&EPA
United States
Environmental Protection
Agency
EPA-600/R-04/032
April 2004	
        Effect of Selective
        Catalytic Reduction on
        Mercury, 2002
        Studies Update

-------
                                          EPA-600/R-04/032
                                          April 2004
Effect of Selective Catalytic  Reduction
on  Mercury, 2002 Field Studies  Update
                            by
                         D. Laudal
               Energy & Environmental Research Center
                    University of North Dakota
                        PO Box 9018
                   Grand Forks, ND 58202-9018
               EPA Cooperative Agreement R-829353-01
                 EPA Project Officer: Chun Wai Lee
                U.S. Environmental Protection Agency
             National Risk Management Research Laboratory
              Air Pollution Prevention and Control Division
                 Research Triangle Park, NC 27711

-------
Abstract
This report summarizes field measurements at six power plants with selective catalytic
reduction (SCR) conducted in 2001 and 2002. Mercury (Hg) measurements were completed
at four coal-fired power plants with SCR systems in 2001, published as EPA-600/R-02/096,
and additional measurements were conducted in 2002 at four plants with SCR, including two
plants tested in 2001 that showed significant Hg oxidation. Speciated Hg concentrations in
flue gas were sampled and evaluated using the wet-chernistry Ontario Hydro method, as well
as near-real-time Hg semicontinuous emission monitors. Sampling was conducted at these
plants at the inlet and outlet of the SCR reactor to evaluate the effects of SCR on Hg
speciation, as well as the  inlet and outlet of the particulate and sulfur dioxide control devices
to evaluate Hg capture. Additional sampling involved the use of selective condensation to
measure sulfur trioxide and EPA Method 27 for ammonia slip. Fly ash, flue gas
desulfurization (FGD) solids, and coal samples were also collected to estimate the Hg mass
balance across the control devices.

These results indicate that SCRs can increase Hg oxidation and improve Hg removal in the
downstream FGD. SCR catalysts appear to assist in converting elemental mercury (Hg°) to
oxidized mercury (Hg2+). This effect appears to be more likely to occur with bituminous
coals, where greater than 90% Hg2+ is possible at the particulate control device inlet. The
three bituminous coal-fired power plants tested with wet FGD systems achieved Hg removals
of 84%-92% with SCR operation, as compared with 43%-51 % without SCR operation.
These increased removal  efficiencies may be due to the combined effects of the SCR system
to increase Hg2+ concentrations and reduce reemissions of the Hg° from the FGD system.
The effect of catalyst space velocity and age are not clear but may have an impact on SCR
Hg oxidation. The only Powder River Basin (PRB) site tested did not show a high oxidation
rate. It is important to note that these findings are based on a relatively small data set and,
thus, should be considered preliminary rather than final conclusions that can be extrapolated
to predict the results at all other similar units. For example, two of the three FGDs tested
were magnesium-lime systems, and the third FGD was a venturi scrubber; thus the combined
effect of SCR and the most common FGD design of a limestone, forced-oxidation system has
yet to be evaluated.

Additional field measurements are being conducted in 2003 to better understand the effects
of coal and catalyst properties and to better characterize longer-term FGD Hg removal,
including the possible impact of SCRs on Hg° re-emissions across the FGD. Full-scale and
side-stream tests are planned by this project team as well as in a separate DOE/Consol study
to further evaluate the combined effect of SCRs and FGDs on Hg removal. To evaluate the
effect of coal properties, measurements are planned at a pulverized-coal-fired power plant
burning a PRB coal, with a second PRB site to be tested around January 2004. Additional
follow-on tests to evaluate catalyst-aging effects  are planned at the two power plants that
indicated significant Hg oxidation in the 2001 tests and were retested in 2002. Tests are also
being conducted at a power plant burning a blend of bituminous and PRB coals. Thus, the
results in this report should be viewed as work in progress, and the reader is encouraged to
follow up and read future reports.

-------
Foreword
   The U.S.  Environmental  Protection Agency (EPA)  is  charged  by Congress with
protecting the Nation's  land, air, and  water resources.  Under a  mandate  of national
environmental laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural  systems to support
and nurture  life.  To meet this mandate, EPA's  research program is  providing data and
technical  support for solving environmental problems  today and  building  a science
knowledge base  necessary to manage  our ecological  resources wisely, understand how
pollutants affect our health, and prevent or reduce environmental risks in the future.

   The National  Risk Management Research Laboratory (NRMRL) is the Agency's center
for investigation of technological and management approaches for preventing and reducing
risks from pollution  that threaten  human  health and the environment. The focus of the
Laboratory's research  program is  on  methods and their cost-effectiveness for prevention
and control of pollution to air, land, water, and subsurface resources;  protection of water
quality in public water systems; remediation of contaminated sites, sediments  and ground
water; prevention and control of indoor air pollution; and restoration of ecosystems. NRMRL
collaborates with both  public and  private sector partners to foster technologies that reduce
the cost of compliance and to anticipate emerging problems. NRMRL's research provides
solutions to environmental problems by: developing and  promoting technologies that protect
and improve the  environment; advancing scientific and  engineering information to support
regulatory and policy decisions; and providing the technical support and  information transfer
to ensure implementation of environmental regulations and strategies at the national, state,
and community levels.

   This publication  has been produced as  part of  the  Laboratory's  strategic long-term
research plan. It is published and  made available by  EPA's Office of  Research and
Development to assist the user community and to  link researchers with their clients.
                                 Lawrence W. Reiter, Acting Director.
                                 National Risk Management Research Laboratory
                                                                                in

-------
Acknowledgment
This report was prepared with the support of EPRI, the U.S. Department of Energy (DOE)
National Energy Technology Laboratory under Cooperative Agreement No. DE-FC26-
98FT40321, and the U.S. Environmental Protection Agency (EPA) under Cooperative
Agreement No. 82935301. However, any opinions, findings, conclusions, or
recommendations expressed herein are those of the author and do not necessarily reflect the
views of EPRI, DOE, or EPA.

The author also wishes to acknowledge the following: Western Kentucky University for
providing the sampling team and equipment for site s4, and WE Energies for providing
analytical support for site s2.

Also, the authors would like to thank the power plants for allowing testing at their sites and
providing the support needed to do the work.
IV

-------
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

TOHIS DOCUMENT  WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC  POWER RE-
SEARCH INSTITUTE,  INC. (EPRI). NEITHER  EPRI, ANY MEMBER OF  EPRI, ANY CO-
SPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF
ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS  OR IMPLIED,
(I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS,
OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON
OR  INTERFERE  WITH  PRIVATELY  OWNED  RIGHTS,  INCLUDING  ANY  PARTY'S
INTELLECTUAL PROPERTY, OR (III) THAT  THIS  DOCUMENT IS SUITABLE TO ANY
PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES,  EVEN IF EPRI OR ANY EPRI  REPRE-
SENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING
FROM YOUR  SELECTION OR USE OF THIS DOCUMENT  OR ANY INFORMATION,
APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

Energy & Environmental Research Center

DOE AND EPA DISCLAIMER

THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY AN AGENCY
OF THE UNITED STATES GOVERNMENT. NEITHER THE UNITED STATES GOVERNMENT,
NOR ANY AGENCY THEREOF, NOR ANY OF THEIR EMPLOYEES MAKES ANY  WARRANTY,
EXPRESS OR IMPLIED,  OR ASSUMES ANY LEGAL LIABILITY OR RESPONSIBILITY FOR THE
ACCURACY, COMPLETENESS, OR USEFULNESS OF ANY INFORMATION,  APPARATUS,
PRODUCT,  OR PROCESS DISCLOSED, OR REPRESENTS THAT  ITS  USE WOULD NOT
INFRINGE  PRIVATELY OWNED  RIGHTS.  REFERENCE HEREIN TO  ANY SPECIFIC
COMMERCIAL  PRODUCT,  PROCESS,  OR  SERVICE BY TRADE NAME, TRADEMARK,
MANUFACTURER, OR OTHERWISE DOES NOT NECESSARILY CONSTITUTE OR IMPLY ITS
ENDORSEMENT,  RECOMMENDATION,  OR  FAVORING BY  THE  UNITED STATES
GOVERNMENT OR ANY AGENCY THEREOF.  THE VIEWS AND OPINIONS OF AUTHORS
EXPRESSED HEREIN DO NOT NECESSARILY STATE OR REFLECT THOSE OF  THE UNITED
STATES GOVERNMENT OR ANY AGENCY THEREOF.

THIS REPORT IS AVAILABLE  TO THE  PUBLIC  FROM THE NATIONAL  TECHNICAL
INFORMATION SERVICE, U.S. DEPARTMENT OF COMMERCE, 5285 PORT ROYAL ROAD,
SPRINGFIELD, VA 22161; PHONE ORDERS ACCEPTED AT (703) 487-4650.

EERC DISCLAIMER

LEGAL NOTICE THIS  RESEARCH  REPORT WAS  PREPARED  BY THE  ENERGY &
ENVIRONMENTAL RESEARCH  CENTER (EERC), AN AGENCY  OF THE UNIVERSITY OF
NORTH DAKOTA, AS AN ACCOUNT OF WORK SPONSORED BY  EPRI,  DOE, AND EPA.
BECAUSE OF THE RESEARCH NATURE OF THE WORK PERFORMED, NEITHER THE EERC
NOR ANY OF  ITS EMPLOYEES MAKES ANY WARRANTY, EXPRESS OR  IMPLIED, OR
ASSUMES  ANY  LEGAL  LIABILITY OR  RESPONSIBILITY  FOR  THE   ACCURACY,
COMPLETENESS, OR USEFULNESS OF ANY INFORMATION, APPARATUS, PRODUCT, OR
PROCESS  DISCLOSED, OR  REPRESENTS  THAT  ITS USE  WOULD  NOT INFRINGE
PRIVATELY  OWNED RIGHTS.  REFERENCE HEREIN  TO ANY SPECIFIC COMMERCIAL
PRODUCT, PROCESS, OR SERVICE BY TRADE NAME, TRADEMARK, MANUFACTURER, OR
OTHERWISE DOES NOT NECESSARILY CONSTITUTE OR IMPLY ITS ENDORSEMENT OR
RECOMMENDATION BY THE EERC.

-------

-------
CONTENTS
ABSTRACT	ii

ACKNOWLEDGMENT	iv

EXECUTIVE SUM MARY	xix

11NTRODUCTION	1-1
   1.1   Potential Impacts of SCR on Mercury Speciation	1-1
   1.2   Pilot-Scale Screening Tests Conducted at the EERC	1-2
   1.3   2001 SCR Mercury Field-Sampling Project	1-3
   1.4   Project Goals and Objectives	1-5
   1.5   Sampling Approach	1-6
     1.5.1   Mercury Sampling Using the Ontario Hydro Mercury Speciation
            Method	1-6
     1.5.2   Mercury Sampling Using Hg SCEMs	1-6
     1.5.3   Other Flue Gas Analyses	1-7
     1.5.4   Mass Balance	1-7
     1.5.5   Plant Operation Data	1-7

2SITES2	2-1
   2.1   Site Description and Configuration	2-1
   2.2   Sampling Approach	2-2
     2.2.1   Flue Gas Sample Streams	2-3
     2.2.2   Other Sample Streams	2-3
   2.3   Process Operating Conditions	2-4
   2.4   Sampling Results	2-7
     2.4.1   Ontario Hydro Flue Gas Mercury Results	2-7
     2.4.2   HgSCEM Results	2-9
vn

-------
     2.4.3   Coal Analysis Results	2-11
     2.4.4   ESP Ash and FGD Mercury Results	2-11
     2.4.5   NH3 Slip and SO3 Flue Gas Results	2-13
   2.5   Mercury Mass Balance	2-13
   2.6   General Observations from S2	2-14

3SITES4	3-1
   3.1   Site Description and Configuration	3-1
   3.2   Sampling Approach	3-2
     3.2.1    Flue Gas Sample Streams	3-3
     3.2.2   Other Sample Streams	3-3
   3.3   Process Operating Conditions	3-3
   3.4   Sampling Results	3-5
     3.4.1    Ontario Hydro Flue Gas Mercury Results	3-5
     3.4.2   HgSCEM Results	3-7
     3.4.3   Coal Analysis Results	3-9
     3.4.4   Mercury Collected by the Venturi Scrubber	3-10
     3.4.5   NH3 Slip and SO3 Flue Gas Results for Site S4	3-11
   3.5   Mercury Mass Balance	3-11
   3.6   General Observations from S4	3-12

4SITES5	4-1
   4.1   Site Description and Configuration	4-1
   4.2   Sampling Approach	4-3
     4.2.1    Flue Gas Sample Streams	4-3
     4.2.2   Other Sample Streams	4-4
   4.3   Process Operating Conditions	4-4
   4.4   Sampling Results	4-9
     4.4.1    Ontario Hydro Flue Gas Mercury Results	4-9
     4.4.2   HgSCEM Results	4-11
     4.4.3   Coal Analysis Results	4-14
     4.4.4   ESP Ash Mercury Results	4-14
     4.4.5   NH3 Slip and SO3 Flue Gas Results	4-15
   4.5   Mercury Mass Balance	4-16
   4.6   General Observations from S5	4-16
Vlll

-------
5SITES6	5-1
   5.1  Site Description and Configuration	5-1
   5.2  Sampling Approach	5-4
     5.2.1   Flue Gas Sample Streams	5-4
     5.2.2   Other Sample Streams	5-5
     5.2.3   Process Operating Conditions	5-5
   5.3  Sampling Results	5-11
     5.3.1   Ontario Hydro Flue Gas Mercury Results	5-11
     5.3.2   HgSCEM Results	5-13
     5.3.4   Coal Analysis Results	5-19
     5.3.5   ESP Ash Analysis	5-20
     5.3.6   NH3 Slip and SO3 Flue Gas Results	5-20
   5.4  Mercury Mass Balance	5-21
   5.5  General Observations from S6	5-22

6 DISCUSSION OF OVERALL RESULTS	6-1
   6.1  The Change in Mercury Oxidation Across the SCR Catalysts	6-3
   6.2  Effect of the SCR on Mercury Oxidation	6-5
   6.3  Effect of SCR Catalyst Age on Mercury Speciation	6-6
   6.4  SCR/Wet FGD Combination for Mercury Control	6-8

7 CONCLUSIONS	7-1
   Future Test Plans	7-1

8 QUALITY ASSURANCE/QUALITY CONTROL	8-1
   8.1  Process Data Evaluation	8-1
   8.2  Sampling Quality Control Evaluation	8-2
   8.3  Evaluation of Measurement Data Quality	8-3
   8.4  Ontario Hydro Method Error Analysis	8-5

PREFERENCES	9-1

A SAMPLING METHODS AND PROCEDURES	A-1
   Ontario Hydro Mercury Speciation Method (OH method)	A-1
   Continuous Mercury Monitors	A-4
     Atomic Fluorescence-Based Hg SCEMs	A-4
                                                                            IX

-------
     Atomic Absorption-Based Hg SCEMs	A-4
     Flue Gas Pretreatment/Conversion	A-5
  Auxiliary Flue Gas Measurements	A-6
     O2 Determination	A-6
     CO2 Determination	A-6
  Chlorides, NH3, and SO3	A-6
  Reference	A-7

B MERCURY MEASUREMENTS	B-1
  B.1   Mercury Measurements Made at Site S2	B-1
     Complete Ontario Hydro Data Set	B-1
     Coal Mercury and Chloride Analyses	B-2
  B.2   Mercury Measurements Made at Site S4	B-3
     Complete Ontario Hydro Data Set	B-3
  B.3   Mercury Measurements Made at Site S5	B-5
     Complete Ontario Hydro Data Set	B-5
  B.4   Mercury Measurements Made at Site S6	B-8
     Complete Ontario Hydro Data Set	B-8

C COMPLETE AUXILIARY FLUE GAS DATA FOR ALL SITES	C-1

D QUALITY ASSURANCE/QUALITY CONTROL	D-1
  Ontario Hydro (OH)  Method	D-1
     Instrument Setup and Calibration	D-1
     Presampling Preparation	D-2
     Glassware and Plasticware Cleaning and Storage	D-2
     Analytical Reagents	D-2
     Blanks and Spikes	D-2
  QA/QC Checks for Data Reduction and Validation	D-9
     Data Reduction	D-9
     Data Validation	D-9
  Sample Identification and Chain of Custody	D-9
  Personnel Responsibilities and Test Schedule	D-9
     Test Site Organization	D-9
  Test Preparations	D-10
     Construction of Special Sampling Equipment and Modifications to the Facility	D-10

-------
     General Services Provided by the Facility	D-10
     Access to Sampling Sites	D-10
     Sample Recovery Areas	D-10
  Test Personnel Responsibilities and Detailed Schedule	D-10

£ SAMPLE CALCULATIONS	E-1
  Volume of Gas Sample	E-1
  Volume of Water Vapor	E-1
  Water Vapor in the Gas Stream	E-2
  Dry Molecular Weight	E-2
  Molecular Weight	E-2
  Average Stack Gas Velocity	E-2
  Isokinetic Sampling Rate	E-3
  Volume of Gas Sample Corrected to 3% O2	E-4
  Mercury	E-4
  Fd	E-4
                                                                            XI

-------

-------
LIST OF FIGURES
Figure ES-1 Mercury Results Comparing Speciation with SCR from 2001 and 2002
    at Site S2	xxiv
Figure ES-2 Mercury Results Comparing Speciation with SCR from 2001 and 2002
    at Site S4	xxv
Figure ES-3 Comparisons of Mercury Speciation with the SCR in Service and with
    the SCR  Bypassed at Site S4	xxvi
Figure ES-4 Mercury Speciation Results Compared for a Unit with SCR and Without
    SCR at Site S5	xxvii
Figure ES-5 Mercury Speciation Results Comparing Units with and Without SCRs in
    Service at Site S6	xxviii
Figure ES-6 Mercury Concentrations at the Inlet of the Particulate Control Device
    with and Without the SCR, respectively	xxix

Figure 2-1  Schematic of Site S2 Showing Sample  Locations from Horizontal and
    Vertical Perspectives	2-2
Figure 2-2 Plant Operation Data for Site S2	2-5
Figure 2-3 Comparison of Mercury Speciation Results 2001 and 2002 for Site S2	2-8
Figure 2-4 Hg SCEM Results for Site S2	2-10
Figure 2-5 Average Hg2+ as  Measured by Hg SCEMs (total  Hg - Hg°) for Site S2	2-11
Figure 3-1  Schematic of Site S4 Showing Sample  Locations from a Vertical and
    Horizontal Perspective	3-2
Figure 3-2 Plant Operation Data for Site S4	3-4
Figure 3-3 Comparison of Mercury Speciation Results with the SCR in Service and
    with the SCR Bypassed	3-6
Figure 3-4 Comparison of Mercury Speciation Results 2001 and 2002 for Site S4	3-7
Figure 3-5 Hg SCEM Results for Site S4	3-8
Figure 3-6 Average Hg2+as  Measured by Hg SCEMs (Total Hg-Hg°) for Site S4	3-9
Figure 4-1  Schematic of Site S5 Showing Sample  Locations for the Unit with  the
    SCR from a Vertical and Horizontal Perspective	4-2
Figure 4-2 Schematic of Site S5 Showing Sample  Locations for the Unit with  No SCR
    from a Vertical and Horizontal Perspective	4-3
Figure 4-3 Plant Operation Data for Site S5 for the Unit with the SCR	4-5
Figure 4-4 Plant Operation Data for Site S5 for the Unit with No SCR	4-7
Figure 4-5 Comparison of Mercury Speciation Results with the SCR and Without an
    SCR at Site S5	4-10

xiii

-------
Figure 4-6 Hg SCEM Results for Site S5 for the Unit with an SCR	4-12
Figure 4-7 Hg SCEM Results for Site S5 for the Unit with No SCR	4-13
Figure 5-1 Schematic of Site S6 Showing Sample Locations for Unit 1 with the SCR
    in Service from a Vertical and Horizontal Perspective	5-2
Figure 5-2 Schematic of Site S6 Showing Sample Locations for Unit 2 with the SCR
    Bypassed from a Vertical and Horizontal Perspective	5-3
Figure 5-3 Schematic of Site S6 Showing Sample Locations for Unit 4 with No SCR
    from a Vertical and Horizontal Perspective	5-4
Figure 5-4 Plant Operation Data for Site S6 for Unit 1 with the SCR in Service	5-6
Figure 5-5 Plant Operation Data for Site S6 for Unit 2 with the SCR Bypassed	5-8
Figure 5-6 Plant Operation Data for Site S5 for Unit 4 with No SCR	5-10
Figure 5-7 Comparison of Mercury Speciation Results for the Three Test Units of
    SiteS6	5-13
Figure 5-8 Hg SCEM Results for Site S6 for Unit 1 with the SCR in Service	5-14
Figure 5-9 Hg SCEM Results for Site S6 for Unit 2 with the SCR Bypassed	5-15
Figure 5-10 Hg SCEM Results for Site S6 for Unit 4 with  No SCR	5-16
Figure 5-11 Average Hg2+ as Measured by Hg SCEMs (total Hg - Hg°) for Site S6
    Unitl (SCR on-line)	5-17
Figure 5-12 Average Hg2+ as Measured by Hg SCEMs (total Hg - Hg°) for Site S6
    Unit 2 (SCR bypassed)	5-18
Figure 5-13 Hg SCEM Results for Site S6 for Unit 2 with  SCR Bypassed	5-18
Figure 6-1 Percent of Oxidized Hg2+ at the  Inlet of the SCR System as a Function of
    Chloride Content of the Coal	6-4
Figure 6-2 Mercury Speciation Results Comparing All the Sites Tested Firing Eastern
    Bituminous Coal	6-5
Figure 6-3 Comparison of Mercury Speciation Results from 2001 and 2002 at Site S2	6-7
Figure 6-4 Comparison of Mercury Speciation Results from 2001 and 2002 at Site S4	6-7

Figure A-1 Schematic of the Ontario Hydro Mercury Speciation Train	A-1
Figure A-2 Sample Recovery Scheme for the Ontario Hydro Mercury Speciation
    Train	A-3
Figure A-3 Schematic of the EERC Pretreatment/Conversion System for Use with
    Hg SCEMs	A-5
xiv

-------
LIST OF TABLES
Table ES-1 Summary of SCR Program Plant Configurations	xxii
Table ES-2 Summary of Coal Analyses for Plants Tested in 2001 and 2002	xxiii
Table ES-3 Change in Mercury Oxidation across the SCR Catalyst	xxix
Table ES-4 Effect of the SCR on Hg° Concentration Across the Wet FGDs	xxxi
Table 1-1 Summary of Plant Configuration from 2001 Test Program	1-4
Table 2-1 Sampling Test Matrix for Site S2	2-3
Table 2-2 Average Auxiliary Flue Gas Data for Site S2	2-7
Table 2-3 Ontario Hydro Average and Percentage of Total Mercury Results for S2	2-7
Table 2-4 Statistical Variation of Mercury with and  Without the SCR in Service
    Based on Hg SCEM Data for Site S2	2-9
Table 2-5 Coal Analysis for Site S2	2-12
Table 2-6 Analysis of ESP Hopper Ash and FGD Material for Site S2	2-12
Table 2-7 NH3 Slip and SO3 Results  at Site S2	2-13
Table 3-1 Sampling Test Matrix for Site S4	3-3
Table 3-2 Average Auxiliary Flue Gas Data for Site S4	3-5
Table 3-3 Average Ontario  Hydro Mercury and Results for Site S4	3-6
Table 3-4 Statistical Variation of the  Mercury with and Wthout the SCR in Service
    Based on the Hg SCEM Data for Site S4	3-9
Table 3-5 Coal Analysis for Site S4	3-10
Table 3-6 Partitioning of Mercury in Material Collected from Venturi Scrubber	3-11
Table 3-7 S4 Flue Gas, NH3Slip, and SO3 Results for Site S4	3-11
Table 3-8 Average Mercury Emission Factors for Site S4	3-12
Table 4-1 Sampling Test Matrix for Site S5	4-4
Table 4-2 Auxiliary Flue Gas Data for Site  S5	4-9
Table 4-3 Average and Percentage of Total Ontario Hydro Mercury Results for S5	4-10
Table 4-4 Statistical Variation of the  Mercury with and Wthout the SCR in Service
    Based on the Hg SCEM Data for Site S5	4-11
Table 4-5 Coal Analysis for Site S5	4-14
Table 4-6 Analysis of ESP Hopper Ash	4-15
Table 4-7 Flue Gas SO3 and NH3 Results for Site S5	4-15
Table 4-8 Average Mercury Emission Factors for Site S5	4-16
Table 5-1 Specifications of Site S6 Units	5-1
xv

-------
Table 5-2  Sampling Test Matrix for Site S6	5-5
Table 5-3  Auxiliary Flue Gas Data for Site S6	5-11
Table 5-4  Average and Percentage of Total Ontario Hydro Mercury Results forS6	5-12
Table 5-5  Statistical Variation of the Mercury Results Based on the Hg SCEM Data
    for Site S6	5-17
Table 5-6  Coal Analysis for Site S6	5-19
Table 5-7  Analysis of ESP Hopper Ash	5-20
Table 5-8  Flue Gas SO3 and NH3 Results for Site S6	5-21
Table 5-9  Average Mercury  Emission Factors for Site S6	5-21
Table 6-1  Summary of SCR Program Plant Configuration	6-2
Table 6-2  Average Analysis of Coals Fired During 2001 and 2002 Field Tests	6-3
Table 6-3  Change in Mercury Oxidation Across the SCR Catalyst	6-4
Table 6-4  Net Change in Hg2+ as Measured at the Inlet to the Particulate Control
    Device	6-6
Table 6-5 Effect of the SCR on Hg° Concentration Across the Wet FGDs	6-8
Table 8-1  Elements of the QA/QC Plan	8-4

Table A-1 Sample Train Components for the Ontario Hydro Method	A-2
Table B-1 Ontario Hydro Mercury Data for Site S2 with the SCR in Service	B-1
Table B-2 Coal Analysis Completed at Site S2	B-2
Table B-3 Ontario Hydro Mercury Data for Site S4 with the SCR In Service	B-3
Table B-4 Ontario Hydro Mercury Data for Site S4 with the SCR Bypassed	B-4
Table B-5 Ontario Hydro Mercury Data for Site S5 for Unit with the SCR	B-5
Table B-6 Ontario Hydro Mercury Data for Site S5 for Unit Without an SCR	B-6
Table B-7 Coal Mercury and Chloride Analyses	B-7
Table B-8 Ontario Hydro Mercury Data for Site S6 for Unit 1 (SCR)	B-8
Table B-9 Ontario Hydro Mercury Data for Site S6 for Unit 2 (SCR bypassed)	B-9
Table B-10 Ontario Hydro Mercury Data for Site S6 for Unit 4 (no SCR)	B-10
Table B-11 Coal Mercury and Chloride Analyses for Site S6	B-10
Table C-1 Auxiliary Flue Gas Data for Site S2 with SCR in Service	C-1
Table C-2 Auxiliary Flue Gas Data for Site S4 with SCR in Service	C-2
Table C-3 Auxiliary Flue Gas Data for Site S4 with SCR Bypassed	C-3
Table C-4 Auxiliary Flue Gas Data for Site S5 for the Unit with an SCR	C-4
Table C-5 Auxiliary Flue Gas Data for Site S5 for the Unit Wthout an SCR	C-5
Table C-6 Auxiliary Flue Gas Data for Site S6 for Unit 1 (SCR)	C-6
Table C-7 Auxiliary Flue Gas Data for Site S6 for Unit 2 (SCR bypassed)	C-7
Table C-8 Auxiliary Flue Gas Data for Site S6 for Unit 4 (no SCR)	C-8
Table D-1  Results of Mercury Speciation Field  Blanks at Site S2	D-3
Table D-2  Results of Mercury Speciation Field  Spikes at Site S2	D-4
xvi

-------
Table D-3 Results of Mercury Speciation Field Spikes at Site S4	D-5
Table D-4 Results of Mercury Speciation Field Blanks at Site S5	D-5
Table D-5 Results of Mercury Speciation Field Spikes at Site S5	D-6
Table D-6 Results of Mercury Speciation Field Blanks at Site S6	D-7
Table D-7 Results of Mercury Speciation Field Spikes at Site S6	D-8
Table D-8 Key Project Personnel	D-11
Table D-9 Test Personnel and Responsibilities	D-12
Table D-10 Typical Test Schedule for a 4-Week Project	D-13
                                                                                 xvn

-------
LIST OF ABBREVIATIONS
AA
AF
CEM
CI
C12
CO2
CVAA
DOE
EERC
EPA
ESP
Fd
FGD
Hg
Hg°
Hg2+
LOT
MW
N
NH3
NOX
OH
pc
PM2.5
PRB
PSA
QA
QC
QMS
Hg SCEM
SCR
SG
SNCR
SO3
SSTP
TiO2
V205
atomic absorption
atomic fluorescence
continuous emission monitor (refers to plant systems)
confidence interval
chlorine
carbon dioxide
cold-vapor atomic absorption
U.S. Department of Energy
Energy & Environmental Research Center
U.S. Environmental Protection Agency
electrostatic precipitator
emission factors calculated from coal analysis - dscf/106 Btu
flue gas desulfurization
mercury
elemental mercury
oxidized mercury
loss on ignition
megawatt
normal is defined at 20°C and 1 atmosphere pressure
ammonia
nitrogen oxide
Ontario Hydro mercury speciation
pulverized coal
particulate matter less than 2.5 jim
Powder River Basin
PS Analytical
quality assurance
quality control
quality management system
mercury semicontinuous emission monitor
selective catalytic reduction
Smith Greenburg
selective noncatalytic reduction
sulfur trioxide
site-specific test plan
titanium dioxide
vanadium oxide
xvin

-------
EXECUTIVE SUMMARY
Introduction

The objective of this report is to document the results and provide a summary of the 2001
and 2002 field tests associated with the "Selective Catalytic Reduction Mercury Field
Sampling Project." The testing was sponsored by EPRI, with additional funds provided by
the utility industry, the U.S. Department of Energy (DOE) National Energy Technology
Laboratory, and the U.S. Environmental Protection Agency (EPA) National Risk
Management Research Laboratory. Hg measurements were completed at six different power
plants, four in 2001 and two in 2002, equipped with selective catalytic reduction (SCR). In
addition, two of the plants tested in 2001 were retested in 2002 for a total of eight data sets.
Testing was also conducted in 2001 at two facilities that employed flue gas conditioning to
improve electrostatic precipitator (ESP) performance and one that used selective noncatalytic
reduction for nitrogen oxide control.

Coal combustion by electric utilities is a large source of anthropogenic mercury (Hg)
emissions in the United States, according to EPA [1]. Recent data indicate  that the total Hg
emission from coal-fired power plants in the United States is about 45 tons/yr [2]. EPA views
Hg from coal-fired utilities as a potential public health concern [3] and, as  a result, is
currently involved in a rule-making process that would require Hg control  for coal-fired
electric utilities by 2008.

Hg emissions from coal-fired boilers can be empirically classified, based on the capabilities
of currently available analytical methods, into three main chemical forms:  elemental mercury
(Hg°), oxidized mercury (Hg2+), and particle-bound Hg (Hgp). These impending Hg regula-
tions require that control strategies be investigated and developed.  The efficiency of Hg con-
trol methods depends largely on the form of Hg (gas vs. particulate) and species of Hg (ele-
mental vs. oxidized) formed upstream  of the control devices. Hgp can be removed from flue
gas by conventional  air pollution control devices such as an ESP or a baghouse. Hg2+
compounds are readily captured in flue gas desulfurization (FGD) units. Hg°  is  most likely to
escape air pollution control devices and be emitted to the atmosphere. Hg°, Hg2+, and Hgp
concentrations are much varied in flue gas, depending on the coal composition, combustion
conditions, and flue gas quench rate. Understanding the speciation of Hg is critical because
control options rely heavily on Hg's form or species. The concentration of Hg°, Hg2+, and
Hgp in the flue gas primarily depends on coal composition and combustion conditions [4].

In addition to Hg, coal-burning power plants are a significant anthropogenic source of
nitrogen oxides (NOx) emissions to the atmosphere. NOx emissions are an environmental
concern primarily because they are precursors to acid precipitation and are involved in

xix

-------
atmospheric reactions that produce fine particles and ozone. The most common NOx
reduction strategy is the use of low-NOx burners. These burners have the capability of
reducing NOx emissions by 40%-60%. However, with possible establishment of stricter
ozone regulations, fine paniculate (PM2.5), and regional haze, there is increased incentive to
reduce NOx emissions to a level below what can be achieved using low-NOx burners. SCR
technology, which can reduce NOx emissions by more than 90%, is, therefore, becoming
more attractive, particularly because catalyst costs continue to decrease and the knowledge
base for using SCR reactors is expanding. It is planned that approximately 100 gigawatts of
coal-fired capacity will have SCR for NOX by 2005 [5].

Potential Impacts of NOx SCR on Mercury Speciation

SCR units achieve lower NOx emissions by reducing NOx to N2 and H2O in the presence of
ammonia. These NOX reactions with SCR are catalyzed by metal oxides such as titanium
dioxide-supported vanadium pentoxide. These SCR units are operated at about 650-750 °F
(340-399 °C). Pilot- and full-scale experience in both the United States and Europe has
indicated that SCR catalysts promote the formation of Hg2+ [6-8]. Therefore, the use of SCR
to reduce NOX emissions has the potential to improve the Hg control efficiency of existing
particulate removal and FGD  systems by promoting Hg2+ formation. Possible mechanisms
that could result in the SCR of NOx impacting Hg speciation include:

•  Catalytically oxidizing the Hg.
•  Changing the flue gas chemistry.
•  Providing additional residence time.

EERC Pilot-Scale Tests (conducted in 2000)

In an attempt to evaluate the effects of SCR on Hg speciation, pilot-scale tests were
conducted at the Energy & Environmental Research Center (EERC) [7]. The general
conclusion reached based on these tests was that SCR has the potential to impact Hg
speciation, but that the effects were coal-dependent. Because of the inherent concerns related
to small pilot-scale tests (surface area-to-volume ratios, different flue gas chemistries, and
time and temperature profiles), the project advisory and research team concluded it was
necessary to conduct sampling at full-scale  power plants. Therefore, EPRI, DOE, EPA, and a
number of utilities began funding the EERC and other contractors to conduct Hg sampling at
power plants with SCR technology.

2001 SCR Mercury Field Sampling Project

The 2001 test program was developed to address the limitations of pilot-scale testing by
applying information obtained from previous work to full-scale electric-generating facilities.
In general, data from 2001 testing indicated that Hg oxidation can be enhanced by SCR
operation, but the effect may be moderated  by a variety of factors, including coal type,
catalyst chemistry and structure, and space velocity. Significant differences in Hg speciation
were observed  among plants with similar coal classifications [8].
xx

-------
Four sites with SCR systems were tested in 2001. Three of these sites fired eastern
bituminous coals and one a Powder River Basin (PRB) coal. Note that for purposes of this
report, the PRB site is referred to as Site SI and the other three as Sites S2-S4. However,
because the PRB site used a cyclone boiler and was operated such that the ash contained a
very high concentration of unburned carbon, it was not considered representative of a typical
PRB site. Two of the three sites that fired eastern bituminous coals showed a significant
increase in Hg oxidation across the SCR unit. These two sites resulted in 89% and 90% Hg
removal downstream of an FGD system. The other test site fired a coal that generated a very
high concentration of Hg2+ at the economizer outlet prior to SCR.

Upon review of the 2001 test results, it was evident that additional data would be necessary
to quantify the effect SCR operation had on Hg oxidation given the diversity of power plant
configurations and coal sources in the United States. The most important data gaps that were
identified included the following:

•  The effect of firing a PRB coal in a more typical configuration
•  The effect of firing a low-sulfur compliance coal
•  The effect of catalyst aging
•  The effect of catalyst type and space velocity

In order to address some of these data gaps, the program was expanded, and additional
testing was conducted in 2002. It should be noted that the highest priority was to test an
SCR-equipped plant that fires a PRB coal. Unfortunately, no plant with this configuration
was available for testing in 2002. However, plans are being made to test two SCR-equipped
PRB plants in 2003 and 2004.

Approach for 2002 Field Test

The principal objective remained the same for the 2002 testing: determine the impact  of SCR
operation on Hg speciation and, ultimately, on Hg emissions. To achieve this objective for
each unit/coal, a sampling plan was developed for various operating conditions so that the
effects of SCR could be determined. At each site, tests were conducted (where feasible)
under operating conditions with and without SCR in operation. This was done either by
bypassing the SCR system or testing sister units, one with and one without SCR.

In addition to the effects of SCR operation, several other factors were identified as contrib-
uting factors to Hg oxidation and removal and were incorporated into the sampling plans  for
2002. These factors included coal type, specifically chlorine and sulfur content,  and catalyst
age. A summary of the configuration of each plant is provided in Table ES-1 for 2001 and
2002 testing. Additionally, a summary of coal data for each plant is provided in  Table ES-2.

Hg measurements were obtained using the manual Ontario Hydro (OH) method as well as Hg
semicontinuous emission monitors (Hg SCEMs). The  sampling plans were set up to obtain
OH samples at the SCR inlet and outlet, ESP inlet and outlet  or, in the case of one plant, a
venturi scrubber and at the stack. The Hg SCEMs were used to measure Hg speciation
primarily at the outlet of the particulate control  device.

                                                                                xxi

-------
    Table ES-1
    Summary of SCR Program Plant Configurations
Plant
S1a
S2
S2-2
S3
S4
S4-2
S5
S6


Year
Tested
2001
2001
2002
2001
2001
2002
2002
2002


Coal
PRB
subbit.
OH bit.
OH bit.
PA bit.
KY bit.
KY bit.
WVbit.
Low-
sulfur
KYand
WVbit.
Boiler Type
Cyclone
Wall-fired
Wall-fired
Tangential-
fired
Cyclone
Cyclone
Wall-fired
Concentric-
fired


Boiler
Size, MW
650
1300
1300
750
650
650
684
700


Low-NOx
Burners
No
Yes
Yes
Yes, with
overfire air
No
No
Yes
Yes


Catalyst Vendor And
Type
Cormetech honeycomb
Siemens/Westinghouse
plate
Siemens/Westinghouse
plate
KWH honeycomb
Cormetech honeycomb
Cormetech honeycomb
Halder-Topsoe plate
Cormetech honeycomb


Catalyst
Age
2 ozone
seasons'3
3 months
2 ozone
seasons
1 ozone
season
1 ozone
season
2 ozone
seasons
3 months
2 ozone
seasons0


SCR Space
Velocity, hr"1
1800
2125
2125
3930
2275
2275
3700
3800


Particulate
Control
ESP
ESP
ESP
ESP
Venturi
scrubber
Venturi
scrubber
ESP
ESP


Sulfur
Control
None
Wet FGD
Wet FGD
None
Venturi
scrubber
Venturi
scrubber
Wet FGD
None


1 Not discussed in detail in this report.
1 The ozone season is from May 1 through September 30.
; One layer of catalyst was replaced after one ozone season.

-------
   Table ES-2
   Summary of Coal Analyses3 for Plants Tested in 2001 and 2002
Parameter
Mercury, |jg/g dry
Chlorides, |jg/g dry
Moisture Content, %
Ash, %
Sulfur, %
Heating Value, Btu/lb
S1
0.10
<60
27.5
3.7
0.19
8960
S2
0.17
1333
7.6
11.7
3.9
11,092
S2-2
0.14
523
6.1
9.4
3.9
12,097
S3
0.40
1248
7.0
14.0
1.7
11,421
S4
0.13
357/1 160b
10.5
9.1
2.9
11,341
S4-2
0.18
270
8.3
9.1
3.0
12,077
S5
0.13
472
4.6
12.1
3.6
12,120
S6
0.07
1020
6.1
11.6
1.0
12,019
a As-received unless otherwise noted.
b First value prior to bypass; second value postbypass.

Description of Sites Tested in 2002

Site S2

Site S2 was tested in 2001 and again in 2002 to collect data after an additional ozone season
(May 1-September 30) of operation on the SCR catalyst. Unfortunately, a number of
operational changes, including addition of SOs mitigation technologies and a change in the
coal (as shown by the chloride values in Table ES-2), between 2001 and 2002 at Site S2 may
have affected the results. In addition, operational problems occurred (plugging of the air
preheater) at Site S2 in 2002 that resulted in a somewhat reduced test plan. The OH and Hg
SCEM data were collected for the  SCR on-line condition, but only Hg SCEM data were
obtained for the SCR off-line condition.

Site S4

Site S4 was tested in 2001 and again in 2002 to collect data after an additional ozone season
of operation on the SCR catalyst. At Site S4, sampling was done with the SCR unit on-line
followed by tests with the SCR unit off-line on the same unit. Based on Table ES-2, there
was significant variability in the coal from one year to the next.

Site S5

Site S5 was selected to provide additional data on the impact of SCR for a facility firing a
high-sulfur eastern bituminous coal and utilizing a wet FGD system for SC>2 control. Hg
sampling at Site S5 was done on two sister units: one with an SCR unit, the other without.

Site S6

Site S6 was selected to represent facilities firing a low-sulfur compliance coal. Hg sampling
at Site S6 was done on two sister units (one with SCR and the other with the SCR unit

                                                                                xxiii

-------
bypassed). As part of another test program that was being conducted simultaneously with the
SCR project, Hg sampling was also done at the stack of a third unit (no SCR).

Mercury Emission/Capture Results for 2002  Field Tests

There were two primary objectives for the Hg testing in 2002. The first was to determine the
Hg oxidation potential of the SCR catalysts at each test site. The second was to determine
what impact SCR had on the Hg removal efficiency  of each pollution control device. The
overall Hg removal (unless otherwise specified) is defined as the Hg measured at the stack
compared to Hg measured at the inlet to the particulate control device.

Site S2

Site S2 was sampled in both 2001 and 2002 to determine the effect of operating SCR over an
additional ozone season on Hg speciation. Units equipped with SCR are required to operate
the SCR unit from May  1 to September 30 (ozone season) for plants such as Site S2 that burn
bituminous coals. To evaluate catalyst aging on Hg speciation, the OH results for 2002 are
compared to those for the 2001 testing. A summary of these results is provided in Figure ES-
1. For this site, results show 54% and 48% Hg2+ at the SCR inlet for 2002 and 2001, respec-
tively. At the SCR outlet, oxidation of Hg across the SCR unit resulted in Hg2+ of 87% and
91% for 2002 and 2001, respectively. Comparing these results shows that the oxidation of Hg
across the SCR did not significantly change from 2001 to 2002. This is also shown by
comparing the ESP inlet sampling results, which was 97% Hg2+ for both years.
              30.0
              25.0-
                                                              fgflC CW20939.CDR
i	1 Particulate-Bound Hg
ESS Oxidized Hg
^H Elemental Hg
                     SCR  SCR  ESP  ESP Stack
                     inlet Outlet  Inlet  Outlet
                     2002 Sampling Results
                           SCR  SCR  ESP  ESP Stack
                           Inlet Outlet  Inlet Outlet
                           2001 Sampling Results
   Figure ES-1
   Mercury Results Comparing Speciation with SCR from 2001 and 2002 at Site S2
XXIV

-------
The overall Hg removal in 2002 across the ESP and wet FGD was 84% compared to 89% in
2001. Operational problems at the plant prevented Hg sampling using the OH method with
the SCR bypassed. Therefore, a comparison of Hg speciation with and without SCR was not
possible using 2002 OH results. However, in 2001, Hg removal was only 51% when the SCR
unit was bypassed.

The Hg SCEMs were operated at Site S2 for approximately 1 month and included the time
the SCR unit was bypassed. Review of the Hg SCEM data illustrates an increase from less
than 0.25 |ig/m3 to approximately 1.0 |ig/m3 Hg° when the SCR unit was bypassed.

Site S4

Site S4 was also tested in both 2001 and 2002. A comparison of the 2001  and 2002 results
are shown in Figure ES-2. For this site, results show 33% and 9% Hg2+ at the SCR inlet for
2002 and 2001, respectively. At the SCR outlet, oxidation of Hg across the SCR unit resulted
in Hg2+ of 63% and 80% for 2002 and 2001, respectively. Although this difference may have
been attributable to a catalyst-aging effect, the coal fired at Site S4 varied, especially with
respect to the chloride content. The measured coal chloride content in 2001 ranged from 350
to 1280 ppm and ranged from 240 to 300 ppm in 2002. Plant personnel indicated that the coal
was from the  same mine for both years. The information collection request coal analysis data
from 1999 for Site S4 also indicated a wide range of chloride concentrations  in the coal.
             20.0
                       Particulate-Bound Hg
                       Oxidized Hg
                       Elemental Hg
             15,0 -
         .2
         "5
         ~  10,0 -
          I
          o
         o
          3
          o
          0>
5.0 -
              0.0
                     SCR
                     Inlet
             SCR    AH
             Outlet  Outlet

             2002 Results
Stack
SCR
Inlet
SCR   AH
Outlet  Outlet
2001 Results
Stack
   Figure ES-2
   Mercury Results Comparing Speciation with SCR from 2001 and 2002 at Site S4
                                                                                XXV

-------
Although there was a substantial decrease in Hg oxidation across the SCR catalyst between
2002 and 2001, downstream of the air preheater and just prior to the inlet of the venturi
scrubber, there was an increase in the percentage of Hg2+. In 2002, 96% of the Hg was
measured as Hg2+ at the outlet of the air preheater compared to 87% in 2001. It is possible
that this difference may be the result of the changing coal composition. The overall Hg
removal efficiency across the venturi scrubber was essentially the same in 2002 and 2001:
93% and 90%, respectively.

Figure ES-3 compares the OH Hg speciation results with SCR in operation and with SCR
bypassed. At the air preheater outlet sampling location, 96% of the Hg is oxidized with SCR
compared to 57% without SCR. In 2001, the comparison was 87% and 56%. As stated above,
the overall Hg removal efficiency across the venturi scrubber was 93%; this is compared to
only 44% when SCR was bypassed. This is  supported by the Hg SCEM data that showed the
average Hg° concentration increasing from 1.1 to 6.4 |ig/m3 when SCR was bypassed.
               20
                Note: Error bars represent standard deviation for total Hg
                                                            EERC DL2252S.COK
           05
           =3.
           03
           c
           CD
           O
           c
           o
           O
           3
           O
                    i   i Particulate-Bound Hg
                        Oxidized Hg
                        Elemental Hg
15 -
                      SCR
                      Inlet
               SCR
              Outlet
 AH
Outlet
Stack
 AH
Outlet
Stack
                        SCR Normal Operation
                                       SCR Bypassed
   Figure ES-3
   Comparisons of Mercury Speciation with the SCR in Service and with the SCR
   Bypassed at Site S4

Site S5

The Hg results for Site S5 are summarized in Figure ES-4. As can be seen in Figure ES-4,
Hg2+ increased from 44% to 81% across the SCR catalyst and was 95% at the ESP inlet
sampling location. For the unit without SCR, the percentage of Hg2+ at the ESP inlet was
80%. The overall Hg removal efficiency across the ESP and wet FGD was 90% for the unit
with SCR compared to 51% for the unit without SCR. It should be noted that the results for
the unit without SCR showed an increase in Hg° (4.7 to 6.1 |ig/Nm3) across the wet FGD.
The increase in Hg° was considerably less (0.7 to 1.0 |ig/Nm3) for the unit with SCR.
xxvi

-------
             Note: Error bars represent standard deviation of total Hg,
           c
           o
           (0
           o
           O
           £•
           ZJ
           o
           OJ
              30-
              25-
              20-
                                                              EERCCW21144.CDR
   Particulate-Bound Hg
   Oxidized Hg
   Elemental Hg
SCR  SCR  ESP   ESP  Stack
Inlet  Outlet  Inlet  Outlet
     Average with SCR
                                                       ESP  ESP  Stack
                                                       Inlet Outlet
                                                     Average Without SCR
   Figure ES-4
   Mercury Speciation Results Compared for a Unit with SCR and Without SCR at Site S5
Site S6
                                                                  2+
The results of flue gas testing from S6 are summarized in Figure ES-5. Hg  increased from
64% to 83% across the SCR catalyst and was 87% at the ESP inlet sampling location. For the
unit with SCR bypassed, the percentage of Hg2+ at the ESP inlet was 69%. However, as
shown in Figure ES-5, there appeared to be more particulate-bound Hg measured when SCR
was bypassed.

The test at Site S6 was conducted to evaluate the impact of SCR on Hg speciation when a
low-sulfur compliance coal was fired; therefore, there was no wet FGD system on either test
unit. Within the variation of the data,  the presence of SCR had no apparent effect on Hg
removal across the ESP (there was little if any for either case). Also, the Hg measured at the
stacks had a high percentage of Hg2+: 92% with SCR and 88% without SCR.
                                                                              XXVll

-------
                Note: Error bars represent standard deviation of total Hg.
                20
                                                               CEflC CW21323.CDR
E
z
            c
            g

            0)
            o
            c
            o
            o
            o
            CD
                15-
                         Particulate-Bound Hg
                         Oxidized Hg
                         Elemental Hg
                      SCR   SCR  ESP  Stack
                      Inlet  Outlet  Inlet
                           Average with
                              SCR
                                      ESP  Stack
                                      Inlet
                                     Average with
                                     SCR Bypass
                                                                  Stack
Average with
  No SCR
    Figure ES-5
    Mercury Speciation Results Comparing Units with and Without SCRs in Service at
    Site S6
Discussion of Overall Results

The primary goal of this project is to evaluate the effect SCR operation has on Hg speciation
and, ultimately, on Hg emissions. The combined results from 2001 and 2002 testing are
discussed below.

Effect of SCR on Mercury Speciation

Table ES-3 presents the results of both the 2001 and 2002 testing. There is an increase in Hg
oxidation across the SCR catalyst for those plants firing an eastern bituminous coal.
However, the amount of oxidation that occurs across the catalyst is highly variable. It appears
to be affected by coal properties as well as catalyst design and, possibly, catalyst age.

Although there is strong evidence that an SCR catalyst does promote Hg oxidation, to
determine the overall effect of SCR, it was useful to conduct tests both with and without SCR
in service at each site. Figure ES-6 shows the comparison. For three of the five sites, there is
a higher concentration of nonelemental Hg (Hg2+ and Hgp) when an SCR unit was present,
based on measurements made at the inlet to the particulate control  device. For the other two
sites, S3 and S6, the percentage of nonelemental Hg was more than 90%, both with and
without an SCR unit in service.
xxvin

-------
   Table ES-3
   Change in Mercury Oxidation across the SCR Catalyst
Site
S1b
S2
S2
S3
S4
S4
S5
S6
Year Sampled
2001
2001
2002
2001
2001
2002
2002
2002
SCR Inlet Hg2+,
% of total Hg
8
48
54
55
9
33
43
60
SCR Outlet Hg2+,
% of total Hg
18
91
87
65
80
63
76
82
Percentage
Increase,
10
43
33
10
71
30
33
22
Point
a%








1 Defined as (SCR Outlet % - SCR Inlet %).
' Site S1 fired a PRB coal; the others were eastern bituminous coals.
                                                                  EERC D12227O.CDR
            100 -

         o
         CD
         .-P  80 H
         0  60 -
         CD

         ?
         15
         CD
         E
             40 -
         _
         CD
         O  20
              0





                          I
                   I     I Inlet to Part. Control Device - No SCR
                   17771 Inlet to Part. Control Device - With SCR



                      S2
S3
S4         S5
     Site
se
   Figure ES-6
   Mercury Concentrations at the Inlet of the Particulate Control Device with and Without
   the SCR, respectively
                                                                                   XXIX

-------
Effect of Catalyst Age on Mercury Speciation

Data indicate that additional Hg oxidation can be expected if an SCR unit is installed on a
unit firing an eastern bituminous coal. A potential concern is "Does the effectiveness of the
Hg oxidation potential of SCR decrease with time?" As has been discussed previously, two
of the facilities, S2 and S4, were tested in both 2001 and 2002 (both burned eastern
bituminous coal). As Figures ES-1 and ES-2 show, there was a decrease in Hg oxidation
across the SCR catalyst in 2002 as compared to 2001. However, the decrease in oxidation
seen over time is less than that seen from coal variability. Additionally, it is expected that
routine replacement of catalyst layers will minimize the effect. Also, mitigating circum-
stances at each plant prevent a definitive conclusion from being developed. At Site S2, the
temperature of the SCR unit was -10 °F cooler due to humidification, and alkali was added
upstream of the SCR unit. At both Sites S4 and S2, the coal chloride concentration was
highly variable. Although there may be an "effect" of aging as measured across the SCR unit,
Hg measurement at the inlet to the particulate control device indicates there was no signifi-
cant difference at either site. To understand if these results are indicative of a catalyst aging
effect, Hg speciation sampling is recommended at these plants for several additional years.

Effect of the SCR on Wet FGD Performance for Mercury Control

The underlying intent of understanding SCR-mediated Hg oxidation is to determine its
potential to improve the Hg collection efficiency of existing ESPs, fabric filters and, in
particular, FGD systems. In general, wet FGDs remove a large percentage (more than 90%)
of Hg2+. However, there has been evidence that some of the captured Hg2+ can be reduced in
the wet FGD to Hg° [9]. Although the sample set is very small (three facilities) and the wet
FGDs tested to date are not representative of the most common FGD design in the United
States (forced oxidation system), the data from this project indicate that some of the Hg2+ is
chemically reduced to Hg° in the wet FGD. This Hg° passes through the FGD and is
therefore not captured, resulting in an increase of Hg° across the FGD. For the purposes of
this report, this effect is termed reemission. As can be seen in Table ES-4, at all the sampling
sites, there is an increase in Hg° across the FGD. Also, the data seem to indicate the operation
of the SCR unit ameliorates possible reemission.
XXX

-------
   Table ES-4
   Effect of the SCR on Hg° Concentration Across the Wet FGDs
   Site
Year Sampled
                              FGD Inlet Hg°   FGD Outlet Hg°
Cone.,
ug/Nm3
Cone.,
ug/Nm3
Hg° Increase,3
   |jg/Nm3
Total Hg
Removal,
                                        With SCR
S2

S2

S4

S4

S5
    2001

    2002

    2001

    2002

    2002
  0.4°

  0.3

  0.5

  1.0

  0.7
  0.9

  1.3

  0.8

  1.3

  1.0
     0.5

     1.0

     0.3

     0.3

     0.3
   89

   84

   90

   91

   91
                                       Without SCR
S2
S4
S4
S5
2001
2001
2002
2002
3.4^
5.6
5.7
4.7
5.0
7.1
8.0
6.1
1.6
1.5
2.3
1.4
51
46
44
51
1 Defined as (FGD outlet Hg° cone. - FGD inlet Hg° cone.).
' For 2001 Site S2 data, the ESP inlet data were used because the FGD inlet Hg concentration values appear to be clear outliers.
Summary

The primary conclusions based on the test results are:

•  For plants firing eastern bituminous coals, Hg° can be oxidized across the SCR catalysts.
   The effect that SCR has on Hg speciation (i.e., extent of additional oxidation that occurs)
   may be dependent upon the coal characteristics and catalyst properties. The percentage
   increase of Hg2+ at the SCR outlet ranged from 10% at Site S3 to 71% at Site S4.

•  At both sites where sampling was done in 2001 and 2002, there appeared to be a decrease
   in Hg oxidation across the SCR catalyst with time. However, at both facilities, the
   decrease was minimal, and other possible explanations related to changes in the plant's
   operation might explain the decrease. These changes do not allow a definitive conclusion
   to be reached concerning the effect of catalyst age (an additional ozone season) on
   SCR/Hg oxidation. It is important to note that the measured Hg oxidation at the inlet to
   the particulate control device was the same (within the variability of the data) for both
   years.
                                                                                  XXXI

-------
•  Based on the limited data at three plants (five total data sets), SCR operation may reduce
   the extent of reemission across the wet FGDs. For the tests with SCR in service, the
   increase appears to be very small and is generally within the variability of the data.
   Nevertheless, five data points show an increase in Hg°. When SCR is not in service, it
   appears that the reemission is more pronounced.

Future Test Plans

Based on a review of these test results, several areas will require further investigation. DOE,
EPA, and EPRI are planning to conduct additional full-scale, as well as bench- and pilot-
scale, studies to address the following:

•  The effect of SCR for a PRB pulverized coal application.
•  The effect of FGDs on Hg capture, in particular Hg reemission.

•  The effect of SCR when PRB-bituminous-blended coal is fired.

•  The effect of catalyst age on Hg speciation.

References

1.  U. S. Environmental Protection Agency. Mercury Study Report to Congress Volume I:
   Executive Summary, Office of Air Quality Planning  and Standards and Office of
   Research and Development, Dec 1997.

2.  EPRI. An Assessment of Mercury Emissions from U.S. Coal-Fired Power Plants; EPRI
   Report No. 1000608,  Oct 2000.

3.  U. S. Environmental Protection Agency. Study of Hazardous Air Pollutant Emissions
   from Electric Utility Steam Generating Units Final Report to Congress: Executive
   Summary; Office of Air Quality Planning and Standards and Office of Research and
   Development, Feb  1998.

4.  ICR Reports, http://www.epa.gov/ttn/uatw/combust/utiltox/utoxpg.html (accessed Oct 7,
   2000).

5.  Cichanowizc, I.E.;  Muzio, LJ. Factors Affecting Selection of a Catalyst Management
   Strategy. In Proceedings of the  Combined Power Plant Air Pollutant Control Mega
   Symposium; Washington, DC, May 2003.

6.  Gutberlet, H.; Schliiten, A.; Lienta, A. SCR Impacts on Mercury Emissions on Coal-Fired
   Boilers. Presented at EPRI SCR Workshop, Memphis, TN, April 2000.

7.  Pilot-Scale Screening Evaluation of the Impact of Selective Catalytic Reduction for NOX
   on Mercury Speciation; EPRI, Palo Alto, CA; the U.S. Department of Energy National
   Energy Technology Laboratory, Pittsburgh, PA; and the U.S. Environmental Protection
   Agency, National Risk Management Research Laboratory, Air Pollution Prevention and
   Control Division, Research Triangle Park, NC,  2000; 1000755.
XXXll

-------
8.  Selective Catalytic Reduction Mercury Field Sampling Project; EPRI, Palo Alto, CA; the
   U.S. Department of Energy National Energy Technology Laboratory, Pittsburgh, PA; and
   the U.S. Environmental Protection Agency, National Risk Management Research
   Laboratory, Air Pollution Prevention and Control Division, Research Triangle Park, NC,
   2002; 1005400.

9.  Nolan, P.; Redinger, K.; Amrhein, G.; Kudlac, G. Mercury Emissions Control in Wet
   FGD Systems. Presented at Air Quality III, Washington, DC, Sept 2002.
                                                                              xxxin

-------
 1
INTRODUCTION
Coal combustion by electric utilities is a large source of anthropogenic mercury (Hg)
emissions in the United States, according to the most recent data, accounting for 45 tons/yr of
total point source Hg emissions [1]. In December 2000, the U.S. Environmental Protection
Agency (EPA) issued an intent to regulate Hg from coal-fired utility boilers [2]. As a result,
many utilities have become proactive in evaluating the effectiveness of current air pollution
control technologies, as well as new technologies for Hg control [1,3-5].

Hg emissions from coal-fired boilers can be empirically classified, based on the capabilities
of currently available analytical methods, into three main forms: elemental mercury (Hg°),
oxidized mercury (Hg2+), and particulate-bound Hg. Paniculate-associated Hg (Hgp) can be
removed from flue gas by conventional air pollution control devices such as an electrostatic
precipitator (ESP) or a baghouse. Hg2+ compounds are readily captured in flue gas
desulfurization (FGD) units. Hg° is most likely to escape air pollution control devices and be
emitted to the atmosphere. Total Hg concentrations in coal combustion flue gas typically
range from 3 to 15 |ig/Nm3; however, Hg°, Hg2+, and particulate-bound Hg concentrations
are quite variable depending on coal composition and combustion conditions [6].

In addition to Hg, coal-burning power plants are a significant anthropogenic source of
nitrogen oxides (NOx) emissions to the atmosphere. NOx  emissions are an environmental
concern primarily because they are associated with increased acidic precipitation, as well as
fine-particle and ozone formation. Depending on the size and type of boiler, the 1990 Clean
Air Act Amendments require specific reductions in NOx emissions from coal-fired electric
utilities. The most common NOx reduction strategy is the  installation of low-NOx burners.
These burners have the capability of reducing NOx emissions by 40%-60%. However, with
possible establishment of fine particulate (PM^.s), regional haze, ozone regulations, and NOx
state implementation plans, there is increased incentive to  reduce NOX emissions to a level
below what can be achieved using low-NOx burners. Selective catalytic reduction (SCR)
technology, which can reduce NOx emissions by more than 90%, is, therefore, becoming
more attractive, particularly because catalyst costs continue to decrease and the knowledge
base for using SCR reactors is expanding.  It is planned that approximately 100 gigawatts of
coal-fired electrical capacity will have SCRs installed by 2005 [7].


1.1    Potential Impacts of SCR on Mercury Speciation

SCR units achieve lower NOx emissions by catalytically reducing NOx to N2 and H2O in the
presence of ammonia (NH?). The catalysts used in SCR units are generally metal oxides such
as titanium dioxide (TiO2)-supported vanadium oxide (X^Os). These units are generally
                                                                                1-1

-------
Introduction

operated at about 650-750 °F (343-399 °C). Initial laboratory-scale testing indicated that
metal oxides, including X^Os and TiO2, promoted the conversion of Hg° to Hg2+ or
parti culate-bound Hg in relatively simple flue gas mixtures [8]. In addition, pilot- and full-
scale Hg speciation measurements in European and U.S. coal-fired boilers equipped with
SCR reactors have shown the potential to promote the formation of Hg2+ [9-11]. Therefore, it
was hypothesized that the use of SCR may improve the Hg-control efficiency of existing air
pollution control devices by promoting Hg2+ or particulate-bound Hg formation.

Possible mechanisms by which SCR operation could affect Hg speciation include:

•  Catalytic oxidation of the Hg. Evidence indicates that vanadium -based catalysts can
   promote the formation of Hg2+ [9-12].

•  Changing the flue gas chemistry. The significant reduction in flue gas NOX and slight
   increase in NH? concentrations associated with SCR may affect Hg speciation. It is well
   known thatNOx, particularly NC>2, has a substantial effect on Hg speciation [13]. The
   gas-phase effects of NH3 on Hg are unknown. SCR units also have the potential to
   catalyze the formation of sulfur trioxide (863) and,  potentially, chlorine (Cb), which may
   then react with Hg [14-18].

•  The SCR unit provides additional residence time for the oxidation of Hg to take place.

•  Changing the fly ash chemical composition. It is possible that SCR operation may change
   the surface chemistry of the fly ash particles such that their ability to adsorb or convert
   Hg species is altered.

•  Increasing wall deposition. SCR systems may result in the deposition of ammonium
   bisulfate and ammonium sulfate in the air preheater and duct walls. It is unknown
   whether increased deposition could impact Hg emissions or speciation.


1 .2    Pilot-Scale Screening Tests Conducted at the EERC

To investigate the effects of SCR on Hg speciation in a coal combustion system, EPRI, the
U.S. Department of Energy (DOE), and EPA funded a pilot-scale project at the Energy &
Environmental Research Center (EERC) [10]. The primary objective for the pilot-scale tests
was to determine whether NH? injection or the catalyst  in a representative SCR system
promote the conversion of Hg° to Hg2+ or Hgp. Although this project was a screening
evaluation and not a complete parametric study, it was designed to evaluate potential
mechanisms for Hg conversion and the various coal parameters (like chemical composition)
that may affect the degree of conversion.

Three bituminous coals and a Powder River Basin (PRB) subbituminous coal were burned in
a pilot-scale  combustion system equipped with an NH?  injection system, SCR reactor, and
ESP. The selection criteria for the four coals investigated were the  significant differences in
their sulfur and chloride contents.
The results from the tests indicated that NHs injection and, possibly, the SCR catalyst
promote the conversion of Hg2+ to Hgp in the coal combustion flue gases for two of the
1-2

-------
                                                                           Introduction

bituminous coals, but this was not the case for the PRB coal. The results were inconclusive
for the third bituminous coal. When the limited data are used in a linear regression analysis, it
appears that the chloride, sulfur, and calcium contents of the coal correlate with Hg
speciation across the SCR unit. Because of the inherent concerns related to small pilot-scale
tests (surface area-to-volume ratios, different flue gas chemistries, and time and temperature
profiles), it was decided that sampling at full-scale power plants was necessary. Therefore,
beginning in 2001, EPRI, DOE, and EPA funded projects with the EERC and others to
conduct Hg sampling at power plants with SCR technology.


1.3    2001 SCR Mercury Field-Sampling  Project

The test program for 2001 was developed to address the limitations of pilot-scale testing by
applying information obtained  from previous work to several full-scale electric-generating
facilities. A summary of plants and their configuration is provided in Table 1-1. The overall
objective of 2001  testing was to evaluate the effects of SCR operation, selective noncatalytic
reduction (SNCR), and flue gas conditioning on speciated Hg emissions at full-scale plants.
More specifically, the objective was to evaluate Hg speciation across the unit as a result of
these technologies. The results of testing conducted for the 2001 program are summarized
below and can be found in "Power Plant Evaluation of the Effect of Selective Catalytic
Reduction in Mercury" [11].

In general, data from 2001 testing indicated that SCR has the potential to increase Hg
oxidation. However, significant differences in Hg speciation were observed between plants
even with similar coal classifications. The possible reasons for these disparate differences
likely include a combination of the following:

•  Coal chloride  concentration - The chloride level in the coal is the most straightforward
   approach to estimating Cl (HC1 and Cb) in flue gas, although it is possible that alkalinity
   in the fly ash may tie up Cl and reduce its availability for some coals.

•  Inlet percentage of Hg2+ - For plants with a high proportion of the inlet Hg already
   oxidized, the potential increase is much lower.

•  Other flue gas constituents  (e.g., alkalinity, SO2, and SO3).

•  SCR system/catalyst properties - e.g.,  space velocity, area velocity, catalyst type, catalyst
   age, or number of catalyst layers.

It was thus theorized that the Hg speciation and associated oxidation of Hg across the SCR is
highly dependent upon coal characteristics.

The primary conclusions from this effort were:

•  At all four sites tested with SCR, an increase in Hg oxidation was observed across the
   SCR unit. It varied from 10% at Sites SI and S3 to 71% at Site S4. SCR units can assist
   in converting Hg° to Hg2+; however, the effect appears to be coal-specific and, possibly,
   catalyst-specific.
                                                                                  1-3

-------
   Table 1-1
   Summary of Plant Configuration from 2001 Test Program
Site
S1
S2

S3
S4
Coal Boiler Type
PRB subbit. Cyclone
OH bit. Wall-fired

PA bit. Tangential-
fired
KY bit. Cyclone
Boiler Size,
MW
650
1300

750
650
Low-NOx
Burners
No
Yes

Yes, with
overfire air
No
Catalyst Vendor
and Type
Cormetech
honeycomb
Siemens/
Westinghouse
plate
KWH honeycomb
Cormetech
honeycomb
Catalyst
Age
2 ozone
seasons
3 months

1 ozone
season
1 ozone
season
SCR Space
Velocity,
1800
2125

3930
2275
Particulate
ESP
ESP

ESP
Venturi
scrubber
Sulfur
None
Wet FGDa

None
Venturi
scrubber
' Flue gas desulfurization.
                                                                                                                                    3


-------
                                                                          Introduction

•  For the two sites with SCR and wet FGD system (S2 and S4), a high percentage of the
   total Hg was removed, 89% and 90%, respectively. It should be noted that at Site S3 the
   percentage of Hg2+ at the outlet of ESP was 83%.

•  Site SI contained significant particulate-bound Hg, which was removed across the ESP,
   resulting in 85% total Hg removal. The high level of particulate-bound Hg may have
   been a result of the high carbon content of the ash (15% to 17%).

•  Based on limited data (one site each), SNCR for NOx control and NH? flue gas
   conditioning for  improving ESP performance appeared to have a fairly small effect on Hg
   oxidation.

Upon review of 2001 test results, it was evident that additional data would be necessary to
quantify the  effect SCR operation had on Hg oxidation, including the following:

•  Determine the effect of firing a PRB in a more typical configuration

•  Determine the effect of firing a low-sulfur compliance coal

•  Determine the effect of catalyst aging

•  Determine the effect of catalyst type and space velocity

In order to address these issues, the program was expanded, and additional testing was
conducted in 2002. It should be noted that the highest priority given was to test a plant with
SCR and a pulverized coal (pc)-fired PRB coal. Unfortunately, no plant could be identified
for testing in 2002 with this configuration. However, plans are being made to test two PRB
plants with SCR units in 2003 and 2004.


1.4    Project Goals and Objectives

The project goal is to determine the impact of SCR operation on Hg speciation and on Hg
emissions. The specific objectives of the 2002 testing were to:

•  Determine the change in Hg speciation across the SCR catalyst as a function of catalyst
   aging (an additional ozone season). Two plants that had been tested in 2001 were retested
   in 2002.

•  Determine the effect of firing a compliance (low-sulfur) coal on  Hg speciation  across the
   SCR catalyst.

•  Determine what  effects SCR has on subsequent Hg speciation and the overall Hg removal
   for the paniculate control device and, if present, the wet FGD.
                                                                                1-5

-------
Introduction

1.5    Sampling Approach


1.5.1  Mercury Sampling Using the Ontario Hydro Mercury Speciation Method

At each facility, the overall sampling approach consisted of measuring Hg across each
pollution control device (SCR, ESP, and wet FGD). In this way, the effect of these devices
on Hg could be determined. To determine the overall effect of SCR on Hg speciation and
subsequent removal, sampling was done at a unit with SCR followed by testing either at a
similar unit without SCR or with SCR bypassed. For example, if a plant had an SCR unit,
ESP and wet FGD unit samples were taken at five locations as follows:

•   SCR inlet
•   SCR outlet (prior to the air heater)

•  ESP inlet (downstream of the air heater)

•  ESP outlet

•  Wet FGD  outlet (generally the stack)

In general, samples were taken in pairs across  each device (i.e., inlet and outlet of the SCR)
and in duplicate or triplicate.

The Ontario Hydro (OH) sampling was done using EPA Method 17, ensuring that the filter
was at the same  temperature as the flue gas. At the SCR inlet and outlet condition, the OH
sample filter averaged between 600 and 750 °F (316 and 399 °C). Following the air heater,
the flue gas temperature was between 250 and 350 °F (121 and 177 °C). Sampling was done
at a single point rather than traversing the flue gas duct. For wet stack locations where the
flue gas temperature was below 250 °F (121 °C), an external heater (Method 5 configuration)
was used to maintain the filter temperature above 250 °F (121 °C).

Based on the OH data, for each plant the following were calculated:

•  The change in Hg oxidation across the SCR unit. This is defined as the difference in the
   percent Hg2+ in the flue gas between the outlet and inlet of the SCR unit.

•  Overall effect of SCR on Hg oxidation. This is defined as the difference in the percentage
   of Hg2+ at  the inlet to the particulate control device with and without SCR.

•  Overall Hg removal. This is defined as percentage based on the difference in total Hg
   measured at  the inlet to the particulate control device and the stack.

1.5.2  Mercury Sampling Using Hg SCEMs

Hg semicontinuous emission monitor (Hg SCEM) testing was done using the PS Analytical
(PSA) or Tekran system with a stannous chloride (wet-chemistry) conversion system.
Attempts were made to operate the Hg SCEMs at inlet and outlet locations; however, it was
extremely difficult to maintain Hg SCEM operation on a continuous basis at locations


1-6

-------
                                                                          Introduction

upstream of particulate removal devices. Therefore, data from Hg SCEMs were collected
primarily from ESP outlet or stack locations. Where applicable, OH data are compared to Hg
SCEM data; however, in general, OH results with appropriate quality control (QC) provide a
more defensible EPA-approved method of quantifying Hg concentration and emissions. The
benefit or advantage of Hg SCEM operation is the real-time nature of the data. When
operated continuously for several days, SCEMs provide valuable information on how Hg
concentration and speciation changes with typical plant operation.


1.5.3  Other Flue Gas Analyses

Other gases sampled for this project include SOs and NH3 slip. SOs was measured using the
controlled condensation method, and NH? slip was measured using Conditional Test Method
27. For most of the sites, SOs was measured at the air preheater inlet (SCR outlet) and at the
ESP inlet locations. The NH3 slip was measured at the air preheater inlet.


1.5.4  Mass Balance

Coal, hopper ash, and FGD samples, where appropriate, were collected at each site and
analyzed for Hg. These results, along with flue gas data, were used to quantitatively evaluate
the fate of Hg throughout the unit.


1.5.5  Plant Operation Data

Each plant provided operational information, such as plant load and flue gas CEM data, for
the purposes of evaluating unit performance and flue gas  chemistry. Differences existed
between plants regarding the type, frequency, and form in which operational data existed.
Therefore, the figures presenting these data are unique to  each plant.  Nevertheless, the
information is useful for comparing Hg SCEM data with plant operational data and
evaluating possible impacts on Hg speciation.
                                                                                 1-7

-------
2
SITE S2
Site S2 was selected to provide data to determine the impact of catalyst aging on the potential
of the SCRs to oxidize Hg.


2.1    Site Description and Configuration

Site S2 fires a high-sulfur Ohio bituminous coal and employs SCR followed by an ESP and a
wet FGD. The SCR unit at Site S2 utilizes a Siemens/Westinghouse plate catalyst with a
space velocity of 2125 hr"1. Although Site S2 had previously been tested in 2001, there were
several operational differences in 2002. To control SOs emission,  alkali was injected just
downstream of the boiler in 2002. Also, a humidification system had been installed upstream
of the SCR unit to lower the SCR temperature -10 °F. Finally, the coal may not have been
from the same mine as that fired in 2001 because the Cl content in the coal was much less,
520 vs. 1330 ppm. The original intent at Site S2 was to operate 2 weeks with SCR in service
and then, following the ozone season, bypass the SCR unit and test without SCR. However,
as will be discussed later, there were operational problems at Site  S2 that changed this test
plan. General information about the configuration of the unit tested at Site 2 are:

•  Fuel type: Ohio bituminous coal

•  Boiler capacity: 1360 megawatts (MW)

•  Boiler type: wall-fired

•  NOx control: low-NOx burner and SCR
•  Particulate control: ESP

•  SO2 control: magnesium-enhanced lime FGD

A schematic of Site S2 including sampling locations is shown in Figure 2-1.
                                                                              2-1

-------
Site S2
                                                                    EERG JT19295.CDR
INI 13 II IJCUIUJI 1 	
Boiler





	



f




/
( SCR \ SCR


' V

/
«F
SCR \ N
Inlet >| AH


i

Outlet

f
^5
^7
ESP
                                             ESP
                                   Inlet
                                                      rfi
                                   FGD
                              ESP
                              Outlet
                                                                        Stack
                                   ESP Inlet
                               Sample Location    cop Outlet
                                                Sample Location
        SCR Inlet         SCR Qut|et
    Sample Location   Samp|e Locatjon
                                        Stack Sample
                                          Location
    Boiler
SCR
ESP
                                                          FGD
                                                                  Stack
   Figure 2-1
   Schematic of Site S2 Showing Sample Locations from Horizontal and Vertical
   Perspectives
2.2   Sampling Approach

The sampling approach at S2 was similar to the previous year and is documented in Power
Plant Evaluation of the Effect of Selective Catalytic Reduction in Mercury [11]. The
2-2

-------
                                                                              Site S2

objective of testing this unit again in 2002 was to evaluate the effect of catalyst aging on
speciated Hg emissions. As stated earlier, changes to the system, including using
humidification to operate SCR at a lower temperature, make comparison and interpretation of
2001 and 2002 data difficult. Also, because of operational problems, the SCR unit was
bypassed earlier than expected, and the complete test program was not able to be completed.
However, Hg SCEM data were collected for approximately 1 month: 2 weeks each with and
without SCR on-line.


2.2.1  Flue Gas Sample Streams

Flue gas Hg speciation was measured at five locations using the OH method. Sample
locations included the ESP inlet and outlet, SCR inlet and outlet, and stack. A test matrix is
provided in Table 2-1. Where practical, OH measurements were conducted simultaneously
across the ESP or FGD in an effort to quantify the effect each had on Hg concentration and
speciation. In addition to Hg, flue gas samples were also taken to measure the total
particulate loading, SOs concentrations, and NH? slip.  The sampling methods used for all the
sites are described in Appendix A.
   Table 2-1
   Sampling Test Matrix for Site S2a
Date
Begin End
07/16/02 07/20/02
SCR In SCR Out ESP In
OH OH OH
332
Stack
OH
2
SCR Out
S03
3
SCR Out ESP Out
NH3 NH3
2 2
a All samples were done with the SCR in service.

Longer-term Hg monitoring was planned using two Hg SCEMs (Tekran): one located at the
ESP inlet and the other at the ESP outlet. However, because of severe plugging of the inertial
filter probe early in the test, both instruments were operated at the ESP outlet. One
instrument measured total Hg and the other Hg°.


2.2.2  Other Sample Streams

Samples of coal, fly ash, and FGD materials were collected in an effort to obtain
representative operational data related to Hg speciation. These samples were analyzed for
total Hg and, along with the flue gas emission data, were used to qualitatively evaluate the
fate of Hg throughout the unit.

Coal samples and ESP hopper ash samples were collected each day of the test. The coal
samples comprised coal from each coal mill that were pulled from the bunker and analyzed
by the plant every 12 hours. The EERC received a split of these samples. The ESP hopper
ash samples were taken as a field composite from the hoppers associated with the first field
of both the upper and lower ESP. The FGD slurry samples were collected from the
blowdown tank of the FGD.
                                                                                2-3

-------
Site S2

2.3    Process Operating Conditions

Plant operational data are presented in Figure 2-2 for Site S2. The most significant change in
plant operations is the bypassing of SCR. The amount of NOx coming out the stack increased
significantly (from approximately 50 to 400 ppm) 390 hours into the test period after NH3
injection was turned off and the SCR unit was bypassed. An increase in Hg° was observed
with the Hg SCEM and will be discussed later in this report. The average 862 collection
efficiency of the FGD during the 4-week testing period was 95%, and the NOx removal
efficiency of the SCR unit when in service was 90%.

Average auxiliary flue gas data including moisture, dust loading, and percent carbon dioxide
(CO2) and oxygen (O2) were collected during the OH sampling from each sample location
(Table 2-2). Values are within expected ranges (complete auxiliary flue gas data are provided
in Appendix C, Table C-l).
2-4

-------
                                                                            Site S2
          7/15/02
                7/22/02
7/25/02
                                                8/5/02
EERCCW21407.CDFI
 8/12/02
1400 -
1200 -
1000 -
 800 -
u
Uy
         600
          13
          12
          11
          10
           9
     uANNi4^M^^
                                                                        - 1400
                                                                        - 1200
                                                                        -1000
                                                                         800
                                   600
                                    13
                                    12
                                    11
                                    10
                                     9
                                  Total Gas-Phase Hg
                          168
                  Air Preh eater
                  Plugging

Figure 2-2
Plant Operation Data for Site S2
                                             504
                                            SCR Bypassed
                                                                       (continued)
                                                                              2-5

-------
Site S2
                                      7/22/02
7/25/02
     8/5/02
EERCCW2UOB.CDR

    8/12/02
                                        168
                                Air Preheater
                                Plugging —-
                                                                                          *-   0
Time, hr
      504

' SCR Bypassed
                                 672
    Figure 2-2 (concluded)
    Plant Operation Data for Site S2
2-6

-------
                                                                                 Site S2
   Table 2-2
   Average Auxiliary Flue Gas Data for Site S2a
Measurement
Location
SCR Inlet
SCR Outlet
ESP Inlet
ESP Outlet
Stack
Flue Gas Moisture,
%
9.8
10.8
11.3
11.0
21.9
Dust Loading,3
gr/dscf
2.7803
3.3547
1.8872
0.0021
0.0016
C02,
%
15.0
14.8
13.9
13.7
13.2
02,
%
3.8
4.6
5.7
5.8
6.5
a Dust loadings were collected as part of the OH method using EPA Method 17 and, therefore, are not for compliance purposes.


2.4    Sampling Results


2.4.1  Ontario Hydro Flue Gas Mercury Results

Average Hg results for flue gas  sampling at Site S2 are presented in Table 2-3. The complete
results are present in Appendix B (Table B-l). As shown in Table 2-3, there was an increase
from 54% to 87% in Hg2+ across the SCR catalyst. This then increased to 97% at the ESP
inlet. Total Hg removal was 84%.
   Table 2-3
   Ontario Hydro Average and Percentage of Total Mercury Results for S2a
Sample Location
SCR Inlet
SCR Outlet
ESP Inlet
ESP Outlet
Stack
Average,
ug/Nm3
Hgp
0.04
0.06
0.03
0.00
0.00
Hg2+ Hg°
6.5 5.5
10.8 1.6
12.2 0.3
11.1 0.3
0.7 1.3
Hg-Total
12.0
12.4
12.6
11.5
2.0
Percent of Total,
%
Hgp
0.4
0.5
0.2
0.0
0.2
Hg2+
54
87
97
97
35
Hg°
46
13
3
3
65
Total Mercury Removal = 84%
1 Hg values are dry and corrected to 3% O2.
3 Total Hg removal is defined as: [(ESP Inlet - Stack)/ESP Inlet] x 100%.
                                                                                   2-7

-------
Site S2
A comparison of the 2001 and 2002 results at Site S2 is shown in Figure 2-3. As shown, in
2001 there was an increase from 48% to 91% Hg2+ across the SCR catalyst, which is a larger
change than seen in 2002. It is unknown whether this decrease in Hg oxidation across the

              Note: Error bars represent standard deviation of total Hg.
              30.0
           E
          Z
          "3)
           c
          .0.
          '*»—*
           TO
           0
           O
           C
           O
          O
           13
           O
25.0 -
20.0 -
              15.0 -
Particulate-Bound Hg
Oxidized Hg
Elemental Hg
              10.0-
                      SCR SCR
                      Inlet Outlet
                 ESP
                 Inlet
          ESP Stack
         Outlet
                     2002 Sampling Results
SCR  SCR ESP  ESP Stack
Inlet Outlet Inlet Outlet
2001 Sampling  Results
   Figure 2-3
   Comparison of Mercury Speciation Results 2001 and 2002 for Site S2

SCR catalyst is a result of operating the SCR unit over an additional ozone season or due to
the modifications that were made for SOs mitigation or changes in the coal in 2002.

Although there was a decrease in Hg oxidation across the catalyst, there was no change from
2001 to 2002 in the percentage of Hg2+ at the ESP inlet location (97% for both years). There
was a small decrease in total Hg removal: 89% in 2001 compared to 84% in 2002.

As stated earlier, several operational problems occurred at the plant during the test.
Approximately midway through the test period, an increase in pressure drop was measured
across the center air heater. To prevent plugging of the air heater, the  economizer outlet
temperature was raised, and flow through the center air heater was greatly reduced. This
eventually required the bypassing of the SCR unit earlier than expected and the load reduced.
Therefore, no OH data were obtained in 2002 with SCR bypassed. However, in 2001, the
total Hg removal efficiency was only 51% with SCR bypassed.

Although the Hg2+ concentration at the inlet to the wet FGD (outlet of the ESP) was 97%,
only 84% of the Hg was captured by the wet FGD. This appeared to result in an increase in
Hg° at the stack, from 0.3 to 1.3  |ig/Nm3.
2-8

-------
                                                                              Site S2

2.4.2  Hg SCEM Results

Two Hg SCEMs were operated at the ESP outlet to gather longer-term variability data. In the
original test plan, one Hg SCEM was to be operated at the ESP inlet and one at the ESP
outlet. However, there was severe plugging of the inertial filter probe because of flue gas
chemistry and particulate loading, so it was not possible to maintain Hg SCEM operation at
the ESP inlet location. Therefore, the two instruments were operated at the ESP outlet
location and configured to operate such that one instrument measured total Hg and the
second measured Hg° continuously. With the exception of occasional maintenance routines
and troubleshooting, data for both total Hg and Hg° were collected during the entire sampling
period. A summary of the Hg SCEM data is provided in Figure 2-4.  As shown in Figure 2-4,
there is good agreement between the Hg  SCEM data and the OH data.

It should  also be noted that a high amount of variability was observed, especially during the
first several days for total Hg concentration. This may be due to continuously switching
between measurement of Hg° and total gas-phase Hg during that period as only one Hg
SCEM instrument was monitoring the ESP outlet location for the first 198 hours of the test.
The variability of the data reduced noticeably after this period when both instruments were
used to monitor each Hg species.

The  SCR unit was bypassed midway through the test. Table 2-4 shows the statistical
variation  of the Hg SCEM data with and without SCR in service. Based on the Hg SCEM,
there was an increase in Hg° when SCR was bypassed. The concentration of Hg° appears to
have increased from less than 0.25 |ig/m3 to approximately 1 |ig/m3  after the SCR unit was
bypassed. In Figure 2-5, the percentage of Hg2+ as determined using the Hg  SCEM (total Hg
-Hg°) is plotted. When the SCR unit is bypassed, the percentage of Hg2+ in the flue gas is
reduced and appears to become much more variable.
   Table 2-4
   Statistical Variation of Mercury with and Without the SCR in Service Based on Hg
   SCEM Data for Site S2
Mercury
Hg(total)
Hg°
Hg(total)
Hg°
Operation
With SCR
With SCR
SCR bypassed
SCR bypassed
Average,
ug/m
4.7
0.1
6.1
0.8
Std. Dev.,
ug/m3
3.4
0.1
2.0
0.5
Upper 90%
Cla
ug/m3
10.3
0.3
9.4
1.6
Lower 90%
C''3
ug/rn3
0.0
0.0
2.8
0.0
' Cl = Confidence interval.
                                                                                2-9

-------
to
                                 7122102
7/29/02
8/05/02
   8/12/02
EERC DL21842.CDR
                                                                                         •   Hg SCEM Total Gas-Phase Hg
                                                                                         O   Hg SCEM Elemental Hg
                                                                                            I OH Total Gas-Phase Hg
                                                                      SCR Bypassed

                                                                                                                                    20
                                                                   - 18


                                                                   - 16


                                                                   - 14


                                                                   - 12


                                                                   - 10


                                                                   - 8


                                                                   - 6


                                                                   - 4
              48   72   96 120  144  168 192 216 240  264 288 312 336  360  384 408 432 456  480 504 528 552  576  600 624 648 672  696

                                                                 Hours into Test
         Figure 2-4
         Hg SCEM Results for Site S2

-------
                                                                             Site S2
                                                            EERCCW22137.CDR
                    7/22
7/29         8/05
     Calendar Time
8/12
8/19
   Figure 2-5
   Average Hg2+ as Measured by Hg SCEMs (total Hg - Hg°) for Site S2
2.4.3  Coal Analysis Results

In an attempt to understand the Hg variability observed at Site S2, all 45 coal samples from
Site S2 were analyzed by WE Energies and the EERC for Hg and chloride content. The
complete data set is presented in Appendix B, Table B-2. The average Hg concentration and
chloride content of the coal were relatively constant during the 4-week test period. The Hg
concentration was 0.12 ± 0.02 |ig/g, and the chloride content was 636 ± 44 ppm. Both values
are similar to those collected during the previous year's sampling. Table 2-5 presents the
proximate and ultimate analyses for the period during which the OH sampling was done.


2.4.4  ESP Ash and FGD Mercury Results

ESP hopper ash samples were collected daily as a field composite of ash from hoppers
associated with the first field of both the upper and lower ESP. These samples were analyzed
for Hg and loss on ignition (LOT). The FGD samples were filtered, and the liquid and solid
fractions both analyzed for Hg. A summary of the analysis of the ESP hopper ash and FGD
samples is provided in Table 2-6.

The average Hg concentration in the ESP hopper ash was low, 0.05 |ig/g. This is consistent
with the particulate-bound Hg measured in the flue gas at the ESP inlet. The LOT was also
very low, 0.8%. One factor which may impact the ability of a fly ash to adsorb Hg is the
carbon content of the ash. Although not directly measured, the very low LOT indicates low
                                                                              2-11

-------
Site S2
    Table 2-5
    Coal Analysis for Site S2a
Parameter
Mercury
Chlorine
Units
ppm (dry)
ppm (dry)
7/17/02 7/18/02 7/19/02 7/20/02 Avg.
0.13 0.11 0.11 0.13 0.12
724 605 593 609 632
                         Proximate Analysis
Moisture
Volatile Matter
Fixed Carbon
Ash
Wt%
wt%
wt%
wt%
5.6
39.0
46.5
8.9
5.4
39.7
45.0
9.9
6.7
39.8
44.1
9.5
6.8
40.2
43.8
9.2
6.1
39.7
44.8
9.4
Ultimate Analysis
Hydrogen
Carbon
Nitrogen
Sulfur
Oxygen
Heating Value
Fd Factor"
wt%
wt%
wt%
wt%
wt%
Btu/lb
dscf/106Btu
5.4
69.6
1.5
3.6
11.0
12,438
9915
5.4
68.3
1.4
4.0
11.0
12,070
10,069
5.4
66.5
1.3
4.0
13.3
11,938
9860
5.4
66.8
1.3
3.8
13.5
11,940
9876
5.4
67.8
1.4
3.9
12.2
12,097
9930
 1 Except where stated, all results are on an as-received basis.
 5 As defined in EPA Method 19.
    Table 2-6
    Analysis of ESP Hopper Ash and FGD Material for Site S2
Date
7/16/02
7/17/02
7/18/02
7/19/02
7/22/02
7/23/02
Average
ESP Hopper Ash
Hg,
M9/9
0.025
0.034
0.072
0.032
0.049
0.140
0.059
LOI,
%
0.97
0.51
1.15
0.68
0.71
1.16
0.86
FGD Material
Hg
M9/9

0.16
0.14
0.16

0.12
0.15
Solids,
%

17.3
18.5
16.5

16.0
17.1
2-12

-------
                                                                              Site S2

carbon content in the ash. Both the Hg and LOT results are consistent with the data obtained
in 2001. For the FGD samples, the average Hg concentration was 0.15 |ig/g.


2.4.5  NH3 Slip and SO3 Flue Gas Results

NH3 slip testing was conducted at the SCR and ESP outlets and the SO3 testing at the SCR
outlet location. A summary of these results is provided in Table 2-7. The NH? slip
concentrations were low, less than 1 ppm. In general, low NH? slip values are representative
of an efficiently performing SCR.
    Table 2-7
    NH3 Slip and SO3 Results at Site S2a
Date
07/16/02
07/17/02
07/18/02
07/19/02
07/20/02
07/20/02
NH3 Slip, S03,
ppm ppm
SCR Outlet ESP Outlet SCR Outlet
0.48
0.47 30.2
983b
1141b
0.52
0.56
 1 Dry and 3% O2.
 5 These two values appear to be outliers.
Unfortunately, there appears to be some contamination in two of the SO3 values. It is
unlikely that the 863 concentration in the flue gas is greater than 900 ppm. A careful review
of the procedures, sample sheets, and analysis did not provide an obvious explanation. One
data point (30.2 ppm) from the SCR outlet was similar to data collected from 2001 and is
consistent with expected values. The average ESP outlet 863 value in 2001 was 33.2 ppm.


2.5    Mercury Mass Balance

The average ESP hopper ash results were compared with flue gas Hg measurements from the
ESP inlet and outlet to determine the Hg mass balance across the ESP. The sum of the ESP
hopper ash (in jig/Nm3 of flue gas as calculated from the dust loading) and the ESP outlet Hg
concentration divided by the ESP inlet Hg concentration results in a balance of 94% (from
Table 2-3).
                                                                               2-13

-------
Site S2

To compare the available Hg in the coal to flue gas measurements, emission factors (Fd
factors) as calculated using EPA Method 19 were used to estimate the coal-based Hg
concentration. Using the average coal Hg concentration and Fd factors results in a flue gas
concentration of 10.9 lb/1012 Btu. This compares to a flue gas Hg measurement at the ESP
inlet of 9.1 lb/1012 Btu, giving a balance of 83%.

The plant did not provide information as to the rate at which FGD material was produced.
Therefore, it is not possible to do a mass balance around the wet FGD. However, using the
factors, the flue gas Hg measurements at the stack averaged 1.43 lb/1012 Btu, which
corresponds to a removal of 87% (based on the coal Hg) compared to a measured Hg
removal of 84%.
2.6    General Observations from S2

•  There was increased Hg oxidation across the SCR catalyst as the percentage of Hg2+ in
   the flue gas increased from 54% to 87%. At the ESP inlet and outlet location, the
   percentage of Hg2+ was 97%.

•  Comparing the 2002 results with those obtained in 2001 indicated a small decrease in the
   Hg oxidation across the SCR catalyst. In 2001, there was an increase from 48% to 91% in
   Hg2+ compared to an increase from 54% to 87% in 2002. It is unknown if this is due to a
   catalyst-aging effect, changes in the operation of the SCR unit (lower temperature), or the
   addition of alkali upstream of the SCR unit. Although there was a slight decrease in Hg
   oxidation across the catalyst, the percentage of Hg2+ at the ESP inlet location was the
   same in 2002 and 2001: 97%. The overall Hg removal at Site S2 averaged 84% in 2002
   compared to 87% in 2001.

•  There appeared to be some reemission of Hg across the wet FGD. The Hg° increased
   from 0.3 |ig/Nm3 at the inlet to the wet FGD to 1.3 |ig/Nm3 at the stack. However, this is
   within the variability of the data.

•  Operational changes during the month resulted in data variability; however, those
   fluctuations resulted from acknowledged changes in system operation and do not
   represent steady state.

•  OH results correlated well with Hg SCEM data and were consistent with repeated
   samples.

•  Very little, if any, Hg was removed across the ESP.
2-14

-------
SITE S4
Site S4 was selected to provide data to determine the impact of catalyst aging on the potential
of the SCR units to oxidize Hg. The mercury sampling and analysis at this site was
conducted by Western Kentucky University and was sponsored by EPRI and the host utility.


3.1    Site Description and Configuration

Site S4 is a cyclone boiler that fires a Kentucky bituminous coal and employs SCR followed
by a combined particulate/SO2 venturi/spray tower scrubber (for purposes of this report, this
will be referred to as a venturi scrubber). The venturi scrubber has a total  of six modules. The
system achieves a high particulate removal using a venturi to generate very small water
droplets that create a high relative velocity between the particle and droplets. Some  of the
scrubbing slurry containing limestone for 862 removal is introduced at the top of the
converging venturi section. However, most of the limestone slurry is introduced through
conventional spray nozzles attached to three spray headers in the annular  spray tower section
of the scrubber. The scrubbed gas and entrained droplets enter a separator before the flue gas
exits the stack. The spent slurry is discharged to an on-site disposal pond.

The SCR unit at S4 has a space velocity of 2275 hr"1 and contains a vanadium/titanium
honeycomb catalyst manufactured by Cormetech. The catalyst is spread into three layers in
the SCR unit. The NH3-to-NOx ratio was specified to be 1.0. The unit is designed to be
operated only during the ozone season (May 1-September 30) (To use all the NH3 on-site,
the SCR was operated 15 additional days in 2002 in October). During the remainder of the
year, the SCR is bypassed, but continually pressurized with heated ambient air. Prior to
testing in 2002, the SCR unit had been operated for approximately two ozone seasons. Flue
gas testing was conducted with SCR operating normally and again with SCR bypassed.
General information about the unit configuration is below:

•  Fuel type: Kentucky bituminous coal

•  Boiler capacity: 704 MW gross

•  Boiler type: cyclone boiler with overfire air to reduce NOx

•  NOX control: SCR

•  SO2 and particulate control: combined particulate/SO2 venturi/spray tower scrubber

A schematic of Site S4, including sample locations, is shown in Figure 3-1.
                                                                                3-1

-------
Site S4
                                                                      EERCJT2Q43Q.CDR
                                            Lime Venturi
                                             Scrubber
                                   Stack
                                       Air Heater Outlet
                                       Sample Location
         SCR Inlet         SCR Outlet
     Sample Location  Sample Location
    Boiler
                   Stack Sample
                      Location
                                                                    Stack
                        SCR
AH
Lime Venturi
 Scrubber
   Figure 3-1
   Schematic of Site S4 Showing Sample Locations from a Vertical and Horizontal
   Perspective
3.2   Sampling Approach

Sampling at S4 was conducted similar to testing conducted the previous year and
documented in Power Plant Evaluation of the Effect of Selective Catalytic Reduction in
Mercury [11]. The objective of testing this unit again in 2002 was to evaluate the effect of
catalyst aging on speciated Hg emissions.
 5-2

-------
                                                                              Site S4

3.2.1  Flue Gas Sample Streams

With SCR in service, flue gas Hg speciation was measured at four locations using the OH
method, the SCR inlet and outlet, the outlet of the air preheater (venturi scrubber inlet), and
the stack. With SCR out of service, sampling was only done at the air preheater outlet and
stack. These locations are identified in Figure 3-1. A test matrix is provided in Table 3-1. To
best quantify the effect SCR and the venturi scrubber had on Hg speciation and
concentration, OH measurements were completed as paired sets across each device. In
addition to Hg measurements, flue gas sampling was done to measure particulate loading,
    concentration, and NH? slip.
   Table 3-1
   Sampling Test Matrix for Site S4
DC
Begin
ite
End
SCR In
OH
SCR Out
OH
ESP In
OH
Stack
OH
SCR In
S03
SCR Out
S03
SCR Out
NH3
                                      With SCR

 09/11/02   09/13/02      3333

                                     Without SCR

 10/16/02   10/17/02                         3        3
Longer-term Hg monitoring was conducted using a Hg SCEM (PSA) located at the air heater
outlet (same as the venturi scrubber inlet). Except for periods of maintenance and when the
unit was down, the Hg SCEM was operated around the clock for the duration of the project.


3.2.2  Other Sample Streams

To determine the fate of Hg throughout the unit, samples of coal and venturi scrubber slurry
were taken and analyzed for total Hg. A coal sample taken from the coal yard was associated
with each day of OH sampling. The venturi scrubber samples were taken as the slurry was
drained to the settling tank.


3.3    Process Operating Conditions

Plant operational data are presented in Figure 3-2 for Site S4. It should be noted, with the
exception of load and boiler Q^ data, the operational data were collected by the plant only
during the period when the OH sampling was conducted (beginning and end of test).  As the
figure shows, the operation at Site S4 during this test program was representative of normal
daily operation at or near full  load,  and there was little variation during the 45 days of the test
program, excluding when the  unit was down for 3 days. The NOx removal efficiency for the
SCR unit averaged 87%.
                                                                                3-3

-------
Site S4
                                                                     EERCCWziserCDR
                     9/15/02
9/22/02
9/29/02
10/6/02
10/13/02
10/20/02
     700
                       144
  312
  480
                                           Time, hr
  648
  814
    Figure 3-2
    Plant Operation Data for Site S4 (note the plant logged some of the data, i.e., SCR inlet
    and outlet NOX, only when the OH samples were being done)
5-4

-------
                                                                               Site S4
The average auxiliary flue gas data for Site S4 are shown in Table 3-2. The complete data set
is shown in Appendix C, Table C-2. The excess C>2 at the boiler exit was 3.7% ± 0.8% over
the entire sampling period of 45 days. However, as is typical at most power plants, there is a
substantial air leak across the air preheater. Based on the measured Q^ at the air heater outlet/
venturi scrubber inlet, the average excess C>2 was 7.2%. The air leakage across the SCR and
venturi scrubber was minimal. The particulate removal efficiency of the venturi scrubber is
more than 99.9%.
   Table 3-2
   Average Auxiliary Flue Gas Data for Site S4a
Date
Flue Gas Moisture,
%
Dust Loading,3
gr/dscf
C02,
%
02,
%
                                     With SCR in Service
SCR Inlet
SCR Outlet
Air Preheater Outlet
Stack
10.2
9.5
8.7
15.2
1.93
1.39
1.10
0.00b
15.0
14.9
9.0
11.4
3.9
3.9
10.7
7.8
Air Preheater Outlet

Stack
          With SCR Bypassed

9.1                 1.24               11.1              7.9

14.0                0.01b               11.1              7.9
a Dust loadings were collected as part of the OH method using EPA Method 17 and, therefore, are not for compliance purposes.
b Measured to only two significant digits (1/100 of a gram).


3.4    Sampling Results


3.4.1  OH Flue Gas Mercury Results

The Hg results for Site S4 with the averages and percentage of each species are shown in
Table 3-3. The complete OH results for Site S4 are shown in Appendix B, Tables B-3 and
B-4.  Figure 3-3 shows a comparison of the data with SCR in service and with the SCR unit
bypassed.  As shown in Table 3-3, significant oxidation occurs across the SCR catalyst, from
33% Hg2+ to 63% Hg2+. The percentage of Hg2+ is further increased to 96% at the outlet of
the air preheater. The overall Hg removal is 91% when SCR is in service compared to only
44% when the SCR unit is bypassed.

A comparison of the 2001 and 2002 results at Site S4 is shown in Figure 3-4. As can be seen
in the figure, there was a decrease in the oxidation across the  SCR catalyst in 2002. In 2001,
the concentration of Hg2+ as a percentage of total Hg increased from 9% to 80% across the
SCR catalyst. This  is compared to only 33% to 63% in 2002.  However, there is no significant
                                                                                 3-5

-------
Site S4
   Table 3-3
   Average Ontario Hydro Mercury and Results for Site S4
Sample Location
Average OH Hg Results
Hgp Hg2+ Hg° Hg(total)
Percent of Total Hg,
Hgp Hg2+ Hg°
                                   With SCR in Service
SCR Inlet
SCR Outlet
AH Outlet
Stack
0.05
0.00
0.06
—
Total Mercury Removal = 91%
4.0
7.1
11.3
0.3

8.3
4.3
0.5
0.8

12.3
11.4
11.8
1.1
0 33
0 63
0 96
— 27

67
37
4
73

With SCR Bypassed
AH Outlet
Stack
0.08
NV
Total Mercury Removal = 44%
7.7
0.5

5.6
7.1

13.4
7.5
1 57
— 7

42
93

               20
                Note: Error bars represent standard deviation for total Hg
                                                              EEflC Ol22S25.COfl
           CO
            E
           _g
           Hh— *
           CO
            0)
            o

            o
           O

            £r
            o
            CD
r~~] Particulate-Bound Hg
     Oxidized Hg
     Elemental Hg
               15 -
               10 -
                       SCR   SCR   AH   Stack
                       Inlet   Outlet Outlet

                        SCR Normal Operation
                                     AH   Stack
                                    Outlet

                                   SCR Bypassed
   Figure 3-3
   Comparison of Mercury Speciation Results with the SCR in Service and with the SCR
   Bypassed
5-6

-------
                                                                                Site S4
               20.0
                                                                 EERC DD11K3 CDR
                         Particulate-Bound Hg
                         Oxidized Hg
                         Elemental Hg
           c
           o
           c
           (I)
           o
           c
           o
           O
               15.0 -
               10.0 -
                                                         SCR
                                                        Outlet
 AH
Outlet
Stack
                            2002 Results
                                                         2001 Results
    Figure 3-4
    Comparison of Mercury Speciation Results 2001 and 2002 for Site S4

difference between 2001 and 2002 results as measured at the air preheater outlet location;
also, the overall Hg removal was the same in 2002 as compared to 2001: 91% compared to
90%. There is some question whether the decrease in Hg oxidation across the catalyst is due
to catalyst aging or some other factor. As will be discussed later in this section, the chloride
content of the coal appeared to vary considerably in 2001 and was more consistent in 2002.
This variation may have affected the Hg oxidation across the SCR catalyst.

It appears that Hg reemission can occur across a wet FGD system. At Site S4, there is an
increase in the concentration of Hg° across the venturi scrubber; however, it is very small
(0.5 to 0.8 |ig/Nm3) and is within the variation of the  data. This is discussed in more detail in
Section 6.4.


3.4.2  Hg SCEM Results

A Hg SCEM was  operated at the air heater outlet location at Site S4. In an effort to gather
longer-term data, the Hg CEM was operated nearly continuously for the duration of the
project except when the boiler was down (the Hg SCEM was operated to  alternate between
total Hg and Hg°). A summary of Hg SCEM data plotted over the entire test period is
provided in Figure 3-5. There is significant variability in both the total Hg and Hg° data.
Table 3-4 shows the statistical variation of the SCEM data with and without SCR in service.
Based on the Hg SCEM data, there was an increase in the average Hg°  from 2.26 to 6.3
|ig/m3 when the SCR unit was bypassed at about 880  hours into the test. In Figure 3-6, the
percentage of Hg2+ as determined using the Hg SCEM (total Hg - Hg°) is plotted. Figure  3-2
clearly shows substantial Hg2+ variability.
                                                                                  3-7

-------
oo
                                                                                                                  EERC DL2184B CDR
                                                                                                                               40
                                                                                               • Hg SCEM Total Gas-Phase Hg
                                                                                                 Hg SCEM Elemental Hg
                                                                                                                             - 35
            240 264 288 312 336 360 384 408 432 456 480 504 528 552 576 600 624 648 672 696 720 744 768 792 816 840 864 888 912 936 960 984

                                                              Hours into Test
         Figure 3-5
         Hg SCEM Results for Site S4

-------
                                                                               Site S4
   Table 3-4
   Statistical Variation of the Mercury with and Without the SCR in Service Based on the
   Hg SCEM  Data for Site S4
Mercury
Hg(total)
Hg°
Hg(total)
Hg°
Operation
With SCR
With SCR
SCR bypassed
SCR bypassed
Average,
ug/m
11.5
2.3
14.6
6.3
Std. Dev.,
ug/m3
4.5
1.7
3.6
1.9
Upper 90% Cla,
ug/m3
18.5
5.1
20.5
9.4
Lower 90% Cl,
ug/rn3
4.1
0.0
8.7
3.2
1 Cl = Confidence Interval
        100
                                                                  EERCCW22190.CDR
          9/16
9/23
9/30         10/07
   Calendar Time
10/14
10/21
   Figure 3-6
   Average Hg2+ as Measured by Hg SCEMs (Total Hg-Hg°) for Site S4
3.4.3  Coal Analysis Results

As stated in Section 3.1, Site S4 burns a medium-sulfur Kentucky bituminous coal in a
cyclone boiler. The analysis of the coal fired is shown in Table 3-5. The coal analysis shows
that the coal was relatively constant in 2002. However, in 2001 at Site S4, the chloride
content in the coal increased from an average of 360 ppm with SCR in service to 1160 with
SCR bypassed. Therefore, it is possible that some of the differences in Hg oxidation across
the SCR catalyst between the 2001 and 2002 testing may be due to  differences in the coal.

                                                                                 3-9

-------
Site S4
   Table 3-5
   Coal Analysis for Site S4a
Parameter
Mercury
Chlorine
Units
ppm (dry)
ppm (dry)
9/11/2002
0.17
269
9/12/2002
0.16
228
9/13/2002
0.15
241
10/16/2002
0.19
270
10/17/2002
0.14
295
                                 Proximate Analysis
Moisture,
Volatile Matter
Fixed Carbon
Ash
wt%
wt%
wt%
wt%
12.3
35.1
44.1
8.5
Ultimate
Hydrogen
Carbon
Nitrogen
Sulfur
Oxygen
Heating Value
Fd,b
wt%
wt%
wt%
wt%
wt%
Btu/lb
dscf/106Btu
6.8
65.7
1.3
2.8
14.1
11,597
10,397
12.6
35.3
43.4
8.7
Analysis
6.9
66.3
1.3
2.7
13.8
1 1 ,468
10,664
11.8
35.5
43.8
8.8

6.8
66.5
1.3
2.8
13.1
11,409
10,733
11.3
35.5
45.0
8.2

5.6
74.6
1.6
2.8
7.0
11,852
11,217
11.2
35.5
45.0
8.4

5.5
77.2
1.9
2.7
4.9
11,848
11,592
 1 Except where stated, all results are on an as-received basis.
 3 As defined in EPA Method 19.
3.4.4  Mercury Collected by the Venturi Scrubber

As shown in Table 3-6, there appears to be less Hg captured by the venturi scrubber when the
SCR is bypassed, as expected.
5-10

-------
                                                                              Site S4
   Table 3-6
   Partitioning of Mercury in Material Collected from Venturi Scrubber

     Date          H9in FGD Material,        Solids %
9/11/2002
9/12/2002
9/12/2002
9/13/2002
10/16/2002
10/17/2002
0.14
0.11
0.10
0.08
0.03
0.04
15.6
16.2
14.3
15.2
12.4
12.5
3.4.5  NH3 Slip and SO3 Flue Gas Results for Site S4

The results for NH? slip and 863 concentrations for each of the tests conditions are shown in
Table 3-7. The NH3 slip concentrations are low (less than 2 ppm), indicating a well-
performing SCR. This is also shown by 87% NOx removal efficiency. Within the statistical
variation of the measured values, the 863 concentration at the SCR outlet and inlet locations
were the same. At Site S4, it does not appear SCR increased the conversion of 862 to 863.
Results were similar to those obtained from testing conducted in 2001.
   Table 3-7
   S4 Flue Gas, NH3 Slip, and SO3 Results for Site S4a
Test Condition
SCR On-Line
SCR On-Line
SCR On-Line
Date
9/11/2002
9/11/2002
9/11/2002
NH3 Slip SCR
Outlet, ppm
0.04
0.18
1.33
S03 SCR Inlet,
ppm
10.9
13.4
—
SO3 SCR Outlet,
ppm
14.4
10.3
—
a All results are reported on a dry, 3% O2 basis.


3.5    Mercury Mass Balance

The Hg balance is determined by comparing the concentration of Hg in sources entering the
plant to the concentration of Hg in the sources emitted from the plant. Site S4 has a venturi
scrubber. Without information regarding scrubber flow, slurry, and blowdown rates, it is not
possible to do a mass balance around the FGD. The average Fd factors for Site S4 are shown
in Table 3-8. A Hg balance comparing the measured Hg in the flue gas at the air heater outlet
location compared to the Hg generated by the coal is 86% with SCR and 99% for the test
conducted with SCR bypassed.
                                                                               5-11

-------
Site S4
   Table 3-8
   Average Mercury Emission Factors for Site S4
Operating Condition
SCR in Service
SCR Bypassed
Coal,
lb/1012Btu
10.8
10.8
Air Preheater
Outlet,
lb/1012Btu
9.3
10.5
Stack,
lb/1012Btu
0.9
6.9
Overall Hg
Removal,
%
92
45
3.6    General Observations from S4

•  There was increased Hg oxidation across the SCR catalyst as the percentage of Hg2+ in
   the flue gas increased from 32% at the SCR inlet to 62% at the outlet. At the air preheater
   outlet location, the percentage was 96%.

•  Comparing the 2002 results with those obtained in 2001 indicated that the percentage of
   Hg oxidation that occurred across the SCR unit in 2001 decreased. It is unknown if this is
   due to a catalyst-aging effect or changes in the coal composition, in particular the
   chloride concentration. Although there was a decrease in Hg oxidation across the catalyst,
   the overall Hg removal at Site S4 did not change:  91% in 2002 compared to 90% in 2001.
•  Comparing the Hg speciation results (at the air preheater outlet location) with and
   without the SCR in service showed that the presence of the SCR unit resulted in
   increased Hg oxidation from 57% without SCR to 96% with SCR.  As a result, the overall
   Hg removal across the venturi scrubber increased from 44% to 91%, when the SCR was
   in service.
•  Although there was an increase in Hg° across the venturi scrubber, it was very small: 0.5
   to 0.8 |ig/Nm3. This is within the variability of the data.

•  There was substantial variability in Hg and Hg speciation as measured using the Hg
   SCEMs.

•  Based on the measurement, it appears that the SCR unit did not result in 862 to 863
   oxidation (note: these results are similar to those generated in 2001).
5-12

-------
4
SITE S5
Site S5 was selected for inclusion in the 2002 SCR testing program to provide additional data
on the effect of SCR on speciated Hg emissions for an eastern bituminous coal with a wet
FGD, in particular, how it impacts the Hg removal across a wet FGD. Two "sister" units
were tested at Site S5. They are essentially the same design except one does not have an SCR
unit.


4.1    Site Description and Configuration

Site S5 fires a West Virginia high-sulfur bituminous coal. The unit with SCR has a plate
configuration catalyst manufactured by Halder-Topsoe. The SCR unit had a space velocity of
3700 hr"1 and had operated for approximately 3 months prior to testing. Both the units tested
operated ESPs for particulate control and wet FGDs to reduce 862 emissions. Information
about the configuration of the two units is presented below:

•  Fuel type: West Virginia bituminous coal

•  Boiler capacity: 684 MW

•  Boiler type: wall-fired pc

•  NOx control:  SCR on one unit; low-NOx burners on both units
•  Particulate control: ESP

•  SC>2 control: magnesium-enhanced lime FGD

Schematics of the two units at Site S5, including sampling locations, are shown in Figures 4-
1 and 4-2. As shown in the figures, the ESP configuration was slightly different. The unit
without SCR had a second ESP in series.
                                                                               4-1

-------
Site S5
                                                               EERCCW21139.CDR
     NH3 Injection
                                                                         Stack
                                                      ESP
                                                     Outlet
                         SCR Outlet
         SCR Inlet     Sample Location
      Sample Location             \
    ESP Outlet
 Sample Location

    ESP Inlet   \  stack Sample
Sample Location \    Location






\
V
w








cr^n

I





'""""""""'
	
H
I 	 I
AH

i









P<5P











COD
\





I
\
i



      Boiler
                                                                   Stack
   Figure 4-1
   Schematic of Site S5 Showing Sample Locations for the Unit with the SCR from a
   Vertical and Horizontal Perspective
4-2

-------
                                                                            Site S5
                                                              EERCCW21157.CDR
                                                                      Stack
                                ESP Inlet
                            Sample Location
                                                               Stack Sample
                                                                 Location
   ESP Outlet
Sample Location
     Boiler
                                           ESP
                                                          FGD
                                                                 Stack
   Figure 4-2
   Schematic of Site S5 Showing Sample Locations for the Unit with No SCR from a
   Vertical and Horizontal Perspective
4.2    Sampling Approach


4.2.1  Flue Gas Sample Streams

The flue gas Hg speciation was measured using the OH method at five locations for the unit
with SCR and three locations for the unit without SCR. A test matrix, which identifies the
location of flue gas measurements, is provided in Table 4-1. Where practical, OH measure-
ments were conducted simultaneously across the various control devices in an effort to
quantify the effect each had on Hg concentration and speciation. In addition to Hg, flue gas
                                                                              4-3

-------
Site S5
   Table 4-1
   Sampling Test Matrix for Site S5
Date
Begin
End
SCR In SCR Out ESP In ESP Out Stack
OH OH OH OH OH
AH In
S03
ESP In SCR Out
S03 S03
SCR Out
NH3
07/26/02

08/15/02
07/28/02

08/23/02
   With SCR

        3

352

  Without SCR
07/26/02
08/13/02
07/28/02
08/23/02
3
343
2 2

samples were collected to measure the total particulate loading and 863 concentrations.
Additionally, NH? slip samples were collected from the unit with SCR to evaluate
performance.

Longer-term Hg monitoring was conducted using Hg SCEMs (PSA) located at the ESP outlet
(same as the FGD inlet) locations for both test units. These data provided semicontinuous
Hg° and total gas-phase Hg concentrations for approximately 3 weeks.


4.2.2  Other Sample Streams

Samples of coal and ESP hopper ash were collected from both test units in an effort to obtain
representative operational data related to Hg speciation. These samples were analyzed for Hg
and, along with the flue gas emission data, were used to qualitatively evaluate the fate of Hg
throughout the units. Coal samples were collected from the coal feeder of both units daily
throughout the test period. ESP hopper ash samples were collected from the first fields of the
ESPs. Ash samples for the unit with SCR were obtained from both Sides A and B of the ESP;
however, the sample collected from the unit without SCR was obtained from only Side A of
the ESP. Plant personnel did not collect samples from the FGD system.


4.3    Process Operating  Conditions

Plant operational data are presented in Figures 4-3 and 4-4 for the two test units. These
figures summarize flue gas characteristics during the test program. Additionally, month-long
Hg SCEM data are included in these plots for comparison with plant operational data. Hg
SCEM data will be discussed later in this report. In general, for the unit with SCR, plant load
remained greater than 80% of full  capacity, with the exception of a 35-hour period approx-
imately 400 hours into the test and a few short reductions at night. These reduced load
conditions did not have  a significant impact on Hg as measured by the Hg SCEMs or OH
4-4

-------
                                                                                      Site S5
                                                                     EERCCW211-IO.CDR
              700
                            8/2/02
                                        8/9/02
                                                    8/16/02
                                                                8/23/02
                            168
                                         336          504
                                            Time, hr
                                                                 672
                                                                             840
Figure 4-3
Plant Operation Data for Site S5 for the Unit with the SCR
                                                                                (continued)
                                                                                         4-5

-------
Site S5
                                                                      EERCCW21U1.CDR
                              8/2/02
                                          8/9/02
                                                     8/16/02
                                                                 8/23/02
                                                                                 700
                      Total Gas-Phase Hg
                               168
                                           336          504
                                             Time,  hr
                                                                  672
                                                                              840
    Figure 4-3 (concluded)
    Plant Operation Data for Site S5 for the Unit with the SCR
4-6

-------
                                                                                Site S5
             700



             600



             500



             400



             300


            2500



     T  , > 2000

     ®n  E

     £ CO  g; 1500
 ~"lg
 •E Si
 3 -i
_0>

c
     O
     CO

     5
     "5
     O  . >

     fees1!
     gwg:

     2
     o
     to
       II
       w
            1000-
       w
          O)

          3.
         O
          O)

          15.
       w
       LLJ
         "
         O
                                                                 EERCCW21143.CDR
                          8/2/02
                                      8/9/02
                                                8/16/02
                                                           8/23/02
                                                                           700
                      Total Gas-Phase Hg
                                                         '
                          168
                                      336         504


                                         Time, hr
                                                            672
                                                                        840
                                                                        15



                                                                        10



                                                                        5



                                                                        0


                                                                        4



                                                                        3



                                                                        2



                                                                        1



                                                                        0
Figure 4-4

Plant Operation Data for Site S5 for the Unit with No SCR
                                                                           (continued)
                                                                                   4-7

-------
Site S5
                             8/2/02
                                         8/9/02
             O 400

               300 -
          O
                10 -
                 8 -
OQ


      0
    1000 -

Q    800 -

^  § 600 -
_Q)  Q.
~  Q. 400-|
CD
     200 -
        L
                                                                                - 500
                                      -  400

                                      -  300

                                        10
                                           vUf^
                                                                 JL
                                         0
                                      - 1000

                                      -  800

                                      -  800

                                      -  400

                                      -  200

                                         0
                              168
336          504
    Time, Hr
                                                        672
840
   Figure 4-4 (concluded)
   Plant Operation Data for Site S5 for the Unit with No SCR

method. The unit without SCR did not experience any significant load reductions below 80%
of full capacity during the test program.

For the unit with SCR, the inlet NOx concentration reduced noticeably beginning around
300 hours into the test from approximately 260 to less than 200 ppmv. In response to that
reduction, the NH3 injection rate was reduced to maintain an SCR outlet NOx concentration
of approximately 50 ppm. Also, note at about 750 hours into the test, there is a substantial
increase in the NIT? injection rate that corresponds to a spike in the SCR inlet NOx
concentration. Overall, an 80% reduction in NOx was measured across the SCR unit.

Boiler CO for both units spiked intermittently from less than 50 to 400-800 ppm. However,
there does not appear to be  any correlation between elevated CO concentration and shifts in
Hg concentration from either OH or Hg SCEM results.

A summary of auxiliary flue gas data, including  percent O2 and percent CO2 for each sample
location, is provided in Table 4-2. The complete data set is located in Appendix C, Table C-
3. In general, the percent moisture, CO2, and O2  were very consistent from day to day.
However, there was air leakage across the SCR unit, air preheater, and wet FGD system that
resulted in the O2 increasing from 4% at the boiler outlet to 7.8% at the stack. Dust-loading
measurements collected at the ESP inlet and outlet location reflect a paniculate removal
4-8

-------
                                                                             Site S5
   Table 4-2
   Auxiliary Flue Gas Data for Site S5a
Date
SCR Inlet
SCR Outlet
ESP Inlet
ESP Outlet
Stack

Moisture,
10.7
9.1
8.9
8.7
13.1
Unit with
Dust,
gr/dscf
3.5652
3.4083
1 .6848
0.0751
0.0073
SCR
C02,
14.3
13.8
13.1
12.9
11.9

02,
5.0
5.7
6.5
6.7
7.8
Unit Without
Moisture, Dust,
% gr/dscf


8.6 0.8287
9.0 0.0453
13.5 0.0063
SCR
C02,


13.6
12.9
12.7

02,


6.1
6.7
7.0
a Dust loadings were collected as part of the OH method using EPA Method 17 and, therefore, are not for compliance purposes.

efficiency of approximately 95% for both units based on an average of inlet and outlet dust
loadings. Based on discussions with plant personnel, the ESPs at S5 are not extremely
efficient, which is reflected by these values.


4.4    Sampling Results


4.4.1  Ontario Hydro Flue Gas Mercury Results

The average Hg results from gas sampling are summarized in Table 4-3. The complete data
sets are presented in Appendix B (Tables B-5 and B-6). Figure 4-5  shows a comparison of
the data for the unit with the SCR and the unit without an SCR. As  shown in Table 4-3, there
is significant oxidation of Hg occurring across the SCR catalyst, from 43% Hg2+ to 76%
Hg2+. The percentage of Hg2+ is further increased to 95% at the inlet to the ESP. It should be
noted that the apparent increase in Hg° across the ESP for the unit with ESP is most likely
due to reactivity with the fly ash across the filter of the OH method.

Comparing the ESP inlet Hg speciation results for the two units indicates that the portion of
Hg2+ was 80% without SCR and 95% with SCR. However, as shown by the error bars in
Figure  4-5, the OH data at this location were highly variable. If the ESP outlet data are used,
the difference is from 63% to 94%.  This is more in line with the overall Hg removal of 91%
with SCR as compared to only 51% for the unit without SCR.

One objective for testing Site S5 was to evaluate the combined effect of an  SCR unit and wet
FGD system on Hg speciation and removal. For the unit without SCR, the measured Hg2+
was 63% of the total Hg. However the Hg removal efficiency of the wet FGD system was
only 51%. This is a result of the Hg° increasing from 4.7 to 6.1 |ig/Nm3 across the wet FGD
system. With SCR, Hg2+ is 94% of the total Hg, and the total Hg removal by the wet FGD
system is 91%. There is still an increase in Hg° (0.7 to 1.0 |ig/Nm3) across the wet FGD
                                                                                4-9

-------
Site S5
    Table 4-3
    Average and Percentage of Total Ontario Hydro Mercury Results for S5
Measurement
Location
Average,
ug/Nm3
Hgp Hg2+ Hg° HgT
Portion of Total,
%
Hgp Hg2+ Hg°
                                     Unit with SCR
SCR Inlet
SCR Outlet
ESP Inlet
ESP Outlet
Stack
0.09
0.02
0.07
0.05
0.02
Average total mercury removal
6.1
10.8
16.8
11.3
0.4
= 91%a
7.8
3.3
0.8
0.7
1.0

14.0
14.3
17.6
12.1
1.5
1
0
0
0
1

43
76
95
94
28

56
24
5
6
72

Unit Without SCR
ESP Inlet
ESP Outlet
Stack
0.05
0.01
0.00
Average total mercury removal
10.8
7.9
0.5
= 51%a
2.7
4.7
6.1

13.5
12.6
6.6
0
0
0

80
63
8

20
37
92

'Average Hg removal is defined: (ESP inlet - stack)/ESP inlet.
           Note: Error bars represent standard deviation of total Hg.
           30-
        o>  25-
                                                               EERC CW21144.CDR
Particulate-Bound Hg
Oxidized Hg
Elemental Hg
                   SCR   SCR   ESP   ESP  Stack
                   Inlet  Outlet  Inlet  Outlet
                        Average with SCR
                                 ESP  ESP  Stack
                                 Inlet Outlet
                               Average Without SCR
    Figure 4-5
    Comparison of Mercury Speciation Results with the SCR and Without an SCR at Site S5
4-10

-------
                                                                              Site S5

system, but it is small. It should be noted that the wet FGD system at Site S5 is a magnesium-
enhanced lime system.

Whether these mercury results would be consistent in the more common limestone forced-
oxidation design is not known since the mechanism for reemission is not well understood.
This is discussed in more detail in Section 6.4.


4.4.2  Hg SCEM Results

Hg SCEMs were operated at the ESP outlet location for both units tested. In an effort to
gather longer-term variability data, Hg SCEMs were operated nearly continuously for 23
days. The Hg SCEM data for the entire test are shown in Figures 4-6 and 4-7.

Significant variability of total gas-phase Hg was observed at the ESP outlet location of both
units. The statistical analysis for the Hg SCEM data is shown in Table 4-4. Correlation
between the Hg SCEM data and the OH method at Site S5 is not very good. Using the OH
method, the average total gas-phase Hg was 9.1 and 9.9 |ig/m3 with and without SCR
operating (dry at actual O2 levels).  This compares to only 5.3 and 5.8 |ig/m3 as measured
using the Hg SCEMs. There were significant problems associated with operating the Hg
SCEMs at Site S5. It took over a week to get the instruments operating, and there were
substantial plugging problems with the sample lines and probe once they began operating. In
general, it is difficult to assess long-term variability of Hg concentration at S5. The abrupt
shifts observed in the Hg SCEM data indicate that Hg concentrations were highly variable at
S5. Complicating  the Hg SCEM data interpretation further is a noticeable increase in Hg
concentration following replacement of the probe or instrument filters.
    Table 4-4
    Statistical Variation of the Mercury with and Without the SCR in Service Based on the
    Hg SCEM  Data for Site S5
Mercury
Hg(total)
Hg°
Hg(total)
Hg°
Operation
With SCR
With SCR
SCR bypassed
SCR bypassed
Average,
ug/m
5.3
0.2
5.8
0.8
Std. Dev.,
ug/m3
3.9
0.4
3.9
0.7
Upper 90% Cl,
ug/m3
11.6
0.9
12.2
1.5
Lower 90% Cl,
ug/rn3
0.0
0.0
0.0
0.0
                                                                               4-11

-------
to
          20
                                                                                                                     EERC DL2184S CDR
          18 -


          16 -
       O)
-
2
c

I
          12 ^
       O   C
       0)
       2
       QJ   6 -
       (A
       05
       03
                •  Hg SCEM Total Gas-Phase Hg
                O  Hg SCEM Elemental Hg
                   I OH Total Gas-Phase Hg
            240  264  288  312  336 360  384  408  432  456 480  504 528  552  576  600  624 648  672 696  720  744  768  792 816  840

                                                                Hours into Test
          Figure 4-6
          Hg SCEM Results for Site S5 for the Unit with an SCR

-------
iu -
18 -
.
CO
£ 16 -
3.
c 14 -
g
2
c 12 -
0)
o
3 10-
1 8 -
£
03 R
c/) O -
to
.C
0.
 4 -
CO
CD

2 -
n -


















• Hg SCEM Total Gas-Phase Hg
O Hg SCEM Elemental Hg
1 1 OH Total Gas-Phase Hg















*
f

\


f




*
t
*
*
*
**
t






I
^
*•






















*

i
•












*
*
* *
*
E
i


»
i£


y
»>
*






4
I
}j
*• 1
*{4
4
* 4

•







*
>
t <
::
r*
t
f%
\

**
* 1 •?







• •
*
•

.















t
1
» t 1 * ';
U'Mj
' \ \ • \'(-
. M ' *
*
• •


' $
*
• .
«
1
{



Li
\
v
•
^
*
l




*
1
1
%








J|
\




.{
I
*
^
*

J
















j




*

*








_ *
*
4
*
.•
* f «








1
t*» •*
* % : «*;
^. .'
j i *
K & ;: c*-
^ ^1 '•• !*
* * +1 <
* \ t
* *r * t


K '! • I %
1
1

i
i \
v;

- 18
.
- 16
- 14

'
- 12
- 10
- 8
- 6
- 4



- 2
. n
       240  264  288 312  336 360  384  408 432  456 480  504  528  552  576 600  624 648  672  696  720  744 768  792  816  840

                                                         Hours into Test
Figure 4-7
Hg SCEM Results for Site S5 for the Unit with No SCR

-------
Site S5
4.4.3  Coal Analysis Results

Coal samples from S5 were analyzed for Hg and chloride. Both the coal Hg and chloride
concentrations were very consistent. For six samples analyzed, the Hg and the chloride
concentrations were 0.13 ± 0.013 and 472 ± 28 ppm, respectively. The analysis of the six
coal samples is shown in Appendix B, Table B-7. Results of proximate and ultimate analyses
for the coal are provided in Table 4-5.
    Table 4-5
    Coal Analysis for Site S5a
Parameter
Mercury
Chlorine
Units 7/28/2002
ppm (dry)
ppm (dry)
0.14
430
8/5/2002
0.12
500
8/15/2002
0.15
480
Proximate Analysis
Moisture
Volatile Matter
Fixed Carbon
Ash
wt%
wt%
wt%
wt%
Ultimate
Hydrogen
Carbon
Nitrogen
Sulfur
Oxygen
Heating Value
Fd Facto rb
wt%
wt%
wt%
wt%
wt%
Btu/lb
dscf/106Btu
5.1
37.7
44.4
12.7
Analysis
5.3
69.4
1.4
3.6
7.6
11,918
10,416
4.8
38.3
45.3
11.6

5.3
69.8
1.4
3.5
8.4
12,164
10,219
3.9
38.7
45.3
12.1

5.3
71.5
1.4
3.8
5.8
12,278
10,460
 1 Except where noted, all results are on an as-received basis.
 5 As defined in EPA Method 19.
4.4.4  ESP Ash Mercury Results

ESP hopper ash samples were collected daily throughout the test period from both test units.
Ash analyses consisted of Hg and LOT determination and are presented in Table 4-6. In
general, the Hg concentration in the ash was less than 0.1 |ig/g with or without SCR. The
LOT for all samples analyzed was less than 7%. Based on these results, only a small amount

4-14

-------
                                                                             Site S5

of Hg was adsorbed by the ash and subsequently removed across the ESP. This is supported
by OH results.
   Table 4-6
   Analysis of ESP Hopper Ash
Unit
With SCR
Without SCR
Hg,
ug/g
0.094 ±0.041
0.068 ±0.016
No. of Samples
16
21
LOI,
%
5.21 ±1.27
3. 96 ±0.87
No. of Samples
5
6
4.4.5  NH3 Slip and SO3 Flue Gas Results

A summary of the NH? slip and SOs results is provided in Table 4-7. The NH? slip was less
than 0.5 ppm for both samples, indicating an efficiently operating SCR. Based on the
expected conversion of 862 to 863 that occurs across the SCR unit, it would be expected that
the unit with SCR would have higher SOs concentrations than the unit without SCR.
Comparing the results at the ESP sampling location, this is indeed the case. However, the
results at the SCR outlet are much lower than seems reasonable, and the ESP inlet 863
concentrations are actually higher. No clear cause has been identified that would explain
these results.
   Table 4-7
   Flue Gas SCh and NI-U Results for Site S5a
   Date
SO3, ppm
NH3 Slip, ppm
                      Unit with SCR
Location
8/1 8/2002
8/1 8/2002
8/22/2002
8/22/2002
SCR Outlet
1.93

1.71
1.76
ESP Inlet

5.28
16.03
18.30
SCR Outlet

0.29
0.34

                    Unit Without SCR
Location
8/14/2002
8/1 5/2002
8/1 5/2002
AH Inlet
10.61
5.06

ESP Inlet
9.92

5.13




1 Dry and 3% oxygen.
                                                                              4-15

-------
Site S5

4.5    Mercury Mass Balance

Average Hg concentration in the coal and Fd factors (Table 4-5) were used to estimate the Hg
emission rate at the various sample locations. For the Hg associated with the ESP hopper ash,
the Fd factors were based on the dust-loading measurements as well as the Fd factor. The
results are shown in Table 4-8.
   Table 4-8
   Average Mercury Emission Factors for Site S5

Unit

With SCR
Without SCR
lbHg/1012Btu
Coal
10.8
10.8
ESP Inlet
13.4
10.3
ESP Hopper Ash
0.2
0.2
ESP Outlet
9.2
9.6
Stack
1.1
5.0
A Hg balance comparing the measured Hg in the flue gas at the ESP inlet location compared
to the Hg generated by the coal is 124% with SCR and 95% for the test on the unit without
SCR.

To determine the mass balance around the ESP the sum of the Hg associated with the ESP
hopper ash plus the Hg in the flue gas at the ESP outlet must equal the Hg measured in the
flue gas at the ESP inlet. The results of this balance for the units with and without SCR are
70%, and 95%. It should be noted that there was substantial variability in the Hg flue gas
measurements. This is particularly true at the ESP inlet location where the standard deviation
was 3.4 lb/1012 Btu for the unit with SCR and 2.9 lb/1012 Btu for the unit without SCR.

Unfortunately, the plant personnel did not provide FGD samples for analysis, and no plant
data  are available (gas flows, slurry feed rate, and blowdown) to estimate the Hg removal
rates by the FGD. Therefore, it is not possible to do a Hg mass balance around the wet FGD
system.


4.6    General Observations from S5

•  There was an increase in Hg oxidation across the SCR catalyst. The amount of Hg2+ in
   the flue gas increased from 43% to 76%. At the ESP inlet and outlet location, the portion
   ofHg2+was95%.

•  Comparing the Hg speciation results (at the ESP inlet location) with and without the SCR
   unit in service showed that the presence of SCR resulted in more Hg2+: 80% without SCR
   and 95% with SCR. There was substantial variability at the ESP inlet locations for both
   units. If the ESP outlet data are used, the difference was 63% to 94%.

•  The overall Hg removal was greater with an SCR: 51% for the unit tested without SCR
   and 91% when SCR was in service.
4-16

-------
                                                                           Site S5
There was an increase in Hg° across the wet FGD system for both units. For the unit
without SCR, the Hg° concentration increased from 4.7 |ig/Nm3 at the FGD inlet to 6.1
|ig/Nm3 at the outlet. The increase was less for the unit with SCR: 0.7 to 1.0 jig/Nm3.

Hg removed by the ESP averages of 0.2 Ib Hg/1012 Btu for the two units.
                                                                            4-17

-------
5
SITE S6
Site S6 was selected for inclusion in the 2002 SCR testing to provide data associated with
burning a compliance low-sulfur eastern bituminous coal. The primary tests were conducted
on two of the four units at Site S6. On one of the units, SCR was operated for the entire
testing period; however, on the other, SCR was bypassed for the majority of the test period.
In addition, at Site S6, additional testing was done at the stack of a third unit (no SCR) as
part of a separate test program. For comparison purposes, the results obtained at this unit are
also presented in this report.


5.1    Site Description and Configuration

Site S6 operates four units consisting of two sets of similar configurations. Two of the four
units have SCR units to reduce NOx, and all four units have ESPs for particulate control. The
SCR catalysts at Site S6 are a honeycomb type and manufactured by Cormetech. The SCR
unit has a space velocity 3800 hr"1. The SCR units have been operating for two ozone
seasons. In between the two seasons, one layer of catalyst was changed. Specifications of the
Site S6 units are presented in Table 5-1. Schematics of the three test units at Site S6,
including sampling locations, are  shown in Figures 5-1 through 5-3.
   Table 5-1
   Specifications of Site S6 Units3
Specification
Fuel Type
Boiler Capacity
Boiler Type
Low-NOx Burners
SCR
Particulate Control
S02
Unitl
KY and WV eastern
bituminous coal
700 MW
tangentially fired
Yes
Yes
ESP
Low-sulfur compliance
coal
Unit 2
KY and WV eastern
bituminous coal
700 MW
tangentially fired
Yes
Operated in bypass mode
ESP
Low-sulfur compliance coal
Unit 4
KY and WV eastern
bituminous coal
900 MW
tangentially fired
Yes
No
ESP
Low-sulfur compliance
coal
1 Site S6 has four units. Unit 3 (no testing was done at this time) is the same as Unit 4.
                                                                                 5-1

-------
Site S6
                                                            EERC CW2-I280.CDR
  NH3 Injection—•
   ESP
                                                                     Stack
                      SCR Outlet
      SCR Inlet      Sample Location
   Sample Location             \        ESP lnlet
                                    Sample Location
                 Stack Sample
                   Location
   Boiler
                        SCR
                                  AH
ESP    ESP
Stack
   Figure 5-1
   Schematic of Site S6 Showing Sample Locations for Unit 1 with the SCR in Service
   from a Vertical and Horizontal Perspective
5-2

-------
                                                                         Site S6
                                                         EERCCW21281.CDR
                                         ESP
Boiler
 SCR
Bypass
                      SCR
                      AH
                              ESP
                              Inlet
                                                                   Stack
                                                      Stack Sample
                                        ESP Inlet         Location
                                     Sample Location






[ 	 I /
1 l n*
1 	 t*
I — I

i VJWI1 | AH





FSP











F^P



J '


Boiler
         Bypass
                                                              Stack
Figure 5-2
Schematic of Site S6 Showing Sample Locations for Unit 2 with the SCR Bypassed
from a Vertical and Horizontal Perspective
                                                                           5-3

-------
Site S6
                                                             EEflCCW212S2.CCfl
                                                                      Stack
                                                       Stack Sample
                                                          Location
   Boiler
                                            ESP
Stack
   Figure 5-3
   Schematic of Site S6 Showing Sample Locations for Unit 4 with No SCR from a Vertical
   and Horizontal Perspective
5.2   Sampling Approach

As stated previously, sampling at S6 was primarily conducted on two units (1 and 2) both
with SCR, but the SCR unit was bypassed on Unit 2. Data collected from these similar units
provided a comparison of speciated Hg emissions from SCR and no-SCR operation.


5.2.1  Flue Gas Sample Streams

The flue gas Hg speciation was measured using the OH method at four locations for Unit 1
(SCR) and two locations for Unit 2 (SCR bypassed). A test matrix, which identifies the
location of flue gas measurements, is provided in Table 5-2. Where practical, OH
5-4

-------
                                                                               Site S6

measurements were conducted simultaneously across the various control devices in an effort
to quantify the effect each had on Hg concentration and speciation. In addition to Hg, flue
gas samples were collected to measure the total particulate loading and SOs concentrations.
Additionally, NH3 slip samples were collected from Unit 1 (SCR) to evaluate performance.

    Table 5-2
    Sampling Test Matrix for Site S6
Date
Begin
End
SCR In
OH
SCR Out
OH
ESP In
OH
Stack
OH
SCR Out
NH3
SCR Out
S03
ESP In
S03
09/22/02
 10/08/02
09/22/02
 10/08/02
09/26/02
10/18/02
09/25/02
10/18/02
        Unit 1 (SCR)

    445

                       7

    Unit with SCR (Unit 2)

                       3

    Unit 2 (SCR bypassed)

              2         7

Unit 4 (no SCR) - Plume Study
10/08/02
1 0/1 8/02
7


Longer-term Hg monitoring was conducted using Hg SCEMs (PSA) located at the stack for
each of the test units. These data provided semicontinuous Hg° and total gas-phase Hg
concentrations for approximately 3 weeks.


5.2.2  Other Sample Streams

Samples of coal and ESP hopper ash were collected daily from the test units in an effort to
obtain representative operational data related to Hg speciation. These samples were analyzed
for Hg and, along with the flue gas emission data, were used to qualitatively evaluate the fate
of Hg throughout the units. Daily coal samples were collected as composites from the
different coal feeders. ESP hopper ash samples were collected from the first fields of the
ESPs.


5.2.3  Process Operating Conditions

Plant operational data are presented in Figures 5-4 through 5-6 for the test units. These
figures summarize flue gas characteristics during the test program. Additionally, month-long
Hg SCEM data are included in these plots for comparison with plant operational data. Hg
SCEM data will be discussed later in this report.
                                                                                 5-5

-------
Site S6
               9/21/02
             800
9/28/02
10/5/02
             EERCCW21327.CDR

10/12/02         10/19/02
                                                          800
    Figure 5-4
    Plant Operation Data for Site S6 for Unit 1 with the SCR in Service
                                                                                   (continued)
5-6

-------
         9/21/02
        800
                       9/28/02
10/5/02
              10/12/02
      Site S6


EEPCCW2T330.CDR

10/19/02

        800
^
55
   Q.
   E
  ,Q>
  _^
   o
  3
  co
 M
o — 2
ZS1
  w 5
  _g> rs
 x^S
oos
Zfcg
  o .9
  w
CD
X,
  |5
  55
   O)
  x.
  |a
  ^_j
  CO
                Total Gas-Phase Hg
                          168
 336

Time, hr
                                                      504
                                672
   Figure 5-4 (concluded)
   Plant Operation Data for Site S6 for Unit 1 with the SCR in Service
                                                                            5-7

-------
Site S6
              9/21/02
            800
9/28/02
10/5/02
10/12/02
 EERCCW2-I329.CDR
10/19/02
         800
     CO
    Figure 5-5
    Plant Operation Data for Site S6 for Unit 2 with the SCR Bypassed
                                                                                 (continued)
5-S

-------
                                                                                 Site S6
         9/21/02
       800
9/28/02
10/5/02
10/12/02
EERC CW22271.CDR

10/19/02
         800
                                        Total Gas-Phase Hg
c/)
                                                                             672
  Figure 5-5 (concluded)
  Plant Operation Data for Site S6 for Unit 2 with the SCR Bypassed
                                                                                   5-9

-------
Site S6
                 9/21/02
              1000 '
9/28/02
10/5/02
10/12/02
EERC CW21326.CDR
10/19/02
        1000
        CO
    Figure 5-6
    Plant Operation Data for Site S5 for Unit 4 with No SCR
5-10

-------
                                                                               Site S6
In general, the plant load at the two primary test units (Units 1 and 2) was near full load
during the day and would drop at night. However, at Unit 4 (no SCR), the plant data showed
a significant load reduction at about 120 hours into the test.

A summary of auxiliary flue gas data, including percent 62 and percent CC>2 for each sample
location, is provided in Table 5-3 (the complete data set is in Table C-3 in Appendix C). In
general, the percent moisture,  CC>2, and C>2 were very consistent from day to day. However,
there was some air leakage across the air preheater that resulted in the 62 increasing from
4.1% at the SCR inlet to 6.5% at the stack for Unit 1 (SCR). The air leakage was about the
same for Unit 2 (SCR bypassed).  Dust-loading measurements collected at the ESP inlet and
outlet location reflect a particulate removal efficiency of more than 99% for both units based
on an average of inlet and outlet dust loadings.
    Table 5-3
    Auxiliary Flue Gas Data for Site S6
Measurement
Location
Moisture,
%

SCR Inlet
SCR Outlet
ESP Inlet
Stack
9.0
8.7
8.9
9.3

ESP Inlet
Stack
8.3
7.8
Dust,3
gr/dscf
Unit 1 (SCR)
3.7306
4.1673
2.7321
0.0165
Unit 2 (SCR bypassed)
4.2279
0.0150
C02,
%

14.7
15.2
13.8
13.1

15.4
13.2
02,
%

4.1
4.7
5.0
6.5

3.7
6.4
Unit 4 (no SCR)

Stack
7.8
0.0388
14.5
4.9
a Dust loadings were collected as part of the OH method using EPA Method 17 and, therefore, are not for compliance purposes.

5.3    Sampling Results

5.3.1  Ontario Hydro Flue Gas Mercury Results

Average Hg results for flue gas sampling at Site S6 are presented in Table 5-4. The complete
results are presented in Appendix B (Tables B-8 and B-9). As shown in Table 5-4, there is an
increase in the concentration of Hg2+ across the SCR catalyst from 60% to 82% Hg2+. There
is only a slight additional increase to 87% at the ESP inlet location (this is within the
variation of the data).
                                                                                 5-11

-------
Site S6
Comparing the ESP inlet Hg speciation results for Unit 1 (SCR) and Unit 2 (SCR bypassed)
shows that the amount of Hg2+ was 69% with SCR bypassed and 87% with SCR. However,
there was essentially no Hg removal across the ESP for either unit. A direct comparison,
including the error bars associated with the total Hg concentration for all three units, is
shown in Figure 5-7.
   Table 5-4
   Average and Percentage of Total Ontario Hydro Mercury Results forS6a
Location
Average,
ug/Nm3
Hgp Hg2+ Hg° HgTotai
Portion of Total,
Hgp Hg2+ Hg°

SCR Inlet
SCR Outlet
ESP Inlet
Stack
0.04
0.03
0.80
0.00

5.8
7.1
8.5
9.3
Unit
3.8
1.5
0.5
0.8
1 (SCR)
9.6
8.6
9.8
10.1
0
0
8
0

60
82
87
92

40
18
5
8
Unit 2 (SCR bypassed)
ESP Inlet
Stack
2.59
0.01

Stack
0.01
6.6
6.0

4.0
0.4
1.3
Unit 4
1.8
9.5
7.3
27
0
(no SCR)
5.8
0
69
82

69
4
18

31
1 All mercury results are on a dry basis corrected to 3% O2.
5-12

-------
                                                                               Site S6
               Note: Error bars represent standard deviation of total Hg.
               20
                                                              EEFIC CW21323.CDR
            E
           z
           1? 15-
            c
           ,2
           "ro
           "c.
            CD
            o
            c
            o
           O
            o
            0)
Particulate-Bound Hg
Oxidized Hg
Elemental Hg
                      SCR  SCR   ESP
                      Inlet  Outlet  Inlet
                           Average with
                              SCR
             Stack
ESP   Stack
Inlet
Average with
SCR Bypass
Stack
                                      Average with
                                        No SCR
   Figure 5-7
   Comparison of Mercury Speciation Results for the Three Test Units of Site S6
5.3.2  Hg SCEM Results

Hg SCEMs were operated at the stack location of the three units tested. The Hg SCEMs were
operated nearly continuously for a month. A summary of Hg SCEM data plotted over the
entire test period for the three units is provided in Figures 5-8 through 5-10. The statistical
average Hg SCEMs results are shown in Table 5-5. For Unit 1 (SCR), the Hg SCEM data
averaged 5.2 |ig/m3 with 90% of the data falling between 2 and 8 |ig/m3. For Unit 2 (SCR
bypassed), the average Hg SCEM result was 5.7 |ig/m3, with 90% of the data points falling
within 1.3 and 10.1 |ig/m3. For Unit 4 (no SCR), the Hg SCEM data averaged 7.5 |ig/m3 with
90% of the data points falling between 5.9 and 9.1 |ig/m3. Although there is some difference
between the OH data and the Hg SCEM averages, the results are within the statistical
variability of the data.

The variability of the Hg2+ is shown in Figures 5-11 and 5-12.  As can be seen, there appears
to be a decrease in the percentage of Hg2+, and there was more variability when SCR was
bypassed. However, as was shown by the OH method, the portion of Hg
more than 90% both with and without SCR.
                                                                 2+
                                          at the stack was
At about 125 hours into the test (see Figure 5-5), there does appear to be a small increase in
Hg° concentration that corresponds with SCR being bypassed. Figure 5-13 presents the Hg
SCEM data during this time period.
                                                                                5-13

-------
o
o
o
O

-------
c
_o
V- '
CD
-*_J



I
o
O
3
o
CD
CD
M
CO

Q_

W
CO

O
20



18 -



16 -



14 -



12 -



10 -



 8 -



 6 -



 4 -



 2 -



 0
                                                                                                               EERCDL21845CDR
                                                      •  Hg SCEM Total Gas-Phase Hg

                                                      O  Hg SCEM Elemental Hg

                                                     I    I OH Gas-Phase Hg
SCR Bypassed
                   *

                   I
  20



- 18



- 16



- 14



- 12



- 10



- 8



- 6



- 4



- 2



  0
        0  24  48  72  96  120 144 168 192 216 240 264 288  312 336 360 384 408 432 456 480 504 528  552 576 600 624 648 672 696


                                                           Hours into Test

   Figure 5-9

   Hg SCEM Results for Site S6 for Unit 2 with the SCR Bypassed

-------
   14


CO



~D>12
 3.
c
,2
1+—*
2
Hh-j
   10 -
 o
O
    8 -
=j  6 -
 CD
 CD
 cfl
 TO
_C
Q_
 CD

O
    4 -
    2 -
            •  Hg SCEM Total Gas-Phase Hg

            »  Hg SCEM Elemental Hg

          f~~~l OH Total Gas-Phase Hg
                                                                                                            EERCDL2i849.CDR
                                                                                                                         14
                                       - 12
                                       - 10
                                                                                                                        - 8
                                                                                                                        - 6
                                                                                                                       - 4
     336     360    384     408    432     456
                                                  480     504    528     552

                                                       Hours into Test
576    600     624    648     672     696
  Figure 5-10
  Hg SCEM Results for Site S6 for Unit 4 with No SCR

-------
Table 5-5
Statistical Variation of the Mercury Results Based on the Hg SCEM Data for Site S6
Mercury
Hg (total)
Hg°
Hg(total)
Hg°
Hg(total)
Hg°
Averaae Std Dev Upper 90% Lower 90%
Unit Operation ™e^e' Stdg/Dme3v'' cia Cl
ug/m [jg/m
1 With SCR 5.2 1.7 8.0 2.4
1 With SCR 0.1 0.2 0.4 0.0
2 SCR bypassed 5.8 2.7 10.2 1.4
2 SCR bypassed 0.7 0.5 1.5 0.0
4 No SCR 7.5 1.0 9.2 5.9
4 No SCR 1.0 0.4 1.7 0.3
a Cl = Confidence interval
EERC CW2218B. CDR
100-
-
90-
-
80-
D) 70-
T3 60-
0)
N
S 5°-
X
0 40-
30-
20-
10-
0 -
^^{^••igja^j^^M|^B^^aMf|lj^.g^gP^^Sj|Br) nfcr*) f) (O ^.
*~^^ * "I**1 ^Wf^^er^^1^^ ®- o
« 0Q
o




























9/22 9/26 9/30 10/04 10/08 10/12 10/16 10/19
Calendar Time
Figure 5-11
Average Hg2+ as Measured by Hg SCEMs (total Hg - Hg°) for Site S6 Unit 1 (SCR on-
line)
                                                                              5-17

-------
Site S6
             100 -

             95 -

             90
          85 -
       D>
      x  so ^
      T3
       N  75 -
      T3
      §  TO H
      vP
      S^  65 -

          60 -

          55 -
             50
              09/22
                                                                EERC CW22273.CDR
   Figure 5-12
   Average He
   bypassed)
                                SCR bypassed
                             O
                               O
                                                  O
                                                         O
                                                         O
                         09/29          10/06          10/13
                                    Calendar Time
10/20
Average Hg2+ as Measured by Hg SCEMs (total Hg - Hg°) for Site S6 Unit 2 (SCR
                                                                    EERC DL21840 CDR
         0.0
           124
   Figure 5-13
   Hg SCEM Results for Site S6 for Unit 2 with SCR Bypassed
                                                                          136
5-18

-------
                                                                                Site S6

5.3.4  Coal Analysis Results

Nine of the coal samples from S6 were analyzed for Hg and chlorides (presented in Table B-
10 in Appendix B). The Hg concentrations were very consistent, averaging 0.066 ± 0.009
|ig/g. However, the chloride content of the coal was somewhat more variable, averaging
1020 ± 300 ppm. These averages are based on all coal samples analyzed regardless of the
unit they were collected from. Plant personnel said that the same coal was fired in all three
units.

Additional analyses were conducted on selected coal samples. Results of proximate and
ultimate analyses (including the Hg and chloride concentrations for that sample) are provided
in Table 5-6. In general, it appears that coal composition from all three units was very
consistent.
    Table 5-6
    Coal Analysis for Site S6a
Parameter
Mercury
Chlorine
Units
ppm (dry)
ppm (dry)
9/24/2002
Unit 1 (SCR) Unit
0.084
1210
9/24/2002
2 (SCR bypassed)
0.052
1520
10/8/2002
Composite15
0.063
1170
10/8/2002
Unit 4 (no SCR)
0.070
1320
Proximate Analysis
Moisture
Volatile Matter
Fixed Carbon
Ash
wt%
wt%
wt%
wt%

Hydrogen
Carbon
Nitrogen
Sulfur
Oxygen
Heating Value
Fd Factor0
wt%
wt%
wt%
wt%
wt%
Btu/lb
dscf/106Btu
5.6
33.1
49.1
12.2
Ultimate
5.2
70.7
1.5
0.9
9.5
11,936
10,357
5.9
34.5
49.1
10.5
Analysis
5.3
71.0
1.7
0.8
10.8
12,142
10,181
6.2
34.8
47.0
12.0

5.3
67.8
1.6
1.2
12.1
12,159
9727
6.5
34.7
47.3
11.5

5.3
68.5
1.7
1.1
12.0
11,837
10,085
a Except where noted, all results are on an as-received basis.
b Composite of all three units.
0 As defined by EPA Method 19.
                                                                                  5-19

-------
Site S6
5.3.5  ESP Ash Analysis

ESP hopper ash samples were collected daily throughout the test period from each unit. ESP
configuration at Unit 1 (SCR) and Unit 2 (SCR bypassed) consisted of two ESPs in series.
Ash samples were obtained from the first hoppers of each ESP (AB and CD) and analyzed
for Hg and LOT. Unit 4 (no SCR) had only a single ESP.  The ESP hopper ash Hg and LOT
averages are presented in Table 5-7. Plant personnel had  indicated that ash characteristics
were substantially different between the first and second  ESP, specifically, the presence of
higher amounts of unburned carbon in the first ESP.
    Table 5-7
    Analysis of ESP Hopper Ash
Unit
Unit 1 (SCR)
Unit 2 (SCR
bypassed)
Unit 4 (no
SCR)

Ash Hg,
M9/9
0.073 ±
0.014
0.152 ±
0.068
0.058 ±
0.017
ESP
No.
Samples
22
20
14
AB
LOI,
3.4
4.8
4.0

No.
Samples
4
5
2

Ash Hg,
M9/9
0.066 ±
0.027
0.118±
0.039
—
ESP CD
No. LOI,
Samples %
28 3.4
25 4.5
— —

No.
Samples
4
5
—
Results from our limited LOI analysis did not indicate a significant difference between the
samples.


5.3.6 NH3 Slip and SO3 Flue Gas Results

NH3 slip samples were collected at the SCR outlet. A summary of these results is provided in
Table 5-8. The NHs slip was less than 0.2 ppm for both samples, indicating an efficiently
operating SCR unit. This is also illustrated by more than 90% NOx removal efficiency
calculated from plant operational data.

SOs testing was conducted at the SCR outlet and ESP inlet of Unit  1 (SCR) and at the air
heater inlet and ESP inlet on Unit 2 (SCR bypassed). A summary of these results is also
provided in Table 5-8.  These data are consistent with what would be expected from a low-
sulfur eastern bituminous coal. It also appeared that (with the exception of the data taken on
October 11, 2002) there was some SOs condensation on the fly ash and possibly deposition in
the air heater. Comparing the data of Units  1 (SCR) and 2  (SCR bypassed) indicates that
there was some conversion of SO2 to SOs across the SCR catalyst.
5-20

-------
                                                                                   Site S6
    Table 5-8
    Flue Gas SCh and NH^ Results for Site S6a
Date
SO3, ppm
SCR Outlet ESP Inlet
NH3Slip, ppm
SCR Outlet
Unit 1 (SCR)
9/23/2002
9/24/2002
13.21
14.07
4.02
4.19
0.11
0.17
                  Unit 2 (SCR bypassed)
10/11/2002
10/14/2002
5.76
8.17
7.68
2.51
a Dry and 3% oxygen.
5.4    Mercury Mass Balance

Average Hg concentration in the coal and Fd factors (Table 5-6) were used to estimate the Hg
emission rate at the various sample locations. For the Hg associated with the ESP hopper ash,
the Hg concentrations were based on the dust-loading measurements as well as the Fd factor.
The results are shown in Table 5-9.
   Table 5-9
   Average Mercury Emission Factors for Site S6
Unit Tested
1 (SCR)
2 (SCR
bypassed)
3 (no SCR)

Coal
5.5
5.5
5.5
lbHg/10
ESP Inlet
6.1
7.3
—
12Btu
ESP
Hopper Ash
0.40
0.93
0.31C

Stack
7.6
5.5
4.5
%
Balance3
Based on
Coal Hg
111
87
—
Balance13
Across ESP
131
88
—
a Calculated balance is based on the coal Hg concentration and the ESP inlet [(Hgcoai - HgEspiniet)/HgCoai].
b Calculated balance is based on the ESP inlet Hg concentration and the stack [(Hgstack + HgEsphopPer)/HgEspiniet].
0 The ESP inlet was not measured; therefore, the inlet dust loading used was that obtained from Unit 2 (SCR bypassed).

As shown in Table 5-9, it appears that the Hg concentration as measured in the coal is low
compared to the flue gas measurements. The variability of the coal Hg was very low with a
relative standard deviation of less than 15%. The variability of the flue gas  data also was low.
The relative standard deviations ranged from 11% to 22%.
                                                                                    5-21

-------
Site S6

5.5   General Observations from S6

•  There was Hg oxidation across the SCR catalyst as the percentage of Hg2+ in the flue gas
   increased from 60% to 82%. At the ESP inlet and outlet location, the percentage was
   87%.

•  Comparing the Hg speciation results (at the ESP inlet location) with and without SCR in
   service showed that the presence of SCR resulted in greater Hg2+: 69% for Unit 2 (SCR
   bypassed) and 87% for Unit 1  (SCR).

•  Hg SCEM data gathered during the time  SCR was bypassed illustrate a small but
   measurable increase in Hg° from approximately 0.1 to 0.75 |ig/m3.

•  There was little, if any, Hg removal across the ESP.
5-22

-------
 6
DISCUSSION OF OVERALL RESULTS
The primary goal of this program has been to evaluate the effect of SCR operation on Hg
speciation with a focus on quantifying the fate of Hg across various pollution control devices.
A total of six plants have been tested over the course of 2 years with two of these plants
being tested in both 2001 and 2002, providing a total of eight data sets from which to
evaluate the effects of SCR operation. A summary of the plant configurations and the type of
coal combusted is provided in Table 6-1. A summary of the coal fired at each facility is
shown in Table 6-2.

It should be noted that additional data had been generated in 2001 for facilities that utilized
either SNCR or flue gas conditioning technologies; however, for the purposes of this report,
only SCR facilities are being addressed. Results from SNCR and flue gas conditioning plants
can be found in the Power Plant Evaluation of the Effect of Selective Catalytic Reduction in
Mercury [11].

As has been stated earlier, the use of SCR to reduce NOx emissions has the potential to
improve the Hg control efficiency of existing particulate  removal and FGD systems by
promoting Hg2+ or particulate-bound Hg formation. As data were compiled at the various
facilities, several factors were identified which may potentially impact the oxidation potential
of SCR. Among these factors, coal type, catalyst type and structure, and catalyst age were
specifically identified as factors that have the potential to influence Hg speciation.

To evaluate the effect of SCR on Hg speciation and, ultimately, on Hg emission at each
plant, the following were determined:

•  The change in Hg oxidation across the  SCR unit.

•  The effect of SCR on Hg oxidation obtained by comparing results with and without SCR
   in service at the particulate  control inlet or outlet.

•  The overall Hg removal with and without SCR.
                                                                              6-1

-------
Oi
K>
Table 6-1
Summary of SCR Program Plant Configuration
Plant
S1a

S2

S3

S4

S5
S6

S2-2e

S4-2e

Coal
PRB subbit.

OH bit.

PA bit.

KY bit.

WVbit.
Low-sulfur KY
and WV bit.
OH bit.

KY bit.

Boiler Type
Cyclone

Wall-fired

Tangential-
fired
Cyclone

Wall-fired
Concentric-
fired
Wall-fired

Cyclone

™er Low-NOx
MW BumerS
650 No

1 360 Yes

750 Yes, with
overfire air
704 No

684 Yes
700 Yes

1 350 Yes

650 No

Catalyst Vendor and
Type
Cormetech honeycomb

Siemens/Westinghouse
plate
KWH honeycomb

Cormetech honeycomb

Halder-Topsoe plate
Cormetech honeycomb

Siemens/Westinghouse
plate
Cormetech honeycomb

Catalyst SJ;R Space
. Velocity,
A9e nr-i
2 ozone 1800
seasons
3 months 2125

1 ozone 3930
season
1 ozone 2275
season
3 months 3700
2 ozone 3800
seasons01
2 ozone 2125
seasons
2 ozone 2275
seasons
Particulate
Control
ESP

ESP

ESP

Venturi
scrubber
ESP
ESP

ESP

Venturi
scrubber
b
1
i
o
Sulfur 3
Control s:
8
None =p

Wet FGDb

None

Venturi
scrubber0
Wet FGDb
None

Wet FGDb

Venturi
scrubber
a Not discussed in detail in this report.
b Magnesium-enhanced lime FGD.
0 Combined particulate  and SO2 limestone venturi scrubber.
d One layer of catalyst was replaced after one ozone season.
e Plant was retested in 2002.

-------
                                                              Discussion of Overall Results

   Table 6-2
   Average Analysis of Coals Fired During 2001 and 2002 Field Tests
a
Parameter
Hg, |jg/g dry
Cl, |jg/g dry
Moisture, %
Ash, %
Sulfur, %
HV, Btu/lb
S1
0.10
<60
27.5
3.7
0.2
8977
S2
0.17
1330
7.6
11.7
3.9
11,092
S2-2
0.14
635
6.1
9.4
3.9
12,097
S3
0.40
1150
7.0
14.0
1.7
11,421
S4
0.13
360
10.5
9.1
2.9
11,341
S4-2
0.16
260
11.8
8.5
2.8
1 1 ,634
S5
0.13
470
4.6
12.1
3.6
12,120
S6
0.073
1020
6.1
11.6
1.0
12,019
a As-received unless otherwise noted.


6.1    The Change  in Mercury Oxidation Across the SCR Catalysts

The percentage of Hg2+ was measured at both the inlet and outlet of the SCR unit at each
facility. It should be noted that all of the OH samples taken at these two locations were prior
to the air preheater; therefore, the temperature ranged from 640 to 700 °F (338-371 °C).
Table 6-3 presents the results of the 2001 and 2002 testing. For all of the plants tested, there
was an increase in Hg oxidation across the SCR catalyst. However, the amount of oxidation
that occurs across the catalyst is highly variable.  Some factors that may affect the level of
oxidation are coal, catalyst chemistry and structure, and catalyst age.

There was substantial variability in the percentage of Hg2+ at both the SCR inlet and outlet
locations. Site  SI fired a PRB coal and had a very high level of LOT, and as such, it is not
surprising it would be different, but there was also variability among the other sites firing
eastern bituminous coal. For example, repeat testing conducted at Site S4 indicated a
substantial increase in the percentage of Hg2+ when the coal chloride concentration increased
from 2001 to 2002 testing. As shown in Figure 6-1, one factor that appears to relate to the
percentage of Hg2+ at the inlet to SCR unit is the chloride concentration in the coal. It appears
there is a threshold chloride concentration at about 300 and 500 ppm chloride above which
40%-60% Hg  oxidation results at the SCR inlet.  What effect this has on overall Hg oxidation
is unclear.

Once the flue gas enters the SCR unit, it would be expected that other factors such as catalyst
type, structure, and space velocity could impact Hg oxidation. Without substantially more
data, it is very  difficult to determine the effects of these parameters. For example, Sites S2
and S4 had "low" space velocities (less than 2300 hr"1); Sites S3, S5, and S6 had "high"
space velocities (more than 3700 hr"1), but there  does not appear to be a clear correlation.
However, as shown in Table 6-1, the catalyst types and structures were also different. An
attempt was made to evaluate catalyst aging effects by retesting two plants in 2002 that had
been sampled in 2001: Sites S2 and S4. The results are discussed in Section 6.3. EPRI is
                                                                                  6-3

-------
Discussion of Overall Results


    Table 6-3
    Change in Mercury Oxidation Across the SCR Catalyst
Site
S1b
S2
S2
S3
S4
S4
S5
S6
Year Sampled
2001
2001
2002
2001
2001
2002
2002
2002
SCR Inlet Hg2+,
% of total Hg
8
48
54
55
9
33
43
60
SCR Outlet Hg2+,
% of total Hg
18
91
87
65
80
63
76
82
Percentage Point
Increase3
10
43
33
10
71
30
33
22
1 Percentage point increase is defined as (SCR Outlet % - SCR Inlet'
' Site S1 fired a PRB coal; the others were eastern bituminous coals.
0 1001
c/)
0
£2
o 80-
o
CL
co 60-
O
0
_ ^
.?=
? 20-
0
0
O)
.2
0 °
y c
EERC DL2I839 COR




S6
S2 - 2002 •
^ 	 __y — . 	 	 	 ~~g •
85 * S4-2002 S2-2001 *
S1 S4-2001

) 200 400 600 800 1000 1200 1400 16
          a.
Average Chloride Cone, in the Coal (dry), (JQ/9
    Figure 6-1
    Percent of Oxidized Hg2+ at the Inlet of the SCR System as a Function of Chloride
    Content of the Coal (note: nonlabeled data points are results from plants without SCR
    units where Hg speciation was measured at the air heater inlet)
6-4

-------
                                                             Discussion of Overall Results

currently in the process of trying to develop models that would predict the effects of the SCR
catalysts based on catalyst properties


6.2    Effect of the SCR on Mercury Oxidation

Although there is strong evidence that an SCR catalyst does promote Hg oxidation, to
determine the overall effect of SCR, it was useful to conduct tests both with and without SCR
in service at each site. For two of the sites (S3 and S4), testing was done on the same unit
with SCR bypassed at the end of the ozone season. For one, S5, sampling was done at two
similar units with only one having an SCR unit. At the final site, S6, two units were tested,
and both had SCR units, but the SCR unit on one was bypassed. Figure 6-2 shows the
comparison for all of the sites firing eastern bituminous coal. Based on Figure 6-2, it appears
(with the possible exception of Site S3) there is increased Hg oxidation as a result of SCR
based on measurements made at the inlet to the particulate control device. Table 6-4
quantifies the change in Hg2+.
0)
CO
0
0)
-=• 80-
c
o 60-
CD
5
"cd
c 40-
0
E
CO
CD
0 20-
z
&"-•

T
1
%
\

T T
X


T
1
T
!





i

r^^l Inlet to Part. Control Device - No SCR
17771 Inlet to Part. Control Device - With SCR

pt-j
!

                                                Site
   Figure 6-2
   Mercury Speciation Results Comparing All the Sites Tested Firing Eastern Bituminous
   Coal
                                                                                6-5

-------
Discussion of Overall Results
Table 6-4
Net Change in Hg2
                     ' as Measured at the Inlet to the Particulate Control Device
Site
S2-01
S2-02
S3
S4-01
S4-02
S5
S6
With SCR,
%
97
97
67
87
96
95
87
Without SCR,
%
73
b
77
56
57
80
69
Percentage Point Increase3
24
—
-10C
31
39
15
18
  Percentage point increase is defined as (SCR Outlet % - SCR Inlet %).
  OH samples were not taken in 2002; however, based on Hg SCEM data, it did not appear the change was as great in 2002 as
  2001 . The Hg SCEM result showed the Hg° concentration only increasing from Oto 0.8 ug/m3 at the ESP inlet locations (with and
  without the SCR in service). In 2001 , the Hg° increased from 0.4 to 3.4 ug/m3 when the SCR was bypassed.
  If the particulate-bound Hg is included, the results are 81 % nonelemental Hg both with and without the SCR.
6.3    Effect of SCR Catalyst Age on Mercury Speciation

Flue gas monitoring was conducted over 2 consecutive years at two power plants to evaluate
the impact catalyst age had on Hg speciation. Sites S2 and S4 were tested in 2001 and again
in 2002 to determine if the oxidation potential of an SCR catalyst was reduced with time,
specifically after one additional season of operation.

Testing in 2001  at Site S2 was conducted after approximately 3.5 months of catalyst age and
about 5 months at Site S4. 2002 testing was conducted after an additional ozone season,
approximately 5 months of catalyst age. Figures 6-3 and 6-4 show there was a decrease in Hg
oxidation across the SCR catalyst in 2002 compared to 2001 for both sites retested. However,
there were mitigating circumstances at each plant. At Site S2, in an attempt to control  SOs
emissions, humidification and alkali injection were done upstream of the SCR unit. As a
result of humidification, the temperature of the SCR unit was cooler by about 10 °F in 2002
compared to 2001. In addition, the coal fired  at S2 was from a different mine than that used
in 2001. At Site S4, the coal chloride concentration was extremely variable in 2001.

Although there were differences in the oxidation across the SCR catalyst, at the inlet to the
particulate control device there was no significant difference between the 2 years for either
Site  S2 or S4. Although there are suggestions that Hg oxidation may have  decreased after an
additional ozone season, the results are considered to be inconclusive. Additional testing is
being planned. Hg speciation sampling is recommended at these two plants for several more
years.
6-6

-------
                                                                Discussion of Overall Results
             Note: Error bars represent standard deviation of total Hg.

            30.0
               ££ftC CW20933.CDR
                    i    i Particulate-Bound Hg
                    C°°~i Oxidized Hg
                    ^H Elemental Hg
                     SCR SCR ESP  ESP Stack
                     Inlet Outlet Inlet  Outlet
                    2002 Sampling Results
SCR  SCR  ESP  ESP Stack
Inlet Outlet Inlet Outlet
2001 Sampling Results
Figure 6-3
Comparison of Mercury Speciation Results from 2001 and 2002 at Site S2
          20,0
                                                                    EERC DDKS! 3 CDR
                      Particulate-Bound Hg
                      Oxidized Hg
                      Elemental Hg
          15.0 -
                   SCR   SCR    AH   Stack
                   Inlet  Outlet  Outlet

                         2002 Results
  SCR    SCR    AH   Stack
  Inlet   Outlet  Outlet

         2001 Results
Figure 6-4
Comparison of Mercury Speciation Results from 2001 and 2002 at Site S4
                                                                                    6-7

-------
Discussion of Overall Results
6.4    SCR/Wet FGD Combination for Mercury Control

The underlying intent of understanding Hg oxidation via SCR technology is to determine its
potential to improve the Hg collection efficiency of existing ESPs, fabric filters and, in
particular, FGD systems. In general, wet FGD systems remove more than 90% of Hg2+.
However, there has been evidence that some of the captured Hg2+ can be reduced in a wet
FGD system to Hg° [10]. Three sites have been tested in 2001 and 2002 that have wet FGD
systems. Sites S2 and S5 employ magnesium-enhanced lime FGDs, and Site S4 is a
combined particulate/SC>2 venturi/spray tower scrubber. Sites S2 and S4 were tested in 2001
and 2002, thus yielding a database of five measurements at three sites. It is important to note
that approximately 60% of wet FGD systems in the United States are limestone forced-
oxidation systems. As can be seen in Table 6-5, there is a measurable increase in Hg° across
the FGD unit at all of the sampling sites when SCR was not in service. For the tests with
SCR in service, the five data points show an increase in Hg°, but the increase appears to be
very small and is generally within the variability of the data.
   Table 6-5
   Effect of the SCR on Hg° Concentration Across the Wet FGDs
Site
Year
Sampled

S2
S2
S4
S4
S5
2001
2002
2001
2002
2002

S2
S4
S4
S5
2001
2001
2002
2002
FGD Inlet Hg° Cone.,
ug/Nm3

0.4b
0.3
0.5
1.0
0.7

3.4b
5.6
5.7
4.7
FGD Outlet Hq° Cone.,
ug/Nm
With SCR
0.9
1.3
0.8
1.3
1.0
Without SCR
5.0
7.1
8.0
6.1
Change,3
ug/Nm

0.5
1.0
0.3
0.3
0.3

1.6
1.5
2.3
1.4
Total Percent Hg
Removed

89
84
90
91
91

51
46
44
51
a Change is defined as: (FGD outlet Hg° cone. - FGD inlet Hg° cone.).
b For 2001 Site S2 data, the ESP inlet data were used because the FGD inlet Hg concentrations were clear outliers.

The mechanism for FGD reemission is not well understood, but it is speculated that sulfite in
the FGD slurry may reduce Hg2+ to Hg°.  The impact of forced oxidation may alter the sulfite
chemistry, potentially giving different results than those obtained for the plants shown in
Table 6-5. Because the mechanism of reemission is not well understood and it is not known


6-8

-------
                                                              Discussion of Overall Results

how SCR units may impact reemission, the reader is cautioned in attempting to extrapolate
the results from these three sites to all FGD systems. Additional studies are recommended
and planned at plants with limestone forced-oxidation FGD systems.
                                                                                 6-9

-------
 7
CONCLUSIONS
The primary conclusions based on the test results are:

•  For plants firing eastern bituminous coals, Hg oxidization occurs across SCR catalysts.
   However, it appears to be variable and most likely related to a variety of factors. Some
   potential factors are coal characteristics, catalyst chemistry, catalyst type and structure,
   space velocity, and catalyst age.

•  It appears that addition of an SCR unit will provide additional Hg2+when an eastern
   bituminous coal is fired. With the exception of Site S3 (where the Hg was essentially all
   H22+ or Hgp, both with and without SCR), all facilities showed increased oxidation at the
   inlet to the particulate control device.  The increase ranged from 15 to 39 percentage
   points.

•  At both sites where sampling was done in 2001 and 2002, there appeared to be a decrease
   in Hg oxidation across the SCR catalyst between the first and second ozone seasons of
   operation. However, at both facilities, there were other possible explanations related to
   changes in plant operation. These changes do not allow a definitive conclusion to be
   developed on the  effect of an additional ozone season on SCR/Hg oxidation. It is
   important to note  that the Hg oxidation at the inlet to the particulate control device was
   not affected by the additional ozone season.

•  Based on the limited data at three plants (five total data sets), it appears there is some
   reemission of the  captured Hg across wet FGDs. For the tests with SCR in  service, the
   increase appears to be very small and  is generally within the variability of the data.
   Nevertheless, five data points show an increase in Hg°. It appears that the reemission is
   more pronounced when an SCR unit is not present.


Future Test Plans

Based on a review of these test results, several areas are recommended for further
investigation. DOE, EPA, and EPRI are planning to conduct additional full-scale, as well as
bench- and pilot-scale studies, to address the following:

•  The effect of SCR when PRB coal is fired in a pc-fired boiler

•  Further testing to  determine the effect of SCR units on Hg capture in wet FGD systems,
   in particular on Hg reemission

•  The effect of SCR when PRB/bituminous-blended coal is fired
•  Further evaluation of the effect of catalyst age on Hg speciation
                                                                                 7-1

-------
8
QUALITY ASSURANCE/QUALITY CONTROL
The EERC is committed to delivering consistent and high-quality research that exceeds its
clients' needs and expectations. To ensure that the goals of this project are realized, an
organizationwide quality management system (QMS), authorized and supported by EERC
managers, is in effect and governs all programs within the organization. The EERC
established and formalized a QMS and QC procedures in August 1988. The Quality Manual
defines the requirements and the organizational responsibilities for each major element of the
QMS and references the supporting documents  needed to provide a comprehensive program.
Compliance with this manual and its supporting documents ensures that the EERC
adequately fulfills governmental and private client requirements relating to quality and
compliance with applicable regulations, codes,  and protocols. This project was required to
follow the Quality Manual, project-specific quality assurance (QA) procedures, and all
revisions. The EERC Quality Assurance Manager implements and  oversees all aspects of
QA/QC for all research, development, and demonstration projects and reviewed the QA/QC
components of this project. The project manager is responsible for  ensuring that project-
specific QA/QC protocols are followed.

To ascertain data quality obtained during the sampling program, the following procedures
were used:

•  Process operating data were examined to ensure that the OH sampling took place during
   steady, representative plant operation.

•  Sampling and analytical analysis protocols were reviewed to ascertain how the data
   compared with other data generated using standard protocols.

•  The reagent blanks, field blanks, and field spikes were reviewed to qualitatively
   determine the confidence that can be placed in the results.

•  The QA/QC data results were then compared with data quality  indicators to qualitatively
   determine the validity of the data in terms of variability and accuracy.


8.1    Process Data Evaluation

Plant operating data were examined to ensure that process operation was stable and
representative during the OH sampling periods. Excessive scatter or significant trends in
relevant process variables can indicate periods of unrepresentative  unit operation. Data
scatter is useful for identifying periods of operational difficulty; data trends indicate periods
when steady-state operation has not been achieved. It was the intent for the Hg SCEMs to be
operated both during steady-state conditions and during any upset conditions that occurred.

-------
Quality Assurance/Quality Control

Plant data, to the extent available, were plotted for each of the test sites. In general, it appears
that all of the OH sampling occurred either when the unit was at or near full-load conditions.
When plant operational upsets occurred during OH sampling, sampling was suspended, and a
new sample was taken after the plant was operating at more normal conditions. This  occurred
at Site S5 and is illustrated by the greater quantity of OH results.


8.2    Sampling Quality Control Evaluation

Sampling precision can be estimated by comparing the results of various parameters  of
replicate samples, notably, velocity, moisture content, and gas composition in the stack.
Sampling accuracy is usually inferred from the calibration and proper operation of the
equipment and from historical validation of the methods. Field blanks are used to determine
any biases that may be caused by contamination or operator errors. A field blank is defined
as a complete impinger train, including all glassware and solutions, which is taken out to the
field during sampling and exposed to ambient conditions. These sample trains are then taken
apart and the solutions recovered and analyzed in the  same manner as those sample trains
used for sampling activities. If the field blank shows contamination above instrument
background, steps are taken to eliminate or reduce the contamination to below background
levels. The results of the blanks can be seen in Appendix B (Tables B-l, B-3, B-5). In almost
all cases, the field blank results were less than detection limits. For the few samples where a
detectable level of Hg was measured, the concentration was low enough to be insignificant
compared to the measured flue gas concentration for that Hg species.

Sampling comparability depends both on whether the samples are representative and on the
use of standard methods consistently applied. All methods used for the project were standard
American Society for Testing and Materials or EPA sampling methods. Sampling
completeness is primarily a function of providing the  requisite number of samples to the
analytical laboratory. In most cases, this consisted of duplicate samples.

The isokinetic  sampling rate is a measure of the operational performance of sampling for
particulate matter. The normal acceptance criterion for isokinetic variation is 10%. With over
90 OH samples taken during this project, five samples were outside the ±10% range. Four
samples were collected at the stack and, based on the very low particulate loading and Hg
concentration,  appeared to have no significant impact on the results. One sample, collected at
the ESP outlet location, had an isokinetic measurement of greater than 100%. Again, the Hg
results from this sample were not significantly affected because of the extremely low
concentration of Hg and low particulate loading. A lower-than-expected isokinetic sampling
rate results in an overestimation of the larger particles, resulting in an inflated dust-loading
estimate. However, for these samples, the dust loading and Hg concentrations were very
similar to the other samples taken. It is believed that this deviation from the accepted
isokinetic value had no significant impact on the overall conclusions.

One known concern with the OH method is a bias that occurs as a result of the  close  contact
between the flue gas and the fly ash collected on the sampling filter. This  is particularly true
at high-dust sampling locations such as SCR inlet, outlet, and the inlet to the particulate
control device. The degree of bias is dependent on the reactivity of the ash collected  on the

8-2

-------
                                                          Quality Assurance/Quality Control

filter and the flue gas temperature. This ash has the potential to adsorb or oxidize Hg. The
only method of determining the extent to which the bias occurs is to compare the inlet and
outlet ESP Hg° results. If there is an increase in Hg° concentration across the ESP, it
indicates some oxidization occurred across the sampling filter. To determine if there was
adsorption of Hg on the sample,  resulting in a high particulate-bound Hg bias, the filter
concentration is compared to the ESP hopper ash samples. Although representative ash
samples are extremely difficult to collect from an ESP, it is possible to obtain an indication
as to whether the filter is biasing the particulate-bound Hg concentration. These comparisons
were made, and the results from  each facility are detailed in the discussion in Sections 2-7.


8.3    Evaluation of Measurement Data Quality

An evaluation of the measurement data quality is based on QC data obtained during sampling
and analysis. Generally, the type of QC information obtained pertains to measurement
precision, accuracy, and blank effects, determined by collecting various types of replicate,
spiked, and blank samples. The specific characteristics evaluated depend on the type of QC
checks performed. For example,  if problems with contamination occur, blank samples can  be
prepared at different stages in the sampling and analysis process to isolate the source of a
blank effect. Similarly, replicate  samples may be generated at different stages to isolate and
measure the sources of variability. Table 8-1 summarizes the QA/QC measures used and the
characteristic information obtained for this project.

As shown in Table 8-1, different QC checks provide different types of information,
particularly pertaining to the sources of inaccuracy, imprecision, and blank effects. In
general, measurement precision and accuracy are typically estimated from QC indicators that
cover as much of the total sampling and analytical process as feasible. Precision  and
accuracy estimates are based primarily on the actual sample media documenting the precision
and accuracy actually obtained, and the objectives serve as benchmarks for comparison. The
effects of not meeting the objectives need to be considered in light of the intended use of the
data. The results of the field and  media spikes that were done as part of this project are
shown in Appendix B (Tables B-2, B-4 and B-6). As can be seen in these tables, the spike
recovery was excellent for field blanks completed. Although blank filters are routinely
analyzed for Hg to ensure no Hg contamination on the sample, no field filter spikes were
completed for the project. However, in the laboratory, known Hg calibration standards are
routinely analyzed.

Other specific QC procedures that were used to measure Hg in the flue gas for this project
are as follows:

•  Instrument Setup and Calibration. The instrument used in the field for Hg
   determination was a Leeman Labs PS200 cold-vapor atomic absorption spectrometer. To
   measure Hg, the instrument was set up for absorption at 253.7 nm with a carrier gas of
   nitrogen and 10% w/v stannous chloride in 10% v/v HC1 as the reductant. Each day, the
   drying tube and acetate trap were replaced and the tubing checked. The rinse container
   was cleaned and filled with fresh solution of 10% v/v HC1. After the pump and lamp were
   turned on and warmed up for 45 minutes, the aperture was set to manufacturer

                                                                                 8-3

-------
  Quality Assurance/Quality Control

     Table 8-1
     Elements of the QA/QC Plan
             QC Activity                              Characteristic Measured

                                         Precision

Replicate Samples Collected over Time     Total variability, including process or temporal, sampling,
under the Same Conditions                and analytical but not bias.

Duplicate Field Samples Collected          Sampling plus analytical variability at the actual sample
Simultaneously                          concentrations.

Duplicate Analyses of a Single Sample      Analytical variability at the actual sample concentrations.

Media-Spiked Duplicates                 Sampling plus analytical variability at an established
                                       concentration.

Laboratory Control Sample Duplicates      Analytical variability in the absence of sample matrix
                                       effects.

                           Accuracy (including precision and bias)

Media-Spiked Samples                   Analyte recovery in the sample media, indicating possible
                                       interferences and other effects. In a single sample, includes
                                       both random error (imprecision) and systematic error (bias).

Laboratory Control Samples               Analyte recovery in the absence of actual sample matrix
                                       effects. Used as an indicator of analytical control.

                                       Blank Effects

Field Blank                             Total sampling plus analytical blank effect, including
                                       sampling equipment and reagents, sample transport and
                                       storage, and analytical reagents and equipment.

Reagent Blank                          Blank effects from reagents used.

     specifications. A four-point calibration curve was then completed using matrix-matched
     standards. The detector response for the given standard was then logged and compared to
     specifications to ensure the instrument had been properly set up. A QC standard of a
     known analyte concentration was analyzed immediately after the instrument was
     standardized to verify the calibration. This  QC standard is prepared from a different stock
     than the calibration standards. It was required that the values obtained read within 5% of
     the true value before the instrument was used. After the initial QC standardizations were
     completed, standards were run every five samples to check the slope of the calibration
     curve. All samples were run in duplicate, and one in every ten samples was spiked to
     verify analyte recovery. A QC chart is maintained at the EERC to monitor the long-term
     precision of the instrument.  The results of these calibrations are available upon request of
     any EERC client.

  •  Presampling Preparation. All data sheets, volumetric flasks, and Petri dishes used for
     sample recovery were marked with preprinted labels. The liquid samples were recovered

  8-4

-------
                                                          Quality Assurance/Quality Control

   into premarked volumetric flasks and logged, then analyzed on-site. The outlet filter
   samples were placed in premarked Petri dishes and taken back to the EERC, where they
   were analyzed using mixed-acid digestion techniques. The labels contained identifying
   data, including date, time, run number, sample port location, and the name of the sampler.

•  Glassware and Plasticware Cleaning and Storage. All glass volumetric flasks and
   transfer pipettes used in the preparation of analytical reagents and calibration standards
   were designated Class A to meet federal specifications. Prior to being used for the
   sampling, all glassware was washed with hot, soapy water, then rinsed with deionized
   water three times, soaked in 10% v/v nitric acid for a minimum of 4 hr, rinsed an
   additional three times with deionized water, and dried. The glassware was then stored in
   closed containers until it was used at the plant. All glassware cleaning solutions are
   periodically checked for Hg. In all cases, the measured Hg concentration was below
   detection limits.

•  Analytical Reagents. All acids to be used for the analysis of Hg were trace metal-grade
   or analytical reagent-grade. The calibration standards used for instrument calibration  and
   the QC standards used for calibration verification were purchased commercially and
   certified to be accurate within 0.5% and were traceable to National Institute of Standards
   and Technology standard reference materials.


8.4    Ontario Hydro Method Error Analysis

The precision of Hg measurements is estimated to be in the order of +10%-30%, depending
upon the total Hg concentration, its proximity to the method detection limit and, possibly,
other flue gas constituents. In addition, flue gas sampling at internal streams such as the SCR
inlet and outlet and the ESP inlet are often single-point samples, may not be uniform
(stratified), and thus may not represent the true flue gas distribution of the power plant. These
uncertainties in the total Hg measurements must be taken into account when data are
interpreted. However, these uncertainties are not expected to have a significant effect on the
overall conclusions of this study.

For example, if the "actual" Hg concentrations at the ESP inlet and the stack are 10 and
1 Ib/trillion Btu, respectively, then the total Hg removal would be 90%. If the measurements
at both the inlet and outlet were assumed to have a +20% error, then the inlet concentration
could be measured at either 8 or 12 Ib/trillion Btu and the outlet concentration at either 1.2 or
0.8 Ib/trillion Btu. The estimated Hg removal could range from as low as 85% to as high  as
93%.

The calculated Hg removal is much more sensitive at sites where the Hg removal efficiencies
are very low. For example, if the "actual" Hg concentrations at the ESP inlet and the stack
were 10 and 9 Ib/trillion Btu, respectively, then the total Hg removal would be 10%. If the
measurements at both the inlet and outlet were assumed to have a +20% error, then the inlet
concentration could be measured at either 8 or 12 Ib/trillion Btu and the outlet concentration
at either 7.2 or 11.8 Ib/trillion Btu. Thus the calculated Hg removal could range from as low
as -48% to as high as 40%.
                                                                                  5-5

-------
9
REFERENCES
1.  U. S. Environmental Protection Agency. Mercury Study Report to Congress Volume I:
   Executive Summary, EPA-425/R-97/003 (NTIS PB98-124738), Office of Air Quality
   Planning and Standards and Office of Research and Development, Dec 1997.

2.  EPRI. An Assessment of Mercury Emissions from U.S. Coal-Fired Power Plants; EPRI
   Report No. 1000608, Oct 2000.

3.  U. S. Environmental Protection Agency. Study of Hazardous Air Pollutant Emissions
   from Electric Utility Steam Generating Units Final Report to Congress: Executive
   Summary; EPA-425/R-98/004A (NTIS PB98-131774), Office of Air Quality Planning
   and Standards and Office of Research and Development, Feb 1998.

4.  Provisions for Attainment and Maintenance of National Ambient Air Quality Standards.
   Public Law 101-549, 1990.

5.  Sloss, L.L. Mercury Emissions and Effects - The Role of Coal; IEAPER/19 International
   Energy Agency: Coal Research Perspectives; London, Aug 1995; p 39.

6.  ICR Reports, http://www.epa.gov/ttn/uatw/combust/utiltox/utoxpg.html (accessed Oct 7,
   2000).

7.  Schimmoller, B.K. SCR Dominates NOX Compliance Plans. Power Eng. 2000, July, 45-
   48.

8.  Galbreath, K.C.; Zygarlicke, C.J.; Olson, E.S.;  Pavlish, J.H.; Toman, D.L. Evaluating
   Mercury Transformation Mechanisms in a Laboratory-Scale Combustion System. Sci. of
   the  Total Environ. 2000, 261 (1-3), 149-155.

9.  Gutberlet, H.; Spiesberger, A.; Kastner, F.; Tembrink, J. Mercury in Bituminous Coal
   Furnaces with Flue Gas Cleaning Plants. VGB Kraftwerkstechnik 1992, 72, 586-591.

10. Gutberlet, H.; Schliiten, A.; Lienta, A.  SCR Impacts on Mercury Emissions on Coal-Fired
   Boilers.  Presented at EPRI SCR Workshop, Memphis, TN, April 2000.

11. EPRI, Palo Alto, CA, U.S. Department of Energy, Pittsburgh, PA, and U.S.
   Environmental Protection Agency, Research Triangle Park, NC. Power Plant Evaluation
   of the Effect of Selective Catalytic Reduction in Mercury, No. 1005400-2002.
                                                                             9-1

-------
References

12. Miller, SJ; Dunham, G.E.; Olson, E.S. Controlling Mechanisms That Determine Mercury
   Sorbent Effectiveness. Presented at the 92nd Annual Meeting & Exhibition of the Air &
   Waste Management Association, St. Louis, MO, June 1999, Paper No. 99-898,

13. Shashkov, V.I.; Mukhlenov, IP.; Beriyash, E.Y.; Ostanina, V.I.; Shokarev, M.M.;
   Vershinina, F.I. Effect of Mercury Vapors on the Oxidation of Sulfur Dioxide in a
   Fluidized Bed of Vanadium Catalyst. Khim. Prom. (Moscow) 1971, 47, 288-290.

14. Carey, T.R.; Skarupa, R.C.; Hargrove, O.W. Jr. Enhanced Control of Mercury and Other
   HAPs by Innovative Modifications to Wet FGD Processes; Phase I Report for U.S.
   Department of Energy Contract No. DE-AC22-95PC95260; Aug 28, 1988.

15. Hitchcock, H.L. Mercury Sorption on Metal Oxides. M.S. Thesis, University of North
   Dakota, Dec 1996; 83 p.

16. Galbreath, K.C.; Zygarlicke, C.J. Mercury Transformations in Coal Combustion Flue
   Gas. Fuel Process. Technol.  2000, 65-66, 289-310.

17. Ghorishi, S.B.; Lee, C.W.; Kilgroe, J.D. Mercury Speciation in Combustion Systems:
   Studies with Simulated Flue Gases and Model Fly Ashes. Presented at the 92nd Annual
   Meeting & Exhibition of the Air & Waste Management Association, St. Louis, MO, June
   1999.

18. Laudal, D.L; Pavlish, J.H.; Galbreath, K.C.; Thompson, J.S.; Weber, G.F.; Sondreal, E.A.
   Pilot-Scale Screening Evaluation of the Impact of Selective Catalytic Reduction for NOX
   on Mercury Speciation, EPRI Report No. 1000755, U.S. Department of Energy Contract
   No. DE-FC26-98FT40321; EERC Publication No. 2000-EERC-12-01, Energy  &
   Environmental Research Center: Grand Forks, ND, Dec 2000.
9-2

-------
A
SAMPLING METHODS AND PROCEDURES
This appendix provides the template for developing site-specific test plans and sampling
protocols.


Ontario Hydro Mercury Speciation Method (OH method)

This is a summary of the sampling and analytical procedures used to conduct the mercury
(Hg) speciation method entitled "Standard Test Method for Elemental, Oxidized, Particle-
Bound, and Total Mercury in Flue Gas Generated from Coal-Fired Stationary Sources
(Ontario Hydro Method)." The American Society for Testing and Materials D22 committee
has accepted the method, and the exact method details are provided on the U.S.
Environmental Protection Agency (EPA) Web page at
http://www.epa.gov/ttn/emc/prelim.html under Preliminary Method 3. All other EPA
methods are also found at the same emission measurement Web address.

The OH method follows standard EPA methods for isokinetic flue  gas sampling (EPA
Methods 1-3  and EPA Method 5/17). Figure A-l presents a schematic of the Hg speciation
sample train.
II Connect to
  \\  Filter Holder


    ili
 Connect to
Vacuum Inlet
                                                                  EERCKG13Wt.CDR
   Figure A-1
   Schematic of the OH Mercury Speciation Train
                                                                          A-l

-------
Sampling Methods and Procedures
Table A-l presents a list of sample train components for the OH configuration.
   Table A-1
   Sample Train Components for the Ontario Hydro Method
        Component
                           Details
Nozzle

Filter

Probe

Connector Line


Impingers 1  and 2

Impinger 3

Impinger4

Impingers 5 and 6

Impinger 7

ImpingerS
Glass, quartz, or Teflon-coated stainless steel

Quartz, in glass or Teflon-coated stainless steel holder

Glass or Teflon, heated to gas temperature

If needed, Teflon line used to connect from probe to impingers, heated
to a minimum of 248 °F (120 °C).

1 mol/L KCI solution; modified Smith Greenburg (SG) impinger

1 mol/L KCI solution; standard SG impinger

5% nitric acid/10% hydrogen peroxide; modified SG impinger

4% potassium permanganate/10% sulfuricacid; modified SG impinger

4% potassium permanganate/10% sulfuricacid; standard SG impinger

Silica gel; modified SG impinger
A sample is withdrawn from the flue gas stream isokinetically through the filtration system,
which is followed by a series of impingers in an ice bath. Particulate-bound Hg is collected
on the front half and filter; Hg2+ is collected in impingers containing 1 N potassium chloride
solution; and elemental Hg is collected in one impinger containing a 5% nitric acid and 10%
peroxide solution and three impingers  containing a solution of 10% sulfuric acid and 4%
potassium permanganate. An impinger containing silica gel collects any remaining moisture.
The filter media are quartz fiber filters. The filter holder is glass or Teflon-coated.  An
approximate 2-hr sampling time was used, with a target sample volume of 1-2.5 standard
cubic meters.

Figure A-2 is a schematic of the sample recovery procedure for the impinger train. The
samples were recovered into precleaned glass bottles with vented Teflon-lined lids. The
following sample fractions were recovered (specific rinse solutions are contained in the
method):

1. The sample filter

2. The front-half rinse (includes all surfaces upstream of the filter)

3. Impingers  1 through 3 (KCI impingers) and rinses

4. Impinger 4 (HNO3/H2O2 impinger) and rinses
A-2

-------
                                                          Sampling Methods and Procedures
5.  Impingers 5-7 (H2SO4/KMnO4 impingers) and rinses

      1. Rinse filter holder and connector with 0,1 N HNO3.
      2, Add 5% w/v KMn04 to each impinger bottle until
        purple color remains.
      3, Rinse with 10% v/v HNO3.
      4. Rinse with a very small amount of 10% w/v
        NH2OH-H2SO4 if brown residue remains.
      5. Final rinse with 10% % HN03.
                                   Rinse with 0.1 N HNCL
                                                    Rinse Bottles Sparingly with
                                                    - 0.1NHNO3
                                                    - 10%W/VNH2OH-H2SO4
                                                    - 0.1NHNO,
                                     HNO3/H2O2
                            Rinse All U-Tubes with 0.1 N HNO
                                                                    EERCDLW139.CDR
   Figure A-2
   Sample Recovery Scheme for the OH Mercury Speciation Train

6.  Impinger 8 (silica gel impinger [note that this sample is weighed for moisture
   determination and not included in the Hg analysis])

The sample fractions were prepared and analyzed as specified in the method and summarized
below:

•  Ash Sample (Containers 1 and 2) - The parti culate catch was analyzed using EPA
   Method 7043 or equivalent (see Table 3) or using a Milestone DMA-80 Hg analyzer.
   However, if the parti culate catch was less than 1 gram (as was the case at the outlet of the
   parti culate  control device), the entire sample of the parti culate collected on the filter
   (including the filter) was subsequently digested using EPA Method 3051, followed by
   analysis using EPA 7471 A.
KCl Impingers (Container 3) - The impingers were prepared using
KMnC>4 solutions as specified in the method.
                                                                     /t, HNOs, and
           2O2 (Container 4) - The impinger solutions were prepared using HC1 and
   KMnC>4 solutions as specified in the method.
                 Impingers (Container 5) - The impinger solutions were prepared using
   hydroxylamine hydrochloride as specified in the method.

Each prepared fraction was analyzed for total Hg by cold-vapor atomic absorption (CVAA).
CVAA is a method based on the absorption of radiation at 253.7 nm by Hg vapor. The Hg is
reduced to the elemental state and aerated from solution in a closed system. The Hg vapor
                                                                                A-3

-------
Sampling Methods and Procedures

passes through a cell positioned in the light path of an atomic absorption (AA) spectrometer.
Hg concentration is proportional to the indicated absorbance. A soda-lime trap and a
magnesium perchlorate trap must be used to precondition the gas before it enters the
absorption cell.


Continuous Mercury Monitors

Four different Hg semicontinuous emission monitors (Hg SCEMs) were used for these tests:
the Semtech Hg 2010, PS Analytical (PSA) Sir Galahad, Tekran, and OhioLumex. These
instruments, when used in conjunction with the Energy & Environmental Research Center
(EERC) or PSA conversion systems, with some caveats as explained in the report, were able
to measure speciated Hg. The instruments are briefly described below.


Atomic Fluorescence-Based Hg SCEMs

The PSA Sir Galahad and the Tekran are fluorescence-based instruments. The Sir Galahad
analyzer was initially used to monitor total Hg continuously in the urban environment and
natural gas. The Tekran analyzer was initially used to primarily monitor ambient Hg. As was
the case for this project, both of these instruments can be used in a variety of gaseous media
including combustion flue gas. These analyzers are based on the principle of atomic
fluorescence (AF), which provides an inherently more sensitive signal than AA. The systems
use a gold-impregnated silica support for preconcentrating the Hg and separating it from
potential interferences that degrade sensitivity.

These instruments require a four-step process to obtain a flue gas Hg measurement. In the
first step, conditioned flue gas is pumped through a gold trap, which is maintained at a
constant temperature. Before the Hg is desorbed from the gold trap, a flushing step is
initiated to remove any flue gas that may be present because it has a damping effect on the
Hg fluorescence. When this is completed, the analysis step begins. The heating coil is
activated, and the gold  trap is heated to desorb the Hg from the trap. The Hg is carried into
the fluorescence detector in an inert gas stream of argon or nitrogen, depending on the Hg
concentration. The gold trap is then cooled in preparation for the next sample. The time for
the entire process is about 5 min.

The systems are calibrated using Hg° as the primary standard. The Hg° is contained in a
closed vial, which is held in a thermostatic bath. The temperature of the Hg is monitored, and
the amount of Hg is measured using vapor pressure calculations. Typically, the calibration of
these units has proven to be stable over a 24-hr period.


Atomic Absorption-Based Hg SCEMs

Both the Semtech Hg analyzer (Semtech Metallurgy AB, Lund, Sweden) and the OhioLumex
instruments are portable Zeeman-modulated CVAA spectroscopes that can monitor Hg°
continuously. These analyzers use Zeeman effect background correction by applying a
modulated magnetic field to a Hg lamp to minimize interferences from the presence of SC>2,

A-4

-------
                                                        Sampling Methods and Procedures

moisture, hydrocarbons, and fine particulate in the flue gas sample. The primary difference
between the Semtech and the OhioLumex instruments is the AA path length. The Semtech
has a path length of about 0.5m compared to 9.7 m for the OhioLumex. The result is a much
lower detection limit for the OhioLumex. The operating range of the Semtech is listed as 0.3
to 160 mg/Nm3 Hg°; however, in practice, the lower limit of quantification is about 2
|ig/Nm3. The OhioLumex has the potential to measure as low as 0.1 ng/Nm3. It should be
noted that the Semtech Hg 2010 has also been certified by TUEV Rheinland for determining
compliance with the German legal limit of 50  ug/Nm3 for total Hg from waste incinerators.


Flue  Gas Pretreatment/Conversion

Whether the Hg SCEM uses CVAA or AF to measure Hg, some form of gas pretreatment is
necessary before accurate measurement of total Hg (or speciated Hg) is obtained. Figure A-3
illustrates the EERC pretreatment system used with Hg SCEMs. For the CVAA-type
systems, only Hg° can be directly analyzed. Therefore, all Hg forms in the flue gas must be
converted to Hg°. For this purpose, SnCb is used as a reductant. To use an Hg SCEM for Hg
speciation measurements, first only Hg° (bypassing the SnCb) is measured, followed by a
measurement of the total Hg by reducing the Hg2+ to Hg° with SnC^ prior to analysis. The
Hg2+ concentration was calculated by difference.
                                       Hg Absorption
                                          Solution
                                            I     Gas + Liquid
                                        Gas f
                                                    ixcess Hg
                                                  Absorption Solution
                                                                         t- To
                                                                         CEM
   Figure A-3
   Schematic of the EERC Pretreatment/Conversion System for Use with Hg SCEMs

For the AF Hg SCEMs, a pretreatment/conversion system is also needed, but for different
reasons. The first reason is to remove gaseous contaminants (HC1, SO3, etc.) from flue gas
prior to the gold trap, thus preventing the trap from becoming poisoned permanently. The
second reason is that both Hg2+ and Hg° collect on the trap; if the instrument is to be used to
                                                                               A-5

-------
Sampling Methods and Procedures

provide Hg speciation data, then the Hg2+ must be removed from the gas stream so that the
Hg° concentration can be measured. To do this, either a heated carbonate trap (the EERC
system) or a basic SnCb trap (PSA system) is used. For all the tests discussed in this report,
the PSA system was used.


Auxiliary Flue Gas Measurements

Auxiliary flue gas measurements were performed using a portable 62 analyzer (as described
below) and H^O by EPA Method 4 (condensation/gravimetric analysis). These measurements
were collected as integral parts of all Hg speciation tests at all locations.


02 Determination

C>2 is measured by a portable 62 analyzer using an electrochemical cell. The gas sample for
the portable analyzer is drawn through a tube inserted in the exit gas of the sample gas meter.
This provides direct analysis of the gas sampled for the Hg test. Care should be taken so that
the 62 sample tube is not inserted so far that it interferes with the meter orifice pressure
differential reading. Calibration procedures for the portable analyzer include the following:

•  At the beginning of each test condition, the instrument is calibrated on ambient air. As-
   found readings are then taken using zero gas and a mid-scale C>2 calibration gas (40%-
   60% of the span to be used to collect readings). If these as-found readings are within 2%
   of span (0.2% 62 if the 10% scale is used), the data are acceptable. If the readings are
   outside of these ranges, the Q^ cell should be replaced, the instrument should be repaired,
   or an alternate instrument should be used.

•  During testing, the calibration of the instrument is checked daily on ambient air. The as-
   found reading is taken, and the instrument is recalibrated on ambient air.

At the end of the test condition, the calibration error step described above is repeated.


CO2 Determination

CC>2 is used for molecular weight determination. At the stack, CC>2 readings are taken from
the plant continuous emission monitor (CEM). If the CEM readings are on a wet basis, they
are converted to a dry basis using the moisture content measured by the Hg train. If the CEM
is out of service or does not provide CC>2 measurements, the CC>2 content is calculated
stoichiometrically from a fuel analysis.


Chlorides, NH3, and SO3

To measure chloride concentrations in the flue gas, EPA Method 26A was used. This method
was designed to measure both the HC1/HF and C^  concentrations in the flue gas. However,
when SC>2 was present in the flue gas, it was found that the method only provides total
chlorides [1]. The impinger train is operated similarly to other sampling procedures such as

A-6

-------
                                                        Sampling Methods and Procedures

EPA Method 5. Once the chlorides are collected in the solutions, they are analyzed using ion
chromatography techniques. For SOs measurements, the controlled condensation technique
was used. For NHa analyses, the flue gas is absorbed in 0.1 N HC1 solution, and the NHa is
measured using a selective ion electrode.


Reference

1.  Sun, Q.; Crocker, C.R.; Lillemoen, C.M. The Effect of Coal Combustion Flue Gas
   Components on Low-Level Chlorine Speciation Using EPA Method. In Proceedings of
   the 92nd Annual Meeting & Exhibition of Air & Waste Management Association, St.
   Louis, MO, June 20-24, 1999.
                                                                              A-7

-------
B
MERCURY MEASUREMENTS
B.1   Mercury Measurements Made at Site S2
Complete Ontario Hydro Data Set

  Table B-1
  Ontario Hydro Mercury Data for Site S2 with the SCR in Service
Date
Hours
Start

07/17/02
07/18/02
07/19/02
58.0
85.6
106.9


07/17/02
07/18/02
07/19/02
58
85.6
106.7


07/18/02
07/19/02
90.1
110.8


07/18/02
07/19/02
90.8
110.9

into Test
End
ug/Nm3
Hgp
SCR Inlet
59.2
87.6
108.9
Average
0.01
0.09
0.03
0.04
SCR Outlet
58.6
87.6
108.7
Average
0.14
0.03
0.02
0.06
ESP Inlet
91.6
112.3
Average
0.00
0.06
0.03
ESP Outlet
92.3
112.9
Average
0.00
0.00
0.00
Hg2+

7.6
5.4
6.5
6.5

11.1
11.8
9.5
10.8

12.2
12.2
12.2

11.3
10.9
11.1
Hg°

4.8
7.5
4.2
5.5

3.3
0.8
0.5
1.6

0.5
0.2
0.3

0.5
0.2
0.3
Hg-rotal

12.4
13.0
10.7
12.0

14.6
12.6
10.1
12.4

12.7
12.4
12.6

11.8
11.1
11.5
                          Stack
07/18/02
07/20/02
91.2
132.9
93.2
134.9
Average
0.00
0.00
0.00
0.5
0.9
0.7
1.4
1.2
1.3
1.8
2.1
2.0
                                                          B-1

-------
Mercury Measurements
Coal Mercury and Chloride Analyses

   Table B-2
   Coal Analysis Completed at Site S2
Date

7/1 5/2002
7/1 5/2002
7/16/2002
7/16/2002
7/1 7/2002
7/1 7/2002
7/1 8/2002
7/1 8/2002
7/1 9/2002
7/20/2002
7/21/2002
7/22/2002
7/23/2002
7/23/2002
7/24/2002
7/24/2002
7/25/2002
7/25/2002
7/26/2002
7/27/2002
7/28/2002
7/28/2002
7/29/2002
7/29/2002
7/30/2002
7/30/2002
7/31/2002
7/31/2002
08/1/2002
08/2/2002
8/03/2002
8/03/2002
8/04/2002
8/04/2002
8/05/2002
8/06/2002
8/06/2002
8/07/2002
8/08/2002
8/09/2002
8/09/2002
8/1 0/2002
8/1 0/2002
8/1 0/2002
8/1 0/2002
Average
Standard Deviation
Chloride
ppm
672
733
577
689
724
717
605
605
593
609
635
638
639
686
704
658
601
656
640
650
602
569
640
637
646
655
684
636
710
624
630
570
648
619
561
680
617
600
591
617
621
654
622
582
561
636
44
Mercury
ppm
0.10
0.12
0.15
0.11
0.12
0.14
0.11
0.11
0.11
0.13
0.12
0.12
0.12
0.13
0.11
0.12
0.10
0.11
0.11
0.13
0.10
0.14
0.12
0.12
0.14
0.10
0.14
0.15
0.15
0.13
0.10
0.17
0.10
0.11
0.07
0.10
0.10
0.13
0.16
0.11
0.11
0.14
0.14
0.13
0.13
0.12
0.02
B-2

-------
B.2   Mercury Measurements Made at Site S4
                                                             Mercury Measurements
Complete Ontario Hydro Data Set

   Table B-3
   Ontario Hydro Mercury Data for Site S4 with the SCR In Service
Plata

Hours
Start
into Test
End
ug/Nm3
Hgp
Hg2+
Hg°
Hg-rotal
SCR Inlet
9/11/02
9/12/02
9/13/02
63.8
86.9
107.7


9/11/02
9/12/02
9/13/02
63.8
87.0
107.8


9/11/02
9/12/02
9/12/02
63.7
82.8
87.4

65.8
88.5
109.0
Average
0.04
0.11
0.00
0.05
SCR Outlet
64.8
88.3
108.8
Average
0.01
0.00
0.00
0.00
AH Outlet
65.7
84.8
89.4
Average
0.03
0.06
0.10
0.06
5.6
3.0
3.2
4.0

12.0
2.8
6.3
7.1

11.6
13.2
9.2
11.3
8.1
7.8
8.9
8.3

3.0
4.6
5.2
4.3

0.5
0.5
0.4
0.5
13.8
10.9
12.1
12.3

15.1
7.4
11.5
11.3

12.2
13.7
9.6
11.8
                                 Stack
9/11/02
9/12/02
9/12/02
63.8
82.9
87.9

65.8
84.4
89.6
Average
— a 0.3
— 0.3
— 0.3
0.3
0.7
0.9
0.8
0.8
0.9
1.2
1.2
1.1
' Not analyzed (all values will be <0.1 ug/Nm3).
                                                                           B-3

-------
Mercury Measurements


    Table B-4
    Ontario Hydro Mercury Data for Site S4 with the SCR Bypassed

UalG
Hours
Start

10/16/02
10/16/02
10/17/02
898.0
891.5
923.0

into Test
End
ug/Nm3
Hgp
AH Outlet
900.0
893.5
925.0
Average
0.14
0.05
0.05
0.08
Hg2+

8.3
8.0
6.8
7.7
Hg°

5.9
6.5
4.5
5.6
Hg-rotal

14.4
14.6
11.3
13.4
                                      Stack
10/16/02
10/16/02
10/17/02
898.3
891.5
923.0

900.0
893.2
924.7
Average
— a 0.4
— 0.7
— 0.3
0.5
6.9
7.2
7.2
7.1
7.2
7.9
7.4
7.5
1 Not analyzed (all values will be <0.1 ug/Nm ).
B-4

-------
B.3   Mercury Measurements Made at Site S5
Complete Ontario Hydro Data Set

   Table B-5
   Ontario Hydro Mercury Data for Site S5 for Unit with the SCR
                              SCR Inlet
                             SCR Outlet
                              ESP Inlet
                             ESP Outlet
                               Stack
                                                           Mercury Measurements
Date
Hours into Test
Start End
ug/NmJ
HgD
Hg*+ Hgu
Hg-Total
08/17/02
08/18/02
08/21/02
538.6
565.1
634.1

540.0
566.6
635.6
Average
0.16
0.08
0.04
0.09
7.5
6.0
4.7
6.1
5.6
8.8
9.1
7.8
13.3
14.9
13.8
14.0
08/17/02
08/18/02
08/21/02
08/22/02
08/23/02
538.4
565.1
634.1
665.1
686.5

539.9
566.6
635.6
666.6
688.0
Average
0.07
0.04
0.02
0.00
0.02
0.03
11.7
10.7
10.3
15.6
10.5
11.8
0.6
3.3
2.4
5.1
2.3
2.7
12.4
14.0
12.7
20.7
12.8
14.5
08/21/02
08/22/02
08/23/02
638.8
665.3
689.0

640.3
666.8
690.5
Average
0.13
0.00
0.09
0.07
11.7
18.4
20.2
16.8
0.8
1.0
0.5
0.8
12.6
19.4
20.8
17.7
07/26/02
07/27/02
07/28/02
08/15/02
08/17/02
08/19/02
08/21/02
08/22/02
12.1
38.3
57.3
493.9
543.0
590.0
637.4
665.3

14.1
40.3
59.3
495.4
544.5
591.5
638.9
666.8
Average
0.09
0.25
0.01
0.03
0.01
0.00
0.00
0.01
0.05
11.7
8.2
7.7
14.2
10.7
8.7
10.4
19.1
11.3
0.6
0.6
0.9
0.6
0.6
0.6
0.9
0.9
0.7
12.4
9.1
8.6
14.8
11.3
9.3
11.3
20.0
12.1
08/17/02
08/21/02
543.6
637.5
545.1
639.0
Average
0.03
0.01
0.02
0.4
0.4
0.4
0.8
1.2
1.0
1.2
1.6
1.4
                                                                        B-5

-------
Mercury Measurements
    Table B-6
    Ontario Hydro Mercury Data for Site S5 for Unit Without an SCR

Date

Hours

Start
into Test

End
ug/Nm3

Hgp

Hg2+

Hgu

HgT
ESP Inlet
08/13/02
08/14/02
08/16/02
446.3
472.2
515.8


07/26/02
07/27/02
07/28/02
08/13/02
08/14/02
08/16/02
08/23/02
11.4
38.3
57.2
446.3
472.2
515.8
682.8

447.8
473.7
517.3
Average
0.01
0.10
0.03
0.05
ESP Outlet
13.4
40.3
59.2
447.8
473.7
517.3
684.3
Average
0.02
0.03
0.01
0.00
0.00
0.00
0.01
0.01
10.7
11.7
10.1
10.8

8.7
7.1
5.6
8.1
9.4
7.1
9.5
7.9
1.4
2.8
3.7
2.6

4.1
4.2
4.4
5.8
5.1
6.0
3.1
4.7
12.1
14.7
13.8
13.5

12.8
11.3
10.0
13.8
14.5
13.1
12.6
12.6
                                    Stack
08/13/02
08/14/02
08/15/02
446.1
473.3
494.0

447.6
474.4
495.5
Average
0.00
0.01
0.00
0.00
0.4
0.7
0.4
0.5
6.7
5.9
5.6
6.1
7.1
6.5
6.0
6.6
B-6

-------
                                                                Mercury Measurements
Table B-7
Coal Mercury and Chloride Analyses
Date
07/26/2002
07/28/2002
08/01/2002
08/05/2002
08/13/2002
08/15/2002
08/19/2002
08/21/2002
08/23/2002
Average
Standard Deviation
Chloride
ppm
450
430
440
500
500
480
490
460
500
472
28
Mercury
ppm
0.14
0.14
0.14
0.12
0.13
0.15
0.13
0.15
0.11
0.13
0.013
                                                                               B-7

-------
Mercury Measurements
B.4   Mercury Measurements Made at Site S6
Complete Ontario Hydro Data Set

    Table B-8
    Ontario Hydro Mercury Data for Site S6 for Unit 1 (SCR)
Date
Hours into Test
Start End
ug/Nm3
Hgp
Hg2+ Hg°
Hg-Total
                                  SCR Inlet
09/24/02
09/25/02
09/26/02
09/26/02
81.9
106.0
129.1
133.3

83.5
107.2
130.6
135.3
Average
0.03
0.05
a
—
0.04
4.1
6.7
5.5
7.0
5.8
3.0
4.0
3.8
4.1
3.8
7.2
10.7
9.2
11.1
9.0
                                 SCR Outlet
09/24/02
09/25/02
09/26/02
09/26/02
82.0
106.0
129.1
133.3

83.5
107.2
130.6
135.3
Average
0.03
0.04
—
0.01
0.03
5.5
7.7
6.6
8.5
7.1
1.1
1.6
1.8
1.6
1.5
6.7
9.4
8.3
10.2
8.6
                                  ESP Inlet
09/24/02"
09/24/02
09/25/02
09/26/02
85.8
87.4
109.9
129.1

86.3
88.9
111.3
130.6
Average
—
0.95
0.70
0.75
0.80
2.1
9.9
7.3
8.3
8.5
0.9
0.3
0.7
0.4
0.5
3.1
11.2
8.7
9.4
9.8
                                    Stack
09/22/02
09/23/02
09/24/02
09/25/02
09/26/02
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
40.0
62.8
85.8
109.8
129.1
424.8
497.0
518.0
542.0
564.0
636.5
661.7

42.0
64.3
87.3
111.3
130.6
426.3
498.5
520.0
543.5
566.0
638.5
663.7
Average
—
—
—
0.00
—
—
—
—
0.01
—
—
—
0.00
6.0
8.9
14.3
8.3
7.5
11.8
10.6
8.2
9.0
8.0
10.6
8.0
9.3
0.2
0.4
0.5
1.2
1.5
0.7
0.5
0.2
0.2
0.6
0.2
3.4
0.8
6.2
9.3
14.9
9.5
9.0
12.6
11.1
8.4
9.3
8.7
10.8
11.3
10.1
' Not analyzed (all values will be <0.1 ug/Nm3).
1 Bold values not included in averages as there was a problem that occurred during sampling.

-------
                                                                                Mercury Measurements
    Table B-9
    Ontario Hydro Mercury Data for Site S6 for Unit 2 (SCR bypassed)

Date

Hours into Test

Start End
ug/Nm3

HgP

Hg2+

Hgu

Hg-rotal
ESP Inlet
10/08/02
10/12/02
10/18/023
425.2 426.5
518.0 519.5
657.8 659.3
Average
3.74
1.44
9.16
2.59
6.0
7.1
1.0
6.6
0.5
0.3
0.1
0.4
10.2
8.8
10.3
9.5
                                                  Stack
09/22/02b
09/23/02
09/25/02
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
39.8
63.3
110.2
424.8
497.3
518.0
542.0
564.0
636.6
661.7


41.8
64.8
111.7
426.8
498.8
520.0
543.5
566.0
638.6
663.7
Average
Standard Dev.
c
—
0.00
—
—
—
0.01
—
—
—
0.01
0.01
6.6
7.1
7.9
7.2
4.6
6.5
6.0
5.5
5.3
6.8
6.0
0.9
0.0
0.2
0.6
1.4
0.8
1.5
1.7
1.5
1.2
0.8
1.3
0.4
6.6
7.3
8.5
8.6
5.4
8.0
7.7
7.0
6.5
7.7
7.3
1.1
a Bold values not included in averages as sample appears to be a clear outlier.
b Data from 9/22 through 9/25/02 was collected prior to the SCR being bypassed for this unit and are not included in the averages.
0 Not analyzed (all values will be <0.1 ug/Nm3).
                                                                                                 B-9

-------
Mercury Measurements


    Table B-10
    Ontario Hydro Mercury Data for Site S6 for Unit 4 (no SCR)
Date
Hours into Test
Start End
ug/Nm3
HgP
Hg2+ Hg°
Hg-Total
                                      Stack
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
424.7
497.3
518.0
542.0
564.0
636.5
661.7

426.2
498.8
520.0
543.5
566.0
638.5
663.7
Average
a
—
0.00
0.02
—
—
—
0.01
5.8
4.6
4.4
3.5
2.0
4.1
3.4
4.0
2.1
1.1
1.2
1.8
2.0
2.2
2.2
1.8
7.9
5.7
5.6
5.4
4.0
6.2
5.6
5.8
 1 Not analyzed (all values will be <0.1 ug/Nm3).
    Table B-11
    Coal Mercury and Chloride Analyses for Site S6
Date
09/24/2002
09/24/2002
09/26/2002
09/26/2002
10/08/2002
10/08/2002
10/12/2002
10/18/2002
10/18/2002
Unit Collected
1 (SCR)
2 (SCR bypassed)
1 (SCR)
2 (SCR bypassed)
1 and 2a
4 (no SCR)
1 and 2a
1 (SCR)
4 (no SCR)
Average
Standard Deviation
Chloride
ppm
1210
1520
871
635
1170
1320
962
794
706
1020
300
Mercury
ppm
0.084
0.052
0.072
0.055
0.063
0.066
0.069
0.070
0.064
0.066
0.0094
 1 Composite sample from Units 1 and 2.
B-10

-------
 c
COMPLETE AUXILIARY FLUE GAS DATA FOR ALL
SITES
   Table C-1
   Auxiliary Flue Gas Data for Site S2 with SCR in Service
Date
Time into the
Test,
hr
Flue Gas
Moisture,
%
Dust Loading,3
gr/dscf
C02,
%
02,
%
                                     SCR Inlet
07/17/02
07/18/02
07/19/02
58.0
85.6
106.9
Average
9.97
9.57
9.79
9.77
0.1609°
3.2881
2.2725
2.7803
14.8
15.1
15.0
15.0
3.7
3.8
3.9
3.8
                                    SCR Outlet
07/17/02
07/18/02
07/19/02
58.0
85.6
106.7
Average
11.00
10.79
10.71
10.83
3.2452
0.82483
3.4642
3.3547
14.8
14.8
14.8
14.8
4.6
4.8
4.4
4.6
                                     ESP Inlet
07/18/02
07/19/02
90.1
110.8
Average
11.36
11.15
11.25
0.03853
1.8872
1.8872
13.8
13.9
13.9
5.6
5.8
5.7
                                    ESP Outlet
07/18/02
07/19/02
90.8
110.9
Average
11.40
10.54
10.97
0.0024
0.0018
0.0021
13.7
13.7
13.7
5.8
5.8
5.8
                                      Stack
07/18/02
07/20/02
91.2
132.9
Average
17.45
26.25
21.85
0.0008
0.0025
0.0016
13.3
13.1
13.2
6.4
6.5
6.5
a Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for compliance
 purposes.
b Dust loadings are lower than expected, attributed to ash loss upon removal of probe and single-point sampling, and not included
 as part of the average.
                                                                           C-1

-------
Complete Auxiliary Flue Gas Data for All Sites


    Table C-2
    Auxiliary Flue Gas Data for Site S4 with SCR in Service
Date
Time into
Test,
hr
Flue Gas
Moisture,
%
Dust Loading,3
gr/dscf
C02,
%
02, %
SCR Inlet
9/11/2002
9/12/2002
9/13/2002
63.8
86.9
107.7
Average
9.5
9.6
11.3
10.1
SCR
9/11/2002
9/12/2002
9/13/2002
63.8
87.0
107.8
Average
8.6
14.0
10.3
10.9
1.3396
2.1084
2.3317
1 .9266
Outlet
2.7819
4.0879
2.7819
3.2172
15.5
14.3
15.3
15.0

15.1
14.5
15.1
14.9
3.5
4.6
3.5
3.9

3.7
4.4
3.6
3.9
Air Preheater Outlet
9/11/2002
9/12/2002
9/12/2002
63.7
82.8
87.4
Average
8.6
8.3
9.0
8.6
1.2303
1.2135
0.8638
1.1025
13.4
13.0
11.2
12.5
5.5
6.0
8.0
6.5
                                                 Stack
9/11/2002
9/12/2002
9/12/2002
63.8
82.9
87.9
Average
15.7
17.3
12.6
15.2
0.0000
0.0000
0.0000
0.0000
11.8
11.4
11.3
11.5
7.4
7.9
8.0
7.8
  Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for compliance
  purposes.
C-2

-------
                                                       Complete Auxiliary Flue Gas Data for All Sites
  Table C-3
  Auxiliary Flue Gas Data for Site S4 with SCR Bypassed
Date
Time into
Test,
hr
Flue Gas
Moisture,
Dust Loading,3 CO2,
gr/dscf %
02,
                                       Air Preheater Outlet
10/16/2002
10/16/2002
10/17/2002
898.0
891.5
923.0
Average
10.2
8.2
8.3
8.9
1.2291
0.9940
1 .4883
1.2371
11
11
11.2
11.1
8.0
8.0
7.8
7.9
                                              Stack
10/16/2002
10/16/2002
10/17/2002
898.3
891.5
963.0
Average
15.4
12.3
14.1
13.9
0.0084
0.0032
0.0080
0.0065
11.2
11.1
11.1
11.1
7.8
7.9
7.9
7.9
Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for compliance
purposes.
                                                                                            C-3

-------
Complete Auxiliary Flue Gas Data for All Sites
    Table C-4
    Auxiliary Flue Gas Data for Site S5 for the Unit with an SCR
Date
Time into
Test,
hr
Moisture, Dust Loading,3 CO2,
% gr/dscf %
02,
                                             SCR Inlet
08/17/02
08/18/02
08/21/02
538.6
565.1
634.1
Average
10.68
10.83
10.45
10.65
16.2456
6.5134
3.5652
8.7747
14.7
13.9
14.2
14.3
4.7
5.6
4.7
5.0
                                            SCR Outlet
08/17/02
08/18/02
08/21/02
08/22/02
08/23/02
538.4
565.1
634.1
665.1
686.5
Average
9.51
9.37
8.31
9.48
8.63
9.06
4.6314
3.7380
3.0199
2.3088
3.3436
3.4083
14.3
13.1
13.4
14.0
14.0
13.8
5.1
6.5
6.0
5.5
5.5
5.7
                                             ESP Inlet
08/21/02
08/22/02
08/23/02
638.8
665.3
689.0
Average
8.67
9.26
8.73
8.89
2.1219
1.7475
1.1852
1 .6848
12.9
13.2
13.2
13.1
6.8
6.4
6.4
6.5
                                            ESP Outlet
07/26/02
07/27/02
07/28/02
08/15/02
08/17/02
08/19/02
08/21/02
08/22/02
12.1
38.3
57.3
493.9
543.0
590.0
637.4
665.3
Average
9.30
8.94
9.08
8.90
9.27
6.79
8.57
8.95
8.73
0.0739
0.1573
0.0669
0.1726
0.0284
0.0082
0.0521
0.0412
0.0751
13.0
13.2
13.2
13.3
12.2
12.2
12.7
13.6
12.9
6.6
6.4
6.4
6.3
7.4
7.4
7.0
5.9
6.7
                                               Stack
08/17/02
08/21/02
543.6
637.5
Average
13.88
12.26
13.07
0.0061
0.0085
0.0073
11.7
12.0
11.9
8.0
7.5
7.8
 Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for compliance
 purposes.
C-4

-------
                                                       Complete Auxiliary Flue Gas Data for All Sites
  Table C-5
  Auxiliary Flue Gas Data for Site S5 for the Unit Without an SCR
Date
Time into
Test,
hr
Moisture, Dust Loading,3 CO2,
% gr/dscf %
02,
                                            ESP Inlet
08/13/02
08/14/02
08/16/02
446.3
472.2
515.8
Average
8.3
8.1
9.3
8.7
0.2117
1.8369
0.4373
0.8287
13.6
13.8
13.3
13.6
6.0
6.0
6.3
6.1
                                           ESP Outlet
07/26/02
07/27/02
07/28/02
08/13/02
08/14/02
08/16/02
08/23/02
11.4
38.3
57.2
446.3
472.2
515.8
682.8
Average
9.1
9.1
8.9
8.5
8.1
8.9
10.7
9.0
0.0711
0.1078
0.0487
0.0091
0.0259
0.0130
0.0419
0.0453
12.8
13.3
13.3
13.0
12.9
12.4
12.4
12.9
6.8
6.3
6.3
6.6
6.7
7.2
7.2
6.7
                                              Stack
08/13/02
08/14/02
08/15/02
446.1
473.3
493.8
Average
13.6
13.2
13.6
13.5
0.0034
0.0111
0.0045
0.0063
12.8
12.7
12.7
12.7
6.9
7.0
7.0
7.0
Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for compliance
purposes.
                                                                                            C-5

-------
Complete Auxiliary Flue Gas Data for All Sites
    Table C-6
    Auxiliary Flue Gas Data for Site S6 for Unit 1 (SCR)
Date
Time into
Test,
hr
Moisture, Dust Loading,3 CO2,
% gr/dscf %
02,
                                        SCR Inlet
09/24/02
09/25/02
09/26/02
09/26/02
27.4
28.4
29.4
29.6
Average
9.04
9.01
9.03
8.98
9.01
3.4195
3.7122
4.2123
3.5784
3.7306
15.2
13.4
15.2
15.0
14.7
4.1
6.1
4.2
4.4
4.7
                                       SCR Outlet
09/24/02
09/25/02
09/26/02
09/26/02
27.4
28.4
29.4
29.6
Average
8.36
8.89
8.74
8.94
8.73
3.7555
3.7580
5.1397
4.0159
4.1673
15.1
15.2
15.4
15.2
15.2
4.3
4.1
3.9
4.2
4.1
                                        ESP Inlet
09/24/02
09/24/02
09/25/02
09/26/02
27.6
27.6
28.6
29.4
Average
10.51
8.82
8.05
8.33
8.93
0.4491
3.1538
2.1157
5.2099
2.7321
13.4
13.4
13.9
14.4
13.8
4.7
4.7
5.6
5.0
5.0
                                         Stack
09/22/02
09/23/02
09/24/02
09/25/02
09/26/02
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
25.7
26.6
27.6
28.6
29.4
41.7
44.7
45.6
46.6
47.5
50.5
51.6
Average
10.09
9.62
9.69
9.17
9.49
9.64
10.15
9.76
9.11
8.65
8.45
8.29
9.34
0.0111
0.0076
0.0059
0.0213
0.0074
0.0053
0.0044
0.0210
0.0335
0.0291
0.0292
0.0221
0.0165
13.5
13.3
13.4
13.0
13.5
12.8
12.8
13.2
12.9
12.8
12.5
13.2
13.1
6.0
6.3
6.2
6.6
6.0
6.8
6.8
6.4
6.7
6.8
7.1
6.4
6.5
 Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not
 for compliance purposes.
C-6

-------
                                                      Complete Auxiliary Flue Gas Data for All Sites
  Table C-7
  Auxiliary Flue Gas Data for Site S6 for Unit 2 (SCR bypassed)
  Date
Time into
  Test,
    hr
Moisture,
    %
Dust Loading,3
    gr/dscf
                                                                      C02,
02,
 %
                                            ESP Inlet
10/08/02
10/12/02
10/18/02
10/18/02
41.7
45.6
51.4
51.5
Average
9.10
9.34
7.72
6.95
8.28
4.0591
3.2689
4.8572
4.7264
4.2279
15.4b
15.4
15.4b
15.4b
15.4
3.5
3.9
3.7b
3.7b
3.7
                                              Stack
09/22/02
09/23/02
09/25/02
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
25.7
26.6
28.6
41.7
44.7
45.6
46.6
47.5
50.5
51.6
Average
8.56
8.32
7.59
7.82
8.42
8.20
7.59
7.05
6.97
7.18
7.77
0.0118
0.0119
0.0080
0.0068
0.0082
0.0039
0.0254
0.0237
0.0148
0.0349
0.0150
13.2
12.8
13.4
13.4
13.3
13.2
13.1
12.9
13.1
13.2
13.2
6.4
6.8
6.2
6.1
6.3
6.4
6.5
6.7
6.5
6.4
6.4
Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for
compliance purposes.
Invalid data: average was used.
                                                                                           C-7

-------
Complete Auxiliary Flue Gas Data for All Sites
    Table C-8
    Auxiliary Flue Gas Data for Site S6 for Unit 4 (no SCR)
Date
Time into
Test,
hr
Moisture, Dust Loading,3 CO2,
% gr/dscf %
02,
                                                  Stack
10/08/02
10/11/02
10/12/02
10/13/02
10/14/02
10/17/02
10/18/02
41.7
44.7
45.6
46.6
47.5
50.5
51.6
Average
7.82
8.93
8.52
7.95
7.28
6.87
6.99
7.77
0.0460
0.0022
0.0265
0.0587
0.0463
0.0521
0.0398
0.0388
14.3
14.5
14.8
14.7
14.5b
14.7
14.3
14.5
5.2
4.9
4.6
4.7
4.9b
4.7
5.2
4.9
  Dust loadings were collected as part of the OH testing using the EPA Method 17 procedure and, therefore, are not for
  compliance purposes.
  Invalid data: average was used.

-------
D
QUALITY ASSURANCE/QUALITY CONTROL
This appendix provides detailed quality assurance/quality control (QA/QC) procedures that
were used for the sampling activities. The most important QA/QC parameter for any
sampling activity is the people who perform the work. All who participated in the sampling
activities for this project had extensive training and experience in the proper procedures.


Ontario Hydro (OH) Method

To provide a high level of QA/QC for this project, all liquid samples (from the OH mercury
[Hg] speciation train impingers as outlined in Appendix A), including those used as blanks
and spikes, were analyzed on-site by the Energy & Environmental Research Center (EERC).
The primary advantage of on-site analysis is that Hg analyses can usually be obtained within
24 hr after the sampling. So if there is a problem, it can be corrected when the sampling
people are on-site. The following are specific QC procedures for the OH sampling.


Instrument Setup and Calibration

A Leeman Labs PS200 cold-vapor atomic absorption instrument was used in the field for Hg
determination. The instrument was set up for absorption at 253.7 nm, with a carrier gas of
nitrogen and 10% SnCb in 10% HC1 as the reductant. Each day, the drying tube and acetate
trap were replaced, and the tubing was checked. The rinse container was then cleaned and
filled with a fresh solution of 10% HC1. After the pump and lamp were turned on and
warmed up for 45 min, the aperture was set to the manufacturer specifications. A four-point
calibration curve was then completed using matrix-matched standards. The detector response
for a given standard was logged and compared to specifications to ensure the instrument had
been properly set up. A QC standard of a known analyte concentration was analyzed
immediately after the instrument was standardized in order to verify the calibration. This QC
standard was prepared from a different stock than the calibration standards. Requirements
stated that the values obtained must read within 5% of the true value before the instrument
was used. After the initial QC standardization was completed, standards were run every ten
samples to check the slope of the calibration curve. One in every ten samples was run in
triplicate and spiked to verify analyte recovery. A QC chart was also maintained by the
EERC chemist to monitor the long-term precision of the instrument.
                                                                             D-l

-------
Quality Assurance/Quality Control

Presampling Preparation

All data sheets, volumetric flasks, and petri dishes used for sample recovery were marked
with preprinted labels. The liquid samples were recovered into premarked volumetric flasks,
logged, and then analyzed on-site. The stack filter samples were placed in premarked petri
dishes, then taken back to the EERC, where they were analyzed using mixed-acid digestion
techniques. The prestack filter samples were placed in premarked containers, logged, and
then analyzed on-site using a Milestone DMA-80 instrument. The labels contained
identifying data,  including date, time, run number, and sample port location, which correlate
back to the data sheets.


Glassware and Plasticware Cleaning and Storage

All glass volumetric flasks and transfer pipettes used in the preparation of analytical reagents
and calibration standards were designated as "Class A" to meet American Society for Testing
and Materials specifications. Prior to being used for the sampling, all glassware was washed
with hot soapy water, then rinsed with deionized water three times, then soaked in 10% v/v
nitric acid for a minimum of 4 hr, then rinsed an additional three times with deionized water,
and dried. The glassware was stored in closed containers until it was used at the plant.


Analytical Reagents

All acids used for the analysis of Hg were trace metal-grade. Other chemicals used in the
preparation of analytical reagents were analytical reagent-grade. The  calibration standards
used for instrument calibration and the QC standards used for calibration verification were
purchased commercially and certified to be accurate within +0.5% and traceable to National
Institute of Standards and Technology Standard Reference Materials.


Blanks and Spikes

As part of the QA/QC, a field blank was associated with  sampling at each location. A field
blank is a complete impinger train including all glassware and solutions that is taken out to
the field during sampling and exposed to ambient conditions. These sample trains were then
taken apart and the solutions recovered and analyzed in the same manner as those sample
trains used for sampling activities. If the field blank showed contamination above instrument
background levels, steps were then taken to eliminate or reduce the contamination to below
background levels.

As part of the QA/QC, a field spike was also associated with each test condition. A field
spike was prepared by the field manager at a level similar to the field samples. These sample
trains were then taken apart, and the solution was recovered and analyzed in the same manner
as those sample trains used for sampling activities. The target range for recovery of the field
spike was +20%.
D-2

-------
                                                            Quality Assurance/Quality Control

The results of the blanks and spikes associated with each of the test sites are shown in
Tables D-l to D-7. With very few exceptions, blanks were at or near detection limits and
results of the spiked samples were within the 20% range required by the method.
   Table D-1
   Results of Mercury Speciation Field Blanks at Site S2
Date
7/17/2002
7/18/2002
7/19/2002
7/20/2002
KG I Solution,
M9
0.04
0.20
0.04
0.03
H2O2 Solution,
M9
<0.01
<0.01
<0.01
<0.01
KMnO4 Solution,
M9
<0.01
0.37
0.09
0.05
                                                                                   D-3

-------
o
Table D-2
Results of Mercury Speciation Field Spikes at Site S2
Date
7/17/2002
7/17/2002
7/17/2002
7/18/2002
7/18/2002
7/18/2002
7/19/2002
7/19/2002
7/19/2002

Measured
Value,
ppb
9.95
4.75
9.37
10.92
4.90
9.80
9.46
4.45
9.02
KCI
Spike,
ppb
10
5
10
10
5
10
10
5
10

Spike
Recovery,
99.51
94.98
93.69
109.20
98.00
98.00
94.64
89.04
90.16
H
Measured
Value,
ppb
0.96
1.06
2.13
1.04
1.04
2.24
1.02
1.14
2.33
2O2 Solution
Spike,
ppb
1
1
2
1
1
2
1
1
2
Spike
Recovery,
96.05
106.25
106.51
103.60
103.95
112.18
101.89
113.78
116.75
^
>
cn
en
1
E
SL
V
KMnO4 Solution §.
.w
Measured
Value,
ppb
8.49
5.12

5.02
4.90

4.98
4.75

Spike,
ppb
10
5

5
5

5
5

Spike o
Recovery, §
84.85
102.30

100.40
98.00

99.60
95.00


-------
                                                                Quality Assurance/Quality Control
   Table D-3
   Results of Mercury Speciation Field Spikes at Site S4a

Date
9/11/2002
9/12/2002
9/13/2002

Measured
Value,
ppb
9.94
9.65
9.76
KCI
Spike,
ppb
10
10
10

Spike
Recovery,
%
99.40
96.50
97.60

Measured
Value,
ppb
9.00
9.53
9.46
KMnO4 Solution
Spike,
ppb
10
10
10

Spike
Recovery,
%
90.00
95.30
94.60
1 Sampling at Site S4 was done by Western Kentucky University.
   Table D-4
   Results of Mercury Speciation Field Blanks at Site S5
Date
7/26/2002
7/27/2002
7/28/2002
8/14/2002
8/15/2002
8/18/2002
8/22/2002
KCI Solution,
I'g
0.18
0.05
0.10
0.04
0.05
0.06
<0.01
H2O2 Solution,
I'g
<0.01
0.01
0.09
<0.01
<0.01
<0.01
0.09
KMnO4 Solution,
I'g
0.05
0.08
0.13
0.06
0.12
0.05
0.01
                                                                                          D-5

-------
to
s
a
Table D-5
Results of Mercury Speciation Field Spikes at Site S5

Date
7/26/2002
7/26/2002
7/26/2003
7/27/2002
7/27/2002
7/27/2002
7/28/2002
7/28/2002
7/28/2002
8/14/2002
8/14/2002
8/14/2002
8/15/2002
8/15/2002
8/15/2002
8/18/2002
8/18/2002
8/18/2002
8/21/2002
8/21/2002
8/21/2002


Measured
Value,
ppb
11.00
5.24
10.00
9.46
4.82
9.30
9.75
4.39
9.01
10.15
5.53
9.87
9.21
7.03
11.23
15.64
10.61
15.71
16.52
16.52
21.84
KCI

Spike,
ppb
10
5
10
10
5
10
10
5
10
10
5
10
10
7
10
15
10
15
15
15
20
H2O2 Solution

Spike
Recovery,
109.99
104.83
100.01
94.62
96.37
92.98
97.48
87.85
90.12
101.49
110.63
98.71
92.06
100.48
112.34
104.24
106.14
104.76
110.13
110.13
109.20

Measured
Value,
ppb
0.92
0.81
1.92
0.99
1.98

0.81
0.87
1.97
1.72
2.00
4.76
1.08
1.13
2.31
1.06
0.87
1.75
1.02
1.21
2.08

Spike,
ppb
1
1
2
1
2

1
1
2
2
2
5
1
1
2
1
1
2
1
1
2

Spike
Recovery,
92.05
81.20
96.08
99.10
99.10

80.96
86.72
98.72
86.10
99.90
95.14
108.15
112.65
115.43
106.19
86.92
87.33
101.79
121.29
104.13
KMnO4 Solution

Measured
Value,
ppb
4.48
4.69

4.96
4.73

4.59
5.13

5.11
4.75

5.25
5.16

7.56
7.84

6.88
10.62


Spike,
ppb
5
5

5
5

5
5

5
5

5
5

7
8

7
10


Spike
Recovery,
89.60
93.80

99.20
94.60

91.74
102.60

102.10
94.90

104.90
103.10

108.00
98.00

98.27
106.21

:urance/{
'.^j
§
^*i

-------
                                                           Quality Assurance/Quality Control
Table D-6
Results of Mercury Speciation Field Blanks at Site S6
Date
9/23/2002
9/24/2002
9/26/2002
9/27/2002
10/9/2002
10/13/2002
10/14/2002
10/15/2002
10/18/2002
10/19/2002
KCI Solution,
I'g
0.02
0.11
0.23
0.04
0.05
0.03
<0.01
0.11
0.01
<0.01
H2O2 Solution,
I'g
<0.01
<0.01
0.07
<0.01
<0.01
0.01
<0.01
0.11
0.02
0.04
KMnO4 Solution,
I'g
0.12
0.15
0.19
0.10
0.04
<0.01
<0.01
0.09
0.02
<0.01
                                                                                    D-7

-------
o

oo
to
s
a
         Table D-7

         Results of Mercury Speciation Field Spikes at Site S6

Date

Measured
Value,
ppb

KCI
Spike,
ppb




H2O2 Solution
Spike
Recovery,
Measured
Value,
ppb
Spike,
ppb
Spike
Recovery,



KMnO4 Solution
Measured
Value,
ppb
Spike,
ppb
Spike
Recovery,
urance/Quality Cc
9/23/2002
9/23/2002
9/23/2002
9/26/2002
9/26/2002
9/26/2002
9/27/2002
9/27/2002
9/27/2002
10/9/2002
10/9/2002
10/9/2002
10/13/2002
10/13/2002
10/13/2002
10/14/2002
10/14/2002
10/14/2002
10/15/2002
10/15/2002
10/15/2002
10/18/2002
10/18/2002
10/18/2002
10/19/2002
10/19/2002
10/19/2002
10.68
5.56
10.46
3.23
4.56
10.13
12.42
5.50
11.10
4.35
5.39
10.91
4.76
4.56
11.06
5.68
6.10
11.82
4.98
5.36
10.98
5.41
5.55
11.11
5.58
6.15
11.08
10
5
10
3.4
5
10
11.6
5
10
5
5
10
5
5
10
5
5
10
5
5
10
5
5
10
5
5
10
106.79
1 1 1 .22
104.61
94.96
91.11
101.27
107.09
109.96
110.98
87.02
107.86
109.09
95.26
91.22
110.57
113.68
122.08
118.16
99.69
107.23
109.76
108.28
110.96
1 1 1 .06
111.58
123.06
110.81
0.83
1.13
2.12
1.04
0.96
1.58
0.97
0.96
1.75
0.88
0.90
1.86
0.89
1.05
1.94
1.05
1.16
2.34
1.18
1.10
1.98
1.39
1.23
2.18
1.64
1.20
2.53
1
1
2
1
1
2
1
1
2
1
1
2
1
1
2
1
1
2
1
1
2
1
1
2
1
1
2
82.93
113.31
106.10
103.89
96.43
78.93
96.75
95.89
87.63
87.73
89.90
93.16
89.33
105.07
96.94
104.81
115.55
116.99
118.20
109.95
98.85
138.56
122.90
109.22
163.57
119.94
126.47
5.15
5.55

12.47
5.63

5.31
5.69

4.57
5.28

4.91
5.06

5.20
5.21

4.68
5.05

5.35
5.45

6.09
5.71

5
5

11.6
5

5
5

5
5

5
5

5
5

5
5

5
5

5
5

103.01
110.99

107.47
112.67

106.13
113.87

91.36
105.64

98.20
101.20

103.90
104.10

93.62
100.98

107.08
108.92

121.76
114.24


-------
                                                        Quality Assurance/Quality Control
QA/QC Checks for Data Reduction and Validation


Data Reduction

Data reduction occurred in two phases. First, preliminary data reduction occurred on the job
site. On-site data reduction may be performed by sampling and analytical personnel or by the
team leaders. Preliminary calculations include velocity, moisture, stack gas flow, sample gas
volume, percent-isokinetic sampling, and flue gas Hg concentrations. Calculations were
performed using spreadsheets on a portable computer; some averaging was done with a
calculator. Standardized spreadsheets were used.

The second phase of data reduction occurred after the team had left the job site. This
included review of the field data and input of laboratory results to complete the calculated Hg
concentrations for the coal and ash samples. In addition, the Hg speciation calculations that
were done in the field were rechecked and put into a predefined data sheet. Equations to be
used in the calculations were contained in the method.


Data Validation

All data, data entry, and calculations were double-checked by the originator and reviewed by
a second person. Reviews included recalculation of results, data entry checks, and calculation
of known and accepted data sets using the existing spreadsheet.


Sample  Identification and Chain of Custody

Samples were identified with unique sample numbers and descriptive notations. Sample
custody was maintained by EERC personnel; samples were stored and taken back to the
EERC. Once the samples were received by the EERC laboratory, sample condition was
checked and then logged into the EERC logging system.

Data sheets were kept in the custody of the originator or the program manager or in locked
storage until returned to the office. The original data sheets were used for report preparation,
and any additions were initialed and dated.


Personnel Responsibilities and Test Schedule


Test Site Organization

Each project comprised a team of personnel able to provide the expertise needed for project
completion. The site-specific test plan (SSTP) that was provided to the company outlines  the
designated management, sampling, and plant personnel required for each project. The key
roles of EERC project personnel for project completion are listed below:
                                                                              D-9

-------
Quality Assurance/Quality Control

•  Project manager

•  Field manager

•  Principal investigator

•  Project chemist
•  Sample custodian

•  Sampling technicians
•  Mercury  semicontinuous emission monitor (Hg SCEM) technicians


Test Preparations


Construction of Special Sampling Equipment and Modifications to the Facility

The correct length of sample probes was made prior to going into the field. No modifications
were needed.


General Services Provided by the Facility

The facility provided safe access to suitable sample ports; process data; 110-V, 20-amp
power at the  sample locations; a suitable location to park test trailers; and power for the test
trailers. In addition, the plant provided restrooms and a clean area for breaks or lunch. The
facility was expected to provide the necessary safety training for the sampling team once they
were on-site.


Access to Sampling Sites

Site visits were conducted to determine, among other things, that all sample ports were
readily accessible. In addition, measurements were taken so that modifications to probes
could be made prior to going into the field.


Sample Recovery Areas

The EERC provided test trailers to set up and tear down  sample trains and do the analysis.
The trailers were situated in an area as free as  possible from ambient dust contamination.


Test Personnel Responsibilities and Detailed Schedule

Table D-8 lists the key project personnel for this project. Table D-9 lists the various
personnel roles and their specific responsibilities. Table D-10 presents a typical test schedule
for a 4-week project. A tentative project schedule with dates and activities was provided in
the SSTP provided to the company prior to sampling.

D-10

-------
                                                           Quality Assurance/Quality Control
Table D-8
Key Project Personnel
Organization
EPRI
DOE
EPA
EERC
EERC
WKU
WKU
QA/QC
EERC
Individual
Paul Chu
Lynn Brickett
C.W. Lee
Dennis Laudal
Jeff Thompson
Wei-Ping Pan
Kunlei Liu
David Brekke
Jeff Thompson
Responsibility
EPRI Project Manager
DOE Performance Monitor
Project Consultant
Project Manager
Principal Investigator
Project Manager
Principal Investigator
QA/QC Manager
QA/QC Oversight for WKU
Phone Number
(650)855-2812
(412)386-6574
(919)541-7663
(701)777-5138
(701) 777-5245
(270) 780-2532
(270)-745-3251
(701)777-5154
(701) 777-5245
E-Mail Address
pchu@epri.com
lynn.brickett@
netl.doe.gov
lee.chun-wai@
epamail.epa.gov
dlaudal@undeerc.org
jthompson@undeerc.org
wei-ping.pan@wku.edu
kunlei.liu@wku.edu
dbrekke@undeerc.org
jthompson@undeerc.org
                                                                                  D-ll

-------
Quality Assurance/Quality Control
    Table D-9
    Test Personnel and Responsibilities
    Staff Assignment
                            Responsibilities
Project Manager
Principal Investigator
Field Manager
Team Leader


Sampling Technician


Project Chemist


Sample Custodian


Plant Engineer
EPRI, EPA, DOE, and the EERC developed and approved the overall test
program, coordinated all test activities, developed the QA/QC test plan,
ensured the project was being completed within budgetary guidelines, provided
data interpretation and completed all reporting requirements, maintained
communication between all test participants, and assisted with other activities
as required.

Worked with the project manager to coordinate all test activities, was
 esponsible for maintaining communications between the plant representative
and the sampling team, provided input into program decisions made by the
 Imding agencies and the project manager, worked with the field manager to
 insure that the objectives for each test program were completed, collected
plant data, completed data reduction and provided input into all reports, and
assisted in other activities as required.

 oordinated or helped perform all sampling activities; coordinated sampling
activities  being conducted by the EERC with those being conducted by plant
personnel; maintained sample custody records; ensured that sampling was
completed so that the objectives of the project were met, including all QA/QC
 •equirements; ensured that all safety requirements were met by the sampling
 earn; provided input into project reports; and assisted other activities as
 •equired.

 °repared and operated  the OH train and Hg SCEMs, recorded and reduced
data, and assisted in sample recovery and other activities as required.

Assisted in preparation  and operation of the sample trains and assisted in
sample recovery and other activities as required.

 erformed all analytical activities at the on-site laboratory, maintained sample
custody records, and shipped samples to off-site laboratory when necessary.

Maintained sample custody records, transferred samples to on-site laboratory,
and assisted in sample  recovery and other activities as required.

Worked with the field manager and principal investigator to facilitate data and
 nformation transfer regarding  plant operations.
D-12

-------
                                                             Quality Assurance/Quality Control
    Table D-10
    Typical Test Schedule for a 4-Week Project
     Day
                                  Activity
1-2

3-4
5-10



11

4-26

19-20
21-26
27-28
Travel to site.

Contact site representative, establish communications, and review unit operation;
coordinate crew safety meeting; and prepare and site sampling trailers.

Set up sample recovery and analysis area, mix fresh reagents as necessary, load
sample trains for sampling, set up field blanks, and collect reagent blanks and do
reagent blank analyses.

Set up Hg SCEMs and pretreatment/conversion systems at the proper locations.

Prepare locations for sampling (i.e., building rails) and conduct preliminary
measurements.

Leak-check sample trains.

Conduct sampling activities for the first test conditions (individual responsibilities
outlined in Table D-9), ensure all blanks and spiked samples meet QA/QC criteria,
and ensure all Hg SCEMs are operating properly and giving good data.

Pack equipment, package samples for transport to the EERC,  and leave site.

1 operator remains to operate Hg SCEMs for the duration of test period.

Perform second round of OH analysis.

Set up sample recovery and analysis area, mix fresh reagents as necessary, load
sample trains for sampling, set up field blanks, and collect reagent blanks and do
reagent blank analyses.

Set up Hg SCEMs and pretreatment/conversion systems at the proper locations.

Prepare locations for sampling (i.e., building rails) and conduct preliminary
measurements.

Leak-check sample trains.

Conduct sampling activities for the second test conditions (individual responsibilities
outlined in Table D-9), ensure all blanks and spiked samples meet QA/QC criteria,
and ensure all Hg SCEMs are operating properly and giving good data.

Pack equipment, package all samples for transport to the EERC, and leave site.
Prior to sampling, 2 days were scheduled for equipment setup. Setup activities included
setting up the equipment at the test locations, verifying power at the test locations, and
conducting a preliminary velocity traverse (assuming the boiler is operating at or near the
target test load). Final coordination with station personnel was done, and safety briefings
were held.
                                                                                    D-13

-------
Quality Assurance/Quality Control

Test team personnel arrived at the plant a minimum of 1.5 hr before the start time of the first
test run on each of the days scheduled for sampling. Pretest activities included final
equipment setup and leak check and verification of target unit operation and communication
links between team members, team leaders, and plant personnel.
D-14

-------
E
SAMPLE CALCULATIONS
Sample calculations are included for each of the calculated parameters. Data were used from
the selective catalytic reduction (SCR) unit inlet location during Day 3 (09/24/2002) from
Site S6.


Volume of Gas Sample

Vm(std)         =  Volume of gas sample measured by the dry gas meter, corrected to
                  standard conditions, dscf

                  K, x Vmc x Pm
Vm(std) (dscf)    =  —	
   '                  Tm + 460
T/ /^i            17.64x30.485xlx30.02  0-7n,,,  ,
Vm(std)         =  	=27.944 dscf
                        117.7 + 460
Where:

Ki             =  17.64 °R/in. Hg

Vmc           =  Vm x Cm = Volume of gas sample as measured by dry gas meter
                  corrected for meter calibration (Cm = meter calibration coefficient)
                  (dcf)

Pm            =  Meter pressure (in. Hg)

Tm            =  Meter temperature (°F)


Volume of Water Vapor

Vw(std)         =  Volume of water vapor in the gas sample, corrected to standard
                  conditions, scf

Vw(std) (set)    =  K2xH2O(g)

Vw(std)         =  0.04715 x 58.9 = 2.777 scf
                                                                         E-l

-------
Sample Calculations


Where:

K2              =  0.04715 ft3/g

H2O(g)          =  Mass of liquid collected in impingers and silica gel (g)


Water Vapor in the Gas Stream

Bws            =  Water vapor in the gas stream, proportion by volume

                        Vw(std)
Bws
                   Vm(std) + Vw(std)
                       2.777

BWS            -   27.944 + 2.777    =°'09°4
Dry Molecular Weight

Md             =  Dry molecular weight of stack gas, Ib/lb-mole

Md (Ib/lb-mole)  =  0.440 x (%CO2) + 0.320 x (%O2) + 0.280 x (%N2 + %CO)

Md             =  0.440 x 15.2 + 0.320 x 4.1 + 0.280 x 80.7 = 30.6 Ib/lb-mole

Where:

%(CO2, O2, N2, CO)  =  Percent (CO2, O2, N2, CO) by volume, dry basis


Molecular Weight

Ms             =  Molecular weight of stack gas, wet basis, Ib/lb-mole

Ms (Ib/lb-mole)  =  Md x  (l - Bws) + 18.0  x Bws

Ms             =  30.6 x (1  - 0.0904) + 18.0 x 0.0904 = 29.5 Ib/lb-mole


Average Stack Gas Velocity

Vs              =  Average stack gas velocity, ft/sec

                                             460
Vs (ft/sec)       = K3xCpx (Ap)1'2 (avg) x
E-2
                                         / C -I- ZLfM I
                               1/2 ,
                                         Psx Ms

-------
                                                                    Sample Calculations
Vs


Where:
Cp


Ap
(Ap)1/2(avg)
Ts



Ps
=  85.49 x 0.84 x 1.0488 x
                                           704 + 460
                        29.37 x 29.46
                                                      1/2
= 87.4 ft/sec
                 =  85.49 ft I secx
                                     Ib
                                  Ib - mole
                                           x in Hg
                                                   1/2
=  Pitot tube coefficient (dimensionless)



=  Velocity head of stack gas (in. Hg)



=  Average of the square root of Ap values



=  Stack gas temperature (°F)



=  Stack pressure (in. Hg)
Isokinetic Sampling Rate


I                =  Percent of isokinetic sampling, %


                    K4 x (Ts + 460) x Vm(std) x 144
If0/
                      Ps x Vs x An x <9x (1 - Bws)



                      0.09450 x (704 + 460) x 27.944 x 144   _  AA
                 —	100.5/0
                    29.37 x 87.4 x 0.01986 x 95 x (1 - 0.0904)
Where:
An



6
   0.09450% (in Hg)(miri)

          °R x sec



   Cross-sectional area of nozzle (in.2)



   Total sampling time (min)
                                                                                E-3

-------
Sample Calculations
Volume of Gas Sample Corrected to 3% O2
Vm*(std)


Vm*(std)
=  Volume of gas sample measured by the dry gas meter (Vm(std)),
   * corrected to 3% oxygen, Nm3
= K5 x Vm(std) x
21-%O2
   18
Vm*(std)
                   21 - 41
= 0.02832 x 27.944 x	 = 0.743 Nm3
                     18
Where:
Mercury

Hg (|ig/Nm3)


Hg


Particulate Hg

Oxidized Hg

Elemental Hg
Fd

Fd
                =  0.02832 m7ft3
  Vm * (std)

  2.259
  0.743
         = 3.04 ng/Nm (note: using the Hg  from Day 3 SCR inlet)
   Sum of mercury from filter and nozzle rinse

   Sum of mercury from KC1 impingers

   Sum of mercury from H2O2 and KMnC>4 impingers (note: all H2O2
   impinger values were nondetects). Since typically less than 5% of the
   elemental mercury (Hg°) is trapped in the H2O2 impinger, the less-than
   values were not added to the total Hg°. Thus the Hg° was calculated
   from the values obtained from the KMnO4 impingers only.
Fd(dscf/10bBtu)  = 106x
   Value relating gas volume to the heat content of the fuel

       [(K6 x %H) + (K7 x %C) + (Ks x %S) + (K9 x % JV) - (Kw x
                                                   HV
E-4

-------
                 = 106x
                                                                      Sample Calculations

                         [(3.64 x 5.23) + (1.53 x 70.74) + (0.57 x 0.86) + (0.14 x 1.52) - (0.46 x 9.46)]
                                                       11936



                 = 10,357 dscf/Btu
Where:

                          dscf
                      '
                      64
                      '
                         %Hx Ib


                          dscf
                          dscf
                    057
                      '
v                        dscf
K9               -
                         %Sxlb



                    014
                      '
                         %Nxlb

                           dscf
HV              =  Heating value of coal (Btu/lb)




% (H, C, S, N, O) =  Percent (H, C,  S, N, O) in coal (as-received from ultimate analyses)
                                                                                   E-5

-------
                                      TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1. RE PORT NO.
      EPA-600/R-04/032
                                2.
                                                                3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Effect of Selective Catalytic Reduction on Mercury, 2002 Field
Studies Update
             5. REPORT DATE
              April 2004
             6. PERFORMING ORGANIZATION CODE
7. AUTHORS
D. Laudal
                                                                8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Energy and Environmental Research Center
 University of North Dakota
 P.O. Box 9018
 Grand Forks, ND 58202-9018
                                                                10. PROGRAM ELEMENT NO.
              11. CONTRACT/GRANT NO.

              CR-829353-01
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, North Carolina 27711
                                                                13. TYPE OF REPORT AND PERIOD COVERED
              14. SPONSORING AGENCY CODE

              EPA/600/13
15. SUPPLEMENTARY NOTES
The EPA Project Officer is Chun Wai Lee, Mail Drop E305-01, Phone (919) 541-7663
16. ABSTRACT
The report documents the results and provides both a summary of tests conducted during the 2002
Selective Catalytic Reduction Mercury Field Sampling Project and also an overall evaluation of the results
from both 2001 and 2002 tests. Mercury (Hg) measurements have been completed at six different power
plants, four in 2001 and two in 2002, with selective catalytic reduction (SCR). In addition, two of these
plants tested in 2001  were tested again in 2002 for a total of eight data sets. Results of the field
measurements made at these six power plants in 2001 and 2002 indicate that SCRs can increase Hg
oxidation and improve Hg removal in the downstream flue gas desulfurization (FGD) systems. This effect
appears to be  more likely for bituminous coal applications where >90% Hg2+ is possible at the particulate
control device  inlet. The three bituminous coal-fired power plants with wet achieved Hg removals of
84%-92% with SCR operation, compared with 43%-51% without SCR operation. These increased removal
efficiencies may be due to the combined effects of the SCR system to increase Hg2+ concentrations and
reduce reemissions of the Hg° from the FGD system. The effect of catalyst space velocity and age are not
clear, but may have an impact on SCR Hg oxidation. The only Powder River Basin (PRB) site tested did not
show a high oxidation rate. These findings are based on a relatively small data set and, thus, should be
considered preliminary rather than final conclusions. For example, two of the three FGDs tested were
magnesium-lime systems, and the third FGD was a venturi scrubber; thus the combined effect of SCR and
the most common FGD design of a limestone, forced-oxidation system has yet to be evaluated.
17.
                                     KEYWORDS AND DOCUMENT ANALYSIS
             DESCRIPTORS
                                                   b. IDENTIFIERS/OPEN ENDED TERMS
                                                                               c. COSATI  Field/Group
Air Pollution
Mercury (metal)
Emission
Elimination
Scrubbers
Electrostatic Precipitators
Performance Tests
Pollution Control
Stationary Sources
13B
07B
14G
                             14B
18. DISTRIBUTION STATEMENT
                                                   19. SECURITY CLASS (This Report)
                                                   Unclassified
                             21. NO. OF PAGES
                                     174
       Release to Public
20. SECURITY CLASS (This Page)
 Unclassified
                                                                               22. PRICE
EPA Form 2220-1 (Rev. 4-77 ) PREVIOUS EDITION IS OBSOLETE
     E-6
                          forms/admin/techrpt.frm  7/8/99 pad

-------