CMOS Pixelated Capacitive Sensor platform for biosensing and many other applications Frans Widdershoven Smart sensors NEREID workshop 21st October 2016 COVER PAGE SUBTITLE PLACEHOLDER COMPANY CONFIDENTIAL
CMOS integration victory: radio Vacuum tubes Transistors Bipolar analog radio IC CMOS digital radio IC (software-defined radio) 2. October 21, 2016 COMPANY PUBLIC
CMOS integration victory: camera Film Vacuum tube (Vidicon) Charge-coupled device (CCD) CMOS image sensor (CIS) 3. October 21, 2016 COMPANY PUBLIC
CMOS integration victory: storage Ticker tape Magnetic tape Hard disk CMOS-based solid- state drive (SSD) 4. October 21, 2016 COMPANY PUBLIC
Widdershoven’s 3 laws of IoT 1) Only non-trivial data need be transmitted 2) Autonomous devices need sensors to generate non-trivial data 3) What can be sensed by CMOS will be sensed by CMOS Examples: Voltage, current, power, temperature, RF spectrum, ambient light, magnetic field, images, radar, information, uniqueness (Physically Unclonable Functions), … … and this trend continues! 5. October 21, 2016 COMPANY PUBLIC
Extend the CMOS integration victory to biosensing? Maybe, but only if we stay close to “standard” CMOS It’s the result of a multi-B$ world- wide aligned development effort, so don’t mess • it up! Exploit its strengths: • • Small feature sizes • High speed & low power • Embedded signal conditioning, A/D conversion, programmability,… • Low-cost and high-yield volume production • … 6. October 21, 2016 COMPANY PUBLIC
Biological length scales CMOS technology nodes in production (October 2016) 7. October 21, 2016 COMPANY PUBLIC
IgG antibody compared to 14-nm finFET TEM cross-section through 2 fins (Intel) 8. October 21, 2016 COMPANY PUBLIC
“Capacitive” sense electrode Cross-section: Equivalent circuit: Electrolyte Captured target 𝑎 𝐹 Probe 𝑎 𝐸𝑀 Linker 𝜇 𝐸 Double layer 𝐷 𝑇 SAM Electrode 𝜇 𝐸 : Debye length (~0.8 nm at 150 mM salt concentration) 9. October 21, 2016 COMPANY PUBLIC
Disturbs Electrolyte Configurations of probes and captured targets Non-specific binding and molecular charge − − SAM defects + − and relaxation Double layer SAM Electrode Intrinsic disturbs Extrinsic disturbs 10. October 21, 2016 COMPANY PUBLIC
Sensitivity scaling 𝑎 1 = ℎ 1 𝑊 1 𝐵 1 𝜏 1 𝐽 ℎ 1 , 𝐵 1 , 𝜏 1 : height, area and complex conductivity of element 1 𝑎 1 𝑎 = 𝑎 𝑇 + 𝑎 𝐶 = 𝑊 𝑊 𝑊 𝑎 2 𝐽 = = 𝑊 𝑊 𝐵 1 𝐾 1 𝐵 1 𝜏 1 𝐹 1 2 𝑍 = 1 𝑎 2 𝜖𝑍 𝐹 1 ∆𝑍 ≈ ∆𝜏 1 = Ω 1 ∆𝜏 1 𝜖𝜏 1 𝑊 Ω 1 = ℎ 1 𝐵 1 (volume of element 1) Sensitivity is proportional to the square of the local electric field strength! 11. October 21, 2016 COMPANY PUBLIC
Model system: semi-spherical metal nanoelectrode C Electrolyte 𝐹 𝑠 B 𝑠 0 Electrode A 2 𝑠 0 𝐹 𝑠 Relative sensitivity = 𝑊 𝜇 𝐸 𝑠 0 = 85 nm Distance above electrode (nm) 12. October 21, 2016 COMPANY PUBLIC
Frequency ranges 𝑔 𝑔 1 2 2 = 𝜏 𝐹,𝐸𝐷 𝑔 2𝜌𝜗 𝐹 𝑔 2 𝑔 1 ≈ 1 + 𝑠 0 𝜇 𝐸 Semi-spherical nano electrode ( 𝑠 0 = 85 nm ) and 150 mM salt concentration: 𝑠 0 = 85 nm 𝑔 1 ≈ 3.3 MHz, 𝑔 150 mM 2 = 360 MHz Frequency range Bulk sensitivity Surface sensitivity Low (“blocking” double layer) A (< 3.3 MHz) High (saturated at low-frequency level) B (3.3 – 360 MHz) Nominal (“transparent” double layer) High (still exceeding bulk sensitivity) Nominal (“vanished” double layer) C (> 360 MHz) Nominal (same as bulk sensitivity) 13. October 21, 2016 COMPANY PUBLIC
Validation by numerical simulations Touching SAM surface Inside double layer Above double layer SAM: 2.5 nm thick; PNA/DNA: 13.2 nm long (40-bp) Federico Pittino, Federico Passerini, Luca Selmi, Frans Widdershoven, Microelectronics Journal 45 (12), December 2014 14. October 21, 2016 COMPANY PUBLIC
Issues Probe molecules Typically much larger than 𝜇 𝐸 (e.g. largely extending above double layer) • Should not stick directly to SAM surface (to avoid denaturing and to keep capturing sites accessible for • target molecules) Other molecules May stick directly to SAM surface (non-specific binding) • Issue At frequencies below 𝑔 2 the sensitivity for non-target molecules and/or SAM surface damage is much • higher than for target molecules Solution Use modulation frequencies ≥ 𝑔 • 2 (or at least as high as possible)* * You won’t find this experimentally by searching for the frequency that gives the highest response 15. October 21, 2016 COMPANY PUBLIC
NXP’s CMOS Pixelated Capacitive sensor chip V T V T Φ T Φ T Φ T C C V N V N V L V L Φ D Φ D Φ D V D V D t 1 t 2 t t = t 1 t = t 2 𝑅 = 𝑀 0 𝑊 𝑈 − 𝑊 𝐷 + 𝐷 𝑄 𝐸 ∆𝑅 = 𝑀 0 𝑊 𝑈 − 𝑊 𝐸 ∆𝐷 Typically operated at 1 – 50 MHz 𝑀 0 : number of charge/discharge cycles 𝐷 𝑄 : parasitic capacitance (~0.4 fF) (runs up to 320 MHz) October 21, 2016 COMPANY PUBLIC 16.
Intrinsic SNR of single sensor cell Cumulative reset noise 2 switch transistors: 𝜏 𝑅2 = 2𝑂𝑙 𝐶 𝑈 𝐷 + 𝐷 𝑄 … and no 1/f noise (at least in principle) Intrinsic signal/noise ratio (SNR): 𝐸 2 ∆𝐷 2 𝑇𝑂𝑆 0 = Δ𝑅 2 = 𝑂 𝑊 𝑈 − 𝑊 𝜏 𝑅2 2𝑙 𝐶 𝑈 𝐷 + 𝐷 𝑄 Reason for using Reason for using nanoelectrodes advanced CMOS 17. October 21, 2016 COMPANY PUBLIC
Cross-section (90-nm CMOS) Contrast Nanoelectrode E-less Au layers moisture barrier nano-electrode Cu via-4 TaN/Ta metal-4 M4 via-3 metal-3 M3 via-2 metal-2 M2 via-1 metal-1 M1 V T Φ T V D V T V D V T contact Φ T Φ D Φ T Φ T Φ D Φ D Φ T Φ T Φ D Early example with e-less Au nanoelectrodes source/drain poly-gate p-well (not used anymore) 18. October 21, 2016 COMPANY PUBLIC
NXP’s CMOS Pixelated Capacitive sensor chip (2) Chip features: 1) 256×256 (= 65,536) nanoelectrodes 2) 4 temperature sensors 3) 8 A/D converters Fluidic seal patch 4) 256 digital data accumulators 3.2 mm × 2.1 mm in 90-nm CMOS (TSMC) Flip over Spring pins Fluid ports Current process: Au-rich AuCu Modified CSP test socket Thermal interface via backside nanoelectrodes, made “the CMOS way” (Aries Electronics part number A1924-314-23) of chip (not shown) October 21, 2016 COMPANY PUBLIC 19.
Reconfigurable counter electrode Active row Counter electrode 20. October 21, 2016 COMPANY PUBLIC
Reconfigurable counter electrode Counter electrode Active row Counter electrode 21. October 21, 2016 COMPANY PUBLIC
Reconfigurable counter electrode Counter electrode Active row Counter electrode 22. October 21, 2016 COMPANY PUBLIC
Reconfigurable counter electrode Counter electrode Active row Counter electrode 23. October 21, 2016 COMPANY PUBLIC
Lower cut-off frequency (region A B) 24. October 21, 2016 COMPANY PUBLIC
Quantitative agreement with simulations Measured Simulated 25. October 21, 2016 COMPANY PUBLIC
Detection of particle conduction type Conducting Insulating 26. October 21, 2016 COMPANY PUBLIC
Detection of nanoparticle binding • Nanoparticles captured by BSA layer on chip surface • Independent verification with AFM Unpublished pictures removed 27. October 21, 2016 COMPANY PUBLIC
Detection of … virus (one of the smallest viruses) Unpublished pictures removed 28. October 21, 2016 COMPANY PUBLIC
Collecting statistics of captured nanoparticles Unpublished pictures removed 29. October 21, 2016 COMPANY PUBLIC
Counting of particles and imaging of living cells 1-µm dielectric particles in water (pH = 3) MCF7 breast tumor cells in growth medium 30. October 21, 2016 COMPANY PUBLIC
Imaging of droplets in water-based emulsions Unpublished pictures removed 31. October 21, 2016 COMPANY PUBLIC
Adding smartness: automatic particle tracking Unpublished pictures removed 32. October 21, 2016 COMPANY PUBLIC
The future: surfing Moore’s Law? 40-nm CMOS design exercise • 3 times cell area shrink • Resolution comparable to that of optical microscopes And what about 14-nm CMOS? 33. October 21, 2016 COMPANY PUBLIC
Thanks to my great collaborators! Udine University (I) Luca Selmi, Federico Pittino, Federico Passerini, Pierpaolo Palestri, Andrea Bandiziol, • Paolo Scarbolo, Andrea Cossettini University of Twente (NL) Serge Lemay, Cecilia Laborde, Christophe Renault, Vincent de Boer, Regine van der Hee, • Jeroen Cornelissen Wageningen University & Research (NL) Maarten Jongsma, Harrie Verhoeven • 34. October 21, 2016 COMPANY PUBLIC
Recommend
More recommend
