A low cost approach to acoustic filters acting as GPR cooperative targets for passive sensing

(1)

approach to acoustic filters acting as GPR cooperative

targets for passive sensing J.-M Friedt & al.

jmfriedt@femto- st.fr What are RADAR cooperative targets ? Why GPR cooperative targets ? Radiofrequency acoustic transducers as cooperative targets Software optimization for dual sub-surface and sensor monitoring Low cost demonstration using off-the shelf radiofrequency filters Conclusion and perspectives

A low cost approach to acoustic filters acting as GPR cooperative targets for passive sensing

J.-M Friedt & A. Hugeat

G. Martin, S. Alzuaga, T. Baron, S. Ballandras FEMTO-ST Time & Frequency, Besan¸con, France

Contact: jmfriedt@femto-st.fr

References and slides athttp://jmfriedt.free.fr

October 23, 2015

(2)

Ground Penetrating RADAR (GPR)

• permittivity and conductivity changes [1]

v = √_µε ¹

2

q 1+ ^σ²

ω2ε2+1 1/2

• dielectric interface:

Fresnel reflection coef.

R=^√_ε

ice−√ ε_rock

√ε_ice+√ ε_rock

2

'

−19 dB

• v ' ^√^c_ε

r:

33 (water)-170 (ice) m/µs

• record 200 ns-5µs

• '1000 samples/trace

[1] G. Leucci,Ground Penetrating Radar: the Elec- tromagnetic Signal Attenuation and Maximum Pen- etration Depth, Scholarly Research Exchange2008, doi: 10.3841/2008/926091

ISTerre – Grenoble

(3)

Ground Penetrating RADAR (GPR)

v = √_µε ¹

2

q 1+ ^σ²

ω2ε2+1 1/2

R=^√_ε

ice−√ ε_rock

² '

−19 dB

• v ' ^√^c_ε

r:

(4)

Ground Penetrating RADAR (GPR)

v = √_µε ¹

2

q 1+ ^σ²

ω2ε2+1 1/2

R=^√_ε

ice−√ ε_rock

² '

−19 dB

• v ' ^√^c_ε

r:

(5)

Results

GPR tracks and depth mesurements interpolated bedrock topography¹

1A. Saintenoy, J.-M. Friedt, & al.,Deriving ice thickness, glacier volume and bedrock morphology of the Austre Lov´enbreen (Svalbard) using Ground-penetrating

(6)

Outline

• What are RADAR cooperative targets ?

• Why GPR cooperative targets ?

• Radiofrequency acoustic transducers as cooperative targets

• Software optimization for dual sub-surface and sensor monitoring

• Low cost demonstration using off-the shelf radiofrequency filters

• Issue of varying operating frequency with soil permittivity, and solution

(7)

jmfriedt@femto- st.fr

What are RADAR cooperative targets ? Why GPR cooperative targets ? Radiofrequency acoustic transducers as cooperative targets Software optimization for dual sub-surface and sensor monitoring Low cost demonstration using off-the shelf radiofrequency filters Conclusion and perspectives

What are RADAR cooperative targets ?

1 target whose backscattered signal is representative of its state (identification, measurement)

2 active targets: radar beacons (racon), IFF

3 passive targets: buried dielectric reflectors, L¨uneberg spheres

4 this work: use of radiofrequency transducers based on surface acoustic wave propagation (RF filters)

H. Stockman,Communication by means of reflected power Proc. IRE36(Oct. 1948) pp.1196–1204

(picture fromhttp://geogdata.csun.edu/~aether/pdf/volume_05a/rosol.pdf)

(8)

What are RADAR cooperative targets ?

A. Glinsky,Theremin: Ether Music And Espionage, Uni- versity of Illinois Press (2005)

P. Wright & P. Greengrass,Spycatcher(1987)

http://madmikesamerica.com/2010/08/the-thing-and-the-curious-life-of-leon-theremin/thing2/

(9)

What are RADAR cooperative targets ?

A. Glinsky,Theremin: Ether Music And Espionage, Uni- versity of Illinois Press (2005)

P. Wright & P. Greengrass,Spycatcher(1987)

(10)

What are RADAR cooperative targets ?

D. J. Thomson, D. Card, and G. E. Bridges, RF Cavity Passive Wireless Sensors With Time-Domain Gating-Based Interrogation for SHM of Civil Structures, IEEE Sensors Journal. 9(11) (Nov. 2009), pp.1430-1438

(11)

What are RADAR cooperative targets ?

C.T. Allen, S. Kun, R.G Plumb, The use of ground- penetrating radar with a cooperative target, IEEE Trans- actions on Geoscience and Remote Sensing,36(5) (Sept.

1998) pp. 1821– 1825

(12)

Why GPR cooperative targets ?

• Complement buried structure information with physical quantity measurement

• Accessible quantities: temperature, stress (pressure), identifier (ID)

• Measurement range: depends on RADAR cross section – typical losses in the 30-40 dB range

Passive cooperative target (no local battery source) for extended life expectancy (limited by packaging)

Envisioned applications:

• pipe tagging/stress/temperature

• soil moisture²

• concrete temperature³

2L. Reindl& al.,Radio-requestable passive SAW water-content sensor, IEEE Trans. Microwave Theory & Techniques,49(4), (Apr. 2001), pp. 803–808

3J. Kim, R. Luis, M.S. Smith, J.A. Figueroa, D.C. Malocha, B.H. Nam,Concrete temperature monitoring using passive wireless surface acoustic wave sensor system Sensors and Actuators A224(Febr. 2015), pp.131–139

(13)

General strategy

Differentiate the sensor response from clutter (passive interface reflexions)

1 load a resonator which slowly releases energy (time constant

Q/(πf0)τclutter): 20·log₁₀(e) = 8.7 but such a strategy is poorly suited to GPR (short pulse unable to load transducer)

2 delay sensor response beyond furthest possible passive reflectors

−8.7 dB/τ

returned power (dB)

t

reflected signals and clutter

sensor response

sensor measurement 3 shrink diemensions by converting the electromagnetic wave

('200 m/µs) to an acoustic wave ('4000 m/s) confined to the surface of a piezoelectric substrate (surface acoustic wave transducers)

4 well known mechanism in radiofrequency signal processing

5 acoustic wave propagation is dependent on substrate properties !

(14)

SAW basics

Demonstrated with dedicated readers for Surface Acoustic Wave (SAW) transducers (Transense, CTR, Sengenuity, RSSI, SenSanna, Frec|n|sys, SENSeOR).

M1

M2 M3 IDT

typical dimensions:

10 mm x 4 mm

@ 100 MHz

• Sensor: ϕ= 2π·D/λ= 2πD·f/c withdc/c(T, σ) known

• GPR is wideband pulse generator⇒delay line architecture.

• This work aims at providing GPR-compatible SAW sensors ...

• ... or divert existing SAW filters for sensing purpose.

(15)

Radiofrequency acoustic (SAW) transducers as cooperative targets

• from a user perspective, an electrical dipole

• physics: conversion of incoming RADAR pulse to an acoustic (mechanical) wave through inversepiezoelectric effect

• acoustic wave propagates on the surface of the piezoelectric substrate at a speed of 3000-5000 m/s

• the acoustic wave is reflected by patterned mirrors, reaches the transducer and is converted back to an electromagnetic pulse (RADAR echo)⇒tagorsensor

• design challenge: match sensor transfer function to incoming pulse spectra

Piezoelectricity islinear process: returned power is a fraction of incoming power, no threshold

Unlike RFID which requires powerwhilebeing operated, SAW transducers are ideally

(16)

Radiofrequency acoustic (SAW) transducers as cooperative targets

• from a user perspective, an electrical dipole

• physics: conversion of incoming RADAR pulse to an acoustic (mechanical) wave through inversepiezoelectric effect

• acoustic wave propagates on the surface of the piezoelectric substrate at a speed of 3000-5000 m/s

• the acoustic wave is reflected by patterned mirrors, reaches the transducer and is converted back to an electromagnetic pulse (RADAR echo)⇒tagorsensor

• design challenge: match sensor transfer function to incoming pulse spectra

Piezoelectricity islinear process: returned power is a fraction of incoming power, no threshold

Unlike RFID which requires powerwhilebeing operated, SAW transducers are ideally suited for GPR interrogation

(17)

Software optimization for dual sub-surface and sensor monitoring

• Separate sensor response from subsurface reflectors by delaying echo beyond any possible interface

reflection (2µs = 4 mm long sensor at 4000 m/s)

• Short term response = buried structures (“clutter”)

• Long term response = buried sensor Available GPR only allows for a small number of samples (<5000) ⇒ long duration measurement must be at slow sampling rate

⇒lost resolution in the shallow region

two-way travel time (ns)

trace number (a.u.) 100

200

300

400

500

10 20 30 40 50

100

200

300

400

500

10 20 30 40 50

4

4J.-M Friedt& al.,Surface Acoustic Wave Devices as Passive Buried Sensors, J.

(18)

Software optimization for dual sub-surface and sensor monitoring

• Long term response = buried sensor Solution: define two time windows, one fo- cusing on the subsurface structures (e.g. 0- 200 ns) and another on the sensor (e.g.

1000-1200 ns)

(19)

Software optimization for dual sub-surface and sensor monitoring

• Long term response = buried sensor Only software re-design, hardware re- mains the same(reconfigure start delay of the stroboscopic measurement)

⇒demonstrated on Mal˚a ProEx GPR

https://sourceforge.net/projects/proexgprcontrol/

(20)

Low cost demonstration using off-the shelf radiofrequency filters

• Dedicated transducers require cleanroom fabrication capability able to handle piezoelectric substrates

• Design considerations:

• transducer transfer function must match GPR pulse spectrum

• delay must be within measurement range of GPR

• reflective transducer operation (most filters operate in transmission) Measurement example, using an TDK/Epcos B3607 filter:

-8 -7 -6 -5 -4 -3 -2 -1 0

100 120 140 160 180 200 220 240 260

|S11| (dB)

frequency (MHz)

0 0.005 0.01 0.015 0.02 0.025 0.03

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

returned power (a.u.)

time (us) ISTerre – Grenoble

(21)

Low cost demonstration using off-the shelf radiofrequency filters

-8 -7 -6 -5 -4 -3 -2 -1 0

100 120 140 160 180 200 220 240 260

|S11| (dB)

frequency (MHz)

0.005 0.01 0.015 0.02 0.025 0.03

power supply (heater) resistance

TX

RX Pt100 probe

monitoring

100 MHz unshielded antennas

SAW sensor

(22)

Low cost demonstration using off-the shelf radiofrequency filters

0 100 200 300 400 500 600 700 800

trace number (a.u.) 2.75

2.8 2.85 2.9 2.95 3

time (us)

-5 0 5 10 15

0 100 200 300 400 500 600 700 800

unwrapped phase (rad)

time (s)

25 30 35 40 45 50

0 100 200 300 400 500 600 700 800

temperature (degC)

time (s) ISTerre – Grenoble

(23)

Signal processing steps

Converting the echo delays to a meaningful measurement:

1 multiple echoes from the sensor to get rid of RADAR-target distance dependence

2 phase measurement of each returned echo: cross-correlation

3 for single echo (eg SAW filter):

phase of the Fourier transform of the main spectrum

component (assuming constant RADAR-target distance)

4 SAW delay lines will usually be more narrowband than RADAR pulse ⇒long echo

⇒improved SNR

5 sub-sampling period resolution:

polynomial fitof the cross-correlation maximum

6 danger: multiple nearby cross-correlation peaks⇒risk of 2π rotation: ϕ= 2πD/λ= 2πD·f/c⇒dϕT =−2πDf/c·dcT/c wheredcT/c(LiNbO3) = 60 ppm/K. IfD= 1 cm, 2πphase rotation

(24)

Signal processing steps

Converting the echo delays to a meaningful measurement:

1 multiple echoes from the sensor to get rid of RADAR-target distance dependence

2 phase measurement of each returned echo: cross-correlation

3 for single echo (eg SAW filter):

phase of the Fourier transform of the main spectrum

component (assuming constant RADAR-target distance)

4 SAW delay lines will usually be more narrowband than RADAR pulse ⇒long echo

⇒improved SNR

5 sub-sampling period resolution:

polynomial fitof the cross-correlation maximum

6 danger: multiple nearby cross-correlation peaks⇒risk of 2π rotation: ϕ= 2πD/λ= 2πD·f/c⇒dϕT =−2πDf/c·dcT/c wheredcT/c(LiNbO3) = 60 ppm/K. IfD= 1 cm, 2πphase rotation on LiNbO3/128 for35 Kat 100 MHz (cLNO128'4000 m/s).

-10 -5 0 5 10

0 20000 40000 60000 80000 100000

phase (rad)

time (s)

72 ns 232 ns

20 30 40 50 60 70 80 90 100

0 20000 40000 60000 80000 100000

temp. (degC)

time (s) time difference

2450 MHz LiNbO3/128 device

(25)

Dedicated software for GPR recording

• Run from (low power) embedded board for long term monitoring of a single spot

• Multiple windows separate in time

• Combination with custom hardware (magnetometer, GPS positioning ...) beyond those provided by Mal˚a

• Custom signal processing (GNU/Octave)

(26)

RADAR pulse central frequency

+350 V

(dipole)emission antenna

HBAR sensor

0 200000 400000 600000 800000 1e+06

0 100 200 300 400 500

FFT magnitude (a.u.)

frequency (MHz)

glacier ice (1) frozen moraine (250)visible water (500)snow ? (315)

two-way time (ns)

trace number (a.u.) -10

0 10 20 30 40 50 60 70

0 100 200 300 400 500 600

ice-rock interface

moraine water soaked moraine

• issue with RADAR pulse generator: central frequency dependent on medium

• in avalanche transistor approach, the capacitor unloads with a time constant given by min(S11).

• f = ^c⁰

2d×√

εr(eff) where ε_r(eff)' ^ε^r₂⁺¹ (ε_r ∈[3..15])

⇒f can vary by a factor of 2

Could we design a transducer operating in all environmental conditions ?

(27)

HBAR for sensors compatible with multiple wavelength antennas

• Stack of an active

piezoelectric thin layer over a low acoustic loss substrate

• Mode spacing given by substrate thicknesst_s: c_s/(2×t_s)×N,N∈N

• Comb of acoustic modes (BAW) associated with the multiple overtones of the standing wave

• Fourier transform of comb is a time domain comb of echoes

• Broad frequency range (up to one decade) to operate in all GPR environments

• BUT loss of interrogation ^-14

-12 -10 -8 -6 -4 -2 0

|S11| (dB)

(28)

HBAR for sensors compatible with multiple wavelength antennas

• Stack of an active

piezoelectric thin layer over a low acoustic loss substrate

• Envelope given by

piezoelectric layer thickness t_p: c_p/(2×t_p)×(2N+ 1)

• Comb of acoustic modes (BAW) associated with the multiple overtones of the standing wave

• Fourier transform of comb is a time domain comb of echoes

• Broad frequency range (up to one decade) to operate in all GPR environments

• BUT loss of interrogation range?! (35 m→1.5 m) ^-400

-200 0 200 400 600

0 5e-07 1e-06 1.5e-06 2e-06 2.5e-06

time (s) ISTerre – Grenoble

(29)

Wireless measurement: time domain approach

1 Radargram, alternatively with & without HBAR

2 Cross-correlate adjacent echoes for precise phase extraction

3 One to one relationship between phase and temperature

-6000 -4000 -2000 0 2000 4000 6000

0 200 400 600 800 1000 1200

received signal (a.u.)

time (ns)

trace number (a.u.) 200

400

600

800

1000

10 20 30 40 50 60 70

-2e+07 -1.5e+07 -1e+07 -5e+06 0 5e+06 1e+07 1.5e+07 2e+07

0 50 100 150 200 250 300 350 400

cross correlation

time (a.u.)

trace 69 trace 45 -6000

-4000 -2000 0 2000 4000 6000

2e-07 4e-07 6e-07 8e-07 1e-06 1.2e-06 1.4e-06 1.6e-06

time (s)

(30)

Wireless measurement: time domain approach

1 Radargram, alternatively with & without HBAR

2 Cross-correlate adjacent echoes for precise phase extraction

3 One to one relationship between phase and temperature

20 30 40 50 60 70 80 90 100

100 200 300 400 500 600 700 800 900

Pt100 probe temperature (degC)

time (s) 167.8

167.9 168 168.1 168.2 168.3 168.4 168.5

0 100 200 300 400 500 600 700 800 900

cross correlation maximum position (pixel) time (s)

167.8 167.9 168 168.1 168.2 168.3 168.4 168.5

20 30 40 50 60 70 80 90 100

cross correlation maximum position (pixel)

temperature (degC) 5.738e−3 pixel/degC

Port1 Port2 HBAR electrode layout

(31)

HBAR figure of merit

• In the RADAR equation P_R

PE

= G²·λ²σ (4π)³·d⁴ the cross sectionσis replaced with insertion lossesλ²/IL

• IL∝k²so coupling coefficient must be maximized

• butPk²of each mode within each piezo layer envelope =k²of the piezoelectric substrate

• k²of piezo layer overtoneN decreases asN²(cf BAW)

• ⇒tradeoff between maximizing k²of each overtone andQ to separate the modes

• figure of merit: Q×k²

0. 050. 10 0. 150. 2 0. 250. 3 0. 35

100 200 300 400 500 600 700 800

admittance(S)

f r eq ( MHz)

5000 10001500 20002500 30003500 4000

100 200 300 400 500 600 700 800

Q(nounit)

f r eq ( MHz)

0 0. 001 0. 002 0. 003 0. 004 0. 005

100 200 300 400 500 600 700 800

k2(nounit)

f r eq ( MHz)

0. 501 1. 52 2. 53 3. 5

100 200 300 400 500 600 700 800

qxk2(nounit)

f r eq ( MHz)

- 0. 20. 20. 40. 60. 801

0 50 100 150 200 250 300 350 400

Re(X)/max(X) Z

Y

5000 10001500 20002500 30003500 4000

0 50 100 150 200 250 300

Q

0. 0020 0. 004 0. 006 0. 0080. 01 0. 012 0. 014

0 50 100 150 200 250 300

K2

0 10 20 30 40 50

Q*K2

(32)

HBAR figure of merit

• keepingQ×k²large requires operating on the fundamental piezo layer mode⇒thin piezo film

• few modes within each piezo layer envelope⇒reduce low loss substrate thickness ...

• butkeep the multimode aspect so substrate thick enough to provide one mode every

5-10 MHz (0.1-0.2µs): 500µm thick sapphire or LNO substrate

- 0. 20. 20. 40. 60. 801

0 200 400 600 800 1000 1200

Re(X)/max(X) Z

Y

0 2000 4000 6000 8000 10000

0 200 400 600 800 1000 1200

Q

0 0. 02 0. 04 0. 06 0. 08 0. 1

0 200 400 600 800 1000 1200

K2

1000 200300 400500 600700

0 200 400 600 800 1000 1200

Q*K2

f r equency ( MHz)

0 0. 002 0. 004 0. 006 0. 008 0. 01

400 500 600 700 800 900 100011001200

K2

f r equency ( MHz)

- 0. 20. 20. 40. 60. 801

0 50 100 150 200 250 300 350 400

Re(X)/max(X) Z

Y

22002400 26002800 30003200 3400 3600

0 50 100 150 200 250 300

Q

0 0. 02 0. 04 0. 06 0. 080. 1

0 50 100 150 200 250 300

K2

0 50 100 150 200 250

0 50 100 150 200 250 300

Q*K2

f r equency ( MHz)

(33)

How not to implement extrinsic load

• Load in parallel to the antenna will kill the resonator contribution

• Measurement range loss in some range of the measurement

• Solution: coupled resonator separates load capability from transduction capability.

Jun Seop Lee& al.,Wireless Hydrogen Smart Sensor Based on Pt/Graphene

(34)

Extrinsic load effect on link budget

• One port dedicated toelectro- magnetic transduction, connected to antenna

• One port dedicated toload effect (capacitive, resistive)

• Change load impedance⇒change phase of returned signal and magnitude

Γ = ^V_V^R

I =^Z_Z^L^−Z⁰

L+Z0 ∈C;

ifZ0∈RandZL∈iR,|Γ|= 1 and ϑ=±π−2 arctan(ZL/Z0) ifZL≷0

• Also applicable to the delay line approach by loading one of the mirrors (minor effect) ⁴

4L. Reindl& al.,Radio-Requestable Passive SAW Water-Content Sensor, IEEE Trans. on

Microwave Theory & Techniques49(4), 2001, pp.803–