Presentazione di PowerPoint · Titolo presentazione Nome relatore Tutorial 4 Low-Noise Impedance...

Titolo presentazione Nome relatore

Tutorial 4

Low-Noise Impedance Sensing: Circuits and Micro-

Technological Applications

Marco Carminati

Politecnico di Milano - Dipartimento di Elettronica, Informazione e Bioingegneria [email protected]

M Carminati - Tutorial 3 - Low-noise Impedance sensing

Motivation: Impedance is Everywhere

Thanks to its versatility, impedance is traditionally leveraged in several industrial sensing applications

2

Strain/Pressure Level Displacement

Resistance Capacitance Inductance


Introduction

Focusing on the example of contactless capacitive sensing,

let’s see the scaling trend from macro to micro size

applications

3


Macro: Contactless Liquid Sensing

4

Air εr = 1

Liquid εr > 1

Electrodes are protected from the liquid, the capacitance increases linearly with the tank level

Range 10-1000pF


Macro: Contactless Capacitance Tomography

5

Plastic bars εr = 3

Plastic beads εr = 3

Air εr = 1

Test structures Obtained Maps

Pipe cross-section

Impurities inside pipes (or droplets in microfluidic channels) can be detected with several electrodes properly arranged

Range 10-100pF


Mini: Capacitive Touchscreen

6

Range 10-100fF


Milli: Capacitive Fingerprint Scanner

7

Range 5-50fF


Micro: Inertial MEMS

8

Capacitive sensing of displacement of micro proof masses

Range sub-aF


Tutorial Outline

• Definitions and spectra diagrams • Microscale impedance measurement • Introduction to electronic noise • Low-noise current sensing • Transimpedanace amplifier and lock-in • Role of the input (stray) capacitance: cables and substrates • Advanced circuit topologies • Example of applications:

• Biological cell counting • Particulate matter detection • Non-invasive light power metering

9


Definitions

Impedance can be measured:

10

v(t)

tT = 1/f

∆t

Ip Vpi(t)

Z Z(f) = eVp

Ip

j2πΔtf

0

Impedance is a complex quantity changing with frequency f:

ReZ: energy dissipation, ImZ: energy storage

Time tracking

t

|Z| Spectroscopy

f

|Z| fixed f


Plotting Impedance Spectra

Two alternative ways to present impedance spectra:

11

log(f)log(f)

Mag

nitu

de(Z

)

Pha

se(Z

)

-Im(Z

)

Re(Z)

Bode Plots Cole-Cole Plot

f

Electronic engineering and automation/control science

Electrochemical & biosensors community

Equivalent!


Microscale Impedance Sensing

Microfabriaction allows the realization of microelectrodes to interface with microscale samples (such as biological cells).

12

The impedance of the micro sample scales with

the size

while

the parasitic impedance of the macroscopic

connections stays fixed.


Noise Becomes Important

If Z increases, given the same excitation voltage (limited by biology or dielectric breakdown), the current flowing through the system decreases:

13

v(t)

i(t)Z

It becomes harder to measure a smaller current

The noise of the instrument (current reader) becomes significant!

The noise is a property of the detection instrument but might depend on the sample and on the connections.


What is Noise?

14

Any electrical signal is affected by disturbances and noise:

Noise is a random fluctuation of the electrical variables due to the physical behaviour of internal

components of the circuit.

Disturbances are signals of external origin that can be filtered.

Described by standard deviation σ


Noise Power Spectrum

Description in the frequency domain:

15

Noise spectral density


Impedance Measurement Techniques

• Sinusoidal stimulation: (Measurand) • Wheatstone bridge (voltage) • Ratiometric (voltage) • Resonant (frequency) • Current sensing + lock-in (current)

most versatile approach, better rejection of parasitics

• Non sinusoidal stimulation: • Multi-sine or pseudo noise FFT-based (current) • Time-domain fitting of step response (current/voltage) • Capacitance charge and discharge (current/time)

16


Impedance Detector Architecture

17

LPF

LPF

cos(2πf0t)

VACsin(2πf0t)+VDC

Sample

sin(2πf0t)Current detector

0VCs1 Cs2

i(t)OUTI

OUTQ

IIR LPF

SampleCurrent detector

OUTI

OUTQ

D/A

n1 bit

A/D

n2 bit

DDS

FPGA

Lock-in digital implementation:

FPGA embedded processing for: • speed • versatility (multi-freq.)

Careful analog design

M. Carminati, IEEE I2MTC. 2012


Transimpedance Amplifier: Noise Analysis

18

Input-referred total current noise:

• Minimization of parasitics • Major motivation for front-

end integration • Accurate modeling of the

sensor impedance

Minimize CStray!

f

Si


Getting Rid of Silicon When Not Needed

19

M. Carminati, “Accuracy and resolution limits in quartz and silicon substrates with microelectrodes for electrochemical biosensors”, Sens. Act. B, 174, 2012.

Connection pad

Spring connector

Passivation layer Silicon substrate

Electrode exposed to the liquid

solution

Silicon substrate

VACSiO2

gold gold

CsCsub2Csub1

isub

VO

RF

isample

Subs

trat

e co

mpa

rison

Amplifier


Transimpedance Amplifier: Limits

20

• gain ∼ RF • noise ∼ √4kT/RF • BW ∼1/(RF CF)

Resolution/speed trade-off

For example: RF = 1GΩ CF = 0.2pF → Si = 4fA/√Hz BW = 800Hz

Need for enhanced topologies

Single design parameter RF

C stray SMD package


1 - Compensated TIA

21

Introduce a zero that cancels the pole: RF⋅CF = RD⋅CD

Wider bandwidth, low noise Manual adjusting of the zero

Ciofi, IEEE Instr.& Meas. 55 2006, M. Carminati, Analog IC Sig. Process. 2013 f

Vout/iin Cascade


2 - Integrator-Differentiator Cascade

22

Remove noisy RF

Robust, linear, no calibration Handling input DC current

Vout/iin Cascade


2.1 - Discrete Time: Reset Switch

23

For example: IIN-DC = 1nA, CF = 1pF → (5V) TMAX = 5ms

Switch periodically closed to discharge the capacitor:

Limits the maximum measurement time

M. Bennati, M. Tartagni , A Sup-pA Delta-Sigma Current Amplifier for Single Molecule Nanosensors”, IEEE ISSCC 2009


2.2 - Continuous Time: Active Reset

24

To achieve continuous-time operation: additional feedback branch H(s) with high gain at low frequency:

G. Ferrari, RSI 78 2007 & IEEE JSSC 2009, M. Carminati, RSI 80 2009.

γ1

CT More complex design


Transfer Functions

25

Two separate paths and gains: • DC and low frequency • High frequency

γπ DCFm RC

f2

1= For wideband impedance spectroscopy,

large RDC is required

DC

AC


Mitigation of Source Noise: Differential Sensing

26

Single-ended Differential

No compensation Compensated (same range)

Compesanted (range reduced)

No compensation Compensated

Capacitance noise [aF rms]

Conductance noise [nS rms]

A

B

C

+

_

RF

+

_

RF

+_

ZC

ZS

Phase Shifter

C

Dummy path

Limited by matching!

(b)

(a)

+

_

+

-1

-G

CF

CFCS

CC

A

B

Sensor

(b)

(a)

_

+

_

-G

C

CFCS

CC

B

Se so

Significant noise reduction

27

50

11

M. Carminati, Proc. IEEE SSD 2014 (Best Paper Award).


Example of Micro-Scale Applications

1.Biological Cells (resistance) 2.Airborne Dust (capacitance) 3.Silicon Photonics (resistance in AC)

27


Applications to Cell Biology

28

• ρPBS = 66Ωcm • εPBS = 78 • CMEM = 0.01pF/µm2

• ρCYTO ∼ ρPBS • εCYTO = 60

Small signal equivalent of a passive cell: the single shell model

At low frequency (<1MHz), the cell can be treated as an insulating sphere (resistivity contrast)


Cell Counting

29

The presence of cells can be sensed by impedance as a perturbation of the electrical field between two electrodes

“Static” “Dynamic”

Two approaches:


Electric Cell-substrate Impedance Sensing

30

Lo, Keese, Giaever, Biophys J. 69 1995

ECIS Technique invented by Giaever and Keese

Cell adeshion, spreading and growth can be

monitored


Comparison of Sensing Geometries

31

r4Rsol

ρ=

20dl rCC ⋅π⋅=

Conformal Mapping


Comparison with HeLa Cells

32


Impedance Flow Cytometry

33

• A novel approach, pioneered by Morgan and Renaud (∼2000) • Tracking impedance between close electrodes over time • High-throughput single-cell analysis (beyond sizing)

Cells

Microfluidic Channel

Flow

Electrode Electrode

Impedance

Time

T. Sun & H. Morgan, Microfluid. Nanofluid. 8 2010 K. Cheung et al., Cytometry 77A 2010


Electrodes Configurations

34


Design of Sensing Electrodes

• Simple technology → coplanar • Ease of alignment → transverse • Target cells 5-15µm → 20 x 40µm2

2 40 Gap


Optimal Electrode Design

36

• The gap sets the vertical extension of the sensing volume • An optimal gap exists for a given cell diameter • Sensitivity scales with cell volume and vertical height


Operating Frequency

100 1k 10k 100k 1M 10M

100k

1M

10M

CS

RSOL

Transition

Impe

danc

e M

agni

tude

[Ω]

Frequency [Hz]

CDL

Four regions in the impedance spectrum:

Operating Frequency Range

Interface

= 0.7pF stray to be minimized

CDL RSOL

CS

Bulk Infinite channel

Shallow channel

37


Enhanced Device Architecture

• Integrated into the device with no size increase • Third electrode available for differential sensing (drift)


Miniaturized Impedance Detector

• On-board signal generator • ASIC 1MHz CMOS lock-in

and 20bit Σ∆ converter

• Real-time FPGA peak detection algorithm

• Digital USB data acquisition


Credit Card Sized Implementation

Ultra compact system

3.3V USB

Power Supply

2 channels: • 2 outlets

• Complex impedance


Zurich Instruments

High-Pass Threshold

FPGA Detected Peaks

Our System

26.426.526.6

26.726.826.9

-1.0-0.50.00.5

20 21 22 23 24 250

1

R [k

Ω]

R [k

Ω]

HP [m

V]

Peak

s

Time [ms]

• Impedance tracking comparable with state-of-art instrument • Error-free real-time FPGA peak counting at 2000 events/s

Performance Assessment


System Validation with Beads

• 10µm polystyrene beads in PBS • 50 beads/s (1µl/min) • VAC = 50mV, f0 = 100kHz, 5kSa/s

0.7 0.8 0.9 1.0 1.10

10

20

30

40

50

60

70

Coun

tsPulse Amplitude [% ∆R/R]

0.4 0.8 1.2 1.6 2.00

50

100

150

200

250

300

Pulse Half Width [ms]

As expected:


Counting Yeast

10µm

• Sacc. Cerevisiae (5µm) • 105 cells/ml in 0.35S/m medium (with BSA) • VAC = 650mV, f0 = 100kHz, 2kSa/s

11.87 11.88 11.89 11.90 11.91 11.92 11.93

107.8

107.9

108.0

108.1

108.2

Resis

tanc

e [k

Ω]

Time [s]

Threshold

10 20 30 40 50 0 Time [ms]

60


Counting Yeast

10µm

• Sacc. Cerevisiae (5µm) • 105 cells/ml in 0.35S/m medium (with BSA) • VAC = 650mV, f0 = 100kHz, 2kSa/s

0 20 40 60 80 100 120 140 160-1.4-1.2-1.0-0.8-0.6-0.4-0.20.00.2

HP V

olta

ge [m

V]

Time [s]

Threshold


Yeast: Peak Analysis

0.1 0.2 0.3 0.4 0.5 0.60

10

20

Norm

alize

d Co

unts

[%]

∆R/R [%]

P.Palatt and G. Saidel, Ann. Biomed. Eng. 7 1979

0.5 1.0 1.5 2.0 2.5 3.0 3.50

10

20

30

40

Norm

alize

d Co

unts

[%]

Peak Duration [ms]

Pulse Width: Amplitude:

Consistent with: • cell velocity (3cm/s) • cell volume (10µm beads/8) • cell size distribution

5µm


A Novel Technique: Electrical Detection

Real-time capacitive detection of single PM particles

46

h

M. Carminati, Capacitive detection of micrometric airborne particulate matter for solid-state personal air quality monitors, Sensors Actuators A 219, 2014.


Choice of the Sensing Configuration

Key design choices: • Air vs liquid • Electrode configuration:

47

(a) (b) (c)Parallel plates Coplanar pair Interdigitated combs

t

C(t)

t(d) (e)

In-flow detectionDepositionΔC

C(t)

ΔtΔC

M. Carminati, Capacitive detection of micrometric airborne particulate matter for solid-state personal air quality monitors, Sensors Actuators A 219, 2014.


Preliminary FEM Simulations

Also in the coplanar geometry, ∆C has a volume dependence

48

1 2 3 4 5 6 7 8 910

0.1

1

10

H = 10µm G = 2µm

εr=2

∆CFE

M [a

F]

Particle Diameter D [µm]

εr=15

G

H

D

L

Resolution limit


Electrode Geometry Optimization

Fine tuning of the electrode parameters with FEM simulations

49

0 20 40 60 801000

1

2

0 5 10 15 201

2

3H = 10 D = 10L = 20εr = 2

∆CFE

M [a

F]

Electrode Gap G [µm]0 20 40

0

1

2

3

4

5

6

7 G = 10 D = 10L = 20εr = 2

(d)(c)(b)

∆CFE

M [a

F]

Particle Height H [µm]

(a)

100 101 102 103 104 1050123456789

G = 10H = 10D = 10L = 20

∆CFE

M [a

F]

Particle Permittivity εr

G = 10H = 10D = 10εr = 2

∆CFE

M [a

F]Electrode Length L [µm]

Gap: min. Height: min. Length: 30µm M. Carminati, Capacitive detection of micrometric airborne particulate matter for solid-state personal air quality monitors, Sensors Actuators A 219, 2014.


Low-Noise Detection Electronics

• Low-noise custom front-end (integrator) • Differential sensing configuration to reduce generator noise • 1.1aF resolution with 1sec response time (5aF with 1ms)

50

1GΩ

90°

Re

Im

VAC

Lock-In Demodulator ADA4817-1

CF = 1.5pF

CS

C1

C2


Static Characterization: PM10 Detection

Validation with calibrated polystyrene 10µm beads

51

∆C consistent with

simulations


Full Detection System

52

AEROSOL GENERATOR PUMP

DUMMY CHAMBER

INLET HEPA

FILTER

AIR PDMS

GLASS

POWDER

DETECTION ELECTRONICS

DETECTION CHAMBER


PM10: Talc Detection

53

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

0.8 0.40.30.20.10.60.4 00.2

∆Cap

acita

nce

[aF]

0Time [s]

Time [s]

(a) (b) (c)

30μm

Large

Small

Large Large

Small

Small

Real-time deposition tracking (∆t = 10ms) εr = 2.2


CMOS ASIC

Single-Chip Detector: Towards Smartphones

54

Smartphone

Capacitance Detection Electronics

Microcapacitor

Sensor Inlet

PM Particle

Improving the resolution by combining on the same chip: • Scaled lithography • Integrated electronics with

ZeptoFarad resolution

Targeting 1 µm PM


Cap. Resolution: Spanning Orders of Magnitude

55

10 100 10 1 100 10 1 100 1 zepto atto femto pico

MEMS Microsensors AFM CMOS Parasitics COTS

General purpose instrumentation

Dedicated discrete-components

Dedicated ASIC

M. Carminati et al. “ZeptoFarad capacitance detection with a miniaturized CMOS current front-end for nanoscale sensors”, Sens. Act. A, 172 2011.

σ [zF] BW [Hz] VAC

Analog Devices 12 0.1 10.5 M.Carminati et al. 5 1.6 6


Integrated Photonics: Context and Motivation

56

Silicon Photonics: great promise for

communications and sensing

but

Large-scale integration of several components is limited by: • lack of a consolidated/standard fabrication approach • lack of local feedback to counterbalance drifts, variations • limited integration with CMOS circuits


A Novel Transparent Light Monitoring Technique

57

Valence band

Conduction band

Traps Ener

gy

hν

hν

defects at the Si/SiO2 interface produce

surface states

F. Morichetti et al., IEEE J. Sel. Top. Quant. Electron., 20 (2014) .

At λ = 1550nm the Si waveguide is transparent but

Photogenerated carriers

Si

Si SiO2


A Novel Transparent Light Monitoring Technique

58

VAC iAC

RWG CA

CA

ContactLess Capacitive Access

F. Morichetti et al., IEEE J. Sel. Top. Quant. Electron., 20 (2014) . PATENTED!

SiO2

1 µm separation

To avoid perturbation of the radiation


Detection Architecture

Optimized electrode layout

59

Metal

SiO2

Si

CA

SiO2

CF

90°

Re[

Y]

Im[Y

]

Low-noise current amplifier Lock-in demodulator

IAC

CA

BWL

fAC

BWL

VAC

CE

CA RWG

DC Reset

Si

ASIC

Force Electrode

Sense Electrode

SiO2

Light

Photonic Platform

Cal

ibra

tion

ContactLess Integrated Photonic Probe

Wav

egui

de L

ocal

P

ower

Est

imat

e

1μm

200μm 100μm

RWG ∼150MΩ ∆G ∼nS CA = 5fF CE = 500fF

100μm

480nm

220nm

1μm

Low-noise impedance tracking platform

Cross section Top view


Non-Invasivity Enables Multipoint Monitoring

No extra light absorption

60

Multisite monitoring in complex circuits

Multichannel ASIC is required


ASIC Design

61

3.2m

m

2.2mm

• 32 channels (4 parallel acquisition chains) • 3.3V, 8mA current consumption, AMS 0.35μm CMOS

• Integrator front-end (CF = 1pF) • DC handling network with sub-threshlod transistors • 8-channel low-parasitics MUX • 2 square-wave multipliers


Performances of the Integrated Setup

62

Readout ASICPhotonic ChipRadical reduction of

parasitic capacitance CE of connections from 500fF to 2fF

Launch Fiber

ASIC

10kΩ Transimpedance

Advantages of integration: • Multipoint monitoring • Miniaturization • CMOS compatibility • Reduction of CE → lower noise

2pS rms

25x


ASIC Characterization

63

100MHz bandwidth Same SNR of external lock-in

External lock-in

On-chip demodulator

640nV rms

565nV rms


CLIPP Characterization: Power Monitor

Monitor response curve

64

1k 10k 100k 1M 10M

1p

10p

100p

1n

-15dBm -10dBm -5dBm 0dBm +5dBm +10dBm

∆Adm

ittan

ce M

agni

tude

[S]

Frequency [Hz]10-5 10-4 10-3 10-2

10-1

100

∆G [n

S]

Local Power [W]

Admittance spectroscopy for a wide range of optical power levels

Min power: -30dBm, 40dB dynamic range


Looking Inside a Ring Resonator

65

100μm

Thermo-optic Actuator

CLIPP

50.6

50.8

51.0

51.2

1549 1550 155150.6

50.8

51.0

51.2

+10dBm +5dBm 0dBm -5dBm -10dBm -15dBm

Y [n

S](b)

2.5V 3V 3.5V 4V

0 0.5V 1V 1.5V 2V

Y [n

S]

Wavelength [nm]

(a)

No perturbative effect of VAC on the resonance!

Tracking input power

Tracking actuator voltage


Closing the Loop

66

Fo

rce

Heater Bias+

IntegralController

ASICFront-End

DitherSe

nse

f2

S. Grillanda et al., Optica., 1 3 (2014)


Summary

67

• Not all the impedance measuring techniques are equivalent in terms of noise and suitability for impedance spectroscopy.

• When measuring impedance at the microscale, the signals decrease and noise (unavoidable property of the electronic components) becomes relevant.

• The current sensing front-end coupled with a lock-in demodulator represents the best solution.

• The input capacitance must be minimized, paying attention to short coax cables and preferably insulating substrates.

• The transimpedance amplifier allows neutralization of the input stray impedance, but is prone to a noise/bandwidth trade-off, that can be relaxed by means of advanced topologies such as the integrator/differentiator cascade.


References

68

• M. Carminati, G. Ferrari, D. Bianchi, M. Sampietro, Impedance spectroscopy for biosensing: circuits and applications, in Handbook of Biochips (Editor M. Sawan), Springer in press.

• M. Carminati et. al. “Theoretical and Experimental Comparison of Microelectrode Sensing Configurations for Impedimetric Cell Monitoring” pp. 75-82, in Lecture Notes on Impedance Spectroscopy: Vol. 4, CRC Press, 2013.

• M. Crescentini, M. Bennati, M. Carminati, M. Tartagni, “Noise Limits of CMOS Current Interfaces for Biosensors: a Review”, IEEE Trans. Biomed. Circuits and Systems, vol. 8, no. 2, pp. 278-292, 2014.

• M. Carminati, G. Ferrari, D. Bianchi, M. Sampietro, “Femtoampere integrated current preamplifier for low noise and wide bandwidth electrochemistry with nanoelectrodes” Electrochimica Acta, vol. 112, pp. 950–956, 2013.

• M. Carminati, M. Vergani, G. Ferrari, L. Caranzi, M. Caironi, M. Sampietro “Accuracy and resolution limits in quartz and silicon substrates with microelectrodes for electrochemical biosensors”, Sens. Act. B, vol. 174, pp. 168-175, 2012.

• M. Carminati, G. Ferrari, F. Guagliardo, M. Sampietro, "ZeptoFarad capacitance detection with a miniaturized CMOS current front-end for nanoscale sensors", Sens. Act. A, vol. 172, pp.117-123, 2011.

• M. Carminati, G. Ferrari and M. Sampietro, "attoFarad resolution potentiostat for electrochemical measurements on nanoscale biomolecular interfacial systems“, Rev. Sci. Instrum., vol. 80, no. 12, pp. 124701/1-10, 2009.


Acknowledgments

Work supported by Fondazione Cariplo Progetto MINUTE

69

and by EU under FP7 project «BBOI» Breaking the Barriers of Optical Integration

Co-workers: M. Sampietro, G. Ferrari, P. Ciccarella, G. Gervasoni

S. Grillanda, F. Morichetti, A. Melloni

. . . Minute

.

Presentazione di PowerPoint · Titolo presentazione Nome relatore Tutorial 4 Low-Noise Impedance...

Documents

Transcript of Presentazione di PowerPoint · Titolo presentazione Nome relatore Tutorial 4 Low-Noise Impedance...