Document Rjr7r5Gwjo5veyoq8a0VN5NmV
1001 G Stn:et. N.W. Suite 500 West Washington. D.C. 2000 1 tel. 202.4 34.4100 jiu 202.434.4646
April 27, 20 15 Via Electronic Mail
Writer's Direct Access
Jean -Cy ril W a lk er
+ I 20 2 43-L-1181 walkcr @ khlaw co m
Ms. Jodi Howard U.S. Environmental Protection Agency Office of Ai r Qual ity Planning and Standards 109 T.W. Alexander Drive Mail Code: E 143-0 I Research Triangle Park, NC 2771 1 l loward.Jodi@epa.gov
Re: pH Calibration Requirements Under National Emission Standards for Hazardous Air Pollutants for Polyviny l Chloride and Copolymers Production ("P VC MACT"), 77 Fed. Reg. 22,848 (April 1 7, 2012)
Dear Ms. Howard:
As they are finalizing several subm issions for the reconsideration of the above-referenced rule, the Vinyl Institute and its PVC MACT Working Group (hereinafter, " Working Group").l have asked that we follow-up with you on an issue of great concern that arose durin g the March 25.20 15, conference call with the Agency. Specifically, Working Group members were dismayed by the discussion on the pH calibration requirements set out at Table 7 to the final rule. Based on this discussion, it would appear that there is a deep and significant misunderstanding of the industry's processes, wh ich has led to the imposition of an overly burdensome pH calibration requirement in the above-referenced rule. Accordingly. the Working Group seeks to address the apparent misunderstandings below in the hope that the Agency will reconsider the calibration freq uency required for pH monitors.
I. Variability in Resin Slate has no Bearing on Scrubbing Media or pH Measurement
Hydrochloric acid (" HCI") is a byproduct of the thermal oxidation of chlorinated organic HAPs, which in the PVC industry is neutralized by using a sodium hydroxide (NaOH) solution or other neutra liz ing media in the caustic scrubbers. Faci lities monitor pH of the caustic liquid.
l tn additio n to Viny l Institute ("VI" ) members Fommsa Plastics Corporation, U.S.A., Occidental Che mical Corpo ration/Oxy Vinyls. LP, Mexichcm Specialty Resins Inc. (formerly PolyOne Corpo ration), Shintcch Inc.. Westlake C hemical Corporation, Lubrizo l Corporation, and Wacker C hemicals, the PVC MAC T Wo rking Gro up has included non-VI me mbe rs Axiall Corporation. The Viny l Institute, Inc., founded in 1982, is a U.S. trade association representing the leading manufacture rs of vinyl. vinyl ch lor ide mono mer. vinyl additives and modifiers, and viny l compound mate rials.
Washingto n. O.C.
Brussels
San Francisco
www.kh law.com
Shanghai
KELLER AND H ECKMAN LLP
Ms. Jodi Howard April 27, 2015 Page 2
using an appropriate probe connected to a continuous analyzer that reads the pH of the caustic liquid at a minimum frequency of once every 15 seconds. The caustic liquid is very effective at neutralizing the HCI gas, and as such the pH of the scrubbing liquid is an exce llent parameter for controlling HCI emissions.
Based on the March 25, 2015 discussions, it would appear that Agency staff believe that an HCI continuous emissions monitoring system (CEMS) is a better method of monitoring HCI because it measures HCI emissions directly. The Working Group submits that an HCI CEMS may not be the best parameter to control scrubber performance because it is in effect a lagging indicator of such performance. In contrast, pH mon itors provide a more direct measure of the neutralization reaction taking place in the scrubber, and allow for earlier control intervention in the process.
More importantly, Agency staff have argued that the need for frequent pH ca libration is due to the variability in PVC resin recipes. The industry has identified variability in product s lates as a driver for flexibility of total organic HAPs with the resin and wastewater limits, which EPA appears to have interpreted to extend to the composition of treated process vent streams. The basis for this position is unclear as industry members have never argued that resin recipes affect the composition of treated process vent streams. Rather, Working Group members noted that most emission profiles sought by EPA would have little value because the recovery system causes process vent HAP concentration to have poor correlation with product s late.
Indeed, the load of the thermal oxidizer is a function of the production volume, and does not necessarily directly correlate to the variability of the resin product slate. As resin production rates increase, the chlorinated organics load to the thermal oxidizer increases, lead ing to greater concentrations of HCI in the thermal oxidizer exhaust. As the Working Group noted in its August 16, 20 13 letter:
A PVC facility is designed to prevent inerts and non-condensable materials from entering the process stream. Inerts require increased pressure in the process, increasing the risk of pressure relief device ("PRD") releases or other malfunctions, and interfere with the recovery of VCM in the recovery system. More importantly, PVC PUs are designed to maximize raw material value by recovering unreacted VCM and reusing it in the process ... Vent streams collected from process equ ipment (both continuously and batch-wise) are recovered overhead and reused. The process is engineered to maximize the percent of the stream that is recovered and minimize the amount that is ultimately vented from the recovery system to a control device.2
As a result, the load of HCI precursors is a function of the efficiency of the recovery system at each facility, and will not vary significantly by product s late. In addition, the fact that HC I is the
'- Docket Document EPA-HQ-OAR-2002-0037-0561 al 52.
K ELLER AND H ECKMAN LLP
Ms. Jodi Howard April 27, 20 15 Page 3 only acidic byproduct of thermal oxidatio n at PVC faci lities supports the expectatio n that nonchlorinated o rganics are not expected to play a role in the format io n of o ther acid ic HAPs. Even if HCI rates do vary, scrubbers at PVC fac ilities are well designed to compensate. The system continually adds caustic or other neutralizing agent, based on the pH measurement (taken at a minimum of 15 second intervals), to e nsure that the HC I is instantly neutralized. As Fig ure I illustrates, pH level will not vary widely at the thermal oxidizer scrubber effiuent.1. This data shows an actual process unit that is in-control and well-controlled over the two week study period. As long as pH of the scrubber liquid is maintained above a certain level, instantaneous neutralization of incoming steady state loads of HCI wi ll be accomplished.
Calibratio n data from PVC faci lities indicate that pH monitors are unlike ly to experience much drift. For example, the cha rt at Appendix B provides 12 weeks of pH calibratio n data from 4 d ifferent continuous pH meters at one PVC facility. As the data indicates, none of the 4 pH monitors was ever adjusted by more than 0.26 pH d uring its weekly calibration and most of the adjustments were less tha n 0. 10 pH for those three months. This data shows that there is very little drift in the pl-I measurement. Thus, more frequent than weekly calibration is simply unnecessary.
1 Note that the two d ips in Fig ure I for pH on March 2 and March 4 are read ings ofa test buffe r solution, rather tha n a spike in incoming HC I in the sc rubber.
fl -1-
-.
>=
----c3
u;::i
8 8
0
t
! ;
-1-
-.~-.
.
Li.,
K ELLER AND H ECKMAN LLP
Ms. Jodi Howard April 27, 2015 Page 5
II. Calibration Every 8 Hours is Overly Burdensome and U nnecessary
As you know, 40 C. F.R. 63 . I I935(c)(5) of the PVC MACT requires that a ll continuous parameter monitoring systems (CPM S) must comply with the calibration frequency set out at Table 7. For pH monitors, the requirement is every 8 hours of process operation , which can impose significant burdens on faci lity operations. We understand that most PVC industry facil ities use pH monitors with a continuous in situ pH probe. The problem is that cal ibration of these devices can be fai rly onerous if performed multiple times w ithin a 24-hour period. The pH analyzer must be taken off-line to remove the pH probe from its holder and perform the calibrations steps summa rized at Attachment A. According to Working Group members, the calibration can take anywhere from 15 to 40 minutes, depending on how many buffer sol utions are used, w hat kind of probe connectors are in place (probe must be removed to perform calibration), access to the probe, adjustment requ irement, instrument manufacturer, and other cons iderations.
In addition, pH calibration every 8-hours would req uire facili ties to hire up to 4 new instrument technicians to staff a 4-shift operation. By way of background, calibration services at PVC fac ilities must be performed by employees w ith specialized training. The instrument technician must have a gene ra l understanding and education in mathematics, chem istry and physics, as well as electrical and e lectronics training and experience. In addition, instrument technicians must attend training offered by instrument manufacturers at each fac ility. Most PVC fac ilities use instrument technicians that work 40 hours a week. By requiring calibration every 8 hours, the Agency is essentia lly compe lling faci lities to move away from having one or two technicians that work 40-hours per week, to a 4-person shift rotation similar to those for operations that run around the clock.1
Cal ibrating pH monitors every e ight hours a lso is unnecessary. Several factors govern the frequency of calibration, such as types a nd amounts of contaminants in the liquid stream being analyzed, type of neutralizing media, and concentrations of caustic or acidic components. In the case of the PVC industry, the o utflow from the thermal oxid izer consists entirely of gaseous byproducts of combustion, w ith little if any particulate matter formed durin g this highly controlled thermal oxidation process. Thermal oxidizer quench a nd scrubber liquids neutralize the HC I byproduct component to very low residual levels carried out in the exhaust gas leaving the scrubber and vented to the atmosphere. And, as noted above, the process stream is re latively consistent, such that PVC facilities do not experience fouli ng and other issues that may require more frequent calibration in other industries.
i Although the calibration work would be fairly light duty for a technician. it would be very difficult to train shift employees for this task, and would impose a significant load on them in addition to their other duties.
K ELLER AN D H ECKMAN LLP
Ms. Jodi Howard April 27, 201 5 Page 6
An info rmal review of industry facilit ies indicates that most facilit ies calibrate their pH monitors at least every two weeks. Spec ifically, of the nine fac ilities that responded to the YI, calibration frequency was distributed as fo llows:
Calibration Frequency
I
Weekly
Every Two Weeks
Monthly
Quarterly
Total
~umber of Facilities 5 I 2 I 9
These calibration inte rvals appear to be consistent with manufacturer recommendations. For example, Mettler To ledo states that:
The calibration interval needed for pH is highly application-depende nt. The more uniform process conditions are (temperature, pressure, composition, freedom from coating, etc.), the more stable a pH sensor wi ll be and the longer the cal ibration interva l may be. Most pH insta llations calibrate between once a week and once a month but shorte r or longer intervals may be appropriate based on expe rience. It is a good practice to begin by calibrating frequently and then gradua lly go to longer interva ls, as stabil ity in the installation and acc uracy requirements a llow.1
Similarly, YSI sensors are recomme nded to be ca librated every two weeks:
Frequency o f ca libration for pH and pH/ORP sensors de pends on the degree o f foulin g and the desired level of accuracy. These sensors begin to dri ft immediate ly after cal ibration. Whi le drift occurs w ithin hours, it does not become immediately noticeable for the a verage use r. For most in-s itu appl ications with average conditions and ty pical accuracy demands (e.g. 0.05 pH units), it's recomme nded that these sensors are cal ibrated at least once every two weeks}1.
III. Frequent Calibration Could Lead to Permit Violations
In addition, the sampling freque ncy could lead to vio lation of permit conditions . In this regard, Table 5 to Subpart HHH HHHH requires fac ilities that use pH monitors as C PMS to monitor a nd record the pl-I of the scrubbe r liquid once every 15 minutes. In addition. the following requirements apply :
i http://us.mt.com/us/en/home/supoortive content/supportive information/support fag/FAO-T HOR-ORPpH .html#questio n I
2 http://www.fondriest.com/pdf/fondricst calibration guidc.pd f.
KELLER AND HECKMAN LLP
Ms. Jodi Howard April 27, 201 5 Page 7
40 C.F.R. 63. l I890(c) and 63.8(c)(4) require the C PMS to be in continuous o peration, unless a spec ific exemption is met.
40 C.F.R. 63. l I 890(c)(2) prohibits the use of data recorded during specific events, including required monitoring system quality assurance or qua lity control activities (including calibration checks and required zero and span adjustments) when reporting emission leve ls.
40 C.F.R. 63. l l 890(c)(2) directs fac ilities to report any period during which the monitoring system failed to collect required data.
Obvious ly, if facilities need to take pH monitors out of service for 30-45 m in, three times a day for calibration in accordance with Tabl e 7, these monitors cannot collect pH data at the frequency required by the rule.
The preceding recitation is re levant because the Working Group is aware of at least one faci li ty w ith a special condition in its permit that makes a deviation any loss of"valid data" due to periods of monitor break down, out-of-control operation, repair, maintenance or calibration above 5% in a 24-hour period. Taking a pH monitor out o f service for 90 minutes per day would cause the unit to exceed the 5% allowable threshold (over 24 hours).
The Working Group submits that EPA should reconside r the requirement that facilities
us ing pH monitors as CPMS calibrate these dev ices every 8 hours . As noted above, the re lative
consistency of the waste gas stream and cleanliness of the caustic liquid stream a re not expected
to require cal ibration o f the pH monitor every 8 hours. Conseque ntly, the Working G roup
submits that requiring weekly calibration or calibration in accordance w ith manufacture r
recommendations should be sufficient to ensure that pH monitors are operating correctly.
Making this change would remove an unnecessary regulatory burden, as well as e liminate a
potential conflict between this provis ion and the data recording requirements in the rule and with
state air permits.
r -Crd Cordially yours,
~/4,__
Jean-Cyri l Wa lker
cc: Penny Lassiter, U.S. Environmental Protection Agency Mark Kataoka, U.S. Environmenta l Protection Agency Richard Krock, The Vinyl Institute The Vinyl Institute PVC MACT Working Group
Enclosure
KELLER AN D HECKMAN LLP
Ms. Jodi Howard Apri l 27, 20 15 Page A- 1
Appendix A
Typical pH Monitor Calibration Procedure at PVC Facility
I.
Record the process pH reading prior to calibration.
2.
Place the pH controller in " manual " so that it does not react to calibration activity.
(Note that putting the controller in manual holds the control valve in its same % open position during the calibratio n process until the controller is taken off manual control.)
3. Isolate the pH probe and pull the probe from the sampler / holder.
4 . Place the probe in lower pH Buffer (typically 4.0) and record result.
5. Place the probe in upper pH Buffer (typically 7.0) and record result. 6. Clean the probe.
7. Place the probe in lower pH Buffer and calibrate as needed to obtain lower result.
8. Place the probe in upper pH Buffer and cal ibrate as needed to obtain upper result.
9.
Place the probe back in the sampler / holder.
I0. Re-align the probe to the process stream.
11. Record result of process stream after cal ibration.
12. Place pH controlle r back in "automatic" mode.
K EL LER AND H ECKMAN LLP
Ms. Jodi Howard April 27, 2015 Page B- 1
Appendix B: PVC Facility Incinerator pH Calibration Records - Effluent Streams
il m & RI I Buffer 4.01 Prior Buffer 4.01 After
to Calibration
Calibration
Buffer 4.01 Adjustment
Buffer 6.86 Prior Buffer 6.86 Afte
to Calibration
Calibration
Incinerator A Effluent Stream
1/5/2015
6.01
4.1
4.01
-0.09
6.76
6 .86
1/12/2015
6.36
4 .04
4.01
-0 .0 3
6.82
6.86
1/20/2015
6.1
4.1
4.01
-0.09
6.83
6.86
1/26/2015
6.14
4.09
4.01
-0.08
6.75
6.86
2/2/2015
6.09
3.93
4.01
0.08
6.8
6.85
2/10/2015
6.25
4.08
4.01
-0.07
6.76
6.86
2/20/2015
6.09
4.1
4.01
-0.09
6.77
6.86
2/25/2015
6.17
4 .07
4.01
-0.06
6.81
6.85
3/3/2015
6.38
4.14
4.01
-0.13
6.81
6.86
3/10/2015
6.21
4 .14
4.01
-0.13
6.72
6.86
3/18/2015
6.29
4.17
4.01
-0.16
6.67
6.88
3/23/2015
6.35
4.1
4.01
-0.09
6.68
6.86
Average adjustment to effluent pH Meter 7031A
-0 .0 8
Incinerator B Effluent Stream
1/5/2015
7.25
4.02
4.01
-0.01
6.87
6.86
1/12/2015
7.6
3.98
4.01
0.03
6.8
6.86
1/20/2015
7.45
3.99
4.01
O.Q2
6.84
6.86
1/26/2015
7.36
3.99
4.01
0.02
6.87
6.86
2/2/2015
7.13
4.04
4.01
-0.03
6 .93
6.86
2/10/2015
7.2
3.99
4.01
0.02
6.85
6.86
2/20/2015
7.24
4.01
4
-0.01
6.87
6.86
2/25/2015
7.46
4.02
4.01
-0.01
6.87
6.86
3/3/2015
7.86
4.03
4.02
-0.01
6.85
6.84
3/10/2015
8.07
3.99
4
0.01
6.84
6.86
3/18/2015
7.86
4.03
4.01
-0.02
6.87
6.86
3/23/2015
7.45
4
4.01
0.01
6.81
6.86
Average adjustment to effluent pH Meter 70318
0.00
Orange JHighlight indicates maximum adjustment for the period
111\lcti-I
Buffer 6.86 Adjustment
0.1
5.99
0.04
6.22
0.03
6.06
0.11
6.12
0.05
6.2
0.1
6.2
0.09
6.06
0.04
6.17
0.05
6.31
- -
0.14 0.21
6.25 6.47
0.18
6.86
0.10
-0.01
7.17
0.06
7.64
0.02
7.5
-0.01
8.4
-0.07
7.28
0.01
7.16
-0.01
7. 15
-0.01
8.25
-0.01
7.35
0.02
7.75
-0.01
6.86
0.05
6.86
0.00
K ELLER AND H ECKMAN LLP
Ms. Jodi Howard
April 27, 20 15
- - Page B-2 I9t:ll h Jr.':Htl
Appo~dl, ~ ' PVC Facility lncln~,ato, pH Callb'.ation R~co'."s - lnfl_u,nt St~~
" " 1 1
t
I
,,.,. - I
1,a.;w, . . . ,_,_ ,,., !'11'.'
.-w. 9':II I .
I t
Ti
1111
1- : 11 11,: I
l WT':11 1 1--:1 I
Incinerator A Influent Stream
1/5/2015
8.13
4.07
4.01
-0.06
6.88
6.85
1/12/201 5
8.61
4 .02
4
-0.02
6.8
6.85
1/20/2015
8.35
4 .09
4.01
-0.08
7.02
6.86
1/26/2015
8 .3 7
4
4.01
0.01
6 .9 5
6.86
2/2/2015
9 .06
4.12
4.0 1
-0.11
6 .76
6.86
2/10/2015
9
4 .12
4.01
-0. 11
6.71
6 .8 5
2/20/2015
7.97
4 .14
4
-0.14
6.91
6.86
2/25/201 5
8.24
4 .25
3.99
-0.26
6.95
6.85
3/3/201 5
8.4
4.02
3.99
-0.03
6.91
6.83
3/10/201 5
8.19
4.11
4
-0.11
6.84
6 .8 5
3/18/201 5
8.47
4.08
3.99
-0.09
6.97
6.86
3/23/2015
8 .2
4.19
4
-0.19
6.98
6.85
Average adjustment to Influent pH Meter 7032A
-0.10
Incinerator B Influent Stream
1/5/2015
8 .6
4.07
4.01
-0.06
6.79
6.86
1/12/2015
8 .61
4.02
4
-0.02
6.8
6.85
1/20/2015
8.4 3
4.03
4.01
-0.02
6.85
6.86
1/26/2015
8.34
4.04
4.01
-0.03
6 .9 9
6 .8 6
2/2/2015
8.83
4.09
4.01
-0.08
6.8 1
6 .8 6
2/10/2015
8.19
4.1
4.01
-0.09
6.93
6.86
2/20/2015
9.14
4.06
4.01
-0.05
6.76
6.87
2/25/2015
8.73
4.05
4.02
-0.03
6.79
6.87
3/3/2015
9 .7 7
4.1
4.01
-0.09
6 .93
6 .8 6
3/ 10/2015
8.39
4.09
4.01
-0.08
6.75
6 .8 6
3/ 18/2015
8.5
4.09
4.02
-0.07
6.77
6.86
3/ 23/20 15
8.35
4.07
4.01
-0.06
6.73
6.86
Average adjustment to influent pH Meter 7032B
-0.06
Orange Highlight indicates maximum adjustment for the period
-
, ,
t
-0.03 0.05 -0.16 -0.09
0.1 0.14 -0.05 -0.1 -0.08 0.01 -0.11 -0.13 -0.04
0.07 0.05 0 .0 1 -0.13 0.05 -0.07 0.1 1 0.08 -0.07 0 .11 0.09 0.13 0.04
. t I
8.22 8.07 8 .9 9 8.4 8.21 8.31 8.15 8.92 8.46 8.09 8.46 8. 11
7.0 1 8.07 8.86 8 .3 8 8.44 8.69 8.95 9.18 8.51 8. 22
8.3 8.51
