Principles of Electronic Nanobiosensors

Transcription

1 Principles of Electronic Nanobiosensors Unit 4: Selectivity Bonus Lecture 4.4: Noise in Transducers By Muhammad A. Alam Professor of Electrical and Computer Engineering Purdue University A. Shakouri nanohub-u Fall

2 Outline Introduction: Noise as a selectivity problem Types and physics of noise sources: Thermal or white noise Absorption-desorption 1/f or pink noise Beating the noise limit and improving SNR Resampling Lock-in Conclusion Alam, Principles of Nanobiosensors

3 Selectivity issue due to biomolecules Competitive binding at steady state N t ( ) = N + N + k N ρ k N ρ kn p pρ p. k k 1 k ρ + 1 ' ' T 0 T T 0 T = + + ' ' T ρt + 1 T ρt + T ' T N p Geom p 2 D ρ=0 α = β = N N T T N + N + N T ' T ' NT + N + N + N ' T Geom Gem o Geom.. α+ β α β β β 1 β β β β α 0 = β β β α β 0 α+ β β β β α 1 Alam, Principles of Nanobiosensors

4 Noise is also a Selectivity Problem Environment Noise Transducer Noise Analyte Receptor Transducer DNA Protein Glucose Antibodies DNA SAMs Electrical Optical Mechanical Output analyzer 4

5 Noise in various lectures Double-gated ph sensor Genome sequencer 2 q kbtntλ I DS f 2 δvg = 1 log 2 + αµ eff Ceff WLCeff gm f1 Alam, Principles of Nanobiosensors

6 Recall: Three types of sensors Potentiometric Fluid Gate Amperometric Ref. & Aux. Electrode Mechanical Gate Charge to current Chemical to current Mass to frequency Alam, Principles of Nanobiosensors

7 Noise in Potentiometric Sensors Electrode noise (A) Fluid Gate Electrolyte noise (A) Absorption desorption Noise (B) Channel noise (A) Pink Noise 1/f (C) Alam, Principles of Nanobiosensors

8 Noise in Amperometric Sensor Alam, Principles of Nanobiosensors

9 Noise in Cantilever Sensors (A) Thermal/White noise (B) Absorption Desorption noise Alam, Principles of Nanobiosensors

10 White noise and Pink Noise Amplitude White time Amplitude Pink time Alam, Principles of Nanobiosensors

11 Outline Introduction: Noise as a selectivity problem Types and physics of noise sources Thermal or white noise 1/f or pink noise Beating the noise limit and improving SNR Resampling Lock-in Conclusion Alam, Principles of Nanobiosensors

12 Thermal (White) Noise in a Resistor kt B CV n = 2 2 V 2 n = = kt C 2 B V n = 2 π k TRB (approx!) B 4 ktrb (correct) B B f = c 1 2π RC υ 2 ( f ) 4k TR n B Alam, Principles of Nanobiosensors

13 White noise from various sources υ 2 ( f ) 4kTR n B c 2 υ n = 4kTR 2 B sol υ = 4kTR n B ch R sol = ( zn I qµ ) 1 2 i 0 R ch = I V D D Alam, Principles of Nanobiosensors

14 Origin of Pink Noise: Trapping/Detrapping I D I = qnµ E Defects I Traps D,0 n0 0 D n µ = + µ n µ time time Alam, Principles of Nanobiosensors

15 I I D n µ = + n µ D, /f noise Number fluctuation mobility fluctuation 2 q k TN λ I 1 A υg ( f) = 1+ αµ WL C I V f f =1 2 2 B T D 2 C0 0 D G f Gn, = υg = 2 1 f V ( f ) df A ln( f f ) Q n = ` CV 0 Gn, SNR = I D I ( f, f ) n 1 2 Alam, Principles of Nanobiosensors

16 Example of 1/f noise υ W=100 nm, L=5 um, C 0 =4.3 x10-7 F/cm 2 N T =2.3x10 18 ev -1 cm -3, λ=.14 nm, a=1 (linear regime) Number of traps Tunneling distance q 2 k TN λ V 2 2 ( f ) = B T = G WL C 2 Hz 0 Smaller transistors/sensors have larger noise! Alam, Principles of Nanobiosensors

17 Outline Introduction: Noise as a selectivity problem Types and physics of noise sources: Thermal or white noise 1/f or pink noise Beating the noise limit and improving SNR Resampling Lock-in Conclusion Alam, Principles of Nanobiosensors

18 Selectivity: A problem of Information theory? DNA sequence [1001] Parasitic molecules Sensor output Homopolymers Sensor noise α + βββ,,, α + β Better S/N ratio by increasing signal strength (PCR), resampling, or by suppressing the noise by tagging Alam, Principles of Nanobiosensors,

19 Resampling for improved SNR Signal S 0 S n 5S < = i 1 n i time Trade-off Speed for accuracy SNR N S N S S N = = = σ( N) σ 0 N σ0

20 White noise and Pink Noise White Amplitude signal SNR time Pink Amplitude SNR time Alam, Principles of Nanobiosensors

21 Signal averaging improves SNR dramatically Signal and Noise mplitude Sample Size Few hundred samples are good enough 21

22 Conclusions We have discussed two classes of selectivity issues related to biomolecules and transducers Of all the noise sources in a potentiometric sensor, 1/f noise is dominant. Thermal noise play an important role in cantilever sensing. Analyte concentration changes slowly compared to the ability to sample the signal. This offers an opportunity to detect signal even below the instantaneous noise level. Careful consideration of noise is key to a robust biosensor platform. Alam, Principles of Nanobiosensors