Silent Discharge Plasma (ENV America Inc at McClellan)

Site and Contaminants Description

The Silent Discharge Plasma technology (SDPT) was demonstrated at the McClellan operable unit D, SVE system for 11 weeks from November 2,1995 to January 12, 1996. The location of the SVE system was placed over an area where 10 waste pits were used for waste solvent and fuel disposal from the early 1940’s to 1981. The waste disposal zone is estimated to extend from approximately 15 to 28 feet bgs. The main off-gas treatment system for this SVE network is a catalytic oxidation unit followed by a particulate and acid scrubber system. The SDPT technology as well as a PDT system were innovative technologies demonstrated on this SVE system prior to the cat-ox treatment of any untreated contaminants or undesirable residuals. A process flow diagram for the SVE network and the innovative technologies is showed in figure 1 below.

Figure 1. Process Flow Diagram of McClellan Site S, OU D SVE System.

Major contaminants in the SVE waste stream and their maximum inlet concentrations are listed in Table 1 below:

Table 1. Maximum VOC contaminant concentrations.
Compound	Concentration (ppmv)
1,1,1-Trichloroethane	188
1,1,2-Trichloroethane	0.81
1,1,-Dichlorethane	4.1
1,1-Dichloroethene	4.37
1,2,4-Trimethylbenzene	38
1,2-Dichlorobenzene	104
1,3,5-Trimethylbenzene	4.03
1,3-Dichlorobenzene	4.35
1,4-Dichlorobenzene	9.33
4-Ethyl Toluene	6.18
Acetone	75.9
Benzene	3.76
c-1,2-Dichlroethene	2.85
Chlorobenzene	1.56
Ethylbenzene	7.09
Freon 113	1.42
m,p-Xylenes	19.2
Methyl Ethyl Ketone	7.27
Methyl Isobutyl Ketone	20.2
Methylene Chloride	6.18
o-Xylenes	6.14
Tetrachloroethene	77.2
Toluene	63.3
Total Xylenes	15.8
Trichloroethene	85.6
Contaminant concentrations were measured by GC/MS EPA TO-14 and GC Method 8010/8020.

Description of Technology

The SDPT generates nonequilibrium nonthermal plasmas to oxidize VOCs in the vapor stream by dielectric-barrier electrical discharges. Nonthermal plasma is a gaseous state of matter at near ambient temperatures and pressures in which a part or all of the atoms or molecules are dissociated to form ions. SDPT produces this plasma in a planar Pyrex ^TM glass cell sandwiched between stainless steel electrodes, using glass as the dielectric medium. Figure 2 below shows a schematic of the plasma reactor.

Figure 2. SDPT Plasma Reactor

The dielectric barrier and the application of alternating high voltages, produces substantial quantities of plasma by a large number of uniformly spread microdischarges in the gas below the electrode area. Without the dielectric barrier, localized intense arcs rather than microdischarges would develop in the gas between the metal electrodes. A mean electrical field discharge area of 1,236 cm² and an active discharge volume of 310 cm³ is given in the planar cell approximate 71 centimeters (cm) in length, 18 cm wide, with a 2.5 millimeters (mm) gap.

As the gases pass through the airtight planar cells, contaminants are exposed to high-energy plasmas. These plasmas generate a series of free radicals such as atomic oxygen and the hydroxyl radical and free radical reactions. These free radical reactions are responsible for the oxidation of the halocarbons into basic products of oxidation such as HC1, CO₂, H₂0 and other reaction byproducts. Field conditions such as temperature, flow rate, humidity, and type and concentration of contaminants affect the quantity and concentration of the products of plasma technology.

During this demonstration, two SDPT reactors were connected in parallel, each having a capacity of 5 cfm. Each reactor consisted of two stacks of planar cells placed in a sealed vessel, with each stack containing 10 cells. The entire SDPT system, excluding an oil-cooled heat exchanger to prevent the cells from overheating, was housed in and operated from a long-bed trailer.

The plasma cells were energized by a variable frequency oscillator amplifier coupled to a set of tuning inductors and a step-up transformer. The electrical and gas measuring instruments were interfaced to a computer-based data acquisition and analysis system The fully loaded trailer (excluding the tow truck) was estimated to weigh less than 4 tons. The SDPT system schematic and layout are presented in figure 3 below.

Figure 3. Schematic of the SDPT.

The following items were included as part of the SDPT package:

• Equipment trailer (including air conditioner and gasoline-powered motor-generator)

• Computer-based data acquisition and control system

• Plasma-processor reactor, including plasma cells

• High-voltage transformer

• Power-factor correction inductors

• Variable-frequency power supply

• Instrumentation and controls

• Oil pump

• Heat exchanger

• Dehumidifier, heater

• An air preparation/drying unit

• A gas temperature control/heating unit

• H₂ supply system

The SDPT system at McClellan AFB was operated for incrementally longer periods of time, since this system had been operated only intermittently and for less than 6 hours on a continuous basis prior to the McClellan demonstration. The system was to be operated for up to 4 hours per day during the first week of the demonstration, 8 hours during the second week, 12 hours during the fourth, and thereafter on a 24-hour basis for an additional 4 weeks. Additionally, the effect of dehumidifying the process gas and introducing hydrogen into the cells would be studied during the demonstration. To compensate for downtime during troubleshooting, the technology vendor demonstrated the system for an additional week.

Two different types of samples were collected and analyzed, namely a relatively inexpensive gas chromatography (GC) method (modified EPA Method 8010/8020) was used to provide fast VOC results ( 2-day turnarounds), and EPA Method TO-14, a gas chromatography/mass spectrometer (GC/MS) method, was used to confirm the VOC results analyzed by GC. The GC method provided results for 18 of the most frequently detected VOCs in the SVE gas stream, while method TO-14 also provided an extended list of VOCs that may be present in the slipstream.

Performance of Technology and DRE

The DRE was calculated under three separate operating conditions; first without pretreatment of the gas, second with hydrogen injection only prior to treatment and finally with hydrogen injection and dehumidification.

Field Performance Data

Listed in tables 2, 3 and 4 below are the DRE for the contaminants of concern at the McClellan SDPT demonstration.

Table 2. Average VOC DRE by SDPT (without pretreatment) analyzed by GC/MS EPA TO-14 and GC Method 8010/8020.
	GCMS			GC Method 8010/8020
	Concentration (ppmv)		Average	Concentration (ppmv)		Average
Compound	Inlet	Outlet	DRE (%)	Inlet	Outlet	DRE (%)
1,1,1-Trichloroethane	155	9.99	93.55	188	28.7	84.72
1,1,2-Trichloroethane	0.81	0.43	>47.41	NR	NR	NR
1,1,-Dichlorethane	3.33	0.11	>96.67	4.10	0.56	>86.32
1,1-Dichloroethene	4.37	0.11	>97.44	4.09	0.52	>87.37
1,2,4-Trimethylbenzene	10.5	0.18	>98.28	8.91	0.47	>94.72
1,2-Dichlorobenzene	31.4	0.92	>97.07	8.47	0.50	>94.16
1,3,5-Trimethylbenzene	3.98	0.10	>97.47	NR	NR	NR
1,3-Dichlorobenzene	2.30	0.11	>95.18	NR	NR	NR
1,4-Dichlorobenzene	6.04	0.21	>96.56	NR	NR	NR
4-Ethyl Toluene	6.18	0.13	>97.98	NR	NR	NR
Acetone	75.9	2.06	97.29	NR	NR	NR
c-1,2-Dichlroethene	2.42	0.11	>95.30	2.85	0.52	>81.83
Ethylbenzene	2.77	0.10	>96.36	3.80	0.57	>84.96
Freon 113	NR	NR	NR	1.11	0.56	>49.70
Methyl Ethyl Ketone	7.27	2.13	>70.70	NR	NR	NR
Methyl Isobutyl Ketone	20.2	.78	>96.12	NR	NR	NR
Methylene Chloride	5.95	.33	94.51	6.18	0.77	87.48
Tetrachloroethene	71.0	1.04	> 96.53	64.8	1.37	>97.88
Toluene	52.2	0.77	> 98.53	47.7	0.85	>98.22
Trichloroethene	81.1	1.21	> 98.51	85.6	1.55	>98.18
m.p-Xylenes	NR	NR	NR	9.35	0.55	>94.18
o-Xylenes	NR	NR	NR	3.19	0.61	>80.94
Total Xylenes	15.8	0.17	> 98.90	NR	NR	NR
Average VOC DRE			> 89.05			>91.39
GC/MS EPA TO-14 method: 4 samples taken from 11/13/95 to 12/15/95.
GC Method 8010/8020: 9 samples taken from 11/9/95 and 12/14/96.
The greater than symbol (>) indicates that the compound was detected in the inlet, but was not detected in the outlet sample on at least on of the collection dates.
NR-not reported

Table 3. Average VOC DRE by SDPT with hydrogen addition analyzed by GC/MS EPA TO-14 and GC Method 8010/8020.
	GC/MS EPA TO-14			GC Method 8010/8020
	Concentration (ppmv)		Average	Concentration (ppmv)		Average
Compound	Inlet	Outlet	DRE (%)	Inlet	Outlet	DRE (%)
1,1,1-Trichloroethane	147	18	87.76	153	20.3	86.75
1,1,2-Trichloroethane	NR	NR	NR	NR	NR	NR
1,1,-Dichlorethane	3.70	0.22	94.05	3.95	0.45	>88.64
1,1-Dichloroethene	3.77	0.22	94.06	2.95	0.45	>84.80
1,2,4-Trimethylbenzene	10.2	0.34	96.71	36.5	1.27	96.54
1,2-Dichlorobenzene	35.6	1.85	94.80	90.6	3.81	95.79
1,3,5-Trimethylbenzene	4.03	0.18	>95.53	NR	NR	NR
1,3-Dichlorobenzene	2.36	0.18	>92.37	NR	NR	NR
1,4-Dichlorobenzene	6.61	0.41	93.83	NR	NR	NR
4-Ethyl Toluene	5.81	0.25	95.78	NR	NR	NR
Acetone	60.9	0.90	>98.53	NR	NR	NR
c-1,2-Dichlroethene	2.51	0.18	>92.83	NR	NR	NR
Ethylbenzene	2.47	0.18	>92.71	5.71	0.32	94.41
Freon 113	NR	NR	NR	1.33	0.20	85.05
Methyl Ethyl Ketone	NR	NR	NR	NR	NR	NR
Methyl Isobutyl Ketone	19.5	0.91	95.35	NR	NR	NR
Methylene Chloride	5.49	0.49	91.02	5.76	0.57	90.08
Tetrachloroethene	55.0	3.13	94.31	65.4	3.84	94.12
Toluene	46.1	2.35	94.90	59.3	2.82	95.25
Trichloroethene	79.0	3.95	94.80	78.7	4.46	94.33
m,p-Xylenes	NR	NR	NR	17.5	0.79	95.50
o-Xylene	NR	NR	NR	5.16	0.26	94.92
Total Xylenes	14.1	0.41	97.10	NR	NR	NR
Average VOC DRE			>92.72			> 92.48
GC/MS EPA TO-14 method: 1 sample taken on 12/19/95.
GC Method 8010/8020: 1 sample taken on 12/19/95.
The greater than symbol (>) indicates that the compound was detected in the inlet, but was not detected in the outlet sample on at least on of the collection dates.
NR-not reported

Table 4. Average VOC DRE by SDPT with hydrogen addition and dehumidification analyzed by GCMS and GC Method 8010/8020.
	GC/MS EPA TO-14			GC Method 8010/8020
	Concentration (ppmv)		Average	Concentration (ppmv)		Average
Compound	Inlet	Outlet	DRE (%)	Inlet	Outlet	DRE (%)
1,1,1-Trichloroethane	162	34.3	78.81	144	8.01	94.43
1,1,2-Trichloroethane	0.81	0.47	41.98	NR	NR	NR
1,1,-Dichlorethane	3.53	0.38	>89.37	3.47	0.20	>94.39
1,1-Dichloroethene	4.23	0.37	>91.32	2.78	0.20	>93.01
1,2,4-Trimethylbenzene	6.13	0.37	>93.95	38.0	0.36	99.05
1,2-Dichlorobenzene	14.7	0.62	95.77	104	2.47	97.62
1,3,5-Trimethylbenzene	2.78	0.34	>87.75	NR	NR	NR
1,3-Dichlorobenzene	4.35	0.28	>93.64	NR	NR	NR
1,4-Dichlorobenzene	9.33	0.34	>96.34	NR	NR	NR
4-Ethyl Toluene	3.91	0.36	>90.79	NR	NR	NR
Acetone	60.3	10.4	82.67	NR	NR	NR
Benzene	NR	NR	NR	3.76	0.18	>95.31
c-1,2-Dichlroethene	2.28	0.34	>85.07	2.28	0.20	>91.46
Chlorobenzene	NR	NR	NR	1.56	0.20	>87.55
Ethylbenzene	2.25	0.36	>84.03	7.09	0.20	>97.25
Freon 113	NR	NR	NR	1.42	0.07	>94.92
Methyl Ethyl Ketone	7.27	2.35	67.68	NR	NR	NR
Methyl Isobutyl Ketone	17.9	1.16	>93.52	NR	NR	NR
Methylene Chloride	6.05	1.15	80.93	5.99	0.24	96.01
Tetrachloroethene	53.9	1.27	>97.65	77.2	0.41	>99.47
Toluene	46.3	0.98	>97.88	63.3	0.47	>99.25
Trichloroethene	78.6	1.47	>98.13	74.1	0.17	>99.77
m,p-Xylenes	NR	NR	NR	19.2	0.16	>99.15
o-Xylenes	NR	NR	NR	6.14	0.13	>97.97
Total Xylenes	11.6	0.42	>96.37	NR	NR	NR
Average VOC DRE			>97.73			>97.24
GC/MS EPA TO-14 method: 3 samples taken from 12/27/95 to 1/10/96.
GC Method 8010/8020: 4 samples taken from 12/27/95 and 1/10/96.
The greater than symbol (>) indicates that the compound was detected in the inlet, but was not detected in the outlet sample on at least on of the collection dates.
NR-not reported

Laboratory studies conducted in October 1995 at Los Alamos National Laboratory, New Mexico indicated that dehumidified contaminated gas streams and the introduction of hydrogen into the gas stream produced higher DREs. This would especially be the case if there was insufficient stoichiometric hydrogen to drive the chlorine in the pollutants to the HCl form.

Treatment Effectiveness

Unlike what is found in most cases, the method TO-14 and method 8010/8020 results obtained in this study are within reasonable bounds of comparability. The overall DREs calculated from these two different methods are nearly identical, although there is significant variability between some of the measured concentrations. Without pretreatment, the overall DRE for VOCs was greater than 89.1 percent. The introduction of hydrogen (when the refrigerator/dryer was not operational) did not have a significant effect or reduced the DRE for some compounds to greater than 92.7 percent. However, only one sample was tested to examine the effectiveness of the hydrogen addition. The addition of hydrogen and dehumidification (i.e. with the refrigerator/dryer operational) resulted in a significant increase in DRE to greater than 97.9 percent. However, the DRE for compounds such as 1,2-Dichlorobenzene (DCB) and 1,4-DCB decreased slightly. The SDPT investigators suggested that this may be attributable to the small aberrations in sampling and analytical precision. Additionally, the DREs for some compounds such as methylene chloride are lower than the average due to a possible recombination of various breakdown compounds. ENV America Inc. suggests that the increase in DREs for chlorinated compounds is due to the introduction of hydrogen gas that produces more free radicals in the cells and reacts with chlorine ions to form hydrochloric acid.

System Reliability

Although the initial up time was low, there were several initial mechanical problems, the system was ultimately demonstrated to run unattended on a continuous (24-hour) basis without any drop in the DRE. The low initial up time was a result of inadequate process control and mechanical problems, which was later rectified. Problems encountered include inadequate or improperly placed relief values, which resulted in an excess vacuum or pressure that cracked the cells. Cracked cells were replaced with new cells and fail-safe measures were added to protect the system. Corrosion of incompatible construction materials probably due to excess moisture, and odd sized fittings which resulted loss of pressure and temperature were other problems encountered with the SDPT demonstration system. Odd sized fittings were replace and corroded materials and fittings were replaced with Teflon components and an additional knockout pot followed by a dehumidifier was added to the system to alleviate these conditions. The addition of a knockout pot and a dehumidifier did not appear to alleviate the generation of acidic condensate in the cells, nor did it seem to reduce the generation of a brownish residue in the reactor cells. Fluctuating temperatures with in the cells resulted in fluctuating off gas capacities and resonance conditions, which triggered frequent shutdowns. However, once a temperature controller was installed, the continuous up time of the system was increased from 6 hours to 24-hours, with no operator.

Ease of Operation

Mobilization, including hooking up to the main system, took less than 2 days, since the system came skid-mounted with practically all its internal plumbing and wiring in place. System startup was simple and the system was stabilized in less than an hour without compromising the high DREs seen during the demonstration. Once the mechanical and engineering errors were addressed, the system ran continuously for 24-hours with no operator.

Maintenance Requirements

The McClellan demonstration showed that the overall engineering design needs substantial improvement for full-scale implementation. Since there are very few moving parts, system maintenance will probably be influenced most by the need to replace parts that fail as a result of corrosion from the acidic components that are generated. Deposition of semisolid residues in the cells limits the usability of the cells to approximately 100 hours of operation. Since these cells are expensive and labor-intensive to replace the McClellan demonstration managers suggest that cell efficiency could be improved by dividing the reactor into several stages. Better control of demineralization and recombination of byproducts could be achieved as each stage is subjected to incremental energy densities. Additionally, constructing the cells from an alternative nonstick, nonporous materials may alleviate the solid residue buildup with in the cells. The demonstration managers suggest that these cells may even be reused after in-place cleaning.

While the dehumidifying process appeared to improve the system DRE, the inside of the dehumidifyer became clogged. It is not clear if this clogging was due to an inadequate capacity or due to corrosion by organic condensation products. During the McClellan demonstration, the cell temperature was maintained by a temperature controller. This controller activates or deactivates an air fan, however this type of control system wears out the fan motor quickly because of frequent startups and shutdowns. The McClellan managers suggest that the installation of a fuzzy logic controller that activates a variable frequency drive motor can regulate the speed of the fan, maintain a more steady temperature, and prolong the life of the fan motor. As a result of the McClellan demonstration manager suggestions, ENV America was in the process of incorporating several design and material changes to the system.

Energy Consumption

Operating costs and capital costs are based on a 472 kW power requirement to treat this contaminant mixture at 250 cfm.

Space Requirements

The trailer containing the SDPT system measured 20’ long, by 7’ wide and 7’ 6” tall.

Worker Health and Safety Issues

(Not available.)

Risk and Consequence of Catastrophic Failure

(Not available.)

Noise/Aesthetics, etc.

(Not available.)

Wastes Produced

Hydrochloric and hydrofluoric acid concentrations in the effluent gas were 150 and 5 ppmv, respectively. The generation of these acids was not expected. In total, approximately 1 gallon of corrosive liquid (pH less than 1.0) was removed from the SDPToutlet. This is equivalent to 0.68 gallons/hour of acidic liquid residue generated per 1,000 cfm of offgas treated. This liquid was approximately 45% nitric acid and 8.8% hydrochloric acid. This liquid acid may have formed from water vapor and condensed liquid water present in the SVE off gas and as a byproduct of the SDPT system adsorbing the acids.

Negligible NO_x formation (less than 2 ppm) was expected as a result of oxidation of the nitrogen in the influent gas, however NO_x measurements were extremely high (~ 68,000 ppm). However, the NO_x method measures NO_x as nitrates and therefore the results could have been corrupted because of interference from nitric acid mists that were generated in the cells. Upon condensing inside the sampling equipment, the nitric acid mists would have been measured as nitrates and consequently as NO_x; therefore, the NO_x generation from this unit was not quantified for this study.

Phosgene was expected to be a byproduct of this system. Two phosgene samples were collected, however the manufacturer of the collection tubes indicated that the tubes may have been defective. Therefore the finding of nondetecable phosgene in the effluent gas is suspect and no conclusions can be drawn regarding the production of this potential by product. Average ozone concentrations in the two samples taken (analyzed by NIOSH method 1300) was 58 ppmv. Average dioxin concentration, based on four samples, was 0.0045 ng/m³.

An equivalent of approximately 1.44 gallons/hour of condensate, per 1,000 cubic feet per minute (cfm) of SVE offgas treated, was generated from dehumidification. One grab sample of the condensate indicated that the total VOC and SVOC concentration was approximately 87,000 mg/l.

A dark, blackish-brown deposit was formed inside the plasma cells. A cursory literature search of similar laboratory and pilot-scale demonstrations indicated that compounds such as acetylene formed solid polymerization compounds when subject to silent discharge plasma. Based on the results of one pH analysis, it is difficult to determine if the cells will be a hazardous waste, and this should be determined during implementation. The McClelan demonstration could not provide conclusions regarding the characteristics nor the causes for the blackish-brown deposit.

Data on Key Parameters

Table 5 below shows the ranges for some key operating paramenters.

Table 5. Average Influent Offgas Conditions
Parameter	Range	Comments
Flow	3 to 10 cfm	6.3 cfm was the approximate average flow rate.
SVE off gas temperature	18 to 59^oC	40^oC was the approximate average temperature.
SVE off gas relative humidity without dehumidification	95 to 100%
SVE off gas relative humidity with dehumidification	20 to 30%
Energy density	4000 to 7000 Joules/Liter (J/l)	DREs were unaffected for energy densities over 4000 J/l
Uptime	26 %	Uptime increased towards the completion of the demonstration.

Capital and Operating Costs

The McClellan demonstration cost analysis included the following assumptions:

· 90 percent process/equipment uptime (average of 22 hours/day) during normal operations.

· Continuous, 24-hour, steady-state operation.

· SVE offgas concentrations in the range of 400 to 500 ppmv of total VOC.

· DREs > 95 percent.

· An organic removal rate of approximately 2 lbs/hr and a flow throughput of 250 cfm.

· A 3-year projected equipment lifetime with 50 percent salvage value.

· Analytical costs reflect only typical and normal operating scenarios.

· Influent flow, temperature, and pressure are maintained to meet the SDPT specifications.

· Operating costs and capital costs are based on a 472 kW power requirement to treat this contaminant mixture at 250 cfm.

Based on these cost assumptions the approximate capital and operating cost for the SDPT system is shown in table 6 below.

Table 6. Approximate Capital and Operating Costs for a 250 cfm SDPT system.
Capital Cost		Annual Operating Cost
SDPT system (installed)	$608,000	Maintenance and supervision	$105,000
Scrubber for acidic gaseous effluent from SDPT	$86,000	Condensate disposal	$16,800
		H2/N2 gas	$4,400
		Cells and parts replacement	$99,600
		Expendables (caustic for acidic gaseous effluent and acid residue)	$8,900
		Electricity ($0.10/kWh)	$378,000
Total	$694,500	Total	$612,700

The estimated operating costs are based on the system and conditions present at McClellan AFB and are sensitive to the following factors:

· Concentration and composition of the SVE offgas, as these concentrations determine the total energy required (joules/liter) to treat a gas stream.

· The cost of energy, since approximately 60% of the operating cost is related to the cost of energy.

· Disposal/recycling cost of the organic condensate and the acidic residue.

· Cell replacement cost.

Data Gaps Identified

· Contact vendor to determine if the changes suggested during this demonstration have been implemented since 1995 and if they have demonstrated the technology elsewhere.

Vendor

ENV America, Inc.

Max Reyhani, P.E.
16 Technology Drive, Suite 154
Irvine, CA 92618
ph: (949) 453-9191, fax: (949) 453-9292

E-mail: envamerica@argotech.net, info@envamerica.com

www.envamerica.com

Demonstration contact

CH2M Hill

McClellan, 1996, Silent Discharge Plasma Technology Technical Memorandum, Environmental Management of Offgas Technology, Operable Unit D, McClellan Air Force Base, AR File Number 3188, October.