Thesis Defense Amir

Terahertz Spectroscopy in Industrial off-gas detection and Analysis

Amirhossein (Amir) AlikhanzadehSupervisor: Murray J. ThomsonCombustion Research Laboratory, Department of Mechanical and Industrial EngineeringUniversity of Toronto, Toronto, Ontario, Canada Investigation of the effects of beam scattering and beam wandering on laser beams passing thorough the off-gas duct of an Electric Arc Furnace (EAF)

Master of Applied Science Thesis Defense University of Toronto, OntarioDecember 10th , 2014

Project Overview2

2

Amirhossein Alikhanzadeh (AA) - Larger fonts, move lower.

3Motivation : Steel industryUses primarily scrap steelLower energy consumptionRegulations on the emissions12 to 14 Billion $ sale in Canada (2012)High production Recycling rate of 40 % to 60 % in Canada (7 million tonnes recycled in 2012)7.5 % of industrial energy use

BENCHMARKING ENERGY INTENSITY IN THE CANADIAN STEEL INDUSTRY (Prepared by Natural resources Canada)EAF 2002 Energy intensity indicator (MJ/tonnes of Hot Rolled Product)

4Electric Arc Furnace Process controlHeats charged material by means of electric arcConsists of three holes plus a fourth hole off-gas extractionOff-gas temperature of around 1400 degrees Celsius

Courtesy of Yuhui Sun [University of Toronto]

5Objective: How important is this study? Zolo - SCAN

Zolo-SCAN System developed for in-situ measurement Difficulty getting the beam over the path length of the exhaust ductCritical Path Length; two reasons

6Objective: How important is this study? LINDARCBased on TDLAS Water cooled rodMaking readings at the centreReasons similar to Zolo-SCAN

7Problems with the systems such as LINDARC and Zolo-SCANMost of the issues come from having the rods:Dust accumulationRod corrosionDamage to the equipment because of molten steelBridging the gap (LINDARC) by means of molten steel

High maintenance expenditure

8Background: Light interaction with medium

8

9Background: What is beam scattering?

Reyleigh sigmas=f*e4*lambda0^4/6/pi/epsilon0^2/c^4*(1/lambda4)Moolecular =AbsorptionAerosoles scattering (has orders of magnitude less in numbers) but Mie is applied and the effect is bigger (coefficient) compared to rayleigh9

10Scattering model: Mie theory-

Attenuation coefficients depend on the dimension, chemical composition and the concentration of particles dispersed in the gasous medium.Assumed spherical and homogenous10

Amirhossein Alikhanzadeh (AA) - Real size, marvel data

Scattering model: Literature validation

11

- Smoother if more orders of the Bessel function are used to generate the draph11

Scattering model: Model implications Changing particle sizes and their distributionConstant scattering areaTotal scattering area is kept the same while constituent particles are varied in size and subsequently concentrationThe assumed particle sizes are 0.5,1,1.5,2,2.5 m in diameter12

BY looking at the plot of Mie attenuation factor it can be said that for larger particles the scattering becomes less dependant on the wavelength and approaches 2 which means that for large enough particles, the attenuation cross section is equal to twice its geometrical cross section.

12

Amirhossein Alikhanzadeh (AA) - Spelling, work alot

Scattering model: Model implications Effects of changing particle sizes and their distribution on light transmission13How NIR transmission is affected when the wavelength is close to the dominant particle sizes

BY looking at the plot of Mie attenuation factor it can be said that for larger particles the scattering becomes less dependant on the wavelength and approaches 2 which means that for large enough particles, the attenuation cross section is equal to twice its geometrical cross section.

13

Amirhossein Alikhanzadeh (AA) - Spelling, work alot

Scattering model: Model implications Effect of refractive index

14

BY looking at the plot of Mie attenuation factor it can be said that for larger particles the scattering becomes less dependant on the wavelength and approaches 2 which means that for large enough particles, the attenuation cross section is equal to twice its geometrical cross section.

Displays the importance of change in refractive index of the particlesAluminum oxide was chosen with refractive index of 1.7682Dust particles from Dofasco have 2.01

14

Amirhossein Alikhanzadeh (AA) - Oscillations, read

15Scattering Experiment: Description/Purpose

15 cm25 cm25 cm25 cm25 cm

15

16Scattering Experiment: ResultModel could not be fully verified with experimentsThe experiment showed that model implications of less scattering at long wavelength is heldAlthough only qualitative measurements for mass, it can be seen that more particles means more attenuation of light

17Scattering model: Real particle sizesMalvern Spraytec Agglomeration of the particles is evident by comparison to Evans result

TalcAluminum oxide

Jamie Loh [University of Toronto]

17

18Scattering : Conclusions

19Beam wandering due to turbulence?

Varying Temperature, Density and Index of refraction through turbulenceAmplitude fluctuations Signal fadesBeam Wandering (Steering) Location movementor Scintillation Distorted beam shape

https://www.youtube.com/watch?v=VEFEQUY-KNA&list=LLSmktf7Lsrml6zbh078Zing&index=2

Beam spreading and wandering due to propagation through air pockets of varying Temperature, Density and Index of refractionResults in random phase and amplitude variations Fading of the signalLens like air pockets result in randomized interference in the warfront of the beamIs seen through beam Wandering (Steering) or Scintillation Distorted beam shapeVideo: Experimental movie of laser beam changes due to atmospheric turbulence. The turbulence was "played" on a spatial light modulator and the resulting beam changes measured on a CCD. As the movie plays, so the strength of the turbulence is increasing. When the turbulence is very strong, one cannot see the original Gaussian beam any longer.Reference : https://www.youtube.com/watch?v=VEFEQUY-KNA&index=2&list=LLSmktf7Lsrml6zbh078Zing

19

20Turbulence: TheoryKolmogorov theory of turbulenceEnergy flow starts from the outer scale and cascades to smaller scaleAct as small lensesLight beam of diameter bigger than the eddy would refract and smaller than the eddy broadens the beam; the net effect is a combination of the two

21Beam wandering: Effects of refractive index change

https://www.youtube.com/watch?v=bW6EcCcjFW0&index=4&list=PLKvrYlykYnYvXdpiffFtD2-jUrv7_2NJE

A well defined wave front will be distorted moving through a turbulent medium; scintillation, beam wander and broadening. (n-1)=79e-6 p/T (p in mili bars and T in Kelvin) , small pressure variations and their quick dispersiondelta(n)=79e-6/(omega-1)*p/T^2*delta (T) and omega= cp/cv=1.4 for air.Given the temperature structure parameter, refractive index structure can be found:Cn=[79*10^-6p/T^2]CT and CT=Root()r^(-1/3).Typical values STRONG-INTERMEDIATE-WEAK (5e-7 == 4e-8 ==8e-9)If the beam diameter is larger than the all the turbuklence scale sizes, the turbules act like weak lenses that deflect the beam ina random way without changing its diameter. If not, diffraction and refraction happens and the beam profile is smeared out.A constantly changing pattern at the end of the turbulent path is formed. If a small detector is placed at the beam, the result is scintillation which is fluctuations in the light intensity.Video: https://www.youtube.com/watch?v=bW6EcCcjFW0&list=PLKvrYlykYnYvXdpiffFtD2-jUrv7_2NJE&index=4

21

Amirhossein Alikhanzadeh (AA) - add reference to the slide not the notes

22Beam wander: Small-scale experiment

- To see how temperature gradient changes the beamNIR laserVIS laserPhoto detectors

TimeTemperature

Amirhossein Alikhanzadeh (AA) - words

Beam wander: Small-scale experiment implication23- Linear relationship between temperature (gradient) and deviation from the detector

24Beam wandering: Experiment goal- The experiment was designed to test if adding a large collecting lens would improve the signal

25

Beam wander: Experiment at the Vertical Flow Reactor

- Mapping the intensity of laser received at the detector

26Beam wander: Near Infrared Laser signal strength map on Vertical axesVertical axis No lensVertical axis With lens

41 % drop9 % drop16 mm6 mm

26

Amirhossein Alikhanzadeh (AA) - add line and see

27Beam wander: Near Infrared Laser signal strength map on Horizontal

Beam wander: Visible Laser signal strength map on Horizontal axis28

28

Beam wander: Visible Laser signal strength map on Vertical axis29

30Beam wander: Frequency of occurrence - Method

The aperture is fixed in position and the intensity of light inside the aperture is received by the detector

http://www.pnas.org/content/111/34/12320.full

From : Single-shot stand-off chemical identification of powders using random Raman lasingHe-Ne after propagating the full 400-m path length30

Beam wander: Frequency of occurrence - Result31- Adding the lens Higher power and peak, narrower profile

Beam wander: Frequency of occurrence - Result- Adding the lens Higher power and peak, narrower profile32

33Conclusion

From : Single-shot stand-off chemical identification of powders using random Raman lasingHe-Ne after propagating the full 400-m path length33

34

Suggestions for future work

35

Future work: Model

The model has limitations in terms of the particle sizes that can be input to the modelDue to the same limitations in particles size parameter, the results of the model are only valid through far infrared region and cannot be used for longer wavelengthsBy adding Rayleigh theory and geometric theory to the model, the dynamic range is vastly improvedThe model assumes a linear relationship between the concentration of particles and the amount of light transmission; a more accurate relationship can be developed thorough experiments which improves the accuracy of the model; the assumption of linear relationship is valid for x~1 but for much higher concentrations may not be valid

35

36Future work: Experiment

Needs improvement on the particle size and concentration measurement; ScatteringNeeds continuous measurements of temperature gradient of the medium to predict the laser beam behaviour; Beam Wandering

36

37Acknowledgments Dr.Zhenyou Wang (University of Toronto)Dr. Arathi Padmanabhan (University of Toronto)Prof.Murray Thomson (University of Toronto)

38Questions

39Background: Rayleigh scatteringResponsible for blue skyResponsible for red sunsetPreferential scattering

Rayleigh Scattering

Reyleigh sigmas=f*e4*lambda0^4/6/pi/epsilon0^2/c^4*(1/lambda4)Moolecular =AbsorptionAerosoles scattering (has orders of magnitude less in numbers) but Mie is applied and the effect is bigger (coefficient) compared to rayleigh39

Amirhossein Alikhanzadeh (AA) - Write the equation, not the image

Scattering coefficient Responsible for blue skyResponsible for red sunsetPreferential scattering

D-MIRD-NIRD-VIS3.2372121.4801410.603516

40

Reyleigh sigmas=f*e4*lambda0^4/6/pi/epsilon0^2/c^4*(1/lambda4)Moolecular =AbsorptionAerosoles scattering (has orders of magnitude less in numbers) but Mie is applied and the effect is bigger (coefficient) compared to rayleigh40