Subsurface wireless chemical sensing
Subsurface wireless chemical sensing strategy compatible with Ground Penetrating RADAR
J.-M Friedt & al Outline Acoustic transducer
Digital is good, analog is bad (TM): timing generator for RADAR systems
Chemical functionalization Sensor measurement
J.-M Friedt1,3 , A. Hugeat1 , S. Lamare2 , F. Ch´erioux2 , D. Rabus3 , G. Goavec-M´erou3 , L. Arapan3 , S. Alzuaga3
Timebase stability Digital ramp synthesis
[email protected], slides at jmfriedt.free.fr/#jmf16n
Conclusion
1
2
FEMTO-ST/Time & Frequency, Besan¸con, France FEMTO-ST/Micro-Nano Sciences & Systems, Besan¸con, France 3 SENSeOR SAS, Besan¸con, France
1 / 12
Subsurface wireless chemical sensing
Outline
J.-M Friedt & al Outline
Subsurface chemical sensing system:
Acoustic transducer Chemical functionalization
t
TX RX
Sensor measurement Timebase stability Digital ramp synthesis
cavity
Conclusion
? [H 2S] 1
2
3 4
acoustic transducer acting as cooperative target (separate sensor echo from clutter) transducer functionalization for chemical sensing: polymer formulation sensor measurement using Groung Penetrating Radar (GPR) sampling rate stability issue and solution for the Mal˚ a ProEx 2 / 12
Subsurface wireless chemical sensing
Outline
J.-M Friedt & al Outline
Subsurface chemical sensing system: δ t~[H 2S]
Acoustic transducer Chemical functionalization
t
Sensor measurement
TX RX
Timebase stability Digital ramp synthesis
cavity
Conclusion
? [H 2S] 1
2
3 4
acoustic transducer acting as cooperative target (separate sensor echo from clutter) transducer functionalization for chemical sensing: polymer formulation sensor measurement using Groung Penetrating Radar (GPR) sampling rate stability issue and solution for the Mal˚ a ProEx 2 / 12
Subsurface wireless chemical sensing
Surface Acoustic Wave transducers as cooperative targets
Outline Acoustic transducer Chemical functionalization
1
2
Sensor measurement Timebase stability Digital ramp synthesis Conclusion
3
4
A RADAR emits an electromagnetic pulse, the cooperative target delays the pulse beyond clutter delay for identifying the transducer response (TDMA) the transducer becomes a sensor if the time delay is dependent on a known quantity (temperature, chemical compound concentration)
0 −20 loss (dB)
J.-M Friedt & al
−40
1−1.5 us delay line FSPL ~ 1/d^4
FSPL sub−surface interfaces = clutter
sensor measurement
−60 −80
delay line
receiver noise level
−100 0
0.5
1 time (us)
1.5
2
shrink delay line dimensions by converting the electromagnetic wave to a surface acoustic wave by using a piezoelectric substrate (acoustic velocity 105 times slower than electromagnetic velocity) 3 / 12
Subsurface wireless chemical sensing J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
Chemical functionalization • Physical transducer: conversion of adsorbed mass to velocity
variation to time of flight variation ⇒ no selectivity • Chemical sensor: adlayer selective to a single compound • Amongst the quantities inducing acoustic velocity variation
(temperature, stress), boundary conditions will define the acoustic velocity • “Microbalance” application of
acoustic transducers: the thicker the loading layer, the slower the wave – sensitivity S A A 1 = ∆v = ∆v S = ∆f f ∆m v ∆m v ρdt ∆v 2 S ' 200 cm /g ⇒ v ' 200 ppm if ∆m = 1 µg/cm2 A
• well known technique for biosensor
applications, but can it be used 8 × 6 mm2 lithium niobate chip, 20 µm for subsurface wireless sensing ? wavelength @ 200 MHz = 5 µm electrodes 4 / 12
Subsurface wireless chemical sensing J.-M Friedt & al
Chemical functionalization
Outline Acoustic transducer Chemical functionalization
• Physical transducer: conversion of adsorbed mass to velocity
variation to time of flight variation ⇒ no selectivity • Chemical sensor: adlayer selective to a single compound
Sensor measurement
• Amongst the quantities inducing acoustic velocity variation
Timebase stability
(temperature, stress), boundary conditions will define the acoustic velocity • “Microbalance” application of d/c T1 T2 T3 acoustic transducers: the thicker the loading layer, the slower the wave – sensitivity S
Digital ramp synthesis Conclusion
A A 1 = ∆v = ∆v S = ∆f f ∆m v ∆m v ρdt S ' 200 cm2 /g ⇒ ∆v ' 200 ppm if v ∆m 2 = 1 µg/cm A
V M1
IDT
M2
M3
pulse
t
• well known technique for biosensor
applications, but can it be used for subsurface wireless sensing ?
2d/c+T1 echo M1
2d/c+T2 echo M2
2d/c+T3 echo M3
4 / 12
Subsurface wireless chemical sensing
Chemical functionalization
J.-M Friedt & al Outline Acoustic transducer
• Physical transducer: conversion of adsorbed mass to velocity
variation to time of flight variation ⇒ no selectivity
Chemical functionalization
• Chemical sensor: adlayer selective to a single compound
Sensor measurement
• Amongst the quantities inducing acoustic velocity variation
Conclusion
−300
• “Microbalance” application of
acoustic transducers: the thicker the loading layer, the slower the wave – sensitivity S A 1 = ∆v = ∆v S= v ∆m v ρdt ∆v 2 S ' 200 cm /g ⇒ v ' 200 ppm ∆m = 1 µg/cm2 A
signal (u.a.)
Digital ramp synthesis
(temperature, stress), boundary conditions will define the acoustic velocity −400 −500 −600 −700 −800 0
∆f A f ∆m
0.1
0.2
0.3 0.4 time (us)
0.5
0.6
0.7
xcorr max
if
4 xcorr(a.u.)
Timebase stability
2 0
• well known technique for biosensor −2 −4 applications, but can it be used for subsurface wireless sensing ? −6
5.7 ns=1/175 MHz 0.26
0.28
0.3 0.32 time delay (us)
0.34
0.36
4 / 12
Subsurface wireless chemical sensing
Chemical functionalization
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement
Based on a well known chemical reaction, design a polymer whose formulation 1
1
Timebase stability Digital ramp synthesis Conclusion
allows for spreading a homogeneous layer with thicknesses of the order of the wavelength (' µm) ... with deposition technique compatible with wafer-scale processing (cleanroom),
2
includes as many sensing sites as possible (low weight polymer matrix),
3
is selective to the targeted compound and rejects interfering molecules.
In this case, hydrogen sulfide (H2 S) looks like water (H2 O), but includes a sulfur reacting with heavy metals (thiolation) ⇒ introduce a reactive heavy metal ion in the polymer matrix 5 / 12
Subsurface wireless chemical sensing J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
Sensor measurement Chemical sensing induces a time delay variation of a few tens of ps: (R − COO)2 Pb + H2 S → 2RCOOH + PbS R: functional alkyl chain PbS nanoparticles (black): visual indicator of reaction Wafer scale functionalization by spincoating the synthesized polymer dedicated to H2 S detection: challenge of uniform spreading sub-µm thick polymer wafer. A • Sensitivity S = dff dm ' 200 cm2 /g (wave property) 3 • ρpolymer ' 1 g/cm & Mpolymer = (131 × 2 + 207) g/mol &
t ' 0.2 µm ⇒ R = ρ/Mpolymer · t=43 nmol/cm2 receptor density • MH2S = 34 g/mol⇒ absorbed mass per unit area: R × MH2S = dm/A ' 1500 ng/cm2 6 / 12
Subsurface wireless chemical sensing
Sensor measurement
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization
• Periodic signal delay measurement as a phase shift • One period τ (5 ns @ 200 MHz) is one full phase rotation 360◦ •
dϕ ϕ
=
df f
⇒ dϕ = ϕ · S ·
Sensor measurement
dm A
◦ = 2πf τ · S · dm A = 20 or 280 ps @ 200 MHz (1.5 µg/cm2 , 1 µs)
Timebase stability
10
Conclusion
phase (deg)
Digital ramp synthesis
H2S
0 -10 -20 -30 -40 -50 0
200
400
600
800
1000
time (s)
Velocity variation = phase variation through ϕ = 2πfd/c with c varying Need to measure time delays with sub-100 ps long term accuracy 7 / 12
Subsurface wireless chemical sensing
Sensor measurement
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization
• Periodic signal delay measurement as a phase shift • One period τ (5 ns @ 200 MHz) is one full phase rotation 360◦ •
dϕ ϕ
=
df f
⇒ dϕ = ϕ · S ·
Sensor measurement
dm A
◦ = 2πf τ · S · dm A = 20 or 280 ps @ 200 MHz (1.5 µg/cm2 , 1 µs)
Timebase stability
140
Digital ramp synthesis
delay (ps)
Conclusion
H2S
0 -140 -280 -420 -560 -700 0
200
400
600
800
1000
time (s)
Velocity variation = phase variation through ϕ = 2πfd/c with c varying Need to measure time delays with sub-100 ps long term accuracy 7 / 12
Subsurface wireless chemical sensing
Timebase stability
J.-M Friedt & al
Static environment, controlled temperature and fixed sensor – time delay as cross correlation between two echo signals:
Timebase stability
0
-200
0
0.2
-300
0.1
-400
0.2
0.4 time (us)
Sensor measurement
fs=6834 MHz
fs=2733.6 MHz
Chemical functionalization
Digital ramp synthesis
0.6
time (us)
Acoustic transducer
Stroboscopic signal generation and cause of drift
-500
1
-600
0.4
1.2
-700
0.5
1.4
-800 0 1000 2000 3000 4000 5000 6000 trace number (s)
Conclusion
• Two echoes
0.3
0.8
0
1000 2000 3000 4000 5000 6000 trace number (s)
delay (cross correlation) between echoes in a stable environment
time (us)
1 3.2
-400
1.1
-500
1.2 1.3
-600
(differential measurement) separated by 300 ns, at 1.3 and 1.6 µs • measure time
5.6 -300
delay (ns)
Outline
2.8
1.4
1.4 -700 0.0 0 1000 2000 3000 4000 5000 6000 trace number (s)
0
1000 2000 3000 4000 5000 6000 7000 trace number (s)
5 ns drift at 300 ns delay = 1.5 %=15000 ppm 200 ppm mass sensitivity or 70 ppm/K T sensitivity
8 / 12
Subsurface wireless chemical sensing
Temperature dependence of the timebase
J.-M Friedt & al Outline
Sensor measurement
fs = 1/dt =
Timebase stability
Vsupply τslow 1 B 1 = · · A δt τRC Vslow δt
Transmitter (hi voltage supply + avalanche transistor)
8
6
Receiver (RF amplifier + A/D converter)
ramps
reset
δt
Tx Rx
TX/RX
freezer
Tx pulse
(freezer) warm−up
10
ProEx control unit timebase generation: voltage to time converter optical fiber link
12
delay (ns)
with an integrator of a constant voltage:
ethernet to personal computer
14
Digital ramp synthesis Conclusion
1
Data
Chemical functionalization
Stroboscpic signal generator voltage to time converter
ADC trig.
Acoustic transducer
dt
2dt
3dt
4dt
...
reset MAX334
+V
C5
4 x1
− −
AD843
2 2000
4000
6000
8000 10000 12000 14000 16000 18000 time (s)
+V
+
AD790 DAC
+
cmp
1 B.A.T.
Johansson, Ground Penetrating RADAR array and timing circuit, Patent US 6496137 (2002)
9 / 12
Subsurface wireless chemical sensing
Temperature dependence of the timebase
J.-M Friedt & al Outline
Sensor measurement
fs = 1/dt =
Timebase stability
Vsupply τslow 1 B 1 = · · A δt τRC Vslow δt
freezer
Tx pulse
(freezer) warm−up
10
ProEx control unit timebase generation: voltage to time converter optical fiber link
12
delay (ns)
with an integrator of a constant voltage:
ethernet to personal computer
14
Digital ramp synthesis Conclusion
1
Transmitter (hi voltage supply + avalanche transistor)
8
6
Data
Chemical functionalization
Stroboscpic signal generator voltage to time converter
ADC trig.
Acoustic transducer
Receiver (RF amplifier + A/D converter)
4
2 2000
4000
6000
8000 10000 12000 14000 16000 18000 time (s)
Operational amplifier rises to > 50◦ C, inducing offset and drift in surrounding passive components
1 B.A.T.
Johansson, Ground Penetrating RADAR array and timing circuit, Patent US 6496137 (2002)
9 / 12
Subsurface wireless chemical sensing
Temperature dependence of the timebase
fs = 1/dt =
Timebase stability
freezer
Tx pulse
(freezer) warm−up
10
ProEx control unit timebase generation: voltage to time converter optical fiber link
12
delay (ns)
Vsupply τslow 1 B 1 = · · A δt τRC Vslow δt
ethernet to personal computer
14
Digital ramp synthesis Conclusion
with an integrator of a constant voltage:
Transmitter (hi voltage supply + avalanche transistor)
8
6
Receiver (RF amplifier + A/D converter)
4
2 2000
4000
6000
8000 10000 12000 14000 16000 18000 time (s)
δt reset
Sensor measurement
1
Data
Chemical functionalization
Stroboscpic signal generator voltage to time converter
ADC trig.
Acoustic transducer
ramps
Outline
Tx Rx
TX/RX
J.-M Friedt & al
dt
2dt
3dt
4dt
...
The (user controlled) slow ramp is generated by a Digital to Analog converter (voltage ref + quartz synchronized), little risk of drift ⇒ what about the fast ramp generation ?
1 B.A.T.
Johansson, Ground Penetrating RADAR array and timing circuit, Patent US 6496137 (2002)
9 / 12
Subsurface wireless chemical sensing
Solution: replace analog timebase with digital timebase
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor 5 measurement Timebase stability
Solution: replace analog timing generator (drifting integrator capacitor) with digital ramp generator ⇒ 100-fold improvement (4 ns → 34 ps) 27 pF Vishay NPO frequency synthesizer
4
delay (ns)
Digital ramp 3 synthesis Conclusion
2
1 std()=34 ps=1.6 K=320 ng/cm2 0 0
500
1000 time (s)
1500
2000
Tektronic arbitrary waveform generator configured for a ramp ranging ±5 V in 3.5 µs, triggered by integrator reset signal. Lab-based, not compatible with field operation ⇒ embedded solution ? 10 / 12
Subsurface wireless chemical sensing
Solution: replace analog timebase with digital timebase
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability
• Replace laboratory equipment with embedded electronics: FPGA or
microcontroller triggers on reset signal and generates ramp • not so obvious ... 8-bit (256 steps) within 5 µs=50 MHz DAC • R-2R network on FPGA output • different clocks for FPGA and GPR ⇒ jitter in reset detection
Digital ramp synthesis Conclusion
50 MHz
Voff b7 2R
reset CPLD
...
R
R
10R
R
− +
R
b2 2R
R
2R
R
2R
2R
b1 b0
reset input 50 ns/div
R out
CPLD out (osc. trig.)
3500/20=175 steps
Analog is asynchronous, digital must be synchronous
±20 ns when FPGA is clocked with 50 MHz oscillator 11 / 12
Subsurface wireless chemical sensing
Solution: replace analog timebase with digital timebase
J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
• Replace laboratory equipment with embedded electronics: FPGA or
microcontroller triggers on reset signal and generates ramp • not so obvious ... 8-bit (256 steps) within 5 µs=50 MHz DAC • R-2R network on FPGA output • clock synchronization on reset signal: use the GPR 16 MHz
reference signal (quartz referenced = ± ppm/K) ProEx 16 MHz
DDS AD9958
x20
160 MHz
Voff
R
2R
R
b7
...
reset CPLD phase synchronized
... R R
b2 2R
R
2R
R
2R
2R
b1 b0
10R − + R out
3.5/160=560 steps
Ramp synchronized with clock, but 160 MHz emission from DDS11 / 12
Subsurface wireless chemical sensing
Solution: replace analog timebase with digital timebase
J.-M Friedt & al Outline
Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
• Replace laboratory equipment with embedded electronics: FPGA or
microcontroller triggers on reset signal and generates ramp • not so obvious ... 8-bit (256 steps) within 5 µs=50 MHz DAC • R-2R network on FPGA output • clock synchronization on reset signal: use the GPR 16 MHz
reference signal (quartz referenced = ± ppm/K) ProEx 16 MHz
Voff
R
2R
R
b7
...
reset FPGA phase synchronized
R
PLL b2 640 MHz
... R
2R
R
2R
R
2R
2R
b1 b0
10R − +
5
R out
4 3.5/160=560 steps
delay (ns)
Acoustic transducer
3 2 std=100 ps std=213 ps
1 0 0
1000
2000
3000
4000
5000
6000
11 / 12
Subsurface wireless chemical sensing J.-M Friedt & al Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
Conclusion and perspectives We have demonstrated • using surface acoustic wave (SAW) transducer as cooperative target and more broadly reference echo generator with ≥ µs delay • sensing capability of SAW transducers when functionalized with the appropriate polymer: H2 S detection yields X00 ps delay @ 200 MHz • issue of drift of the sampling rate
reference of the stroboscopic measurement ... • ... solved by replacing the analog
timing generator with a quartzsynchronized digital timebase. Perspectives: • Replace R-2R network with a “real” DAC with voltage reference • investigate PLL jitter impact on sampling rate stability Who else ? anyone using the phase, i.e. beam focusing/time reversal Project repository: sourceforge.net/p/proexgprcontrol/wiki/Home/ Further reading: J.-M Friedt, Passive cooperative targets for subsurface physical and chemical measurements: a systems perspective, IEEE Geoscience and Remote Sensing Letters 14 (6), 821–825 (2017), available at jmfriedt.free.fr/ieee_gpr.pdf 12 / 12
Subsurface wireless chemical sensing J.-M Friedt & al
Conclusion and perspectives
Outline Acoustic transducer Chemical functionalization Sensor measurement Timebase stability Digital ramp synthesis Conclusion
Acknowledgement: Mal˚ a Geoscience has supported this research by lending and donating equipment for research purposes. • issue of drift of the sampling rate
reference of the stroboscopic measurement ... • ... solved by replacing the analog
timing generator with a quartzsynchronized digital timebase. Perspectives: • Replace R-2R network with a “real” DAC with voltage reference • investigate PLL jitter impact on sampling rate stability Who else ? anyone using the phase, i.e. beam focusing/time reversal Project repository: sourceforge.net/p/proexgprcontrol/wiki/Home/ Further reading: J.-M Friedt, Passive cooperative targets for subsurface physical and chemical measurements: a systems perspective, IEEE Geoscience and Remote Sensing 12 / 12 Letters 14 (6), 821–825 (2017), available at jmfriedt.free.fr/ieee_gpr.pdf
