical grade silicone (Bluesil RTV 3040, Bluestar Silicone USA Corp.) using a 3-D printed mold.

B. Characterization and in vitro RAA

The distal end of the soldered stainless steel wires were connected to a custom PCB and TDT ZIF-Clip® (Tucker Davis Technology) to enable electrical connection with the encapsulated PCB. Devices were characterized by EIS (PGSTAT302N, Metrohm Autolab) and channel-to-channel DC resistance (B2985, Keysight). EIS spectra were obtained using a 3-electrode setup with an ACA test-structure via/wire working electrode (n = 21, per device), Pt-wire counter electrode, and a 3 M KCl Ag/AgCl reference electrode at 100 mA from 100 KHz to 1 Hz in phosphate-buffered saline (PBS) electrolyte (Sigma Aldrich). No DC offset with respect to the open circuit potential was applied. Channel-to-channel DC resistance was measured between every contact pad (n = 441 measurements per device) using a custom made multiplexer at 100 mV. RAA soak tests were performed in PBS with 10 to 20 mM H2O2 at 77 °C to assess in vitro device durability [15], [16]. The H2O2 degrades rapidly at 77°C necessitating replenishment with concentrated H2O2 to maintain a constant 10 to 20 mM concentration that was verified using a titanium (IV) oxysulfate colorimetric assay. RAA experiments were performed in 500 mL 4-neck round bottom flasks under reflux with both the EIS and Ch-to-Ch resistance measurements performed longitudinal in the RAA solution. Shorted channel pairs had a Ch-to-Ch resistance of ~20 to 30 Ω, and a pair was considered “failed” when the Ch-to-Ch resistance increased to >106 Ω consistent with isolated Ch pairs. All devices were tested in RAA for 24 days, which is estimated to be equivalent to a 12-month implantation (ASTM F1980-16). In addition, EIS was performed post-RAA at room temperature in PBS, as before.

III. Results and Discussion

A. Device Fabrication and Packaging

The bonding of thin-film test-structures to PCBs was successful with 96.7% of shorted channel pairs effectively connected to the PCB with a corresponding dc resistance (Rc) of 23.1 ± 4.7 Ω measured from 5 devices (N = 5), with 30 shorted channel pairs (n = 6, per device). Optical inspection revealed that the single failed connection was due to an absence of a conductive particle at the contact junction between the PCB and test-structure. The remaining isolated channels (n = 4, per device) maintained a high resistance >106 Ω between all other channels indicating the ACA did not cause any undesirable shorting between adjacent pads that were not connected via the thin-film test-structure. Pre-RAA EIS analysis revealed high-impedance values across the entire frequency spectrum with an average of 29 MΩ at 1 KHz and corresponding capacitive –phase of 88°. These results are expected because the silicone fully encapsulated the PCB-ACA test-structure with no metal sites exposed to the PBS electrolyte.

B. Longitudinal RAA

RAA experiments are typically performed at 87°C [15], [16]; however, the temperature of all accelerated aging tests must be below the onset of any thermally induced changes that devices would not experience during implantation. Differential scanning calorimetry (DSC) analysis of the ACA used in this study determined that the glass transition temperature (Tg) was between 90 to 100°C, so we elected to lower the bath temperature to 77°C to avoid any possible Tg induced failure of the ACA connected pairs. Longitudinal analysis of devices revealed that the ACA rapidly failed during RAA with ~59% remaining functional channel pairs after only 2 days and only ~10% after 4 days corresponding to only ~1-2 months in vivo (ASTM F1980-16) (Fig. 2). We believe the ACA failed due to rapid water diffusion through the silicone package and into the ACA epoxy resulting in swelling/expansion that broke the physical contacts made between the conductive ACA particles and the PCB/thin-film contact pads.

Figure 2.

Figure 2.

Open in a new tab

% Functionality of ACA contact junctions during RAA. Data points were determined by longitudinally measuring device Ch-to-Ch DC resistance of contact pairs shorted via ACA particles. Device failure was demonstrated by a large increase in Ch-to-Ch resistanct as the conductive ACA particles lost contact between the PCB and flexible test-structure. Data points represent the arithmetic average ± standard deviation (N = 5, n = 6)

Longitudinal EIS analysis showed a steady decrease in both the impedance magnitude and –phase angle over time with the most prevalent shifts occurring at lower frequencies (Fig. 3). It is believed that the measured decreases are due to penetration of the silicone packaging with moisture and ROS that lowered the dielectric properties of the encapsulation and corroded the solder. After 24 days in the RAA solution, devices showed an average –phase angle < 20° at 1 KHz indicating a more resistive pathway has been exposed to the PBS electrolyte. Furthermore, the Nyquist plots for the majority of working electrodes had a characteristic loop with two time constants indicative of metal site corrosion. Optical inspection confirmed that the metal sites were corroded; however, the silicone encapsulation did not display any cracks or other bulk pathways for water ingress so it can be reasonably asserted that the EIS behavior and corrosion was due to water diffusion to the metal sites and not bulk failure of the package.

Figure 3.

Figure 3.

Open in a new tab

EIS summary of RAA-tested devices highlighting the impact to the back-end packaging. (A-B) Average impedance magnitude and –phase angle of devices before and after 24-day RAA testing. All EIS spectra were collected in PBS at room temperature. RAA-induced changes in the EIS behavior are due to water penetration and corrosion of the package. Data points represent the arithmetic average ± standard deviation (N = 5, n = 21). (C) Nyquist plot post 24-day RAA with a characteristic inductive loop due to metallic corrosion of the solder/metal sites (observed visually). (D) Longitudinal EIS at three separate frequencies during RAA showing a steady decrease as moisture diffused into the bulk silicone. Data points represent the arithmetic average ± standard deviation (N = 5, n = 21).

IV. Conclusion and Future Work

All polymeric packaging material is susceptible to some degree of moisture penetration during implantation. We have described the use of reactive-accelerated aging to assess the functionality of anisotropic conductive adhesive (ACA) and silicone-rubber encapsulant to protect an implantable electronic package. The rapid failure of the ACA contacts suggest that the ACA may not be reliable for any chronic in vivo implants that do not use a hermetic package. Furthermore, even though silicone is widely used as an encapsulation material by many in the bioelectronic community, we have shown that it is prone to rapid diffusion of moisture during RAA. This fact should be considered as a possible failure mechanism by all in the community when performing accelerated-aging experiments or electrical characterization. With that said, more testing is needed to validate additional ACA manufacturers and adhesive materials/compositions and to determine if the rapid failure shown during RAA experimentation is replicated under non-reactive accelerated aging (AA) in PBS with no H2O2 and exactly how well the RAA platform replicates and thus predicts in vivo failure modes.

Acknowledgment

The microfabrication reported here was enabled by the facilities and staff of the University of Florida Research Services Centers.

This work was sponsored by the Defense Advanced Research Projects Agency (DARPA) Biological Technology Office (BTO) Electrical Prescriptions (ElectRx) program under the auspices of Dr. Eric Van Gieson through the DARPA contracts Management Office, Pacific Cooperative Agreement: No. HR0011-15-2-0030.

References

  • [1].Vitale F and Litt B, “Bioelectronics: the promise of leveraging the body’s circuitry to treat disease,” Bioelectron. Med, vol. 1, no. 1, pp. 3–7, 2017. [Google Scholar]
  • [2].PR P et al. , “Chronic in vivo stability assessment of carbon fiber microelectrode arrays,” J. Neural Eng, vol. 13, p. 066002, 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Vetter RJ, Williams JC, Hetke JF, Nunamaker EA, and Kipke DR, “Chronic Neural Recording Using Silicon-Substrate Microelectrode Arrays Implanted in Cerebral Cortex,” IEEE Trans. Biomed. Eng, vol. 51, no. 6, pp. 896–904, June. 2004. [DOI] [PubMed] [Google Scholar]
  • [4].Simeral JD, Kim SP, Black MJ, Donoghue JP, and Hochberg LR, “Neural control of cursor trajectory and click by a human with tetraplegia 1000 days after implant of an intracortical microelectrode array,” J. Neural Eng, vol. 8, no. 2, 2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Chestek CA et al. , “Long-term stability of neural prosthetic control signals from silicon cortical arrays in rhesus macaque motor cortex,” J. Neural Eng, vol. 8, no. 4, 2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Hukins DWL, Mahomed A, and Kukureka SN, “Accelerated aging for testing polymeric biomaterials and medical devices,” Med. Eng. Phys, vol. 30, pp. 1270–1274, 2008. [DOI] [PubMed] [Google Scholar]
  • [7].Stieglitz T, Beutel H, and Meyer JU, “‘Microflex’ - A new assembling technique for interconnects,” J. Intell. Mater. Syst. Struct, vol. 11, no. 6, pp. 417–425, 2000. [Google Scholar]
  • [8].Khan S, Ordonez JS, and Stieglitz T, “Reliability of spring interconnects for high channel-count polyimide electrode arrays,” J. Micromechanics Microengineering, vol. 28, no. 5, p. aaaf2c, 2018. [Google Scholar]
  • [9].Viventi J et al. , “Flexible, foldable, actively multiplexed, high-density electrode array for mapping brain activity in vivo,” Nat. Neurosci, vol. 14, no. 12, pp. 1599–1605, 2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Bisoni L et al. , “Investigation on the hermeticity of an implantable package with 32 feedthroughs for neural prosthetic applications,” Proc. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. EMBS, vol. 2016–October, pp. 1967–1970, 2016. [DOI] [PubMed] [Google Scholar]
  • [11].Qin Y, Howlader MMR, Deen MJ, Haddara YM, and Selvaganapathy PR, “Polymer integration for packaging of implantable sensors,” Sensors Actuators, B Chem, vol. 202, pp. 758–778, 2014. [Google Scholar]
  • [12].Hassler C, Von Metzen RP, Ruther P, and Stieglitz T, “Characterization of parylene C as an encapsulation material for implanted neural prostheses,” J. Biomed. Mater. Res. - Part B Appl. Biomater, vol. 93, no. 1, pp. 266–274, 2010. [DOI] [PubMed] [Google Scholar]
  • [13].Cvancara P, Lauser S, Rudmann L, and Stieglitz T, “Investigations on different epoxies for electrical insulation of microflex structures,” Proc. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. EMBS, vol. 2016-October, no. Dlc, pp. 1963–1966, 2016. [DOI] [PubMed] [Google Scholar]
  • [14].Meyers CH, “Pressure-volume-temperature data for oxygen,” J. Res. Natl. Bur. Stand , vol. 40, pp. 457–466, 1948. [DOI] [PubMed] [Google Scholar]
  • [15].Takmakov P, Ruda K, Phillips KS, Isayeva IS, Krauthamer V, and Welle CG, “Rapid evaluation of the durability of cortical neural implants using accelerated aging with reactive oxygen species,” J. Neural Eng, vol. 12, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Kuliasha CA and Judy JW, “In Vitro Reactive-Accelerated-Aging (RAA) Assessment of Tissue-Engineered Electronic Nerve Interfaces (TEENI),” in 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC’18), 2018, pp. 1–4. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES