Examination of reactivity for metallurgical coke using micro-CT analysis David Jenkins, Merrick Mahoney, Alex Deev, Jason Donnelly, Rowan Davidson, Sherry Mayo, Richard Sakurovs CSIRO Australia NIER, University of Newcastle, Australia

Talk outline Background on coke Project objectives Reaction and imaging Results and conclusions Acknowledgements

What is coke? made from crushed (metallurgical/coking) coal coal heated in absence of air undergoes transformation: coal -> plastic phase -> coke particles soften gas evolves, bubbles form, grow and coalesce re-solidifies to form a strong, hard substance used for reduction of iron ore in an ironmaking blast furnace part of the integrated steelmaking process

source: World Steel Association (worldsteel.org) (2016)

Source: Resources and Energy Quarterly, June 2017 (www. industry. gov Source: Resources and Energy Quarterly, June 2017 (www.industry.gov.au)

Our approach – building understanding through examination of fundamental processes How does a packed bed of coal transform to a solidified porous composite to form the final coke? blends of coal from different sources Coal chemistry and structure  plastic layer behaviour Formation of structures – plastic layer behaviour Weak / strong structures incl. reactivity

Objectives Develop a detailed understanding of the effect of coke reaction with carbon dioxide upon the microstructure of coke. Focus on (industry standard) test for reactivity. Examine the selective degradation/reaction of the coke components. Link the change in coke microstructure due to reaction with change in coke strength through: finite element modelling, pore structure analysis and measurement of strength.

Coke microstructure | David Jenkins samples coke samples: 5 different cokes CSR lumps 6 samples from each coke react 3 samples at 1100oC react 3 samples at 900oC Coke microstructure | David Jenkins

1100oC 30 mins reaction time CSIRO Labs Clayton CO2 4 laps

stressing of samples CSIRO Labs Clayton 5th lap

3D image acquisition high resolution 3D microstructure imaging: Imaging and Medical Beamline Australian Synchrotron 3D CT scan of coke CSR lumps ~9mm voxel size ~15 mins/scan ~10Gb image per sample >150 images

coke reaction component carried out by CSIRO Mineral Resources, Clayton lab testing/ validation phase before main experiment 2 furnaces 1100oC 900oC CO2 flowrate 3.5 l/min protocol: 20 mins cool zone 30 mins reaction zone 10 mins cool zone remove

Coal/coke properties

C187 particle 1 @ 1100oC 26. 4% mass loss (c. f C187 particle 1 @ 1100oC 26.4% mass loss (c.f. CRI of 22 in standard test) (1) registered the image of the unreacted particle with the image of the particle after 30, 60, 90, 120 minutes of reaction (2) then subtracted reacted particle image from the original

unreacted particle

reacted 30 mins difference 0 -> 30

reacted 60 mins difference 0 -> 60

reacted 90 mins difference 0 -> 90

reacted 120 mins difference 0 -> 120

C188 particle 1 @ 1100oC 45. 5% mass loss (c. f C188 particle 1 @ 1100oC 45.5% mass loss (c.f. CRI of 41 in standard test) (1) registered the image of the unreacted particle with the image of the particle after 30, 60, 90, 120 minutes of reaction (2) then subtracted reacted particle image from the original

unreacted particle

reacted 30 mins difference 0 -> 30

reacted 60 mins difference 0 -> 60

reacted 90 mins difference 0 -> 90

reacted 120 mins difference 0 -> 120

Why the difference? Porosity? Seems to be the case that lumps with higher porosity have flat reaction profiles lumps with lower porosity have steep reaction profiles Indication that gas transport through pores is limiting factor

Overall porosity change

Why the difference? Porosity? limitation due to resolution of image Seems to be the case that lumps with higher porosity have flat reaction profiles lumps with lower porosity have steep reaction profiles Indication that gas transport through pores is limiting factor limitation due to resolution of image smallest pores ~10-20mm across some porosity “missed” Limited study – only 5 cokes

3D stress analysis Solve elasticity problem for loaded particle Effectively “squeezing” particle 3D microstructure used as input Very large computation Calculate distribution of von Mises stress Compare high stress in outer shell (~2mm) with core of particle

What causes difference in inert reactivity? Mineral content? Inert pore structure? 2017 ACARP proposal: “Imaging gas penetration inside coals and cokes at nanoscale and determining its influence on coke reactivity” Sherry Mayo (CSIRO) & Merrick Mahoney (NIER)

Discussion high resolution micro-CT imaging of progressively reacted coke shows selective reactivity of IMDC and RMDC different reaction behaviour for different inerts shows transport (pore diffusion limited) control of reaction important indications that pore diffusion limited by “micro” and “macro” pores indication of higher stress in outer shell due to transport limited reaction Next step to image gas penetration into inerts

Acknowledgements Funding from ACARP Coal/coke from various coal companies Merit-based access to IMBL/Australian Synchrotron Peter Tyson, CSIRO – high performance computing support Hannah Lomas & Gareth Penny (Newcastle), Karryn Warren & Lukas Koval (CSIRO) – sample preparation and imaging