MEMS Piezoelectric Vibration Harvesting

Transcription

1 ENERGY HARVESTING: MEMS Piezoelectric Vibration Harvesting Thermoelectric Harvesting Lindsay Miller, Alic Chen, Dr. Yiping Zhu, Deepa Madan, Michael Nill, Dr. Rei Cheng Juang, Prof. Paul K. Wright & Prof. James W. Evans University of California, Berkeley

2 Multi source Energy Harvesting Industrial Pump Smart Roll Thermoelectric Wireless Sensor Node Smart Stamp Piezoelectric Wireless Sensor Node

3 Smart Stamp Energy Storage Wireless Sensor Micro device MEMS Sensor Sensor 1cm Energy storage Radio Energy harvesting Piezoelectric Energy Harvesting Cable Output Voltage Radio 1cm Magnetic Field Piezoelectric MEMS Cantilever Microscale Magnet

4 Progress made in past 6 months: PIEZOELECTRIC Modeling completed Device optimization completed P > 1 μw W at matched hdfrequency Investigating methods for frequency tuning THERMOELECTRIC 1 st printed 50 couple prototype with 75µW/cm ΔT = 20K Future work on materials processing can improve device performance Exploring alternative geometries

5 Piezoelectric operating principle DEFORMATION (mechanical energy) VOLTAGE (electrical energy) PIEZOELECTRIC MATERIAL CANTILEVER BEAM + V - S. Roundy, PhD Thesis UC Berkeley 2003 Figure: Wikipedia, 2010

6 Where we left you 6 months ago: PIEZOELECTRIC VIBRATION ACCELERATION (MEASURED) Tested 8 beams on 8 ambient sources reliably produce low power Successfully printed proof mass on harvesters to modify resonance frequency BEAM OUTPUT (MEASURED) BEAM OUTPUT (CALCULATED) TRANSFER FUNCTION (CALCULATED) Hz BEFORE Hz 2.5 mm 1.5 mm Hz AFTER 1.5 mm 1.5 mm

7 Progress: Optimization & redesign Power out for 30 Hz sinusoidal accel input: 10 μw Power out for ambient accel input: 3.5 μw Powe er, μw Pow wer, μw Beam length, mm Beam thickness, μm Beam length, mm Beam thickness, μm am length, mm Be Constrained area = 1 cm 2. mm Be eam length, Beam thickness, μm Beam thickness, μm

8 Progress: Optimization conclusions 1. P > 1 μw is attainable if optimize for specific vibration source 2. If optimized harvester is moved to different source, power drops off 3. A broadband or tunable device is needed Optimized geometry Schematic Old geometry Photograph Top view of MEMS chip Top view of MEMS chip

9 Progress: How to deal with the need to match harvester and source frequencies 1. Measure vibration source a priority, customize harvester 2. Make an array with harvesters of different resonances 3. Design broadband device 4. Active tuning External applied force magnetic, electrostatic Stiffness modification Axial mechanical preload 5. Passive tuning Mechanical stoppers Nonlinear spring stiffness Bi stable oscillator

10 Multi source Energy Harvesting Industrial Pump Smart Roll Thermoelectric Wireless Sensor Node Smart Stamp Piezoelectric Wireless Sensor Node

11 Smart Roll MEMS Sensor Rad dio & Sensorr Radio Energy storage Radio gy Energy harvesting Energy Storage 1cm Wireless Sensor Micro device Thermoelectric Energy Harvesting

12 Thermoelectric (TE) Operating Principles Thermoelectric (TE) Energy Harvesting Current ΔT Heat Sink N P Heat Source Resistive Load Substrate Hot Side Cold Side Metal Interconnects N Type P Type Semiconductor Semiconductor Sources of Waste Heat Location Source Temp. Gradient Residential Boilers, Dryers, Freezers, Oven 10 30K Factories Exhaust pipes, Boilers, Condensers 10 80K Aluminum Smelter Boiler Vehicles Engine, Exhaust pipes pp 60K >100K Airplanes Cabin to External 10 50K Dryer

13 Thermoelectric Device Design Substrate Heat Flow Metal Interconnects N Type P Type Semiconductor Semiconductor Low Aspect trti Ratios Requires structural support Labor intensive assembly Heat Flow High aspect ratio pillars High density arrays 900+ couples for D = 1cm Takes advantage of printing process Flexible Substrate N type Semiconductor Metal Interconnects P type Semiconductor

14 Where we left you 6 months ago: Printable TE Materials epoxy resin add active particles mix ink THERMOELECTRIC Printable semiconductor/epoxy thermoelectric materials synthesized Small scale prototype Flexible Polyimide Substrate Printed 10 couple prototype which produced 0.85µW for 20K temperature difference Printed TE Materials Leg Dim.: 5 mm Length, 500 µm width, 200 µm thick

15 Progress: Device Scaling & Fabrication Voltag ge (mv) μw ΔT = 20K Voltage Output Power Output Evaporated Metal Contacts Printed Thermoelectric Lines utput ( W) Power O Density ( W/cm 2 ) Maximu um Power Flexible Polyimide100 Substrate 2 50 Ideal Model Fitted Model Measured Prototype t Electrical Leads Coiled 50 couple Device 75 μw/cm Current (ma) 5mm Temperature Difference (K) Device Prototype: 50 Couple Device (100 elements) ΔT = 5, 10, 20 Kelvin Element Dim.: 5mm x 640µm x 90µm Device Resistance ~ 2.5kΩ Power Density ~75µW/cm ΔT = 20K

16 Progress: Harvesting from Hot Pipes Printed P type Semiconductor Printed N type Semiconductor Pip pe Hot Water TEG w Pipe Hot Water Central heating unit Hi h i ill High aspect ratio pillars High density arrays 100+ couples for D = 10cm Takes advantage of printing process Rings can be stacked

17 Progress made in past 6 months: PIEZOELECTRIC Modeling completed Device optimization completed P > 1 μw W at matched hdfrequency Investigating methods for frequency tuning THERMOELECTRIC 1 st printed 50 couple prototype with 75µW/cm ΔT = 20K Future work on materials processing can improve device performance Exploring alternative geometries

18 Thank you! Any questions? THERMOELECTRIC Alic Chen Deepa Madan, Dr. Rei Cheng Juang, Michael Nill 1.3 cm PIEZOELECTRIC Lindsay Miller Dr. Yiping Zhu Prof. Paul K. Wright & Prof. James W. Evans Acknowledgements: California Energy Commission, CITRIS, Berkeley Manufacturing Institute, Berkeley Wireless Research Center