Update on the Engineering Design Studies Evaluated in the NRC Report, 'Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot': Letter Report (2002)

Advisers to the Nation on Science, Engineering, and Medicine

National Academy of Sciences

National Academy of Engineering

Institute of Medicine

National Research Council

When the Committee on Review and Evaluation of Alternative Technologies for the Demilitarization of Assembled Chemical Weapons: Phase II (ACW II Committee) finished the National Research Council report Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot (hereinafter referred to as the Blue Grass EDS Report) in September 2002, several engineering design tests had not been completed in time to be included (NRC, 2002). In this letter report, the committee provides its assessment of the results of these tests, which have since become available, before the Defense Acquisition Board completes its technology selection process and reaches its Record of Decision.

As reported in the Blue Grass EDS Report, the Assembled Chemical Weapons Assessment (ACWA) program is considering three technologies as alternatives to incineration for the destruction of the assembled chemical weapons at the Blue Grass Army Depot in Richmond, Kentucky (NRC, 2002):

• The AEA SILVER II™ process (with participant CH2MHill) destroys agents and energetic materials by oxidation with Ag²⁺ generated electrochemically. Metal munition parts are washed with dilute nitric acid and sent to a metal parts treater along with undissolved fuzes for decontamination to a 5X level. Dunnage and DPE suits are decontaminated to a 5X level in a continuous steam treater.

• The General Atomics Total Solution (GATS) uses hydrolysis with water or caustic to destroy the agents and the energetics, followed by supercritical water

oxidation (SCWO) of the resulting hydrolysates. All dunnage is slurried and also passed through a SCWO reactor. Metal parts are treated to a 5X decontamination level by an electrically heated discharge conveyor.

• The Eco Logic technology package (with participants Foster Wheeler, Kvaerner, and El Dorado Engineering) also uses hydrolysis with water or caustic to destroy the agents and the energetics, followed by supercritical water oxidation of the resulting hydrolysates in a transpiring wall-SCWO reactor (TW-SCWO) developed by Foster Wheeler. Metal parts and dunnage are decontaminated to a 5X level, and gaseous effluents from the hydrolysis processes and the thermal reduction batch processors are treated in the gas-phase chemical reduction (GPCR™) system.

While the statement of task for this letter report asks the NRC to assess only the outstanding test results of the Eco Logic technology package, the committee decided to include an analysis of results from the AEA tetrytol testing (also not completed in time for the Blue Grass EDS Report) so that its evaluation of the AEA technology is complete. Mention is made of the General Atomics technology insofar as the results from the TW-SCWO have a bearing on the General Atomics SCWO operation.

Table 1 lists final reports received after January 2002 when the ACW II Committee concluded its data gathering for the Blue Grass EDS Report. The final versions of the first three reports listed in Table 1 did not change significantly from the draft versions that were used by the committee in the evaluations presented in the Blue Grass EDS Report.¹

The only test results from the AEA technology package addressed in this letter concern the destruction of tetrytol with the 12-kW SILVER II™ system. However, these results do not alter the committee’s evaluation and findings for this technology, presented in the Blue Grass EDS Report (NRC, 2002).

The test results for the evaporator/crystallizer/filter press and the TW-SCWO units of the Eco Logic technology do lead the committee to further findings, presented in this letter report. Moreover, the TW-SCWO reactor test results indicate that the problems encountered could also be relevant to the scaling up of the General Atomics SCWO reactor design from the 3.5-inch diameter of the test reactor to the proposed 18-inch diameter of the full-scale reactor.

In presenting this update, the ACW II Committee first reiterates its overarching observation that hydrolysis with water or caustic is capable of destroying 99.9999 percent of the chemical agent and, in combination with secondary treatment steps, 99.999 percent of the energetic materials. However, the resulting hydrolysates require additional

treatment to destroy the Schedule 2 compounds that are specified in the Chemical Weapons Convention treaty.

In early 2002, the Program Manager for ACWA (PMACWA) authorized a test of the destruction of tetrytol (TNT and tetryl) to validate a destruction and removal efficiency (DRE) of 99.999 percent for this energetic material in the AEA 12-kW EDS test unit (AEA, 2002). The purpose of the test was to demonstrate continuous operability, reliability, and maintainability during tetrytol destruction (AEA, 2002). The test unit ran 134.5 hours at steady state and 41.2 hours in a rundown mode to simulate the polishing operation. During this time, 101 kilograms of tetrytol were fed to the test unit and a DRE of 99.997 percent was achieved for tetryl and 99.987 percent for TNT. Because the full-scale design incorporates additional processing steps and better condensation of off-gases, AEA estimates and the committee analysis agrees that the adjusted DREs would be in excess of 99.999 percent for both TNT and tetryl.

During steady-state operation with a continuous feed of tetrytol to the anolyte vessel, the electrochemical efficiency was maintained at 40–50 percent. As would be expected, the efficiency rapidly declined to approximately 4 percent once the feed was stopped and then gradually to about 1 percent as the organics concentration dropped. Test results also indicate a small transfer of organics from the catholyte to the anolyte toward

the end of the rundown period, when the anolyte organic concentration dropped below the catholyte concentration.

The equipment operated with no significant failures or problems. The hydrocyclones removed approximately 85 percent of the solid energetics material from the liquid feed to the cell and successfully prevented plugging of the cells; incidentally, they also reduced the cell electrochemical efficiency by lowering the organics concentration. At the end of the run, the cell membranes, which now have spacers and a membrane support grid of polytetrafluoroethylene as a result of earlier tests, were found to have only a small amount of swelling, and no silver was seen on the cathodes, as had been observed in the earlier dimethyl methylphosphonate test. The committee notes that none of the EDS II tests were run with the main circuit and impurities removal system operating concurrently.

Finding (Blue Grass Ltr) 1. Based on the 12-kW system-processing test with tetrytol, and assuming that the additional processing steps will be incorporated into the full-scale design, the committee agrees that the AEA SILVER II™ technology should be able to achieve the required tetrytol destruction efficiency of 99.999 percent.

Finding (Blue Grass Ltr) 2. Although the 12-kW test unit operated reliably during the test with tetrytol, the committee continues to believe that the complexity of the full-scale plant still poses major maturation challenges that might not allow consistent achievement of similarly reliable overall performance. The findings in the Blue Grass EDS Report concerning the AEA SILVER II™ technology remain valid (NRC, 2002). See the discussions of agent and energetics destruction in that report, as well as the text following General Finding (Blue Grass) 4, which gives seven specific reasons in support of the committee’s conclusion that major difficulties related to maturity will be encountered.

One part of the Eco Logic EDS II study plan that was not completed prior to the issuance of the Blue Grass EDS Report was the effluent evaporation testing for the Foster Wheeler TW-SCWO (Eco Logic, 2001a; NRC, 2002). The overall objective of the testing was to generate design data for the evaporator/crystallizer/filter press system that would be employed in a full-scale Blue Grass facility (Foster Wheeler, 2002a).

Aquatech International Corporation performed three bench-scale tests (laboratory boil-up tests) with liquid effluents from the TW-SCWO reactor that had been generated during each of the 500-hour GB, VX, and HD operability tests. The specific objectives were to (1) determine the chemical and physical properties of each effluent, (2) observe and record the physical characteristics of the effluent during evaporation, and (3) observe and record the characteristics of the solid residue once it had cooled to room temperature.

The boil-up experiments indicated that foaming and bumping² did not occur in any of the samples. Fouling (or salt buildup) did occur at the liquid/air interface. However, the residual solids at this interface were easily washed back into solution after the tests were completed.

In February 2002, a pilot-scale test of the forced circulation evaporator/crystallizer/filter press system was carried out at the Dugway Proving Ground simultaneously with the 500-hour VX TW-SCWO test. The pilot plant was operated in a batch mode on a slipstream of the TW-SCWO effluent. The major objectives were to (1) confirm the viability of the evaporator/crystallizer/filter press system design presented in the Eco Logic engineering design package (EDP) (Eco Logic, 200 1b), (2) determine the maximum achievable degree of concentration of the effluent, (3) measure heat transfer performance, (4) measure filterability of the crystals in the slurry, (5) determine the physical characteristics of the filter cake and filtrate, and (6) optimize the operating parameters of the filter.

A total of 221 hours of pilot plant operation with 5,924 kilograms (13,060 pounds) of feed to the evaporator produced 100 kilograms (220 pounds) of solids. The operation was interrupted for 4.5 hours after 110 hours to correct a problem with a liquid level control in the hot well of the unit. Although most of the hot well was free of any solids accumulation, the lower-level transmitter was 75 percent covered with salt, which is thought to have caused pump cavitation during the run. Plugged sample ports and pressure gauges were cleaned during the shutdown and the process restarted for the remaining 111 hours of operation. Foaming was not observed, and suspended aluminum oxide solids from the caustic dissolution of aluminum fuze end caps and rocket casings did not cause any problems. Any fouling or salt precipitation that occurred within the heat exchanger was easily cleaned with water flushing, and the effectiveness of the cleaning was confirmed by heat-transfer coefficients determined from the temperature change and other measurable parameters before and after the test. Also, the heat-transfer coefficients reached a steady level during the two individual runs of the evaporator.

During the pilot testing with the VX TW-SCWO effluent, scaling or fouling of the evaporator was not significant and could be removed with periodic flushing cycles to dissolve the precipitated solids. The pilot test run for VX TW-SCWO effluent was not long enough to determine the required frequency of the flush cycles, but based on the limited results of the boil-off test, the behavior of the heat-transfer coefficient during the test, and the similarity of the results for the first 110 hours and the second 111 hours of operation, the technology provider estimates that once a month will be adequate.

The technology provider had anticipated less water remaining in the salt cake than was observed in the test, and this led it to modify the EDP material balance. Problems were experienced with the filter press, because although it was the smallest available commercial unit, it was still too large for the low flow rates used in the pilot test. The filter incorporates an air drying step in which air is blown through the filter cake to remove interstitial filtrate. Because a relatively small amount of solids was being filtered,

some of the air bypassed the solids, leading to high water concentrations in the salt cake. Otherwise, the cake quality was good, indicating an acceptable filtering technique. Although the evaporator system was a commercial unit, the materials of construction were not described (Foster Wheeler, 2002a). Other than mentioning the absence of significant fouling, the report provided no other information relevant to the condition of the heat exchanger or other parts of the system after testing operations.

The committee continues to have concerns about corrosion, particularly when chloride ions are present. Also, the pilot evaporator was a single-effect unit, while the full-scale plant evaporator will have a double-effect unit. Foster Wheeler claimed that operation of the single-effect pilot evaporator was adequate to meet the test goals, and the committee concurs.

Finding (Blue Grass Ltr) 3. While normal scale-up processing issues remain for the evaporator/crystallizer/filter press system, the process appears acceptable for incorporation into a final plant design for the disposal facility at Blue Grass. Concerns about corrosion and salt management would need to be addressed during development of the full-scale design.

Recommendation (Blue Grass Ltr) 1. Further investigation of potential corrosion in the evaporator/crystallizer/filter press system should be carried out to identify effective materials of construction for the full-scale operation.

Only a draft of the long-term test results for the TW-SCWO reactor was available when the data-gathering activities for the Blue Grass EDS Report were completed in January 2002. At that time, the committee was very concerned about the observed presence of carbon monoxide (CO) and hydrocarbon (HC) spikes in the effluent from the SCWO reactor and the likelihood of scale-up problems.³ This led the committee to the following findings regarding the TW-SCWO reactor in its Blue Grass EDS Report (NRC, 2002):

Since then, Foster Wheeler has completed long-duration tests of the TW-SCWO reactor, mostly with hydrolysates of VX and HD simulants but also with hydrolysate of actual VX agent, and final reports have become available (Foster Wheeler 2002b, 2002c, 2002d).

The EDS II optimization testing of the TW-SCWO technology in support of the Eco Logic EDP for Blue Grass was performed during March and April 2001 (Foster Wheeler, 2001a, 2001b). This testing was performed using the TW-SCWO system at Dugway Proving Ground that had previously been employed for the ACWA Demo II testing program in 2000 (Foster Wheeler, 2001 a). Following the optimization tests, the TW-SCWO system was modified for 500-hour operability tests on the three agent/energetic hydrolysates listed in Table 2.

Hydrolysates of agent simulants were used for most of the GB and VX test campaigns and for the entire HD campaign. Some GB and VX agent hydrolysates were also used during the campaigns. The modifications to the TW-SCWO system for the GB campaign were these:

• use of oxygen (rather than air) as the oxidant

• design and construction of a new TW-SCWO reactor with an inside diameter of 6.07 inches and a length of 61 inches

• modification of the reactor outlet piping to process high levels of solids in the reactor effluent

• modifications to the reactor feed system

• installation of a new, online process analyzer for analyses of total organic carbon (TOC) and Schedule 2 compounds in the liquid effluent that was more sensitive and accurate than the analyzer used in the Demo II testing (Foster Wheeler, 2002e)

The following additional changes were made to the reactor for the 500-hour VX campaign:

• The upper outer shell of the reactor was replaced with the shell used during the Demo II tests, but with a new transpiring-wall liner installed. This liner contained three additional nonfunctional dummy platelet layers at its inner diameter to allow for additional corrosion.

• The injector was replaced with the injector from the Demo II and EDS optimization testing, which was modified to accommodate oxygen instead of air as the oxidant.

For the final 500-hour campaign with HD simulant hydrolysates, three additional changes were made to the system:

• The injector used in the GB campaign was reinstalled in the reactor. Foster Wheeler judged this injector to be in better condition than the injector employed in the VX campaign.

• A degassing system was installed in the reactor outlet piping to enable accurate measurement of the effluent flow.

• A static mixer was installed in the liquid feed line immediately upstream of the reactor to mix the hydrolysate and the propylene glycol (PG) fuel. Foster Wheeler thought that this would alleviate the HC and CO spiking.

The operability testing during the GB campaign was begun with a hydrolysate flow rate of 250 lb/hr instead of the planned 500 lb/hr rate because the installed oxygen supply system was found to be inadequate. CO and HC concentration measurements, which normally would each be 50–100 ppm, spiked to 500–1,000 ppm at 5 to 15 minute intervals in both cases. Each spike lasted about a minute. When the flow rate was increased to 500 lb/hr, the spikes increased in both frequency and magnitude. A 50 percent increase in the quench water flow at the bottom of the reactor appeared to reduce the spiking problem, possibly by improving smooth salt transport.

One way in which Foster Wheeler attempted to improve the stability of the TW-SCWO system was by continuing to use the start-up flow of hot water (80 lb/hr) and

kerosene (12 lb/hr) during normal operation. This change was begun during the 500-lb/hr workup runs to resolve the loss-of-reaction problems that were attributed to the low reactivity of the PG. Subsequently, Foster Wheeler suggested a change from PG to hexylene glycol. At some point, nitrogen flow was increased to improve system stability and purging.

During the initial VX campaign workup, the hydrolysate and PG were premixed in the feed tank, but the spiking was not alleviated. Nevertheless, in subsequent testing of VX and HD hydrolysates, a portion of the PG was mixed with the hydrolysate in the feed tank. The rest of the needed PG was introduced directly into the reactor. When hexylene glycol was temporarily substituted for PG, no difference in the spiking was observed, so PG was used in subsequent testing during the VX and HD campaigns.

Other changes made for the VX testing included (1) reducing the quench water flow to the design value, (2) increasing the size of the feed strainer, (3) decreasing the diameter of the piping from the reactor outlet to the pressure control valve, and (4) increasing the nitrogen flow to increase the flow velocity in the piping.

During workup, because it was observed that HC spiking was less for hydrolysate feed rates below 350 lb/hr, further tests were conducted at this reduced flow rate. The 350 lb/hr of hydrolysate was premixed with 50 lb/hr PG for a total of 400 lb/hr. After these changes, Foster Wheeler reports that the CO spikes were “considerably alleviated” (Foster Wheeler, 2002c). The hourly average of measured HC was reported to be consistently less than 100 ppm except for occasional excursions between 100 and 200 ppm. The CO spikes were one-tenth to one-fifteenth as high as the HC spikes.

For the tests with HD hydrolysates, the feed rate was reduced from 300 lb/hr to 210 lb/hr to reduce corrosion rates. The hydrolysate feed was premixed with 35 lb/hr of PG, for a total feed rate of 245 lb/hr. The amount of excess oxygen was also increased to improve mixing of oxygen with the feed hydrolysate below the injector and to increase the effluent velocity from the reactor to the pressure control valve. The use of a static mixer installed in the feed hydrolysate line just upstream of the reactor produced no discernible effect on HC and CO spiking. The 10-hour rolling average concentration for the HC spikes for the last 400 hours of operation was only a few hundred ppm. Foster Wheeler claims that most of the spikes were due to instrument calibrations. Accordingly, the committee believes that operability and permitting issues associated with spiking remain to be resolved. Each issue and its applicability to the TW-SCWO as well as the GATS SCWO reactor are discussed in the following paragraphs.

Foster Wheeler has tried several approaches to solving the HC and CO spiking problem experienced with the TW-SCWO system of the Eco Logic technology package. These include changes in the feed composition and preparation, the feed rate, the feed strainer size, the quench water flow, and the configuration of the reactor outlet piping. Some of these changes alleviated the spiking, but overall the problem appears to have been inadequately resolved. The committee concludes that Foster Wheeler has not clearly

identified the root cause or causes of the problem. The cause may lie within the TW-SCWO reactor itself and might be resolvable only by developing a better understanding of the heat transfer, mixing, fluid flow, and chemical kinetics inside the reactor. A good understanding of these phenomena is essential to the design and operation of any larger, higher-throughput SCWO unit, whether it be the solid wall configuration used by General Atomics, the transpiring-wall configuration used by Foster Wheeler, or any other configuration.

The committee expects that a better understanding of the spiking problem will require additional testing. For example, the spiking might be caused in part or entirely by the buildup and periodic sloughing off of salts in the reactor or the outlet piping. As the salts pass unevenly through the pressure control valve, they might cause pressure fluctuations, which in turn would cause the observed HC and CO spiking. Foster Wheeler expects that changes to the outlet piping in the full-scale system, including the installation of the pressure control valve directly under the reactor, will eliminate or further reduce the spiking problem, but this is unproven.

Foster Wheeler recently reviewed its TW-SCWO effluent HC and CO measurements and carried out additional data reduction (Foster Wheeler, 2002f). One-hour averages for the HC and CO concentrations in the TW-SCWO effluent were calculated after removing the instrument calibration periods. The resulting HC averages were in the tens to hundreds of ppm, and the CO concentrations were considerably lower. DREs for total organics content ranged from 98.88 to 99.96 percent. DREs for Schedule 2 compounds were greater than 99.9999 percent for the GB and VX runs and 99.61 percent for the HD run. After a review of these results, Foster Wheeler decided to install a catalytic oxidizer downstream of the SCWO reactors to complete the oxidation of HCs and CO.

The lack of understanding of the cause or causes of the spiking problem raises questions about the use of the TW-SCWO reactors in the Eco Logic technology package at a Blue Grass facility, even if they are used at the size and throughput rates recently tested. Foster Wheeler’s plans to scale up the reactors, with the inside diameter increasing from 6.07 inches to 11 inches, will probably exacerbate the problem.

The CO and HC spiking may impact the permitting of the TW-SCWO system by regulatory agencies. A SCWO treatment unit would most likely be permitted as a miscellaneous treatment unit under RCRA regulations at 40 CFR Part 264, Subpart X. Subpart X permit provisions are tailored to be technology- and site-specific, thereby enabling new and innovative technologies to receive RCRA permits. For those Subpart X units resembling conventional units, the permit may incorporate performance requirements from one or more of the Subparts I through O, W, AA, BB, CC, and DD, plus Part 266, Subpart H.

Emission standards for a SCWO unit could be drawn from one or more of the following: 40 CFR Part 266, Subpart H, and/or 40 CFR Part 264, Subpart AA (process

vents), Subpart O (hazardous waste combustion maximum achievable control technology standards), or Subpart X (risk criteria, 40 CFR Part 264.601). 40 CFR Part 264, Subpart AA, contains the applicable standards that establish limits on total HC emissions from process vents. It is anticipated that the permitting authority would select performance criteria and establish operating and emission limits based on the following:

• Standards at 40 CFR Part 264, Subpart AA, for emissions of volatile organic compounds in vent gas require that such emissions be reduced to below 3 lb/hr and 3.1 tons/yr or reduced by 95 percent through the use of a control device. More stringent additional limits for volatile organic compound emissions will be evaluated using site-specific risk analysis.

• Standards at 40 CFR Part 264, Subpart O, for HC and CO emissions from the exhaust of hazardous waste combustors require that emissions of total HC be limited to 20 ppm corrected to 7 percent O₂ on a 1-hour rolling average, and that emissions of CO be limited to 100 ppm corrected to 7 percent O₂ on a 1-hour rolling average.

The emissions spikes of HC and CO may or may not be significant relative to permitting and health risks. The final permitting approach and the applicable and relevant rules used can only be completely defined by the permitting authorities when they review the permit application. There is some concern on the part of the committee that current performance will need to be improved in order to meet HC emissions standards if the hazardous waste combustor standards are applied. It will also be important to quantify HC emissions and speciation in order to evaluate the potential health impacts.

The corrosion pattern in the 500-hour HD test was considerably different from the pattern in the prior GB and VX test campaigns. Most of the corrosion in the upper liner was toward the bottom rather than the top, and the lower liner now experienced a very high corrosion rate. In the GB and VX tests, there had been practically no corrosion in the lower liner. Similar results had been obtained during the optimization testing in 2001 (Foster Wheeler, 2001b). Corrosion of the lower liner had been observed during testing of the HD hydrolysate simulant but not during testing of the GB and VX simulants.

At the beginning of the 500-hour HD test the lower liner was the same one used during the GB and VX campaign 500-hour tests, where it was exposed to about 1,400 hours of operation. No corrosion was visible at the beginning of the HD test. This liner was constructed with three 0.004-inch-thick dummy platelet layers and five 0.010-inch-thick inner platelet layers inside the critical metering platelet layer. The total liner thickness was 0.164 inch.

At the 176-hour shutdown, there were some local regions of corrosion in the lower liner. In the corroded areas, all three dummy platelet layers had corroded away, to a total depth of 0.012 inch. At the 268-hour shutdown, the corrosion had spread, but the

maximum depth of corrosion still appeared to be 0.012 inch. At the end of the 500-hour test, the liner corrosion was more extensive. It was heaviest near the top of the lower liner. In the worst area, the three inner dummy platelet layers (each 0.004-inch thick) and three of the five inner platelet layers (each 0.10-inch thick) had corroded to a total depth of 0.042 inch. This level of corrosion leaves a 0.020-inch margin of corrosion before the critical metering platelets in the platelet stack become exposed. The technology provider prefers a 0.030-inch margin and would add three more 0.004-inch dummy platelet layers to the liner used for disposal of the H at Blue Grass.

The TW-SCWO reactor design has also been modified to five liner segments of equal length rather than a short upper liner and a longer lower liner. Foster Wheeler believes that this will simplify disassembly and liner replacement and provide for better support of the liners.

At the higher throughputs utilized in the latest testing with HD hydrolysate, the corrosion profile has moved down the reactor to the bottom of the upper liner and the top of the lower liner. The corrosion rate with HD hydrolysate is also considerably higher than with the other hydrolysates. Because the HD hydrolysate contains chlorine and the GB and VX hydrolysates do not, it may be that the presence of chlorine contributes to the accelerated corrosion. Accordingly, the design has been changed to increase corrosion allowances (more layers of dummy platelets) and accommodate periodic (500-hour) liner changeouts. The Eco Logic design for Blue Grass now includes five TW-SCWO reactors, one of which is a spare, so that one reactor can undergo a liner change while the others are running, avoiding an overall process shutdown. This seems like a reasonable strategy if corrosion during disposal operations at Blue Grass proves to be at about the same level as experienced in the test run. However, given the unknown factors that affect corrosion rate, scale-up of the reactor could increase this corrosion rate further, exacerbating the maintenance requirements.

Finding (Blue Grass Ltr) 4. The technology provider has proposed several explanations for the spiking in hydrocarbon and carbon monoxide levels in the TW-SCWO reactor effluent, and it has made several changes that appear to have reduced the severity of the spiking but not eliminated it. Based on its members’ experience, the committee believes that these spikes are symptoms of an underlying instability that is not yet understood. This problem needs to be thoroughly understood and resolved before the TW-SCWO reactor is implemented in the larger size and at the increased throughput proposed for Blue Grass.

Recommendation (Blue Grass Ltr) 2. Foster Wheeler should present convincing evidence that it has identified the root cause or causes of the spiking problem before scaling up the design of the TW-SCWO reactor.

Finding (Blue Grass Ltr) 5. The results of the Foster Wheeler TW-SCWO reactor tests indicate that making changes in the design or operating conditions (such as flow rate of feed or choice of oxidant) of a SCWO reactor is not straightforward. Therefore, implementation of full-scale SCWO reactors with different operating conditions or designs may result in unexpected performance problems.

As another example, General Atomics is planning to scale up its solid-wall reactor from a 3.5-inch diameter and 6-foot length to a full-scale 18-inch diameter and 18-foot length. This is a 3-fold linear scale-up, a 26-fold cross sectional area scale-up, and an 81-fold volumetric scale-up. The committee believes that this degree of scale-up introduces uncertainties that could also lead to problems and instabilities other than or akin to those experienced by Foster Wheeler with the TW-SCWO reactor.

Finding (Blue Grass Ltr) 6. Based on its experience and knowledge of permitting regulations, the committee believes that the hydrocarbon and carbon monoxide spiking may affect permitting of the SCWO process. The final permitting approach and the applicable and relevant rules can only be defined by the permitting authorities upon review of the permit application. The current performance may need to be improved to meet hydrocarbon emissions standards if the hazardous waste combustor standards are invoked.

Recommendation (Blue Grass Ltr) 3. Consideration should be given to establishing a performance standard for hydrocarbon and carbon monoxide emissions from any SCWO reactor system selected for implementation at Blue Grass. The basis for selection of these performance standards should be discussed with permit writers to ascertain likely permitting issues. While the committee first recognized this concern as a result of the Foster Wheeler EDS II testing, it is potentially a concern for any SCWO system or other oxidation system scaled up to a size appropriate for treatment of the Blue Grass stockpile.

Finding (Blue Grass Ltr) 7. Significant corrosion of the lower liner was experienced during testing on HD hydrolysate but not during testing on the GB and VX hydrolysates. This may be due to chlorine, which is present in the HD hydrolysate but not in the GB and VX hydrolysates. While the amount of H at Blue Grass is relatively small compared with the other agents, increased maintenance will be required during its destruction campaign.

All SCWO testing experience to date indicates that the operating environment associated with the SCWO reactions poses corrosion and material durability challenges per se, regardless of the SCWO process. Therefore, the committee believes that use of the SCWO process will require that the operator have an aggressive monitoring program in place to ensure that planned completion schedules are not severely compromised by higher than expected maintenance and repairs.