Hg 2014_v0_slideshare

transcript

Mercury Speciation Measurements from

Boiler to Stack

Presented by Clean Air Engineering

There is a growing need to measure speciatedmercury throughout air pollution control systems at coal-fired power plants.

There are two major drivers for this…

Emissions Standards for Boilers and Process Heaters and Commercial/Industrial Solid Waste Incinerators

aka, the Boiler MACT

National Emission Standards for Hazardous Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating Units

aka, the EGU MATS

Both of these rules require reductions in mercury emissions

So-called “co-benefit” control from existing air pollution control devices is one approach to reduce mercury emitted to the atmosphere

Another would be add-on controls, such as sorbent injection

APHFGD

ESP/FFSCR

Boiler

HgP Hg+2

Sorbent SorbentSorbentSorbent

700˚F + 130˚F +Flue Gas Temperature

Sorbent

Consequently, there is much interest in the fate of mercury across the various air pollution control systems of power plants

Co-benefit and add-on controls generally rely on operational measures to…

• Promote oxidation of Hg0 to Hg+2

• Promote adsorption of Hg onto particles

To understand these relationships requires knowledge of the different species of mercury throughout the APC system…

i.e., from boiler to stack

This is difficult thing to model…

It relies on a variety of factors…

Such as…

– coal

– ash

– flue gas temperature

– flue gas chemistry

– retention time in APCDs

– scrubber chemistry

– boiler operation

And previous efforts to measure it have been handicapped by a lack of standardization in the test methods

Not to mention the characteristics of the flue gas being measured…hot, dirty, interferences such as NH3, HBr, etc.

The traditional test method to measure speciated mercury is the Ontario Hydro Method(ASTM D6784)

It looks like this schematically

This method has a reputation of being cumbersome and difficult to reproduce...

And there is a known issue with mercury partitioning in the particulate fraction…

And it takes several hours to perform. The cost of data is relatively high.

All of which has led to the prevalence of using Sorbent

Trap Methods to measure Hg

EPA Method 30B is the sorbent trap

reference method

Method 30B

Spiked

Section 1 Section 2

Method 30B does not speciate Hg. For that, we turn to the FAMS approach (aka, Modified Method 30B)…

FAMS Method

Potassium Chloride

Activated Carbon

PM FilterGas Flow

Glass Wool Plug

Section 2 Section 3 Section 4 Section 5

Oxidized Mercury Section (S1-S3)

Glass Wool Plug

Section 1

Elemental Mercury Section (S4-S5)Particulate Mercury(semi-quantitative)

Hg0Hg+2

FAMS has its limitations though…

You can measure particulate-bound mercury with FAMS…

…by adding a filter ahead of the traps.

Filter KCl Trap Carbon Trap

But there is a problem with this when used upstream of particulate control.

Unless the duct is traversed and sampling is performed isokinetically, particle size variation will bias the particulate collected.

Does the particulate captured from this location using only one test port mean anything?

Probably not!

Does the particulate captured from this location using only one test port mean anything?

Probably not!

But often, that is the approach taken.

There are other limitations to FAMS, such as temperature…

Section 1 Section 2 Section 3 Section 4Section 5

PM Filter KCl CarbonAGS

The carbon should not exceed about 450°F

The KCl should stay below 250°F

There are also known flue gas matrix interferences…

Speciation TrapsLimitations

Such as NO/SO2 oxidation effects

NO + SO2 → NO2

Hg0 + NO2 → Hg2+

This is mitigated by the addition of an acid gas scrubber (AGS) ahead of the sorbents…

But do we know how the AGS affects speciation? The assumption is that only oxidized mercury collects on it. Is that true? Does the AGS promote any conversion?

And then there are the interactions between mercury, flue gas, and particulate that collects on the pre-filter…

For example…

Species Flyash Effects:

HgP Increase from adsorption of Hg2+ and Hg0

Hg2+ Increase from catalytic oxidation of Hg0

Hg0 Loss through adsorption

HClNOx

LOIIron{

And the location of the filter also has an effectns

In hot gases, being at the front can reduce these impacts.

And the location of the filter also has an effect

To combat the temperature limitations, some tests have been done with the traps located outside of the hot flue gas…

Like this...

Temperature controlled

300-700°F 250°F

APHFGD

ESPSCR

Sometimes this data does not make sense.

Location HgP

Run 1 SCR Inlet 1% 11% 7% 9% 9%

AH Inlet 25% 21% 88% 19% 19%

ESP Inlet 79% 44% 30% 13% 23%

WFGD Inlet 35% 6% 17% 6% 7%

Stack 15% 12% 20% 17% 18%

Run 2 SCR Inlet 66% 2% 42% 4% 5%

AH Inlet 32% 8% 2% 5% 8%

ESP Inlet 86% 8% 42% 6% 11%

WFGD Inlet 4% 1% 76% 3% 3%

Stack 5% 13% 12% 5% 5%

Run 3 SCR Inlet 44% 1% 54% 3% 4%

AH Inlet 22% 12% 2% 12% 12%

ESP Inlet 90% 43% 9% 24% 35%

WFGD Inlet 59% 4% 1% 3% 4%

Stack 1% 65% 2% 6% 6%

Relative Deviation

Paired Trap Agreement (%RD)

And paired trap agreement is poor

The problem is that mercury will re-partition and convert as it traverses the temperature gradient of the probe.

And what you collect in the cool traps is different than what was in the gas stream.

Problem Statement

So we need to design a sorbent trap probe to allow Hg speciation measurements in flue gas with:

– high dust

– high temperature

– reactive interferences

A solution is to filter hot, shorten the gas path prior to adsorption, and cool the traps in-situ.

Here is one way…

a Forced-Air Cooled Probe

300-700°F

And take some additional measures…

Minimize test duration

TemperatureTest Duration

↑↓

Particulate

Minimize sample flow rate

TemperatureSample Rate

↑↓

Particulate

Move pre-filter up front

Potassium Chloride

Activated Carbon

PM FilterGas Flow

Glass Wool Plug

Section 1

Add a dust shield

Also…

Placement of trap thermocouples is critical

Potassium Chloride

Activated Carbon

PM FilterGas Flow

Glass Wool Plug

Section 1

And…

Strict QA/QC is critical

Potassium Chloride

Activated Carbon

PM FilterGas Flow

Glass Wool Plug

Section 1

You cannot waive your hands at QA/QC for FAMS. Just as in 30B…

If the goal is to measure speciated mercury, PAIRED FAMS traps should be collected…

Common industry practice of pairing a total trap with a FAMS trap does not give adequate QC metrics.

These are important QC metrics for each species measured.

• Relative Deviation

• Breakthrough

• Spike Recovery

Our recommendations…

Specification

Applicable

Sample

Fraction Acceptance Criteria Frequency

Consequences if

Not Met

Paired sorbent

trap agreement

HgP None N/A N/A

Hg+2 ≤15% RD if Hg+2>1 μg/dscm, ≤25% RD or

≤0.25 μg/dscm absolute difference if

Hg+2≤1 μg/dscm

Every run Invalidate run for

Hg0 ≤10% RD if Hg0>1 μg/dscm, ≤20% RD or

≤0.2 μg/dscm absolute difference if Hg0≤1

μg/dscm

≤10% RD if Hgi>1 μg/dscm, ≤20% RD or

≤0.2 μg/dscm absolute difference if Hgi≤1

μg/dscm

HgT or Hgvap

Breakthrough & Spike Recovery

QA/QC Specification

Applicable

Sample

Fraction Acceptance Criteria Frequency

Consequences if Not

Oxidized

Breakthrough

(KCl sections)

Hg+2 Section 3 KCl contains ≤10% of

the Section 1 and 2 masses if

Hg+2>1 μg/dscm, Section 3

contains ≤20% of the Section 1

and 2 masses if Hg+2≤1 μg/dscm

Every trap Invalidate trap data for

Hg+2 and Hg0

Elemental

Breakthrough

(carbon sections)

Hg0 Section 2 carbon contains ≤10%

of the Section 1 mass if Hg0>1

μg/dscm, Section 2 contains ≤20%

of the Section 1 mass if Hg0≤1

μg/dscm,

Every trap Invalidate trap for Hg0,

HgT, and Hgvap

Field Recovery Test Hg0 Average recovery between 85%

and 115% for Hg0

Minimum three

spikes per

program

Flag data

Following these measures, good data that makes sense can be obtained…

APHFGD

ESPSCR

Location HgP

Run 1 SCR Inlet 2% 20% 7% 1% 1%

AH Inlet 1% 1% 3% 0% 0%

ESP Inlet 5% 0% 1% 0% 0%

WFGD Inlet 1% 1% 7% 1% 1%

Stack 13% 8% 6% 3% 5%

Run 2 SCR Inlet 1% 6% 3% 4% 4%

AH Inlet 0% 2% 6% 0% 0%

ESP Inlet 27% 1% 6% 1% 2%

WFGD Inlet 1% 2% 2% 3% 3%

Stack 8% 32% 3% 6% 6%

Run 3 SCR Inlet 0% 4% 0% 1% 1%

AH Inlet 1% 5% 11% 1% 1%

ESP Inlet 3% 0% 9% 0% 0%

WFGD Inlet 0% 1% 27% 2% 2%

Stack 3% 1% 8% 6% 8%

Relative Deviation

Breakthrough

Location Trap A Trap B Trap A Trap B

Run 1 SCR Inlet 0% 0% 0% 2%

AH Inlet 0% 1% 7% 0%

ESP Inlet 0% 0% 0% 0%

WFGD Inlet 0% 0% 0% 0%

Stack 0% 0% 8% 0%

AH Inlet 0% 0% 0% 0%

ESP Inlet 0% 0% 0% 0%

WFGD Inlet 0% 0% 0% 0%

Stack 0% 0% 0% 0%

AH Inlet 2% 0% 0% 0%

ESP Inlet 0% 0% 29% 0%

WFGD Inlet 0% 0% 0% 37%

Stack 0% 0% 15% 0%

Breakthrough

Fraction Hg0 Fraction

10% maximum (20% at stack) targeted acceptance limit

Spike Recovery (Hg0)

Location Run 1 Run 2

SCR Inlet 110.8% 87.7%

AH Inlet 93.6% 113.6%

ESP Inlet 101.4% 111.9%

WFGD Inlet 108.0% 101.5%

Stack 103.2% 100.1%

Program Average 103.4% 102.9%

85% - 115% targeted acceptance range

The take-away…FAMS has potential, but standardization of a FAMS protocol is needed to ensure good results across the industry.

There may be better alternatives…

Source: Apogee Scientific, Inc.

An Inertial Probe removes particulate without filtration

Source: Apogee Scientific, Inc.

A sorbent trap placed downstream of the inertial probe should not suffer from filtration effects

This approach needs more work to determine its viability.

What about Bromine?

Potassium Chloride

Activated Carbon

Activated CarbonGas Flow

Glass Wool Plug

Section 1

1.8ng2.0ng ND

7.6ng ND

Here are data from a trap downstream of a wet scrubber with brominated PAC added upstream

Not much mercury anywhere…

Potassium Chloride

Activated Carbon

Activated CarbonGas Flow

Glass Wool Plug

Section 1

1.8ng2.0ng ND

7.6ng ND

Here are data from a trap downstream of a wet scrubber with brominated PAC added upstream

Not much mercury anywhere…except in the prefilter

13.9ng

Bromine creates a problem for speciation, with or without particulate matter present.

A mercury monitor may be a solution

Source: Ohio Lumex Co.Source:ADA-ES, Inc.

Mercury AnalyzerConvertor

Scrubber

Hg0 + Hg+2

Hg0Hg+2→ Hg0

Hg CEMS measure Total Vapor-Phase Mercury downstream of the convertor

Hg0 + Hg+2

Scrubber

Hg0Hg+2→ Hg0

Hg0Hg0 + Hg+2

Or elemental mercury if the convertor is bypassed and the gas is scrubbed

Hg0 + Hg+2

Scrubber

Hg0Hg+2→ Hg0

Hg0Hg0 + Hg+2

Oxidized Mercury is the difference in the two configurations

Hgtotal

Hg-Elemental Hg-Oxidized Hg-Total

Summary

• Mercury speciation measurements are crucial 1st step in compliance strategy

• FAMS is practical method of choice

Summary

• Mercury speciation measurements are crucial 1st step in compliance strategy

• FAMS is practical method of choice

but…

• Temperature and particulate effects must be minimized to prevent speciation bias

• Rigorous QA/QC needs to be imposed

• Industry standardization needed

And…

• There’s more than one way to skin a cat…different forms of filtration and mercury CEMS need consideration.

For more information on measuring mercury and fate-of-

mercury balances, contact CleanAir today.800-991-3300

contact@cleanair.com

Thanks!