Designing For Safety And Efficacy 331 Leadless Injectable Package

In most electronic systems, the largest assembly costs and most likely sources of failure tend to be in the physical connections between components. This is particularly so when those connections must function in a pool of warm salt water that is constantly moving.6 These considerations led us to design a system based on individual modules with the following key properties:

• Small enough to be located directly at the site where stimulation and/or sensing is required, with self-contained electrodes;

• Form factor suitable for injection through a hypodermic needle;

• Powered by inductive coupling of energy from an externally generated magnetic field;

• Capable of receiving and transmitting data by modulated radiofrequency (RF) telemetry;

• Digitally encoded device address and stimulus parameters.

The color figure shows the package for the BION1 stimulation module, which consists of a cylindrical glass capsule with two rigidly mounted electrodes on its ends. The electronic subassembly in the capsule is dominated physically by the antenna coil, which consists of about 200 turns of 1 mil insulated copper wire wound over a cylindrical ferrite form to maximize the capture of energy from the magnetic field. It is wound so as to be self-resonant at the 2 MHz carrier frequency generated by the external primary coil, which uses Class E amplification and synchronous modulation to achieve high magnetic field strength with high electrical efficiency.7 The cylindrical ferrite is actually a sandwich of two hemicylindrical ferrites glued to a custom integrated circuit (IC) mounted on an alumina ceramic printed circuit board (PCB), as shown in the center portion of Figure 3.1. This arrangement maximizes the surface area available for the IC and provides a stable platform for its wire bonds plus a discrete diode chip and soldered termination of the coil windings.

The PCBs are built in wafers of 16, using two-sided, gold-plated through-hole metallization on an alumina wafer. Before metallization, the wafer is laser-drilled in a pattern that becomes the edge-conductive detents at the ends of the PCBs after the wafer is diced. One end of the wafer makes electromechanical contact with a platinum-iridium washer welded to the tantalum stem of the tantalum electrode. The other end makes contact to a gold-plated elgiloy spring that provides an electromechanical connection to the iridium electrode at the other end, via a hollow tantalum feedthrough (Figure 3.1, right). The electronics are sealed into the capsule by sliding them into an open-ended capsule, which is formed onto the Ta electrode (Figure 3.1, left), and welding the tubular feedthrough (Figure 3.1, right) to the glass capillary walls of the capsule.

3.3.2 Achieving and Demonstrating Hermeticity

The most important requirement for the package is to protect the electronic circuitry from the deleterious effects of water. Sophisticated electronic circuitry such as the tuned RF power and data receiver and the digitally controlled stimulus pulse generator are particularly vulnerable to condensed moisture. Failure mechanisms include detuning of the self-resonant receiver coil, poisoning of semiconductor junctions from solubilized contaminants, and shorting of circuitry as a result of corrosion and dendrite formation between conductors with voltage differences.6 At body temperature, water vapor will condense when it reaches 6% concentration (dew point) in the free space of the capsule, which is about 50% of the total inside volume of the glass capsule (0.01 cc).

FIGURE 3.1 The fabrication sequence of the BION1 implant consists of three major subassemblies (top, left to right): capsule subassembly built onto the Ta electrode, electronic subassembly stacked onto the ceramic micro-printed circuit board (^PCB), and feedthrough subassembly built onto a Ta tube. After the subassemblies are brought together, the final fabrication process includes hermeticity testing and vacuum bake-out through the still-open Ta tube followed by final closure of the Ta tube and attachment of the Ir electrode. The last process is anodization of the Ta electrode, which is accomplished by probing the Ir electrode and passing current backward through the output circuit with the Ta electrode dipped in 1% phosphoric acid. Numbers refer to ten critical joints accomplished with four different technologies as shown in the key at the lower left. The glass-to-metal (3, 5) and glass-to-glass (4, 8) seals are made by melting the glass in an infrared laser beam that provides well-controlled and very localized heating. The metal-to-metal joints (1-2, 9-10) are made in an argon plasma needle arc, which provides the intense but localized heat needed to melt Ta (2700°C) and Ir (2400°C). A resistance weld is used to attach the spring to the Ta tube without loss of temper (6) and the fine copper coil windings are terminated by microsoldering to the ^PCB (7).

FIGURE 3.1 The fabrication sequence of the BION1 implant consists of three major subassemblies (top, left to right): capsule subassembly built onto the Ta electrode, electronic subassembly stacked onto the ceramic micro-printed circuit board (^PCB), and feedthrough subassembly built onto a Ta tube. After the subassemblies are brought together, the final fabrication process includes hermeticity testing and vacuum bake-out through the still-open Ta tube followed by final closure of the Ta tube and attachment of the Ir electrode. The last process is anodization of the Ta electrode, which is accomplished by probing the Ir electrode and passing current backward through the output circuit with the Ta electrode dipped in 1% phosphoric acid. Numbers refer to ten critical joints accomplished with four different technologies as shown in the key at the lower left. The glass-to-metal (3, 5) and glass-to-glass (4, 8) seals are made by melting the glass in an infrared laser beam that provides well-controlled and very localized heating. The metal-to-metal joints (1-2, 9-10) are made in an argon plasma needle arc, which provides the intense but localized heat needed to melt Ta (2700°C) and Ir (2400°C). A resistance weld is used to attach the spring to the Ta tube without loss of temper (6) and the fine copper coil windings are terminated by microsoldering to the ^PCB (7).

There are well-developed sealing techniques and test methods for implantable electronic devices such as pacemakers and cochlear implants, but these devices are much larger than BIONs (see nomogram in Figure 3.2). The various test methods differ in their sensitivity, which is expressed as a gas leakage rate in cc/s for one atmosphere of pressure gradient. The most sensitive method is the "helium sniffer" in which a high vacuum is applied to one side of the seal to be tested and helium gas is squirted over the other side of the seal. Trace helium gas leaking through the seals is detected by a specially tuned mass spectrometer in the vacuum line. The practical limit of this test in the laboratory is about 2 x 10-11 cc atm/s; the sensitivity level for production testing is usually derated to 1 x 10-9 cc atm/s. The equivalent leak rate for water vapor would be about 1/2 that of helium (it is related to the square root of the molecular weight of the gas). Thus it is possible to compute the length of time it would take for water vapor to reach the dew point in a capsule of a given volume if there were a leak in one of the seals at a rate just below the detection limit of the test method. For the tiny BION, hermeticity testing at even the laboratory limit could not guarantee a functional life of more than one year.

The problem of guaranteeing sufficient moisture resistance for each manufactured BION has been addressed by incorporating a "getter" that absorbs water vapor. It consists of anhydrous magnesium sulfate powder molded into a small cylinder of silicone elastomer that fits over the spring inside the BION capsule (see Chapter 3, Color Figure 1* and Figure 3.1). Magnesium sulfate is one of a class of salts that can be used as desiccants because its crystal structure tends to bind water molecules in its interstices. Fully hydrated magnesium sulfate holds up to 8 moles of water per mole of salt. Even the small amount of magnesium sulfate in the BION getter (10% of the package volume containing 2.3 ^g = 2 x 10-5 moles) can absorb a very large volume of water vapor (>1 cc). In effect, the internal volume of the package for water vapor is about 100 times larger than the volume of the BION package, resulting in a projected 30-yr survival time even if the package is leaking at just below the sensitivity limit for production testing (10-9 cc atm/s).

Extensive qualification testing of the completed BIONs and of individual components of the seals indicates that the seals are, in fact, hermetic to a degree far beyond the sensitivity of laboratory leak-testing. BIONs built without a water getter have been soaked in saline at 160 atm pressure for over four months without evidence of detectable condensation when cooled. Other BIONs without getters have been operated for over a year with continuous output pulsing in saline while cycling between 37°C (3 h) and 77°C (9 h). In another test, BION capsules were fitted with a bare PCB in place of the usual electronic subassembly, resulting in an open circuit between the adjacent traces of metal on the PCB connected to the output electrodes. High impedance spectrograms at room temperature were compared at various intervals during one year of soaking in saline at 85°C, with no detectable change (this method is sensitive to a layer of condensed water vapor only 10 molecules thick).

The hermeticity of the glass-to-metal seals in the BION probably depends on chemical bonding between the borosilicate glass (Kimbel N51A) and the native oxide on the tantalum metal of the Ta electrode stem and the Ta tubular feedthrough.

* Chapter 3, Color Figure 1 follows page 112. © 2001 by CRC Press LLC

FIGURE 3.2 Nomogram to determine the minimal guaranteed life time (s) of an implanted device from its volume (cc) divided by the minimal leak rate (cc atm/s) that can be detected during testing of its hermetic seals, as shown by log scales. Volume: Product names of various commercial implants are positioned over scale to indicate total package volume; black solid arrows indicate the volume of water vapor required to reach dew point at 37°C (6%) assuming 50% free volume. Leak Rate: Assumes 1 atm pressure head for test and use conditions. The equivalent sensitivity ranges of various conventional test methods are shown in the boxes above and parallel to the leak rate scale. Life Time: Straight lines from a given volume through a leak-rate detection limit intersect the life time scale to show the minimal reliable working life of the device; log time scale calibrated in calendar time (above) and seconds (below).

FIGURE 3.2 Nomogram to determine the minimal guaranteed life time (s) of an implanted device from its volume (cc) divided by the minimal leak rate (cc atm/s) that can be detected during testing of its hermetic seals, as shown by log scales. Volume: Product names of various commercial implants are positioned over scale to indicate total package volume; black solid arrows indicate the volume of water vapor required to reach dew point at 37°C (6%) assuming 50% free volume. Leak Rate: Assumes 1 atm pressure head for test and use conditions. The equivalent sensitivity ranges of various conventional test methods are shown in the boxes above and parallel to the leak rate scale. Life Time: Straight lines from a given volume through a leak-rate detection limit intersect the life time scale to show the minimal reliable working life of the device; log time scale calibrated in calendar time (above) and seconds (below).

© 2001 by CRC Press LLC

In order to prevent excess oxidation, the glass-to-metal seals are made by melting the glass using a CO2 infrared laser under an argon gas curtain (Model F48-2-28W, Synrad). The hermeticity of these seals can be lost if excess residual oxygen or longitudinal grooves are left in the Ta metal from the drawing process. Earlier versions of the BION package used a tubular feedthrough of 90%Pt-10%Ir, which also produced seals that tested hermetic but tended to fail catastrophically during prolonged soaking and temperature cycling in saline because of differences in the coefficient of thermal expansion between the glass (5.5 x 10-6/°C) and PtIr (8.7 x 10-6/°C). Excess residual stress in the walls of the sealed glass capsules was measured using the photoelastic effect on the rotation of polarized light (Model 33 Polarimeter, Polarm-etrics, Inc., Hillsborough, NH). By contrast, the coefficient of Ta is 6.5 x 10-6/°C. Even low amounts of residual stress, however, can result in catastrophic package failure if there are "stress-risers" in the package, such as scratches or irregularly melted regions of the glass capillary. Fortunately, these failures can be greatly accelerated by soaking in saline pressurized to 160 atm. At this pressure, the dipolar water molecules insinuate themselves rapidly into the glass defect, resulting in a propagating crack that leads to catastrophic failure of the package almost immediately.

Each seal in the BION package is tested for hermeticity and robustness at several points in the fabrication process. Each open capsule subassembly is mounted in an O-ring fixture to perform a He sniffer test of the bead-to-stem and capillary-to-bead seals. Each closed capsule is tested for hermeticity of both glass-to-glass and glass-to-Ta seals by fixturing on the open Ta tube in the He sniffer (Model ASM 110 Turbo, Alcatel). Any residual moisture in the capsule is removed by vacuum bake-out and back-filling with an inert gas mix containing 25% He. The final seal is made by melting the Ta tube closed in a plasma needle arc welder (Ultima 150, Thermal Dynamics, West Lebanon, NH) (Figure 3.1). Slow leaks in the final seal are screened by trying to detect this captured He escaping from the Ir electrode end of the package. Finally, all finished BIONs are bombed in saline at 160 atm pressure for 24 h and examined for visible cracks or condensed moisture before being cleaned and sterilized.

The mechanical integrity of the BION package and seals has also been tested in several destructive qualification tests. The glass capsule is most susceptible to breakage by three-point bending over its long axis; it will fail catastrophically if forces greater than 2 kg are applied at rigid fixation points. The maximal stresses that the BION can experience in situ are limited, however, by the structural strength of the soft muscle tissue in which it is intended to be implanted (about 30 N/cm2 even in tetanically contracting fast-twitch muscle8). Using a combination of static and dynamic load testing and instrumentation of the capsules with strain gauges, we have demonstrated that the capsule has a safety margin of at least 4:1 once it is surrounded by at least 1 cm thick muscle. We have also been unable to induce failures of the hermetic seals when loaded axially to 1.3 kg force, which is over four times the maximal insertion force. BIONs have also been subjected to multiple episodes of random severe handling such as might occur prior to implantation, including dropping bare devices onto a steel instrument tray from a height of 20 cm and dropping packaged devices onto a tile floor from a height of 1 m, all without failure. The same devices have then survived five cycles of autoclaving and freezing without loss of function or hermeticity.

Was this article helpful?

0 0

Post a comment