DAQ-2024-006989

1

DAQC-277-24
Site ID 10007 (B5)
MEMORANDUM

TO: CEM FILE – HOLCIM

THROUGH: Harold Burge, Major Source Compliance Section Manager

FROM: Rob Leishman, Environmental Scientist

DATE: March 19, 2024

SUBJECT: Source: Kiln
Contact: Clint Badger – 801-829-2122
Mark Miller – 972-221-4646
Location: Devil’s Slide, Morgan County, UT
Test Contractor: Holcim In-House Staff
FRS ID#: UT0000004902900001
Permit/AO#: Title V operating permit 2900001004 dated November 18, 2021
Last revised March 25, 2022
Subject: Review of RA/PST Protocol dated March 18, 2024

On March 18, 2024, DAQ received a protocol for a RA/PST (relative accuracy/performance specification
test) of the Holcim Portland Cement Kiln at Devil’s Slide, Morgan County, UT. Testing will be
performed on May 2, 2024, to determine the relative accuracy of the HCL monitoring systems.

PROTOCOL CONDITIONS:

1. Test Method EPA PS 15 - Measurement of Gaseous Hydrogen Chloride Emissions at
Portland Cement Kilns by Fourier Transform Infrared (FTIR) Spectroscopy, analyte
spiking procedure: OK

DEVIATIONS: No deviations noted.

CONCLUSION: The protocol appears to be acceptable.

RECOMMENDATION: Send attached protocol review and test date confirmation notice.

1 8 2

Holcim (US) Inc.
6055 E. Croydon Rd.
Morgan, Utah 84050

Phone [phone]
www.holcim.com/us

Sent via email

March 18, 2024

Bryce C. Bird, Director
Utah Division of Air Quality
195 North 1950 West
Salt Lake City, UT 84116

Re: HCl Performance Test Notification for Holcim (US), lnc.'s Devil's Slide Cement
Manufacturing Plant in Morgan, UT

Dear Mr. Bird,

Holcim (US) Inc. ("Holcim") owns and operates a Portland cement manufacturing plant located
in Morgan, UT. The Devil's Slide Plant operates under Utah Division of Air Quality issued Title V
Permit, No. 2900001004. The Plant is subject to the National Emission Standards for
Hazardous Air Pollutants for the Portland Cement Industry ("Portland Cement NESHAP"), 40
C.F.R. Part 63, Subpart LLL, and is a major source of HAPs.

Holcim will perform testing and evaluation of the ABB ACF5000 FTIR to meet the HCl
continuous emission monitoring requirements in Subpart LLL. This site-specific HCl CEMS
performance evaluation test plan is prepared to satisfy the requirements of 40 CFR 63, Subpart
A §63.8(e) for the required monitor performance test (i.e., “performance evaluation”). The
performance evaluation will follow the requirements in EPA Performance Specification 15 for
FTIR-based CEMS. Spike testing will be completed as outlined in the protocol with an example
of previous spike test showing run times from each subsequent calibration of 12 spikes as
attachment A. The test is currently scheduled to be conducted May 2, 2024 by Holcim Devil’s
Slide Plant personnel.

If you should have any questions or need additional information, please contact me at
702.358.7280 or by email at [e-mail].
Sincerely,
Mark Miller
Mark Miller
Environmental Director

Enclosure test protocol and attachment A
Site-Specific HCl CEMS Performance Evaluation Test Plan for
Holcim (US) Inc. Devil’s Slide Plant

Holcim (US) Inc. (“Holcim”) owns and operates a Portland cement manufacturing plant
located in Morgan, Utah. The Devil’s Slide plant is subject to the National Emission
Standards for Hazardous Air Pollutants for the Portland Cement Industry (“Portland
Cement NESHAP”), 40 CFR Part 63, Subpart LLL, and is a major source of HAPs.

The Devil’s Slide plant will perform testing and evaluation of the ABB ACF5000 FTIR to meet
the HCl continuous emission monitoring requirements in Subpart LLL. This site- specific HCl
CEMS performance evaluation test plan is prepared to satisfy the requirements of 40CFR63,
Subpart A §63.8(e) for the required monitor performance test (i.e., “performance
evaluation”). The performance evaluation will follow the requirements in EPA Performance
Specification 15 for FTIR-based CEMS.

Schedule
The performance evaluation is expected to be conducted on May 2nd, 2024, starting at
10:00 A.M. Holcim will notify UDAQ as soon as practical of any scheduling changes or
delays.

Holcim Contacts
Questions or comments about the test program should be directed to:

Clinton Badger
Environmental Manager
Holcim (US) Inc.
801-829-2122
[e-mail]

Test Services
Holcim will provide the necessary performance evaluation services. Devil’s Slide Plant
personnel will conduct the performance evaluation. These individuals have extensive
experience working with CEMS and HCl emission measurements at the Devil’s Slide
plant.

Objectives and Approach
The objective of this test program is to demonstrate that the ABB ACF5000 HCl CEMS
meets the applicable requirements of Performance Specification 15 (PS 15) for FTIR-
based CEMS. If successful, this will complete the requirements for the monitor
performance test (i.e., “performance evaluation”). The monitor can continue to be used
to acquire HCl emissions data to demonstrate compliance with the emission limitation
and will be subject to ongoing quality assurance requirements in 40CFR60, Appendix F,
Procedure 1.
Subpart LLL, §63.1350(l)(1) states:
“(1) If you monitor compliance with the HCl emissions limit by operating an HCl
CEMS, you must do so in accordance with Performance Specification 15 (PS 15) of
appendix B to part 60 of this chapter, or, upon promulgation, in accordance with
any other performance specification for HCl CEMS in appendix B to part 60 of this
chapter. You must operate, maintain, and quality assure a HCl CEMS installed and
certified under PS 15 according to the quality assurance requirements in Procedure
1 of appendix F to part 60 of this chapter except that the Relative Accuracy Test
Audit requirements of Procedure 1 must be replaced with the validation
requirements and criteria of sections 11.1.1 and 12.0 of PS 15.”

The above referenced Section 11.1.1 and calculations in Section 12.0 of PS 15 require
that the performance of the CEMS be validated by conducting analyte spikes (also
known as “dynamic spikes”) rather than by the more traditional method of acquiring
independent effluent measurements and calculating the relative accuracy. In addition,
PS 15 does not include a 7-day zero and upscale calibration drift test.
However, it does require injection and analysis of a calibration transfer standard (CTS)
on each day of the test program and it requires that the monitor response agree within
± 5% of the CTS value. CO2 tri blend gas will be used as the CTS gas. The CTS gas may
include other constituents such as SO2 or NO.

Analyte Spiking Procedure
HCL spike test will consist of 12 sequential up and down analyte spikes which will
typically be conducted over a 2.0 hour period with 2 minutes between spikes. Holcim
has provided an example of typical spike sequencing as attachment A. Analyte spiking is
the quantitative addition of a known amount of calibration gas into the effluent gas
stream. The exact concentration of the calibration spike gas is determined by directly
analyzing the calibration gas by the FTIR (or using the tag value if trusted). The
calibration gas is then spiked into the CEMS probe immediately upstream of the
particulate matter filter (similar to a normal CEMS system calibration.) The difference
between analyte spiking and system calibrations is that only a fraction of the total flow
is injected into the effluent stream instead of conducting a “probe flood” or similar
procedure that results in 100% calibration gas in the analyzer cell.

Spike gas is introduced at a rate of no greater than 10% of the total sample flow rate:
i.e., 10:1 dilution so that the sample gas matrix remains mostly unchanged. (The native
effluent concentrations are diluted by the introduction of the spike gas.) For example if
the total system flow is 10 ppm, then the spike flow can be no greater than 1 lpm.
Dynamic spikes that introduce less spike gas than 10:1 (e.g., 15/1 and 20/1) are also
acceptable. The nominal HCl concentration of the spike gas cylinder will be between 40
and 70 ppm. The spike gas will also contain about 330 ppm N 2O ppm as a tracer to
determine spike flow to total flow rate according to equation 3.2 below. The goal is to
quantitatively add approximately 4 to 7 ppm of HCl with the stack gas effluent and then
examine the CEMS response to determine if the sampling and analytical systems are able
to produce reliable HCl monitoring data. The specific procedure is as follows: Analyze the
spike gas cylinder directly with the FTIR. Name the response values as the cylinder
reference value (i.e., “tag value”) for both HCl and N 2O concentrations. Alternately, use
the gas supplier provided tag value if it is a known and trusted value. Obtain stable
sequential spectra while sampling stack gas only while the process is operating at stable
conditions and name this as unspiked sample number 1.

Begin injecting HCl spike gas at a 10:1 ratio1 or less until stable readings are obtained.
(ACF5000 values are updated approximately every 35 seconds and are the result of the
co-addition of x FTIR scans. VIM DAS compiles 1-minute averages but is not synchronized
to ACF5000 analysis cycles.) Collect several sequential measurements and name the
initial final stable response as spiked sample number 1. Collect an additional
independent value (separated by 1 or more 1-minute VIM DAS averages) and name this
as spiked sample 2. Shut the spike off and allow the measurements to stabilize back to
unspiked effluent only (“baseline”) concentration. Then, record two independent
baseline stable values as unspiked sample numbers 2 and 3. Repeat this series of spiked
and unspiked samples until there are twelve sets of spiked and unspiked effluent
measurements.

Variations in the effluent concentrations of either HCl or the tracer gas N2O during the
analyte spike procedures can adversely affect the results. It is important to conduct
these tests during operating periods when variability of these concentrations is low. In
the event that excess variations in the HCl or N2O concentrations are observed, it may
be necessary to discard affected spike runs and conduct additional runs.

Sulfur hexafluoride (SF6) is often used for FTIR analyzers as a tracer because SF6 has a
large infrared absorption, is easily measured at low concentrations (especially for liquid
nitrogen cooled MCT detectors) and is not present in the effluent. 5 ppm of SF6 is a
good cylinder concentration for some FTIRs. However, ABB prefers use of N2O as a
tracer since it was already a compound included in their TUV certified analysis routines.
The concentration of N2O in the cylinder must be significantly higher due to the weaker
N2O absorption characteristics. In addition, the calculations of spike recovery must take
into account that there is a low level of N2O present in the effluent samples at most
cement kilns. ABB previously submitted documentation and procedures supporting the
N2O spike procedure to Holcim.

These documents are available upon request. Jim Peeler of EMI derived HCl spike
recovery calculations for the use of N2O as a tracer during the Holicm 2012 HCl CEMS
Field Evaluation. These procedures were validated during the Holcim 2012 study and
also used during the Holcim 2015 evaluation of the ABB ACF5000 CEMS. The dilution
factor calculation when the tracer gas is present in the effluent samples is now included
in the Performance Specification 18, Appendix A “Standard Addition Procedures” as
equation A3.

The HCl spike recovery calculations for systems that also measure N2O as used in the previous
Holcim studies and now found in PS18, Appendix A are as follows:

1 Adjust the spike gas flow rate to provide a spiked N2O, value to a level of 0.1 (10%) or less of the direct analysis value
of N2O. For example, if a direct analysis is 330 ppm N2O, then adjust the spike gas flow rate until the observed stable
value of the spiked effluent N2O measurement is about 33 ppm plus the native N2O concentration.
Dilution Factor: 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 = � N2OObserved − N2ONative � N2OBottle − N2ONative N2OObserved − N2ONative CExpected = CNative �1 − N2O Bottle − N2O � Native N2OObserved − N2ONative + CSpike Gas � N2O Bottle − N2O � Native
Where:
CExpected = HCl expected concentration response for dynamic spike (i.e., analyte
spike) based on the sum of the diluted native concentration and spike addition
CNative = HCl concentration present in the sample gas without spike added
N2OObserved = Concentration of N2O observed during dynamic spike gas injection
concurrent with stable spiked HCl response
N2ONative = N2O concentration in the sample gas without spike added N2OBottle =
Concentration of N2O in spike gas determined from direct analysis of the spike gas
CSpike Gas = HCl concentration of spike gas as determined from direct analysis of the spike
gas

The definition of the dilution factor above is the reciprocal (i.e., 1/DF) as defined in
PS 15.

Calculate the amount of added HCl in the spiked gas:
Cs = Dir HCl/DF in PS15 Eq. 5 is therefore modified to be:
Cs = Dir HCl*DF
Calculate the Bias:
B = Sm – Mm – Cs

Where:
Sm= mean of average of spiked samples
Mn = mean of average of unspiked samples

Calculate the Standard Deviations:
Calculate the standard deviation and RSD of the spiked samples (and unspiked
sample) differences.
Where:
Di2 = the difference between spiked and unspiked sample pairs; S1-S2 and U1-U2
RSD = Sds/Sm
Calculate the CF:
CF = 1/(1 + B ) Cs Calculate the composite standard deviation of the combined differences: SD = �Sds2 + Sdu2
Determine If Bias is Statistically Significant – Compare t = 2.201 B t = SD
Spreadsheet calculations have been developed for these calculations.

Data Quality Objectives
The data quality objectives are:
• Demonstrate acceptable CTS gas analysis results for each day of the test
program where the ACF5000 response is within ± %5 of the gas value.
• Acquire a minimum of 12 valid spiked and unspiked pairs of samples where
the spike gas flow rate is 10% or less of the total sample flow rate as
determined by the N2 O tracer concentration and the measurements are not
adversely affected by temporal variations in the effluent concentrations
• RSD < 50% for the spiked (and unspiked) paired differences
• If the apparent bias is determined to be statistically significant, the bias
correction factor (CF) shall be between 0.7 and 1.3

CEMS Description and Quality Assurance
An ABB ACF5000 FTIR-based CEMS will be evaluated to determine its ongoing ability
to acquire accurate and precise HCl effluent measurements. The CEMS is a full
extractive (i.e., non-dilution) measurement system. It consists of an M&C sample
probe with provisions for injecting calibration gases and spike gases upstream of the
PM filter, approximately 96 feet of heated sample line (also containing calibration
gas lines and communications) and the ABB FTIR analyzer. The sample lines and
analyzer gas cell are maintained at 180°C (356°F), while the sample probe and PM
filter may be operated at a higher temperature of ~230°C.

The ABB sample probe acquires samples from the stack at the same level used for all
stack sampling activities and a point approximately 1 meter from the stack wall.
RATAs of SO2, NO, O2 and CO2 CEMS have been performed at this location for many
years and there is no evidence of gas concentration stratification and no reason to
expect that such stratification is present.

The ACF5000 will communicate with the VIM DAS that is used to record all emission
monitoring data. All effluent measurements that are used to determine the
ACF5000 performance will be acquired from the VIM DAS records.
Due to delays and technical problems encountered by NIST and EPA, NIST Traceable HCl
calibration gases with an uncertainty of ≤5% are not available. HCl calibration gases with
nominal concentrations of 40 and 70 ppm HCl and 330 ppm N2O will be used. These
gases are being certified in accordance with a broadly applicable EPA approved
alternate method designated Alt-114 which was based on an alternative method
request and supporting information submitted to EPA by Air Liquide. EPA’s approval is
contained in a letter from Steffan Johnson, Group Leader, Measurement Technology
Group, OAQPS, dated Feb. 22, 2016. The designation of these gases are HCl GMACS and
are certified by Air Liquide (now Airgas) by dual interlocking gravimetric and classical
gas analysis methods, each with a total combined uncertainty of <5%. Additional
information regarding the HCl GMACS and an example certificate of analysis are
available upon request.

Compressed HCl calibration gases will be introduced to the ACF5000 using Silconert
Treated dual stage low volume stainless steel regulators and HCl Gas Delivery Panel
Prototype VI and Nitrogen Purge & Purification Panel acquired from Air Liquide.
These devices allow the gas regulators and gas delivery panel to be completely purged
with ultra-dry nitrogen prior to and after the introduction of HCl mixtures to greatly
reduce the potential for internal corrosion and/or HCl losses and to reduce the time
required to achieve stable HCl responses from the ACF5000.

After the performance certification of the ACF5000 for the measurement of HCl
concentrations is successfully completed, the Devil’s Slide plant will continue to operate
the HCl CEMS in accordance with the requirements of 40 CFR 60, Appendix F, Procedure
1. The span value of the HCl CEMS will be 20 ppm. Daily zero checks will be performed by
injection of zero gas and daily upscale calibration checks will be performed using an
internal gas cell. Quarterly cylinder gas audits will be performed using HCl GMACS
compressed gases at 20-30% of span and 50-60% of span. In accordance with the explicit
requirements of §63.1350(l)(1), dynamic spike evaluations will be performed annually.
Run Date time Action HCl
Cylinder
[ppm]
HCl
FTIR
[ppm]
N2O
Cylinder
[ppm]
N2O
FTIR
[ppm]
HCl
Expected
[ppm]
%
Recovery
(all)
Cs = HCl
Added
[ppm]
HCl
Change
[ppm]
%
Recovery
(Change)
DF
Spike 1 Up 11/4/2022 12:22 PM baseline 0.26 0.8
11/4/2022 12:13 PM spike response 40.2 3.90 333.0 33.2 4.16 93.9%3.92 3.67 93.5%9.8%
Spike 2 Down 11/4/2022 12:24 PM baseline 0.20 0.9
11/4/2022 12:16 PM spike response 40.2 4.04 333.0 31.8 3.92 103.0%3.74 3.86 103.2%9.3%
Spike 3 Up 11/4/2022 12:46 PM baseline 0.35 0.9
11/4/2022 12:37 PM spike response 40.2 4.72 333.0 36.2 4.59 102.9%4.27 4.41 103.1%10.6%
Spike 4 Down 11/4/2022 12:48 PM baseline 0.25 0.7
11/4/2022 12:40 PM spike response 40.2 4.32 333.0 33.9 4.24 101.9%4.02 4.09 102.0%10.0%
Spike 5 Up 11/4/2022 13:06 PM baseline 0.40 1.0
11/4/2022 12:57 PM spike response 40.2 3.89 333.0 30.0 3.88 100.3%3.51 3.52 100.4%8.7%
Spike 6 Down 11/4/2022 13:10 PM baseline 0.23 0.7
11/4/2022 12:59 PM spike response 40.2 4.13 333.0 32.0 3.99 103.4%3.79 3.92 103.6%9.4%
Spike 7 Up 11/4/2022 13:32 PM baseline 0.26 0.6
11/4/2022 13:19 PM spike response 40.2 4.80 333.0 36.6 4.59 104.7%4.35 4.57 104.9%10.8%
Spike 8 Down 11/4/2022 13:35 PM baseline 0.21 0.7
11/4/2022 13:22 PM spike response 40.2 4.53 333.0 34.0 4.22 107.4%4.03 4.34 107.8%10.0%
Spike 9 Up 11/4/2022 13:53 PM baseline 0.30 0.6
11/4/2022 13:44 PM spike response 40.2 4.01 333.0 30.8 3.93 102.2%3.65 3.74 102.3%9.1%
Spike 10 Down 11/4/2022 13:56 PM baseline 0.24 0.5
11/4/2022 13:46 PM spike response 40.2 4.08 333.0 31.1 3.92 104.1%3.70 3.86 104.4%9.2%
Spike 11 Up 11/4/2022 14:25 PM baseline 0.32 0.8
11/4/2022 14:12 PM spike response 40.2 4.76 333.0 36.1 4.56 104.4%4.27 4.47 104.7%10.6%
Spike 12 Down 11/4/2022 14:42 PM baseline 0.37 0.7
11/4/2022 14:16 PM spike response 40.2 4.57 333.0 34.9 4.47 102.3%4.14 4.24 102.4%10.3%
Mean 102.5%3.95 102.7%
Spike No.Spiked Response Sn-S(n-1)Sdi*Sdi
Unspiked
Response Un-U(n-1)Udi*Udi
Cs (HCl
Added,
ppm)
1 3.90 0.26 3.92
2 4.04 0.14 0.020 0.20 -0.06 0.004 3.74
3 4.72 0.35 4.27
4 4.32 -0.40 0.160 0.25 -0.10 0.010 4.02
5 3.89 0.40 3.51
6 4.13 0.24 0.058 0.23 -0.17 0.029 3.79
7 4.80 0.26 4.35
8 4.53 -0.27 0.073 0.21 -0.05 0.003 4.03
9 4.01 0.30 3.65
10 4.08 0.07 0.005 0.24 -0.06 0.004 3.70
11 4.76 0.32 4.27
12 4.57 -0.19 0.036 0.37 0.05 0.003 4.14
Mean Sm 4.31 Mean Mm 0.28 Mean Cs 3.95
SDs 0.171 SDu 0.065
RSDs 3.97%RSDu 23.1%
Composite SD 0.18
Composite RSD 23.4%Precision Acceptable RSD < 50%
Bias = Sm-Mm-Cs (ppm)0.08
Statistically Significance Test t 0.441 Bias Not Statisticaly Significant if t < 2.201
Bias is not statistically significant
Correction Factor CF 0.980 Correction Factor Is Not Required
Table 1: Dynamic Spike Results for Holcim (US), Inc. Devil's Slide ACF5000 CEMS
Analysis from VIM DAS Data ASTM D 6348 Method 321
Table 2: PS 15 Calculations for Devil's Slide ACF 5000 HCl CEMS