Low-cost Printed Electronic Nose Gas Sensors for Distributed Environmental Monitoring

Transcription

1 Low-cost Printed Electronic Nose Gas Sensors for Distributed Environmental Monitoring Vivek Subramanian Department of Electrical Engineering and Computer Sciences University of California, Berkeley RD

2 Distributed environmental monitoring Need for distributed monitoring - Identification of environmental hazards - Triggering of proactive action - Development of accurate environmental models Sensor Requirements - Ultra-low-cost - Ease of dispersal - Trainability / adaptability Our Approach: Arrayed organic FETs - Easily arrayed at low-cost via printing - Flexible for easy dispersal - Trainable via electronic nose architecture

3 Commercial E-noses ppbrae Plus $ For homeland security - Detects toxic agents, mildew Cyranose $ Can be trained to detect a wide range of odors: alcohols, chemicals, oil, food

4 Commercial Gas Sensors Vernier O 2 sensor $186 Minimax Pro H 2 sensor $199 Gas Alert Micro 3 H 2 S sensor $612

5 Arrayed Gas Sensors Molecule in ambient Map Responses Substrate A B C D Sensor Parameter Responses to different molecules Pixel Array Index Generate Chemical Signatures

6 Printing: a pathway to low-cost No lithography No vacuum processing (CVD, PVD, Etch) Reduced abatement costs Cheap substrate handling Reduced packaging costs

7 Organic Gas Sensors Gas sensing with OTFTs is a good match - Good sensitivity - Synthetic richness - Easy array integration - Low performance requirements - Short-term applications available The New York Times, April 4, 2002, illustration by Mary Ann Smith

8 OTFT Gas Sensing Absorbed through grain boundaries and reactive molecular sites Film expands Analyte changes hopping barrier height Source Odors + + Active + Material + T T Gate Dielectric Gate T Drain Analyte donates carriers or activates existing donors Analyte introduces traps and scattering sites

9 Low-cost Fabrication Inkjet deposition of organic material allows integration of sensor array Ultra-low cost requires integration of supporting circuitry

10 Printed Transistors Gate electrode is printed using gold nanocrystals Polymer dielectric is deposited via inkjet Low-temperature anneal forms S/D stripes and connections (in plane of page) Substrate Substrate Substrate Source / Drain contacts are printed using gold nanocrystals Various active layers are deposited via inkjet Low-temperature anneal forms gate stripe to edge of array (out of page) Substrate Substrate Substrate

11 Baseline sensor screening process The channel is exposed to the analyte, resulting in performance changes S G D Materials are characterized using a substrate-gated architecture (easy fabrication for rapid screening) A silicon substrate enables easy I/O via an edge connector

12 Sensor Characterization Switching between individual sensors is performed via a switch matrix PCB Agilent 4156 To ensure accuracy, measurements are performed with a calibrated precision semiconductor parameter analyzer.

13 Experimental Setup Valve Valve Agilent 4156 N 2 Mass Flow Controller Analyte Delivery Valve Sensor Chamber Mass Flow Meter Bubbler Exhaust Exhaust

14 Sensor Repeatability Id-Vd (Zoomed) -5.00E E-08 Id (Amps) -1.50E E E E E E E E E+00 Vd (Volts) Baseline Toluene Regen Multiple cycles can be performed with full regeneration

15 Multi-parameter sensing Transconductance Mobility baseline 60 sccm regen 100 sccm regen 1.00E E E E E E+00 gm (S) baseline 60 sccm regen 100 sccm regen 4.50E E E E E E E E E E+00 u (cm2/vs) Threshold Voltage Drain Current 0.00E E E E E E E+00 V (Volts) -6.00E E E-07 Id (Amps) baseline 60 sccm regen 100 sccm regen -8.00E E+01 baseline 60 sccm regen 100 sccm regen -1.20E E E-07

16 Sensor dynamics transient response Change in Drain Current Under Toluene Exposure -5.00E E E-09 Id (Amps) -2.00E E E E E E E Time (Minutes) Sensor response can be very slow, due to slow analyte absorption. Speed can be increased by reducing film thickness

17 Interaction Mechanisms - Sensors show a wide range of interactions, complicating analysis. Interactions include: Polar group interactions Chain / bulk interactions Swelling Percent of Baseline value I on Change in P3HT due to Acetic Acid 12 nm 40 nm 75nm 100 nm 12 nm 40 nm 75nm 100 nm Baseline Time (min) Percent of Baseline Value Baseline 20 0 I on Change in P3HT due to Ethanol 12 nm 75 nm 100nm Time (min)

18 Differential sensitivity pathway to an electronic nose? Nose Response to Water (Pentacene) Nose Response to Water (P3HT) Id (Amps) -1.00E E E E E E E Time (Minutes) Id (Amps) -7.00E E E E E E E E E Time (Minutes) Nose Response to Water (P3OT) -1.50E E-08 Id (Amps) -2.50E E E E E Time (Minutes)

19 Demonstration of basic electronic nose functionality Nose Response to Water and Milk Percentage of Baseline Current Milk Water Pentacene P3OT P3HT Time (Minutes)

20 Organic Transistor Stability Mobility (cm 2 /Vs) mu (sat) mu (lin) Ion (sat) -4.E-07-3.E-07-2.E-07-1.E-07 Drive Current (A) 0 7/14 7/15 7/16 7/17 Time (Date) 0.E+00 Implication: We must either improve dielectric interface or use V T -insensitive differential sensing method

21 Sensing Circuits Amplify sensor response Desensitize against operational drift Integration of encapsulated and unencapsulated OTFTs Integration of sensing OTFTs with supporting OTFT or silicon CMOS circuitry

22 Sensing Circuits Crone et al, J. Appl. Phys, vol 91, pp , 2002

23 Conclusions & Future Work Organic FET-based sensors show promising responses, including transient behavior and cycle life Work remains to optimize structure and process flow, particularly in terms of stability and reliability Future Work: - Integration of latest sensing materials into printed device architecture - Deployment in testing of environmentally-relevant analytes - Enhancement of specificity through functionalization / doping