TRE with Append 04_08_22 red.pdf

Toxicity Reduction Evaluation

For

Industrial Wastewater Treatment System (Bioplant), Radford Army Ammunition Plant

March 2022

Revision 0

BAE Systems, Inc. Not Cleared for Export 2 | P a g e

Revision Table

Revision No.

Date Reason for Revision Author

0 March 2022 Initial document S. Andrews

BAE Systems, Inc. Not Cleared for Export 3 | P a g e

Acronyms and Abbreviations

Bioplant Biological Wastewater Treatment Plant

BMI Biological Monitoring, Inc.

BOD Biological Oxygen Demand

C. dubia or Cd Ceriodaphnia dubia

COA Certificate of Analysis

Coastal Coastal Bioanalysts, Inc.

COD Chemical Oxygen Demand

DMR Discharge Monitoring Report

EDTA Ethylenediaminetetraacetic Acid

EQ basin Equalization basin

HLB SPE Hydrophilic-Lipophilic-Balanced Solid Phase Extraction

I&I Inflow and Infiltration

LC50 Lethal concentration 50%

MC Methyl Centralite

NG Nitroglycerin

NGDS Northrup Grumman Defense Systems

NOV Notice of Violation

P. promelas or Pp Pimephales promelas

RBC Rotating Biological Contactors

RFAAP Radford Army Ammunition Plant

SPE Solid Phase Extraction

TC Total Carbon

TIE Toxicity Identification Evaluation

TKN Total Kjeldahl Nitrogen

TRE Toxicity Reduction Evaluation

TUa Toxicity Units (acute)

VDEQ Virginia Department of Environmental Quality

VPDES Virginia Pollution Discharge Elimination System

WET Whole Effluent Toxicity

BAE Systems, Inc. Not Cleared for Export 4 | P a g e

Table of Contents Background

History of Permit Excursions at Outfall 029

Information and Data Gathering

Review of Effluent Toxicity Data

Industrial Wastewater Sewer Design and Operation

Maintenance Activities

Distribution System

Facility Performance Evaluation

Bioplant Operation Evaluations

Exploratory Bioplant Sampling and Analyses

Process Evaluation of Bioplant

Nutrient Evaluation of Bioplant

Nitroglycerin Adsorption

Solids Within the EQ Basins

Coagulation/Polyaluminum Chloride (PAC)

Microbial Treatment (EQ Basins and RBCs)

Overview of the Importance of Microbial Treatment

Microscopic Evaluations

Seeding

Phosphoric acid

Source Identification/Reduction

Inflow and Infiltration

Solvents (Ethanol, Acetone, Ether)

NG

Graphite

Toxicity Identification Evaluation

Evaluation of 1st Quarter 2021 TIE Testing

Evaluation of 2nd Quarter 2021 TIE Testing

Evaluation of 3rd Quarter 2021 TIE Testing

Evaluation of 4th Quarter 2021 TIE Testing

Suspected Toxicants Investigations

Methyl Centralite

Sources of Ionic Constituents

Toxicity Control Implementation

Continuing Updates

Sources Cited

BAE Systems, Inc. Not Cleared for Export 5 | P a g e

Appendices

Appendix A. Timeline of Outfall 029 Permit Excursions

Appendix B. Laboratory COA (Reports) for Outfall 029 Acute WET toxicity Testing

2021 01 05 WET CD 029

2021 04 29 WET CD 029 Coastal

2021 04 29 WET CD 029 Pace

2021 08 24 WET CD PP 029

2021 11 16 WET CD Coastal

2021 11 16 WET CD Biological Monitoring

Appendix C. Data Profile for Lab Results, DMR Monitoring, Operational Data

Appendix D. Noncompliance COA (Reports) for Outfall 029 Toxicity Exploratory Testing

2022 02 01 CLE CD PP BMI

2022 02 01 CLE CD PP Coastal

2022 02 22 CLE CD PP BMI

2022 02 22 CLE CD PP Coastal

2022 02 22 CLE CD PP San Diego

Appendix E Operations and Maintenance Manual for Outfall 029 – Biological Wastewater Treatment

Plant

Appendix F. Bioplant Collection System

Appendix G. Process Evaluation of Bioplant Fall 2021

Appendix H. Nutrient Evaluation of Bioplant Fall 2021

Appendix I. Nitroglycerin Adsorption Testing Fall 2021

Appendix J. Graphite Testing 2021 and 2022.

Appendix K. Toxicity Identification Evaluations (TIE) Manipulation Results

Appendix L. TIE Reports

Outfall 029 TIE results Q1 2021

Outfall 029 TIE results Q2 2021

Outfall 029 TIE results Q2 2021 more testing

Outfall 029 TIE results Q3 2021

Outfall 029 TIE results Q4 2021

Appendix M. Investigative Reports

2.11.21 BAE NG Screening CD WET test

12.7.21 CLE 029 tox and NG degradation analysis

10.22.21 MC testing

BAE Systems, Inc. Not Cleared for Export 6 | P a g e

Background The Radford Army Ammunition Plant (RFAAP) operates under a Virginia Pollution Discharge Elimination System (VPDES) (Permit Number VA0000248) issued by the Virginia Department of

Environmental Quality (VDEQ). Wastewater is treated at the Biological Wastewater Treatment Plant (Bioplant) before being discharged to the New River at Outfall 029. The receiving industrial sewer lines are sourced from various areas within the plant including nitrocellulose, nitroglycerin (NG), acids, solvents, research activities, and various tenants The treatment train of the Bioplant consists of two equalization basins, rotating biological contactors (RBCs), and secondary clarification.

History of Permit Excursions at Outfall 029 Recently, the Bioplant has experienced a sequence of failing acute Whole Effluent Toxicity (WET) tests from 2020 to present date. The VPDES permit1 for the facility dictates that acute WET testing shall occur at Outfall 029 on a quarterly basis, with TUa (acute Toxicity Units) limit of 1.0. Results are reported on the Discharge Monitoring Report (DMR). The acute tests are 48-hour static tests using

Ceriodaphnia dubia (Cd). Once a year, tests are conducted using both Cd and Pimephales promelas (Pp). The test, which results in the lowest LC50 (i.e. highest TUa) is reported on the DMR. Appendix

A outlines a timeline of permit excursions and Notice of Violations (NOVs) for Outfall 029 starting in

January 2021. Appendix B contains the Laboratory Certification of Analysis (COA) for compliance toxicity analyses received in 2021. Table 1 outlines acute WET test results at Outfall 029.

Table 1. Historical TUa results for Outfall 029 from past six (6) years. Cd = Ceriodaphnia dubia, Pp =

Pimephales promelas.

Date of Toxicity

Sampling DMR Reported TUa Species

1/12/2016 < 1 Cd

4/12/2016 < 1 Cd

7/12/2016 < 1 Cd

10/4/2016 < 1 Cd

1/10/2017 < 1 Cd

4/4/2017 < 1 Cd

8/8/2017 < 1 Cd

10/3/2017 < 1 Cd

1/9/2018 < 1 Cd

4/3/2018 < 1 Cd

7/17/2018 < 1 Cd

10/2/2018 < 1 Cd

1/8/2019 < 1 Cd

4/2/2019 < 1 Cd

7/18/2019 < 1 Cd

10/1/2019 < 1 Cd

1/7/2020 < 1 Cd

4/7/2020 3.74 Cd

8/11/2020 < 1 Cd

10/13/2020 < 1 Cd

1/5/2021 4.05 Cd

4/29/2021 5.275 (average) Cd

8/24/2021 1 Cd

BAE Systems, Inc. Not Cleared for Export 7 | P a g e

8/24/2021 3.13 Pp

11/16/2021 17.085 (average) Cd

Failing WET tests prompted the facility to begin the process of conducting a toxicity reduction evaluation (TRE) to evaluate sources of effluent toxicity. A TRE is defined as “a site specific study conducted in a stepwise process designed to identify the causative agents of effluent toxicity, isolate the sources of toxicity, evaluate the effectiveness of toxicity control options, and then confirm the reduction in effluent toxicity”2. Historically in Virginia, a TRE plan would be have been required to be submitted in the situation that a permittee failed to pass acceptable toxicity criteria or permit limits.

However, DEQ currently opts to not be directly involved in the TRE process, allowing the burden of reduction and elimination of toxicity to fall on the permittee3.

Information and Data Gathering The primary data sources utilized to collection information during the TRE include the following:

- Bioplant Daily Logsheet:

o Consists of Operator input data at Bioplant to track operational/monitoring/measuring data.

o Data is recorded on a shift, hourly, and daily basis.

- Discharge Monitoring Reports (DMR):

o Submitted via online eDMR portal.

o Parameter values are tracked in spreadsheet, “Monthly eDMR Outfall Tracker.” Outfall

029 has a dedicated tab in this spreadsheet.

- Other spreadsheets, reports, records from exploratory sampling:

o As the investigation progressed, other exploratory sampling and subsequent data analyses occurred both within the treatment system of the Bioplant and within the entire collection system. The following sections and data within the Appendices explain such documents.

Appendix C (Data Profile) contains a cross-referenced list of measured parameters aligning with toxicity sampling. Note that compliance and noncompliance (exploratory) samples are designated.

A cross-functional working group was initiated during March 2021 and has been meeting on a weekly basis since that time. The team members participating in the cross-functional working group include

Safety Health Environmental (SHE) Manager, Utilities Operator in Charge, Utilities Process Engineer, Production Area Process Engineers, Environmental Manager, Environmental Water Compliance

Specialist, Environmental Specialists, Maintenance Supervisor, and local Army personnel. This meeting is an opportunity to discuss corrective actions, evaluate data/information, and plan next steps. Goals/objectives of the Wastewater Permit Excursion Working Group include the following:

- Share information and updates to help understand the possible sources and issues within the treatment systems.

- Review past corrective actions and implement new actions to address wastewater permit excursions.

- Discuss limitations and barriers to achieving our goals to better sustain wastewater treatment.

- Make recommendations that can be elevated to higher entities to help us achieve our goals.

Review of Effluent Toxicity Data Al review of every laboratory certificate of analysis (COA) report occurs upon receipt from the subcontractor laboratory. Laboratory COAs from 2021 are contained in Appendix B. When reports were received after sampling events, there were no issues identified within the quality parameters reported by the laboratories that would warrant discrediting or flagging the data reported.

BAE Systems, Inc. Not Cleared for Export 8 | P a g e

The Division of Consolidated Laboratory Services maintains an online database of accredited commercial laboratories that is searchable by parameters. Certifications for both subcontractor laboratories used to analyze compliance samples can be found in Table 2.

Table 2. Verification of laboratory certification for subcontract laboratories performing toxicity testing for RFAAP.

Lab Name VELAP

ID

Certificate

No.

Method Analyte

Biological

Monitoring

Inc. (BMI)

460015 11321

EPA 1000 (FW CHRONIC 7D S,

G)

Pimephales promelas

EPA 1002 (FW CHRONIC 3-

BROOD S, R)

Ceriodaphnia dubia

EPA 2000 (FW 48H LC50) Pimephales promelas EPA 2002 (FW 48H LC50) Ceriodaphnia dubia

Coastal Bioanalysts, Inc. (Coastal)

460030 11375

EPA 1000 (FW CHRONIC 7D S,

G)

Pimephales promelas

EPA 1002 (FW CHRONIC 3-

BROOD S, R)

Ceriodaphnia dubia

EPA 2000 (FW 48H LC50) Pimephales promelas EPA 2002 (FW 48H LC50) Ceriodaphnia dubia

RFAAP has historically used Pace Laboratories in Beaver, West Virginia to analyze compliance sampling that includes toxicity samples. Because of persistent and unacceptable performance issues, a new laboratory (Enthalpy Analytical) was contracted beginning in March 2021 with an overlap of approximately a month with both laboratories analyzing samples for RFAAP. Enthalpy uses Coastal

Bioanalysts, Inc. (Coastal) as a lower tiered subcontractor for their toxicity sampling.

Split sampling of the 4th quarter 2021 and 1st quarter 2022 samples occurred to explore the possibility of laboratory discrepancies. Toxicity results and laboratory measured parameters (measured upon receipt of sample) are listed Appendix C (Data Profile). Comparison of results from split samples helped the cross-functional team evaluate the performance of each laboratory.

Evaluation of split sampling results indicated that Coastal reported higher TUa results compared to Biological Monitoring Inc. (BMI). As a confirmatory step, a three-way analysis was conducted

February 22, 2022 on noncompliance samples from two waste streams at RFAAP. The third laboratory contracted was Enthalpy Analytical San Diego. Care was taken to ensure homogeneity of samples during pour-ups to reduce potential variability. RFAAP also dictated that all three laboratories run the same dilution series and begin testing on the same day. These steps helped to minimize possible effects from sample storage and allow us to compare WET test results in good faith. After analyzing the results from the three-way split (Appendix C and reports (Appendix D), it was determined that

Coastal results are biased high for TUa. From the laboratory measured parameters, Coastal also reported unusually high values for conductivity (Appendix C).

The results from the February 22, 2022 three-way split were not received in time to make a definitive conclusion about switching the dedicated subcontract laboratory, therefore the 1st quarter 2022 compliance sample (collected March 15, 2022) was split between both Coastal and BMI. A three-way split was attempted earlier in the month (February 1, 2022), however a major ice storm in the mid-west prevented the 3rd sample from arriving in San Diego within the allotted hold time, resulting in the samples only being analyzed by Coastal and BMI. Information gathered from all split sampling events is being compiled for a justification to use BMI as the primary lab to evaluate future toxicity samples.

BAE Systems, Inc. Not Cleared for Export 9 | P a g e

Industrial Wastewater Sewer Design and Operation Details about the processes and operation of the Bioplant can be found in the Appendix E (Operations and Maintenance Manual for Outfall 029 – Biological Wastewater Treatment Plant). There have not been any major design or operational changes/upgrades that would be identified as a causal factor for toxicity issues. This an aging facility and maintenance repairs and upgrades are happening on a continuous basis. However, there were several major maintenance improvements that were made in 2021, described in the section below. None of these improvements or upgrades directly correlated to an elimination of failed toxicity tests.

Maintenance Activities

There were several important maintenance activates that were completed in 2021 for various components of the Bioplant. The age of the facility presents challenges for completing repairs and occasionally specific parts are manufactured onsite in order to complete a repair. Below are the impactful maintenance activities completed in 2021:

- RBCs: various RBC’s located both inside and outside needed repairs and replacements, including shaft repairs, gear box repairs, gear reducer repairs, small drive sprocket repairs, electrical re-wiring with new conduit, and repairs to concrete platform.

- Recirculation pump: the P-91 recirculation pump in the EQ basin was replaced.

- Baffled effluent weirs on the gravity thickener: onsite fabrication was required to reproduce weirs, new weirs were installed.

- Variable Frequency Drive (VFD): new drives installed on recirculation pumps to increase recirculation efficiency.

Distribution System Appendix F contains a flow diagram of the collection system from all production areas that ultimately discharge to the Bioplant. The monitoring stations are indicated by yellow boxes with the station identification number. Most monitoring stations record TC (total carbon), pH, and flow with data being continuously logged. Lift stations are noted by blue boxes.

Facility Performance Evaluation During the ongoing investigation in 2021, efforts for understanding toxicity were divided into two categories: Bioplant operation evaluation and influent source identification/reduction. The overall challenge for the TRE over the course of 2021 and early 2022 has been the continued evidence that there is not one sole event or toxicant that is causing the failed toxicity tests at Outfall 029. It is important to record and track the changes that have occurred to evaluate and improve the performance of the Bioplant. Additionally, the identification of influent constituents and further control of inputs to the Bioplant is of equal importance for this TRE.

Bioplant Operation Evaluations Extensive evaluations of Bioplant functionality occurred in the second half of 2021, utilizing additional expertise from a wastewater engineer outside of the facility. Additional sampling was planned to aid in further evaluating functionality of the Bioplant. The exploratory sampling, in conjunction with in-depth evaluations of each Bioplant function, allowed RFAAP to make important decisions in 2021 to optimize treatment performance.

Exploratory Bioplant Sampling and Analyses

Multiple sampling plans and evaluations occurred in late 2021 to better understand the treatment system within the Bioplant. Sampling locations and events varied depending on the sampling plan.

Data can be found in subsequent Appendices.

BAE Systems, Inc. Not Cleared for Export 10 | P a g e

Process Evaluation of Bioplant

Process evaluation sampling involved measuring COD, available Phosphorus, and Total Kjeldahl

Nitrogen (TKN) throughout the plant (Appendix G) over the course of several weeks. The conclusions from this sampling effort determined that there is a lack of available orthophosphate. This is crucial for sustaining healthy microbial populations

- Solution: Additions of phosphoric acid, according to a recommended dosage, was added to the EQ basins. For the past several years, fertilizer has been added to the system. However, fertilizer also has the potential to add excess ammonia to the system and limited the amount of orthophosphate.

Nutrient Evaluation of Bioplant

A microbial nutrient evaluation was performed at the beginning and end treatment at the Bioplant to determine if there are deficiencies for microbial populations across the entire treatment train

(Appendix H) The most important nutrients, vital to supporting microbial communities, were present with the system. However, micronutrients may either be lacking or below detection limit of the analyses performed. If phosphoric acid additions do not result in enhanced microbial treatment, then

Iron, Zinc, and Copper levels will be further evaluated with lower detection limits.

Nitroglycerin Adsorption

Nitroglycerin (NG) adsorption to particulates was assessed at multiple upstream stations and within the Bioplant treatment train (Appendix I). It was concluded that NG is not immediately sorbing onto particulates (i.e. graphite) within the system.

Solids Within the EQ Basins

The Bioplant contains two covered concrete Equalization Basins (labeled East EQ Basin and West EQ

Basin) with a volume of approximately 3.6 million gallons each. Solid levels are controlled through the operation of recirculation pumps. There was a large volume of solids that had settled over time in the

East basin. As part of a major corrective action, subcontractors were engaged to perform a full removal of solids from the East basin. This work was completed at the beginning of January 2022 and the basin was placed back in service January 28, 2022. This improvement also supports the microbial health of the EQ basin by allowing for increased retention time.

Coagulation/Polyaluminum Chloride (PAC)

The source of PAC was investigated to see if there were any changes around the time of failed toxicity tests. No changes in the vendor/supplier were observed. Increased TSS values have not been a major issue and results of TIE testing conducted in 2021 did not reveal metals or other toxicants that would be settled out during coagulation as a causal factor for failed toxicity

Microbial Treatment (EQ Basins and RBCs)

Growth and microbial treatment is achieved in both the EQ basins and the RBCs. Microbial communities are supported in the EQ basins by aeration, phosphorus additions, and seeding/nutrient additions. RBCs operate as an aerobic attached-growth biological treatment process used to remove organic matter from the wastewater. The RBCs consist of a series of closely spaced circular disks of high-density polyethylene with carbon black added for ultraviolet resistance. The approximately twelve-foot diameter disks are slowly rotated through the wastewater, allowing the biological growth on the disks to come into contact with the organic matter in the wastewater and oxygen in the atmosphere.

The EQ basins were not designed to be biological reactors4. However, there is evidence that the majority of treatment and removal of organic constituents occurs within the EQ basins. Microbial communities are fed and maintained by the materials that are input. This is an issue when inputs vary broadly and sporadically due to the fluctuations of propellant manufacturing of many different products and the their subsequent waste streams at RFAAP. Achieving stability will always be a challenge when there are a multitude of products produced at RFAAP. The best way to counteract the

BAE Systems, Inc. Not Cleared for Export 11 | P a g e effects of these sporadic inputs are by having the most robust microbial population possible.

However, the EQ basins were not designed to be biological reactors4.

Overview of the Importance of Microbial Treatment

Because of the design of the Bioplant, review of microbial health was a major focus of improvement efforts. With proper microbial communities, nutrient support, and retention time, constituents that should undergo microbial degradation through the treatment process include solvents (ethanol, acetone, ether), Nitroglycerin, and other various compounds and chemicals present in the wastewater as the result of many different propellant production activities.

Microscopic Evaluations

In the past, observations of the microorganisms occurred but were ceased many years ago with changing personnel. A microscope was received by the facility during Fall 2021 to resume observations. Samples are collected from the Bioplant EQ basins and the RBCs on a routine basis and immediately viewed on slides for organism counts.

Seeding

Beginning nearly a decade ago, micronutrients (Microsolve 118 and Microclear M100) are added to the EQ basins on a routine basis. Exact microbial communities within Microsolv products is proprietary. Other microbial retailers also have formulations that are used in industrial wastewater treatment. Novozymes was contracted to provide additional proprietary seeding material (Bioremove COD), given the solvent-laden wastewater streams at the Bioplant. Seeding with this material began in September 2021. Later, in December 2021, an additional product was added to the seeding dosage. This product is called MB2 and was recommended to try to enhance NG treatment. A third Novozymes product, MB MN, was added beginning in January 2022 to add micronutrients to the system. The effectiveness review of these products is continuing into 2022.

Phosphoric acid

After exploratory sampling that occurred in Fall 2021, it was determined that the Bioplant system lacked available orthophosphate which is crucial to supporting microbial communities. The recommendation was made to switch from inputting bagged fertilizer to liquid phosphoric acid. A trial dosage of phosphoric acid began in February 2022. This was deemed an effective method for provided additional phosphorus inputs to support microbial health. Phosphoric acid is added at the influent pump station and flows into the EQ basins.

Source Identification/Reduction In conjunction with the evaluations and improvements made within the Bioplant treatment system, influent sources and specific chemical parameters were also researched, identified, and brought under more systematic control.

Inflow and Infiltration

Stormwater (inflow) and groundwater (infiltration), known as I&I in the wastewater industry, can have a major influencing on wastewater treatment systems. Storm events can cause increased flows to the Bioplant as well as a hydraulic overloading, that can cause a disruption in treatment for the microbial communities in the EQ basins and RBCs. Multiple efforts were completed in 2021 to address

I&I pulses to the Bioplant.

There were several major water break repairs that occurred during 2021 that were in close proximity to the intake to lift station 906-8. Prior to the repairs, the lift station pumps were running continuously for a period of approximately 3 months. After the repairs, the cycling of the pumps at this lift station returned to normal. The exact flow rates at this lift station are not measured, but are estimated to be substantial.

In early 2021, there were major repairs of the guttering system that occurred in the area of the facility that feeds Station 12. Several buildings and areas were identified with gutters that

BAE Systems, Inc. Not Cleared for Export 12 | P a g e inadvertently diverted rain into the trench system connected to the industrial sewer system. Once repairs were completed, rain was properly diverted to gutters and downspouts, leading to the stormwater system. This improvement was also not measurable, but thought to have had a substantial effect on decreasing I&I entering the Bioplant during rain events.

Solvents (Ethanol, Acetone, Ether)

Solvent-based production of energetic material occurs at RFAAP. The entire collection system and the monitoring stations within the Bioplant monitor are designed to monitor and check the solvent levels through TC (Total Carbon) monitors. Historically, the only pretreatment limits that exist for the collection have been TC limits for two areas of the plant that are known to pulse potentially high amounts of solvent-laden wastewater to the Bioplant.

Removal of organic material, including solvents, occurs within the EQ basins by microbial activity. By observing the data, it is clear that when the microbial community is suppressed, there is detected levels of ethanol, acetone, and/or ether at the clarifier. When microbial treatment is optimized, these constituents are non-detect at the clarifier. These solvents will have a direct effect on the BOD and

COD values monitored at Outfall 029 for the DMR, see Appendix C.

Although monitored pulses of solvents occur as part of the varied manufacturing schedule, there were two very large peaks that occurred in February 2022 (Figure 1). An investigation commenced into finding the sources of these pulses.. Multiple leaks of alcohol were draining into the pit system (under operational control) within one of the production buildings. Another pump/trench system in the back of the same building was correlated with these pulses based on operator interviews. Historically, the pump has not been under operational control and is switched on/off as needed by the operators.

These deficiencies were immediately corrected and more operational control gained through the use of lock/key, procedure changes, and re-training of area operators. Now both the pit system and the back pump must be sampled for TC by Bioplant operators and will be allowed/disallowed to pump to the Bioplant based on the TC results.

Figure 1. Monitoring data from station 7b measured from 2/15/2022 to 2/18/2022. The red line indicates TC (ppm) and the blue line indicates flow in gallons per minute.

More stringent TC limits were also deployed beginning in January 2022 for an onsite tenant, Northrup

Grumman Defense Systems (NGDS). This tenant has a solvent still recovery process that discharges to the Bioplant if it is not recirculated. Before discharging, NGDS must have a Bioplant operator analyze a sample for TC and the sample must be below the allowable limit. This allowable limit was further reduced in January 2022.

NG

Early in the investigation, amounts of Nitroglycerin (NG) measured throughout the Bioplant have been considered as possible toxicant relating to failing toxicity tests, however there is not a direct correlation between pulses of NG and failed toxicity tests.

BAE Systems, Inc. Not Cleared for Export 13 | P a g e

Nitroglycerin was identified as possibly creating treatment issues at low solids levels by previous research at the Bioplant4. NG has a 48-hour LC50 for Ceriodaphnia dubia at 17.83 mg/L5. Initial data reviews and reduction efforts focused on NG as a possible toxicant causing the failed toxicity tests.

However, NG is not the sole factor causing failed toxicity tests (Appendix C). There were sampling occasions with higher NG measured in the clarifier but no failed toxicity (sample dates 8/24/2021 and

1/11/2019).

The exact interactions of NG within the Bioplant treatment train are not fully understood. Although

NG can be microbially degraded, the fate of NG within the EQ basins could take several paths. The NG that enters the EQ basin could be sorbed onto solids (although exploratory sampling suggested that this does not occur immediately(Appendix I)), vaporized in a carrier solvent, or degraded into NG degradation products4.

Figure 2 shows that total summation of NG that is both entering the Bioplant (RawC) and detected in the clarifier (CLE). NG levels steadily increased in 2021 in the raw influent. There was also a shift that occurred during 4th quarter 2020 and 1st quarter 2021 in which there was more NG detected in the clarifier than what was entering the Bioplant. This would support the idea of NG sorbing/desorbing from solids within the treatment system.

Figure 2. Monthly summation data calculated from daily NG values measured from Bioplant influent (RawC) and Bioplant clarifier (CLE). Green lines indicate when a non-toxic, passing acute WET test was measured (for C. dubia). Orange lines indicate when a failed acute WET test was measured.

Currently, NG is not being readily removed from within the Bioplant system. Historical data of NG monthly sums illustrates a shift in treatment starting January 2019 (Figure 3). From 2015 to January 2019, NG was below detection limits within the clarifier Effluent. The reason for the shift has not been identified, but is under investigation. The validity of the lab results, collection methods, equipment, etc. have been questioned and explored multiple times, but there does not seem to an ancillary reason for this shift.

BAE Systems, Inc. Not Cleared for Export 14 | P a g e

Figure 3. Historical monthly summation data calculated from daily NG values measured from Bioplant influent (RawC) and Bioplant clarifier (CLE).

Early in 2021, an effort was begun to understand the sources of NG entering the Bioplant. There is an industrial sewer line that runs underneath the New River bridge. Production wastewater is collected from all areas located on the Horseshoe bend area of the plant into one lift station before being transported across the bridge. It was determined that the wastewater from the NG production area was the major source of NG-laden wastewater. In the first quarter of 2021, it was thought that a clean-out of EQ tanks located in the pretreatment system for NG-2 would reduce NG entering the waste stream. The results of pre- and post-cleanout samples indicated that this did not reduce NG entering the waste stream.

Table 3. Sampling results from NG and toxicity testing at NG-2 pretreatment areas.

Sample Point Pre-clean out NG mg/L

Pre-clean out Toxicity (TUa)

Post-clean out NG mg/L

Post-clean out Toxicity (TUa)

906-11 – EQ

tanks at NG (cleaned out 3/26/21)

52 (production shutdown)

23.62 75.9 (production running AA2)

24.1

9410 – acid pretreatment

446 No sample 313.20 No sample

In February 2021, a basic compound screening test was performed by Pace Laboratory on solutions created using RFAAP NG standards at low levels. This information would help the facility better understand if NG was the source of toxicity at low levels, see report in Appendix M. The LC50 for Ceriodaphnia dubia is 17.83 mg/L5, but the daily NG levels detected in at the clarifier throughout

2020 and 2021 typically range from 2.0 to 15.0 ppm. Therefore, the NG screening tests used concentrations of 0.5, 1.0, 2.5, 5.0, and 10.0 ppm. The results were that the solutions “passed” the screening compared to a control. One C. dubia death (out of five test organisms) was observed in the

10.0 ppm solution. NG can still have a negative effect at ranges from 10.0 ppm to 20.0 ppm that could lead to failed toxicity tests. Testing performed later in 2021, examined NG values for the day of sampling, and concluded that C. dubia deaths in the sample were likely not solely from the effects of NG (Appendix C). The numbers of death expected at the levels of NG should be much lower. This corroborates the conclusion that there is not one sole toxicant responsible for the recent failed toxicity tests at the Bioplant.

BAE Systems, Inc. Not Cleared for Export 15 | P a g e

During TIE testing that occurred in 4th quarter 2021, an aliquot of sample was analyzed at a speciality laboratory for NG and degradation products. The results can be found in the report located in

Appendix M and the table below. The toxicity of specific NG degradation products is unknown at this time. The subcontracted agreed to attempt exploratory, compound testing for degradation products.

However, acquiring standards for NG degradation products was not feasible. However, testing confirmed that there were NG degradation products present within the Bioplant treatment system in the later part of 2021.

Table 4. Results for aqueous analysis of Nitroglycerin and degradation products for the sample collected from Bioplant clarifier (CLE) on 12/7/22.

Target Analyte Units Replicate 1 Replicate 2 Replicate 3

1-Mononitroglycerin mg/L 0.98 0.98 0.99

2-Mononitroglycerin mg/L 1.15 1.15 1.16

1,2-Dinitroglycerin mg/L 7.4 74 7.4

1,3-Dinitroglycerin mg/L 3.8 3.8 3.8

Nitroglycerin* mg/L 22 22 22

*On the same date, RFAAP Main Lab measured 19.6 mg/L NG on the same sample.

Source reduction explorations are still ongoing for the NG production area.

Graphite

Graphite has previously been identified as having an impact on treatment efficiency at the Bioplant4..

In addition to settling in pipes, straining pumps, and taking up space within basins (and effectively reducing retention time), graphite can have an impact on microbial treatment within the Bioplant.

Graphite coats surfaces within the system, including the RBCs, and can inhibit biofilm attachment and growth. The coating limits movement of oxygen and nutrients through the biofilm, effectively smothering the microbial community. Graphite can also adhere to suspended microbial populations, causing settling and reducing contact with nutrients.

Due to the impact that graphite has on efficiency of the Bioplant, the cross-functional team is evaluating setting limits on the quantity of graphite allowed to be discharged from various production areas. A month-long sampling campaign was designed to evaluate if graphite concentrations could be estimated by other parameters that are easier to measure, such as TC or TSS (Appendix J). Graphite concentrations were variable, depending on the production schedule. Similar to the Cherry and

Boardman study4, TC does not have a good correlation with graphite concentrations. Moderate correlations were found in relations to TSS, depending on the production schedule of the areas that utilize graphite.

Toxicity Identification Evaluation USEPA has developed Toxicity Identification Evaluation (TIE) procedures to use as tools for TRE studies. These TIE manuals describe procedures for characterization, identification, and confirmation of effluent toxicants2.

TIE manipulations can involve a wide range of tests to determine the general classification of a contaminant (organic, divalent cation, etc.). Toxicity evaluations can be performed on any of the test organisms, and we have had TIE tests completed with both of our VPDES permitted reference organisms (C. dubia and P. promelas). The subcontracted laboratories have cited the EPA’s Method for Phase I Toxicity Characteristics Procedures for the various TIE tests that were completed in 20216.

BAE Systems, Inc. Not Cleared for Export 16 | P a g e

The results of the TIE tests performed in 2021 and 2022 are summarized in the spreadsheet found in Appendix K. The laboratory reports can be found in Appendix L. The following descriptions summarize the manipulations performed:

Initial – These are the results of the original WET test.

Baseline – Because time has passed since the initial toxicity test, a new toxicity test is performed in order to evaluate any changes that may have occurred since the original test.

Due to the timing of the tests, results of TIE manipulations are referenced back to these numbers.

SPE Column – Solid Phase Extraction Column. This column has material that will sorb many hydrophilic organic compounds and other filterable toxicants.

HLB SPE Column – Hydrophilic-Lipophilic-Balanced Solid Phase Extraction column. This type of column extraction is suspected to adsorb Nitroglycerin and more soluble degradation products.

EDTA – Ethylenediaminetetraacetic Acid can be added to the sample in varying concentrations in attempts to ‘remove’ certain metals that may be present in the sample

(EDTA complexes to divalent cationic metals such as copper, cadmium, nickel, mercury, and lead). EDTA is typically added at least 2 hours prior to testing to allow time for it to react.

Sodium Thiosulfate – Sodium Thiosulfate can be added to the sample in varying concentrations in attempts to ‘remove’ some metals and many oxidants.

Aeration – Samples are aerated for 60 minutes using carbon-filtered air.

Filtration – Samples are filtered through glass fiber filters (GFFs) with 1 µm pore size. What passes through the filters (dissolved compounds and any particles smaller than 1 µm) is evaluated with a toxicity test.

pH adjustments – pH 7 tests involve bubbling CO2 gas through the sample. ‘Hi pH’ conditions

(pH of 9.0) are created by adding sodium hydroxide to the sample. ‘pH 11’ conditions are also created by the addition of sodium hydroxide. ‘pH 3’ adjustments are made by additions of hydrochloric acid to the sample.

HDPE Cube Container, Glass, or PVC – Due to the differing sorption properties of glass vs.

plastic, different container materials were evaluated to see if contaminants would sorb to the container over a period of time.

Evaluation of 1st Quarter 2021 TIE Testing A Toxicity Identification Evaluation (TIE) study was conducted on the effluent of Outfall 029 collected on 1/26/2021 by Pace Analytical (Beaver, West Virginia). A baseline test was conducted to identify initial toxicity, followed by three focused tests in an effort to eliminate the major toxic components to yield a passing toxicity test. This would help to identify the broad constituents that may be the cause of the toxicity issues. The three focused tests were as follows:

EDTA test to detect toxicity from certain cationic metals, Sodium Thiosulfate test designed to detect toxicity from metals and some oxidants including chlorine, bromine, iodine, ozone, and chlorine dioxide, Post C18 Solid Phase Extraction test designed to detect toxicity from non-polar organics, some metals, and surfactants.

These tests were selected after conversations with the Pace Laboratory, to target the reduction of toxicity by reducing total metals and/or NG. Unfortunately, none of these manipulations reduced toxicity. The Pace report did note the detection of aluminum.

BAE Systems, Inc. Not Cleared for Export 17 | P a g e

Evaluation of 2nd Quarter 2021 TIE Testing In May of 2021, Coastal Bioanalysts, Inc. performed TIE testing on samples collected 4/29/21. The TIE procedures used were selected based on the continued suspicion that NG may be present in trace amounts and chemical analyses performed by Pace Analytical in January 2021. Results suggested that toxicity due to aluminum and NG were not likely. The report also suggested some additional testing that may shed light on potential sources of toxicity. BAE Systems moved forward with the suggested analyses. The focused tests were as follows:

Baseline – Sample toxicity increased slightly after being stored for 7 days in original HDPE cube container. This would point to a more stable compound, not a volatile compound, causing toxicity.

Filtration – No toxicity reduction.

SPE column extraction – No toxicity reduction.

pH 7 adjustment – This test was suggested based on the presence of aluminum from the Pace report from January 2021. Aluminum toxicity can be influenced by pH and hardness.

Because this test did not reduce toxicity, effects from aluminum is unlikely.

Three days additional storage in either HDPE cube container, Glass, or PVC – No toxicity reduction and no difference based on the storage container. Some toxicants can adsorb onto the plastic of bottles. The lack of sorption by the PVC trial suggests that NG and NG degradation products are not the source of toxicity, as those products tend to adsorb to PVC and would have been removed from the sample.

Coastal Bioanalysts, Inc. suggested additional testing and BAE Systems moved forward with the suggested analyses that occurred with the original sample collected 4/29/21.

Baseline “2” – Test occurred 27 days after collection. Sample toxicity decreased significantly after being stored for 27 days in original HDPE cube container. This result contradicts the result from the first round of testing in which toxicity increased slightly after 7 days of storage. Decrease in toxicity could be the result of a breakdown or off-gassing of the original toxicant. The contradiction of results could be demonstrating the existence of multiple toxicants.

SPE HLB column – Did not decrease toxicity, indicates toxicity was not likely due to organic compounds

UV-irradiation – Did not decrease toxicity, indicates toxicity was not likely due to filamentous algae or bacteria.

Mock sample – A “mock” sample is created by estimating major cations and anions and had significantly reduced toxicity compared to the 029 effluent. However, given the measured conductivity of the sample versus the estimated conductivity from the ion analyses, only 66% of the total ions present were accounted for in the mock solution (K+, Na+, Mg2+, Ca2+, Cl2-, SO4

2- and HCO3

Evaluation of 3rd Quarter 2021 TIE Testing During the 2021 third quarter toxicity analyses, Outfall 029 failed toxicity for Pimephales promelas with a value of 3.13 TUa while toxicity testing using Ceriodaphnia dubia passed with a value <1.00 TUa. Please note per the facility’s VPDES permit requirements, a test for both species must occur once a year at Outfall 029. Upon failure, RFAAP immediately engaged Coastal Bioanalysts, Inc to perform TIE testing on the effluent using P. promelas with the following tests:

1. Baseline – Did not decrease toxicity (contradicts 2nd quarter TIE testing).

2. Aeration – Did not decrease toxicity

3. Filtration – Did not decrease toxicity

4. EDTA chelation test – Did not decrease toxicity

5. pH 7 test – only slightly increased toxicity

6. pH 9 adjusted – Did not decrease toxicity

BAE Systems, Inc. Not Cleared for Export 18 | P a g e

7. HLB SPE – Reduced toxicity significantly.

8. Thiosulfate reduction test – Did not decrease toxicity

9. pH 3 adjusted – slightly decreased toxicity, but was not statistically significant

10. pH 3 aeration – slightly decreased toxicity, but was not statistically significant

11. pH 3 filtration – Did not decrease toxicity

12. pH 3 HLB SPE - Reduced toxicity significantly.

13. pH 11 adjusted – Did not decrease toxicity

14. pH 11 aeration – Did not decrease toxicity

15. pH 11 filtration – Did not decrease toxicity

16. pH 11 HLB SPE - Reduced toxicity significantly.

The HLB SPE columns removed toxicity completely at pH 3, pH 11, and ambient pH. This was the first manipulation that achieved drastic toxicity reduction since RFAAP began exploratory testing in January 2021. Therefore, we knew that the toxicant in this sample was likely an organic compound, and not an easily filterable toxicant (because the physical filtration did not reduce toxicity). It must be noted that these manipulations were tested using P. promelas. All previous TIE tests in 2021 were conducted using C. dubia.

The extracted contents from the columns were submitted to the laboratory to analyze for semi-volatile compounds (SVOC) using GCMS (Method SW-8270) with a library search. The compound identified in all three (pH 3, pH 11, and ambient pH) columns was 1,3-Dimethyl-1,3-diphenylurea, also known as Methyl centralite. Methyl centralite is used as a coating agent at RFAAP for only one particular propellant product. RFAAP began an investigation into this possible toxicant. bis(2- Ethylhexyl) phthalate was also reported from the library search, but can be a common laboratory contaminant.

Evaluation of 4th Quarter 2021 TIE Testing It was suspected that Coastal was providing WET results that were biased high compared to other subcontracted labs. The suspicion was mostly centered around testing of a separate wastewater stream (acidic) and separate Outfall (007) located at RFAAP. Because of these suspicions, the 4th quarter 2021 compliance sample was split between two laboratories. When the reports were received, Coastal reported a higher TUa for Outfall 029. But with no obvious deficiencies in their report or analyses, we used their result in a calculated average to report on the DMR.

Based on the previous results from 3rd quarter, Coastal suggested trying to find a laboratory to analyze NG degradation products and performing a subsequent WET test with multiple dilutions.

Nitroglycerin and degradation products would not have been detected in the library scan. Although the onsite RFAAP Main Lab analyzes NG, the presence or amount of NG degradation products within the Bioplant system are not fully understood. Coastal also suggested reanalyzing with both C. dubia and P. promelas to attempt to gain more understanding regarding the differences in response of the two species.

C. dubia baseline – This test was run 16 days after sample collection after storage in an HDPE cube container. Although the initial test was very toxic (23.98 TUa, LC50 4.17), the toxicity greatly increased during the storage time to 69.93 TUa (LC50 1.43). This contradicts the results from the 2nd quarter 2021 TIE test when a storage time of 27 days decreased toxicity.

P. promelas baseline – This test was run 16 days after sample collection after storage in an

HDPE cube container. When tested with this species, toxicity was greatly reduced at 3.30 TUa

(LC50 30.3). This contradicts the result from 2nd quarter 2021 TIE testing when toxicity was greater with P. promelas compared to C. dubia.

Suspected Toxicants Investigations The data from the TIE manipulations was not able to provide conclusive evidence for a specific toxicant, except for the TIE testing conducted in the 3rd quarter of 2021. This testing showed a significant decrease in toxicity when the effluent was filtered through an HLP SPE column. The

BAE Systems, Inc. Not Cleared for Export 19 | P a g e column material was extracted and the elutant run through a library scan in an attempts to identify organic constituents in the elutant that could be toxic. The scan had detections of Methyl Centralite

(MC) and bis(2-Ethylhexyl) phthalate. The phthalate can be a common lab contaminant derived from plastic. Phthalates as a group have not been eliminated from investigation consideration because some production areas utilize phthalates. Phthalates that are part of the SVOC suite (EPA Method SW-

846 8270) have been below detection limits during previous evaluation. Further exploration of this group of contaminants will be considered after other avenues have been exhausted.

Methyl Centralite There is a lack of information on MC in the current literature and toxicity studies are not available for this compound (MC is not included on EPA’s ECOTOX website). BMI laboratory was contracted to determine an LC50 for MC during late 2021. The report from this testing can be found in Appendix M.

The LC50 for C. dubia was determined to be 297 mg/L and P. promelas was determined to be 277 mg/L. It should be noted that MC is not very soluble in water, but could be more readily mobilized in the presence of solvents. MC was mixed with strong alcohol and alcohol were also tested during this initiative. Because the source of MC entering the industrial wastewater system was traced back to a particular coating activity, the discharge from two separate production areas (sweetie barrel discharge and water dry discharge) associated with this activity were tested. Table 5 lists the LC50 for each source that resulted from this exploratory testing. The only source that was identified as having acute toxicity was the water dry discharge sample. The next exploratory steps will focus on what constituents in this waste stream are causing toxicity. The production of this product and others at the facility use potassium nitrate, potassium sulfate, and potassium perchlorate. It is possible that these constituents, through chemical reactions occurring within the wastewater system, are producing the ions that were essentially “missing” from the mock sampling TIE testing. Further research and evaluation are needed to explore this possibility.

Table 5. Results from acute toxicity compound testing conducted in October 2021.

Material tested for acute toxicity

Organism LC50 (mg/L) Confidence Intervals

(mg/L)

Sweetie Barrel Discharge C. dubia 2608 2106-3230

P. promelas 4910 4058-5941

Methyl Centralite C. dubia 297 254-348

P. promelas 277 236-324

Methyl Centralite with 0.2% Strong Alcohol

C. dubia 330 301-362

P. promelas 218 189-251

Strong Alcohol C. dubia > 10000 NA

P. promelas > 10000 NA

Water Dry Discharge C. dubia 16.60% 14.5-19.0%

Sources of Ionic Constituents The mock sample composing lab stock solutions that tried to recreate the ratio of commonly measured ions did not prove to be toxic.The “missing” 33% of ions that were not accounted for in the test could possibly contribute to failed toxicity results. Additional exploratory sampling and product source investigation will be ongoing in 2022.

Toxicity Control Implementation Evidence from the data review and TIE testing revealed that there is not one sole toxicant responsible for failed toxicity tests. Therefore, a holistic approach has been utilized to both control constituents

BAE Systems, Inc. Not Cleared for Export 20 | P a g e entering the Bioplant and to bolster the treatment capabilities of the existing systems within the Bioplant. When viewing all the compliance and monitoring data for particular sampling events, it is evident that multiple constituents increased over the course of 2021.