Characterization of Ionic Winds from Flames and Corona Discharges

Derek Dunn-RankinMechanical and Aerospace Engineering

University of California, Irvine

California Institute of Technology, Pasadena, CA – February 25, 2005

Creating Ionic Wind via Corona

• Chattock wind• Ion wind• Ionic wind• EHD flow• Electric wind

Negative DC Corona Discharge

1 Td = 10-21 V m2

E = electric field strength (V/m)N = neutral density – (1/m3)

Drawing not to scale – from Chen and Davidson, Plasma Chemistry and Plasma Processing, Vol. 23, No. 1, March 2003.

Ion Winds and The Electric Fly

• Wilson (1750) built spinning pinwheel

• Cavallo (1777) claims similar-charged air repels points– does not work in vacuum

• Electric fly used to discuss ion wind theory: Michael Faraday (1838), James Clerk Maxwell (1873), among others

• Most recently discussed by Leonard Loeb in Electrical Coronas (1965)

• Electric fly motion can be observed before any wind motion– spins at voltages well

below breakdown

Rotational Speed Linear with Voltage

Electric Fly Speeds

0

500

1000

1500

2000

7 9 11 13 15 17

ABS Voltage (kV)

RPM

Applications of Ionic Winds• Coronas

– Ionic Breeze – fan and particle removal– Air delivery without moving parts– Electrostatic precipitator enhancement– Heat transfer and evaporation– Plasma actuators

• Flames– Dynamic flame control– Heat direction in microgravity– Earth-based microbuoyancy

Ionic Wind Applications

HVSharperimage.com

PIV Experiment

45o

1

2

3 4

5

6

7

HV

1 Laser 4 Ion wind generator 7 Mist-seeded airflow2 Sheet optics 5 Mist canister3 Camera 6 Rubber tubing

Air vent45o45o

1

2

3 4

5

6

7

HV

1 Laser 4 Ion wind generator 7 Mist-seeded airflow2 Sheet optics 5 Mist canister3 Camera 6 Rubber tubing

Air ventAir vent

• Copper Vapor laser sync’d to 2kHz camera captured high-speed video• Camera shifted 45º forward of perpendicular to maximize intensity• Personal Air Purifier electrodes powered by variable -HV power supply• Water/glycerin mist traced air flow (seeded downstream of electrodes)

High-speed video• HV=-10.3 kV (max before breakdown) shown for subtle nozzle ( ID

= 9.5 mm )• Actual orientation: directly upwards• Capture rate: 1,825 frames per second

– Real time elapsed during 500 images: 1/4 seconds• Viewing rate: 25 frames per second

– Total time to view 500 images: 20 seconds

PIV results -- Velocity profile at exit

0

0.5

1

1.5

2

-1 -0.5 0 0.5 1r/ro

Velo

city

[ m

/s ]

10.3 kV

9.6 kV

8.9 kV

8.1 kV

7.4 kV

5.9 kV

4.4 kV

Point-to-Cylinder Ionic Wind

2R

V

Charged needle

Earthed ring

High Voltage Power Supply

i

- HV Gas Flow

Acrylic tubeNozzle (optional)

Xo Yo

Optimizing the Ion Wind

0

5

10

15

20

0.3 0.6 0.9 1.2Yo, Length of cylindrical ground electrode [ in ]

Sta

gnat

ion

Pre

ssur

e [P

a]

Xo = 0 [in]Xo = 0.25Xo = 0.5

0

1

2

3

4

5

6

7

8

0 0.2 0.4 0.6 0.8 1

Exit Area / Inlet Area

Mea

n Ve

loci

ty [

m/s

]

0

1

2

3

4

5

6

7

8

Volu

met

ric F

low

Rat

e [ S

LPM

]

Ideal

Ideal is velocity needed to maintain 6.4 l/min flow rate as nozzle exit area decreases

Flow Limit of Single Stage

Stacking Stages to Increase Pressure

Multi-Stage Ion Blower

Pressure Delivered

5.5 Pa per stage

Staged System Performance

Ionic Wind Fan Curves

Future Activities

• Swirl Generator– ESP enhancement

• Microscale Gas Delivery– Point of application flows

• All Electric Burner• Modeling

Applications of Ionic Winds

• Coronas– Ionic Breeze – fan and particle removal– Electrostatic precipitator enhancement– Heat transfer and evaporation– Plasma actuators

• Flames– Dynamic flame control– Heat direction in microgravity– Earth-based microbuoyancy

Flame Derived Ion WindFlame Derived Ion Wind

V

e-e-

++

+

++ + +

+++

+++

++

+

++++

++

++

++Ions travel from the flame towards ground

Electrons conduct through the flame to the capillary

Neutralization by electrons at the cathode

Force and Velocity from Ion Wind• K is mobility of the charge carrier -- 1 cm2/s/V• g is gravitational acceleration; ρ is the density• x is the distance over which the field acts• j is current density – flame: 1 µA/ cm2

xKjve ρ

=gK

jFF

b

e

ρ=

Order of 1 m/s over 1 cm distance

Order of 1-2

Effect on Small

Diffusion Flame

Methane fuel flowrate: 9 cc/minCapillary diameter: 1.7 mmElectrode Spacing: 7 cm

Conditions1. Naturally Buoyant – 0 Volts2. Microbuoyant – 2880 Volts3. Negatively buoyant – 4300

Volts

0 1000 2000 2500 3000 4000 5000Applied Potential (Volts)

Double Exposure Holographic Interferometry Images (setup #1)

0 Volts 2242 Volts 2765 Volts

0

200

400

600

800

1000

1200

1400

1600

2 4 6 8 10 12Radial Position (mm)

Tem

pera

ture

(K)

0 Degrees5 Degrees10 DegreesDiffusion Profile

Temperature Profiles

X

Y

Flame Current

0

0.5

1

1.5

Curr

ent

(mic

roam

ps)

0 1000 2000 3000 40000.7

0.8

0.9

1

Applied Potential (Volts)

Rela

tive

CH

430

nmEm

issi

on

Zone 1

Zone 2

Zone 3

Zone 1

Zone 2

Zone 3

• Saturation current can be characteristic of the flame

• Current provides a detection and/or decision opportunity (e.g., ion probe)

saturation plateau

Toward Feedback Control

Saturation

Σ D(s)-

+ e H(s)

Controller Plantry y

what do we measure for feedback? CH*

earthed

ComputerHV

Amplifier

PMT

430nmfilter

Amplifier

lensiris

•commonly accepted that CH* is a “good” measure of the global heat release in a flame

•CH* emission is linearly related to the fuel flow rate•the electronically excited states are confined to a thin region of the primary flame surface

Sine Tracking

0 2 4 6 8 105.5

6

6.5

7CH

Che

milu

min

esce

nce

DesiredActual

0 2 4 6 8 100

2

4

App

lied

Pote

ntia

l (kV

)

Time (sec)

Refined Apparatus•• Acrylic chamber Acrylic chamber

•• Ion current measured at the cathode (ground plane)Ion current measured at the cathode (ground plane)

•• Temperature, pressure, and relative humidity are monitoredTemperature, pressure, and relative humidity are monitored

•• Continuous running via air flushingContinuous running via air flushing

FUELCAPILLARY

FLAME

Earthed

+ HV

ELECTRODESV probe

Rref DMM

MICROSCOPE

CAMERA (VIDEO OR STILL)

AIR INLET(optional)

EXHAUST

INLET DIFFUSER

SCRUBBER

THERMOCOUPLE &HYGROMETER

MANOMETER

CERAMICSTANDOFFS

FUELCAPILLARY

FLAME

Earthed

+ HV

ELECTRODESV probe

Rref DMM

MICROSCOPE

CAMERA (VIDEO OR STILL)

AIR INLET(optional)

EXHAUST

INLET DIFFUSER

SCRUBBER

THERMOCOUPLE &HYGROMETER

MANOMETER

CERAMICSTANDOFFS

Schlieren Imaging Illustrates ComplexitySchlieren Imaging Illustrates Complexity

•• Identify buoyancy regimesIdentify buoyancy regimes

•• Determine relative electrical Determine relative electrical characteristics between fuelscharacteristics between fuels

•• ZZ--type type schlierenschlieren apparatusapparatus

•• Conventional and high speed Conventional and high speed videovideo

1000 VBuoyant

(negligible current)

2000 Vunder balanced(subsaturated)

2400 VBalanced condition

(saturation)

3000 VPartially neg-buoyant

(some 2ndary ionization)

5300 VNegatively buoyant

(much 2ndary ionization)

Ethane Fuel [12.7 mg/min]

1000 VBuoyant

(negligible current)

1000 VBuoyant

(negligible current)

2000 Vunder balanced(subsaturated)

2000 Vunder balanced(subsaturated)

2400 VBalanced condition

(saturation)

2400 VBalanced condition

(saturation)

3000 VPartially neg-buoyant

(some 2ndary ionization)

3000 VPartially neg-buoyant

(some 2ndary ionization)

5300 VNegatively buoyant

(much 2ndary ionization)

5300 VNegatively buoyant

(much 2ndary ionization)

Ethane Fuel [12.7 mg/min]

LIGHT SOURCEIRIS

MIRROR (f = 48 in)

MIRROR (f = 48 in)

KNIFE EDGE OR IRIS

LENS

LENS

VELLUMSCREEN

VIDEOCAMERA

6 in

Z – TYPE SCHLIEREN APPARATUS

LIGHT SOURCEIRIS

MIRROR (f = 48 in)

MIRROR (f = 48 in)

KNIFE EDGE OR IRIS

LENS

LENS

VELLUMSCREEN

VIDEOCAMERA

6 in

Z – TYPE SCHLIEREN APPARATUS

Ion Current ProfilesIon Current Profiles

•• PointPoint--toto--plane distribution plane distribution (Warburg Law) = cos(Warburg Law) = cos55ΘΘ

•• Integrated current provides ion Integrated current provides ion density boundary density boundary conditonconditon

•• Current distribution for model Current distribution for model validation

0

0.2

0.4

0.6

0.8

1

1.2

-40 -30 -20 -10 0 10 20 30 40

Radius [mm]

new data set 2new data set 3Jan 21 dataWarburg Law

validation

+ HV

V probe

Earthed

Rprobe DMM

TRAVERSE

INSULATED ION PROBE

Rground plane

+ HV

V probe

Earthed

Rprobe DMM

TRAVERSE

INSULATED ION PROBE

Rground plane

60 40 20 0 20 40 600

0.001

0.002

0.003

0.004

j θ( )

θ

60 40 20 0 20 40 60500

600

700

800

900

Temp θ( )

θ

60 40 20 0 20 40 602

1.5

1

0.5

0

Fratio θ( )

θ

Point to Plane Does not Produce Uniform BalanceWarburg curve

Electric force

TemperatureInverse of buoyant force

Force ratio – Fe/Fb

Balance condition

Numerical SimulationNumerical Simulation

oo

e ceVεε

ρ==∇2Gauss’s Law

( ) ( ) EceuTTgPuu

tu o

o +∇⋅∇+⎟⎠⎞

⎜⎝⎛ −+−∇=∇⋅+

∂∂ µρρρ 1Momentum

( ) 0=⋅∇+∂∂ u

tρρ

Mass

( ) ( ) ( )TkPuTcutTc

pp ∇⋅∇+∇⋅=∇⋅+∂

∂ρρEnergy

( ) ( )VceKcDcutc

∇+∇⋅∇=∇⋅+∂∂

+Ion concentration

Computational Model

• FEMLAB simulation• Finite element method• General equation solver• Staged solution

1. Electrokinetic (V, c)2. Electrodynamic (V, c)3. Hot flow (V, u, v, P, T, c)

• 2-D axisymmetric• Assumptions

– Balance condition– Isotropic Fickian diffusion– Single ion species (positive ions)– Temperature is constant within the flame (hot flow)– Uniform ion concentration at the flame– Ion properties are of H3O+

Gravitydirection

Sym

met

ry b

ound

ary

Anode boundary

Cathode boundary

Infinite condition

RADIAL DIRECTION

AX

IAL

DIR

ECTI

ON

hcap

Rflame (3.0 mm)

Rcap (0.5 mm)

63.5 mm

Gravitydirection

Sym

met

ry b

ound

ary

Anode boundary

Cathode boundary

Infinite condition

RADIAL DIRECTION

AX

IAL

DIR

ECTI

ON

hcap

Rflame (3.0 mm)

Rcap (0.5 mm)

63.5 mm

Electrokinetic Potential Field

• Ion motion is decoupled from the field (electrokinetic assumption)• Flux pathlines illustrate no flux beyond 20 mm away from the centerline

Role of Ion

Diffusion

• Diffusivity is varied to match experimental profile– D(T) varies by a factor of approx 20 from 300-1500 K

• Results– True D/K approximately 1 V– Ion drift speed more than 1000 times diffusion velocity– Ion diffusion cannot account for measured profile– Space charge dispersion is significant

0

0.2

0.4

0.6

0.8

1

-40 -30 -20 -10 0 10 20 30 40Radius [mm]

Nor

mal

ized

ion

curre

nt

h=60mm 13.8mg/minh=35mm 13.8mg/minnorm uniform ion concentrationekf D/K = 1000ekf D/K = 100ekf D/K = 10ekf D/K = 1

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

300 600 900 1200 1500Temperature [K]

Diff

usiv

ity

775.19, 10 TD airi ×= −

[m2 /s

]D

iffus

ivity

(cm

2 /s)

Including Space Charge EffectsIncluding Space Charge Effects

• Gauss’s law is solved including space charge• Space charge drives ions to travel further from the centerline

Effect of Ion ConcentrationEffect of Ion Concentration

0

0.2

0.4

0.6

0.8

1

-40 -30 -20 -10 0 10 20 30 40Radius [mm]

Nor

mal

ized

ion

curr

ent

h=60mm 13.8mg/minh=35mm 13.8mg/min1e8 ions/cc K.00032 D(T) 3kV1e9 ions/cc K=.00032 D(T) 3kV1e10 ions/cc K=.00032 D(T) 3kV1e12 ions/cc K=.00032 D(T) 3kV

Overall ion concentration in flame is varied– Lower concentration (108 ions/cc) – profile narrows– Higher concentration (1010-1012 ions/cc) – profile broadens

Buoyant Flow Preliminary Results

Hot Flow at Balance Preliminary Results

c = 109 ions/ccV = 2800 Volts

Flame Derived Ion WindFlame Derived Ion Wind

Induced electric field (causes horiz. migration)

V

Re-circulation driven by neutralized buoyant gas redirection• Source of unstable T and O2• Observed in shadowgraphs

Neutralization by electrons at the cathode

Ions travel from the flame towards ground

Electrons conduct through the flame to the capillary

High ion concentration at the center decreasing radially outward

+

++++++ +

++ ++ + + + +++

+ ++

+

++++++

+

+++++++++

+

+ ++

+

++++++

+

+++++++++

+

+ ++

e-e-

++

+

++ + +

+++

+++

++

+

++++

++

++

++

Final Thoughts• Ionic winds produce modest flows (a few m/s)

with no moving parts• Positive and negative polarity produces flow in

the same direction (allows AC approaches)• Efficiencies (flow power out/electrical power in)

are low but not far worse than standard fans at small sizes

• Flame source ion winds require far less electrical energy (100 times less) than corona winds

• Optimizing ionic wind systems should be possible by modeling systems with various ground plane geometries

Acknowledgments

• Matt Rickard (ionic wind) and Mike Papac(electric flame) – current UCI graduate student researchers

• Ben Strayer (flame control) and Jonathan Posner (holographic interferometry) – past UCI graduate students

• Prof. Felix Weinberg and Dr. Fred Carleton –Imperial College

• NASA Microgravity Combustion Program