"Non-equilibrium chemically active plasma: modeling with Chemical Workbench

Deminsky MaximKintechLab23 July 2015

Outline

Chemical Work Bench (CWB) – toll for conceptual design of chemically oriented phenomena

Problems arising during elaboration of plasma model

Collection of plasma model in CWB environment

Coupling of plasma models with other models

Data needed for simulation: construction of plasma-chemical mechanism in CWB

Recovering of unknown characteristics of elementary plasma-chemical processes in CWB

Example of modeling of mercury-free light sources

Example of modeling plasma-assisted combustion

CWB computational environment

Substance and process properties, kinetic mechanisms

Kinetic mechanisms

Substance and elementary process

properties

Substance properties

KhimeraKhimera

Chemical workbenchChemical workbench

Automated data importQuantum chemistryQuantum chemistry

KintechDB

CWB software tools

Kintech Lab develops methods and special software tools for development of the predictive kinetic mechanisms and conceptual design of complex combustion and plasma systems:

Kintech Lab develops methods and special software tools for development of the predictive kinetic mechanisms and conceptual design of complex combustion and plasma systems:

Chemical Workbench – an integrated environment for the development and reduction of chemical mechanisms, and conceptual design of the chemistry intensive technological processes

Khimera – a unique tool for calculating microscopic parameters from first-principles calculations

KintechDB – a database of evaluated data for properties of substances, elementary processes and chemical mechanisms

Chemical Workbench©

Combustion Plasma Chemistry Pollution Control Waste Treatment and Recovering Metallurgy General Chemical Kinetics and Thermodynamics High Temperature In-Organic Chemistry Thermal and Plasma Hydrocarbon Pyrolysis Processes Education

Integrated modeling environment for kinetic modeling, kinetic mechanism development and conceptual reactors design in the fields of

Problems arising during elaboration of plasma model

Design of modelDesign of model

Chose ofappropriate model

Chose ofappropriate model

Collection of data set (“mechanism”)

Collection of data set (“mechanism”)

Thermodynamic dataThermodynamic data

Kinetic data (rate coefficient, cross sections)

Kinetic data (rate coefficient, cross sections)

Transport dataTransport data

Adequacy of model to discharge type Adequacy of model to discharge type

Coupling with electric network

Coupling with electric network

Coupling with chemistryCoupling with chemistry

Well Stirred Reactor (WSR), 2 modelsPlug Flow Reactor (PFR), 3 modelsCalorimetrics Bomb Reactor (CBR), 4 models Calorimetric Reactor with Deviation (CRD)/Sensitivity (CRS), 4 modelsPremixed Flame, 1 model

Kinetic models and Surface kinetic modelsKinetic models and Surface kinetic models

Full Thermodynamic Equilibrium Reactor (TER), 8 modelsStoichometric Equilibrium Reactor (STR), 4 models

Thermodynamic modelsThermodynamic models

Plasma models and Plasma models with Surface kineticsPlasma models and Plasma models with Surface kinetics

CWB model’s collection

Detonation and aerodynamic modelsDetonation and aerodynamic models

Chapman-Jouguet Reactor (CJ), 1 modelsZel’dovich-von Neumann-Doering Reactor (ZND), 1 modelExhaust Reactor (EXH), 1 model

CWB Plasma models CWB Plasma models

CWB plasma model’s collection

Is…

0D or 0D(+) dimension

uniform media (T,P,E/N, [Ci])

hydrodynamics time >> plasma & chemical times

coupling of EEDF solution with chemical reactions

Is not…

2D or 3D dimensions

for non-uniform distributed characteristics (T,P,E/N)

Navier-Stokes for CFD

Thus, CWB models is for investigation of plasma-chemical kinetic mechanisms and conceptual design of complex chemically active plasma systems

Thus, CWB models is for investigation of plasma-chemical kinetic mechanisms and conceptual design of complex chemically active plasma systems

EEDF solution with chemistry

The Boltzmann kinetic equation is solved with the use of two-spherical harmonics expansion of electron velocity distributed function, which gives following equation for EEDF:

Qel, Qrot, Qin, Qsup, Qatt, Qee elastic, rotational, inelastic, superelastic, attachment and electron-electron collision integrals

- calculation of rate constant of non-elastic processes

,

,

[ ][ ] j lai

i j lj l

d Ck C

dt - solution of balance equations for chemical species

iter

ativ

e

CWB plasma model’s collection (Types)Types:

Plasma Model – EEDF solution with Chemical reactions

Calorimetric Bomb Reactor (CBR) –0D, uniform, time dependent model

P type – pressure is constantQ type – given heat exchangeV type – volume is constantT type – temperature is constant

Plasma & Surface – EEDF solution withChemical reactions in gas and surface

Calorimetric Bomb Reactor with surface(CBRS) –0D(+), uniform, time dependent model

CWB plasma model’s collection (Subtypes)

Subtypes:

The reactor model is based on numerical solution of the Boltzmann kinetic equation for electron energy distribution function (EEDF) and determination of rate coefficients of electron induced chemical reactions, energy distribution and electron’s swarm parameters in gas discharge. Gas composition in the reactor is changed as a result of chemical and vibrational kinetics plasma.

Nonequilibrium plasma reactor models available for different electric circuit configuration:• Current is given (J)– reactor with specified fixed value of the discharge current. Corresponds to electric circuit with plasma-gap connected in series with current generator (high voltage generator with high internal resistance).• L-C-R Circuit – the dependence of reactor electric field intensity and current density is determined by external LCR circuit. The plasma-gap is connected in series with a resistor, capacitor and inductance. Initial voltage on the capacitor is used as initial voltage on gap. It is assumed that the initial current at zero.• E/N is given (U) – time dependence of the reduced electric field is specified.• U-L-C-R Circuit – the plasma-gap is connected in series with resistor, capacitor, inductance and voltage source. It is assumed that the initial current and initial voltage on the capacitor is zero. Time dependence of the voltage at the voltage source is specified.• V-R – the plasma-gap is connected in series with a resistor and voltage source.

Extension of plasma models capabilities by flow sheet simulations

Non-uniformity: treatment by many streamersNon-uniformity: treatment by many streamers

1) treatment by plasma

2) extension,mixing with gas

3) relaxation &chemistry

1st pulse 2nd pulse

Time1) Plasmamodel with E/N(t)

2) Well Stirred Reactormodel with

3) Plug flow reactormodel

Loop for number of pulses

Admixing of surrounding gas

FlowRate2

FlowRate1

Need to know:a) FlowRate1/ FlowRate2 ~ Streamers Volume / Total Volumeb) Mixing time tmix

tmix

Data needed for simulation: construction of plasma-chemical mechanism in CWB

Tree ofplasma-chemical

processes

Quantum chemistry--------------------------------------------

(Gaussian®, GAMESS®, Jaguar®)

Microkinetics---------------------------------------

(Khimera®)

Kin

tec

hD

BChemical kinetics

---------------------------------------(CWB®, Chemkin®)

Applications-----------------------------------------------

-(CWB®, TRASS®, Chemkin®,

Fluent®, ANSYS CFX®, Star-CD®)

KintechDB - databank of physical-chemical data and information system for multidisplinary R&D projects

Database content

Particle properties

Thermodynamic properties of individual substances

Particle properties

Thermodynamic properties of individual substances

Elementary processes characteristics

Elementary processes characteristics

Kinetic mechanismKinetic mechanism

Data analysis and visualization tools

Data analysis and visualization tools

KintechDB data analysis and visualization tools

Thermodynamic and kinetic data. Analysis and visualization

• Substance thermodynamic functions visualization and comparison

• JANAF, TPIS table generation

• Thermochemical reaction analysis

• Forward/reverse rate constants calculation

• Rate constants temperature/pressure dependence visualization

•Rate constants for different reactions/sources comparison

Operation with data: How construct mechanism?

1. Putting data by hands in the calculation from external sources2. Data export from database of processes and substances

3. Mechanism export from database of mechanism

3 general ways:3 general ways:

• Khimera: model library– Chemistry of heavy particles

• Direct Bimolecular Reactions

• Bimolecular reactions via long lived Intermediate complex

• Multi-channel unimolecular reactions

• Dissociation of diatomic molecules

• Ion - molecular reactions

• Gas - Surface reactions

– Electron molecular reaction • Excitation

• Ionization

• attachment

– Vibrational Energy Transfer • VV and VT exchange

– Photochemical Reactions • photo dissociation

• quenching

• isomerization

– Classical trajectories methods– Surface diffusion – Multicomponent thermodynamic properties model– Multicomponent gas transport properties model

Khimera© or “What to do if there is now data?”

Example: Te ionization cross sectionThe cross section of the reaction is evaluated in the framework of Born-Compton similarity function method. Three subshells give the main contribution into the total atomic ionization cross section, namely, 5p4 (IP=9 eV, N=4), 5s2 (IP=17.84 eV, N=2) and 4d10 (IP=47 eV, N=10). The account of the contributions of these subshells to total atomic cross section is sufficient for the incident electron energy up to 200–300 eV.

0 50 100 150 2000

2

4

6

8

10 Te+e=Te++e+e

, A

2

E, eV

experiment, R.S.Freund et al. 1990 Born-Compton resultsCross section of the process . Results of

calculations described are shown by red line, experimental data is shown by black squares (R.S.Freund et al. Phys.Rev.A, 41, 3575 (1990)).

Transport properties calculation

Data Base of interaction potentials and collisional integrals

Transport properties calculation

0 5000 10000 15000 20000 25000 300000

1

2

3

4

5

6

7

T, K

Thermalconductivity

totalW/m/K

Khimera

Cressault

Narayanan

Capitelli

0 5000 10000 15000 20000 25000 300000

2000

4000

6000

8000

10000

12000

14000

16000

T, K

Electricalconductivity

S/m

Khimera

Narayanan

Capitelli

0 5000 10000 15000 20000 25000 300000

0.0001

0.0002

0.0003

Khimera

T, K

ViscosityPa*s

Cressault

Narayanan

Capitelli

Transport coefficients are calculated by the accurate formulas of the Chapman-Enskog method with account for higher approximations 14, here is the number of approximations, i.e. the number of retained terms in Sonine polynomials expansions.

Example: calculation of transport properties of Air at P=1 atmExample: calculation of transport properties of Air at P=1 atm

Example of modeling of mercury-free light sources

Boltzmann equation for

the EEDF

Cross sectionsdata base

Rate coefficientsdata base

System of kineticequations for charged

and neutral species

Electric circuitequation

Рисунок лампы с травлением

GaI3(pellet)

GaI3 + e =>GaI3(-)=>….=>…GaI+e.=>Ga + I,I2(-)

Ga, GaI2, GaI3 (wall)

evaporationetching

condensation

I2(-) +M*=>I2 + e +M

Ga +e=>Ga*=> Ga + hjj

Рисунок лампы с травлением

GaI3(pellet)

GaI3 + e =>GaI3(-)=>….=>…GaI+e.=>Ga + I,I2(-)

Ga, GaI2, GaI3 (wall)

evaporationetching

condensation

I2(-) +M*=>I2 + e +M

Ga +e=>Ga*=> Ga + hjj

2.00 Torr Ar-ZnT ~ 400oC,

Zn pressure ~ 10 mTorr

R~1.3 cm, J~300 mA

Candidates:Halides of Ga, Zn, In,Cu, Al, Cd, Sb, Bi, Tl

Hierarchy of the processes leading to Ga formation

GaI3

GaI2

GaI

Ga +3I

GaI4-

GaI3-

Ga+3I-

+e

+e

+e

+e

+2e

+GaI2

GaIy ,Ga, I +Wall GaIn(wall)

Kinetic Modeling and Approach Validation

GaI+e=>Ga(4p3/2)+I+eGaI3+e=>GaI+I2+eGaI2+e=>GaI+I+e

GaI+e=>Ga(4p1/2)+I+eGa(4p3/2)+e=>Ga(4d)+eGa(4p1/2)+e=>Ga(4d)+eGa(4p3/2)+e=>Ga(5s)+eGa(4p1/2)+e=>Ga(5s)+e

Ga(4p3/2)=>Ga(Wall)I=>I(Wall)

GaI2=>GaI2(Wall)GaI=>GaI(Wall)

Ga(4p1/2)=>Ga(Wall)Ga(4p1/2)+GaI3=>GaI+GaI2

-0.6 -0.4 -0.2 0.0 0.2 0.4

Sensitivity

X A

xis

Titl

e

48

1216

20

0

10

20

30

40

6070

8090 100

Em

issi

on E

ffici

ency

, %

Sensitivity analysis Optimization of Emission

Calculation of emissivity properties of Ar-GaI plasmaCalculation of emissivity properties of Ar-GaI plasma

Comparison with experiment: emission spectra of GaI plasma

260 280 300 320 340 360 380 400 4200

1

2

3

4

5

6

6s => 4p3/2, 1/2

4d => 4p1/2

4d => 4p3/2

5s => 4p1/2

Em

issi

on In

tesi

ty, a

.u.

Wave length, nm

Simulation Experiment

5s => 4p3/2

380 384 388 392 396 4000.0

0.2

0.4

0.6

0.8

1.0 Simulation Experiment

Rel

ativ

e in

ten

sity

Wave length, nm

Atomic emission Molecular emission

[1] J. Phys. D: Appl. Phys. 40 (2007) 3857–3881 Multiscale multiphysics nonempirical approach to calculation of light emission properties of chemically active nonequilibrium plasma: application to Ar–GaI3 system, S Adamson, M Deminsky, et al.

[2] Journal of Physics D Applied Physics 05/2015; 48(20). Comparative nonempirical analysis of emission properties of the Ar–MeIn glow discharge (Me = Ga, Zn, Sn, In, Bi, Tl) M Deminsky at al

Modeling plasma-assisted combustion for turbine appl.

air

swirl fuel

plasmans

afterglow1 ms

flame0.5 ms

downstream30 ms

plug flowperfectlystirred

plug flowBoltzmann

electron-impact cross-sections on

air + methane

Mechanism* for natural gas combustion, including NOx chemistry+ low-temperature extension for methane+ plasma species reactions (ions, excited)

plasma afterglow waiting forignition

Vibrkin CBR WSR

t1 – dischargeduration

t2=L/v – afterglow time

tresid = 0.5 msec, Tburn = 1900 K

E/N=200 Tdt1~0.55 nsecEinp≈320 J/g

t2=3 msec

P=18.6 atmTgas inp=700 K

Process

Parameter

Conditions

CBR

t3=30 msec

burnout

Discharge model. Calculated E/N and current

0 50 100 150-80

-40

0

40

80

120

160

200

240

1st pulse

151st pulse Experiment

Air, 1 atm, T = 20 s, C = 33 pF

Dis

char

ge c

urr

ent,

A

Time, ns

С

Pulse generator

Electric chain

The waveform of the calculated form changes within about 20 pulses. In the established form first and second half cycles are nearly equal. Experiment and theory reasonably agree.

0.0 0.5 1.0 1.5 2.0 2.5 3.0

-50

0

50

100

150

200

250

Air, 1 atm, T = 20 s, C = 33 pF

Dis

char

ge c

urr

ent,

A

Time, ms

151st pulse1st pulse

Extension of combustion limits

equ

iva

len

ce r

atio

lean

rich

10–5 10–4 10–30.1

1

0.4

320 J/g, 200 Td pulsed plasma

residence time in recirculating flame zone (s)

0.31

noplasma

2 [CH4]

[O2]

equivalenceratio

=

[1]. Russian Journal of Physical Chemistry B, 2013, Vol. 7, No. 4, pp. 410–423. LowTemperature Ignition of Methane–Air Mixtures under the Action of Nonequilibrium Plasma, M. A. Deminskii at al ,

Effect of additional NOx production by plasma

25 ppm9 ppm3 ppm

25 ppm9 ppm3 ppm

flame temperature (K)

NO

x (p

pm)

2400200016001

10

100

1000 0.35 0.5

0.7 0.9

0.45

0.6

0.8

equivalenceratios

plasma off

320 J/g, 200 Td pulse plasma

Optimization: flame stabilization vs NOx generation

ΔNOx (ppm)

ΔT

turn

dow

n (K

)

10 100 100010

100

1000

100 J/l0

50 J/l0

100 J/l0

240 J/l0

Thank you !

info@kintech.ru for general service and product informationsupport@kintech.ru for all questions concerning Kintech Lab software

http://www.kintechlab.com +7 (499) 704 2581