Field measurements of mercury and SO3 from the Fabric Filter at the Callide Oxyfuel Project during Air-Oxy Transitions

Rohan Stanger*a, Timothy Tinga, Lawrence Beloa, Chris Sperob, Terry WallaaUniversity of Newcastle, bCallide Oxyfuel Project

5th Oxyfuel Combustion Research Network MeetingWuhan, CHINA29th October 2015

ANLECR&D Research areas at COP to reduce cost and risk

• SO3 formation → acid dew point corrosion + temperature• Hg capture → affected by oxy-conditions or carried to CPU

→ Hg oxidation uncertain

• SO2 capture → pH and NaOH usage in high CO2→ SO2/NOx reactions (N2O) in CPU

• Removal of Hg2+ ?

• NOx capture in CPU → kinetic reaction to NO2 or emitted as NO→ stability of condensates

• Hg capture in CPU → dependant on NOx→ product identification and stability→ risk to brazed Al-HEX cold box

Fabric Filter(not ESP)

NaOH Polishing (not FGD)

Compression(no AC bed)

The Callide Oxyfuel Project Retrofit

Picture courtesy of Yamada, IHI APP Oxyfuel Course 2010

No deSOx / deNOx / de-HgAll impurities sent to CPU

64-74% Hg captured in

fly ash

~70% Hg2+

in flue gasTo CPU

Hg results from Macquarie UniTrace Impurity Study, Nelson 2012

Transition Schedule and associated measurements

0:00 - 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00 - 8:00 10:00 12:00 14:00 16:00 18:00MONDAY 30/06/14 TUESDAY 1/07/14

Sampling interval

AIR Mode

OXY Mode

Ash Samples taken

12

3

4

5

6

7

89

10

11

12

1314

Low NOx burners used Original burners

used

• Transitions between air-oxy modes occur without CO2 capture• All flue gas to stack (except recycle) no CPU• Real oxyfuel emissions • Ideal time for studying Hg behaviour under changing O2

Heated Lance for SO3 and Total Hg(30 minute interval sampling during transitions)

• 2 heated zones• Collects 1 SO3 + 2 Total Hg samples• 10 minute “change-over”

Non-heated lance for continuous “Online Hg”• Designed with manual “switch” between

Hgtotal and Hg0

• In practice SnCl2 not responsive enough for Hgtotal

• Ohio Lumex 915+ Hg analyser• 1-20,000ng/m3 Hg in real time

Probe Assemblies

PFA sampling tubes with AC beds

Heated Probe gas conditioning

Online Hg Probe gas conditioning with PFA switches

Location on stack and sampling equipment

Hg measurements with original burners

AIR

OXY

(as Hg0)

sampling interval

Hg measurements with Low NOx burners(2 out of 4 burners)

(as Hg0 and Hgtotal)

(shaded Online Hg as HgTotal)

(note change in scale from previous slide, was 5µg/m3)

Bigger Picture – Historical trends in CO & Hg

(as Hgtotal)Higher CO

WithLow NOxburners

Summary of Hg exiting the Fabric Filter

• Hg affected by burner configuration• Similar capture in air and oxy firing modes

AIR OXYLow NOx Original Low NOx Original

HgTotal gas µg/m3 0.07 0.53 0.15 4.4* → 2

Hg2+ 0% 77% 65% 87%

HgTotal gas µg/m3

Corrected to 12%CO2

0.06 0.45 0.03 1.2* → 0.42

Hg Captured(by mass balance) 98 % 93% 99% 93%

* Highest point in transition corresponding to higher O2

SO3 measurementsSample Description SO3, ppm

1 OXY steady 0.55

2 OXY → AIR 0.56

3 OXY → AIR <0.27

4 OXY → AIR <0.23

5 AIR steady <0.06

6 AIR → OXY <0.25

7 AIR → OXY <0.26

8 AIR → OXY 0.82

9 OXY steady <0.14

10 AIR steady <0.06

11 AIR → OXY <0.28

12 AIR → OXY <0.26

13 AIR → OXY <0.28

14 OXY steady <0.10

Previous Measurements

31/05/2014 OXY steady 3.65

27/06/2014 OXY steady 1.60

28/06/2014 OXY steady 0.58

• Very low concentrations

• 0.6-3.6ppm SO3• 0.08-0.5% conversion SO2• ADP in oxy 118-136°C• Most below detection limits

• Disappointing for researchers

• Great results for Operators

• BUT is this the whole story?

Recall the Gas Cooler on COP flowsheet

• Gas Cooler used in oxy mode only

• Potential for cold surfaces during transitions

• Speculation that SO3condensing as H2SO4

Bypassed in air mode

oxy mode

Conclusions

• The capture of Hg at Fabric Filter >90%• Burner configuration has over-riding influence on

• Hg total• In particular Hg 2+

• Oxy-firing showed higher concentrations of Hg 2+• Similar to air when corrected to 12% CO2• Expected to be beneficial for downstream capture

• Combined methods of Total Hg (AC) & Online Hg0

allowed trends in Hg 2+ to be observed

• SO3 measurements indicate very low concentrations 0.6-3.6ppm• May be indicator of acid condensation during transitions

Thanks for Listening

further questions?

rohan.stanger@newcastle.edu.au

Overview

• Current work at UoN• Experimental• Results• Key Findings

Comparison of SO3 measurements(& estimates)

Comparison of Hg measurements(by gas phase mass balancing-has advantage of online)

Comparison of Hg measurements(by ash balancing- a poor method for transition tests)

Fabric Filter Trial Outcomes• SO3 levels below detection limit in transitions• SO3 level 0.6 - 3.7ppm in steady state oxy-mode

• Hg affected by burner configuration

AIR OXYLow NOx Original Low NOx Original

HgTotal gas µg/m3 0.07 0.53 0.15 4.4* → 2

Hg2+ 0% 77% 65% 87%

HgTotal gas µg/m3

Corrected to 12%CO2

0.06 0.45 0.03 1.2* → 0.42

Hg Captured 98 % 93% 99% 93%

* Highest point in transition corresponding to higher O2

Oxyfuel research at UoN

• Part of Callide Oxyfuel Project Feasibility Study 2005-6• Coal reactivity 2007-9• Sulfur impacts 2009-11

– Review, SO3 formation, catalytic impacts, ADP

• NOx in compression 2012-15– Effect of pressure residence time, capture in H2O– Laboratory compression system, mass balancing

• Hg in compression 2012-15– Impact of NOx + ∆P + ∆t– Mass balancing → Recovery methods

• Emissions from condensates (depressurised)

Further Research Needs

• Hg-NOx product identification• Different sections of compression • Higher temperature (<200°C directly after ∆P, dry)• After cooling+ water condensation

• SO2-NOx combined capture in compression• Optimised• Formation of N2O minimised or accounted for

• CPU liquid product recovery

Research Needs in the COP CPU

• Australian context– Low Coal sulfur, no deSOx– Low Hg + Fabric Filter, no AC guard bed– No SCR/nSCR

• For a Australian retrofit– Caustic polishing at low pressure– CPU to passively remove NOx + Hg as condensates– Cold Box to remove remaining NOx from product CO2

and recycle•

ANLEC R&DAustralian National Low Emission CoalResearch & Development

• Combined Federal Government Funding & Australian Coal Association• Addresses:

– The near term risk reduction and technology developments necessary for successful demonstration of LECT in Australia.

– The delivery of skills, data and knowledge to assist key stakeholders understand the benefits and deployment risks of LECT's.

– Support for, and investigation of, issues affecting the performance of the early demonstration projects.

• Program overall funding……across 2012 to 2020 ASK NOEL

Publications• Stanger, R, T. Ting, L Belo, C Spero, T Wall (2015), Field measurements of NOx and mercury from oxy-fuel

compression condensates at the Callide Oxyfuel Project, International Journal of Greenhouse Gas Control 42: 485-493, 2015

• Stanger, R., T. Ting, C. Spero and T. Wall (2015). Oxyfuel derived CO2 compression experiments with NOx, SOx and mercury removal—Experiments involving compression of slip-streams from the Callide Oxyfuel Project (COP). International Journal of Greenhouse Gas Control 41: 50-59, 2015

• Stanger R, T Ting, T Wall, High pressure conversion of NOx and Hg and their capture as aqueous condensates in a laboratory piston-compressor simulating oxyfuel CO2 compression, International Journal of Greenhouse Gas Control 29, 2014

• Ting T, R Stanger, T Wall, Oxyfuel CO2 compression: The gas phase reaction of elemental mercury and NOx at high pressure and absorption in nitric acid, International Journal of Greenhouse Gas Control, 29 2014

• Ting, T., R. Stanger, and T. Wall, Laboratory investigation of high pressure NO oxidation to NO2 and capture with liquid and gaseous water under oxy-fuel CO2 compression conditions. International Journal of Greenhouse Gas Control, 18(0) 2013

• ANLECR&D Report website: www.anlecrd.com.au/projects/gas-quality-impacts-assessment-and-control-in-oxy-fuel