Measuring trace elements during the December 2012 Callide Oxy-Fuel - PowerPoint PPT Presentation

Measuring trace elements during the December 2012 Callide Oxy-Fuel trial Peter Nelson a *, Anthony Morrison a , P. Sargent Bray a , Hugh Malfroy b , Rohan Stanger c , Chris Spero d a Graduate School of the Environment, Macquarie Unibversity, NSW, Australia, 2109 b Malfroy Environmental Strategies, 18/37 Nicholson Street, East Balmain, NSW, Australia, 2041 c Faculty of Engineering and Built Environment,,Newcastle University, University Drive, Callaghan, NSW, Australia, 2308 d CS Energy/Callide Oxyfuel Services Pty Ltd, Level 2 540 Wickham Street, Fortitude Valley Queensland Australia 4006

Project Parameters • Examination of trace metal fate and concentration • Four coal feeds (Coal C, Blend 1, Blend 2, Coal M) • Combination of sampling: – solids inputs and outputs (coal and ash) – at the stack exhaust (under both air and oxyfired conditions) – at various points of the CO2 Processing Unit (CPU) • Sampling targets : – Mercury (using both sorbent traps and continuous analysis) – Other trace metals (As, B, Be, Cd, Co, Cr, Cu, Mn, Ni, Pb, Sb, Se, Zn) – Halides (HBr, HCl, HF, Br, Cl, F) – Particulates

Stack, Ash and CPU Sampling

Stack Sampling Methods • Sampling probes and collection trains complied to the following standards where practicable: – Metals : USEPA Method 29 – Halides : USEPA Method 26/26a – Particulates: Australian Standard 4323.2 – Mercury : USEPA Method 30b ( sorbent trap modified for measurement of Hg ++ )

CPU Sampling Methods • Pressurised process stream - no sampling probes used - collection trains identical to those on the exhaust stack • Continuous Tekran mercury analyser: – deployed at several locations within the CPU – parallel determinations of mercury using sorbent tubes

CPU Sampling Locations Coldbox Dryer Outlet Outlet Compressor Outlet Blower Outlet

Outcomes Halides: Location STACK CPU (mg/Nm³) (mg/Nm³) Air Oxy HBr <MDL 0.06-0.61 <0.10 <MDL HCl 10 - 67 9 - 18 <MDL HF 9 - 36 4 - 12 Br <MDL <MDL Cl <MDL <MDL F <MDL <MDL *MDL Minimum Detection Limit for Method

Outcomes Metals: Oxyfiring Airfiring 3.0 mg/Nm³ 2.0 1.0 0.0 B Cr Mn Ni Zn 0.20 0.15 mg/Nm³ 0.10 0.05 0.00 Cd Co Cu Pb Se The elements Sb, As, Be were All elements were at or around MDL below MDL under both air and values in the CPU beyond the first oxyfiring conditions, Cd was below low pressure scrubber. the MDL under airfiring conditions.

Summary of Outcomes • Halides and Metals – variations in stack gas concentrations appear influenced by both firing method (oxy or air) and coal type; • Halides – detectable in stack gases; all species below method detection limits beyond blower outlet (i.e following initial low pressure gas scrubbers) • Metals - elemental concentrations at very low levels in the CPU at the blower outlet and beyond; often below method detection limits

Mercury Mass Balance Mercury analysis at the following stages of the process: • Coal Input • Ash Outputs (fly ash, bottom ash, rear pass and air heater ash) • Flue gas concentration at stack • Process gas concentration in CPU – Blower outlet – Compressor Outlet – Dryer Outlet – Coldbox outlet

Total mercury concentration in flue gas 0.1- 0.8 µg/m³ 5-13% Hg in coal feed 22.9 -39.4 (ng/g) Mercury Distribution Airfiring Coal feed stockpile Slag & Ash 100% 87-95%

UBC in Flyash (%) a Location STACK Air Oxy Coal C 4.4 1.0-2.6 Blend 1 10.6 4.0 Blend 2 6.0 3.5 Coal M b 11.0 12.5 a Weighted averaged from 8 hoppers b Results likely to be sub-optimal due to failure of coal swirler during testing period

Total mercury concentration in process gas Approx 80% 0.4-0.9 µg/m³ of flue gas Hg Total mercury to CPU concentration in removed at flue gas scrubber 2.7-4.9 µg/m³ 23-33% Approx Hg in coal feed 10% of 22.9 -39.4 (ng/g) total flue gas flow Total mercury concentration in process gas <0.1 - 2 ng/m³ Coal 2.0-2.9% feed stockpile Mercury Distribution Slag & Ash 100% Oxyfiring(%) 64-74%

Mercury loss from CPU process gas Total mercury concentration Approximate Mercury Loss from in flue gas input 2.7 - 4.9 µg/m³ 0% Process Gas 20% Total mercury concentration 40% in process gas Total mercury concentration 0.4 - 0.9 µg/m³ in process gas 60% <0.1 - 2 ng/m³ 80% 100% Flue Gas IN Scrubber Blower Compressor Scrubber Dryers Coldbox HP LP CPU Process Operation

Gas Phase Mercury Speciation • Estimations of Hg ++ using KCl segments in sorbent traps; • Imperfect technique for Hg ++ as some breakthrough in the KCl occurred even at low sampling flow rates; • Will result in small underestimation of Hg ++ ; • Total mercury estimation unaffected, negligible breakthrough to second activated carbon segment ; • Total of 115 sorbent tubes analysed (71 stack, 34 CPU). KCl segments (Hg ++ ) FLUE GAS FLOW Activated carbon segments

Averaged Ratio Hg ++ /Hg total Location STACK CPU Air Oxy Blower Outlet Coal C 0.52 0.68 0.08 Blend 1 0.65 n.a. n.a. Blend 2 0.47 0.68 0.11 Coal M a 0.63 0.72 0.06 a Results likely to be sub-optimal due to failure of coal swirler during testing period n.a not available

Hg increases* following CPU depressurisation TEKRAN LOG 12th Dec 80 Compressor 70 Restored 60 Hg Concentration (ng/m³) 50 Compressor 40 Trip 30 20 10 0 00 02 04 07 09 12 14 16 19 TIME (HRS) * Sampling location CPU compressor outlet

Summary Hg outcomes • Hg was more likely to report to slag and ash under air firing conditions, most likely a result of generally higher UBC (%) in flyash when air firing; • Ratios of Hg ++ /Hg total seem relatively unaffected by firing condition (air or oxy); • Approximately 80% of Hg in CPU process gas was removed by the initial low pressure scrubber; • Following depressurisation of the CPU, evidence of significant Hg concentration increases was observed using the Tekran continuous Hg analyser; • Final CPU process gas Hg approached the levels measured in ambient air (<2 ng/m³).

Conclusions • Successful sampling campaign; • Initial analysis of data completed; • Minimal transfer of trace elements beyond the first scrubber in the CPU; • Hg levels in CPU produced process gas approach those measured in ambient air.

Acknowledgements • The authors acknowledge the significant input to the success of the project made by the sampling team from ECS Pty Ltd, Simon Newbigin, Michelle Yu, Henry Diona and Dante Mude. • The authors wish to also acknowledge financial assistance provided through Australian National Low Emissions Coal Research and Development (ANLEC R&D). ANLEC R&D is supported by Australian Coal Association Low Emissions Technology Limited and the Australian Government through the Clean Energy Initiative.

Outcomes Metals: Location STACK CPU CPU (mg/Nm³) (mg/Nm³) (mg/Nm³) Air Oxy Blower Outlet Compressor Outlet Antimony <MDL <MDL <MDL <MDL Arsenic <MDL <MDL <MDL <MDL Beryllium <MDL <MDL <MDL <MDL Boron 0.60 - 1.11 1.22 - 2.03 0.004 - 0.011 <0.006 Cadmium <MDL <0.006 <0.001 <MDL Chromium 0.34 - 0.98 0.01 - 1.22 <MDL <0.002 Cobalt 0.005-0.019 0.003-0.013 <MDL <MDL Copper 0.041-0.110 0.024-0.065 <0.0013 <0.005 Lead 0.008-0.021 0.009-0.017 <0.0008 <0.0008 Manganese 0.085-0.390 0.097-0.375 0.001-0.039 0.001-0.019 Nickel 0.205-1.700 0.009-0.649 <0.001 <MDL Selenium <0.015 <0.043 <MDL <MDL Zinc 0.073-0.410 0.047-0.172 0.003-0.010 0.002-0.008 *MDL Minimum Detection Limit for Method

Method Detection Limits Stack CPU MDL (mg/dsm³) MDL (mg/dsm³) Halides HBr 0.038 - 0.077 0.016 - 0.033 HCl 0.19 - 0.39 0.082 - 0.17 HF 0.19 - 0.39 0.082 - 0.17 Br 0.031 - 0.077 0.016 - 0.033 Cl 0.15 - 0.39 0.082 - 0.17 F 0.031 - 0.077 0.016 - 0.13 Metals Antimony 0.0017 -0.0043 0.00055 - 0.00082 Arsenic 0.0042 - 0.0110 0.0014 - 0.0021 Beryllium 0.0017 -0.0043 0.00055 - 0.00082 Boron 0.0042 - 0.0110 0.0014 - 0.0024 Cadmium 0.0017 -0.0043 0.00055 - 0.00082 Chromium 0.0017 -0.0043 0.00055 - 0.00082 Cobalt 0.0017 -0.0043 0.00055 - 0.00082 Copper 0.0017 -0.0043 0.00055 - 0.00082 Lead 0.0017 -0.0043 0.00055 - 0.00082 Manganese 0.0017 -0.0043 0.00055 - 0.00082 Nickel 0.0017 -0.0043 0.00055 - 0.00082 Selenium 0.0043 - 0.0110 0.0014 - 0.0021 Zinc 0.0042 - 0.0110 0.0014 - 0.0021