|
|
|
A recent trial burn at an Ohio cement kiln provides additional evidence that burning hazardous waste in cement kilns threatens the environment and human health. The burn -- dubbed by some the most comprehensive cement kiln trial ever conducted -- should lay to rest industry claims that cement kilns are safe incinerators for hazardous waste.
The kiln -- Southwestern Portland Cement, a subsidiary of Southdown, Inc. -- has burned a diverse mixture of hazardous wastes for several years at its Fairborn, Ohio, location. Citizens have organized the Greene County Environmental Coalition to stop the practice.
The results of the trial burn -- despite severe limitations in scope and design -- make clear once again that cement kilns do not "destroy" hazardous waste but instead emit significant quantities of heavy metals and combustion by-products into the environment. The attached comments discuss the test results, with an emphasis on the following findings:
The trial burn did not include routes of toxic release other than the smokestack (fugitive emissions, releases during transport, releases of contaminated kiln dust, and shipments of contaminated cement products). Further, the trial burn was conducted under carefully controlled and monitored conditions using a homogenous waste stream. The results of the trial burn thus present an overly optimistic picture of day-to-day operations.
For further information, please contact Joe Thornton at Greenpeace.
GREENPEACE
Toxics Research and Information Unit
139 Townsend Street
San Francisco, CA 94107
415/512-9025
COMMENTS ON
Entropy Environmentalists, Inc., for Southdown, Inc.
"Stationary Source Sampling Report #8114:
Southwestern Portland Cement Company, Fairborn, Ohio
April 16-23, 1991"
prepared by Joe Thornton, Research Analyst
The trial burn clearly demonstrates the failure of the tested cement kiln to "destroy" hazardous wastes. When wastes were burned, the kiln was found to emit large quantities of toxic chemicals to the environment, including the extremely toxic dioxins and furans. Total emissions of unburned hydrocarbons are estimated at up to 190,000 pounds per year when wastes are burned.
The kiln failed to demonstrate a destruction and removal efficiency (DRE) of 99.99 percent for all specified waste chemicals tested (called principal organic hazardous constituents, or POHCs). On a legal basis, this failure in itself necessitates the denial of permission to burn hazardous waste.
Emissions of heavy metals when wastes were burned increased significantly over emissions when coal only was burned. Total emissions of the 14 metals measured were 167 percent higher when waste was burned than when coal-only was burned. Lead emissions were more than 6 times greater, and zinc emissions were 11 times greater.
Emissions of dioxins and furans were more than 8 times greater when wastes are burned than when coal only was burned. Annual dioxin and furan emissions when the kiln burned wastes exceeded the annual "acceptable dose" calculated by EPA for approximately 1.5 billion adults.
Emissions of carbon monoxide and total hydrocarbons -- parameters thought to indicate the effectiveness of the incineration process -- were significantly worse when wastes were burned. Both parameters exceeded the limits specified in EPA regulations. The failure to achieve the levels EPA considers indicative of "good operating practice" confirms the view that the kiln is not effective in destroying hazardous wastes.
Flaws and limitations in the design of the monitoring program limit the applicability of the results for making policy. The results of the test are certainly inadequate to form the basis of a risk assessment or to defend any assertion that the burning of waste is "safe" or "acceptable."
1. Idealized conditions. The operation of the kiln during the trial burn was controlled and monitored far more carefully than under routine daily operation, when variations in waste type and quantity, human error, equipment malfunction, and combustion upsets all lead to increased emissions. As a result, the trial burn results provide an overly optimistic estimate of emissions that can be expected during routine operation.
2. Major omissions. In addition, the test measured only smokestack emissions, excluding the following routes by which toxic chemicals are released from kilns burning waste:
As a result, any risk assessment based on the test results will be so incomplete as to be irrelevant.
3. Failure to prevent hysteresis effect. Southdown was warned that a failure to allow adequate time between test runs would result in the delayed emissions from one run "contaminating" subsequent tests. For instance, delayed emissions from Condition 2 could contaminate "coal-only" Condition 3, while resulting in overestimates of the system's ability to remove metals and POHCs during Condition 2.
In its early test plans, Southdown noted that cement kilns require as much as one week to reach equilibrium after feed concentrations of metals and chlorine are changed. Southdown failed to follow its own recommendations, as shown below. After waste-burning Condition 1 (ending at 0405 hrs on 4/18), only 11.5 hours passed before waste-burning Condition 2 began at 1535 hrs on the same day. After Condition 2 ended (1622 hrs on 4/19), coal-only Condition 3 began just 72 hours later -- an insufficient amount of time for the system to stabilize itself under "coal only conditions."
Because sampling devices were shut off when wastefeed ceased at the end of each condition, delayed emissions were not measured or reported. As a result, estimates of metals emissions during conditions 1 and 2 are underestimated and DRE calculations are overestimated.
The test results make clear that significant hysteresis effects occurred from Condition 2 to Condition 3. For instance, cadmium and mercury (important metals in hazardous waste) were detected in condition 3 emissions, despite the lack of any measurable cadmium or mercury in either the coal or the minerals fed to the kiln. Further, as additional time passed within Condition 3, emissions of key parameters (metals, THC, CO, etc.) continued to fall, suggesting the waning influence of delayed emissions as time passed.
WANING INFLUENCE OF HYSTERESIS DURING CONDITION 3
Condition 3 (coal only)
Run 1 Run 2 Run 3
Date 4/22/91 4/23/91 4/23/91
Time 1815-2245 910-1310 1351-1727
-----------------------------------------------------------------------
Carbon monoxide ppm 189.2 166.0 162.9
Total hydrocarbons ppm 89.0 87.1 75.3
Particulate (lb/hr) 5.74 1.15 .905
PCDD/F (lb/hr ^-9 TEQ) 5.55 3.19 3.86
Arsenic (^-3 lbs/hr) 6.4 < 5.8 < 5.8
Barium (^-4 lbs/hr) 10.6 2.8 9.1
Chromium (^-4 lbs/hr) 5.28 5.26 3.18
Lead (^-4 lbs/hr) 9.54 4.18 6.21
Mercury (^-3 lbs/hr) 1.39 2.04 2.60
Nickel (^-4 lbs/hr) 9.15 <5.26 <5.26
Zinc (^-3 lbs/hr) 3.17 17.6 2.65
4. Last-minute changes. For unknown reasons, Southdown deviated significantly from the test designs published in late 1990, which had been subject to public comment and review and was subsequently approved by RAPCA. Some of these changes appeared in the eleventh-hour plan printed just before the test (March 27, 1991); some do not. In either case, these last minute changes are a violation of public process and -- in my opinion -- completely invalidate the test as a decisionmaking tool in a community's political process. It also betrays a profoundly cynical attitude by Southdown and RAPCA concerrning the value of public participation.
For instance, Southdown did not conduct tests to separate the effects of burning tires from those of burning hazardous waste, as specified in the approved plan. As a result, denial of permission to burn wastes based on the trial burn must apply to both hazardous waste and tires.
In addition, Southdown changed its selection of waste chemicals (principal organic hazardous consitutents, or POHCs) from that specified in the final plan. Further, the quantities of liquid hazardous waste burned were significantly less than those published in the final test plan, as shown below:
QUANTITY OF LIQUID HAZARDOUS WASTE BURNED (lbs/hr)
Test Actual
plan test
-------------------------------
Condition 1 7040 5605
Condition 2 7040 5573
Of course, the values used in a trial burn provide the maximum operating values that may be permitted. As a result, the maximum wastefeed rate that could be permitted based on this trial burn is 5605 pounds per hour, some 20.5 percent less than the amount specified in Southdown's proposed permit conditions.
Southdown failed to follow the published test design for metals testing. According to the test plan, "cadmium and lead will be spiked into the fuel stream at a sufficient level to approximate worst case conditions. Cadmium and lead are known to be among the most volatile metals present at greater than trace level in the fuel stream. This will demonstrate the removal efficiency (RE) of the system." This important element of the testing plan was not carried out, as shown below.
METALS IN HAZARDOUS WASTE FEED
(in lbs/hour)
Maximum Actual feed Actual feed
permitted Condition 1 Condition 2
-------------------------------------------------------
Lead 24.0 18.4 15.6
Cadmium 1.4 1.3 0.5
Nickel 16.8 8.3 6.0
The kiln failed to meet the standard of 99.99 percent destruction and removal efficiency for monochlorobenzene. Reported DREs for Condition 1 are <99.97 and <99.98. This finding in itself disqualifies the kiln from receiving a permit to burn hazardous waste.
We anticipate that Southdown will attempt to discount this result by arguing that DRE in this case was "probably" better than 99.99 percent but detection limits combined with a failure to feed enough monochlorobenzene to the kiln precluded such a finding.
The law states that a permit for a combustion device may be granted only if the facility demonstrates the actual achievement of 99.99 percent for all POHCs tested -- not only some of the POHCs, and not only those POHCs for which the test was designed with sufficient foresight to measure a DRE of 99.99 percent. Southdown cannot legally be granted a permit on the basis of this trial burn.
Dioxin and furan emissions were far higher in the waste-burning tests than in the coal-only test, as shown below. Given the lack of any chlorine/organochlorine content in most grades of coal, we believe that the presence of dioxin in the coal-only run is probably due to delayed emissions from condition 2, which took place less than 72 hours previously.
PCDD/PCDF emissions are 8 and 5 times higher in waste-burning conditions than in coal-only conditions.
PCDD/F EMISSIONS
(Toxic equivalency factors, g/yr)
Condition 1 (waste) .238
Condition 2 (waste) .130
Condition 3 (coal) .029
Such emissions are of great significance. Because dioxins and furans are extremely persistent and bioaccumulative, emissions of apparently dilute quantities build up in the environment over time. In light of dioxin's extreme toxicity at low doses, emissions measured from the kiln during waste-burning runs are cause for great concern. Low doses are thought to have their greatest effect upon the reproductive and immune systems and upon the neurological and sexual development of infants).
Based on the emissions during waste-burning condition 1, the kiln's dioxin emissions are equal to EPA's "acceptable" dioxin dose for 1.5 billion persons. The acceptable dose (.006 picograms/ kilogram/day) is based upon a one-per-million risk of fatal cancer, and this calculation assumes an adult weighing 70 kilograms and kiln operation of 8000 hours per year. Under waste-burning condition 2, the kiln's dioxin emissions are equal to the "acceptable" dose for 848 million persons. Of course, not all the PCDD/F emitted from the kiln will be directly ingested by humans; this calculation serves only to put the apparently small quantity of this highly persistent chemical into toxicological perspective.
These dioxin emissions are comparable to those of the second tier of combustion sources analyzed in EPA's National Dioxin Study. While less than those measured at a secondary copper smelter and several municipal waste incinerators, the kiln's dioxin emissions are of the same order as those from a hazardous waste incinerator, a boiler burning salt-soaked wood, an industrial waste incinerator, and a kraft paper recovery boiler. The kiln's dioxin emissions are one or more orders of magnitude greater than all other combustion sources analyzed in the study.
Measurements of total hydrocarbons (THC) during waste-burning runs totaled 190,000 and 151,000 pounds per year respectively (assuming 8000 hours of operation per year). The actual total of PICs is presumably significantly higher, since THC has very poor efficiency for measuring chlorinated species, as well as non- and semi-volatile PICs.
It is interesting that very few of the PICs analyzed for were present in quantities greater than detection limits. As a result, only a tiny fraction of the unburned chemicals emitted from the stack were identified. Over 99 percent remain mystery chemicals. The few compounds identified cannot serve as the basis for a meaningful risk assessment.
The identities of the PICs that make up the THC measurements is thus largely undetermined. Presumably, the identities of PICs emitted are different when burning waste from those when coal is burned, due to the lack of chlorine in coal. However, it is also notable that the quantities of THC emitted are significantly greater during waste-burning runs than during coal-only conditions.
THC EMISSIONS
Lbs/hr ppm
Main + bypass Main stack only
---------------------------------------
Condition 1 (waste) 23.8 109
Condition 2 (waste) 18.9 81
Condition 3 (coal) 18.5 83
The concentrations of THC measured in the main stack in all runs exceed by a factor of 4-5 the regulatory standard of 20 ppm used for both hazardous waste incinerators and waste-burning kilns. Clearly, the kiln is not an adequate combustion environment to destroy hydrocarbons (i.e., hazardous wastes).
It is also notable that CO emissions are significantly worse during the burning of waste than when coal-only is burned. Increased carbon monoxide emissions are commonly regarded as evidence of incomplete combustion.
Further, the main stack emissions of CO again greatly exceed the regulatory standard of 100 ppm required for hazardous waste incinerators and for waste-burning kilns (with some loopholes).
CO EMISSIONS
(main stack only, in parts per million)
parts per
million, volume
Condition 1 (waste) 232
Condition 2 (waste) 202
Condition 3 (coal) 173
As noted above, the failure to account for the hysteresis effect results in underestimates of metal emissions for waste-burning runs and a possible overestimate during coal-only runs. Nevertheless, metals emissions are significantly higher when waste is burned. Total emissions of all metals analyzed were 167 percent and 52 percent higher when wastes were burned than when coal-only was burned.
Not surprisingly, the increases were most significant for the more volatile metals. Those metals for which hazardous waste makes a relatively small contribution, increases were small or non-existent (i.e., barium, thallium, vanadium). But those metals of greatest concern -- lead, mercury, cadmium, nickel, and zinc -- showed great increases when wastes were burned. Lead emissions in condition I were 6 times greater than in condition 3, and zinc emissions were 11 times greater.
METALS EMISSIONS
Percent increase when burning waste over coal-only conditions
Condition
1 2
--------------------------------------------
Cadmium 184 64
Chromium 41 (-2)
Lead 549 150
Mercury 162 197
Nickel 185 90
Selenium 351 100
Silver 51 05
Zinc 1035 627
ALL METALS 167 52
| · To Jym's Greenpeace Page · | · To Jym's Home Page · | · To Greenpeace's Website · |