FTIR CEM Performance Specification
Modification Considerations
MKS Instruments
On-Line Product Group
2 Tech Drive, Suite 201
Andover, MA 01810
Tel: 978-482-5364
MultiGas™ FT-IR
Automated HCl CEM Systems
EPA/ICAC Emissions Measurements
Roundtable Meeting
September 17, 2013
Peter G. Zemek
MKS Instruments
Marty Spartz
Prism Analytical
Technologies
Performance Specifications for
FTIR CEMS
PERFORMANCE SPECIFICATIONS:
• PERFORMANCE SPECIFICATION 15 - FOR EXTRACTIVE FTIR
CONTINUOUS EMISSIONS MONITOR SYSTEMS IN STATIONARY
SOURCES
• RM RATA RUN COMPARISONS – FOLLOW PERFORMANCE
SPECIFICATION 2—SPECIFICATIONS AND TEST PROCEDURES FOR
SO2 AND NOX CONTINUOUS EMISSION MONITORING SYSTEMS IN
STATIONARY SOURCES
• DRAFT PERFORMANCE SPECIFICATION 18 AND TEST PROCEDURES
FOR HCL CONTINUOUS EMISSION MONITORING SYSTEMS IN
STATIONARY SOURCES
FTIR Reference Methods for
CEMS
REFERENCE METHODS: TO VALIDATE UPON INSTALLATION OR RATA
• METHOD 301—FIELD VALIDATION OF POLLUTANT MEASUREMENT
METHODS FROM VARIOUS WASTE MEDIA
• METHOD 320* - MEASUREMENT OF VAPOR PHASE ORGANIC AND
INORGANIC EMISSIONS BY EXTRACTIVE FOURIER TRANSFORM
INFRARED (FTIR) SPECTROSCOPY (INCLUDES FTIR PROTOCOL)
• METHOD 321** - MEASUREMENT OF GASEOUS HYDROGEN CHLORIDE
EMISSIONS AT PORTLAND CEMENT KILNS BY FOURIER TRANSFORM
INFRARED (FTIR) SPECTROSCOPY
• METHOD 318 - EXTRACTIVE FTIR METHOD FOR THE MEASUREMENT
OF EMISSIONS FROM THE MINERAL WOOL AND WOOL FIBERGLASS
INDUSTRIES
• ASTM D6348-12 STANDARD TEST METHOD FOR DETERMINATION OF
GASEOUS COMPOUNDS BY EXTRACTIVE DIRECT INTERFACE
FOURIER TRANSFORM INFRARED (FTIR) SPECTROSCOPY.
*1 of 2 RM for EGU
**Only acceptable RM for Portland Cement MACT
Confusion and Errors
Between Methods/Specifications
TEST METHOD 320 - MEASUREMENT OF VAPOR PHASE ORGANIC AND
INORGANIC EMISSIONS BY EXTRACTIVE FOURIER TRANSFORM INFRARED
(FTIR) SPECTROSCOPY - (This is already a “self-Validating” Compliance
(Reference) Method with its own QA/QC)
PS-15 states- For System Validation, “Since FTIR CEMS have
multicomponent capability, it is possible to perform more than one RM
simultaneously, one for each target analyte.”
“?”
Interpretation, If use EPA M320, 321, 301, etc, ASTM D6348 you can RATA
w/FTIR
Comparison PS/RM for FTIR
Parameter PS-15 PS-18 Draft EPA M320 ASTM D6348-03,12 M301
Dates Promulgated Feb 2000 Feb 2014? Feb 2000, Protocol 1995 Feb 2003, '10,'12 June 1991, '04, '11
Cal Gas Direct Accuracy 5%, 7% w/bias NA 2% or 5% 5% NA
Dynamic Spike Recovery +-30% +-15% +-30% +-30% +-30% to 50%
# of Spike Runs 24 (M301), 3 (M320) 6? 3 1 24
Interference test No (just mentions them) <3% span combined No bias check No
Method Relative Accuracy (DQO)
20% (RM in denom) 15%
(Emiss Limit in denom)
SO2 (10% Emiss Limit)
PS2 < 15% 2.5% 2%
Acc 10%
Prec 5% t-test comparison
"Bias" Correction Yes ?
">30% is not valid
method", but apply a corr
factor May Yes, above 10%
Calibration Error Test 10%
3 pt, <5%, intercept
<15% of Span No One Knows No depends on RM
Dry Cal check
YES (All Compounds)
Daily ? Yes, but select set..no HCl
Yes, but select
set..no HCl depends on RM
7-Day Drift No 3% of span/day No No No
Detector linearity 3 point, 2%/5% yes? 3 point, 2%/5% 3 point, 2%/5% NA
Detection Limitations depends on RM LOS, MDC#3, LOD MAU, OFU
MDC#1, MDC#2*,
MDC#3 PLQ*, LOD
Cell Volumes per Data Pt. 10 spiking, 5 sample 5? 5 5 NA
Conf Intervals depends on RM 99%? 95% 95% 99% LOD, 95% t-test
* Reviewed/Revised
Manual Data Verification NA (Yes RATA RM) NA (Yes RATA RM) YES YES NO
Calibration Transfer Standard (CTS) 5% ? 5% 5% NA
*MDC #2 = PLQ (no longer used)
FTIR is a Special Case
• Most of the Methods/PS were written with traditional CEMs in mind
• Traditional CEMS/GC are destruction techniques and as such change
daily due many factors, i.e. ionization plates, fuel, gas conditioning
• FTIR is an optical technique where sample is not changed in any way
– Therefore, it DOES not change unless it is not operating properly
• If one compound does not perform 1 day, all of the compounds will be
biased, but if one compound does perform, they will all perform
• EPA has stated, “If EPA library reference spectra are not available,
use calibration standards to prepare reference spectra according to
Section 6 of the FTIR Protocol”
• EPA is stating that, “if the spectra are good, the numbers will be
good” – Important point! Initial certification pass = very little on-
going QA/QC needed
FTIR Do not Drift or Change
Response
MKS Instruments 7
FTIR has come far (err…mid)
FTIR Requirements for CEM has been burdened to the point
that much of the daily QA/QC and initial certification is
unwarranted and will prevent its use in industry as newer
technology QA/QC becomes more efficient (FTIR Rules
Outdated)-Discriminatory
Consistent with current practices, the Agency would continue to expect CEMS to achieve
high data availability.., monitor downtime in excess of 5.0 percent may warrant appropriate
investigation and follow-up activities.
75.33(d) would make it a violation of the primary measurement requirement § 75.10(a) to
allow the annual monitor data availability to drop below 80.0 percent for SO2, NOX, or
CO2. (Docket A–97–35, Item II–B–16).
40 CFR Parts 72 and 75 [FRL–6007–8] RIN 2060–AG46 Acid Rain Program;
Continuous Emission Monitoring Rule Revisions
Uptime = Credits = $
Current PS-15 Requirements
not necessary today
PS-15 Required Reason Whats Changed
Detector Linearity Manufacturer performed
and Performance Based
New Boards and Algorithms
Cell leak test CTS will determine if
leaks present
Redundant
X-axis accuracy Will fail relative accuracy
and dynamic spike test
Only 1 manufacturer has a problem
Resolution Check Will fail relative accuracy
and dynamic spike test
Redundant
Audit Spike Dynamic spike with
Audit is Good if conc
correct
Used in place of Dynamic Spike
100%T Line 5% Req Current algorithms
correct for this
Faster computers/More stabile
Temperatures
Audit Spectra Silly Not possible unless all spectral
functions are provided
Submit Spectra for Analysis
to EPA
Really? Need multiple soft wares and training.
No one else is required to do this
Signal to Noise Calculation Meaningless Will be determined by LOD/LOS
Daily Dry Zero, Mid, Span or
ALL TARGET COMPOUNDS
Does not represent
actual target compound
response (Need H2O),
Contaminates System
Need to measure HCl, NH3, CH2O, HF,
HCN and other polar species at very
low levels
If dry cal gas spans/etc are required, then certain cmpds will take hours
Cement Plant CEM Data with no
Filter Change/Back-Purge for
3 Months- End of Period Test
A steady-state is reached until the moisture changes
Once dry, HCl is liberated in very high concentrations
Recommended On-Going PS-15
On-Going QA/QC Performance Specifications (After Initial Installation
Certification):
QA/QC Checks that should be required on a daily basis or other basis are as
follows:
• Background – Times are at the Discretion of Manufacturer
• Daily – CEM System Zero (2% of 1st Span)
• Daily - CTS indirect gas check, preferably using Ethylene or dilution of
CO2/H2O (System Check)
• Daily - Dynamic spike check based on a minimum of at least 1 target
compound. (Preferably the component that is expected to be the hardest to
recover during the dynamic spiking test due to its chemical or reactive nature,
adherence to the wetted components, or has the highest spectral
interferences) (20% Recovery at 10% DF @ 50% expected or 5x MDL *
Special Cases Options should exist for HCl/NH3 streams)
• Semi-Annual - RATA test with an acceptable reference method (Beware
traditional CEMs such as Chemiluminesense or CO/CO2
Calibration Transfer Standard
(CTS)
The FTIR CEM, once validated is accurate and precise if
functioning properly
Indirect CTS gas checks the instrumentation and sampling
system functionality
Used to determine:
• Leaks within the sampling system
• T95 Rise and Fall times
• Instrument performance results for
• Resolution
• Signal to noise ratio (SNR)
• Linearity
• Frequency accuracy
• Sample cell path-length
• Instrument Line Shape (ILS)
• Noise Equivalent Area (NEA)
• Detection Limit (DL)
• Spectral limits
• Ensures that the correct Analytical Method is loaded
CTS – Ethylene – Wet or Dry
Detection Limits - Which One?
Needs Modification and
Consistent Terminology
MKS Instruments 14
1 set of data
PS-15/PS-18 Detection Limitations
Modifications Updating (~LOS)
M301 “The PLQ is a limit determined by the standard deviation of an estimate
of a concentration; if the standard deviation of the estimate exceeds a
threshold, then that estimate is unacceptable. The LOD is a limit determined
by the estimate of the concentration itself” (therefore PLQ = MDC#2)
The LOD is clearly a threshold. An estimate that cannot be distinguished from
one resulting from a blank sample is unlikely to provide meaningful results.
y = 0.0017x + 0.0201
0
0.005
0.01
0.015
0.02
0.025
0.03
0 0.2 0.4 0.6 0.8 1 1.2
HCl Std Dev as f(conc)
Std as f(c)
Linear (Std as f(c))
LOD = 0.02 ppm
1. Prepare 7 samples standard LOD1
2. Std Dev = S1
3. LOD0 = 3S1
4. LOD1 ≤ 2LOD0, then S0 = S1
5. LOD1 > 2LOD0, then:
6. 2 additional standards <LOD1
7. 7 samples std dev each = S2, S3
8. Plot S1, S2, S3 as f(conc)
9. Draw Regression Line
10.Extrapolate to zero
Span? Must be Defined for FTIR
Each Cal
Point is
a Span
Multi-Span
in 1 Method
Must Define
Span PS15
For Your Consideration?
Typical dynamic spike involves diluting the CEM sample flow by 10% to get 50% addition
5 ppm native, use 25 ppm HCl cylinder diluted 10x = ~2.5 ppm added to 5 ppm native
Theoretical CEM response = ~ 7.5 ppm HCl
0102030405060708090
100
2 52 102 152 202 252 302
% R
ecov
ery
Spike Level (ppm)
Recovery vs. Spike Level (5 ppm native HCl)
(10 percent dilution)
70
81
8.1 ppm added to 5 ppm native HCl
Greater than 50% Rule (162%)
Matrix Effect - is it wrong?
High NH3
slip
25
Fail because not 10% DF? Why? Still reports 5 ppm at stack – correct
50% of Native Rule or 10% Flow Rule Wrong?
40
ASTM allows
up to150% of
Native Conc.
Recommendations
MKS Instruments 18
• Rewrite/Modify/Update PS15
• Allow substitution of ASTM D6348-12 for M320 or
Correct/Modify/Update M320
• Standardize on Terminology
• Establish a sub-committee of outside interested
parties/stakeholders made up of experienced field
technicians/scientists/engineers in FTIR to provide
recommendations to EPA on changes to PS15 and M320
• Make it understandable to the layman
• Try to keep the Politics Out of the Science