Appendix · the complete record · 56 grants
The Catalogue
Every granted US patent in the corpus, in order of first filing, grouped by line of work. Open any entry to light its case: the hero drawing, a plain-language account, and the way to the exhibit page and the official record.
flagship major notable supporting
Hermetic Packaging & Feedthroughs · 20 grants
An implantable device is a promise: nothing leaks in, nothing leaks out, for a decade or more, in warm salt water. This cluster is thirty years of keeping that promise — from the 1993 housing that let a pacemaker be assembled like a microchip, to laser-bonded feedthroughs that pass signals through sapphire walls with nanometer-scale seals, to testing structures that prove the seal is real (US7902851). If the corpus has a thesis statement, it is here.
1993 US5535097A Implantable medical device including a first enclosure portion having a feedthrough in a second interior surface Origin patent (1993, first-listed); 100 direct citations — second only to the accelerometer among direct-citation counts.
Filed in 1993, this is where the corpus begins. Pacemakers were then built as welded clam-shells crowded with separate positioning parts; this design reshapes the can's interior so battery, circuit, and feedthrough each drop into a ready-made seat before one lid seals it shut — fewer parts, fewer steps, and a rounded contour kinder to the tissue it lives in. A hundred later patents cite it: the seed of thirty years of packaging work.
2005 US7288847B2 Assembly including a circuit and an encapsulation frame, and method of making the same
Encapsulant faces a paradox: it must flow freely enough to infiltrate a circuit's every crevice, yet not flood the board — and molded walls leak at post holes and cost too much. Ruben's first-listed answer is geometry: a flat frame, thermally matched to the substrate, whose through-hole corrals the encapsulant around the components, with extra openings that deliberately keep contacts and ID codes exposed for later assembly steps. A dam instead of a mold — cited 13 times, including by Shindengen.
2007 US7682878B2 Encapsulation circuitry on a substrate
Protecting a finished circuit means covering it in encapsulant that must be runny enough to seep around every component yet thick enough not to pour off the board. This continuation extends the encapsulation-frame idea: a flat frame, matched to the board's thermal expansion, whose through-hole wall dams the liquid right where it is needed. Even its extra holes are practical, leaving contacts and ID codes exposed for the next step on the production line.
2009 US7902851B2 Hermeticity testing
Moisture is the quiet killer of implanted electronics, and the seam between two bonded chips is exactly where a micro-package is most vulnerable — yet conventional tests can't see into that buried interface. So the test is built into the seal itself: tiny chain circuits on both chips mate at bonding into rings around each sealed region, and moisture creeping in breaks or shorts a chain at the perimeter before the electronics are ever at risk. Not making the seal — proving it, electrically, from the inside.
2013 US10464836B2 Hermetic conductive feedthroughs for a semiconductor wafer
How do you carry a signal out through a glass wall without giving body fluid a path back in? Here the conductor is entombed in the wafer itself — its outer face never reaches the surface directly, but takes a deliberate dog-leg through an embedded trace to a recessed cavity, so fluid can never follow the interface back inside. Filed in 2013, it founded the glass-wafer feedthrough line for making implants at wafer scale.
2015 US9832867B2 Embedded metallic structures in glass
Implant packages have to keep shrinking without giving up their airtight seal, and wiring that lives only on the surface wastes the walls. Here two glass wafers are fused with filled vias, conductive traces, and an insulating layer between them, turning the glass itself into three-dimensional wiring — even an antenna can be etched straight into the wafer, so the wall becomes a working radio component instead of mere enclosure. The interconnect industry noticed: four of the five visible follow-on citations come from connector maker Samtec.
2015 US9865533B2 Feedthrough assemblies (parent of the laser-bond feedthrough line) Founds the laser-bond feedthrough line (priority 2014-12-24).
Every implantable device faces the same paradox: its electronics must be sealed away from body fluid absolutely, yet wires have to pass through the wall. This grant founded his answer — instead of baking glass and metal together at furnace temperatures, a laser traces a ring-shaped bond around each wire's passage, fusing it shut while the rest of the part stays cool. It grew from a provisional application filed on Christmas Eve 2014 and seeded a decade-long line of patents.
2017 US10535596B2 Feedthrough assemblies and methods of forming same Core articulation of the laser-bond feedthrough claim.
This is the core articulation of the laser-bond feedthrough claim: a conductive via passes through a non-conductive substrate, and the contact capping it is hermetically sealed by a bond that surrounds the via — an interface that can be just tens of nanometers thin. Because the light delivers heat only where the joint forms, the wall can be glass, sapphire, or even silicon carbide, and the device can shrink without ever compromising the seal.
2018 US10420509B2 Sealed package and method of forming same
Conventional hermetic housings hang their fate on feedthroughs built from ferrules, glass seals, and high-temperature brazing — complexity and stress concentrated at the one joint an implant cannot afford to lose. This grant, part of a four-patent chain with Craig L. Schmidt, continues the sealed-package alternative: a cavity in the substrate covered to form the enclosure, with each external contact laser-bonded in a ring around its via, heating only a whisper-thin surface layer.
2018 US10813238B2 Hermetically-sealed packages including feedthrough assemblies
Where its sibling grants claim the feedthrough alone, this sole-inventor patent claims the whole sealed package built around it: the substrate carrying the laser-bonded contacts is integrated into the device's hermetic housing itself. It descends from one of the twin provisional applications filed on Christmas Eve 2014 — the branch of the feedthrough line he pursued under his name alone.
2019 US11103714B2 Sealed implantable medical device and method of forming same
Feedthroughs are where hermetic seals leak, and high-temperature joining can alter the very materials it connects. This axial design sidesteps both: the external contact is sealed at one end of the housing, and electricity crosses to the electronics inside through a conductive member held under compression — connection by pressure, not bonding. The housing can even be transparent glass or sapphire, letting an optical sensor read the body straight through the wall with no extra opening.
2019 US10765372B2 Sealed package and method of forming same
In this continuation of the sealed-package line with Craig L. Schmidt, the package stops being a box that contains the device and becomes the device's own body: a cavity carved into the substrate, a cover layer sealing it shut, and a via carrying signals out to a contact laser-bonded in a ring around it. The laser heats only the surface layer where the joint forms — and the finished bond is stronger than the bulk substrate around it.
2019 US11485670B2 Hermetic conductive feedthroughs for a semiconductor wafer
As implants shrink, the wired doorways through their walls have to shrink too — and conventional feedthrough builds don't miniaturize reliably. Here the conductor is entombed inside a glass wafer, its outer face never reaching the surface directly: the path out takes a deliberate dog-leg through a buried trace to a recessed cavity, so body fluid can't follow the wire back in. Defense in depth for electrons, built at wafer scale.
2020 US11419552B2 Sealed package and method of forming same
Traditional hermetic feedthroughs are little assemblies of ferrules, glass seals, and high-temperature brazing — manufacturing complexity and material stress at the most critical joint in the device. This continuation refines the alternative: a substrate with a sealed cavity that is itself the device body, each electrical contact laser-bonded in a ring around its via. The laser heats only the surface layer it touches, and the finished bond is stronger than the bulk of the substrate around it.
2020 US12610487B2 Sealed package including electronic device and power source
An implant's insides are full of joints — solder, adhesive, weld — and every one must stay stable for the life of the device. Here the electronics sit on a see-through substrate that is laser-sealed to the housing holding the power source, and the electrical connection between them is never bonded at all: the assembly's own geometry presses the contacts together and keeps them there. It is the newest grant in the corpus, issued April 2026, and 13 patent families already cite it.
2020 US11950387B2 Methods for forming hermetically-sealed packages including feedthrough assemblies
Every implantable device needs wires to pass through a wall that must stay absolutely sealed against body fluid. This sole-inventor grant covers the methods behind that trick: a laser traces a ring around each tiny via, welding the contact down so the doorway itself is fused shut instead of brazed in a furnace. Granted in 2024, it carried a line seeded by twin provisional applications filed one Christmas Eve in 2014 into its second decade.
2020 US11865639B2 Hermetic assembly and device including same
Joining a ceramic-like insulator to a metal frame classically demands either perfectly polished mating surfaces or brazing heat that stresses both parts. The move here is a translator layer: a patterned metal film is laser-bonded onto the insulator first, and the metal ferrule is then conventionally welded to that film — so the fragile insulator never feels the welding heat. And because the bonding works from one side, the frame can be attached after the electronics are already in place, something brazing temperatures would never allow.
2022 US12082354B2 Electronic package and device including same
A 2024 addition to the hermetic-packaging cluster: an electronic package built around shaped conductor blocks — the scalloped metal pieces at the base of the stack that carry the package's electrical connections. The drawings assemble it like a kit, layer by layer, from the covered components down to the arched blocks beneath the substrate.
2022 US11744518B2 Sealed package and method of forming same
Traditional implant housings are built up from ferrules, glass seals, and high-temperature brazing, each one a source of manufacturing complexity and material stress. This grant is the mature form of a different idea: hollow a cavity into the substrate itself, seal a cover layer over it, and laser-bond each electrical contact in a ring around its via — the package is the device body. The laser heats only the surface layer it touches, and the finished bond is stronger than the surrounding material.
2024 US12233477B2 Hermetic assembly and device including same
Joining an insulator to a metal frame classically demands perfectly polished mating surfaces or high-temperature brazing that stresses the parts. This 2025 design laser-bonds a patterned metal layer onto the insulator first, then conventionally welds the ferrule's flange to that layer — so the fragile material never feels the welding heat, and because the bonding is one-sided, the frame can be attached after the electronics are already in place. An adapter pattern in metallurgy: one layer translating between the laser-bond world and the weld world.
Laser Materials Processing · 11 grants
The craft thread: putting heat exactly where the joint is and nowhere else. It opens in 2000 with a laser fired through the eye of a bonding tool to weld a ribbon without vibration, and matures into kinetically-limited diffusion bonds — sapphire fused to titanium across an interface thinner than 1000 nanometers, co-invented with Corning. These techniques are the enabling hands behind the packaging cathedral.
2000 US6501043B1 Apparatus and method for laser welding of ribbons Sole-inventor founding of the laser craft arc; 35 forward citations.
Microelectronic connections were made by ultrasonic scrubbing — vibration that shook parts out of position and left weak, inconsistent bonds. Working alone, he threaded the ribbon under a bond head and fired a laser through an aperture in the tool itself: the head holds, the light joins. Thirty-five later patents cite it, and it opened a twenty-five-year arc of using focused light as a precision tool.
2002 US6717100B2 Apparatus and method for laser welding of ribbons
A sole-inventor continuation of the ribbon-welding patent that opened the laser-craft arc: a bond head holds a conductive ribbon in place while the laser fires through an aperture in the tool itself, so the weld forms without the vibration that used to shake delicate parts loose. The refinement went on to be cited by 38 later patents — more than the original.
2005 US7872208B2 Laser bonding tool with improved bonding accuracy
Laser bonding is only as good as the tool's aim — the weld has to land exactly where the foot is pressing. This patent, with Ruben listed first, refines the bonding tool's tip so the laser strikes its target more accurately, a workhorse improvement in his laser materials-processing line. A dozen later patents cite it.
2010 US9171721B2 Laser assisted direct bonding
Conventional direct bonding fuses two polished parts by heating them past 400°C in bulk — and that heat warps and cracks the very joint it is meant to strengthen, especially between mismatched materials. This method simply presses the parts into contact to form a weak room-temperature bond, then traces a laser along the seam to strengthen it locally while everything else stays cool. It is the keystone of the corpus's laser craft — the bridge between generic direct bonding and the whole family of laser-bonded packages that followed.
2010 US8796109B2 Techniques for bonding substrates using an intermediate layer
Some pairs of materials simply refuse to bond directly, and heating a whole stack would wreck whatever is inside. The trick here is a film just tens of nanometers thick sandwiched between the two surfaces: laser light passes clean through the top substrate and heats only that film, welding the pair right at the seam. There is even built-in quality control — the tinted film turns optically clear where the bond succeeds, so a good seal can be inspected by eye.
2015 US10124559B2 Kinetically limited nano-scale diffusion bond structures and methods Kinetically-limited nano diffusion bond; Corning co-assignment = external validation.
Joining sapphire to titanium — a transparent gem to an opaque metal — normally demands a furnace at 600 to 1000 °C. Invented with Corning, this technique fires a laser through the sapphire so its energy lands only at the buried interface, in a flash too brief for the heat to wander, leaving a bond thinner than 1000 nanometers between two materials that never felt an oven. The Corning co-assignment is the corpus's clearest external validation: a materials-science giant putting its name beside his.
2018 US10981355B2 Kinetically limited nano-scale diffusion bond structures and methods
A laser pulse fired through sapphire deposits its energy exactly at the interface with the metal beneath, forming a diffusion bond thinner than a thousand nanometers. 'Kinetically limited' is the heart of it: the flash is too brief for heat to migrate, so diffusion happens only in that nano-zone and the parts never really get hot — with named applications reaching past medical packaging to vacuum windows, photonics, and spacecraft. The patent carries both Medtronic's and Corning's names — a materials-science giant's endorsement written into the record.
2019 US11548092B2 Surface texturing using energy pulses
An implanted electrode wants the most surface area in the least space, but texturing metal in a gas environment can't make the features fine enough. The answer: fire the energy pulses at the metal while it sits under liquid, whose faster heat transfer quenches the melt quickly enough to freeze in texture at micro scale or smaller. The bath does double duty — its oxygen and nitrogen react with the hot metal so texture and surface chemistry arrive in the same pulse, for electrodes, stent struts, and housings.
2020 US11999014B2 Laser cutting system
Laser-cutting a board with buried conductors has a hidden failure mode: the cut smears conductive residue across the exposed face, leaving invisible paths for leakage. This system cuts with one beam and follows with a second, angled to clean the cut surface — and it watches for the plasma flash when the first beam strikes metal, using that flash to locate the conductor and aim the cleanup beam automatically. Late-career work, granted 2024: even the cutting step gets closed-loop intelligence.
2021 US12454117B2 Kinetically limited nano-scale diffusion bond structures and methods
This is the 2025 form of the diffusion-bond line David shares with glassmaker Corning — one of the few places in the corpus where another company's name sits beside Medtronic's on the patent itself. A laser fires through a transparent material into the buried face where it meets metal, forming a kinetically limited bond only nanometers deep while the bulk of both materials stays cool and undisturbed. The line has drawn 34 citing patent families.
2022 US11969821B2 Surface texturing using energy pulses
An electrode works better the more surface it packs into a tiny footprint, and texturing in a gas environment can't make the features fine enough. The method here fires energy pulses at metal sitting under liquid: the liquid quenches the heat faster, leaving micro-scale-or-smaller texture, while its oxygen and nitrogen react with the hot metal so texture and surface chemistry form in the same pulse. A 2024 continuation of the underwater-texturing line, aimed at electrodes, stent struts, and device housings.
Implantable Sensors · 9 grants
A therapy device is blind without senses. This cluster gives implants their proprioception: shock-proof piezoelectric accelerometers that let pacemakers feel a patient's motion and pace the heart to match (the verified rate-responsive lineage), capacitive pressure sensors robust to parasitic effects, and — in the pending pipeline — multi-analyte biosensor arrays. The arc runs from feeling motion to reading blood chemistry.
1997 US5885471A Shock resistant accelerometer for implantable medical device First-listed; the beam-stop that made implantable accelerometers survivable. THE most-cited patent in the corpus: 203 direct citations + 30 citing families.
A rate-responsive pacemaker senses its patient's activity through a tiny piezoelectric beam — a diving board that generates voltage as it flexes — but early beams were so fragile that dropping the device from a few feet could snap them. This invention adds a stop that leaves normal flexing untouched yet catches the beam just before a shock would break it. It is the most-cited patent in the corpus, with over 200 later patents building on it.
1997 US5911738A High output sensor and accelerometer implantable medical device cited 105 times
Part of the line that taught pacemakers to feel motion: a piezoelectric sensing element, built with James M. Sikorski, engineered to deliver a stronger electrical signal as it flexes with the body's movement — a clearer read on patient activity for rate-responsive pacing. With 105 citing patents, it ranks among the most-cited work in the corpus.
1998 US6038475A High output sensor and accelerometer for implantable medical device cited 85 times
The second of two grants with James M. Sikorski refining the accelerometer line's high-output sensing element — the piezoelectric beam whose flexing with body motion tells a pacemaker its patient is up and active. A continuation that sharpened the earlier design, it has been cited by 85 later patents.
1999 US6216537B1 Accelerometer for implantable medical device Solved the sensitivity-axis/manufacturing conflict for activity sensing.
A motion-sensing beam measures along only one axis — and the axis that matters clinically points into and out of the patient's chest, while sensors naturally lie flat on the circuit board, aiming the wrong way. This surface-mount package uses multilayer conductive end caps to stand the sensing element upright, pointing its sensitivity where medicine needs it while staying mass-producible and testable. More than 130 later patents cite the solution.
2005 US9616223B2 Media-exposed interconnects for transducers
Some sensor connections cannot hide behind a hermetic wall — the strain of a beating heart is too much for sealed designs — yet bare connections corrode. This one stacks two conductive layers whose stresses pull in opposite directions, one compressive and one tensile, cancelling to almost no net load, and builds the contacts from valve metals like titanium whose surface oxide resists corrosion in saline. Filed in 2005 and granted twelve years later, it is cited by Endotronix, the company behind an implantable heart-failure pressure sensor.
2008 US7591185B1 Pressure sensor configurations for implantable medical electrical leads
A cardiac lead is normally just wiring; this patent puts a working pressure sensor inside the lead itself. Behind an insulating sidewall, a pressure-sensitive diaphragm moves one plate of a tiny gap capacitor read by an on-board chip, with a grounded conductive layer shielding the measurement. It is the moment the wire becomes an instrument.
2009 US7886608B2 Pressure sensor configurations for implantable medical electrical leads
A cardiac lead is normally just a wire; this work builds a pressure sensor directly into it, so the wire itself becomes an instrument. Inside the lead's insulating wall, a pressure-sensitive diaphragm forms one plate of a gap capacitor read by an integrated circuit, with a grounded conductive layer shielding the diaphragm from interference. It's a continuation in a sensing line whose ideas were picked up by 46 later patent families.
2011 US8424388B2 Implantable capacitive pressure sensor apparatus and methods regarding same
Tiny capacitive pressure sensors are haunted by stray capacitance, and the problem is worst exactly where implants live — surrounded by conductive tissue and fluid. This design flips the obvious layout: the electrode facing the body is the grounded, pressure-flexing diaphragm, while the signal electrode sits isolated on an insulator behind it, so the patient's own conductivity becomes shielding instead of interference. Design judo — the hostile environment is recruited into the circuit.
2022 US11725995B2 Pressure sensor assembly for use in implantable medical device (substrate with via along via axis)
Shrink a capacitive pressure sensor far enough and its signal drowns — stray capacitance and manufacturing variation swamp the tiny reading, and protective films only add stress. The fix removes the middleman: the flexing membrane carries one capacitor plate while the second sits on the integrated circuit itself, so the sensing capacitor forms straight along the axis of a via. The membrane can even be part of the enclosure — the sensor's own package becomes the sensing element.
Power Sources · 8 grants
Every battery change is a surgery. This cluster attacks that arithmetic from every direction — wafer-scale packages with power sealed inside, stacked thin-film battery fabrication, and the boldest line: nuclear micro-power. Medtronic implanted plutonium pacemakers in 1970; one ran 34 years. David's 3D nanostructured betavoltaic converters are that dream rebuilt with quantum dots and metal foam — a radiation sponge in place of a thermal slug.
2011 US8666505B2 Wafer-scale package including power source Wafer-scale package w/ internal power: 50 direct citations + 126 citing families — top-tier measured influence in the leadless-device era. Promoted on sweep data 2026-07-22.
Direct wafer bonding runs hot enough to destroy a battery — yet a device isn't truly complete at wafer scale until its power source is sealed inside. This design puts the control chip and the battery together in a cavity between two bonded substrates, then closes the seam with low-temperature laser-assisted bonding so nothing inside is harmed. It became one of the most-cited pieces of work in the corpus, with 126 later patent families drawing on it as tiny self-powered leadless implants became an industry.
2012 US9252415B2 Power sources suitable for use in implantable medical devices and corresponding fabrication methods
A single thin-film battery cell is only about 14 microns thick — far too little energy on its own to power an implant. So the cells are made like chips, in arrays on wafers, then stacked: routing layers carry each cell's contacts out to the side, vias gang the aligned contacts in parallel, and micron-scale cavities engineered between the cells give them room to swell as they charge. IBM and France's atomic-energy agency are among the citing parties — battery stacking with relevance well beyond medicine.
2014 US9318400B2 Wafer-scale package including power source Continuation of the wafer-scale power line; 151 citing families — the corpus max. Promoted on sweep data 2026-07-22.
Fusing two wafers into one sealed package normally takes 400°C or more — heat that would destroy any battery inside — yet a device is not truly wafer-scale-complete without its power source built in. This design encloses both the control chip and its battery in a cavity between two bonded substrates, hermetically sealed by low-temperature laser-assisted bonding, and explicitly contemplates leadless device configurations. With 151 later patent families citing it, it is the most-cited grant in the corpus.
2015 US10096393B2 Nuclear radiation particle power converter Founds the 3D betavoltaic line reviving Medtronic's nuclear-power heritage.
A betavoltaic cell works like a solar panel lit by radiation instead of sunlight — but a flat panel lets much of each particle's energy slip away before it is collected. This design crumples the collector into a three-dimensional sponge of coated nanostructures, so nearly every particle strikes a working junction before its charge can fade. It founded the patent line reconnecting Medtronic to its nuclear-pacemaker heritage: the pursuit of an implant whose power source never needs replacing.
2018 US10811157B2 Nuclear radiation particle power converter
A flat radiation-to-electricity converter wastes much of what its fuel emits: charge carriers recombine before they can be collected. This continuation carries the 3D answer forward — a sponge-like collector of nanorods, nanotubes, or foam, coated with charge-separating layers such as quantum dots, with the isotope infiltrated into the structure itself, even by exposing the finished sponge to pressurized tritium gas. It is a solar cell rebuilt around beta decay, with the fuel soaked into the electrode.
2019 US10818811B2 Power source and method of forming same (radioluminescent: radioactive to light to photovoltaic)
Instead of converting radiation straight to electricity, this power source takes a detour through light: a hermetically sealed cavity holds radioactive material beside a phosphor that glows under its particles, and a photovoltaic layer harvests the glow. The design even accounts for the helium that tritium decay produces, using a gas-permeable substrate so the sealed cavity doesn't slowly pressurize itself over the decades — a battery designed around its own exhaust. It founded the radioluminescent branch of the corpus's nuclear-power line.
2020 US11189390B2 Power source and method of forming same
A nuclear battery is only as good as its fuel storage, and tritium — a workhorse isotope — is a gas, which limits how much power can be packed into a small space. The answer here is a loaded substrate: a carrier material holds the isotope densely, thin deposited barrier layers keep it from wandering, and the whole assembly sits inside a sealed enclosure. Contain the fuel properly, and the rest of the battery can do its job.
2020 US11881325B2 Nuclear radiation particle power converter
A radioisotope is a battery whose lifetime is set by nuclear physics rather than chemistry — the only power source that can promise decades of output sealed inside a body. This grant is the latest and broadest form of the converter: a three-dimensional current collector coated with charge-separating layers, with the radiation-emitting fuel placed so its particles strike inside the sponge rather than a flat panel. It issued in January 2024, just as nuclear microbatteries were making startup headlines, on work with priority stretching back to 2014.
Flow Control & Surgical Navigation · 3 grants
The smallest cluster carries the heaviest human stakes: MEMS-actuated shunt valves for hydrocephalus — closed-loop pressure control for a condition where device failure often means brain surgery, frequently a child's — plus accelerometer-corrected magnetic tracking for surgical navigation. Engineering applied where the margin for error is a person.
2010 US9913693B2 Error correction techniques in surgical navigation
Magnetic tracking pinpoints a surgical instrument faster than the patient can be re-imaged, so every heartbeat and breath shifts the anatomy and leaves the instrument drawn in the wrong place on a frozen picture. The fix pairs the magnetic sensor with an accelerometer: when acceleration spikes, the system falls back on trusted earlier positions, and frequency analysis can even tell a heartbeat from a breath from the surgeon's own hand. It is his accelerometer craft resurfacing thirteen years later in the operating room, now cited by 145 later patent families.
2020 US12059542B2 System and method for valve control
In hydrocephalus, excess fluid must drain from the brain through an implanted shunt — and the valve doing that job has to hold exactly the right pressure. This grant continues the closed-loop valve line: a MEMS micro-machine inside the flow-control assembly adjusts the valve according to the selected pressure settings. It belongs to the family the exhibit ranks as carrying the heaviest human stakes in the corpus.
2020 US11701503B2 System and method for valve control MEMS closed-loop hydrocephalus valve; heaviest human stakes in corpus.
For a person with hydrocephalus, an implanted shunt valve quietly decides how much cerebrospinal fluid drains from the brain — the heaviest human stakes anywhere in the corpus. This valve is not a simple spring but a layered machine: tiny MEMS actuators flex a membrane to tune the opening, while a pressure sensor reports back so the implanted controller keeps adjusting until measured flow matches the doctor's target. The settings arrive wirelessly from a handheld programmer outside the body.
Interconnects & Materials · 5 grants
The quiet cluster: corrosion-resistant contact stacks using titanium-niobium, metallization with deliberately tailored (even negative) thermal expansion so joints don't tear themselves apart across temperature swings, freeform substrates cut to the device's interior shape, and terminal arrays that replaced hand-wiring. Nearly all sole-inventor work — the personal toolbox underneath everything else.
2000 US6787891B2 Freeform substrates and devices
Substrates come out rectangular because saws cut straight — but implantable devices are curved, so corners of precious interior volume sit empty. Ruben's sole-inventor answer, filed in 2000, was to cut the silicon freeform to the device's interior contour using a water-jet-guided laser, shaping the circuitry to match. Every reclaimed cubic millimeter is that much less bulge in a patient's chest.
2002 US6963780B2 Implantable medical device including a surface-mount terminal array First-listed; terminal array cited ~31 times by effectively the whole rival CRM/neuro industry (Boston Scientific, BIOTRONIK, Pacesetter, Cardiac Pacemakers, Greatbatch) plus Verily and TDK. Promoted from major when citation data landed (2026-07-22).
Welding and wire-bonding straight onto a circuit board kept damaging the board itself: solder contamination, heat cracking, costly rework. This molded terminal array is a translation layer — each terminal offers a small solder face downward for tight-pitch fillets and a large bond head upward for the bonding tools, so each joining technology gets the surface it wants. Rivals noticed: the patent has been cited roughly 31 times by effectively the whole competing pacemaker-and-neurostimulator industry, from Boston Scientific and BIOTRONIK to Verily and TDK.
2004 US7335530B2 Freeform substrates and devices
A continuation of Ruben's sole-inventor freeform-substrates idea: instead of a rectangular circuit board leaving the corners of a curved implantable device empty, the substrate is cut to the shape of the device's interior. Even the integrated-circuit components are formed to follow that contour, freeing room for more electronics — or letting the whole device shrink.
2007 US8461681B2 Layered structure for corrosion resistant interconnect contacts
Electrical contacts that sit in blood corrode, slowly degrading an implanted sensor over the years it must survive. This sole-inventor design armors them in a three-layer sandwich — titanium-niobium, niobium, titanium-niobium again — chosen for corrosion resistance and for thermal expansion that can be tuned to match the silicon beneath. It was filed the same day as its twin, the negative-expansion weld pad, as two complementary answers to the slow thermal-mechanical death of interconnects.
2007 US8141556B2 Metallization with tailorable coefficient of thermal expansion
Welding a wire to a thick metal pad on silicon can crack the chip, because the heat makes pad and substrate expand at different rates. The sole-inventor answer is a composite pad containing a material that shrinks when heated — niobium and niobium pentoxide, blended so the mismatch cancels to zero — and the patent hands over the design equation as a recipe. The idea traveled: four of its visible citations come from Lockheed Martin, defense-aerospace uptake of an implant metallurgy trick.
Still in motion
The record is not closed: 8 notable applications remain pending as of 2026 — analyte biosensors, packaging, and power among them. The machine is sealed; the work is not.