US Production Of Manufactured Gases Assessment Of Past Disposal Practices

4J fo k km • -o 4 a— — E *• 03 k O k — • o o a— • • «• «" C i e — o o k •>• • • 4* — O O 4» U 4l — •' e • '• 5 '-o 44* o x • e <• '• 4> x a? u 44* k • • U • -0 XX O T5 XX* 4< 4> — C 4> • 4> • X 4 • «- • • 4> k NO X « • N 3 » 4j « e •— * 4* k X • 0 k • 4 J|4» ^> k .- 5g" U C •— 4 U 5 3 ku c 4 .- v. -o e 4 — • « S • f — X > -004 X E u — a,4> u 4> • • e • k OX 0—0 0X0 M It X 1 • t •• "o •o • • t ?2 — X r: VI O X 4 k ^j | J" ^ J, O « c 0 * u > 1 4 3 U ** a « 0 — ^ II fl ii a x e <-i E 4 k • 34) k • a t ax C 4) — e k Tl •o • 4* *• • U 4< • i a — M O «l 4 e o •^ 4 O U 0 C "• 0 •• ot 4> e 4» — Ii e e • 4»4» * See 1.2.1 O C 4 — OH • U • • en a— • • X 4> • e u — T c 3 O J> 3 3 — • •o L S a w {•J l-ii 5J; u N ^ X - C 4) k 4 41 U O 4 a x k O 4t x'o'c ux — 41 U ; K ^ C * H g « e — • o a - • [•3 t o 4) «• j 1 •K S •v ... o %• o c 0 4) N k 5 k • • * '• e ll — • — u o a O S5' e • 'k o a 4»-"o. •) — E •2 S * — ak "3-S5 • 5-5 O O *o ok §5-s r^: • CO sss- 0 k 3 1 4 k 4> * e • £ ' ° 4* • • "O .- • M 4» • X > O • — — e 4> 4> 4 U — 4TI • • e ja " a o • c u o — — 4l 4> — U k — a • o. k 0 — 0 ^— a k o £a ax S e 41 « — k — X U •0 • 4 M 5? "Si • o •ON e k o o U k — — t 4» H ant 4 4> o • > c — » k • S . S3 a • • — U k 41 04 e 4> • O en 4> u o *^ • 8k a u 0 — — o a. • o— ax O 4* < U *04» « • m ^ •^ • k"J a^ o x k • ax 4> m » u • • X 2- a • • 3 5.- 4> • • k 4> M • k 7ai 3 4» * • f x x i ^ <• €3 e ox — 4 !•> ^. ^. « * « 4» O. o k a 4j U M X X a Is • a. 4> e -> 4» it • • e 4 0 • U 4 o — Jx «- e o — • u X— 4) 41 <•? O 4> ft a e k ^9 o e A • 4) 4) VI ~x O 4> * o t u e e O 4 • k k O • u > u e 4 0 u e o « • — I* • a ruid M» >on Her in lations. • k| — 4 o e k 0 U k u 5 ••» fl — * a II * -°t o > k > e « a — c — 4> a e £ -81 ^ •o 3 C •• e o u 213
------- „ "0 3 C C O U a 01 c k o "c » o. o c *i - « k " *> *> "• • 3 c i a •o • e k o o — k • • Ok a • e 0*0 k U. o * A a-o e c • k • o • i • •~ A V) 0 • e — — a. 'o 0 * • u e • e j: • ai k e »• o — o u — u & e 4 E o • — e * • •- k i9 W > — «J C — • e • * u * — c u XX X * •O k c • • «> — » u • .. x • XI — u o 4 • "i • XXX k ** o k m x u c e * o — k — U x — > u -o e • o ox H- U •» N * C 4 k •o »» C • ?l — u r: II x e e «> o — •OUCH 3 • — •Ok • C. *i -S> • • • • o u 3 e — i-oo .- — o j • • e • 0 * k U a. a— *» *> a— u • e J • •- — e • a. 3-0 x a a— *> a. — 0*0. X 0 T> C - k C o a • • u — »'k • • *i • > e c *>•- u So T> • "3 c c o • •— a k k — • o — •. •J ••• 0 x e o e o n k ^ c o u X I • •o • 4 k 3 C g • Ol > C 4 — X 01 k *> 3 4 1 *> 12 3 'S*" — 01 k e a— k — a— a k Q e 01 k 3 a ^ k 0 k a. c • ~ X *J k fi > k e 0 U 0 • C k ^ •» J* TS • ** 4 • • 3 •- II *. e e 4 §: o — o o U 4 k • • U *> e • -o > •o k e e o 4 0 U*& k 4 01 •> •0 II ii e * M N X « • ^i ^ e > • 4 k 4> 9 *~ e e x •x •> 00. C k-o" 2T) • • A ^ »•— •• 00.8 •- o a > a 4 k • • • k 4 • 3 • X7J k J • a u • 0 > O k U 4 41 U • 1 214
------- I In a manufactured-gas site investigation in Wallingford, Connecticut, ground-penetrating radar was used to estimate the location, extent, and char- acter of tar ponds, in cases where no records were available. The ground- penetrating radar demonstrated that the tar had migrated well beyond the orig- inal pond location and the site boundary. Magnetometer surveys were used to locate buried pipes extending from the tar pond to a former lake bed, which could later be investigated by a grid of soil-test borings. Additional geo- physical tools used in this investigation included seismic refraction to assist in the definition of the depth to bedrock (a potential controlling factor in the subsurface migration of high-density contaminants; see Brattle- bo ro case study, Chapter 3) and electrical resistivity to outline locations of potential groundwater contamination (Quinn et al., 1985). Ground-penetrating radar also has potential for estimating the location and extent of lighter hydrocarbons that may be floating on the groundwater table (Stanfill and McMillan, 1985). Soil-gas sampling has potential for delineating contamination at a gas plant site when the more volatile fractions of gasifier tar (e.g., benzene, toluene, xylenes, naphthalene) are present at a site. An investigation con- ducted at the Spencer, Massachusetts, town gas site illustrates this potential applicability. During test pit excavation, site air was screened for volatile organics using a photoionization meter. These measurements were made to assess potential air quality impacts' of excavation activities, which were demonstrated to be minimal. However, air in the test pits had substantial concentrations of volatile organics (>200 ppm), levels of concern from the standpoint of occupational safety (Perkins Jordan, 1984). Although the small size of this site would limit the value of using soil-gas sampling as a site investigation technique, the levels of volatile organics suggest that it may be used to help guide sampling and analysis activities at larger, more complex sites. A discrepancy commonly encountered in the gasworks site investigations reviewed by RTI is insufficient information on the processes that operated at the specific sites. Most site assessments reported that gas was produced by coal pyrolysis or carbonization (i.e., retort or coke-oven gas); most of these sites actually were carbureted water-gas (CWG) plants. The difference is 215
------- significant, both in terms of waste characteristics and byproduct utilization practices (see Chapter 1). For instance, nitrogen and sulfur compounds are more prevalent in coal carbonization tars than in tars from CWG processes. Tar emulsions produced by CWG processes were hard to dewater. As a result, they were not reused and were disposed onsite, especially in smaller plants. Spent oxides from CWG cleanup processes often do not have the brilliant blue color often considered a characteristic of spent oxides because of the absence of significant levels of ferrous ferricyanides. One site assessment report reviewed under this study identified a mixture of yellow and red cinders, but it failed to recognize the material as spent oxide from the small CWG plant. It was not sampled or analyzed, but it could have been a source of contami- nants at the site. Historical background information of the gas industry is invaluable in planning and conducting gas plant site investigations because it can provide data on the characteristics and likely disposition of potential contaminants at site. 2.2.3 Recommendations for Site Investigations 2.2.3.1 Introduction-- As discussed in the previous section, site investigation techniques employed for hazardous waste site investigations are generally applicable to former manufactured-gas sites. However, some special considerations should be taken into account when conducting site investigations in order to focus the investigations on characteristic features of these sites. First, as described in Chapter 1 of this report, contaminants, especially gasifier tar and oil, often are contained in below-ground structures that were covered over and left when the plant was decommissioned. Gasworks site investigations initially should concentrate on identifying these structures because they often contain almost pure contaminants. Because such contaminants are contained, they are relatively easy to remove, and because they may be relatively pure, the mate- rials may be reused as supplementary fuel or chemical feedstocks (see Platts- burgh Case Study, Chapter 3). In addition, it is especially important to take extreme care not to damage these structures during site investigation or reme- diation because this could result in the release and spread of contaminants, complicating and increasing the expense of cleanup operations. 216
------- t S Second, it is important to determine the real extent of contamination on and off a site as wastes, especially solid wastes from gas cleanup operations (e.g., woodchips, spent oxides). Such wastes were often disposed in areas adjacent to but not actually on the original gas plant site. In addition, gas plant sites were usually sited in low-lying areas (to facilitate gas distribu- tion) and were adjacent to streams, lakes, or wetlands. In many cases, wastes were accidentally or deliberately discharged into these areas; recent releases into streams, lakes, and rivers have resulted in site discoveries in many cases. It is important, therefore, to investigate wetlands and waterbodies adjacent to gas plant sites for potential contamination. Third, it is important to recognize that organic contaminants with vari- ous densities commonly occur at gasworks sites. Multiple-density contaminants can result in complex contaminant migration patterns in the subsurface (Section 2.1.1) and can complicate the design and implementation of site in- vestigation and groundwater monitoring. The relative density of potential contaminants should be known, at least qualitatively, during the planning stages of site investigation activities. Fourth, it is important to understand the variety of methods used to produce the gas and the resulting variability of byproducts and waste prod- ucts. By knowing the gas production processes used at a given manufactured- gas site, it is possible to determine the most appropriate chemical analyses for development of the site investigation plan, thereby resulting in lower investigation costs. For example, an assessment plan being developed for a site that used a coal-carbonization process should include analysis of pheno- lic compounds, nitrogen heterocyclics, ammonia, and cyanides. The analysis of these substances at carbureted water-gas and oil-gas production sites is less important because they usually were produced in low amounts in these proc- esses. In addition, it is important to determine the potential toxicity and other hazards that may be associated with gas plant wastes (e.g., the carcino- genicity of coal tar and the tendency of spent oxides to spontaneously com- bust) so that adequate provisions may be made for the health and safety of onsite workers and the general public during site investigation and remedia- tion. The following is a general approach for planning and conducting site investigations at abandoned town gas sites. Most of the site investigation 217
------- ) techniques and procedures are the same as those applied to investigate any ; ground contamination situation; therefore, details of the techniques are not addressed. The approach below recommends a chronological sequence of optional activities that may have applicability to gasworks sites. The discussion focuses on describing how characteristics of gasworks could influence the planning of a site investigation. Because of the heterogeneity of gasworks sites, specific and detailed site investigation plans must be developed on a site-by-site basis. 2.2.3.2 Information Collection and Review-- Because of the age of these sites and the fact that most of the visible evidence on the site (including storage tanks and waste disposal areas) have been destroyed, it is important to review as much available information as possible. Information collection efforts should concentrate on the following: • Identification of the processes and operating practices that were used at a site, including plant size, gas production pro- cesses, types of feedstocks, gas cleanup processes, waste types, waste disposal practices, and byproduct recovery opera- tions. The entire history of the site should be covered, if possible. • Locations of structures inch as retort houses, water-gas pro- duction facilities, gas cleanup facilities, storage tanks, etc. Also, locations of waste disposal and fuel stockpiles. • Information on the activities and historical condition of prop- erties adjacent to the plant, focusing on likely areas for waste disposals (e.g., wetlands). • Information on the geology of the site (e.g., from old con- struction borings) and regional geological information. , • Past incidents of contamination release into adjacent bodies of water or encounters with contaminants during construction on the site. • This information can be very helpful in developing a field investigation plan (e.g., locating surface geophysical survey lines, soil borings, and moni- tor wells). By collecting this information early on during site assessment efforts, one can maximize the efficiency and effectiveness of subsequent site investigation efforts, both in terms of cost and utility of the data collec- ted. 218
------- i Information collection and review should begin by obtaining the actual records of the town gas site, including business records, construction plans, geotechnical reports, tax and insurance records, utility location plans, and town plat maps. Old insurance maps can be especially valuable for determining the locations of old buildings and other structures on the site. These maps were published for most towns in the East and in California until the 1950's. They were published by street address, have a scale of 1 inch for 10 feet, and were updated at 10-year intervals. The most recent versions of these maps can be obtained from the Sanborn Map Company, Plattsville, New York. Earlier versions are available from the Library of Congress on microfilm. An excellent source of information about past practices at manufactured- gas plant sites is interviews with old-timers who worked at these sites. Often these persons can provide a wealth of information that is not recorded anywhere. In several cf the case studies reviewed, old-timers supplied valua- ble information on past waste disposal practices, especially information on the locations of old waste disposal areas. Another important source of information to review when investigating abandoned town gas facilities is old aerial or ground-level photographs of the site and surrounding area. These old photographs generally provide the best record of past site activities. If one is fortunate to obtain photographs spanning several years of the town gas operation, it may be possible to accu- rately locate sources of potential contamination. As an example, Figure 61 shows the Seattle gas plant on Lake Union late in its operational period (1959); Figure 62 shows it more recently after it was developed into a park. By comparing these photographs, one can associate areas of vegetational stress in Figure 62 with gas plant operations in Figure 61. U.S. Geological Survey (USGS) and Soil Conservation Service (SCS) maps and publications, information from State geological surveys, geotechnical records, and geological publications should be consulted during a site inves- tigation for background information on local and regional hydrogeology. Finally, a walk around the site often can prove valuable during informa- tion review efforts. Even if structures have been removed above ground, often ground-level evidence remains, such as circular features marking the sites of old gas holders. Often waste disposal areas can be identified, as can surfi- cial contamination by spent oxides (especially when they contain ferric ferro- 219
------- 0> a a. V* •C i vt co O to e g, 220
------- •£ £ i S o N CO 221 . t-.*.,^s~.^*r£*n*? f'l? ** .lit'"
------- cyanides) and tars and oils. Odor also can be used to identify areas of likely contamination during these walk-around preliminary site assessments. 2.2.3.3 Field Investigation Plan Development-- Once background information has been obtained for a site, a field inves- tigation pU"i should be prepared. This effort should be closely coordinated with local, State, and Federal environmental agencies to ensure that all en- vironmental concerns are properly addressed and that State and Federal site investigation requirements are satisfied. Prior to beginning the hydrogeolog- ical site investigation, it may be necessary to submit the field investigation plan to the various environmental agencies for their review, comment, and possibly their written approval. The plan should consist of a detailed site description, past site activi- ties (including a list of known chemicals used or produced at the site), statement of work objectives, description of proposed fieldwork activities, and proposed laboratory analyses. Also, a detailed health and safety plan should be included. The health and safety plan should be prepared by a qualified industrial hygienist who should characterize the site for the potential risk to human health by field personnel conducting the site investigation. Safety precau- tions, including the level of respiratory and dermal protection, should be addressed. Emergency plans and procedures also should be included in the health and safety plan. The following text describes the field activities that are specified in the field investigation plan. The actual field investigation may deviate from the original plan if unexpected site conditions warrant. 2.2.3.3.1 Surface geophysical survey—Conducting a surface geophysical survey can be an excellent "first step" in a field investigation because it can provide preliminary information about the subsurface conditions of the site. This information may be used to modify the field investigation plan by locating areas where more detailed subsurface investigation may be necessary. The surface geophysical survey is a valuable tool for investigating old town gas sites for two reasons: • It provides a method for locating buried storage tanks, buried lagoons, and other buried structures that may contain contami- nants.
------- • It provides a method for delineating contaminants (coal tar and other chemicals) in the soil and in groundwater. However, the ability to detect hydrocarbon compounds in soil and groundwater is limited generally to areas-where only high concentrations of these compounds are present. A number of surface geophysical techniques (ground-penetrating radar, electro- magnetics, electrical resistivity, magnetometry, and seismic surveys) can be used to provide preliminary information about subsurface conditions at contam- inated sites. Ground-penetrating radar can be and has been used to detect and delineate pools of organic compounds below ground. Howeve' , site conditions can inter- fere with the operation of tnii equipment, and it is difficult to predict where it can be used successfully. Applicability, cost, and equipment availa- bility may be factors determining its utilization at specific sites. Electromagnetic (EM) conductivity is an excellent technique for making a fast and efficient site survey of subsurface anomalies. It can locate old excavations (buried lagoons), buried tanks, pipes, and other metal objects. This equipment also can detect hydrocarbon compounds (tars and oils) in the ground if the compounds are present in high concentrations. Such concentra- tions are typically represented by low-conductivity measurements at the ground surface because these compounds inherently have very low electrical conductiv- ities. Although EM equipment can locate subsurface anomalies, it may not be able to determined accurately the size, depth, and subsurface condition caus- ing an anomaly. An electrical resistivity survey can be conducted in conjunction with an EM survey to confirm the EM anomalies and to better define the size and depth of the anomalies. Also, utilising the electrical resistivity equipment in a sounding and profiling array can help to define subsurface geologic conditions at a site. Electrical resistivity surveying can be used to delineate the depth of the water table as well as the presence of subsurface layers or len- ses of different permeability that have contrasting resistivities (e.g., clay and sand layers). However, electrical resistivity methods cannot be applied in certain geologic settings where general subsurface resistivity is relative- ly high; these methods are best used in areas (e.g., the Atlantic Coastal 223
------- if Plain) where electrical resistivities of subsurface materials contrast strong- ly (White and Brandwein, 1982). Further information on electrical surveying may be found in reports by the U.S. Environmental Protection Agency (1978) and| Freeze and Cherry (1979). Magnetometry may be used to detect buried metal objects at a site. Pipes, drum-j, buried tanks, and other metal objects may be detected by this method. At one gasworks site, a magnetometer survey was used to locate out- fall pipes running from a waste lagoon to a la«e adjacent to the site. Seismic refraction surveys can give valuable information about the depth to bedrock, the subsurface bedrock topography, and the condition (fracturing) of the bedrock (Cichowicz et al., 1981). In addition, the seismic velocity of a geologic material is altered by the degree of weathering and water satura- tion and therefore can provide information about the variability of these parameters in the subsurface. However, because of the multitude of variables that can affect a material's characteristic seismic velocity, seismic results can be difficult to interpret, especially in areas with complex subsurface geology or in areas where there is little contrast in seismic propagation velocities in the subsurface. For this reason, limited exploratory drilling usually will be necessary in conjunction with seismic surveys to confirm interpretations based on this technique (Cichowicz et al., 1981). More ( detailed information on seismic refraction surveying may be found in Dobrin (1960). The selection of geophysical techniques depends to a large degree on the geologic setting (White and Brandwein, 1982) and local site conditions. In general, surface geophysical methods can be utilized on most town gas facili- ties. However, there are certain sites where geophysical methods may not be appropriate because of local site conditions. Proximity to power lines, metal fences, railroad tracks, and buried utilities may make it difficult to proper- ly interpret geophysical data. In many cases, the type of geophysical tool best suited for a specific site is often difficult to determine without onsite testing. Further information on the application of surface geophysics to groundwater investigation may be found in Zohdy et al. (1974). 2.2.3.3.2 Soil sampling—Soil sampling includes soil-test borings and test pits, soil-water sampling, and soil-gas sampling. These activities are 224
------- the most important means to determine the extent and nature of contamination at a gasworks site. They provide samples for contaminant analyses and docu- ment the subsurface conditions at the site; extensive soil sampling is neces- sary prior to planning remedial actions at a site. A soil sampling program should be directed toward determining subsurface stratigraphy, properties of the subsurface materials that are important to contaminant transport (e.g., permeability, clay content, primary and secondary porosity), and obtaining representative samples of wastes and contaminated soil and water for analyti- cal characterization. This section briefly reviews the important aspects of a soil sampling program. For more information on soil sampling and monitoring, see U.S. EPA (1984a). A particularly important activity in a soil sampling program is to deter- mine the proper number, location, and depth of the soil borings. Existing information collected during the initial phase of a site investigation as well as surface geophysical results are extremely valuable in planning a site-spe- cific test-boring program. This program should be directed toward delineating the extent and characteristics of contamination at the site and in determining the characteristics of the subsurface soil and rock material. Soil-test bor- ings are typically drilled using hollow stem augers so that the borings can be converted easily to groundwater monitor wells. Also, this drilling technique minimizes the potential for aquifer contamination compared to other drilling processes. Down-hole geophysical methods can be utilized in soil-test borings where complex geology (including multiple aquifer systems) is anticipated. Various geophysical tools can be used to provide a variety of continuous down-hole data that is useful in determining the presence of contamination and inter- preting soil stratigraphy. Down-hole geophysical methods are especially help- ful in delineating relatively thin clay and sand layers that may not be detec- ted by discontinuous soil-boring sampling methods (Keys and MacCary, 1971). Test pits, usually constructed using backhoe excavators, allow for more complete inspection of subsurface conditions than do soil borings. Features such as vertical fractures or sand lenses, which may present pathways for contaminant transport and can be difficult to detect in soil boring, can be readily observed in test pits. Test pits offer a means to determine the 225
------- continuity and persistence of such features in the subsurface. They also may be used to delineate pockets of contamination and to investigate buried structures on the site. Test pits require the excavation of considerable amounts of soil. Because this soil can be contaminated, adequate provisions should be made prior to excavation for the safe handling, transportation, and storage of contaminated soil. Other reconnaissance techniques that may be used during soil sampling efforts are soil-gas monitoring and soil-water sampling in open boreholes and in the vadose zone. So~"!-gas monitoring is generally accomplished in one of two ways. One method •» /olves penetrating the partially saturated and capil- lary fringe zones au 'e the water table with a pressure-driven probe or auger through which soil gas is withdrawn and collected. Soil-gas samples are then analyzed for volatile components onsite, commonly with mobile gas chromatog- raphy, or taken to a laboratory for later analysis. An alternative soil-gas sampling method requires that passive vapor collectors be installed within 5 feet of the ground surface. The vapor collectors remain buried for a period of days to weeks; when exhumed, they are taken to a laboratory where the vapors are released and analyzed. Although both methods are relatively quick and inexpensive ways of qualitatively characterizing subsurface organic con- taminants, they are limited to compounds with relatively low water solubili- ties and high vapor pressures that are capable of diffusing through porous media. In general, soil-gas monitoring has little utility at sites that lack the more volatile fractions of coal tar, e.g., benzene, toluene, xylene, or naphthalene. If these components are present, however, soil-gas monitoring may prove successful in qualitatively characterizing the extent of contamina- tion at a site. Soil-water sampling is very similar to soil-gas sampling except that a water sample is collected. Drill-stem sampling collects the sample in open boreholes at the top of the water table. Drill-stem sampling offers some advantages over soil-gas sampling in that dissolved nonvolatile and volatile organic and inorganic contaminants can be measured. The method offers cost savings when compared to conventional groundwater monitoring techniques using permanent well installations. Soil-moisture profiling in the partially satur- ated or vadose zone can be accomplished by a modified soil-gas sampling probe j 226
------- or by a number of geophysical methods (e.g., neutron scattering or gamma-ray absorption). Once a sufficiently moist horizon is located,'suction lysimeters with porous clay cups can be installed in the vadose zone for sampling soil water. In practice, soil-gas sampling probes have a water-sampling capabil- ity, so the advantages of both methods can be combined. As stated previously, it is very important to take special care when using invasive site investigation techniques (e.g., borings, test pits) at abandoned gasworks sites to avoid penetrating or otherwise damaging buried structures such as tanks, gas holder foundations, or tar separators. These structures often contain tars, oils, or other contaminants. Structural damage could result in their release and spread of contaminants, complicating cleanup efforts. 2.2.3.3.3 Groundwater monitoring—The major objectives for installing a groundwater monitoring system are to: • Measure watpr levels for the purpose of determining gradient and direction of groundwater movement Perform in-situ permeability tests • Sample groundwater for chemical analysis. This section discusses the means to achieve these objectives with specific emphasis on monitoring considerations for abandoned gasworks sites. More detailed information on the design and installation of groundwater monitoring systems may be found in Barcelona et al. (1983), Barcelona et al. (1985), Todd (1980), Fetter (1980), Freeze and Chowy (1979), Johnson Division (1975), Villaume (1985), and NV/WA/API (1984). The number, spacing, depth, and well screen length of monitoring wells may be determined based on background information collected about a site and on the findings of the soil sampling and surface geophysical monitoring pro- grams. It is important to properly space the monitor wells across the site so that the gradient and direction of groundwater movement can be measured to determine groundwater flow directions and velocity at a site. On small sites it may be necessary to locate monitor wells offsi'te to discern measurable differences in groundwater levels. If multiple aquifers or perched water 227 r •;
------- table conditions are suspected, it is suggested that nested piezometers be installed at selected depths to measure vertical gradients. In-situ variable head permeability tests should be performed in selected monitor wells repre- senting various geologic conditions across the site. The permeability meas- urements along with the groundwater gradient data are useful in estimating the average velocity of groundwater movement across the site. For groundwater quality sampling and analysis, it is important to have a good distribution of monitor wells upgradient and downgradient from the suspected source of groundwater contamination. The upgradient monitor wells provide the background (uncontaminated) water sample. It may be necessary on small old town gas sites to use offsite wells upgradient of the site as back- ground wells. The downgradient monitor wells should be well spaced and have variable-depth well screens for the purpose of determining the vertical and lateral extent of groundwater contamination. It is also suggested that a downgradient monitor well be placed near the property boundary to determine if the suspected contaminant plume has migrated offsite. Variable density contaminants have been observed in the subsurface inves- tigations of several manufactured-gas sites and can result in complex contami- nant migration patterns in aquifers. The potential for variable density fluids needs to be recognized to the appropriate design of groundwater moni- toring systems at manufactured-gas sites. Adequate groundwater monitoring in flow fields with significant density contrasts requires careful monitoring well design and placement to avoid costly redrilling efforts or the creation of undesirable conduits for contaminant migration. Although single well in- stallations that are properly screened within a groundwater flow system may be adequate for some variable density situations, it may be necessary to supple- ment single wells with multiple-level sampling to fully characterize the ver- tical extent of contamination. It is also important to compensate measure- ments and sampling activities for differences in density where significant contrasts exist. Because the variable density contaminants commonly occur at abandoned town gas plants, special monitoring considerations for immiscible, multiple density fluids in gro- ndwater are discussed below. The relative density of potential contaminants at a gasworks site should be understood, at least qualitatively, before implementation of a groundwater 228
------- monitoring program. In some cases, the relative density contrast may be obvi- ous, such as with low-density (coal oil) or high-density (coal tar) immiscible contaminants. However, soluble components of the contaminant also may be present, especially when low-density immiscible contaminants occur (as discus- sed in Sections 2.1.1 and 2.3.3.3.5), and these need to be considered in the design of the monitoring system (Figure 63). In this example, the downgradi- ent well closest to the source area may encounter immiscible and soluble com- pounds, whereas further downgradient, the monitoring well will encounter only soluble compounds. A multilevel groundwater sampler would be useful in this example to detect migration of the soluble component and its stratification within the groundwater. Lysimeters or similar in-situ pore-water samplers might be useful in delineating the dimensions of the contaminant plume above the water table (Figure 63). In other situations, contaminants migrating from a gasworks site muy consist of constituents with multiple densities (Figure 64). In this example, downgradient well A will detect an intermediate density zone, and well B will detect the higher density zone. A multilevel sampler (well C) can be used to further delineate the two relative density zones. The position of the screened interval of monitoring wells (or intake ports of multilevel samplers) is one of the most important aspects of detect- ing variable density contaminants in the subsurface. This is illustrated in Figure 65 where examples of appropriate and inappropriate monitoring tech- niques are compared for variable density contaminant situations in a uniform flow field. In example 1, the high-density contaminant solution could be overlooked as a result of shallow screen settings of the monitoring wells. Deeper-screened settings would be more appropriate in this example (nested wells A, B, and C), or a multilevel sampler (well D) would allow for more complete definition of the vertical extent of contamination. For example 2, the low-density immiscible contaminant could be largely overlooked if screened intervals were too deep below the water table. Shallow monitoring wells would be more appropriate in this situation, particularly for defining the depth of the depressed water table. In example 3, the contaminant solution has a simi- lar relative density as the groundwater, but it is not detected by the shallow screen setting of well A. The long screen interval of well B intercepts the 229
------- 230
------- Of_ < (L X \ - t_ t> 1 xi-.-'l T '•"••' i •'•;• • '.'"'i.' •>-.»> ;•>•>?.• JL»' • .• f*-'*""'Si \; V*"* ' t SiiiffiMl'' • /^sffi-^i ' ; ; liO-xw^S* • ,*.-;fi»** » I t l*&>j£i»i'*"l'ir-'il* I . 1 - : > ^™***?*'Js-» *^'W*. *» i • * * t y .i«?^ji:^:V / !^i^d:^ , . , . / .:: ».,*;«•:/ 1 ... i/ i-i.:''*. « . 4,^1 '^•:1«^- ^rfS— 1 « g '£ O «S2T •s's .? u XO ' III III III III III III III !!! lit i'l'i' I'i'i1 !! I'i'i1 rr I I'l'l1 iii i'l'i ill IM'I1 ill I'l'l 114 I'l'l' "' "I1 III 111 (MM 1 ' ' 111 III III Ml'l' i.i.l, 111 ft! Ill '••.I i'l'i1 »' " ' >" I1!1!1 :!:!'! Ill I'l'l' 1 III II !'!; i i 111 ill II iti 'I'i'i i.u HP HP Hi in 'i'l'i iiifi iii ri'i1 I'l'i' IM'M id ijiji 1.1,1 i'i't' Hi i1!1! 111 I'IM' iii mi ijiji iji'ij K !!! it iii Hi LI!I!I 231
------- Inappropriate Monitoring Appropriate Monitoring Example 1. High relative density contaminant solution. Example 2. Low relative density immiscible contaminant. Example 3. Contaminant solution with similar relative density as ground water. Source: Alexander, 1984. Figure 65. Comparative groundwater monitoring of variable density contaminants in uniform flow field. 232
------- contaminant plume, but it also draws in uncontaminated groundwater, as does well C. The results of groundwater analysis from these wells may not reveal the presence of contamination because of dilution of the samples. More appro- priately, the carefully screened intervals of wells A and B would detect the contamination, but that would require prior knowledge of the plume's vertical extent. This knowledge could be gained by the installation of a multilevel sampler (well C). The presence of high-density tars at gas sites in the subsurface requires special care when constructing monitoring wells into deeper aquifers below a site. These wells can provide pathways for such tars to move, under density gradients, into deeper aquifers, even against an upward hydraulic gradient between the confined aquifer and the surface. At St. Louis Park, Minnesota, coal tar flowed down a multiaquifer well, resulting in contamination of multi- ple aquifers (see Chapter 3). Because of this contamination potential, moni- toring wells for aquifers beneath a gasworks site should not pass through zones of tar contamination. If tar is encountered during the construction of such a well, the well should be moved to an area with no underlying coal tar. If this is not possible, extreme care should be taken to seal off the tar- containing zone to prevent migration of tar into the borehole and down into the aqui fer. 2.3 SITE REMEDIATION 2.3.1 Introduction Remediation options for gasworks sites are basically the same as those for other Industrie.; hazardous waste sites: no action; onsite containment, with or without stabilization or fixation; removal and disposal of contamina- ted material; in-situ treatment; removal and treatment or destruction of con- taminated materials. The selection and implementation of remedial alterna- tives for specific gasworks sites are the same as for other hazardous waste sites. This discussion does not go into detail about site remediation. Instead, it concentrates on the unique features of gasworks sites that may affect site remediation, case studies of actual gasworks site remediation, and listing remedial action alternatives for specific gasworks wastes. For more information on the selection and evaluation of remedial action alternatives 233
------- for specific sites, the reader is referred to Cochran and Hodge (1985a, 1985b), Boutwell et al. (1985), U.S. EPA (1982), Ehrenfeld and Bass (1983), and Sims et al. (1984). 2.3.2 Factors Affecting Site Remediation Gasworks sites have certain unique features that can influence the selec- tion of remedial alternatives. First, the sites are old: Many were abandoned more than 50 years ago, and almost all are more than 30 years old. This age can affect remediation in several ways. It can result in a low-priority rank- ing for the site in terms of cleanup. If the site owner can demonstrate that there is no history of contaminant migration and that wastes currently are remaining onsite, it is possible that site remediation efforts could be post- poned without damage to human health or the environment. The fact that a site has existed for decades without problems may be taken as evidence that post- poning remediation will cause no further problems. If cleanup is postponed, however, groundwater monitoring should be employed to detect contaminant release, and measures such as restricted site access should be taken to avoid exposure of the public to contaminants at the site. On the other hand, the age of these sites can afford a long period of time for contaminants to move offsite, thereby resulting in a significant spreading of contaminants and an increase in the volume of material that must be cleaned up. This was the case at Brattleboro, Vermont, where coal tar has moved through a porous gravel layer along a bedrock surface, underneath a river adjacent to the former plant site. At St. Lou.is Park, Minnesota, where a coal tar refinery operated for more than 50 years, contaminants have spread to several aquifers to a depth of over 900 feet, and a plume of contaminants extends over one-half mile from the site. At Ames, Iowa, lighter tar constit- uents from a gas plant closed in the 1930's have contaminated the municipal well field, resulting in the closure of five municipal wells since contamina- tion w.->s first detected in 1927. In contrast, at Stroudsburg, Pennsylvania (Brodhead Creek), favorable geological conditions resulted in the containment of over 8,000 gallons of free coal tar in the subsurface for about 40 years, until excavation of the adjacent creek bank caused release of the tar into the creek. 234
------- Many of the case studies reviewed in this study (see Chapter 3) illus- trate the fact that gasworks were often built in low-lying areas adjacent to . ( waterbodies or wetlands. In some cases, discharges into these waterbodies resulted in a site discovery. Proximity to waterbodies or wetlands can require barrier construction to prevent surface water contamination during site remediation. In addition, contaminants may have been disposed of or migrated into these waterbodies, which can result in accumulation in river or lake sediments. This could necessitate underwater cleanup operation, compli- cating and increasing the cost of site remediation. Gasworks also usually occur in downtown areas or old industrial dis- tricts. The recent trend to redevelop these areas has resulted in the discov- ery of many former gasworks sites across the country. Redevelopment pressures and priorities can affect site remediation efforts and vice versa. The following cases illustrate how redevelopment and remediation were handled in different areas of the country. In Newport, Rhode Island, two multimillion dollar apartment buildings were being constructed across the street from one another when tar from a former gas plant was discovered in the subsurface at both construction sites. One building was being constructed on pilings. The only contaminated material removed from this site was that actually excavated for the pilings. It was disposed offsite, and the lower floors of the building were designated for nonresidential use (parking garage). At the other site, a buried concrete structure was discovered and accidentally ruptured during construction of the foundation. It was full of coal tar. In this case, the structure was repaired, the coal tar left in place, and a ventilation system installed to prevent organic vapors from accumulating in the basement of the apartment buiIding. In San Francisco, California, coal-tar contamination was encountered during construction of an addition to EPA's Region 9 headquarters. This mate- rial was removed and disposed in a secure landfill. There was suspicion that the soil under the existing building also could be contaminated, but this has not been verified. Cases of contamination discovery under existing buildings constructed after a gas plant was removed were not uncovered in this study. However, the 235
------- downtown location of most plants makes the existence of such a situation pos- sible, if not probable. The presence of an existing building over a contami- nated gasworks site would be considerably complicated and could prevent reme- diation of a site. In such cases, onsite containment may be the best option. Case studies in Chapter 3 that illustrate the interaction of redevelop- ment and site remediation are GasWorks Park, Seattle, Washington; Brattleboro, Vermont; Plattsburgh, New York; Everett, Massachusetts; and Mendon Rd., Attle- boro, Massachusetts. When gasworks were decommissioned, surface structures often were removed, but structures below the surface usually were left in place. These structures often contain contaminants, usually tars, oils, or tar/water emulsions. Because of this, it is important to determine the locations of these struc- tures during a site investigation and to consider their locations when plan- ning site remediation activities. In some cases, free tars and oils occur in these structures; such gasification byproducts may be reused as supplementary boiler fuel or chemical feedstocks. If reuse is not a viable alternative, careful recovery of the material from the structures results in a more concen- trated waste stream for treatment or disposal. If subsurface structures are damaged during remediation efforts, contamination can spread into surrounding soils, increasing the expense and complexity of remediation efforts. Another feature of gasworks sites that can affect remediation efforts is the presence of injection wells that were used for waste disposal (e.g., for tar residues and emulsions). At least one site reviewed in this study, Stroudsburg, Pennsylvania, may have had one of these wells. Research by the Stroudsburg site investigators suggested that other gasworks in the area may have used wells for waste disposal. Maps for the Lowell, Massachusetts, plant showed a "deep well" on the site. However, it is not clear whether this well was used for waste disposal. Additionally, it is important when reviewing old site maps not to confuse tar wells, which are underground structures containing tar, with injection wells used for disposing of wastes. The location and depth of all wells on a site should be •>.••*c.-.:ined during remedial investigations. These wells nay be reopened and sampled for contami- nation. Care should be taken during reopening to prevent them from adding to the spread of contaminants. If no contamination is detected, they should be 236
------- properly closed and sealed to prevent them from becoming pathways for contami- nant migration. If contaminated, they can complicate site remediation efforts. However, if wastes were pumped down a well, it may be possible to pump them back out. This was accomplished at Stroudsburg, where over 8,000 gallons of free coal tar was removed from the subsurface. However, consider- able tar remains bound up in subsurface material at Stroudsburg; this necessi- tated containment (slurry wall) to prevent migration of contaminants offsite. 2.3.3 Remedial Action Alternatives 2.3.3.1 Introduction-- As previously stated, remedial action alternatives for gasworks sites are similar to those for other uncontrolled hazardous waste sites. Containment, removal and disposal, and treatment all are applicable. Some containment generally will be required for all remedial actions to prevent the release and spread of contaminants. Slurry walls and caps have been used to contain gas- works wastes. Removal and disposal is a simple but expensive option that also has been used to clean up gasworks sites. Treatment to stabilize, detoxify, or destroy gasworks wastes has not been employed to a great extent, but it is attractive because it can destroy a waste's hazardous nature, enabling safe disposal of residues in nonhazardous waste landfills and eliminating future liability. Treatment alternatives with potential applicability to gasifica- tion wastes are summarized in Table 47. The following discussion focuses on remediation techniques actually ap- plied to gasification wastes or similar substances. For more general informa- tion on the evaluation and selection of remedial action alternatives, the reader is referred to the Deferences listed at the end of Section 2.3.1. 2.3.3.2 Oils, Tars, and Lampblack-- The most prevalent and persistent contaminants at gasworks sites are organic byproducts of the gas manufacturing process--tars, oils, and lamp- black. Tars and oils could be produced in any process; lampblack was most commonly produced in oil-gas processes. Tars and oils can contaminate soils and groundwater (see following sections), but they also occur as free products at gasworks sites, especially in buried tanks and other structures, buried lagoons, and in coarse sands and gravel in the subsurface. Lampblack may 237
------- TABLE 47. POTENTIAL TREATMENT TECHNOLOGIES FOR CONTAMINATED SOILS Method In-situ methods Neutralization Solvent extraction Chemical oxidation Immobilization Attenuation Description Applicability Addition of base to soil to neutralize acid; base could be spread or injected into soil as a solution or spread as a powaer and tilled into soil Flush with chemical solution to remove contaminants, then collect and treat solvent; solvent could be acidic, basic, or surfactant, injected or percolated into soil and collected in drain or with- drawal wells Addition of chemicals such as ozone or peroxide to break down compounds into harmless forms or forms more readily attenuated by natural microbial activity; lack of selectivity may lead to high dosage requirements Reduces rate of release of con- taminants into environment; pH adjustment or chemical addition promotes sorption or precipita- tion onto organic materials such as sawdust or agricultural byproducts; may have already occurred at gas manufacturing sites through reaction with organic "fluff" Mixing of contaminated soils with clean soil, municipal refuse, or sewage sludge; may be acceptable for low-risk wastes, also may promote natural biological degradation Acids or acid-forming wastes Organics or metals, depending on solvent Primarily organics, may mobilize metals, requiring leachate collection and treat- ment Metals and organics Compatible wastes of low mobility and toxicity (continued) 238
------- TABLE 47 (continued) Method Description Applicability Biological oxidation Occurs naturally in soils; may be enhanced by addition of nutrients, oxygen, or specially developed microbes; contaminants are metabolized by bacteria and/or fungi to harmless forms Methods requiring excavation Thermal treatment Evaporation Incineration Chemical treatment Biological treatment Landfarming Composting Contaminated soils heated to drive off volatiles that are destroyed in an afterburner Entire waste matrix heated to over 1,000 °C to destroy con- taminants Neutralization, extraction, oxi- dation, immobilization similar to description under in-situ methods, carried out in a reactor under controlled conditions Waste incorporated into upper layers of soil, biological degradation stimulated, cover or livestock feed crops grown Waste biologically stabilized above ground, may be mixed with municipal refuse or sewage sludge; result may be used as a soil amendment Primarily organics although sulfur and nitrogen also may be oxidized Organics, cyanides, sulfides; auxiliary fuel required Same as evaporation Various wastes Organics, cyanides; not suitable for wastes containing heavy metals, which may build up in soil or crops Primarily organics or cyanides; disposal of metals depends on final disposition of product SOURCES: Sims et al., 1984; Hoogendoorn, 1984. 239
------- occur at or below the surface. If these contaminants can be recovered in pure form, they may be reused as supplementary fuels or chemical feedstocks. Alternatively, they are easily incinerated. Removal of tars or oils from underground containment structures is rela- tively straightforward, but care must be exercised to avoid rupturing the structure. Once the substances have been removed from tanks, the tanks can be either cleaned using steam or aqueous surfactants or removed and disposed of. At Stroudsburg, 8,000 gallons of free tar was pumped from the ground using techniques developed for control of distribution of tar in the surface (Villaume et al., 1983; Roberts et al., 1982). Poor understanding of these phenomena at Stroudsburg resulted in overestimation of free coal tar in the subsurface and overscaling of the coal tar recovery system. Original esti- mates of 35,000 gallons of free tar resulted from a failure to realize that the tar was present in several different "phases" or zones. Much of the tar was held up in the subsurface by capillary forces so that no coal tar could be removed by pumping, or it was associated with water in a fashion that would result in recovery of coal tar and water if this zone was for heavy oil recov- ery (see case study in Chapter 3). In this case, recovery by pumping was possible because the tar was contained in a coarse, highly permeable aquifer that enabled it to move relatively freely. The feasibility of this approach at other sites may be determined from the characteristics of the porous medium (e.g., porosity and permeability), the characteristics of the tar (e.g., vis- cosity, density, interfacial tension between tar and water, and wetting angle of tar on aquifer material in the presence of water), and an awareness of how viscous and capillary pressure forces can be pumped. Figure 66 illustrates the zoned distribution of water and coal tar in the subsurface at Stroudsburg, inferred from capillary pressure theory, and it indicates the types of material that may be pumped from the different zones. Failure to perform this sort of analysis can result in overestimation of the amount of free tar in the subsurface; tar in water emulsions and tar held by capillary forces in the subsurface material may be included in the free coal tar estimates. As previously mentioned, free products recovered from gasworks sites may be used as fuel (as at Str.oundsburg) or as chemical feedstocks. It also may 240
------- PUMP PRODUCT WATER WATER WATER AND COAL TAR COAL TAR WATER LIQUID PHASE(S) T GROUND WATER WITH DISSOLVED ORGANIpS GROUND WATER AND TRAPPED COAL TAR GROUND WATER AND COAL TAR COAL TAR AND 'IRREDUCIBLE' WATER GROUND WATER WITH DISSOLVED ORGANICS POROUS MEDIUM '-". *. •:..'.'.. V • •::*:'.• :.••• . » • . « . «• . , :••.»:•:•:••'• -::••••••••'••• ,.'•••.' • .* M K °.* * » • » V ( ••:•••. V.:V /.•>:: Y.-V-V::.-: •.•••*. •!•;.- • • • •*. .»*"•>. • ; t ?>>:d?:-: * •* B *"*"••* * * ". ' *f ' * * * ••*.*^ »••« .•..•..••oc ..:/;. •••:-••: °.t-v" .V'f.'o k-:' • - • . z «... • '..;•:• :.• < .'.-*.• "•:.• :•«•• .;..v, Q ....... • . o . . Z «' • :•.•..-. <•..*-.. ••••"« 0} '.•"• ••••-• T: ; .. .-,.. -,>v. . •*.'.' •••". * • -. •.•:.:.-•;.-.•. • .:::>.:>::••.: •:::•••.•.•••••'» » * • • • • • *• * * • • • • • * «. • • * » ' o k • ...-;. •".••••..• •-, • •*. *••. ;** •. ••.•.•:'•".;*••• * *.• •* ""*•« "• \. ••-.V.'.Ai:.".-...^ '• -..':.?• V " '*•'•.• *r' SILTY i-K, &.- SAND .X. 5;fe|%fvi^> :.0::;-r-Vv-'- •••>••:..•.•.'..•.':•'>.*•-; !;Ur • • ' » • • j' *.' Source: Villuame et. al., 1983. Figure 66. Ideal distribution of coal tar in porous materials at the Stroudsburg contamination site, as inferred from capillary pressure theory. 241
------- be easily incinerated (heating value -17,736 Btu/lb) or may be amenable to land treatment, as described in the following section on contaminated soil. Lampblack is solid, sooty material that was commonly produced in oil-gas plants. It is composed of very heavy organic compounds, including PAH. It is essentially immobile and insoluble in the subsurface. Because of this, it can be safely contained onsite, as was done at an unnamed site in southern Cali- fornia. If it is removed, it can be briquetted and used as solid fuel or possibly used as a blackening agent in certain industrial processes. Alterna- tively, it may be easily incinerated. Lampblack contains PAH's and is car- cinogenic; its powdery form makes it necessary to exercise care to prevent dust emissions when excavating and handling the material. Inhalation and skin contact also should be avoided. 2.3.3.3 Spent Oxide Wastes-- Spent oxide wastes, as described in Section 2.1.2, are extremely hetero- geneous in nature from site to site and within specific sites. This variabi- lity occurs both in terms of the wastes' physical characteristics and types of contaminants that may be present. Because of this variability, and because they have not been extensively characterized by composition or occurrence, it is difficult to evaluate remedial alternatives for these wastes. This discus- sion concentrates on the characteristics of the wastes that can affect their treatment and handling during remedial actions and on two cases in which sites containing spent oxide wastes were remediated. Spent oxide wastes are pyrophoric, i.e., when exposed to air they have a tendency to self-heat and spontaneously combust. For instance, Downing (1932) reports: t The disposal of spent oxide is a vexatious problem for many gas plants. Because of a possibility of fires starting through heat generated by revivification, it is necessary to hold the spent mate- rial at the plant until this danger is past. As soon as city authorities learn of this menace the material is prohibited at pub- lic dumps. Continuous storage on gasworks land eventually becomes impossible. The material makes excellent filling for roads or pri- vate property when properly handled. It should be covered with ashes or soil immediately to prevent the access of air and conse- quent combustion. 242
------- M This pyrophoric nature probably is due to the presence of reduced sulfur com- pounds that oxidize exothermically when exposed to air. At one unnamed gas- works site, a gas cleanup box that was left full of oxide material years ago when the plant closed was opened during site cleanup activities. It subse- quently caught on fire. In this case, the fire was easy to extinguish because it was contained. However, care should be taken to avoid combustion when excavating, moving, or storing spent oxide wastes at a gas plant site. The material should be covered as much as possible with soil, plastic, or other material to prevent contact with air. In addition, when it is to be stored or transported, it should be carefully placed and compacted into the pile or transportation vehicles to prevent, air from permeating the waste materials. Alternatively, it may be possible to separate combustible materials (e.g., woodchips) from the sulfur-containing oxides to prevent combustion of these materials. Physical separation, followed by incineration of the combustible material, may be an appropriate alternative for treating these wastes. Spent oxides can have elevated levels of arsenic associated with wastes from the Thylox gas cleanup process. They also have significant acid-generat- ing potential, leachates from these wastes having a pH of 1.5. This low pH can result in release of arsenic or other trace metals. At the Birmingham, Alabama, gasworks site, arsenic levels of 8.0 mg/L were reported for 1.5 pH leachate from spent oxide wastes that contained 160 ppm arsenic (Harry Hendon and Associates, Inc., 1982). Total cyanide levels as high as 8,900 ppm were measured in spent oxides at the Birmingham site. However, the highest levels of free cyanides in water reported at sites contaminated with these spent oxides was 2.6 ppm for a sam- ple with a pH of 1.5 (Harry Hendon and Associates, Inc., 1982); free cyanide levels less than 1 ppm were more commonly associated with spent oxide wastes at Birmingham. This is because most of the cyanides are present as complex iron cyanides. These compounds are very stable in the environment and have a low toxicity. They do appear to release small concentrations of free cya- nides; however, these concentrations are well below the 200 ppm level that limits degradation of free cyanides in aerobic soils, and most are below the 2 ppm limit for the anaerobic degradation of free cyanide (Fuller, 1984). The persistence of complexed ferric ferrocyanides remaining for decades in spent 243
------- oxide wastes disposed at or near the surface is further evidence of their stability in the soil environment. The persistence of the cyanide materials in spent oxides and the relative stability of ferric ferrocyanide compounds is an encouraging observation from the standpoint of treating these wastes. Although the complete destruction of cyanides in spent oxide might be the most ideal solution, the cost associated with destruction options, along with potential for the liberation and release of free cyanide during treatment, may make stabilization or fixation a more desirable choice. The long-term survival of ferric ferrocyanides at gas plant sites, along with the use of this material in table salt, highway deicing salt, paints, pigments, and laundry bluing, suggests that treatments to elimi- nate any hazards under the Resource Conservation and Recovery Act (RCRA), and containment onsite or disposal in a municipal landfill may be an environmen- tally acceptable and cost-effective alternative for dealing with these wastes. At the Alabama Gas Corporation Gas Works site in Birmingham, Alabama, in-place stabilization was selected as the remedial alternative for an onsite spent oxide disposal area (Harry Hendon and Associates, 1982). Stabilization of the 2.4-acre site involved excavating and stockpiling the contaminated material, then mixing agricultural lime (CaCC>3) and soil in 1-foot lifts across the site, not exceeding 80 tons of lime per acre-foot of soil. In addition to lime, fertilizer and sewage sludge was added to the top 6 inches to promote the growth of vegetation. The lime neutralized the acidic condi- tions formerly present at the site, thereby reducing trace metal (As) release to environmentally safe levels. The remediation plan was successful: The once barren site has been revegetated, and soil samples indicate that acidic conditions and high arsenic concentrations have abated. The cost of remedia- tion was about 5100,000; removal, disposal in a secure landfill, and refilling was estimated to cost S2 million to $5 million. At the Mendon Road site in Attleboro, Massachusetts, 1,083 yd3 (about one-third of the volume of material at Birmingham) of spent oxide material from gasworks manufacture had been disposed in an abandoned gravel pit. The site was discovered during residential development of the area. The waste was similar to that found at Birmingham (pH = 1.61; total cyanide = 7,500 ppm, free CN~ = 0.7 ppm) except that high arsenic levels were not detected and low 244
------- ppm levels of PAH compounds were found. The waste was excavated, removed from the site, and disposed in a secure hazardous waste disposal facility ?t a cost of over SI.6 million. The difference in costs in the two spent oxide site remediations is not insignificant. In-place stabilization appears to be a desirable remedial alternative for cyanide-containing spent oxide wastes on both technical and cost bases. If site use plans rule out onsite stabilization as a viable alternative (as at Mendon Road), removal, stabilization, and disposal at a nonhazardous waste landfill may be an environmentally acceptable alternative that is more economical than disposal as a hazardous waste. Studies demon- strating the low mobility in soils of ferrocyanides in municipal waste leach- ate suggest that stabilization and disposal in municipal landfills may be acceptable (Fuller, 1984). However, more research is needed on the mobility of complex iron cyanides before this can be proven safe. In addition, the extreme variability and heterogeneity of spent oxide wastes necessitate waste- specific evaluations of remedial alternatives. Other methods for treating cyanide-containing wastes are discussed in the following section on remedia- ting contaminated soils. The characteristic blue color of complex ferric ferrocyanides can be used ^ both to identify areas of spent oxide contamination during site investigations™ and to guide remediation efforts; however, some question exists as to cr^or- threshold-contaminated levels. At the Mendon Road site, color was used to delineate contaminated soil with greater than 2 ppm total cyanide during cleanup efforts. Wilson and Stevens (1981) report that blue color may be detected in soils containing about 270 ppm total cyanide (or 500 ppm ferric ferrocyanide). Further analyses of samples of soil contaminated with complex iron cyanides is necessary to resolve this discrepancy. Spent oxide wastes that do not contain complex cyanides are usually red to yellow. They may be more common at U.S. gas plant sites than are cyanide- containing wastes because of the prevalence of water-gas and oil-gas processes that produced gas that characteristically had low levels of cyanide compounds. The major hazards associated with these wastes is their acid-producing poten- tial and their potential to release toxic trace elements. These hazards may be reduced by additives, such as CaC03, that can reduce acid and limit trace metal release. 245
------- Spent oxide materials may be contaminated with tar and/or may have been codisposed with tar-contaminated shavings from shavings scrubbers used during gas cleanup to tar mist prior to the oxide boxes. At one site visited by the authors, oyster shells contaminated with tar were seen onsite; these probably were used in place of shavings for tar removal. Methods for treating solid materials contaminated with tars and oils are discussed in the following sec- tion. 2.3.3.4 Contaminated Soil-- Our review of gas plant site investigations revealed that the most com- monly occurring soil contaminants are byproduct tars and oils from gas manu- factured. Spent oxide waste containing complex iron cyanides, sulfur com- pounds, and arsenic is another significant but less prevalent soil contami- nant. Treatment techniques that may have applications at gas plant sites are summarized in Table 46. A complete review of treatment technologies for contaminated soils is beyond the scope of this study. The following discus- sion considers techniques actually applied on contaminated soil from gasworks plants or on soils contaminated with substances similar to gas plant wastes (i.e., creosote). More information on soil treatment techniques in general may be found in Sims et al. (1984), Hoogendoorn (1984), Cull inane and Jones (1984), Spooner (1984), Rulkens and Assnik (1984), and Wagner and Kosh (1984). 2.3.3.4.1 Land treatment—The land treatability of PAH-contaminated soils and PAH-containing sludges has been demonstrated for petroleum refinery wastes (API, 1983) and for creosote used by the wood-preserving industry (Sims, 1984; Sims and Overcash, 1983; Umfleet et al., 1984; Patnode et al., 1985; Ryan and Smith, 1986). The fractional distillation of creosote from coal tar (creosote has a 200 to 400 °C distilling range), suggests that land treatment will be effective in treating soils contaminated with gasifier tars and oils. Comparison of contamination removal rates for creosote wastes and refinery wastes shows good agreement (Ryan and Smith, 1986); this implies that the land treatability of PAH-containing hydrocarbons is similar regardless of their source. Currently, the wood-treating industry and the U.S. EPA are sponsoring studies to demonstrate the land treatability of creosote sludge and creosote- 246
------- contaminated soils [Ryan and Smith, 1986; R. C. Sims, Utah Water Research Laboratory (UWRL), personal communication, 1986]. At one site in Minnesota, bench-scale and pilot-scale field tests have demonstrated the feasibility of land treatment of creosote-contaminated soils (Patnode et al., 1985; Ryan and Smith, 1986). Important results of this study are: • Percent removals of benzene-extractable hydrocarbons averaged about 40 percent over 4 months, with a corresponding first- order kinetic constant of 0.004. • Complete toxicity reduction appeared to fall between 2.5 and 5.0 percent benzene-extractable content. Two out of five test plots were nontoxic after 4 months (those with lowest initial application rates). All plots showed significant degradation. • Microbial assays suggested that initial concentrations of creo- sote compounds would kill soil microorganisms and inhibit de- gradation. This did not occur. In addition, seeding plots with adapted microorganisms did not significantly enhance de- gradation. This implies that an active, adapted microbial population naturally developed in the contaminated soil. • Within the range of loading rates tested (4 to 10 percent ben- zene extractables), no correlation between loading rates and kinetic rates was observed, with the exception of 4+ ring PAH compounds, which showed a slight inverse relationship between loading rates and kinetic rates. • All loading rates tested (4 to 10 percent benzene extractables) were feasible. • loxicity reduction occurred at a faster rate at 4 to 5 percent initial loading rate than at higher loading rates. • Greater kinetic rates were observed after waste reapplication to a treated soil. • At this site, 3 to 5 years would be necessary to treat 12,500 tons of contaminated soil. • Waste application rates of 2 to 3 pounds of benzene extracta- bles per ft3 of soil per 2 months can be degraded. This study demonstrates the feasibility of land treating sandy soils contami- nated with creosote wastes in Minnesota. Treatment times should be lower in warmer areas with a longer growing period. Preliminary results from an on- going study in California suggest similar kinetic degradation rates in clayey soils (Ryan and Smith, 1986). 247 u.
------- TABLE 48. COST ESTIMATES FOR REMEDIAL ACTION ALTERNATIVES AT A CREOSOTE IMPOUNDMENT Alternative Land treatment (onsite) Landfill Incineration (onsite) Incineration (offsite) a!2,500 tons contaminated material. SOURCE: Patnode et al., 1984. Unit cost (S/ton) 51 200 184 1,900 Total cost9 (SI, 000) 738 2,500 2,300 23,750 248
------- One of the most significant results of the Minnesota study is that onsite land treatment is very cost-effective. Table 48 compares the cost estimates of land treatment with other options (i.e., landfill and incineration); land treatment cost estimates were lowest at 551/ton. If onsite conditions are not amenable to land treatment, costs will increase as a result of transportation costs to a suitable treatment site. However, even if this results in costs higher than landfilling, land treatment will still be preferable because it can detoxify the waste, thereby eliminating long-term Viability. Comparison of onsite land treatment costs with onsite incineration (Table 48) demon- strates that land treatment is more cost-effective. Our review of remedial alternatives for soil contaminated with tars and oils from gas plant manufacture indicate that land treatment is the best demonstrated treatment technology. It appears to be cost-effective, as well as effective in detoxifying the wastes. The age of all gasworks sites further supports this conclusion because soil microbes capable of degrading tar and oil compounds will have had time to evolve. The Ames, Iowa, case study (see Chapter 3) demonstrates this; organisms capable of degrading PAH compounds have evolved in the groundwater at Ames. Several questions remain unanswered with respect to applying the results of the creosote studies to gas plant residuals. First, creosote is a distil- late fraction of coal tar; the tars and oil at former gas plants tend to have a broader boiling point range. In addition, creosote is derived from coal tar; most gas plants operated water-gas processes, which produced tars with different composition (e.g., no tar acids or bases), it is not clear how this will affect soil toxicity and degradation rates. It does seem possible that soil microbes will have adapted to whatever tar constituents are present at a site. Other soil contaminants present at gas plant sites also could affect the land treatability of contaminated soil. Complex iron cyanides are not amenable to land treatment (Hoogendorn, 1984); free cyanides are rapidly broken down by soil microbes at concentrations below 200 ppm; and, as long as complex iron cyanides do not release free cyanides at rates sufficient to elevate soil levels to above 200 ppm, they may not affect degradation. Sulfur and arsenic compounds also may be present and could influence degradation rates. Another question is the volatilization of volatile components in coal 249
------- tars and oils during land treatment operations. These questions can be addressed by site-specific land treatment demonstrations such as those required for permitting a facility under RCRA (40 CFR 264). Studies to demonstrate the treatability of contaminated soils and tars and oils should include bench-scale and pilot-scale tests to evaluate the effect of various design and operational parameters on the treat.abil ity of the wastes in question. These parameters include: • Soil characteristics • Waste characteristics • Treatment supplements Climate Initial loading rate • Reapplication rate Soil lift thickness • Frequency of tilling. Treatability studies should be directed toward determining the effects of these parameters on the reduction of organics, PAH's, and toxicity for the wastes or contaminated soils to be treated. M When conducting a treatability study, soil conditions that promote the degradation of hydrocarbons should be maintained. These conditions include (Ryan and Smith, 1986): • Soil pH of 6.0 to 7.0 in the treatment zone • Soil carbon-to-nitrogen ratios of 25:1 • Soil moisture .near field capacity. Other criteria that have been recommended for land treatment of creosote wastes include: • Small and frequent fertilizer applications • Waste reapplication only after initial applications have been effectively degraded. U.S. EPA has published general guidance on land treatment demonstrations (EPA, 1984a; EPA, 1983a; and EPA, 1983b). EPA also has released a draft tech- nical guidance manual on hazardous waste land treatment demonstrations for 250 ™
------- public comment (EPA, 1984b). This latter document currently is being revised to address and incorporate the public comments (R. C. Sims, UWRL, personal communication, 1986). It should be stressed that each of these EPA documents presents guidance only and not regulations. The detailed design of a land treatment unit for gasifier wastes will depend on the conditions at the specific site. Although onsite land treatment is most economical, the location of many former gas plants in populated, urban areas may preclude onsite treatment. Regardless of whether treatment is to be conducted onsite or offsite, the contaminated soil to be treated must be excavated and stockpiled at the treat- ment site. The stockpile may be covered and placed on a liner to prevent spread of contamination. The treatment area should be lined, and a leachate collection system installed, to prevent migration of leachate. The contami- nated soil is then laid down in 1 to 1.5 foot lifts, and soil amendments and water are added as necessary to reach and maintain optimum soil condition for degradation (determined in bench-scale and pilot-scale studies). It may be necessary to blend clean soil with the waste or contaminated soil to achieve the desired contaminant loading rate. The soil should be cultivated regularly during the treatment process; soil conditions (moisture, pH, nutrients, etc.) should be carefully monitored and controlled. Once the initial lift has been detoxified, a second lift is placed on the previous lift, and so on until all the soil is treated. Leachate collected from the land treatment facility may be treated or discharged without treatment, depending on the level of contaminants. At the Minnesota creosote treatment site, the State and EPA permitted discharge of leachate either into the Mississippi River or into the municipal sewage sys- tem, depending on the level of PAH compounds in the leachate. This implies that dissolved PAH's may be successfully treated in municipal wastewater treatment plants. Land treatment is therefore a well demonstrated, effective technology for degrading PAH compounds. Field and bench-scale treatability studies on creo- sotes have demonstrated that a range of initial loading rates are acceptable and that degradation time increases with increasing loading rate. The selec- tion of loading rate should balance land area requirements and time require- 251
------- merits for completing the treatment process (Ryan and Smith, 1976). Lower loading rates decrease the time required for degradation, and higher loading rates decrease the land area requirements. Further information on the design A and demonstration of land treatment may be found in Overcash and Pal (1979), API (1983), and EPA (1983a and b, 1984a, b, and c). 2.3.3.4.2 Extraction or thermal treatment of excavated contaminated soil--Hoogendoorn (1984) and Rulkens and Assnik (1984) reported on the succes- sful pilot-scale use of a hot aqueous alkali solution to clean gasworks soil contaminated with free and complexed cyanides. The process (Figure 67) has been scaled-up to 25 tonnes/hr and is estimated to treat soils at a cost of S24.80 to S99.20/m3. Soil is pretreated to remove large objects (wood and stones) and to break up clods. It is then extracted with a lye solution, the soil and cleaning agent are separated, and the extraction agent is cleaned by pH adjustment, coagulation, flocculation, sludge separation, sludge dewater- ing, and a second pH adjustment. The sludge, containing free and complexed cyanides, may be landfilled or incinerated; hydrolysis also may be practical. However, there is little experience in applying incineration and hydrolysis to these sludges. The alkali extraction process should be applicable to soils contaminated with PAH compounds as well (Hoogendoorn, 1984). Current applica- tions are limited to clean sands; difficulty in applying extraction techniques" to loamy soils include difficulty in separating clay/silt suspensions and strong adsorption of contaminants and clay particles. The excavation and extraction of contaminated soils is economical in the Netherlands because of the high cost and intensive utilization of land and the high demand for clean fill. In the United States, this alternative may not be the most cost-effective one. The in-situ extraction of organics by alkali solutions has been demonstrated for industrial sludges (Kosson et al., 1986). This technique should be more economical than excavation and extraction, may be applicable to organic-contaminated soils at gasworks, and may be more cost- effective than excavation and extraction. However, in-situ alkali extraction should not be used when cyanide contamination is present at a site because strong alkalies can dissociate complex iron cyanides into free cyanide com- pounds. 252
------- LCNTAMlNAItU SOU AQUEOUS ALKAII * ACIDS SOIL P«l TRtATMfcNT » SOU »- uiATf U 1 / EXTRACTION < SEPARATION COARSE SANO i ? 0£ WATERING I COARSE SANO ** NEUTRAL rSAT ION **^ AOUEOUS RE-OEPCS-TO. < ' " ALKALI — ••**• OF CLEANED 1 SOIL : A SEPARATION FINE SANO i PRECiPi RECYCLING j WASHING AND OEWATERING "FINE SANO -t ^ f < i TA T ftfMLJ IA I nJN » OCWATFR.NT .SLUOGt SUJOGE 0tWATERING ^ICYANOE CONTAW«MG) SEWERAGE **TER * ' SYSTEM Source: Hoogendoorn, 1984. NOTE: The arrows indicate the level of original coal tar injection. Figure 67. Treatment of soil by extraction with an aqueous solution. 253
------- \ Thermal treatment methods (high-temperature evaporation and incineration) also are applicable to soils contaminated with cyanides and PAH's (Hoogendoorn, 1984). Unlike the alkali extraction process, both sandy and clayey soils are amenable to thermal treatment methods. Evaporation at 850 °C has been used to clean cyanide and PAH-contaminated soil excavated from a gasworks site at Tilberg in the Netherlands. However, these techniques require excavation of the soil and are more expensive (after excavation) than is alkali extraction (Hoogendoorn, 1984). Thus, they may not be cost-effec- tive even though they are technically effective. 2.3.3.4.3 Fixation—A novel, patented process for fixating wastes has been applied to gasworks wastes at Dortmund in the Federal Republic of Germany (U.S. Patent 4,456,400). Remedial investigations at the Dortmund site revealed extensive contamination. Liquid coal tar was clearly visible to a depth of 10 meters along with volatile hydrocarbons and sulfur compounds. Large quantities of spent iron oxide (containing sulfur and complexed cya- nides) from gas purification were also present. Remediation at this site involved excavating and treating the contami- nated soil, contaminated water, and waste by mixing it (onsite) with lignite fly ash using the patented process (Heide and Werner, 1984). The treated material was finally disposed in a specially designed plastic-lined pit loca- ted on the site. This site cleanup was the first application of the technol- ogy on such a large scale. This cleanup approach is expected to result in considerable cost savings over an alternative plan involving removal of the contaminated material to an offsite licensed disposal facility. The treatment/solidification process relies on the pozzolonic properties of the brown lignite fly ash. The ash used at this site was obtained from local power plants burning brown lignite coal. The contaminated soil, tars, and water are mixed with the ash in a three-stage reactor along with addition- al water. The exothermic reaction must be controlled carefully to maintain a continuous flow through the mixers. The product exiting the final mixing stage is a freely flowable slurry and is conveyed directly to the lined pit. Within approximately 30 minutes, the slurry hardens to a solid material that is claimed to be virtually impermeable to water «10~8 cm/sec). Data from numerous tests indicate that metals, sulfates, cyanides, and organics are f j . 254 i i
------- bound tightly in the treated material and are not leached even under rigorous conditions. Solid wastes, fluid suspensions, and sludges can all be treated by this process, being combined with the fly ash in amounts up to 50 percent by weight. Between 20 and 40 percent water is required in the process. From the standpoint of gasworks waste, the process is attractive because it can fix organic contaminants, cyanides, and sulfates. The German governmental authorities granted approval for the site cleanup plan after 2 years of reviewing the data to support the proposed process and considering other alternatives. Protection of groundwater was the major con- cern. After the remediation is completed, the site will be used again for heavy industry. The pit containing the solidified waste will be monitored to ensure that there is no leaching of contaminants. One limiting factor in the process is the availability of sufficient quantities of the lignite fly ash, which must be trucked in from local power plants. Brown coal ash is different from the ash of U.S. bituminous or anthracite coals because of its higher content of alkali metals (e.g., Na, K) and alkaline Earth elements (e.g., Ca, Mg). Brown coal ash contains about 10 percent CaO; it also contains calcium ferrite and calcium sulfate (Heide and Werner, 1984). It is this high concentration of calcium that is responsible for its pozzolanic properties. The ash of Western coals also tends to have higher calcium contents; however, the availability of fly ash from these coals is limited. It is possible that other fixation agents could be identified with similar properties or could be made up (e.g., by combining conventional coal fly ash and lime). The effectiveness of the fixation process may be evaluated by leaching tests such as EPA's EP or TCLP in soils. It may be the method of choice for remediating contaminated soil at gas sites. 2.3.3.5 Contaminated Groundwater-- The most significant groundwater contaminants at gasworks sites are light aromatics (e.g., benzene, toluene, xylene, ethylbenzene, naphthalene, acenaph- thene indene). Incidents of significant offsite migration of gasworks contam- inants in groundwater (e.g., Ames, Iowa; Dover, Delaware) have involved the lighter components of gasworks tars and oils that are easy to detect at ppb levels by the water's tdste and odor. The concentrations of the heavier PAH compounds (three or more aromatic rings) in groundwater are generally lower, 255
------- being controlled principally by their aqueous solubilities. PAH concentra- tions tend to drop off rapidly beyond the coal tar source; the persistence of these heavier compounds in groundwater beyond the immediate site area has not been documented. 2.3.3.5.1 Source control--The most important step in the remediation of contaminated groundwater is destruction or removal of the source of contamina- tion. Until this is successfully accomplished, the success of groundwater cleanup will be limited by continuing contaminant release at the source. It is especially important to identify and remove any lighter organics (i.e., oils) present at a gasworks site because their higher solubilities and usual occurrence above the water table give these organics a high potential to con- taminate groundwater. The heavier tars tend to cause localized groundwater contamination that is localized around the area of tar contamination. How- ever, it is important to clean up free tars or to ensure that they will be effectively contained onsite; free tars can migrate significant distances from the site under certain subsurface conditions (see Section 2.1). Coal tars, produced in processes that involve coal pyrolysis, have more potential to contaminate groundwater than do water-gas or oil-gas tars because they contain significant quantities of more soluble tar acids (e.g., phenols, cresols, and xylenols. Inorganic contaminants that can contaminate gasworks sites include sul- fates (which can acidify groundwater) and trace elements (e.g., arsenic) asso- ciated with gas manufacture. The source of these contaminants includes spent oxide wastes and other solid waste from gas manufacture. Control of these contaminant sources may be accomplished by removal or treatment; in many cases, pH adjustment with limestone may be adequate treatment. Neutralization reduces acidity, raises pH, and thereby controls trace metal release. The potential for groundwater contamination by cyanides from solid wastes at gas- works sites also must be considered; however, no cases of significant contami- nation of groundwater by cyanides was found in this study. At the Birmingham, Alabama, site, leachate from untreated spent oxide wastes had free cyanide levels well below the level that can be effectively degraded by soil microbes (200 ppm), in spite of a'pH
------- 2.3.3.5.2 Selection of groundwater treatment a1ternat1ves--In devising remedial actions for contaminated groundwater, one must consider the follow- ing: • Containment control to prevent the further spread of contaminants and to collect groundwater for treatment • Treatment to destroy or remove contaminants in the groundwater. Both of these factors must be addressed when devising remedial actions for groundwater contaminants because the long times required to treat contaminated groundwater necessitate the containment activities, and it is often necessary to collect the groundwater prior to treatment. Groundwater control measures for contaminant containment include physical barriers and hydrologic barriers. Selection of appropriate technologies depends on the hydrogeologic characteristics of the site and the extent of contamination. For instance, physical barriers such as slurry walls, grout curtains, and sheetpile cutoff walls and hydrologic barriers such as intercep- tor trenches or subsurface drains are appropriate for sites where contamina- tion is confined to the near surface (25 to 50 feet deep) and underlain by a low-permeability layer into which the barrier may be keyed. Examples of the use of physical barriers (slurry walls) during gasworks site remediation may be found in the case studies for Stroudsburg, Pennsylvania, and Plattsburgh, New York, in Chapter 3. When contamination extends to greater depths, or where there is no natural barrier to vertical (downward) migration of the contaminant plume, hydrologic barriers using pumping wells may be the only appropriate control strategy. A hydrologic barrier using pumping wells was employed to control contamination from the Ames, Iowa, gasworks (see Case Studies, Chapter 3). Groundwater collection strategies include subsurface drains and intercep- tor trenches, which are appropriate for shallow contamination, and pumping wells, which may be used for shallow or deep contaminated groundwater. Sub- surface drains were used at Plattsburgh, New York, to collect incoming ground- water to prevent breaching of the slurry wall. The drain system also served to collect contaminated groundwater leaving the site (see Chapter 3). Pumping wells were used to collect free coal tar at Stroudsburg, Pennsylvania, and 257
------- contaminated groundwater at Ames, Iowa, and may be employed to control and collect contaminated groundwater at St. Louis Park, Minnesota (see Chapter 3). For further Information on the selection, evaluation, and design of groundwater control strategies, see U.S. EPA (1982), Ehrenfleld and Bass (1983), U.S. EPA (1984d), Boutwell et al. (1985), Schafer (1984), Xanthakos (1979), and D'Appolonia (1980). Treatment alternatives for groundwater contaminated with aromatlcs from byproduct tars and oils include physical methods (e.g., carbon adsorption, reverse osmosis), chemical methods (e.g., wet air oxidation, ozonation), and biological methods (Ehrenfield and Bass, 1983). At St. Louis Park (see Chapter 3), the groundwater remediation plan includes the use of granular- activated carbon to clean up contaminated groundwater. At Ames, Iowa, recovered contaminated groundwater was used, without treatment as boiler make- up water at a nearby power plant. Microbes capable of degrading PAH compounds were discovered in the contaminated Ames groundwater (see Case Study, Chapter 3). This suggests that, where groundwater is contaminated with organic compounds from gas plant wastes, indigenous microbes capable of degrading these organics may have evolved. In these cases, in-situ remediation may be possible by containing the groundwater and allowing natural degradation to take its course, with or without enhancement through the addition of oxygen (or air) and nutrients. For more information on biological treatment methods for contaminated groundwater, see Parkin and Calabria (1985). 2.3.4 Conclusions The following conclusions can be made concerning the investigation and remediation of town gas sites. • Site investigation techniques used at abandoned town gas plants do not differ significantly from those used at other uncontrolled hazardous waste sites. Because of the age of the sites, collection of historical Informa- tion from company records, insurance maps, interviews with plant personnel, aerial photos, etc., is an important first step in site investigations at abandoned town gas plants. • Surface geophysical techniques can be used to Identify buried struc- tures, pipes, and subsurface zones of coal-tar contamination at abandoned town gas plants, and they can help guide further site Investigation activities. 258
------- It is important to identify buried structures because these can contain tars, oils, emulsions, and other contaminants. Care should be exercised to avoid damage to these structures when using invasive site investigation techniques or when conducting remedial actions. If care is not taken, these substances may be released. The probable presence of multiple-density contaminants (i.e., tars and oils) should be considered when planning site investigation activities and when evaluating remedial action alternatives. The long-term stability (i.e., no release of hazardous substances over a period of years) of some sites may make no-action a viable alternative at some sites. Free tars and oils recovered at a site often may be sold for bene- ficial use as fuel or chemical feedstocks. Land treatment has been proven effective In treating soil contamina- ted with byproduct tars and oils. Other treatments used for remov- ing or destroying heavy organics in soils also may be applicable. Spent oxide wastes and soils contaminated with complex cyanide compounds have been treated successfully by immobilizing with lime, or with a combination of lime and pozzolonic material, and evapora- tion at elevated temperatures. The presence of indigenous microbes capable of degrading aromatic compounds in the groundwater at Ames, Iowa, suggests that in-situ biological treatment may be feasible for groundwater contaminated with compounds from byproduct tars and oils. 259
------- 3.0 SITE INVESTIGATIONS OF SPECIFIC TOWN GAS SITES 3.1 INTRODUCTION This chapter describes the specific town gas sites reviewed by RTI. It is divided into two sections: Section 3.2 describes the sites visited by RTI personnel, and Section 3.3 discusses case studies of town gas sites that have been described in recent literature. This chapter is designed as a overview of existing town gas sites, types of contaminants, and remedial actions. In its review, RTI also collected some historical data from pre-1960 sources on specific sites that sometimes conflicted with the site information reported by other investigators. These contradictions are also examined in this chapter. 3.2 SITE VISITS PERFORMED BY RTI 3.2.1 Introduction Mr. Scott Harkins of RTI visited six gas sites and one iron oxide dis- posal site to permit RTI personnel to collect data and site assessments on specific sites during the course of the project. Site assessments were avail- able for only two of these gas sites (Lowell, Massachusetts, and Spencer, Massachusetts) and the spent oxide disposal site (near Attleboro, Massachu- setts). One site was chosen because the authors were familiar with it, and because many of the structures were still present on the site (Richmond, Vir- ginia). One other site (Taunton, Massachusetts) was recommended by the Massa- chusetts Department of Environment Quality Engineering (DEQE), and the other two were selected because they were within traveling distance of the other sites examined (Pawtucket, Rhode Island, and Worchester, Massachusetts). All of these sites and the information obtained during the site visits are described in the next section. 261
------- 3.2.2 Colonial Gas Company, Lowell, Massachusetts This Colonial Gas Company 1n Lowell, Massachusetts, was visited on March 3, 1986. The site examination consisted of reviewing the Phase 1 site assessment, visually examining the plant site and surrounding area (without entering the site), and collecting an early site map of the plant. This 17-acre site produced coal, water, oil, and LP gas for the town of Lowell, Massachusetts. The plant began as a coal-gas plant in 1849, added carbureted water gas during the 1870's, converted to oil gas between 1950 and 1951, and was placed on standby in 1951. It operated intermittently between 1951 and 1975 to supplement natural gas supplies. The site is currently used as an operations center and storage and gas distribution center by the Colonial Gas Company (formerly the Lowell Gas Light Company). The site is approximately 300 feet from the Pawtuckett Canal, which removes water from the Merrimack River, flows through the town of Lowell, and then returns to the river. An 1876 map of Lowell (available at a local national park gift shop) clearly shows the plant layout, with five large buildings and four masonry gas holders. Two buildings on this map currently remain onsite. A vacant area is seen next to the plant and is now part of the plant site. A Phase 1 site investigation (problem definition and site history) of the site was completed in December 1985 by M. Anthony Lally Associates, and a Phase 2 site investigation (problem evaluation and field investigation) is currently planned. These investigations were in response to observed volatile contamination of soil and groundwater during an investigation of PCB contami- nation on the property adjacent to the site. VOC's were detected at 65.1 mg/L in groundwater flowing from the gas site. Soil samples were taken and organic vapor concentrations measured from shallow depths (0 to 3 feet) around the plant. Organic vapor concentrations from the probe hole varied between 0 and 96 ppm, and soil concentrations were between 0 and 37 ppm. Analyzed soil samples showed contamination by benzene (0.013 mg/g), toluene (0.004 mg/g), ethyl benzene (0.030 mg/g), xylenes (0.23 mg/g), and assorted PAH compounds (1.09 mg/g). RTI's examination of the site area found two small sources of oil flowing into the canal from the canal wall nearest the gas site. The water in the canal was lowered for routine mainte- nance during the visit. The canal itself, and several areas around the plant, 262
------- had fairly strong gaseous odors, probably from gas plant wastes. Diagrams of the plant site contained in the Phase 1 site assessment indicate a "deep well" was present. This well was possibly used for waste condensate disposal because any liquid wastes dumped into the canal would flow through the center of town and pass through water-powered factories. A literature review by RTI revealed that several articles were written by engineers working at the Lowell plant. One article on oxide purification of gas stated, "Because of the possibility of fires starting through the heat generated by revivification, it is necessary to hold the spent material at the plant until this danger is past. As soon as city authorities learn of this menace ths material is prohibited at public dumps. Continuous storage on gasworks land eventually becomes impossible. The material makes excellent filling for roads or private property when properly handled. It should be covered with ashes or dirt immediately to prevent the access of air and conse- quent combustion. ...The plant is indeed fortunate it has a place to store the spent oxide and doubly so if a transportation company will agree to remove it without charge because of its value as a filling material" (Downing, Super- intendent of Manufacturing, Lowell Gas Light Company, 1932). Evidence of tor and oil contamination of the site was also located in an article on gas plant wastes. "That large quantities of gas house waste can enter the ground is strikingly shown by investigations made at the Lowell, Massachusetts, gas works in 1905 and 1906 by A.T. Stafford and W.H. Clark, who estimated that there existed within the ground and within an area of a few acres 1,600,000 gallons of tarry and oily wastes. Some of these consisted of accumulations in old drains and porous gravel, which when tapped by excava- tions flowed out in springs. Much waste was regularly finding its way into sewers, and from the sewers it entered cellars along the lines of sewers at even remote distances from the works" (Hansen, 1916). RTI has yet to locate the articles Hansen referred to, but if accurate, they indicate possible wide- spread contamination from the facility. 3.2.3 Massachusetts Electric Company, Spencer, Massachusetts The Massachusetts Electric Company in Spencer, Massachusetts, was exam- ined on March 4, 1986. The site examination consisted of viewing the fenced portion of the site through the fence, making an examination of the perimeter 263
------- of the site, and examining the site assessment prepared by Perkins Jordan in January 1984. The site came to the attention of the Massachusetts DEQE when a truckload of soil (removed so that a drainage culvert could be installed) was delivered to a landfill during a routine inspection of the landfill. The inspector recognized the materials as being gas production wastes and ordered that they be returned to the site. A subsequent site investigation by Perkins Jordan used nine test pits, seven borings, and two test wells. Tins also was the site of a very small carbureted water-gas plant. (Brown's Directory and the 1917 report of the Massachusetts Board of Gas and Light Commissioners show it to be a carbureted water-gas plant, but the site assessment identifies it as coal-gas plant.) It was constructed between 1885 and 1887 and operated into the 1950's. The site is approximately 0.4 acres and adjacent to a small stream, the "Muzzy Meadow Brook." It is currently fenced off and is the site of a power substation. Wastes typical of carbu- reted water-gas plants (coal, coke, ash, slag and tars) were identified at the site. About 15 to 20 feet of soil rests on top of the bedrock at the site. The depths of the test well and pits were limited by the bedrock under the site. Soil samples were found to be contaminated with toluene, benzene, ethyl- M benzene, PAH compounds, and xylenes. Groundwater samples contained low levels of PAH and volatile compounds. Table 49 shows the measured concentrations of volatile and semi volatile compounds in soil and w, samples from test pits and brook samples. Table 50 shows the same analy -oil samples from borings, and Table 51 shows concentrations from mo1
------- to Ul ^ vt § o o cc to 1 t- 0. in Ul 1- 3 O cc n in T t- CO in Ul • a 5 >• o < £ Ul -1 CO ^ t- E a a ^«» •^ 1 n _• & G «l t. 3 A | ^ a je I VI • 1 " — - "o in E J< l. O * t. c m > o •o E -x f C i. O*J W H ma 3 r^S 1 N »- ^ CM WO ;J i 'I** 0. 1 1-10 ^ OM< 1 H<* 1 IO OIO) a. i ON. CL 1 IO a. i -•n 1 CM CL 1 HV) •1 •fl e 3 compo _ u "i u 5<<< Illlllllllll II z 2 z 111111111111 ii <<<< ititiiiiiiii ii ZZZZ Illlllllllll ll •< < < < SI IVSSSSSSI I IV CM S,r*«- l U> n ZZZZZZZZZZZZ ZZ til S • • s s s •<<•<< IISOISISSIII I0> ZZZZ 1 1 -r r- a> 1 CO 10 1 1 1 i CM Q S CM CM ^ 9 s s s s s s SSi*. 222222222222 22 <*) Qk * O Q Q> /N /\ 2522 5 °. S S 8- ". S S "? E ! ', !» •» Q f. M Ol ID . <<<< IIIIIIIIIIIS II ZZZZ liiiiliiiii»i ii o s ZZZZ lOCoSoivCMOCMODI |xa*L C O •• • X X — ^ • • I. C — • 4> U X O • • C • O • • • • X -C 9 JE i. e — • c 3 i. e e • j: — • o. c o • — — • — x • • c*> x*» — x o N *>x.c^o» x t. •,£jiax*-»**% • C MX4>^^«-s*A>*X — O>4>— 4>O4>W — •••«— ^JZJCMCCC^ •— ••3O.fl* — c c M » oo.a^^««»c»j:x JCJB EX — — C * — E»»CCI-33«UO.»1IQ.I*> «^ OOC'tX •OO»»i — — JEX« — — — O > t- m ui x c/)<u.u.o-a.ZQCD o K • CO C 01 0 •J e u ^ ^> o *> • -) • • C L. •1 99 ^9 J) 'jt — » *» e u • *> — • e • a • a. • TJ — 1> i' o o J • u z z t- vi § " " " n o < i a. U) Z 1 K V> 265
------- TABLE 50. VOLATILE AND SEMIVOLATILE ORGANIC ASSAY RESULTS FROM BORING SAMPLES Chemical compounds Volatile organics Toluene Benzene Ethylbenzene Xylenes Semi volatile organics Acenapthylene Acenaphthene Benzo(k)fluorene Benzo(a)pyrene Chrysene Fluoranthene Fluorene Phenanthrene Pyrene Naphthalene Diethylphthalate Bis{2 ethylhexyl)phthalate Di-n-butylphthalate Total polynuclear aromatic hydrocarbons (PAH's) B-2 S-42 & 43 5.80 75.0 41.0 53.0 NA NA NA NA NA NA NA NA NA NA NA NA NA NA Soil samples B-2 S-41 NA NA NA NA 80.0 46.0 11.0 100 31.0 130 170 370 100 670 -- -- -- 1,708 (mg/kg) ppm B-3C S-48 NA NA NA NA 4.60 9.30 — 1.30 -- 6.70 8.40 21.0 8.0 39.0 — -- -- 98.3 B-4 S-52 NA NA NA NA -- -- 0.520 0.720 2.00 — 1.60 1.40 -- -- -- -- 6.24 SOURCE: Perkins Jordan, 1984. NA = Not analyzed. -- = Not detected. B = Boring identification. S = Sample number. 266
------- TABLE 51. VOLATILE AND SEMIVOLATILE ASSAY RESULTS FROM MONITORING WELLS Groundwater samples (mg/L) ppm (taken on 11-30-83) Chemical compounds MW-1 MW-2 Volatile organics Toluene 0.0095 0.120 Benzene 0.071 0.410 Ethylbenzene 0.015 0.480 Xylenes 0.068 0.610 Semi volatile organics Acenaphthylene ND 0.041 Acenaphthene ND 0.032 Anthracene ND 0.004 Fluorene ND 0.030 Naphthalene ND 1.000 Total polynuclear aromatic hydrocarbons (PAH's) ND 1.147 SOURCE: Perkins Jordan, 1984. MW = Monitoring well. ND = Not detected. 267
------- A complete investigation (documented in transcripts in "The Affairs of the Richmond Gas Works," [no author] 1896) followed a major scandal at the gas plant shortly after the conversion, thereby marking that year. The scandal involved several plant supervisors who were dumping ash-coke mixtures from the water-gas generators before most of the coke was converted to gas and ash. This allowed the ash to be hauled from the plant and the coke recovered and sold for the profit of those who stole it. Also during this period, the plant operated for 9 months with no down-run on the carbureted water-gas apparatus (the down-run valve had burned out). This caused poor heating of the appara- tus and resulted in the inadequate cracking of carburetion oils. The creek behind the plant was reportedly full of oil that overflowed from the relief holder with condensate. The report states that the fences were rotted, the roofs decayed, the coal benches were clogged and had to be rebuilt, the water-gas plant needed to be relined, new castings and valves were needed, the purifying house oxide boxes were rusted and leaking, the condensers were broken, employees were A mismanaged, and coke was constantly stolen. The report indicates that the plant sold coke, tar, sulfate (probably ammonium sulfate), lime, and junk (the type of junk was not defined). When the plant switched to carbureted water gas, they also switched from lime purification to the use of iron oxide (the new purifier house was erected in 1894). They had previously used 9,000 to 10,000 bushels of lime per month (415 to 460 ft^). This use dropped to 80 bushels per month after converting to iron oxide purification. This gas plant, also referred to as the lower gasworks, is shown on maps in the library. One map from 1888 clearly shows another gas plant along the river, closer to the center of the city. An 1876 map of the gasworks (Figure 68) clearly shows the plant layout and structure. The round object below the coal shed is labeled as a retort by the mapmaker, but it is actually a gas holder for the plant. The plant continued to produce water gas until the early 1950's, when the plant was converted to LP gas for peak loads and standby operation. Most of the buildings present in 1950 remain on the site, i.e., the gas house, com- pressor building, purifier buildings, coal shed, and gas holder. The purifier^ building has been converted into a welding shop and classrooms for the 268
------- Figure 68. Fulton Gas Works (1876). 269
------- current, municipally owned gas company. The other buildings are not used and probably would have been removed many years ago if the plant were not city owned. (Cities do not collect property taxes from their buildings.) All of the buildings, except the purifier building and the gas holder, were to be removed during the spring of 1986. The site is adjacent to a concrete culvert (formerly a creek) that flows into the James River about 600 feet from the site. The area between the gas plant and the creek shows substantial signs of being a dump area for the plant, with contaminated woodchips, ash, coke, firebricks, and tar present. No site or environmental assessment of the plant is currently planned prior to removing site structures. The entire site was flooded with about 6 feet of water from the James River during the fall of 1985. 3.2.5 Hendon Road Spent Iron Oxide Disposal Site, near Attleboro, Massachusetts The Mendon Road Spent Iron Oxide Disposal Site near Attleboro, Massachu- setts, was visited on March 3, 1986. The site was evaluated by visiting the site during site remediation and by examining two reports of the site, a geo- " hydrologic study by Clean Harbors, Inc. (May 1985) and a hazardous waste evaluation by Hydrosample (November 1984). This was not a gas site, but a site where some spent oxide waste was disposed. The site was originally a gravel pit, but it later became a dump and was recently filled and houses were constructed on the site. When the land was purchased, the buyers sent a sam- ple of the waste to the State health department to approve construction. The. perk tests revealed additional waste, and this information was sent to the Massachusetts DEQE. After two subsequent site investigations, removal of the spent oxide wastes began with funding from the State Superfund. The waste is spent iron oxide (mixed with woodchips) from coal-gas manu- facture. The waste was apparently used as fill at the site, with other fill material above and below the waste "seam." The waste material consisted of contaminated woodchips with high concentrations of PAH compounds, iron cyan- ides (total CN 7,500 ppm, soluble CN 0.7 ppm), and low pH (1.7 to 3.8). It passes the EP toxicity but has a high total metal content. The waste was a seam of material with a maximum thickness of about 3 feet, covered by between 1 and 4 feet of clean topsoil. The site remediation was to remove all cyanide™ 270
------- to a concentration of 2 ppm in the soil. Clean topsoil was removed and put aside; the waste and an additional foot of soil below the waste were removed, stabilized with calcium sulfate, and transported to a hazardous waste landfill in Alabama. The resulting holes were filled with clean fill dirt. An early estimate of the necessary remediation was removal of 2,500 ft^ of contamina- tion at a cost of SI.6 million. The solubility of the iron cyanide compounds in water was evidently very small. The cyanide wastes were removed from the equipment used in the remediation by physical means only. The equipment was hosed off with water, and the water was drained, into a holding tank (approximately 200 gallons). The solids were allowed to settle to the bottom of the tank, and the clear water was removed from the top of the tank. This water was then run through a sand filter, and the resulting water was discharged without further treatment. The cyanides were essentially all removed by settling and filtration with sand. The material that settled in the tank, and the tank itself, were to be discarded in the Alabama landfill at the end of the remediation. A similar disposal of spent oxide wastes is on the ground surface just across the Rhode Island border. 3.2.6 Pawtucket, Rhode Island The Pawtucket, Rhode Island, site was examined on March 5, 1986. It was evaluated by only a visual examination of the plant site and by data from Brown's Directory. This is a fairly large gas site that produced both coal and water gas during operation and had an electrical power plant as part of the site. The site occupies 20 to 40 acres between a residential neighborhood and the Seekonk tidal basin, just south of Pawtucket, about 3.5 miles from the Attleboro road site in Massachusetts. Part of the site is currently used as an electrical substation and for the distribution of natural gas. There were several areas of the site that contained spent oxide wastes similar to that at Mendon Road (e.g., woodchips from spent oxide, and blue areas of soil from ferrocyanides). A substantial amount of waste from the gas production and power generation was visible on and around the site, evidently as fill. 271
------- 3.2.7 Taunton, Massachusetts ™ The Taunton, Massachusetts, site was visited on March 3, 1986, with a representative of the Massachusetts DEQE. The site was examined visually. It is a mid-sized gas plant that primarily produced coal gas but later produced water gas (Brown's Directory, as reported by Radian Corp). Constructed around 1890, the plant added water gas around 1920. The site, approximately 15 acres, is in an industrial area south of Taunton, adjacent to the Taunton River. All of the structures were removed in the early 1960's, but the site has never been properly decommissioned. Gas-holding tanks were cut off at ground level and filled with soil from the site. It is very unlikely that any underground structures were removed. The plant was located at the northern end of the site, and the southern part of the site was evidently used as a waste disposal area. The State OEQE was called by the gas company in the early 1970's to stop waste materials from eroding into the river. Eventually, the southern half of the site was capped with a layer of clay soil and top- soil. This southern half has a small stream that crosses it and currently flows through a lined culvert. The northern half has remained uncovered. No A waste materials have been removed from the site, and approximately 1 to 3 feet of mixed wastes are under the capped area. Heavy tars, ash, and spent oxide wastes are visible in the uncapped area. The site is currently fenced, and the local gas company operates a standby LP gas facility across the street. No additional remedial actions or in-depth site studies are currently planned for the si te. 3.2.8 Worchester, Massachusetts The Worchester, Massachusetts, site was visited on March 4, 1986. The city was chosen because it is large, happens to be close to Spencer, and is listed in Radian's 1984 compilation of U.S. gas sites. Some information and maps of the plant site were located in the Worchester public library. The Worchester Gas company, chartered in 1849, moved to a 9-acre site on Quinsigamond Avenue in 1869. It produced both coal and water gas. Currently, the site is used by the Commonwealth Gas Co. as a gas storage and distribution facility. The entire site has been capped with approximately 3 feet thick of construction refuse and fill. The site has no noticeable wastes and only a ^ 272
------- slight odor. An EPA pollution control project (Project C250.347-04), a new S7.5 million (S5.5 million Federal, SI million State) sewage treatment facility, is to be newly constructed on the gas site (as indicated by a sign on the property). 3.3 CASE STUDIES OF TOWN GAS SITES 3.3.1 Introduction The case studies in this chapter were selected to demonstrate the types, modes of occurrence, and persistence of contaminants at abandoned manufactured- gas sites, as well as applicable remedial measures for these sites. The case studies are presented to support the material discussed in the preceding chap- ters. They were collected from published literature, State and Federal agen- cies, and previous work at RTI. Differences in detail between the studies reflect different amounts of information available for specific sites. In addition to six former gasification sites, two byproduct tar utilization facilities, a creosoting plant (Pensacola, Florida), and a coal-tar processor (St. Louis Park, Minnesota) are included. These two studies offer well- documented evidence of migration and degradation of coal-tar derivatives in the subsurface that is relevant to contamination at gas plants. The case studies were compiled from the references presented at the beginning of each study. 3.3.2 Norwich, Great Britain (Wood, 1962) The Norwich, Great Britain, site is the oldest site found during this study, having groundwater contamination from tar present for over a century. It illustrates the potential persistence of gasworks tar in the subsurface environment, both in terms of the tar's appearance and its potential to con- taminate groundwater. In 1950-1951, a 36-inch bore was sunk into the chalk aquifer underlying Norwich for water-supply purposes. Although it produced water of sufficient quality for its intended use, the well's yield was inadequate. To remedy this, a horizontal adit was drilled from the bore into the chalk at a depth of 150 feet below the surface. Shortly after, the water acquired a tarry taste and thus was rendered unusable. Subsequent colorimetric analysis Indicated 273
------- that the water contained about 0.2 ppm total phenols, which appeared to be largely cresols. Thiocyanates were below the detection limit of 0.01 ppm. Inspection of the adit by descent into the well showed black tarry matter eluding from the adit roof. Samples of the tar contained a small proportion of volatile matter, which had a trace of phenols, but was mainly composed of a yellow oil with a blue fluorescence in benzene solution (suggesting the pre- sence of aromatic compounds). The larger portion of the tar sample was non- volatile, tarry in consistency and odor, and contained particles of solid carbon. The source of the tar was originally a mystery because the site was far from the Norwich gasworks. However, subsequent investigation revealed that the first gasworks plant in Norwich was constructed over this site. That plant, which operated from 1815 to 1830, produced gas from destructive distil- lation of whale oil by the Taylor process. Thus, the well constructed in 1951 was polluted by tar that had been lying in the ground for over 120 years. This case study illustrates that tar from town gas processes can persist and retain its potential for environmental damage for over a century. The amount of tar degradation that may have occurred is impossible to estimate because there is little information on the original tar composition. However, of signiTicance is that at least some of the tar acids (phenols and cresols) have persisted in spite of their high solubility, and they have contaminated groundwatcr. The absence of thiocyanates is expected because of the low sul- fur content of whale oil. The tar's appearance and odor is similar to that of coal tar, illustrating that, with the exception of the formation of sulfur and nitrogen compounds, the gas production process is more important than feed- stock composition in influencing tar formation. The "steam-volatile matter" reported by Wood (1962) probably corresponds to the naphtha or light oil frac- tion of tar, and it may be responsible for much of the observed groundwater contamination. 3.3.3 Ames, Iowa (Siudyla, 1975; Yazicigil, 1977; Yazicigil and Sendlein, 1981; Burnham et al., 1972; Burnham et al., 1973; Ogawa et al., 1981) The Ames, Iowa, case study illustrates long-term contamination of a water supply by town gas wastes from a relatively small gas plant that served about 15,000 customers. Groundwater contamination was first detected by taste and 274
------- odor problems in 1927 and has persisted into the 1980's. This case study illustrates the following: • Site discovery through odors in water caused by ppb levels of dissolved tar constituents in groundwater • Contamination of groundwater by lighter tar fractions (tar oils) that are less dense than water and more soluble than heavier tar components • Contaminant sources resulting from town gas waste disposal practices • Contamination by tar wastes from a water-gas process, notable by their lacK of tar acids (phenols, cresols, xylenols) • Migration of contaminants through cracks in soil to the water table • Influence of pumping wells on the migration of dissolved coal tar constituents in the groundwater Use of historical data in a site investigation • Degradation of PAH's by microbes naturally occurring in ground- water at Ames • Remediation through removal of contaminant sources, instal- lation of barrier wells, and controlled municipal well pumpage. 3.3.3.1 Site History-- According to Siudyla (1975), who interviewed long-time residents of Ames, town gas was produced in Ames from 1911 until 1927. The original gas plant was in operation from 1911 until 1920, and it was located in the western sec- tor of the Ames well field. In 1920, the plant was moved to its final loca- tion. Although there was a waste pit at the original plant site, 70 feet of glacial drift isolated this source from the underlying buried channel-sand aquifer. However, the drilling of a municipal well in 1968 through the pit and into the underlying aquifer resulted in some contamination of the aquifer by PAH's. Contaminant levels at the well have decreased over the years because the well has been pumped (Siudyla, 1975). Brown's Directory indicates that the Ames plant operated from 1912 until about 1932 when gas lines were completed from Boone, Iowa. There is no men- tion of the plant's 1920 move, but 1t is indicated that Iowa Railway and Light 275
------- purchased the gasworks in 1925. According to Brown's, the plant produced carbureted water gas over most of its history, with some coal gas being produced from 1916 to 1918. The directory notes that bituminous coal was used to fuel the plant after 1924. The operating data from Brown's is compiled in Table 52. No mention is made of fuel type prior to this entry, although coke, anthracite, or bituminous coals would be used in the generator, and gas oil or fuel oil would be used in the carburetor. The appearance of a disagreeable taste and odor in groundwater from two city wells first occurred in 1927. The timing of the appearance could be related to the change in plant management in 1925, which could have affected waste disposal practices, or it may just reflect the time it took the contami- nants to reach the wells from the source. In the early 1930's, three auger holes showed increasing concentrations of contaminants toward the waste pit at the second gas plant site, which was then recognized as the source of contami- nation. At that time, investigators determined that abandoning contaminated wells and drilling new wells farther from the source was the best solution. This practice was followed until 1961, when the wastes from the second pit were removed to a sanitary landfill in an attempt to mitigate the problem. It did not. By the late 1960's, five wells had been abandoned and several were restricted to limited pumping. In 1975, Siudyla interviewed a former gas plant employee and discovered that an overflow channel not visible on any city maps had once flowed from the waste pit to the Skunk River. Although now buried with fill, the channel was described as once being "odorous...containing pools of coal tar wastes" (Siudyla, 1975). Subsequent sampling showed that oils had collected in two low areas in the former channel and were floating on top of the water table at these locations. These areas were identified as the contamination sources of the city's water supply aquifer. The type of organic contamination was thus discovered, and its oily nature is consistent with the disposal of waste con- densate (and floating oils) from carbureted water-gas manufacture. 3.3.3.2 Extent of Contamination-- As previously described, the taste and odor problems in Ames' groundwater have existed since 1927. Originally attributed to phenolic compounds, analyt- ical work in the early 1970's showed a notable lack of phenolics. The 276
------- < o (/I 1 t- § 8 ce *• v> o CM u> UJ CO ^ « T) L ^ O •t | O M 3 C O 3 a a. L -X O ^ O S c o u 3 "O 0 t. 0. M U O 1. a. L t V •o *> X o — c u • *> t- II C O ^ t-t •— • n M u IS. < a. • § S § in 3 03 a. QU « CM IB f^^ X 2-O-D "O-U ^ "O • c «» • ^ ^ ^ •» ^ —- ^ ^ —— •>• 5OOOOOOOOOJ ZZZZZZZZZ-J ooc»otoissscaSi-if-> COVCOCDOlOiCnOlCnCAOk M o fl I . • < ^N u • > o • 1. > JIS4 U *^ O) — • C _ -o .- JJ — > 3 — i. 0. O • « M • C * 0 • QUO ^ CO •1 » II m 171 CM m •* CM 3 0- CO S Q IA NO X C 0 ^ c * I-l 01 X t. u • • a. jt a no u c u o U — M C — C •0 "HO o *> £. OSS 3 S O Q "O Q T S o - - U> U ID CM Q S Q U) S IS •* •* r* o> o • u •o M • » III 9 m w 5 € Q M C*l S U3 6> CM *4 (Q • O) V M CO <-i f* 33 3 *> *J *J (L CO 00 (X. ffi w 1.J Q B Q 2g §2 S to r-- in w M « L L. C • « If « C >s C >% W >\ C C L C — c — c e c o t- oh- « i- « CO S rl 11 'M o> o> o> r^ r1* r* • 0 • 0 — x c • o *> 0 3 •-—OH u <» e • • u— o> a 3 E • -0 3 M U O 41 C L — O 1- O..O — c — T3 M » • — • c a o a> o £ 4» S §o 5 5 S IA »T> *«n ia • o> *•« 23 §•» CO •» U) f-l ^H Q S IO S 0> 1C oT -i 3 3 .*> ^ m o s s Ul ^ m ^ X • c « • I«/-N H • L C U • C • *J C *J it J CM » CM CM O> Ol IH IH ^^ ] C *> c 0 u ^s 277
------- T? 3 C '^ C 0 u v_s CM 10 UJ % K n K T) | X *> o M i c ; 0 c o 4 3 a o a. L X m £ CO 3 • — C • oc -o — — -o a •- «> O £ tl _l O • A • i. 3 c injc— n t u> a*) o 01 a c o • c O CM — — O U O O 1-1 o» w .0 — x«i— u — X OIMO) * 01 * C M M II j> x c — aoc-*>o> — u o -< — i. *> o u -O3O*>r- HOT 4J3S • ~1**- IA CC*^Q1A 1>CO L •» - 0> • us O r~ 3 — » m m i c 01 T M am CT W O «l X •< 3 1-1 D CO Si £ CO Q CO «-« CM •-" §n CM ^ rn CM •-> m CO CM 01 33 3 CO CD CD CS) U) U3 U) CM CM U) U3 LO c L. *> • h- HI- • • * *•» J: »-» o > CO CD S CM CM 0) 01 Ol 01 X a • t c — o — o * CD CC E * 0 S t. O ••- HI • 0 • c • X 0 *> O— X (.1101 3 .- .- a > _j M ^ ail n o in r- en «* t~i CO CM sT 3 O u> CM U> o> CO •o tt M ** n 0 0) u 3 U U ^ « if N f-4 M X II I. 01 o 41 t, • u «xn • a *> L. II •». •O — L « C • 4» X t C « jJ O • * i. — on O 3 U *J CD L a.— ~^ • « -o 3 u u c 4» • m Ul II k U u • n O 0~OO. 4» 10 -i x oeo 278
------- predominant contributor to taste and odor was found in the neutral component cf the groundwater organics, which included several aromatic compounds. Table 53 presents concentrations of these organics. No basic organic compounds were found. The lack of tar acids (i.e., phenolics) is consistent with the water- gas process that operated at the site; water gas does not produce significant tar acids. Recent analysis of Ames1 groundwater for heavier PAH compounds [e.g., phenanthrene, benzo(a)pyrene] has shown these compounds to be present, but at very low concentrations (Tom Neumann, Ames Municipal Water Department, per- sonal communication, 1986). The concentrations of heavier PAH's in water from the dewatering wells were slightly higher than those in water use wells, but no wells showed total PAH levels above 100 + 80 ng./L, and all levels were below levels of concern and World Health Organization (WHO) water quality criteria. The low level of the heavier PAH's is consistent with their low solubility in waters. The source of contaminants in the Ames1 aquifers was the waste pit and the overflow channels. There is no information on the type and disposition of contaminants in the original disposal pit prior to its removal in 1961. The overflow channel did receive some pit wastes, but these may largely represent the lighter floating components of the tar and wastewater disposed in the pit. Soil auger borings and test pits were used to investigate the overflow chan- nel. The borings showed four levels of contamination: (1) odor, oil, and tar; (2) odor and oil; (3) odor alone; and (4) no odor. Determination of the vertical extent of contamination from the soil bor- ings was difficult because of contamination of the auger as it passed through the upper levels of oily and tarry materials. Test pits, dug to 10 feet, showed that the contaminants had moved downward through vertical cracks in the alluvial materials and that oil was floating on the groundwater table (Yazicigil and Sendlein, 1981). Subsequent excavation of the contaminated material indicated that heavier contaminants (heavy oil and tar) had moved below the water table and that pockets of tar in an almost solid state existed in the excavated material. Excavation depths were limited to 15 feet because of the high water table (at 8 feet). However, the lighter oil, floating on the water table, was probably largely responsible for the taste and odor 279
------- TABLE 53. NEUTRAL COMPOUNDS IN A CONTAMINATED AMES, IOWA, WELL Concentration Std. Compound (ppb) Dev. Acenaphthylene 19.3 1.4 1-Methylnaphthalene 11.0 0.6 Methylindenes 18.8 0.8 Indene 18.0 1.5 Acenaphthene 1.7 0.2 2-2-Benzothiophene 0.37 0.11 Isopropylbenzene Ethyl benzene Naphthalene 2,3-Dihydroindene 15 Alkyl-2,3-dihydroindene Alkyl benzenes Alkyl benzothiophenes Alkyl naphthalenes SOURCE: Burnham et al., 1972. 280
------- i problems in groundwater, and this was removed by the excavation of the channel area. The well field for Ames, to the north of the site, has been contaminated by tar constituents in spite of a regional hydrologic gradient to the south- east. Pumping of municipal wells appears to have locally reversed the grad- ient, causing contaminants to flow northward from the source to the municipal wellfield. Burnham et al. (1973) demonstrated that total concentrations of aromatics at a given well in 1972 were directly proportional to the demand placed on the well (total pumpage) over a period from 1935 to 1972. Drawing on this conclusion, Yazicigil and Sendlein (1982) modeled the Ames1 aquifer system and various remediation alternatives. Based on their investigation, they suggested removal of the source materials, installation of pumping wells to create a hydrologic barrier between the source and the wellfield, and con- trolled municipal well pumpage to control the problem and prevent further well contamination. Ogawa et al. (1981) studied the degradation of aromatic compounds in samples of Ames1 groundwater. They found that, at a 25 to 150 /jg/l concentra tion, acenaphthylene, acenaphthene, 2-methylnaphthalene, 2-methylindene, 3- methylcindene, and indene were almost totally degraded at ambient temperature within 3 days. Decay rates were highest for acenaphthylene and lowest for indene. Additionally, acenaphthylene was degraded even when spiked into the Ames1 well water at its solubility limit (3 mg/L). Degradation could be pre- vented by filtering the groundwater through a 0.45-/im filter. Samples of distilled water and uncontaminated Ames1 groundwater that were spiked with acenaphthylene (at 80 /jg/L) showed no degradation of this compound after 18 days. However, when similarly spiked samples when inoculated with water from a contaminated Ames' well, the acenaphthylene was degraded within 9 days. Inoculation with anaerobic and aerobic bacteria from a sewage treatment plant resulted in no degradation. These results suggest that a population of microbes capable of degrading aromatic compounds has adapted in the contaminated Ames' groundwater. Cell mass measurements and microorganism counts further support this conclusion. Correlated with the decrease in aromatic constituents, Ogawa et al. (1981) 281
------- observed an increase in both cell mass (from 2 to 20 mg/L) and microorganism count (from 102 to 104 cells/ml). Ogawa et al. (1981) also measured the degradation of heavier PAH com- pounds spiked into aged Ames' groundwater that was formerly contaminated. Acenaphthene, phenanthrene, and fluoranthene (added at a 150 /*g/L level) were degraded within 36 days, Pyrene, which had the same concentration, was 56 percent degraded in the same period. Thus, degradation rate of the PAH compounds decreases as the number of rings increases, as was also illustrated by degradation rates for the lighter PAH compounds (discussed previously). The Ogawa et al. (1981) study demonstrates that dissolved PAH compounds, at concentrations up to their solubility limit, can be degraded by microbes naturally occurring in groundwater and that these microbes do not normally occur in groundwater, but may adapt in groundwater contaminated with PAH compounds. These conclusions are important for the remediation of abandoned coal gasification sites. Degradation of compounds by microbes suggests that cleanup of groundwater contamination may be possible by somehow enhancing this degradation, either by aeration and adding nutrients to the groundwater and/or by enhancing the degradation rates of these microbes by breeding more active strains. Additionally, groundwater with no PAH-degrading microbes may be inoculated with water from groundwater systems where microbial degradation is occurring. 3.3.3.3 Site Remediation-- To date, site remediation at Ames has consisted of following the recom- mendations of Yazicigil .and Sendlein (1982), i.e., removal of the source of contamination, installation of two dewatering wells to form a hydrological barrier between the source and the wellfield, and careful management of the pumpage in the individual city wells. The sources in the overflow channel were removed in 1980-1981 by excavat- ing a 30 x 70 x 15 foot deep trench, removing contaminated material to a land- fill, and replacing it with clean fill along the length of the channel. Two dewatering wells, installed to the north of the channel to permit this excava- tion, are now pumped to create a hydrologic barrier between the overflow chan- nel and the weM field to the north. Water from these wells is used at a near- by power plant. It is too early to assess the effectiveness of the removal 282
------- action in mitigating the contamination, but increases in contaminant levels have not been observed in either the barrier wells or in the last two wells closed in the municipal wellfield (Tom Neumann, Ames Municipal Water Depart- ment, personal communication, 1986). Questions remain about whether the source of contamination has been suf- ficiently removed because the depth of the excavation of the channel was lim- ited to 15 feet due to a high water table. If the source were removed or reduced to a size that results in a slower, low-level release of contaminants, it is possible that microbial degradation may eventually reduce contaminant levels in the aquifer. Otherwise, it may be necessary to continually pump the dewatering wells and carefully manage pumping of the aquifer to control con- taminant migration. 3.3.4 STROUDSBURG, PENNSYLVANIA (Adaska and Cavalli, 1984; Berg, 1975; Campbell et al., 1979; Hem, 1970; Hult and Schoenberg, 1981; Lafornara et al., 1982; McManus, 1982; Schmidt, 1943; Unites and Houseman, 1982; Villaume, 1982; Villaume et al., 1983) The Stroudsburg, Pennsylvania, town gas site, located next to Brodhead Creek, was in operation from the mid-1800's until 1939. During plant opera- tions, the production byproducts (mainly byproduct tars) were disposed in open trenches and later in an underground injection well located onsite. After severe flooding in 1955 from Hurricane Diane, the Army Corps of Engineers modified the Brodhead Creek Channel. In 1980, the channel was deepened to prevent undercutting of the levee. At this time, black tarry globules were observed emanating from the base of the dike along the western bank of Brod- head Creek. The site was reported to the National Response Center, and the EPA initiated an investigative study. The study found that tar was present in the subsurface at the site; the tar was confined primarily to coarse clean gravels and had collected in a large depression underlain by a fine silty sand. The site was listed as a priority Superfund site and was the first one in the nation to receive em- rgency Superfund money. This case study illus- trates the following: • Site discovery through discharge into an adjacent stream • Role of capillary pressure in controlling the movement of coal tar 283
------- • Recovery of free coal tar in the subsurface by pumping through a 30-inch gravel-packed well • Increasing the efficiency of tar recovery by pumping the over- lying groundwater to create a negative pressure and make the tar upwel\ • Construction of a 648-foot bentonite-cement slurry cutoff wall on the streamside of the western le^ee to contain the contami- nation and prevent further seepage into the streambed • Possible misinterpretation of historical data, leading to erroneous conclusions about the site, the nature of the contam- ination, and site remediation (see the next section). 3.3.4.1 Site History-- In light of the information collected during RTI's historical literature review, some of the previous site historical information about the Stroudsburg plant appears to be incorrect. This section compiles the site history and processes reported in the current literature, and Section 3.3.4.4 addresses the contradictions between this section and data collected by RTI. The Stroudsburg coal gasification site is located in the borough of Stroudsburg, Pennsylvania, along the western bank of Brodhead Creek (Figure 69). The geology of the area consists of limestone bedrock overlain by a valley-fill-type deposit. The valley-fill-type deposit is made up of an underlying, well-sorted, fine, silty sand overlain by both stratified and unstratified, well-sorted, coarse glacial gravels. .Inside the western levee is a single, steep-sided, gravel-filled depression, probably a kettle feature. The median depth to groundwater previous to any remediation was 10 feet, the hydraulic gradient was 0.015 foot per foot, and the groundwater generally flowed to the southwest at the rate of about 2 feet oer day (Villaume et al., 1983) (also see Figure 70). The plant was built in the mid-1800's and was in operation until 1939. The coal gas was manufactured by heating pulverized coal in a reaction vessel to drive off the volatiles. Superheated steam was then passed over the hot coal to produce a gas-steam mixture that was blown into a large holding tank. In this tank, the steam condensed, leaving the gas at the top and a liquid containing coal tar at the bottom. The major byproducts of this procedure 284
------- Figure 69. Stroudsburg site map with top-of-contamination (dash) and ground water (dot-dash) contours (in feet) shown. The groundwater data are for June 12,1981, prior to slurry wail construction. Almost no free coal tar occurs beyond the 374-foot contour. Source: Villaume et al., 1983. Figure 70. Top-of-sand contours (in feet) for the Stroudsburg coal-tar contamination site. Source: Villaume et al., 1983. 285
------- were the coal tar left In the reaction vessel and the liquid containing coal tar in the holding tank. lafornara et al. (1982) estimate that as much as 16 million gallons of coal -tar could have been produced over the 100-year operating life of the Stroudsburg gas plant. Initially, the reaction vessel coal tar was disposed in open trenches that ran along the western edge of the site, eventually discharging into Brodhead creek, and the water and tar that collected in the holding tanks were blown down onto the ground next to the tanks (Lafornara et al., 1982). In the early 1900's, as coal-tar reprocessing technology devel- oped, the coal-tar wastes were purified onsite to remove the commercially valuable constituents. The remaining wastes were disposed in an underground injection well onsite. This method of disposal continued until the plant shut down in 1939. Brodhead Creek experienced severe flooding in 1955 as a result of Hurri- cane Diane. Between 1958 and 1960, the Army Corps of Engineers had to modify the stream channel by straightening several reaches of the stream and placing the channel within a floodway lined by riprapped levees. Within the next 20 years, the levees experienced significant downcutting, causing officials to deepen the riprap another 10 feet in 1980 to protect the levees from under- cutting. During this work, coal tar was identified in the open trenches along the western bank of Brodhead Creek. In 1981, the site was reported to the National Response Center. The EPA ordered all affected .-roperty owners to conduct a study to determine the extent of the contamination and a method of rectifying the damage. The Stroudsburg, Pennsylvania, site appears on the expanded list of 418 priority Superfund sites (which currently number 388) and was the first site in the nation to receive emergency Superfund money (Lafornara et al., 1982; McManus, 1982; Unites and Houseman, 1982; Villaume, 1982). 3.3.4.2 Extent of Contamination-- Based on the 1981 investigative studies, up to 1.8 million gallons of free coal tar is estimated to be distributed over an 8-acre area (Figure 50). The contamination extends vertically downward only to the top of the silty sand deposit. This deposit currently cannot be penetrated by the coal tar because of the extreme capillary-pressure forces that must be overcome. An 286
------- accumulation of up to 35,000 gallons of nearly pure coal tar was estimated to occur in a single stratigraphic depression located just below the old gasification plant (Figure 51). Capillary pressure (P) is defined by the equation: P = 27 cos 0 / R where 7 = the interfacial tension between the coal tar and water 0 = the contact (wetting) angle formed by the coal tar against a solid surface in the presence of water R = the radius of the water-filled pore that the coal tar is trying to enter. The displacement of water by coal tar is most difficult when the capillary pressure is high, by definition indicating a high interfacial tension and low contact angle. Once the interfacial tension and contact angle are set, the pore size of the rock determines whether the coal tar can move into the media. Using Hobson's Formula (Berg, 1975), the critical height of coal tar needed to overcome the capillary pressure is calculated to be more than 10 meters. The maximum thickness of coal tar in the contaminated zone at any location onsite does not exceed 5.5 meters. The high capillary pressure and lack of critical column height of the coal tar explains why the silty-sand deposit serves as an effective barrier to the coal tar. Hydrodynamic dispersion would be expected under onsite groundwater flow conditions. Shallow groundwater samples from throughout the site indicate the presence of dissolved contaminants. Partial analysis of the Stroudsburg coal tar is shown in Table 54. The polynuclear aromatics were generally detected at the ppb level or within the range of known aqueous solubilities of the individual chemical species involved. Table 55 shows that the principal control on the concentrations of these contaminants in the groundwater is their aqueous solubility and not their concentration in the coal tar. There is not enough data at this time to determine whether a relationship exists between solubility and distance of transport; however, there appears to be a rapid decrease in concentration just beyond the free coal-tar plume in the downgradient direction. The only contaminant detected at this point is naph- thalene at less than 10 ppb. 287
------- TABLE 54. PARTIAL ANALYSIS OF THE STROUOSBURG COAL TAR Parameter Value Units Naphthalene 3.60 I Fluoranthene 3.20 % Phenanthrene 2.30 % Anthracene 2.30 % Dimethyl naphthalenes 2.15 % Trimethyl naphthalenes 1.78 \ Methyl phenanthrenes 1.50 % Trimethyl benzene 1.30 ',- Fluorene 0.98 % Acenaphthylene 0.74 \ Acenaphthene 0.72 % Pyrene 0.56 i Benzo(a)anthracene 0.31 % Chrysene 0.31 I Benzo(a)pyrene 0.10 % Other 7.84 % Total 29.69 % Acidity pH Free carbon (Carbon I) Ash Total carbon Total hydrogen Total nitrogen Sulfur Chloride Ammonia Cyanide Iron Copper Manganese Zinc Nickel Cadmium Lead Arsenic Aluminum Vanadium Barium 0.62 4.6 <0.01 0.00 90.77 8.12 0.17 0.65 50.0 0.26 0.18 50.3 2.48 2.11 0.13 0.19 0.01 0.5 12.7 22.4 1.6 0.5 mg KOH standard 0 o Q ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm SOURCE: Villaume et al., 1983, 238
------- ce UJ t- | ^ o cc , O — _l 3 J3 Ol Cf_3 £ K "o ^ A — *J 3 JC $ .- o » 2 * A "^ t. O u. ^> c A C 'i A C O lACDincoinscorocn'-inm&S) SSScM'iSSsSSSSSS foQcassssstatassistB •*/ v HliIiiiS!iSv?c5 m CM CM m A A -Q CO -O A AHACDIACMCM <*) t C O. *J U C 0 x-v 0 A 0 A — -O A O U C I. X U L C X 0 O I. 1 o 00 00 *> 1.30(0 0— c cc c x — a - — C X 0 000 A O. <•- ^ CM A0XX CUX O O O — U — *> J C C JC *> H0NMX" ^>AXXCU*>C CCCCO)-^ 0«>AAl.l.ALCCQnaiCOOC CXCCOXCO0I XI IN0 I O.003^03CCMI.V^rc-D 0AOO— CX— X"X --0C A • * • • CD •-« S m * to 01 CM O> •-• 6» (9 ^ 1 1 1 u u u (9 ro O> CO CO U) f". U) « — i CO tO — . M « CO CM XTT T JU fL, co r- co u u o c o «« J 0 u c A 0 t- N <•- c 000 • f C _^ k t «U — 0 0 — — N 3 X *» C — X A 0 O *> — CO t- UJ o > ro CD a> «x - **• A «> 0 1 3 A «— — .— ^ .. UJ \J cc. g w o o X a o u u 0 a M M VI A X X a A O) o o i_ X u M A O CM CO O) X i-i jQ M CO I ^ — VI 3 -0 Ol I. C A •o •P C M • A — f- *» •o r- V) C 0) • — r* «•- O U 3 3 m A — A CO X 0 O> 4> M I. f* A 3 X "O CD - *» C « VI A «— U X C — ?A O- J< — 3 O U •»> < vi «. A g - 2 Z 2 •IEEE 2O O O III. A <•• <^ t- y •- A A A C A A A SO U l-tO Q CS • A X- I M 4> Z 289
------- Of the volatile organic fractions, only benzene, ethylbenzene, and tolu- ene were found in the shallow groundwater. No acid fraction organics, most notably phenol, were found in the shallow groundwater. These materials were also detected in extremely low levels in the coal tar itself and were attri- buted by Villaume (1982) to either their original absence or to prolonged leaching by groundwater. Although the latter interpretation was supported by the Villaume, our investigation found that the plant operating at the site was a water-gas plant, which would produce tars with very low levels of tar acids (phenols, cresols, and xylenols), supporting the hypothesis that these com- oounds were not initially present in tars. Elevated levels of certain metals and traces of cyanide were detected in the shallow groundwater at the site. In sorre of the sampled wells, aluminum, iron, manganese, and cyanide were detected at levels as high as 218, 460, 25.5, and 0.30 ppm, respectively. By comparison, these contaminants were measured in the raw tar at levels of 22.4, 50.3, 2.11 and 0.184 ppm, respec- tively. Sodium also was found in the groundwater at 26.2 ppm, but it was never analyzed in the tar. Cyanide, probably as either HCN or NH4CN, is a byproduct of the gas cleanup and was typically removed from an iron salt (see Chapter 1). The source of the aluminum, on the other hand, is more problema- tical and, at such high concentrations, is probably present as a precipitated solid (Hem, 1970). The high sodium levels may be the result of sodium hydrox- ide usage at the plant. Even higher levels were-found in the aquifers around the coal-tar distillation plant studied by Hult and Schoenberg (1981), who attributed them to such a source. The toxic effects of tar seepage into Brodhead Creek were assessed using a macroinvertebrate and fish survey, tissue analysis, and in-situ toxicity testing of caged trout. These analyses revealed no apparent biological accum- ulation of the tar constituents. Also, tar contaminants were not found in the mixed stream flow as measured by gas chromatograph/mass spectrographic anal- ysis. 3.3.4.3 Site Remediation-- In 1981, the State's investigative study recommended the construction of a slurry trench cutoff wall to contain the coal tar and prevent further migra- tion into the streambed. Also recommended was the installation of a recovery 290
------- well system to collect tar wastes for removal. Because of the nature and extent of contamination, the State applied for and received funds for the | remedial work under the Superfund program. The cutoff wall was constructed of a bentonite-cement slurry. The com- pleted wall is 648-feet long, 1-foot wide, and 17-feet deep. The wall extends down through the contaminated gravel stratum and 2 feet into the silty sand layer, which serves as an effective barrier to the coal tar. The upstream end of the wall is tied into a sheet-piling gate that is part of the existing flood dike, and the downstream end is tied into an impermeable cement-benton- ite grout curtain (Adaska and Cavalli, 1984). Initially, it was estimated that 35,000 gallons of free pumpable tar had accumulated in the single stratigraphic depression below the old coal gasifi- cation plant at Stroudsburg. This is tar that has displaced virtually all of the initial pore water in the gravel. Some tar also occurs above the pure coal tar in the depression, but it is associated with free water (water not held by strong capillary pressure forces), which could be picked up during any pumping operation. To recover the full tar, a 30-inch gravel-packed well cluster was installed at the deepest point in the depression. It consists of four 6-inch wells screened only in the coal-tar layer. In the center is a single 4-inch monitoring well, which is screened over its entire length. Originally, prod- uct recovery was accomplished by pumping only the tar at a very slow rate. Using this method, approximately 100 gallons per day of nearly pure material were recovered, although this rate decreased drastically over time .as the volume of tar in the vicinity of the well was depleted. To increase the efficiency of the coal-tar recovery, the central monitoring well was modified by the installation of a packer at a depth between the static groundwater and static tar levels, thus isolating the lower part of the well. When groundwater is pumped from the uppermost layer, the resulting pressure reduction combined with the density difference between the two fluids causes the tar to upwell. If the tar is pumped at the same time as the overlying groundwater, the tar flows into the recovery well at an increased rate. Using this setup, a two-fold increase in the recovery rate 291
------- was achieved. To date, approximately 8,000 gallons of product with less than 1 percent water content has been recovered. The initial estimate of total free coal-tar contamination at Stroudsburg is probably too high because it was based on an assumed 30 percent porosity for the contaminated gravels and on the assumption of complete coal-tar satu- ration. The majority of this porous material is probably only poorly saturated. This is evidenced by field observations that could not be explained at the time they were made, but they are consistent with the capil- lary pressure model presented by Villaume et al. (1983). Had this been under- stood earlier, justification for the expense of building the containment wall may have been questioned. The amount of tar in the stratigraphic depression below the old gasifica- tion plant also was overestimated. The overestimation occurred because of well-screening practices that did not account for the characteristics of the various coal-tar phases and because these phases are virtually indistinguish- able in split-spoon samples. Had the estimation been closer to the actual amount present, the recovery operation may not have been undertaken or may have been scaled down considerably. Currently, the pumping operations have been stopped, with a total of 10,000 gallons of tar recovered. The site is still on the National Priority List (NPL) (ranked at 388), and it is uncertain whether further cleanup action will be required. 3.3.4.4 New Historical Data on Stroudsburg-- During RTI's historical literature review of the town gas industry, sev- eral items were uncovered that will result in reevaluations of previously reported information about the Stroudsburg site. These observations concern (1) the gas production processes used at the plant, (2) the previously report- ed method of waste disposal (injection well), (3) the source of the tar con- tamination, and (4) the nature of tar products from the site. The Stroudsburg site has always been reported as a coal-gas production site. Table 56 shows the gas production at the site as compiled from Brown's Directory, which lists the gas production process as oil and steam (1891 to 1394), Van Syckel oil process (1894 to 1904), and Lowe carbureted water gas (1912 to 1952). The process specifics for the oil and steam gas production 292
------- h-l z >- VI UJ 0. • o ce 1 en w ^ 2 hi O ce 0. V) o • to UJ UJ si J* a n is L 0 X *> o u u Q 4j w 3 U e o « 0 0. ^-s. ^ >^ m* v ID ••« C 'JJ u 3 •o 0 u a. in IA Proce L A >. O U +3 X Ol _J •o o C VJ n n vi a a o U X O V) 2 to CM CM T a. \J s CM i-* E a 0 VI c !«.. O Ol CM 00 Ol 09 CO «^ O 0 • ^J £ Ol I/I <- n 3 O JJ IA « TJ C 3 0 0 N U • • J,} 43 I/} ,« U IA G> U) ii * CM •» •» «0 Q Q O O S S S S Q S S S S CM Q u> ^ r* f, r. CM CM CM CM CM CM CM CM CM CM UJ • n ^ to **• <0 _ — _ — -o 0 0 0^— 0^-. 0n je Jt Jt in jc in *> n u U on u a 0a> X X X Ol X Ol U n n no no «n > > >^ >^ o* Ol ® Q O (9 GD Ol Ol Ol Ol O 'u x jJ 4> O» — U t- U 03 n — jo a UJ n n •o o •* 3 0 L in i. « n •** TJ O W — 0 c o S . 1 — & s *> 0 W U) Ul CM CM CM CM r- CO •o *~. 0 « n • *> n no 001 o w L ••"' JJ 0 0X U JJ > O n n on u * -J3 § 2 01 Ol « • V o w n O) • X C M x *> o e 'u '*» — U A U — n a. 3 « a n -a a n u o U 18 u as V. 0 CO 9 "^ W * •« — « o O X f" JC t. t§ *" * T) *• 6t 0 O r- « IS Q Q Q co O Q Q O M r* o> Q W in • * * #H r* rH §C& Q O §03 S (S O Q (S s" o" o sT CM •3 +> 6) CM CM CM S CO (O CM CM CM CM <-i|A 51 CM CM ^- ^ v>4 r* »•• ^ CH ^^ ^•^ ^^ <*• ^N • M • IA • VI • M • HO no no no no nu ou o o O u uu U «»* *»* ^^ Ni^ SM' • x 0 X 0 X 0 X v X 0U * o * o * u *u sn on on on on 03 _I3 _J3 -12 -J3 _l CM T tO B V 1-1 « f-< CM C4 Ol 0> 01 Ol 01 ,-1 1-1 ^H !->!-> L 0 c •— o o •^ & — n c 'o 0 n o c • no 0 *> • n o) •*> u o a 3 U « U •o «« n o 0 e L t- *> o 0 x a n — jc *> t- — o c vi a n u n n O) 01 vi M n c <*> c c • U 0 Ol O — 0 — O *> f» *> 0 *» 0 «^ « 3 3 *o o o * co 0 O) r-1 O> T IA D s (& » C*4 o 3 CD s Ul Ol • ^ IA •o « M *> n 0 Ol V. a i. n n w * CO CM Ol ^x •o 0 3 C ~ 4) C o u >M ' 4 293
------- ^ 3 C 49 C 0 u to Ul Ul m K It n •o u f O M L I O 41 M 3 U C o .« 4} It 3 a 0 a. *••* L X to H_ (O s r* C O 4l U 3 •^ O L. 0. VI VI 0 U o u a. u £ «*» O X • U /-K flj • .«. 0 «. ^5 ^J •* •" CO- (BOO — J3 — — VI 0 3 V. 4> It 3 12. o 3 0>"~ 0 TO VI U ID VI It J3 it « n O) O It 01 W u CM c n v- i-. o C — O CO 4> 0 It «•- - N L i-i n — 4> 13 r- CM U> & O to •o 0 ID 4> It 0 O> L. 3 C 01 0 It It U * CM n Ol VI *^ It 9 VI — 3 It O — — W n o con u O 3 — « 0 « 4> It C 3 U u oi— • o n 3 E— C 4> •O HI 3 0 » Ol O 4> — JC It O) it 0 VI O Ol CO VI U 3 C U t. r- c 0 ««- o r- O Ul O C 4> C Q *. -4> 0 u o r- (*1 Ok 0 <*) 4l • "O ui 10 ~* Ul *r 0 fH Ql •• fN f* M n o 13 CO Ul r- « CM O U] O CM *H 3 CO N CM U) 7 «^ to TJ O in *> it 0 0) 3 L. ^) 0 U 4) n it *j > to CO Ol •^ 0 « 3 — It — O — It U • • •«• Q c o u n • 0 3 VI — « VI O V. 4> It 3 CO y 01 o — 4» 3 c En •o w — 31. 0 C E *> 0 1. 01 vi it .a 0 in It 01 — C C O) VI — O — S C O 4> U to 0 JO -0 o CM 4> to 0 N- . C IO M U> i-l — r* 3 •O T 1O C 1-1 tO !•» VI It L. 3 1. U 41 It It U t ta o> •M — c n • .. — .- o — co on O T3 O — W 0 M U J It W 3 U 01 3 O0 3 C 4> t) VI — O C It E U VI It M O •• JO 4) It Ol C 4J — C 0) o n 0 n IQ CM 4> t. 3 C U ro 0 <•. o v» o co r- c 4> c v- ««•« 0 U 0 m CM 0) 0 CM 45 ^ u> » — CM. 0 -i .1 — h- 09 « O D 03 J Ol CM S IO n ^4 3 CO S w u> to Ul CO •o 0 M 4> It 3 I. L 4> n it VJ % T V Ol f-4 C C ^9 *O 0 0 VI M — 3 3 e'o 0 0 O 4> *> 4> • O U 0 Ol It • 3 t. 1. 0 C 4> 4) — U v> n vi o vi • I* O> C 4> C H Ol O It O L U Ol 0 — O ^ 03 C CJ — ». .en 0 09 o O f- O>f* J3 O CM « 1-1 VI 3 3 CO CM CM Ul CO ^ Ol 0 VI 4) n 0 0> u 3 t. ^a 0 L 4l n n u > OD « 01 w •o n 0 w — u n 3 L. w 41 n n us Z - JS 0 L 4J CM Ul 01 CO % 294
------- (and the Van Syckel oil-gas process) were not found during this study, but processes of this type generally sprayed oil and steam into an externally heated retort. The oil cracked into lighter gaseous hydrocarbons, and the steam reacted with carbon to produce CO and H2- The Lowe carbureted water-gas process is described in Section 1.2.3 of this report. There is no indication in Brown's that coal carbonization ever occurred at the Stroudsburg site. Brown's also shows the Stroudsburg plant as operating into the 1950's, with natural gas being installed sometime between 1952 and 1956. According to these data, the plant was operated primarily as a carbureted water-gas plant. It has been reported that an injection well was used to dispose of waste tars at Stroudsburg. When tar was produced and separated from town gases, it was usually stored in an underground tank until sold or used. These tanks were called "tar wells," in that tar was placed into the tanks and pumped out as if one were removing water from a well. The tar wells were labeled as "tar well" on plans and maps of the sites. They were also sometimes completely underground, with only a pipe visible from the surface for removal and filling of tars from the tar well. Unless the notation on the site map was clearly labeled as a tar disposal well or an injection well, it is possible that it was actually a tar storage well. There are two other possible sources that could have caused the subsurface tar contamination. Leaks of tar and oils from carbureted water-gas plants were very common. Underground tar wells (for tar storage) were often constructed of masonry and leaked. Underground liquid storage tanks were sometimes constructed of"wood. Tars were frequently placed in the gasholder for storage (gas sometimes blew around the tar-water seal for the holder, blowing tar out of the holder and onto the ground). The bottom of the gasholder was frequently below the groundlevel and also was prone to leaks. Underground pipes also leaked oil and tar materials into the ground. The second likely source of the tar contamination is the disposal trench ' described by Lafornara et al. (1982). The tars and emulsions -draining into : the ground from the trench would flow downward until stopped, and they would '; have accumulated in the area where the subsurface tars were located. The f f amount of tar produced by the plant in 1936 was 15,000 gallons (this is about [' 10 percent of the gas oil used that year). Thus, finding 10,000 gallons of 295
------- free tar underground (and maybe 5,000 to 20,000 gallons of tar [this estimate is a guess) left in the ground] is approximately 1 to 2 years of tar produc- tion during this period. The Stroudsburg tar (as described by VilVaume, 1982) is a carbureted water-gas tar, not a coal tar. It is only slightly more dense than water (P = 1.02 g/cuP), contains very little nitrogen (0.17 percent), has no tar acids, and has a viscosity of 19 cp (45 8F). Coal tar would be denser (1.1 to 1.2 g/cm^), contain more nitrogen, have some phenols, and be more viscous. The density of the tar is so close to that of water that it would be very difficult to separate a tar-water emulsion. Lafornara states that "Treat- ability studies performed on a coal tar/water emulsion pumped from the back- water revealed that no cost-effective method could be found to separate the emulsion and treat the water." This is precisely why the water gas tar was originally disposed during plant operation. Such an emulsion would frequently be disposed. The distillation curve (90 percent at 662 °F) shows that the tar did not contain very much heavier boiling organics, which probably indicates they were removed in the washbox and not disposed with this tar. If this tar could have been successfully recovered at the plant, it either would have been burned or added to the carburetion oils. The water-gas plant bought large quantities of oil that were poorer carburetion oils than was the recovered Stroudsburg tar. 3.3.5 Pittsburgh, New York (Thompson et al., 1983) The coal-gas and carbureted water-gas plant in Plattsburgh, New York, was in operation from 1896 to 1957. The plant was located on 11 acres of land on the south bank of the Saranac River. Byproduct tar was disposed in unlined ponds just above the river. Over several decades, coal tar could be periodic- ally observed on the south side of the riverbed as globules and as a film along the riverbank. This case study illustrates the following: • Site discovery through discharge into an adjacent waterbody • Coal-tar migration during active disposal by slow downward movement through subsurface soils along a dense till layer and from occasional overflow of the ponds during heavy rainfall 296
------- • Various influences on contaminant migration including seasonal groundwater fluctuations causing changes in pore pressure, increased temperatures during summers causing coal tar to become more mobile due to decreased surface tension and viscos- ity, and increased river flow causing a flushing of the contam- inants from the soil • Remediation by containing the contaminants onsite (Two con- tainment structures include cells built of a soil-bentonite slurry wall keyed into an underlying, low-permeability till layer and capped with a 36-mil Hypalon liner covered with 15 centimeters of sand, topsoiled, and seeded, and a second cement-bentonite wall built along the riverfront to prevent migration of contaminant not contained within the soil- bentonite cells.) • Remediation with a groundwater collection system being built to collect waters draining from the uncontained contaminated site (These waters will be pumped to water treatment equipment, treated, and discharged into the Saranac River.) • Use of the 4 acres of reclaims that lie along the river as part of the City of Pittsburgh's riverfront park system. 3.3.5.1 Site History— A coal and carbureted water-gas plant was operated within the city limits of Pittsburgh, New York, from 1896 to 1957. The New York State Electric and Gas Corporation (NYSEG) purchased the site and coal gasification plant from Eastern New York Electric and Gas Corporation in 1929. The plant was located on 11 acres of land on the south bank of the Saranac River. The topography falls gently in steps from an approximate elevation of 125 to 130 feet mean sea level (MSL) along the south edge of the site to 102 to 107 feet MSL along the Saranac riverbank. Other than a narrow band of trees and bushes adjacent to the river, most of the site has been cleared and filled. Two structures that cross the site are a 24-inch diameter concrete sanitary sewer and an active transmission line (owned by the Plattsburgh Municipal Lighting Dis- trict) (see Figure 71), This land consists of two parcels. The larger parcel (approximately 9 acres owned by NYSEG) lies uphill to the south and is the old site of the gas plant. The smaller parcel (approximately 2 acres) is a long narrow strip of land that fronts the Saranac River just downhill (to the north) of the NYSEG gas plant. This parcel was given to the City of Plattsburgh in 1981 by 297
~~------- I Reproduced from [best available copy. Figure 71. Pittsburgh, New York, general site plan. Source: Thompson et al., 1983. 298~~
------- as a contribution to the city's long-range plan for recreational development of the Saranac River inside the city. Table 57 is a list of the gas productions as recorded in Brown's Direc- tory. This plant produced primarily water gas over its history, although I notations in 1906, 1924, and 1936 indicated that coal was also carbonized at : the plant. t I Byproduct tar and condensate from the gas production was disposed in unlined ponds on the NYSEG property just uphill from the Saranac River. No records of the amount and times of tar disposal into the unlined ponds could be found. After the plant shut down in 1957, the ponds were filled with ran- dom material and covered with layers of cinders and ash. Over the years, this coal tar migrated downhill across the property now owned by the city and into the Saranac River. This migration occurred via two routes: by slow downward movement through subsurface soils, and from occasional overflow of the ponds during periods of heavy rainfall. Tar can be observed periodically on the south side of the riverbed both as globules of coal tar and as film along the riverbank. This problem, which has been in existence for some years, has been attributed to seepage of the tar from the previously existing tar-ponding " areas on the site. To address the problem, NYSEG conducted a geotechnical investigation during the summer of 1979. This fieldwork and laboratory testing, together with preliminary, alternative strategies for site remediation, were completed in early 1980. Following review of this work, a supplementary program of soil boring and testing was undertaken in November 1980. Actual site remediation occurred between September 1981 and September 1982. Remediation activities were coordinated with the City of Pittsburgh's long-range plans for recrea- tional development of the Saranac riverbank, including the parcel given to the city by NYSEG. Construction plans include building scenic overlooks for fish- ; ing during trout season and a pedestrian bridge to cross the river. 3.3.5.2 Extent of Contamination-- To define the site geology, hydrology, and area of contamination, a total of 53 boreholes were drilled across the site. In addition to these boreholes, three test pits were excavated to obtain bulk samples of the tar and soil for 299
------- 0 £ UJ . 5 CC 03 to t- < CL t- < t- O O O CC CL CO 3 . f- U) UJ 03 ^ ^» It n •o u c JC o M I o 3 U c o *> n 3 OP O CL u <^ .^ 0 s c o 4J U ^3 O u CL M M O o L 0. • u (1 C e t o CL "* II C I • |3 .c o _i j» Ol L 3 J3 M *» O « 0. r) CM r» u) p-< Q 0> co v u> - SO S Q S Q S S Q Q Q SI S> S Q <3 S S Q Q S S B Q CO* Co" 03 CO 0* S* «" 5>" 0.0. CL'W'WTTT 1 1 1 1 1 in ui to CM CM CM CM ® CMCM CMCMCMCNCMCM U) CM • rt rt CM C ni t * * * * > a X n oo ooooooo -J -J -1 _l _l _J -1 _J 0 O> « CM^SMVO CD O) Ok Ol Q 6) O Q "S 2222 ^ ^ "O «l « 3 N .. C — 0 -O •~ c vt 4> II II u u a 3 •o — « one LOO 0. O — VI VI A a c 0> Ol O ^ O t- CM O rH Ol •*• ^f * U) r«- ^3 • Ol • f* r-1 M = r- a Q to oi »-• cj co *r (O Q r^ u> O> O) C3 Q »-* C*l T (O Q Q Q Q CO f) CO Q Q Q O rH »«H r* O r-» »—* f-H O S Q 0. U 3 3 3 3 3 3 4> 4> *> 41 *» *J ^r 0.03 0.03 Qj.03 0.03 0.03 03 1 CM U) 03 CD S t Q MIA QCM CM" <9>H O) CM S CM CM (O CM tO CM ffl CM tO — • 3 U U Q O — 3 C U T) «l — 4> o e E U L. O 3 « a— a UJ M n-o • (10) ill > 01 n L Z T C « t- UJ O — e o in *> — us- « o « to u> .a *» T3 r« oj « • C4 (0 a « UJD 10 00 03 m ** (N 3 00 s U) U) • tc 1-1 1 9 O CO CM 01 1-1 •o 9 3 e •~ c u *•«•.* 300
------- •a 0 3 C 4} C o u r*. m Ul 03 H™ n A ^j L ^ O 1. g o n 3 U e 0 •« ^> n 3 a o 0. u >» «"" to (0 6) ^* c o "j u 3 O L ft. M Proce; t. n > U L. O O «•- «*• — it n 0 O C 0 U U o — M 10 M *> It 3 3 U 01 0 0 3 C C •o in — o c E E <- O 3 3 a— *> »» *J _ .- u — « i» .a o jo l» O) *» 01 n s- - 0 f^ CO OlF^ •Q 03 U) r- 0 CM 09 U) ID CM (O CO _T M en n m" (•4 9 CO O) U) 1^, j^- fM D, o -I (S m ••4 M n 0 .4 U •— U 4} u 0 ^ tu o Jj n VI 4^ u M 0 — je i. O O JO ^* • o * u 1 c II a. E o u V. >, 1C *> -o 3 *-v (XO *> i c 019 M S • « X CM a m 2 CM L 0 > «i_ ^* ^v It 0 • C •on it /-» C 3 U HH O — It M « « « C2—U.C IIM it <• n 3*)C4J Ollt Ol O) 0) ^ 0 — w O> 3 1. U L L 0*OCMjC 0— 0 0 0 j o m .a 4>n *> 4j *> o u a> •- no n it n _ICu~* * U * * * CM 01 0) 01 . 0 VJ • ^ m e >. a g u o> i o a 41 U • > O 0C1 O 0 *> *> I. .^ r t a c • i-« •o— u W 0 0 « e 0 *> o -O * u — arc O 3 U A) "O (. -O II 0 «~ « co u a— -^*> it n -o 3 — u u c -u e It 03 O ui n i. u u 0 II II ce »|0 u ^ o o"5 a. *» 3 M _IX U 03 D 301
------- laboratory testing. Nineteen standpipe piezometers were installed to monitor groundwater levels across the site. The borings indicated the presence of an extremely dense till underlying the entire site. This till consist? of silt and fine sand intermixed with medium- to coarse-grained sand and gravel. The till appears to have served as a barrier over the years, halting vertical migration of the coal tar on the site. No tar was observed below this till anywhere on the site. However, in the sandy soil and fill layers above this till, tar contami- nation was found over most of the site. In the area or the original tar ponds, contaminated soils were found as deep as 4 meters.- From this region of maximum soil contamination, the thickness of the contaminated soil gradually lessened toward the NYSEG property boundaries except for a layer of contamina- tion extending across the city's parcel to the north and into the riverbed of the Saranac River. The data from the borings indicated that the subsurface movement cf tar from the ponds had been downward through the permeable sands and gravels and then laterally along the top of the till toward the river. No tar was observed below the till layer (Thompson et a!., 1983). A laboratory testing program was undertaken to further characterize the contamination. Tar content (percent dry weight) in contaminated soils was found to be as high as 9.6 percent with an average content of 1.5 percent. Tests to determine total Teachable salts in the soil/coal tar showed low con- centrations of metals (although Teachable arsenic was reported at 2 and 3 ppm and lead at 0.9 and 1 ppm in two samples). Determination of total Teachable salts in tar reported for three samples showed high chemical oxygen demand (COD) and total organic carbon (TOC) at 850, 900, and 935 ppm. Leachable phenol was as high as 4 ppm in a tar sample taken from the Saranac River (Thompson et al., 1983). The investigations determined that tar migration has decreased exponen- tially since disposal of tars was halted in 1957. When active disposal was in progress, the sands, silts, and gravels beneath the ponds became saturated with tar. The higher viscosity of the tar and its immiscible properties allowed the tar to migrate in density currents as a separate phase from the groundwater. With continued disposal, movement of the tar occurred relatively rapidly downgradient along the top of the till layer into the river. Once the'' 302
------- tar disposal stopped, the rate of migration gradually decreased. Thompson et al. (1983) believe that the majority of the tar currently onsite is being retained within the pores and matrix structure of the soil grains by capillary forces, and that the mechanism causing the tar migration today is different from that when the ponds were in operation. Although difficult to quantify, the mechanism causing tar migration today is most likely influenced by one or more factors, including seasonal groundwater fluctuations causing changes in pore-water pressure, increased ground and groundwater temperature during summer causing the tar to become more mobile due to decreased surface tension and viscosity, and increased riverflow causing a flushing of the contaminants from the soil. 3.3.5.3 Site Remediation-- Site remediation occurred in two phases. The Phase I Project focused on arresting the subsurface migration of coal tar away from the area of the orig- inal disposal ponds. The Phase II Project addressed the cleanup of the Saranac River and the city-owned property to the north. Phase I began in the fall of 1981 with the installation of a soil-benton- 4 ite slurry wall around the main tar pond area (735 feet in perimeter). This wall was keyed into the underlying impervious till that was 4 to 6 meters below grade in the main-pond area. This main-pond area was then capped with a temporary 20-mil polyvinyl chloride (PVC) liner. It was estimated that approximately 80 percent of the onsite coal tar was encapsulated within this containment cell. A well was placed within the cell to monitor the effective- ness of isolation. Phase II remediation activities began in June 1982 with the installation of a temporary, portable fabric cofferdam in the Saranac River. Behind this cofferdam, tar contamination in the riverbed was excavated in the dry. Water was pumped from the area of excavation into a triple-compartment settlement tank before being discharged back into the river. Riverbed cleanup was per- formed in two stages moving from upstream to downstream. The temporary PVC liner that had been placed as a cap over the previously constructed containment cell was perforated, and the contaminated material excavated from the river was placed on top. Additional contaminated materials were placed in an area just to the southwest of the original containment 303
------- Later, this additional area was also surrounded with a soi1-bentonite slurry wall and thus represented an enlargement (almost a doubling) of the size of the original containment cell. After excavation of all visible contamination in the riverbed and along the riverbank, the riverbed and bank were reestablished to grade with imported clean fill. To prevent continued migration of remaining uncontained tar into the riverbed area, a cement-bentonite cutoff wall was constructed through the clean fill for approximately 213 meters along the riverbank. A cement- bentonite wall was used in this area (instead of soil-bentonite wall used previously on the NYSEG property) because a higher strength wall was consider- ed necessary to meet the city's plans for recreational development of this area. To intercept drainage of groundwater from the uphill area above the cement-bentonite wall paralleling the river, a groundwater collection system was installed. This system consists of a 15-centimeter perforated drainpipe 0.6 meters below grade and 3 meters upgradient of the cement-bentonite wall. This drainpipe discharges into a precast manhole at the midpoint of the line. Water collected by this system is pumped back uphill to water treatment equip- ment located in the vicinity of the coal-tar containment cell. Treated groundwater has been discharged into the Saranac River since September 1982. After grading the contaminated soil in the areas inside the walls of the containment cells, the cells were permanently capped with a 36-mil Hypalon liner. This liner was then covered with 15 centimeters of sand, topsoiled, and seeded. This site work was completed in September 1982. Because so much tar contamination has simply been contained onsite, future use of both the NYSEG and City of Plattsburgh parcels will have to be carefully guarded. Specifically, certain restrictions to onsite development have been mandated by the NYSDEC, and other restrictions have been suggested by NYSEG, who will remain responsible for maintaining the slurry walls, con- tainment cell, groundwater collection and treatment system, and monitoring network on both parcels. These restrictions are: • Sale of the lands on which the containment cell was constructed is prohibited by NYSDEC. 304
------- • No structures or other activities that could result in rupture to the Hypalon membrane may be placed or performed on the con- tainment cell. • All trees or shrubs will be maintained at a distance from the slurry walls such that their mature drip line will not inter- sect the slurry walls. • All construction on or near the cement-bentonite partial cutoff wall and/or groundwater collection system must have prior engi- neering approval of NYSEG. 3.3.6 Seattle, Washington (Cole, 1972a and b; Cole and Machno, 1971; Drew, 1984; Haag, 1971; Royer, 1984; Mayor's Committee on Gas Works Park, 1984; Orth, 1984; Steinbrueck, 1971) The Seattle Gas Works plant was in operation for approximately 50 years. A large portion of the waste byproducts were disposed offsite, but large quan- tities of lampblack were disposed onsite, building up the shoreline into the adjacent Lake Union in Seattle, Washington. This case study illustrates the following: • Site discovery through redevelopment as a park • Large stockpiling of lampblack filling in Lake Union • Conversion of the site into a public park by partial building demolition, composting of contaminated soils in preparation for planting, without removal of onsite contaminants • Closing of park • Present ongoing investigations to determine whether further remediation is necessary. 3.3.6.1 Site History— The GuS Works Park is located on a point projecting into Lake Union in Seattle, Washington. The park occupies about 20.5 acres, which includes some 1,900 linear feet of waterfront. The surrounding area is mainly industrial property. The Lake Union site known as Brown's Point, once a popular spot for pic- nicking, was developed in 1906 by the Seattle Lighting Company as a gas plant. The location of the plant on Lake Union made it ideal for the barge delivery of local and imported coal (and later, oil) for gas production. Eventually, 305
------- the site became known as the Gas Company Peninsula, built by a slow process of filling in Lake Union with cinders, unusable coal and coke, and gas production wastes. The Seattle Lighting Company became the Seattle Gas Company in 1930 and eventually was made part of the Washington Natural Gas Company (WNG). The original plant on Lake Union produced illuminating, heating and cook- ing, and industrial gases for the growing Seattle community. Coke ovens were operated, and retort gas and carbureted water gas were produced. During the mid-1930's, six water-gas sets were in operation with a total daily capacity of 6,600,000 ft3 of gas (Steinbrueck, 1971). The byproducts of the gas plant operations were ammonia, light oils (benzene, toluene, xylenes), various other hydrocarbons, and tar, which was refined into creosote. Tar and creosote produced by the Seattle Gas Company were delivered to the American Tar Company, which was located adjacent to the Seattle Gas Company until about 1920. The tar company refined the coal tar into various grades of tars and pitches using steam distillation (Orth, 1984). In 1937, oil replaced coal carbonization as the basis for gas production. The plant continued to produce wafer gas. Table 58 shows the gas production and byproducts from Seattle as compiled from Brown's Directory. Oil-gas tars contained more asphaltene-type compounds than did the coal tars produced earlier and were not suitable for the products derived from the coal tars. Thus, the oil-gas tars were generally used as fuel for steam production. The tar emulsion from the Jones crackers was over 90 percent water and had to be concentrated before it could be burned. Naphthalene and related .aromatic oils were collected in the condensation from this process. The naphthalene was sometimes combined with creosote oils and sold, but it often was simply dumped offsite (Orth, 1984). The lampblack from the oil-gas cracking operation was dried for bri- quetting and used to replace coke in the water-gas sets. However, the bri- quets would often break during the firing. As a result, thers was consider- able waste. The lampblack production far exceeded the use, and the excess was piled next to the lake. The pile of lampblack grew to nearly 100 feet high and covered several acres (Orth, 1984). There were frequent complaints of odors from the plant and from the wind dispersal of the lampblack. 306
------- (9 0) . « oi- -5 • * °t:c 00. *J °^ .f * o>c3 UJ COW •*•& en-o «i .— .- 0 I/I O C • S - c S. - S 3 8 I ji .- *> O O •*» 2"u A .- « wo wi IS 11 (O tn S s tA •o n II •si « n a O u a an o oo u c n " o I I ! I ! ! 307
------- x* 3 C *— +3 C o 03 in LU m 1- n *> a •o t_ o •*> o t t 3 VI 3 U C o « — n 3 a 0 a. X ^^ m «*. (0 s^ C O u 3 •o O L a. VI VI O u o L. o. u e H .« i. O U 19 19 *J • C VI *» •« on — — o> — « n o j .. .. o — n u in -Q o u e 3 — C « O O OO— DCC6O i-u — -ooonu a no n ai i. — — n in IA c a. n n .a e co *> Si CM CM UP) - U Q co *^ o to *^ «, * « Q -in o P> M r>- - -O a - i. co T in -« 0 r- CM a in in t <•> VI X D 03 PI f*. CD in to C3 1 IS CM P> 3 CO in Q in CM C» PI T S> CM ••4 -« U1 • f-t c 0 o — 0 0 n * -x O 0 O 0> TJ — 0 n N — 0 .- ... c o CO— C • •0 — VI C J3 «l t) 3 O U B 0 t> 0 — n 01 «i 0 C in *j o 3 in .- . .. n — u n 3 E in -o *> c 3 — C 0 3V.00 0 -onojc 0 *> « u jt in | OO— O J« • 3OC6 w u — u o js — ijuon o. n n u o o — incnviu vi j i, vi — in we con c acnjo Ol^JCfl^HOU^ •-»!*) M U *> 0 *> f< Q U CO -CO 0 — tO "0 O f S P) o CM s co c n — oo 0 3 to S Si >»0)^cn0p)ooiMi3aa - - • O) - CAW JO - 3 0 -M CO O tO CM CM P) ri CLVCniA M X D ffi ^ in p> CO s" m p> 3 03 in S in O) IA r- r- M CO • *-t O 0 1 c •a 0 c > n o — 0 n jt o o uu s PI 0> •o 0 N — — .- e c o — •• e O i. ig 0 — n O) vi *> 0 3 U « 3 — C « •o n o je O o — o •- o — o a. n n en vi we c n o o o oiJj s *> U 1C -« O f CM o a" 0) PI 9 03 O» in 1? T CO f4 M n vi O) A O) u 0 — *> n « o 36 0 u at C •* 3 tN 0 N TS M TO v+ O • • U CO CO CLP) X CO w •o — e - .- N .- 0 — — o — C X C C O .. o — > — c j> 0 n x O U VI 0 T) t} > — It 0 JC 0 0 II *> o — w n 0 • U •DM3 3 JC » 3 — C L -a n « o o o n « o o c — jt je e — I. U O — O O O» a — n u w u — o vi _ o> U — X wen n o « • n o Q)^* c 4^ c vi o) c cn*J sonou •- I.COIN> 0 — in -o OCDl~-TincCM— •» 0 s- CM -M n 0 in o - n -Minm oii- jo R 3 *^J ^ 5) - - ^ *>4 0 ^ CM U> 00 P) -" IA D s P) r>- isT to o p> to csT O) p> 3 co ^ Q in P) Q in f« CM in •. *H M « C n 0 en > o c 0 O *> JC n o 36. W> to 01 rt •o 0 3 C ^} C O u \^ 308
------- in it * c o « •o a « •> — o « 20* •* '^.-«« i r.r. V» UVs. p. i*5v 'o •> - i.S 0 ss .» U ..- "2-2 H I. M O C — c 0 — a •»> a « c > a .- w jj 0 E 0 — w .- II 0 01 M « C C o o t. II ".._•- o '^VSxc1 « S e« „ *2 c a •» £ 5 »~ CN r^ U cn u in co ' i. 5 in i- i « T.S S ! n-o^'S co -co' -Sco-s^ -S2:.-- acs« | CD 3 I « 15 n « 01 o 0 Q 3 S »- o a u i* • « c n it « — Ol OI — c • O OI 0 I I 3 ' o 01 ^ O> 0 O> « *» *.: r- « CN (O m ID o 03 s 01 s Q r) r- ID CD (O •o OI I" II l_ OI 0 n — a M A I. OI 0 *J — f o O U (O n Oi s a> 309
~~------- 310~~
------- The company continued to produce gas until 1956, when a natural gas pipe- line was extended to Seattle. After that, VING used the site for storage and other activities. During the plant's operation, the shoreline on the penin- sula had been extended some 24 meters into Lake Union. Eventually, the site was almost flat down to the lake's edge where there was a 2.4 meter drop. In 1962, the City of Seattle purchased the peninsula for development as a public park. A bond resolution passed in 1968, providing funds for park development, and planning for the park was initiated. The city hired a land- scape architect, Mr. Richard Haag, to propose a master plan for the park. After a study of the site, Haag determined that traditional park development would be impractical and proposed a controversial plan that allowed for the restoration and reuse of some of the gasworks structures. The plan for the site demolition (to be done by WNG in 1971 under the 1962 purchase agreement) called for leaving six generator towers, the pre-cooler towers, a boiler house, and an exhauster building. Haag concluded that it would not be pos- sible to remove all of the underground piping and existing soil from the site, nor to cover the entire site sufficiently to permit the growth of large trees essential to a more traditional park design. Despite the controversy over allowing the former plant structures to remain, the city council finally approved Haag's plan in 1972. ' 3.3.6.2 Extent of the Contamination-- Some 50 years of heavy industrial use at a time when there was little concern for environmental contamination had left the site on Lake Union heav- ily contaminated with residues from production, spills, waste materials, and air pollution fallout.- Haag, the landscape architect, expressed concern for the ability of the site to support vegetation, noting that there was no "natu- ral" soil on the site. He described the condition of the soil as a sterile layer cake of hydrocarbon contamination that supports no vegetation (Haag, 1971). Studies were undertaken by the Seattle Engineering Department and by Dr. Dale Cole and Peter Machno of the University of Washington to characterize drainage patterns and soil conditions at the site. 3.3.6.3 Site Remediation— The description of the remediation activities below is summarized from information contained in a document made available by the site manager in the 311
------- U.S. EPA Regional Office. The document is not commonly available but was probably prepared in 1984. After the removal of the above-ground structures by WNG in 1971, consid- erable site preparation work was still needed. The primary intent was to stockpile and/or bury onsite much of the excavated material and demolition rubble. The stockpiling was in the central portion of the site. Portions of the stockpile were later buried onsite. Several existing structures consid- ered potential safety hazards were removed. WNG was required to purge certain pipes in 1973. The mound area in the southwest portion of the site consisted of excava- tion materials from offsite. This fill had been brought to the site during the 1960's and early 1970's. It was thought at one time that this fill mate- rial could be used to cover the entire site following the demolition of the above-ground structures. However, the "Great Mound" became a major element of the master plan for the park, and it was cleared, grassed, and opened to the public for the purpose of viewing the ongoing park development. Work contracted by the Parks Department included the following tasks: • Demolition and burial in the northwest section of the rubble from 13 concrete purifiers that were located just east of the tower area • Removal and stockpiling of the contents of the purifiers (i.e., woodchips coated with iron oxide and residue from the purifica- tion process) • Removal and burial in the northwest section of the concrete slab remaining from the 2 million ft3 storage holder Demolition of remaining concrete foundations and piping • Excavation and removal or stockpiling onsite of approximately 20,000 to 30,000 yd3 of badly contaminated soils • Regrading of demolition areas to match the surrounding ground level. In the process of removing contaminated material and burying rubble and debris, there was concern of increased pollution to surrounding areas, partic- ularly Lake Union. Of particular concern was the excavation of the contami- nated soil in the southwest area. The contract specifications cautioned the 312
------- contractor responsible for this work of the conditions there. The contract stated, "Excavating oil-gas contaminated material at the southwest property edge shall be performed with extreme care. This excavation extends to the lake level and shall commence 30 feet or more inland from the water's edge. Demolition work and pipe removal shall be completed prior to any excavating of this 30 foot wide levee. When the inland area Is excavated, filled and/or graded to the proposed grade the levee at the lake's edge shall be removed." One part of the site preparation work involved efforts to improve growing conditions by an application of a compost-like mixture containing dewatered- sludge cake as the primary ingredient. The mixture was applied over approxi- mately 10 to 12 acres of the southerly half of the site (about 100 tons per acre, wet) and then worked into the top 18 to 24 inches by periodic plowing. Sawdust and leaves were also applied and worked into the surface soil. The surface was reworked, fertilized, and sown with a cover crop of grass about 2 weeks after the compost treatment. The first crop was plowed under, and the area was finally rehydroseeded. The actual park improvements were undertaken upon completion of the site preparation work. Phase I of the park development consisted of the following actions: • Renovation of the former boiler house for use as an indoor/out- door picnic shelter • Renovation of the former exhaust building for use as a "Play Barn" • Creation of a grassed picnic "Bowl" projecting to the water's edge • Construction of paths • Further development of an existing 170-car parking area • Deter access to the towers and remove miscellaneous structures • Regrade mound and hydroseed • Plant trees and shrubs and provide sod in one small section of the picnic area. 313
------- The work delineated above was completed, and the official park opening was held during the summer of 1976. Additional improvements were completed in 1978. Plans for further improvements were being finalized when the U.S. EPA began an investigation of contamination at the site. Soil testing during the park development was directed primarily at horti- cultural aspects of the design. The park did not include any significant amounts of fill. Cuts were made primarily in the southeast quadrant and between the mound and tower areas. Considerable soil was removed from the site, part of which was known to contain arsenic. No work was undertaken in the water areas surrounding the site. According to the Gas Works Park history: It appears that the development was directed at reusing the site in what was felt at the time to be an environmentally sensitive manner. Both the general design concept and the budget were important i -:- tors in the decisions that were made. The major controversial issues centered on the retention and reuse of structures associated with the former gas plant. Most of the discussion concerning the levels of pollution centered on what would and would not grow on the site. Public health was an issue, more in terms of access to the towers, aquatic activity from the park, and use of the Play Barn, than in terms of general use of the site (Gas Works Park, no date). Recognizing the severity of the buried contamination at the gasworks site, concern was expressed by some members of the community that opening up the soils of the Gas Company Peninsula could only worsen the potential for irreversible ecological damage to Lake Union. Notable among those voicing this concern was Mr. Otto Orth, Jr., a distinguished chemist and lifelong citizen of Seattle, who in 1984 recounted in a letter to the Seattle Times a history of the operations at the gasworks (Orth, 1984). During 1983 and 1984, Environmental Protection Agency and University of Washington investigators began to sample for toxic materials in offshore sediments and surface and subsurface soils. Because of the high levels of polyaromatic hydrocarbons [i.e., benzo(a)pyrene] and other contaminants reported, Mayor Charles Royer temporarily closed the park on April 21, 1984. He established a Health Advisory Committee that reopened portions of the park considered safe for the public. The committee agreed it would be prudent to conduct additional testing and investigations at the site. Tetratech, a consulting firm, was hired to carry out soil-sample and groundwater 314
------- investigations. A summary of the maximum polyaromatic-hydrocarbon concentrations found onsite is presented in Table 59. The groundwater investigation is still in progress. 3.3.7 Brattleboro-Hinsdale Bridge: Brattleboro, Vermont (E. C. Jordan Co., __ The Brattleboro, Vermont, site illustrates the following: • Site discovery during site investigation for a road construc- tion project Movement of dense tar components by the action of gravity along a subsurface bedrock surface, from the original disposal area to beneath a riverbed • Movement of tar in a coarse sand and gravel deposit • Limited groundwater contamination from the wastes 3.3.7.1 Site History-- During initial site explorations associated with constructing a bridge across the Connecticut River, the State of New Hampshire discovered "odorous, oily materials" in soil borings. Subsequent analysis indicated that the mate- rials were similar in composition to coal tars. Further investigation indi- cated that the site was the location of a town gas facility that was closed around 1949. One of the original gasworks buildings remains in use as a dis- tribution center for bottled gas. The planned bridge abutment is to be built between this building and the river. No detailed site history has been compiled on this plant. Table 60, which give the gas production data as compiled by the Radian Corp., shows that the plant produced carbureted water gas. Currently, a site contamination audit has been completed, including recommendations on how to remove and safely dispose of contaminated materials encountered during construction of the bridge. i 3.3.7.2 Extent of Contamination-- The initial exploratory borings indicated that there might be tar < ! contamination at the site, and the site contamination audit confirmed this 1 hypothesis. This investigation showed that the site was underlain by 5 to 15 feet of fill material that grades into alluvium as one proceeds out under the 315
------- TABLE 59. MAXIMUM CONTAMINANT LEVELS: GAS WORKS PARK, SEATTLE, WASHINGTON Soil Compound (PPm) Naphthalene !•& Acenaphthylene U 1° Acenaphthene U 20 Fluorene 7.4 Anthracene 10 Phenanthrene 26 Fluoranthene 65 Benzo(a)pyrene 28 Pyrene 10° Benzo(b)fluoranthene 28 6enzo(a)anthracene 26 Chrysene 33 Benzo(k)fluoranthene 11 Benzo(g,h,i,)perylene 29 Dibenz(a,h)anthracene 3.1 Indeno(l,2,3-c,d)pyrene 25 U = Undetected at the detection limit shown. 316
------- TABU 60. GAS PRODUCTION AT BRATTLEBORO, VERMONT Gas Production Year process (106 ft3/yr) 1890 Lowe 6 1900 Lowe 5 1910 Lowe 15 1920 Lowe 24 1930 Lowe 41 Lowe = Carbureted water gas. SOURCE: Radian Corp. from Brown's Directory. 317
------- river (Figure 72). Underlying this layer is about 10 feet of sand and gravel that rest upon weathered bedrock (phyllite) and also extends out under the river. The bedrock surface slopes downward under the western portion of the river from about 220 feet above mean sea level (AMSL), to the eastern bank where the bedrock surface is at about 70 feet AMSL. Figure 72 illustrates the extent of contamination under the old gasworks site and under the river, as far as Bridge Pier 1. It also illustrates how the contaminants have collected in the coarse sand and gravel immediately overlying the bedrock under the old gasworks building, and it shows that the contaminants have migrated through this coarse layer, down the bedrock surface, and under the river to the site of Bridge Pier 1. A borehole to the east, at the site of Bridge Pier 2, indicates that the coal tar has continued to migrate along the bedrock surface under the eastern portion of the river, where it occurs under 45 feet of sediments. This contaminant distribution clearly illustrates that the tar moved by way of density currents along the surface of the bedrock. The high permeability of the sand and gravel layer above the bedrock has enabled this migration to occur. Migration distance is at least 360 feet laterally and 150 feet deep from the contaminant source. Maximum contaminant levels for soil, river sediment, and groundwater are presented in Table 61. Maximum levels in soil were found to the east and to the west of the gasworks building (B-107, B-108, B-110). Maximum levels in sediment were found at the site of Pier 1 (BIOS, B106). Maximum groundwater contaminant levels occurred both onshore (MW-107) and at the Pier 1 site (B-105, B-106). Sediment contamination levels at the site of Pier 2 were about five times lower than those presented in Table 61; no PAH's were detected in the groundwater at this location. 3.3.7.3 Site Remediation-- To address the contamination previously described, the following recom- mendations were made: Any contaminated soils excavated during construction of Abut- ment A or Pier 1 should be removed and disposed in a secure hazardous waste landfill. • Suspended soil and visible contamination in water removed from the above construction areas should be removed. The water may then be discharged into the river without further treatment; no 318
------- ••'mi f WEST ABUTMENT AREA morn.! c •> I ••race wen i /.WEATHERED BEDROCK FINE SANDY SILT LEGEND I I O-IOPPM I . I IO • IOO PPM I I IOO . PPM SANOS AND GRAVELS fVIUIVI ^*^^^ "•' '• AND I *"~ PILU I ^^IT"— -.—; -L -TA^ ""^^ PRORLE' A-A' rirt*j»e a Source: E. C. Jordan, Co.. 1984. BEDROCK (PMVLLITCI FIGURES INTERPRETIVE ZONE OF COAL TAR CONTAMINATION SITE CONTAMNATON AUDIT -CONSTRUCTION PrtOCEXJBeS BRArOEBOnCHWSOALEBBOOe NEW IIAMPSltre DEPARTMENT OF PUDUC WORKS AND HOHWAYS • ECJORDANOar Figure 72. Brattleboro—Hinsdale Bridge. 319
------- TABLE 61. MAXIMUM CONTAMINANT LEVELS: BRATTLEBORO, VERMONT Compound Benzene Toluene Ethylbenzenc Xylenes Naphthalene Acenaphthylene Acenaphthene Fluorene Anthracene, Phenanthrene Fluoranthene Benzo(a)pyrene Pyrene Benzof luoranthene(b, k) Benzo(a)anthracene/Chrysene Benzo(g,h, i , )perylene Indeno(l,2,3,-c,d)pyrene Soil (ppm) 1.3 4.8 32 64 140 85 140 100 190 64 9.8 43 10 21 2.3 2.2 River sediment (ppm) 0.025 -- 0.130 0.27 180 1.3 28 22 240 72 4.8 77 4.8 8.5 1.4 i.r Groundwater (ppm) 0.15 0.20 0.27 0.79 5.5 0.27 0.84 0.051 H.037 0.0097 0.011 0.0094 0.01 0.0095 — — 320
------- NPDES permit will be required (New Hampshire Water Supply and ' Pollution Control Commission). • Pilings should be used to support the bridge at Pier 1 to mini- mize the removal of contaminated material. • Site safety and contingency plans should be developed to mini- mize worker and public exposure to contaminated material. The report concluded that the bridge could be constructed without signi- ficant environmental or public health impacts and that removal of all contami- nated materials would not be necessary. Since the report, the New Hampshire Department of Public Works and High- ways has decided to use pilings for both Abutment A and Pier 1, thereby avoid- ing any excavation. However, there is also the possibility of moving the bridge site upstream (for reasons other than site contamination), thus avoid- ing the contaminated area entirely. Vermont's Agency of the Environment con- siders the site to be of low priority because of low potential for release and contamination of groundwater, surface water, or air. 3.3.8 St. Louis Park, Minnesota (Barr Engineering Co., 1976; Ehrlich et al., 1982; Harris and Hansel, 1983; Hickok et al., 1982; Hult and Schoenberg, 1984; May et al., 1978; Minnesota Department of Health, 1938, 1974; Rittman et al., 1980; Schwartz, 1936; Schwarz, 1977; Sutton and Calder, 1975; U.S. Forest Products Laboratory, 1974) The Reilly Tar and Chemical Corporation operated a coal-tar distillation and wood preserving plant (80-acre site) in St. Louis Park, Minnesota, from 1918 to 1972. The plant wastes, consisting of solutions of phenolic compounds and a water-immiscible mixture of PAH's, were discharged into a network of L ditches emptying into an adjacent wetland. The contaminants entered under- : lying aquifers via the wetlands and multiaquifer wells in the area. In 1932, I the first well was shut down due to contamination, followed by others until | over 35 percent of St. Louis Park's water supply was shut down. In 1975, the Minnesota Pollution Control Agency conducted a study to assess the extent and magnitude of the contamination. Since then, the Reilly site has been desig- nated as the State of Minnesota's highest priority Superfund site. This case study illustrates the following: 321
------- !T • Site discovery through groundwater contamination • Contaminant transport via spill of drippings onsite, surface runoff, plant process-water discharge into adjacent wetlands, and movement of coal tar directly into bedrock aquifers through one or more deep wells used to drain creosote from the site and through one well that had experienced a spill into the well • Contamination of several aquifers due to other water wells in the area extending through several aquifers, thereby providing a pathway for the contamination to travel between aquifers • Contaminant migration in aquifers influenced by pumpage of water supply wells • Removal of phenolic compounds in groundwater by biodegradation and naphthalene concentrations being reduced due to sorption • Plan of remediation including a gradient-control well pumping system, a granular-activated carbon-filtering system, repair of leaking multiaquifer wells, removal of coal tar from any con- taminated wells (in particular W23), establishment of source control wells, and monitoring of all contaminated aquifers over a set period of time. 3.3.8.1 Site History-- The Reilly Tar and Chemical Corporation operated a coal-tar distillation and wood preserving plant (80-acre site) in St. Louis Park, Minnesota, from 1918 to 1972 (Figure 73). The plant wastes, consisting of solutions of phenolic compounds and a water-immiscible PAH mixture, were discharged into a network of ditches discharging into an adjacent wetland. The contaminants entered underlying aquifers via the wetlands and a 909-foot deep, plant site well (W23) (see Figure 73). Well W23 was drilled in 1917 as a source of cooling water for the plant. In 1932, the first St. Louis Park village well was drilled 3,500 feet from the plant. After only several weeks of operation, the well was shut down because of odors attributed to phenols. An investigation done by McCarthy Well Company (USGS files) concluded that the contaminants were entering the groundwater through old wells used to drain creosote from the site. One of the wells, W23, had experienced a spill of tar into ihe well, leading to con- tamination of several aquifers. By 1938, the Minnesota Department of Health (MDH) reported nine wells contaminated with phenolic or tar-like taste. The 322
------- Reproduced from best available copy. PM9 W1J WETLANDS AREA Source: Ehrlich et al.. 1982. figure 73. Location of former plant site, wetlands area, hydrologic section, water table configurations, and location of key wells at St. Louis Park, Minnesota. Generalized potentiometric surface, June 5, 1979, shown. 323
------- well farthest from the plant site was originally 280-feet deep {into the St. Peter aquifer; Schwartz, 1936). This well was deepened another 130 feet, extending into the Prairie du Chien-Jordan aquifer, and it immediately yielded a distinct tar-like taste. Throughout the 1960's and 1970's, the MDH and St. Louis Park monitored municipal, commercial, and industrial wells for phenol. In 1975, the Minne- sota Pollution Control Agency (MPCA) conducted a study to assess the extent and magnitude of contamination. The study concluded that soil and shallow unconsolidated sandy aquifers near the old Reilly site were seriously contami- nated and were the source of contamination to deeper bedrock aquifers. In 1978, PAH's, including benzo(a)pyrene, were found in several St. Louis Park municipal wells located 1/4 to 1/2 miles north of the site. These wells were closed down, followed by - re well closures in 1979 and 1981 until over 35 percent of the city's water supply capacity was shut down. In 1978, a USGS study of private wells in the St. Louis Park area, including Reilly's deep Well W23, revealed a down-hole flow of contaminated water from shallow aquifers to the Prairie du Chien-Jerdan aquifer. The flow was estimated at greater than 150 gallons per minute (gpm). The well was plugged to stop continuing downward water contamination. In 1982, the MPCA cleaned out Well W23, removing over 150 feet of coal-tar wastes and debris. All of the closed municipal wells draw from the Prairie du Chien-Jordan aquifer, as does 80 percent of the water supply to Minneapolis-St. Paul, of which St. Louis Park is a suburb. The Reilly site is designated as the State of Minnesota's highest priority Superfund site. 3.3.8.2 Extent of Contamination-- The vertical strata, including five major aquifers in the area, are shown in Figure 74. The Platteville Limestone is a nearly flat-lying, dolomite limestone. Fractures and solution channels contain water that yield small supplies to wells. The Glenwood Shale underlies the Platteville Limestone and serves as a confining bed except in locations where the shale has been eroded away. Glacial drift consisting of glacial till, outwash sand and gravel, lake deposits, and alluvium of several ages and provenances overlies the Platte- ville Limestone. The detailed stratigraphy of the drift at St. Louis Park is complex, but three areally persistent units have been identified. Directly 324
------- DEPTH BELOW LAMD HYDROGEOLOGIC SURFACE. UNIT IN FEET 100- 200- 300- 400- 500- 600- 700- 800- 900- 1000 J Drift Piaiteville aquifer Glenwood ^~ confining bed / St. Peter aquifer Basal St. Peter confining bed Prairie du Chien- Jordan aquifer S.t. Lawrence- Franconia confining bed Ironton-Galesville aquifer Eau Claire confining bed Mount Simon- Hinckley aquifer *'•*.'.'•'*.'' ,*'•.'•• •"•*. * • mi •*.*.•'•'.';."•"• K:; •':•> ^^^ C^"' ^' — ~^ ?PI ^^ Hi ife :-:-:-:: •vAv'V'v:'- GEOPHYSICAL LOGS Natural . .. Current Gamma Cahper meter ? 10-inch / casing \ ^*" 7 7 - i n c h ( casing """""">• — r—y. Open _^^ hole Present (1979) depth 595 feet Open-hole well bore filled with coal tar. / sand, and (or) other material Original (1917) depth 909 feet »*• 3 -Water ' 'entering r— ' well here —Waler leaving well here L » 4 i — r~r~' — r~> — i f 6 10 14 18 C Well Vc diameter, i in inches pe /Water level in welt Direction L^.^ lof flow Note: Flow was measured before installation of t emporary packer 1 ' I > ' i 60120 ilocity, i feet r minute Figure 74. Hydrogeologic and geophysical logs of Well W23 ("Hinckley" well on the site). Source: Hult and Schoenberg. 1984. 325
------- overlying the Platteville Limestone are (1) a unit of till, outwash, valley- fill deposits, and deeply weathered bedrock; (2) a middle unit of glacial sand and gravel called the Middle Drift aquifer; and (3) an uppermost unit of lake deposits and till. Below the Glenwood confining bed lies the St. Peter aquifer, the Basal St. Peter confining bed, the Prairie du Chien-Jordan aquifer, the St. Lawrence-Franconia confining bed, the Ironton-Galesville aquifer, the Eau Claire confining bed, and the Mount Simon-Hinckley aquifer. The movement of the groundwater and, consequently, contaminants over the 50 years of plant operation has most probably varied with time because of a number of factors. A major control in groundwater movement is the draw-down created by water demand in communities as they have grown and diminished in population. The continuity of confining beds plays an important role in that a conduit for water and contaminant exchange between aquifers occurs where confining beds have been eroded. The presence of glacial valleys filled with coarse-grained deposits may provide preferential pathways for movement of groundwater or contaminants. Also, multiaquifer wells (wells hydraulically connecting two or more aquifers) provide an avenue of transport for contami- nants and water, and they can locally change potentiometric surfaces of con- necting aquifers. Multiaquifer wells result from original open-hole construc- tion, leaks in casing, or flow in annular space between casing and borehole. In the St. Louis Park area, Hult and Schoenberg (1984) found that the water level in each aquifer is higher than the level in the underlying aquifers, causing water flow through multiaquifer wells to be downward. The major contaminant from the Reilly plant was creosote, a complex mix- ture of chemical compounds. Typically, creosote contains 85 percent PAH [i.e., naphthalene, anthracene, phenanthrene; some of which are carcinogenic (at least 12 have been identified as carcinogenic, U.S. EPA, 1980a)] and 2 to 17 percent phenolics (i.e., phenol, methylated phenols). The remaining con- tents consist of various nitrogen- and sulfur-containing heterocyclic com- pounds (U.S. Forest Products Laboratory, 1974). In addition to creosote, the Reilly plant discharged approximately 80,000 gallons of 70 percent NaOH into ponds from 1940 to 1943, as well as some sul- furic acids. [For more detail, see Table 4 in Hult and Schoenberg (1984).] 326
------- The distinction between transport processes of most natural constituents of groundwater and transport of coal tar is that many compounds of coal tar are relatively insoluble (Sutton and Calder, 1975; Schwarz, 1977). PAH's tend to adsorb strongly to soil particles and have low aqueous solubilities (Hickok et al., 1982). Phenolic compounds are generally more soluble in water than PAH's. The solubility of phenol is more than 10 g/L at 25 °C and pH 7.0, while the solubility of naphthalene under the same conditions is only 0.032 g/L (May et al., 1978). Solubility behavior of hydrocarbons is poorly under- stood. In Hult and Schoenberg (1984), dissolved constituents are defined as those not removed by filtration through a 0.45-micrometer filter. Many coal- tar derivatives are non-ionic and may exist as microscopic aggregates of individual monomers known as micelles. Micelles are considered part of the aqueous phase, and their movement is controlled by critical pore size. Micelles may move as though they were ideal solutes or become trapped, forming a hydrocarbon fluid phase at some distance from the source. This complicates contaminant movement and explains the wide variation of contaminant concen- tration throughout the area. When creosote is mixed with water, two phases generally emerge: a light- er aqueous phase enriched in phenolics and a more dense hydrocarbon phase enriched in PAH's. Because the second phase has different properties (i.e., density and viscosity) from the aqueous phase, the hydrocarbons may move at a different rate and in a different direction than does the groundwater. At St. Louis Park, the dense hydrocarbon phase has percolated downward relative to the direction of groundwater flow, allowing contaminants to dissolve in the flowing groundwater and to be transported downgradient. The major transport mechanism is in the aqueous phase, whether as solutes or as micelles (Hult and Schoenberg, 1984). There are three major paths for contaminant transport. The first is by spill or drippings onsite, which infiltrated and percolated through the unsat- urated zone to the water table. This has resulted in extensive contamination of the unsaturated zone on the 80-acre Reilly site. The contaminants reaching the groundwater vary in composition from area to area because the coal tar used throughout the plant's operation came from different suppliers and 327
------- subaereal decomposition of the coal-far constituents produced degradation products dissimilar to those produced in the saturated zone. The second path for contaminant transport is surface runoff and plant process-water discharge to depressions and wetlands found south of the plant site. Natural surface drainage was toward the site and south to Minnehaha Creek. Since approximately 1938, the drainage has been disrupted by roads and other manmade structures. Therefore, surface runoff and plant process-water were discharged through ditches and culverts to water table ponds near Well W13 (see Figure 73). If the rate of discharge becomes greater than the rate of evaporation, mounding in the water table occurs and vertical movement of the contaminated water and hydrocarbon-fluid phase into the underlying, confined drift aquifers occurs. Visible contamination extends at least 50 feet below the water table south of the plant site near Well W13 (Minnesota Department of Health, 1974; Barr Engineering Co., 1976). Since approximately 1938, surface water inflow to the ponds recharged to underlying peat and the Middle Drift aquifer. Inflow included 30 to 60 gpm of wastewater (Minnesota Department of Health, 1938) and as much as several hundred gpm of runoff during peak periods, increasing the vertical leakage. Also included in the plant discharge were sodium hydroxide and sulfun'c acid occasionally used in plant processing. The third path for contaminant transport is movement of coal tar directly into bedrock aquifers through one or more deep wells onsite. The main pathway is through the 909-foot deep Well W23, drilled in 1917. At some time, a coal- tar spill into this well occurred and is probably the source of early contami- nation reported in the Prairie du Chien-Jordan aquifer. The well was tempo- rarily plugged and is now 595-feet deep. An unsuccessful removal of the vis- cous material was attempted in 1958. 3.3.8.3 Site Remediation-- In 1980, the available data were studied to assess the feasibility of (1) controlling movement of contaminated groundwater by pumping wells, (2) excavating or otherwise remedying contaminated soils, and (3) treating and disposing the residual waste products. A system of 12 to 15 wells in 5 to 6 aquifers was designed to flush the groundwater system. Hickok et al. (1982) estimated that the contaminated areas could be flushed in a few decades with 328
------- minimal sorption effects. However, leakage from the overlying drift, and especially from the "source zone," could continue to cause significant contam- ination of the bedrock aquifers for thousands of years, even with gradient control wells. Ideally, management of the "source zone" would include excavating the highly contaminated surficial peat, removing the associated fluid, and pumping out the body of hydrocarbon fluid generally underlying the peat in the Middle Drift aquifer. Hickok et al. (1982) surmised that, at the time of their study, too little information on the actual contaminant distribution was available to design a complete remedial program for the "source zone." As far as disposal of the "source" material, Hickok et al . (1982) con- cluded that the hydrocarbon fluid could not feasibly be treated for discharge to the Mississippi River or other surface waters. They concluded disposal would probably entail transport by truck or rail tank car to a secure land- fill, a reprocessing plant, or another option depending on the total volume of hydrocarbon fluid. The disposal of the peat-associated fluid probably would be similar. In a subsequent study, Harris and Hansel (1983) completed an evaluation of groundwater treatment and potable water supply alternatives for the City of St. Louis Park. As part of this study were bench-scale tests conducted to determine the efficiency of various water-treatment technologies in removing PAH's and other coal-tar derivatives from groundwater. Of all the technol- ogies tested, only tr-ee were shown to be effective in removing PAH compounds to below the treatment goal of 280 ng/L total "other" PAH compounds. These three technologies were: granular-activated carbon (GAC) , ozone/ultraviolet (03/UV), and hydrogen peroxide/ultraviolet (H202/UV). At raw-water concentra- tions of about 7,000 ng/L, GAC appears to be the most cost-effective, and a GAC pilot plant was set up and successfully operated in the pump station at one of St. Louis Park's contaminated wells. These three technologies achieved compliance with project-specific treatment goals and provided effluent water quality adequate for use in a potable water distribution system. Phenolic compounds and naphthalene are disappearing downgradient from source points (i.e., Wells W13 and W23) faster than expected if only dilution were occurring. A study by Ehrlich et al. (1982) concludes that phenolic 329
------- compounds in groundwater are being converted to methane and carbon dioxide by anaerobic bacteria. Naphthalene also shows an attenuation in concentration, but this appears to be due to sorption rather than biodegradation. Ehrlich et al. (1982) believe that the contaminated drift is acting as a treatment zone for removal of phenolic compounds that have penetrated the aquifer. They characterize this zone as a continuous flow bioreactor consisting of a fixed- film microbial population fed by a multiple nutrient stream as envisioned by Rittmann et al. (1980). To date, a portion of the surface contamination has been removed and infilled with clean topsoil. The State of Minnesota is planning to build a highway interchange that would cover an area of contamination that has not yet been removed. If the State builds the interchange, the construction plans will include removal of the contaminated soils. If the interchange is not built, the Reilly Tar and Chemical Corporation is responsible for this surface contamination removal. Upon approval by all parties involved, a remedial action plan will go into effect. The plan includes a gradient-control well pumping system, a GAC filtering system, repair of leaking multiaquifer wells, removal of coal tar from any contaminated wells (in particular W23), establishment of source control wells, and monitoring of all contaminated aquifers over a set period of time. The entire remedial action plan has not been completed and is still being drafted. The Minnesota Pollution Control Agency is coordinating the remedial action planning. 3.3.9 Pensacola, Florida (Ehrlich et al., 1982; Franks et al., 1985; Mattraw and Franks, 1984; McCarty et al., 1984; Troutman et al., 1984; Wilson and McNabb, 1983) American Creosote Works Inc., an abandoned wood-treatment plant near Pensacola, Florida, was chosen by the U.S. Geological Survey in 1983 as a field laboratory to study the transport and environmental fate of creosote constituents in groundwater and surface water. Also, the site was chosen as being appropriate to apply the latest techniques for characterizing hazardous waste problems. To quote the National Priority List (NPL) description: The American Creosote Works, Inc., Site covers 1.5 acres in Pensa- cola, Florida, about 0.3 miles north of where Bayou Chico and Pensa- cola Bay meet. The facility treated wood with creosote and penta- 330
------- chlorophenol (PCP) from the early 1900s to late 1981 or early 1982. PCP-contaminated waste water was discharged into two ur.lined 80,000- gallon percolation ponds. In February, 1981, the U.S. Geological Survey identified phenols in ground water associated with American Creosote Works. At present, no drinking supply wells are within the known zone of contamination. This case study illustrates the following: • Contamination of a sand-and-gravel aquifer from direct contact with creosote waste • Insignificant attention of contaminants through sorption onto aquifer materials • Anaerobic degradation of phenolic compounds in the groundwater environment • Degradation of quinative to 2-quinolinone in groundwater by microbial oxidation • Utilization of novel onsite groundwater sampling and analysis method to map the extent of microbes responsible for contami- nant degradation, and by reference, the extent of contamination (Report is a selective summary of the USGS findings and is entirely based on the three referenced documents) 3.3.9.1 Site History-- The wood-treatment facility located within Pensacola, Florida, had been in operation from 1902 to 1981. Over this time, wood-preserving chemicals were discharged into two, unlined surface impoundments. Prior to dewatering and capping in 1982, the impoundment wastewaterb were in direct hydraulic contact with an underlying sand-and-gravel aquifer. The aquifer was up to about 300-feet thick and~consisted of deltaic, fine-to-coarse quartz sand deposits interbedded with locally confining, discontinuous clays and silts (Troutman et al., 1984). The impoundment wastes, in general, consisted of the wood preservative creosote, a coal-tar derivative. In addition to creosote, diesel fuel and pentachlorophenol (PCP) were discharged to the surface waste impoundments. 3.3.9.2 Methods of Investigation-- 3.3.9.2.1 Soils and qroundwater samplinq--Nine test borings were drilled in 1981 to investigate the hydrostratigraphy beneath the site and to survey 331
------- groundwater quality close to the facility. Borings were later (1983) completed and developed as groundwater monitoring wells. At each boring site, a well cluster of two to five wells was constructed with each well set at dif- ferent depths. Details of well construction and materials, sampling proto- cols, and the results of groundwater sampling for creosote constituents and PCP's are given in Troutman et al. (1984) and Mattraw and Franks (1984). 3.3.9.2.2 Microbiological investigations--The aerobic degradation of quinoline in soils derived from the site was evaluated by standard laboratory batch techniques. The anaerobic degradation of phenolic compounds was also studied using enriched bacterial cultures from contaminated groundwaters at the facility (Mattraw and Franks, 1984). 3.3.9.2.3 Experimental/innovative investigative techniques--The research site was used to test the practicability of several experimental, nonconventional groundwater sampling methods:. • A multilevel "bundle" piezometer for sampling groundwater and measuring hydraulic heads at discrete vertical intervals within an aquifer (Mattraw and Franks, 1984) • A reconnaisance groundwater sampling method, whereby ground- water within the hollow-stem auger is sampled and analyzed by an onsite high-performance liquid chromatograph (HPLC) for dissolved methane (Troutman et al., 1984; Franks et al., 1985). 3.3.9.3 Extent of Contamination Findings— r Results of the 1983 groundwater analyses by gas chromatography/mass spec- it troscopy (GC/MS) indicate the presence of approximately 80 organic contami- nants in groundwaters p?ar r.he facility. For classification purposes, three compound groups were iaentified: phenols (up to 2 ppm); PAH's (up to 2 ppm); ; and heterocyclic compounds containing oxygen, nitrogen, or sulfur (up to 1.5 ; ppm). Based on these general groupings, two contaminant zones were observed I at the waste site: :: • A highly contaminated water-table aquifer plume to approxi- ' ; mately 36 feet depth i • A relatively less contaminated, confined, or semiconfined aquifer plume extending to a maximum depth of 75 feet. 332
------- Questions concerning the transport of pure creosote within the unsatu- rated zone and within the aquifer were not directly addressed in any report. However, pools of denser-than-water, black, oily material were reported to be ™ seeping from a stream approximately 450 feet downgradient of the waste impoundment (Mattraw and Franks, 1984). PCP was not observed to be present in groundwater downgradient of the waste site at concentrations greater than 0.01 ppm. Vertical distributions of contaminants at well clusters near the impound- ments and approximately 450 feet downgradient show that contaminants have, in general, moved en masse (though in a dissolved state) with little or no "chromatographic separation" of compounds because of their differential reten- tion on the aquifer media. Based on these observations, the reports conclude that retardation of organics because of sorption on aquifer materials and soil organic matter provides little or no control of contaminant transport at the site. This is not surprising considering that aquifer materials are predomi- nantly clean sands, with minimal clays and organic matter. Individual contaminants such as phenols do, however, decrease in concen- tration downgradient, presumably because of microbial degradations. Phenol biodegradation under anaerobic aquifer conditions is well established (Ehrlich M et al.f 1982; Wilson and McNabb, 1983; McCarty et al., 1984), and results at the Pensacola creosote site replicate these findings specifically. Godsy and Goerlitz (Mattraw and Franks, 1984, pp. 77-84), found a sequential disappear- ance of C-* through C^ carboxylic acids, phenol and benzoic acid, 3- and 4-methyl phenol , and 2-methylphenol "during downgradient movement within the aquifer." In laboratory digesters containing enriched bacterial cultures from contaminated groundwaters at the site, the same sequential disappearance was observed with concomitant methane and carbon dioxide production. The extent of the dissolved methane plume, and thus the extent of methane-generating bacteria and their degradation products, was later addressed in 1985 using an innovative drill-stem groundwater sampling techni- que and an onsite HPLC analysis (Franks et al . , 1985). These findings indi- cate a much wider distribution of methane in the aquifer and that some of the byproducts of microbial degradation may have migrated farther in the aquifer than did the more readily degraded organic contaminants. Thus, selected 333
------- contaminant plumes may extend well beyond the trace of the specific target (or "indicator") compounds (e.g., total phenols) if lower molecular weight organic and inorganic byproducts of the target compounds are considered. No evidence was presented for the anaerobic microbial degradation of PAH's or heterocyclics, nor were any studies undertaken to examine the aerobic microbial degradation of any compound except quinoline. In one study by Bennett et al. (Mattraw and Franks, 1984, pp. 33-42), groundwater samples were collected and found to contain appreciable amounts of 2-quinolinone, a princi- pal aerobic degradation product of quinoline. Subsequent soil samples and surface water and groundwater samples were found to contain large numbers of aerobic bacteria that convert quinoline to 2-quinolinone. These organisms were identified and counted. 3.3.9.4 Site Remediation (as of July 1983)-- According to the NPL description: In March, 1982, American Creosote sold all the equipment onsite and later filed for bankruptcy under Chapter 11 of the Federal Bank- ruptcy Act. The state has negotiated a Consent Order requiring American Creosote to restore the discharge areas and install onsite monitoring wells. The company constructed higher berms around the ponds to prevent overflow during heavy rainfall. EPA recently completed a remedial plan outlining the investigations needed to determine the full extent of cleanup required at the site. EPA plans to fund (1) a $290,000 remedial investigation/feasibility study to determine the type and extent of contamination at the site and identify alternatives for remedial action and (2) an $85,000 initial remedial measure involving fencing the site, posting warning signs, reconstructing the berms, and controlling flooding from the waste ponds. The work is scheduled to start in the third quarter of 1986. 3.4 CONCLUSIONS Each of the gas sites visited showed surface contamination by tars, ash, and other wastes associated with gas manufacture. The amount of visible con- tamination varied from site to site, but it appeared more widespread at the larger sites. Blue ferrocyanide contamination was visible at the Mendon Road, Taunton, and Pawtucket sites. Each of these sites was known to produce gas by coal carbonization. Spent oxides were discovered at the Spencer and Richmond 334
------- plants. This spent oxide showed signs of sulfur and iron, but no ferro- cyanides. Both of these were principally water-gas plants. Some oil contamination of the water in the Pawtucket canal (in Lowell, Massachusetts) was visible. This contamination was from the general direction of the gas plant. No other oil contamination of surface waters was seen at the other former gas sites. Substantial gas odors were noted at the Lowell, Richmond, Taunton, Paw- tucket, and Mendon Road sites. The odors indicate that contamination may be substantial at these sites. Only slight odors were noted at the Spencer and Worchester sites. The plant at Spencer was very small, and the Worchester site was capped with construction refuse and soil. The case studies indicate that sites are "discovered" when (1) surface water is contaminated, (2) construction activities disturb the site or ground around the site, (3) redevelopment of the site is attempted, or (4) municipal groundwater sources are contaminated. Phenol and PAH compounds appear to degrade in the groundwater when they are present in dilute concentrations. In raw tars, however, the microorgan- isms cannot survive, and the tar components do not degrade. This means that tars can remain substantially unchanged over time. Tars (heavier than water) sink within groundwater systems until stopped by low permeability strata. Oils can float and spread on the surface of groundwater, contaminating a band of soil and thereby serving as a source of contamination to underlying groundwater. Cases of significant groundwater contamination usually can be attributed to the lighter, more soluble aromatics found in oiIs. Local pumping of groundwater wells can affect the flow and transport of tars and contaminated water. Controlled pumping can be used to limit the spread of groundwater contamination. Much of the historical data reported about the Stroudsburg site appears to be incorrect. The "coal tar" at Stroudsburg actually appears to be a tar from the production of carbureted water gas. The density of the tar is very close to water, which later separated. The low carbon content and absence of high-boiling organics imply that the tar was condensed after the washbox removed the higher boiling organics. The lack of phenols and the low 335
------- • content of this tar identify it as a water-gas tar. The existence of an injection well for tar disposal also has been questioned because the term "tar well" was frequently used to describe underground tar storage tanks. None of the case studies examined a plant that produced gas only by coal carbonization. Possible explanations for this include the fact that the coal carbonization plants produce tars that are not as prone to tar migration, 1t may only indicate the widespread adoption of the carbureted water-gas process, or coincidence. Coal carbonization tars were generally more dense and more viscous than carbureted water-gas and oil-gas tars. Tar viscosity decreases with temperature, and surface tars generally become more mobile during the summer months. The principal remediation employed at town gas sites is containment. Slurry walls, caps, and collection wells have been used. Site contamination differs with the processes employed for gas manufac- ture. The principal contamination at the Seattle plant was lampblack, which was produced in substantial amounts by oil-gas production. At carbureted water-gas plants, the principal contaminant of concern was relatively M,ob1le tar. The waste disposal practices at the sites examined were generally quite poor. Although tars were frequently recovered, the liquids that disposed were either placed into the nearest body of water or, if they could not be disposed into water, placed into lagoons, trenches, or allowed to flow across the soil until absorbed. Solid wastes either were used to fill in areas along the shoreline or piled in a dump beside the plant. 336
------- 4,0 STATE STATUS CF MANUFACTURED-GAS SITES 4.1 INTRODUCTION This portion of the project was undertaken to determine the current sta- tus of manufactured-gas sites on a national basis. Originally, this determi- nation was to be made by comparing the Radian list of manufactured-gas sites (compiled from Brown's Directory of American Gas Companies) to the national Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) CERCLIS list of sites (reported to EPA by the individual States). Such a com- parison would have produced a list of manufactured-gas sites that Individual States viewed as sufficiently hazardous for inclusion in CERCLIS. The resulting list could then be used to assist in planning further EPA efforts in the area. The organization and nature of the information on the two lists prevented approaching the problem as planned, and an alternative approach was used to determine the status of manufactured-gas sites within States. Each EPA region was contacted to identify which States had placed manufactured-gas sites on CERCLIS and to determine what the status of the sites was. For most regions, the persons responsible for placing State sites on CERCLIS within individual States had to be consulted. Section 4.2 explains why the originally planned list comparison was impractical, and Section 4.3 describes the information acquired on the status of gas sites within States. Section 4.4 discusses the Radian list of manufactured-gas sites. 4.2 COMPARISON OF THE RADIAN LIST AND CERCLIS The original task of comparing the Radian list and CERCLIS of manu- factured-gas sites proved infeasible because the data included in each were incompatible. Figure 75 illustrates the type of data contained in the Radian list of town gas manufacturing sites. These data were compiled from Brown's Directory at 10-year intervals between 1890 and 1950. The information 337
------- f T I 2 1 ' s! 3 : 4 • •- 1 m Jl * 33 Wl — Jl _ c. w n 3 >• — H 5 •» : i 3| •n 4 d^ » wt «t ? ii] J 1 Sji»»«8 S s = " ii I si 1 g - 5 «* = fe i 3t S z u -5 jj K S "" 9 ^ N j s!i 1 S3 siuvl"u g : j jl^C'^i H 1 If 4 •* 4 — '1 ^ u i 7 i i 1 ^ 40 ^1 *^ IfJ «• • A. - *"-as:S2 , "•""•a^s- .3 -j i « >1 IU >M •*• — ---•«•« -| | UJ 2 " S Mk tu «\* M> 143 **•• a. „ 2 ? ' ' 3 s J | 3 w* , vt s i anufactured- E *•* in S 1 cc in 1 iZ ^ r\ 01 i i 338
------- reported Includes the city where the plant was located, company name, plant status, production, and byproducts. The only information recorded on plant location is the city name where the plant was situated. Figure 76 provides data from CERCLIS. These data indicate the EPA identification number, site name, address, county, latitude, and longitude for each site. The list includes no information on the type of contamination at the sites or on any operations at the site resulting in contamination. The site name of sites on CERCLIS can be used to determine if listed sites were former manufactured-gas sites, but only when the site is listed specifically as a gas plant or as owned by a gas company. Many of the sites in the list have names that do not indicate anything about the source of site contamination. Thus, merely compiling a list of the sites with site names that indicate they might be manufactured-gas sites would produce many omissions and inaccuracies. The only basis that could be used to compare the Radian list and CERCLIS would be to compare the cities on each list and produce a list of CERCLIS sites in cities that also had manufactured-gas sites. Table 62 shows the number of sites resulting from this approach for the State of Alabama. There * were 164 CERCLIS sites in cities that had manufactured-gas sites in the Radian list. The inability to match Radian and CERCLIS sites within cities made this type of comparison essentially worthless, so an alternative approach had to be found to examine the status of manufactured-gas sites in the States. 4.3 EXAMINATION OF MANUFACTURED-GAS SITE STATUS IN STATES As an alternative, individual EPA regions and States were contacted to collect information on "manufactured-gas sites within States. Table 63 lists the results of the inquiries and the current status of sites within each State. The information was collected from employees of either the EPA or State agencies who were "in a position to know" the status of CERCLIS waste sites within their areas. Consequently, the absence of known gas-manu- facturing sites on CERCLIS may either indicate that there are actually none on the list for that State, or merely that the individuals contacted were not aware of any. Table 63 summarizes the information collected from regions and States on the status of manufactured-gas sites. Tables 63 through 72 list the sites ^ 339
------- L.I - SITE LOCATION LISTING 06/07/1985 EPA 10 SITE NAME STREET CITT COUNTY CODE COUNTY NAME ZIP CODE LATITUDE LONGITUDE SMSA HTORO UNIT PHATTVILLE AUGUSTA 001 AL0980710370 CALLAHAN PROPERTY MHY 82 ROUTE 4 BOX 2 66 PRATTVILLE 001 AUTAUGA 36067 36067 322748.0 862830.0 5240 322748.0 862830.0 5240 AL09S0556245 SOUTHERN RAILWAY DERAILMENT SITE HP 178.9 FREEHONT 001 AUTAUGA 36784 315442.0 674424.0 ALOOOS557004 UNIOII CAflP COSP MONTGOMERY MILL SITE JENSEN RO PRATTVILLE C01 AUTAUGA 36067 ALD980495667 BALDWIN COUNTY LANDFILL PO BOX 150 BAY NINETTE 003 BALDWIN 36507 ALD980495709 BAY MINETTE CITY DUMP M 7TH ST BAY HINETTE 003 BALDWIN 36507 ALD940727929 BOLON PROPERTY RABUH RO BAY fllNETTE 003 BALDWIN 36507 AL0980727747 BRANTLEY E R NEWPORT PARKWAY BAY HINETTE 003 BALDWIN 36507 ALD000652941 051 TRANSPORTS INC HWY 47 N BAY HINNETTE 003 BALDWIN 36507 AL0001874254 KAISER ALUMINUM 1 CHEMICAL CORP HWY 31 S BAY niHETTE 003 L4RPT1 - PREPARED BY OPfl 305300.0 874624.0 5160 305300.0 874624.0 5160 3150201 3150201 3150203 322520.0 862820.0 5240 3150201 305300.0 874624.0 5160 3140106 3CSJOO.O 874624.0 5160 3140106 3140106 305300.0 874624.0 5160 3140106 3140106 Figure 76. CERCLIS waste sites. 340
------- r TABLE 62. COMPARISON OF RADIAN TOWN GAS SITES TO CERCLIS FOR ALABAMA County Barbour Calhoun Colbert Dallas Etowah Jefferson Jefferson Lauderdale Madison Mobile Montgomery Morgan Talladega Tuscaloosa City Eufaula Ann is ton Sheffield Selma Gasden Bessemer Birmingham Florence Huntsville Mobile Montgomery Decatur Talladega Tuscaloosa Number of CERCLA sites in city 1 13 6 10 6 5 34 4 16 29 18 14 1 7 Total 164 341
------- V) LU H* tn z X H IH Ul l-( ^) to 5 S u J 1 u. o £ f- • m LJ 2 H n 4> C | U o z M M C -1 o M n c O 0 •o u II c o o f 4) • 4> J, 4> c u £ L F X O» c 'i L, C •o ffi c L 0 m U • fl "O "• «tf • n M ••" al o u """ « A > « C O > z ? • Jt X M ••t ^ C • c c z z « M • • >• > W C M O C N • |^ ^ | U 4» n — c ic • *- c c c a •- o o w u c u c C 4) > 4> • 4> . jj .- J ._ to u 13 x n "a w *H « C 4> C _i c 4» c jc a c VJ C U 3 O O X oe o c o • u 4> ui o c u— o» O C JO V c o c a o o c u u — E Z o c u • x > C •- C c L a • « U X C • T> X — O 4> 0 • HI — 4> -T> 4> •o — c c -o — •1 • C M O E • •> -e • .- § 4> n • ui— L. L. ill • — • » — o » 0%. « o — • 0.0 MA C A A A "™ W U» B •• V BL C C < CM «- <• e • M • c o £ ^ *"" *— C •"• C • U C C 4* « L U O U J3 M O C O — H. O — X C*- (. H- «l — •1 J3— U C — X 1. * • C 4> • 0 c > u— o jc m c -o c — T X L -O C U 49 O c • c 1 o • O • 4> L *> . 4> 3 «l 3 6 X e— oiixo** Z Kl WO— M4>l- *" ' t M C > z M M >• > M •^ c o C 13 o • S- C o "• "o w w C 1 O "DC c — « 3.A *> C 0 (1 •- > ut- X-D •(/) a c c — • o * <4^ 0 W « « 4» H -4^ L *'"' 0 C N « — a 1. Q, • L > O Q. > O U O O CO • • A f V J* « « 4» U 4> C CUM 4> O 4> • U — O. W • « n c 41 c 3 U C 4» *» cr • c c — C 4> C E C « — • 0 M • o~ « c • 4> — u » a » x o a. • e « — *J n o — a c x— c C 4» — 1. 4> O u HI m » — c c e H- "5 "« 'i'i ° « — • c c — • T9 C U O — u ao a w UK 1. a L i. e C Jx 4> O « 0- 4> * IA Z 3Z 0 »<0 00 M 0 >• « « >• >- 4> 9 U 4> 1. U • c • c — o c w o i - '- U EC 4> 1 • 4* M E 4> • o c k - o o-o i. • u c c c L 3 > O X O O t- «l 1. U JC • • o>— • n > • ML •- H 0. C C • - u c x n a c c r™ E L X» 4> 41 «. a. n •- a^. N < n a. »u 4> u < 4> a. c • c u — c 4» co. •» c «. • *«^ > u U O X • 3 — ••(/) 4>— O -0 o — a. > c 4) M • 4> « 4> L 4»X C C-0 W M U C 4) C C — II 6 • II 4) — • « k — H -HI C 4> C C « — x » • o -o • U 4> U a: -o c 3 u o • OH C — •o t» — au. • C 4>-0 M n S • c c M C >• ^ • • k 0 n • 10 g. ^ x 3 C • c -a c c > > o o U i H C — L O n • S ••" C 01 0 u — 0 4> •'I o c A* A ^C O u — L • Jj 1- Ou <• c — M • HI O • <_) n o o • Z • M M 0 O C >• Z 5- * '« •- 0 u • x J> * • -o n 9 4J •— y| a 4J U o 0 •o O) c X 43 c u L. 3 U 4> C 9 n •- • o e c * •• •• — •« — c M M s*. •o 3 C 4> O u c .0 4> o TJ e c 4l * M A ^J 0 i 342
------- •a e 'J c o c m to in * e ( c 0 Z CO M w 0 ^ *» U — o: « UI M n c o o •o 3 o J c rt u 0 4J • *> a at c c — — * • 9 9 *» II J3 J3 C C -*li c — n « E c 0 -D n 0 « — « L n 0 0 k • — « 3 — C — « — n E— 0 k 0 — a. c 0 e k a a a. .- . — •o « o. • OS— 0 10 e — »-« c n — -l o a. 0 w 60) K « o ai 0 o k o *> o •- xw C M *J o — -o k — 0 0 0— C *> X *> C — *> 3 » (0 O — x a X J3 *> • X — -o -o — O 0 0 *> k C C C •— 0 o a E k « 0 • k A k X 3 2 a 0 u «•" « e * l J3 J> M M • gt V M ^ • « 1 e 0 • X Jt O u c e * 0X X *> *T3 O 0 k C o — x e M k H . C VI ^"\ ^ -0 0 — 0 -*> M A) •— •v n 9 O» X ^ Ik W > k 0 0 > « s- c — -o x C k • ^ ^ V • 0 *^ M * 0 0 0 U ••X 0 0 U C *» w M 0 0 a * * It O II — C »-o o o — O 0 *5 * O 0 • 41 V c c z 2 n « • M V 0 X C JC • o 3 W 4J e 3 0 0 X -1 £ •- 0 c • 0 k 0 k 09 0 C ffl k C * II a k C a s 0 c n •— •— 0 * X- 0 _l II 4^ • c-o 0 C J? II 4> 0 k M O « a. * c X — k • 0 e k 'I 0 4) L. " a. to « M 35 0 c a k o X • eo — 0 J- ™~ • • J"" L a « •^ t/> • IS o — u « 9 49 >x x o « V. 0 ~* 491V' — X II •1 0 *> 4) *. 3 .- X ** • T3 M C — C « — O — X c n k • *> a oe « o •0 0 4> M M — 3 *- t^ CM n 0 n 55 •o c 4 •• X k j o» -o c c — 1 • c n Q M ^ X— 0 k kX- II II 0 • 0 M U 0 k TJ a c CO 0 «•- X X O. 9 —!_ -o *» a — — c c - i) • 0 — *» a k g *» 1 0 •0 *> 3 — k a: «u 0 9 O k > c u a. ui 0 c • c • — J> o n 0- 0 « ^ k — *» 0 • — •-*> c •- « — 0- E M *» — M « C • *- «« a o -O 0 • 0"0 0 -o x c 0 E— o— k • 0 x k S a M k o a • a*> a x 04> *> o— an — k 0 3 M .- TO *> CX c • u « c 3 e 0 • o k k n /-. 0 O» 0 « TJ o a o > a 0 k 3 k M 0 JO — — 0 0 0 ii- ex _ *> 4> 4> 0 E 4> a • c * o *» *> *> 0 o o < v>v) 6_i e k k 3 u C ffl 0 n 3 U H 41 M * 1 n k M • 0 e i o « x H e * n *> • 1) O X4> 0 O « N U k k • * k O) O *» _K C C O O *> — « • — S- U M — 0) g a-o I 01 0 C x — c c — o .C 0 0 — • 0 *> — k e ax — 3 a c M a o * TJ - "" * 3 * • 0 a.,-, 0 CMS 9 -O 9 — t-l k C— X _| o O — — *» xo •»..-.- — k a: k 4> > — M 0 uj 0 • e > *» > u CL— 0 n •o 0 M — -o c w « k 0 — 0 o • e o *» k X 0 —09 O -O CE TJ « CO u c/-x e M U C • 0 0 • 0 • W — *> — (_> «l 0 0.0 • ax 0 IB x * X O 4» — *> 1- oS'ii k g — k o ii k a M 0 u e *> > *• -u o w c «o~.- O 0 (7 M k f X c *> *» *> E *> JK — -o e 0 k *" • t Q, ° e 0 • • * ^"2 *~ 0 n -o M « — « 1- C — N •*> 0 4> — 0 O II — X — U. E W • Ml- M c c c c c e e X J3 « « n n « « ii a • a c *> — — * • 0 » k • je o> M ii 3 e n k • o c • • c -O > .. .-..-.-000 a a a a 3 ± i t^\ iy 9 3 e e o ^•^ . 0 J 0 1 <*) n ffl ^5 o e 343
------- •o 0 3 e '^> c o en O ui i n e A u i » M '« S n <3 § u II c o 5 «. » n c o •o 0 H a £ w. a. a c Q X X 4) c u ^ 3 U n 4) c g n 0 M n x • u « e v .§ " 0 3 V" O a. •»- T y) *" > ^ L X M | X J • 0 1. c 0 0 *> 0—0 X M X *> > X « II 4> 4> X « « o — 0 U> O> X 0 "» "0x a •»- a E n — — x • •DOC u o 00 Q> •« a.*) •a < » C 0) • 4> n x « 4> m > — I. C tl 0 — ti t. • * e 0) 0 -0 • Z c n e 30 0 C *> > — x— 0 — n « -o 4» ( 0 C X *J II C M <•- U 1. I. ••* O 3 0 u. — o -> *.. g « e M M 0 0 X O 0 U » •— L X * t 1 I L 0 S) 0 (^ *> 0 XX .QK 0 0 •U 0 « «• • x n 33 c — 0 « u 1. 0 3 *J U « 0 S L. • n 3 J£ O 0 1. N * • Z c C > • ^ B^ 4> n u 0 • • *> c « M 4^ Hi- fi • M 000 *» *» H X * 0 M CO 4> P.. 0 0 M C C 0 0 O n C — • — 4> 0 O u je jt mm c u o o >- XX *> *> S Ik 1 2 £ * n 0 •« n 0 c •~* E x 0 L O 0 II U o 0 M 4> U O t- m O e 0 x x 4> e L c u 0 e * o o S o ^K •o 0 U 0 •« n c 0 u 0 L III 0 M «l n 0 J M «•» M* e •_ •o ae jt • H u u 0 •o — x a 0 e o >• 9 x JC 0 • 4} X O) •« 1 •o A L. 0 o a. u •»• ._ u a 0. x x 1 c 0 II 0 4) fl ^ c 4) c "a. 0 o ^ VI > • >• e o u o 1 M L — 3 U 9 «P • C •*" V) 0 « 0 4> a— u u M ax * H «l — c -a c — 3 « O > 0 t- — ^ ^ >% — OT « M C C • 0 • « a. 4> 0 »»>«•- •<• >^ o ^ M 4} CUM • 3 — .- o — • U a c o o*- 0 X "O t^ «> 0r» u §m 0 a— L 0X «K C « ah- ^ a « '«'0 2xJ 0 c2 o c o S n > >• il ._ c II X n c 0 a. 0 L e • • 0 V) 0 M C U -1 0 0 ^ > a: TJ o u c o n 4» C L — 003 X X 0 E t 3 •« • u c ••* • •o • -o e 0 0 • U 4> — • II w — M— TS • f 0 « I • •000-0 o a a; c x n ig 0; .- _ •o • n - CM 000 0-0-00 c — i. -o 0X00 •o oxx '> C »l 0 O 4J . L L. 4> 4> O-— 0 «. • 0 M 0 1 Z 0 > « 4) • n e •— X fl ••• • a O 0 0 *» K H 4> *M M 0 0 « c e 0 00 ,, C •o •- • C — 4> II O O — U Jt M U O 0' XX T) 4> 4> 0 33 js o o o: w in 9 •^ «l M II O) • t. o ^_ c n •M •o II K X 0 c 0 •o c i M fH 0 jj . n x • t X c£ 'e - — e "0 4» b n 0"* O *M 00 0 c c c Z° I Z° ? 2 £ ! N S S X ex • ^^ T3 0 3 C •« 4> e 0 u ^ • • o ^5 c 0 4> M 5 § (0 344
------- o> m 1 C S . a^- •- c a t. « E « 3 L u c a a — i- • OQ - u t - JxS •o o c 'i « 0. • « n 3 „ >- Z O « c — i " U <- « W o a c .c .a • o •»--• * "- c < « - a M o s I. • CT C U. c f • O o. JC +> ~ i H u L a OOD •O 0*73 c J-8S.2 c-o • n .o t. *» • *- a I • «_ c-*j->>T12 »° ->-t;c«c«? .5^, Slto^5 !» .» u t o*> e± feZ . . • *» z — D CL C - J 11 I. O 00 L -X L o I- 3 01 ••>. ,0 U 4» Jt l. s * «• J4.--g « 3 C • — o *> — »Jt C ? w — < s & .2 O) • • o 6 2 2 • «*• X O — c •o • e 4> O u « CO I. a u c J U 3 W , • O «l : t- ^ ai « ne *> O • a. o TJ — • O U u • • i ' If a w M P-H ^d — a « iu <_/ n «ss •o 4J u • 4> e o u n £ n I • I c O -o C g M w > M • w M a S i e s o +> m 4> CO 41 C § • (t c 'a > o> c ? -» I s 5 o '»
------- TABLE.' t.«. P,V. Site G i 1 ro y Hollister Monterey Sal inas San Luis Obispo Santa Cruz Watsonvf lie Bakersfield Chi co Chico Coalinga Colusa Fowler Fresno Fresno Grass Valley Grass Valley Lodi Madera Marysville Marys vi lie Merced Modesto Nevada City Oakdale Orovi 1 le Red Bluff Redding Sacramento :.A\ SITES IN ciC GAS AND No. 408-1 418-9 418-1 418-2 418-4 408-7 408-8 335-1 210-1 210-1A 325-8 212-1 325-2 325-3 325-3A 215-1 215-1A 316-6 325-4 212-2 212-2A 325-5 316-2 215-3 316-3 212-3 213-1 213-2 206-2 CALIFORNIA COMPILED BY ELECTRIC COMPANY County Santa Clara San Bern" to Monterey Monterey San Luis Obispo Santa Cruz Santa Cruz Kern Butte Butte Fresno Colusa Fresno Fresno Fresno Nevada Nevada San Joaquin Madera Yuba Yuba Merced Stanislaus Nevada Stanislaus Butte Tehena Shasta Sacramento (continued) 346
------- TABLE 64 (continued) Site Sacramento Selna Stockton Tracy Turlock Wi 1 lows Woodland Eureka Eureka Eureka Santa Rosa Santa Rosa Okiah Benicia Daly City Livermore Los Gatos Napa Napa Oakland Oakland Petaluma Pittsburg Redwood City San Francisco San Francisco San Francisco San Francisco San Francisco San Francisco No. 206-2A 325-6 316-4 316-7 316-5 210-2 206-3 119-1 119-1A 119-18 104-6 104-6A 104-B 104-1 508-2 601-1 408-3 104-3 104-3A 601-2 601-2A 104-4 601-3 508-1 502-1 502-1A 502-1B 502-1C 502-10 502-1E County Sacramento Fresno San Joaquin San Joaquin Stanislaus Glenn Yolo Humboldt Humboldt Humboldt Sonoma Sonoma Mendocino Solano San Mateo Alameda Santa Clara Napa Napa Alameda Alameda Sonoma Contra Costa San Mateo San Francisco San Francisco San Francisco San Francisco San Francisco San Francisco (continued) 347
------- TABLE 64 (continued) Site San Francisco San Francisco San Francisco San Francisco San Francisco San Fruncisco San Jose San Jose San Leandro San Rafael San Rafael Santa Clara St. Helena Vallejo Vallejo No. 502-1F 502-1G 502-1H 502-11 502-1J 502-1K 408-5 408-5A 601-4 104-5 104-5A 408-6 104-7 104-9 104-9A County San Francisco San Francisco San Francisco San Francisco San Francisco San Francisco Santa Clara Santa Clara Alameda Mar in Mar in Santa Clara Napa Solano Solano 348
------- TABLE 65. DELAWARE GAS SITES Dover Gas Light (DES7) Wilmington Coal Gas Co. Coal Gas Holder Site New Castle Gas Co. Smyrna Gas-Coke Co. Georgetown Gas Co. Lewes Gas Co. Sussex Gas Co. 349
------- . TABLE 66. FLORIDA GAS SITES Location Walkover Inspec- tion (PCAP)< Comments PER District Office NW District Pensacola (Municipal) Yes Tallahassee (Municipal) Yes NE District Jacksonville (Peoples/ Yes Container Corp.) Gainesville (Gainesville No Gas Co./Poole Roofing Co.) No No Yes No No visible problem. No visible problem; known as Cascades Park. Coal tar present onsite, CAP'S being prepared. Palatka (Municipal) St. Augustine (Municipal) SW District Tampa (Peoples) Lakeland (Peoples) St. Petersburg (Peoples; site owned by City) Bradenton (Southern Co.) Clearwater (Municipal) Winter Haven (Central Florida Gas; No No No No No No No No No No No No No No No No Location not known. Location not known. Coal tar was shipped offsite. Field and parking lot. Coal tar may have been barged offsite; stadium constructed onsite. Coal tar sold and decomposed by bacteria. Now a parking lot. Adjacent to lake. (continued) 350
------- TABLE 66 (continued) Location Walkover inspec- tion (PCAP)' Comments St. Johns River District Orlando (Peoples) Sanford (FL Public Utilities) Ocala (Gulf Natural Gas Corp.) Deland (FL Public Utilities) Daytona Power & Light South Florida District Key West Ft. Myers (Municipal) SE Florida District Miami (Peoples) Ft. Lauderdale (Peoples) Miami Beach (Peoples) West Palm Beach (FL Public Utilities) No No No No No No Yes No Yes Yes No No Yes Yes No Office and parking lot. Up for sale. No No No No No No Location not known. No No visible problem. Soil and groundwater sampling by ERM; no visible problem, low concentrations of coal tar constituents in ground- water. CAP has been prepared, but not approved by DER and DERM. Office and parking lot aPCAP = Preliminary contamination assessment plan. 351
------- TABLE 67. MARYLAND GAS SITES Annapolis Plant (MD141) Bayard Station (MD166) Canton Station (MD159) Spring Garden Station (MD145) First Plant (M0147) Second Plant (MD148) Scots St. Station (MD191) Cranberry Run Substation Westminster Plant (MD146) Cambridge Town Gas (MD165) Fredrick Town Gas (MD164) De Grace Town Gas (MD162) Salisbury Town Gas (MD163) Cumberland Gas Light (MD190) Frostbury Gas Light Elkton Gas Light Chesterton Gas Light Hyattsville Gas & Electric Crisfield Gas and Light Easton Gas and Light Hagerstown Gas and Electric 352
------- • e ~ o IfZ k *> 3 U a e k k • • o u • *» a x— o i c u CM 9 a. «J e • • > k c -a e o i. •- — g -J e • a o • •- .- -a .j *> e « — 3 • •^ O > •a <- e < at — u e u o < v c « .U •> o — \n • -o • 3 •o a e • • • O 41 < — — — TJ a *• « c E — «. 3 • •o ^ o « •001. •f a c « o • 5-5 r L a. — e • a u — 3 a (E X e *> : : i i < V) e o a *> e at c k .o L. •< e e .-*>.- •D •« r — «> e 5 ^ 13 a. 2 J .- — c a — o • > o : ^ i S 353
------- V) H t/> V> 111 M X "1 * z . at 4» n 4» a a I. C C — .0 jC 4» — E M * — Jj*z£ f.' OB oi s» *"* • x e * | x x • o w w ~ M 0 •91 U *> U UJ 0 u > i. u — J3 « ^M • 0 u > 0 C x u u 4> .0 o • . c U 4> n 4) *> <• • c u c c M x o A -- • al 4» 4> — a • .e < < a. tux O i-l N K1 V) . •< •T U> W ^ I. O 3: • 4» 0 C 3 a! • « •< c 4> c • *P j3 i 0 II X U r-04 c: • o — *> — • — at > M n I. L JJ *> o o c e **_« "•^^^ je v. -x o 0 • V) u *» • e M M 0 e • e M B O U C <-> O X X n ^r u> «3 0 > o o t- o o o ja o A J «l « M • n M Jl U n o jc 3s i. O 4» 4> JC • • u *> «.— u « • I/) W O O 0 c • o e • •a cui JC fli O> ^3 •*• ^3 M xxo* ^i QQ Q| Q) rt _ A .4) C u X M 0 •" > X e £ u « o II a *» « jJ a 6 ?3 •0-HI--OC* «l « 3 • C — 1-4 oaaoa.i/)a, U N JC U C O 4> Jc i. o • S c u o S — it * jt mo. 4> U HO — c » o-o • X M § O • 4>) •** (Q t* C !• c e • — 4> J 2 S a- S ^^ w w ^ to to X 4> e a 0 i. x a u x u n jt Jt O i .1 • L C o o k* w QI o o c"! x o o < e c c » 4> O O 3 • l_ _f* *-* • ' _ft !• J» *• *•• «D Ol • 4) C C 13 'TJ ^ 3 9 O O — — O 1. «. O ea VIKK 3: f- 00 01 S r* (M « c f C9 L h X 4> C « • e x— • T) > -o a. u • u k o — o — e L. a oo OK o X4> o (-• c k * i. c *> >— a • c. o • • u o aa>o o _i |H CM «) T IO sJils! o f» co o> s «i r< T* 0 je o M V 9 M a e o o O VI 1C UO O 1 • • • i • -5 c c « X 0 • U 1. • • £. a. o o 4^ C fl ^S JC — U JC 4> 4> 4) M • 0 HI * S • 3 • - * XX • u M O X o e e -c .c -a o > 4> 4> E 4) o • •< a 4>. ja ja x c e j; • • f. • — o 4> M N 4> * E 4> • uiiuia.KiZ z — rl CM HI T U) « UJ • o 49 C _ a e o 41 ^ U •*• II at .. o u i u o ^ o II u i 4> • • L a. • X c a E O u 0 X ^ *^ c u ft 0 o •o e c o t. 4> 3 A n 5 e k o 4> j; e >o M M* IU o X u o „ u u 5 e u b M 5 e u o 5 *M I/) JA j n • g| M •^ O u 354
------- TABLE 70. NEW YORK STATE GAS SITES New York State Electric and Gas Oneonta site Mechanicville sites (2) Pittsburgh site Cayuga Inlet site Cortland-Homer site Ithaca-Court Street site Ithaca-First Street site Elnrira site Geneva site Niagara-Mohawk Power Corporation South Glens Falls site Glens Falls site Gloversville site Saratoga site Harbor Point site Rochester Gas and Electric Lower Falls site 355
------- i o w 5 £ X c 3 O »-> X (/> - UJ W t- Hl t/1 5 *£ M «( ^ _l ^B (/) 2 2 UJ 0. • »"< 4^ "• S UJ U j £ t- 1 z 5 U 05 f- 8 ^ 00 CD CD tt to h> ID O C& O CD CO 03 Ol O) 01 SO O < < ft O_ ft ^ Ik U* • en i. E .- — 3 £ n — 0 — O _J CO O c £ O « -* 4> C U co- co* — 4» I. — — 0 < < CD . w 0 4) > M «< • C TJ 4> O O V) O C * J£ D — « 0 O 41 •o c \u jc. O 4> USA (/) r* * c * o c _ ^ 1 ^J _ 0. 0 0) HI U <*- II 3 O I. t» C i* « * 004 a n L n • o c o ^J O o c M 4» C § < o! w ?! S 03 S S W OD 1 1 c o 4> a "a 2 — 4l X U — 1 Z § i v v O O _ _ •£ .C "• 4! 03 CD • •o at 0 w •» L o -o f I. 0 > -X «-• o n ^ a. E 0 L. X « 0 E 4> 0 0 £. 0 0 X 0 — ca a. a oi u t- (^ 5; § S u a •o o 0 U — HI S S O> CO i I 1 1 • "i » 3 -X O 0 u oa Ol U C X O E L 0 X CD 03 • « 4> (O X n 4> • x «| 4> C X 0 *> > X 0 4, (L IA II O O) u 3 X in £ Q 0 ca — a. o ^J O) «*. n* j d • L a a. o c C M JO gj SiO oi g|A s CO 00 1 1 • •o c c C II s ~ JC X II 3 J O f — •o — o U L. • •o K. je u 0 > 4} 4 O X x en 4> — x x at 4> a. 4> i/) 4> 00 — a. o U O) 3 4> C V) 4> n i. 0 X O 4> 3 4 0 _! C n 4> 0 OL "» « — •• i. o • u — 1 n M. ^ L a. • u n <». O) • L a QL 0 c w e M 0. « 00 0> s a CD 00 1 S L. 4> "§) S •~ u •C C 0 A -J — 1 • 7 4 3 •- H E 4> — • II W 4> 43 (/) M 4) «M 3 — c a a 4 4> 4 C O 4> 1. c u. o L II 3 er 3 M M 5 n u i 4) 4) .. •. a. a. a a •»- •*- M M <3 <3 a a ^ ^ 0 0 U VJ HI HI n co N f> S r- i 1 1 1 X 1. 01 4) ^ C L. O O 2 V L. ^ e M 0 O Jt X U 0 o c X HO. X K c M n o • n e o ff ^ (0 T3 c •o a ce — w U HI 0 > t. *"™ 9 K C M 3 V U T CD M C C c 3 ^ CO 1 o u 4> X Ol IB 0 -I 4 0 JC U .3 $ o 0 a. g o 4) II M C 1 1 3 « •V CO CS CO n m 00 M S O 01 0) s s A n u* tu c o c 4) O t4> . 1 X II X X 4) 4) U U O 0 z z c c o o W M • II UJ UJ 4> • (^ 4> V) C 0 X 0 *' U m o 4 • z *» S 0 ^ L (0 U. c o 4^ « n UJ • * 1 4) M a. a V 3 4> M 5 5 cZ o. c o 3 ° n HI • u> Ol O) 2 § « i c w t. « 0 U N C 3 • -J _J e 0 i — o — *> X n 0 •a x i. « (i N » TJ — UJ UJ . H to H • VI 0 9 > X 4 X 4> X 3 Ol § -E X oi * 4) 4) •M •« o, a. a a a 5 81 n • • O O a a U L 0 0 o w HI HI § § 10 00 11 « O) ao §00 « r- to S S Ok Ol s s c o *> ot c c •- x a « 3 a • * 0 en L 3 X- II M E a i. C I. — » UJ X • n V) >^ ^ c. 0 « ja X- 3 a 2 c H 4> -O 0 5 i£ m « 4> O — 0. c * Ol O O 0 s - L N • • X X r* 00 1 S « u e M 4> U) V. C^ ^ 01 a. o • flu Of Z* au n •o o ]g ^ • 4> z Z o w en 4> 2 01 -I « 4 <3 u a k u £ o Ii 356
------- o z o t-l ^ u cc Ul >, 4) c 3 0 u X 4> u •o 0 3 C • w *k C 0 u rl f*fc UJ cd <-tl «** h- • • 4) i 0 4> W e» *r *^ K) Q r*» m « (9 Q CO CO Q o c o CL 1% X O ^ Jk o « 2 * c * 0 4> • t "O « « — « X X M 4> lf> ^) 0 c en to •* 1 * < f a "" • 3 U X C o — 4> a. L, • M — O « X • 1 « •O — 0 0. C 0 CD x •»• 2 4) Cl M _J f«t ^ t «fl a u ^j c M e CO 1- ~ ^ 0> *V O 0> ^T w m CM (Q i1* CO co co s Q O O L 0 O 4> C C IA L, ii n M J£ C 3 U « a t t- 0 L. 4) 0 JC 43 4> U 0 M O x n — no ^-% ro CO ce 1- 4» ^ 01 — .- o JJ L, 4) — t. c A O A ** U U 4> 4i U J3 C O •- M C « (0 4> m— a. u 0 O — a. a. n ^ CL Cb O 0 «•- 002 c « X • <5 X • 0 a CL U 4> U c •H •— M 01 m m CM W ^ 00 r* QQ ^ t-4 (O m ro « O O O O L « C 4> C — M 4> o — n «i X* •*- U 0 Vp **• C t (••-no 0 2 -J U. C C 0 J 41 O 0! X 4» N U Ol Bl 4> 3 i M 4> C/) m 4> T m •» • 3 0 00 — > 0 -0 *> c ce o) 4 CA ra 0 — 4> 4> 4J E C A a 9 o o 4J — L, C ' ^ 0- «l n U c o 4> ^ 0) 49 4J — x c n. 0 « _i — e» a. »- 4» *> 2 a o o. ^ « • j; * c ° H c x a « U 3 c — .a c * »i x to ^ f^. s ^ CO o ^ c n ^_ u 0 1 3 X L 0 c o •~ 2 • ^3 i 0 ,« • u 4) (/> 4} L 0 ^ .4. a 0 u L ^ ^g « c5 c c Q N CD (*• ^> fh» « 2 SC9 0) S 9 3 09 O) C C 'I 'i O U X X 0 KB X — (. > 1 t O 3 CD 0 4> 4> C C o o •S 2 Ul , 0 > < L 0 c Ol 5 J 10 •* J J -o — jC IS S ,•• tO CD C u •- 0 T> C — 0 o II X * •o w u « n * o o w o L, L, 4> 4> * 3: M n S <5 c c c c T N h* 3 M x n 4> U L C o a 2 J V U X II O O ~) 4> 4> 2 2 • n 4> to 0 C 4> to > c 4 *~ A TJ 2 u S u> fl CO ce CM M CL M fl O X O ™ 1 *> 4> C 2 l» — O D. U W 4> (t OJ O -I i u • a. c 4> 0 2 a. 00 >-l 0> ^ fO ii Cf> *3k W « CO ^ a s * OS CO rt O) Ok Q Q Q O 01 C 0 C t. 6 L, « O 0 V U N II X 3 0 -I -J CD C O 0 £ -X C) o — U <- X — CO U 4J c c * 3 n 0 S Z Z m 41 to to IO 41 CO 0 3 JC C CO fl It f« 2 3: r* 0 L. 0 TS O 4> x — Q. X u n c « 3 O a 0 0 0 -X U <_> o u -o L. — O 4» 41 L. * C 0. •1 CL c5 C3 O U W 0 £ «• C l-t 4> « O 3 •v <0 SIN <0 CO d) m a Q 5 s s < < u M 0 JC — U 4> •J 3 CC CD 0 II M — • O JC L. L, 4l 0 0 i *> U) CL we » •P S M 1. to 0 a u o 2 * CO C e ° S CO CM 4fe U C M O y| L. 0 CL a o 4) :*: a. M a u •to 3 2 -0 0 M C C9 °" a • L O O O o c 3 « 5 a. CM 0) S t* § 9 4 •_ X a "o •o A «« •« x a. n •^ x a. •J •o a x — je x 3 ^ u to ^g| , 0 ^ je T x n • M > A »•• 0. K n ^ U e o je e x n 0. IA o s 5 s CO Ol A •_ f a "J •o ~> «•! X a. 0 •• x CL 0 II X • n 4> M 4> 0 JC L •o L ro CM * II O C » O H 0 N ffl 4> e *o 0. CO T> ro « 6) 3 1- C s •- f— 4> O e So u § ~ A •• X a "0 •o » ^ !•• X a. A •^ X Cv "0 •o II — ~2 t z ^| • 0 0 L 5 « - a L • 0 M 4 O c £ •o 0 ce 357
------- £ o M u K LU U 9810378 980S391K9 ••« co a> 3 3 0. 0. •o c • C O u u 0 o. n .« 0 0) Ol c a » cc c 0 c. u > M >_ 0 O) c c • a -X a _l c c a a jc k. "• JS ~ 3 o c o 4J it a L. X 3 U V) a •c c c a a O o *> o k o O) L .0 If. •o 3 O C - Q -a c *J >, L A C -D c X a «T o X c. JJ c 0 u c 0 a a * c. o c o c. 4> « c. o 00 Q. X c *) U o o 4> n X C C C a « — •- a c 'E o X N 0 !_ U ffi M N 2 (B N N 3 3 _l _J U « t, A ID H 2 -O M I O U X *) I • a 4> U o & » a •o c 'Z c o u Ul m n *> •o n o — L. « — c °a o o U U. M k CD O K CO «^ n *> V) 0 • t. 0> •< U ffi «« « *> x U U « -o -0 *> c 10 uj (/) 0 O> *> j= 0) x _ 4) 0 jc «/> — c — (/>— *> « C .< !• mi.— TJ *> U.M0-O.A e 2 «i •*» o o — «l 3 CO » • *-o e • C. — — 0 1. *> co JB *> X co 4> L -P co « *> » *» CO C. 0 > co co > < -o at e * o )_ a en 41 C C L. CO •»•-• oe — — — — M — C 0 3 4) *> a. « a *> O a. *> 0. o u • O *j CO 0) C. — a. M 3 C to o u *> — a. »i • U je - O A c 0 X4) to - a. o uot JC U < — a. M • o x 1. 3 ^> C 3 w o — HI — o « — OL Z O • o • (. t, • ffl 0 jc *> — a. *> to o c. — — — — a. 1. « a cn— <_> u * uo 0 -P o 0 o.*> a i. a 0 u c o. u — L. E 3 04> xii * Z a 0 u.- oa u 0 e U *» u U C o *» M « tu M • o *> co 0 « 0 ae o C. 358
------- TABLE 72. VIRGINIA GAS SITES Site Status Danville Town Gas Craghess St. RR Depot Danville, VA 24541 Fredericksburg Town Gas 400 Charles Street Fredericksburg, VA 22401 Fulton Bottom Town Las Fulton & Williamsburg Road Richmond, VA 23201 Lynchburg Town Gas Black Water Street Lynchburg, VA 24501 Newport News Town Gas Terminal Blvd. & 22nd Street Newport News, VA 23601 Norfolk Town Gas Monticello & VA Beach Rd. Norfolk, VA 23501 Portsmith Town Gas Gust Lane Portsmouth, VA 23701 Roanoke Town Gas NE Klmbeil & Rutherford Ave. Roanoke, VA 24001 Suffolk Town Gas Hill Street Suffolk, VA 23434 Alexandria Town Gas City Yard Town Gas Discovery (PA) Discovery (PA) Discovery (PA) Discovery (PA) Discovery (PA) Discovery (PA) Discovery (PA) Site inspection Discovery (PA) Discovery (PA) PA = Preliminary assessment. 359
------- that have been located, are currently under investigation, or have been listed by the States. 4.4 EVALUATION OF THE RADIAN LIST OF MANUFACTURED-GAS SITES I The list of gas production sites compiled by Radian is a faithful compi- lation of the site material contained in Brown's, but it has several short- comings, most of v;hich result from the way Brown's compiled and reported information on the manufactured-gas industry. Sitss were listed in Brown's corporate designation. Whenever two plants merged their management, Brown's usually stopped listing one plant, even though it was often still in production. In Radian's compilation of the data from Brown's, plants that merged with larger plants showed no production at the site, even though gas was still produced there. The listing for Platts- burgh, New York, is a good example. The plant merged with New York State Electric and Gas Corporation in 1932, and subsequently its production was included with that of Ithaca, New York. The Radian compilation shows that no gas was produced under the Plattsburgh listing in 1940 and 1950, although the plant actually operated into the 1950's. Brown's Directrry includes only gas producers who sold their gas to con- sumers. Facilities that supplied gas to a limited market (e.g., a large hotel or an individual factory) did not appear in the directory. Many universities also had their own gas plants at one time; however, because they did not sell gas to consumers, they were not listed in Brown's. Brown's also did not list gas production at factories that generally manufactured producer gas for onsite heating purposes. An estimated 11,000 such gas producers were in operation in 1921 (Chapman, 1921). Most sites using producer gas would probably have several gas producers on each site, so the actual number of possible sites would be much lower than 11,000. Brown's Directory, however, reported none of these. Brown's Directory also did not record the movement of plant operating sites. It was common for gas companies to operate a small plant initially, outgrow it, and then expand to a larger facility. Brown's recorded the company's production as occurring at a single site rather than at two sites and, as a result, the records Radian compiled indicate only a single site. 360
------- Brown's generally included substantial information on plant byproducts marketed by individual companies (in later operating years), but Radian did not generally compile this information. The data available in Brown's could be very useful in evaluating individual sites, but a very large effort would be required to compile the data for all listed sites. The Radian compilation apparently did not include any gas purchased by gas companies from byproduct coke ovens. This was gas produced by coal car- bonization, which was not manufactured by a gas company, but was sold (generally locally) to a gas company by a coke manufacturer. From a waste or site standpoint, it makes no difference if the gas were produced by a coke company selling gas as a byproduct or by a gas company selling coke as a by- product. A town having a gas company that produced some gas and purchased additional gas from a local coke manufacturer would have had at least two gas production sites, but it would be reported only as one in the Radian compila- tion. When the data were compiled from Brown's at 10-year intervals, signifi- cant variations in rates of gas production were overlooked. The production of gas dropped sharply after 1930, and it did not recover until World War II. This would have produced errors i i the total amounts of gas reported, particularly for the production of carbureted water gas. 4.5 CONCLUSIONS Many States currently have active programs to examine nanufactured-gas sites specifically for possible environmental hazards. In most cases, the existing owners are requested to perform preliminary site assessments to determine the extent of site contamination. Any necessary remedial actions are determined only after the extent of contamination is known. Several States have used the Radian list of manufactured-gas sites to assist them in locating gas sites within their States. In most States, the environmental authorities are initially satisfied with determinations that no significant amounts of waste materials are moving off a site and that no significant groundwater contamination has occurred. Remediation is generally not performed at sites until some waste material moves offsite or additional use of the manufactured-gas site is planned. The site owners are generally content with leaving the sites as monitored (but 361
------- unreinediated) because the cost of carrying the site as undeveloped land Is small compared to the costs of remediation and redevelopment. In many cases, the sites have remained undeveloped land since the surface structures were removed. In summary, the Radian list of manufactured-gas sites presented several problems. Not all gas-manufacturing sites appeared in Brown's; hence, the list is incomplete. Brown's listed gas manufacturers by corporate designa- tion, so some companies listed as single sites in Brown's were actually com- posed of several operating plants. In addition, several p'lu-it sites were listed as only one when plants moved within cities. Cities having operating coke plants (which produced gas that was sold to gas companies) and gas compa- nies were reported as having only a single gas production site. The Radian list is a good starting point for locating gas plants because most of the towns listed had a gas-manufacturing plant. Local sources of information, however, should not be overlooked, and they should take pre- cedence over both information in Brown's Directory and in the Padian list. 362
------- REFERENCES Adaska, W. S., and Cavalli, N. J. 1984. Cement barriers. In: 5th National Conference on Management of Uncontrolled Hazardous Waste Sites, October 31-November 2, 1983, Washington, DC. Hazardous Materials Control Research Institute. Silver Spring, MD. pp. 126-130. Addridge, W.M. 1951. Biochem. J. 48:271. AGA. 1919. Statistics of the artificial gas industry, 1919. AGA. Arlington, VA. AGA. 1920. Statistics of the manufactured gas industry for the year 1920. AGA. Arlington, VA. AGA. 1921. Statistics of the manufactured gas industry for the year 1921. AGA. Arlington, VA. AGA. 1923. Statistics of the manufactured gas industry for the year 1923. AGA. Arlington, VA. AGA. 1924. Statistics of the manufactured gas industry for the year 1924. AGA. Arlington, VA. AGA. 1925. Statistics of the manufactured gas industry for the year 1925. AGA. Arlington, VA. AGA. 1928. Annual statistics of the manufactured gas industry, Statistical Bulletin No. 6. AGA. Arlington, VA. AGA. 1929. Annual statistics of the manufactured gas industry of the U.S., Statistical Bulletin No. 7. AGA. Arlington, VA. AGA. 1930. Annual statistics of the manufactured gas industry of the U.S., Statistical Bulletin No. 8. AGA. Arlington, VA. AGA. 1931. Annual statistics of the manufactured gas industry of the U.S., Statistical Bulletin No. 9. AGA. Arlington, VA. AGA. 1932. Annual statistics of the manufactured gas industry of the U.S., Statistical Bulletin No. 10. AGA. Arlington, VA. AGA. 1933. Annual statistics of the manufactured gas industry of the U.S., Statistical Bulletin No. 11. AGA. Arlington, VA. AGA. 1934. Annual statistics of the manufactured gas industry in 1933, Statistical Bulletin No. 12. AGA. Arlington, VA. 363
------- AGA. 1935. Annual statistics of the manufactured gas industry in 1934, Statistical Bulletin No. 17. AGA. Arlington, VA. AGA. 1936. Annual statistics of the manufactured gas industry in 1935, Statistical Bulletin No. 21. AGA. Arlington, VA. AGA. 1937. Annual statistics of the manufactured gas industry in 1936, Statistical Bulletin No. 26. AGA. Arlington, VA. AGA. 1938. Annual statistics of the manufactured gas industry in 1937,. Statistical Bulletin No. 32. AGA. Arlington, VA. AGA. 1939. Annual statistics of the manufactured gas industry in 1938, Statistical Bulletin No. 36. AGA. Arlington, VA. AGA. 1940. Annual statistics of the manufactured gas industry in 1939, Statistical Bulletin No. 40. AGA. Arlington, VA. AGA. 1941. Annual statistics of the manufactured gas industry in 1940, Statistical Bulletin No. 44. AGA. Arlington, VA. AGA. 1942. Annual statistics of the manufactured gas industry in 1941, Statistical Bulletin No. 48. AGA. Arlington, VA. AGA. 1943. Annual statistics of the manufactured gas industry in 1942, Statistical Bulletin No. 52. AGA. Arlington, VA. < AGA. 1945. Annual statistics of the manufactured gas industry in 1943, Statistical Bulletin No. 55. AGA. Arlington, VA. AGA. 1955. Annual statistics of the manufactured gas industry in 1944, Statistical Bulletin No. 58. AGA. Arlington, VA. AGA. 1961. Historical statistics of the gas industry. AGA. Arlington, VA. AGA. 1966. Gas facts-.-a statistical record of the gas utility industry in 1965. AGA. Arlington, VA. AGLA. 1985. Discussion of lime purification. Proc. Am. Gas L. Assn. 2(5):140-145. Alexander, J.O'D., and Macrosson, K.I. 1954. Squamous epithelioma probably due to tar ointment in a case with psoriasis. Br. med J. IV: 1089. Alexander, W. J. 1984. Technical issues regarding variable density fluids with considerations for ground-water monitoring at treatment, storage, and disposal facilities. Draft White Paper Report prepared for U.S. Environmental Protection Agency. Washington, DC. Alrich, H. W. 1934. Gas tanks now and then. Am. Gas J. (5):75-79. 364
------- American Health Association. 1980. Cyanide: Standard method for the examination of water and wastewater, 15th ed. American Health Association. Washington, DC. API. 1983. Land treatment practices in the petroleum industry. American Petroleum Institute. Washington, DC. Arcos, J. and Argus, M.F. 1974. Chemical induction of cancer. Vol. 11A. Academic Press, NY. Auebach. 1897. What buyers of spent oxide want. Am. Gas Light J. (6):1035- 1036. Bains, G. B. 1921. Treatment of contaminated water at Reading. AGA Mon. 3(5):302-304. Ball, C.B. 1885. Tar cancer. Trans. Acad. Med (Irel). 3:318-321. Barcelona, M. J., Gibb, J. P., and Miller, R. A. 1983. A guide to the selection of materials for monitoring well construction and ground-water sampling. EPA-600/52-84-024. U.S. Environmental Protection Agency. Ada, OK. Barcelona, M. J., Gibb, J. P., Helfrich, J. A., and Garske, E. E. 1985. Practical guide to groundwater sampling. EPA-600/2-85-104. U.S. Environmental Protection Agency. Ada, OK. Barlow, T. S., and Kennedy, J. S. 1922. Separation of condensates in water gas plants. AGA Mon. 4(10):639-640. Barr Engineering Co. 1976. Soil and ground-water investigation, coal-tar distillation and wood preserving site, St. Louis Park, MN: Consultant's report to Minnesota Pollution Control Agency, Minneapolis, MN, 152 pp. Bateman, E. 1922. Coal-tar and water-gas tar creosotes: their properties and methods of testing (U.S. Dept. of Agriculture Bulletin No. 1036). U.S. Government Printing Office. Washington, DC. Bell, J. 1876. Parafin epithelioma of the scrotum. Edinb. Med. J. 22:135-137. Bennett, J. G. 1935. Dehydration of gas-plant tar emulsion and disposal of liquid wastes. Proc. Penn. Gas Assn. pp. 101-114. Benson, J.M., Hill, J.O., Mitchell, C.E., Newton, G.J., and Carpenter, R.L. 1982a. Toxicological characterization of the process stream from an experimental low Btu coal gasifier. Arch. Environ. Contain. Toxicol. 11:363-371. 365
------- 11 Benson, J.M., Mitchell, C.E., Royer, R.E., dark, C.R., Carpenter, R.L., and Newton, G.J. 1982b. Mutagenicity of potential effluents from an experimental low Btu coal gasifier. Arch. Environ. Contain. Toxicol. 11:547-551. Berenblum, 1. 1945. 9,10-Dimethyl-l,2-Benzanthracene as a highly potent carcinogen for the rabbit's skin. Cancer Res. 5:265-268. Berenblum, I., and Schoental, R. 1947. Carcinogenic constituents of coal tar. Br. 0. Cancer. 1:157-165. Berg, R. R. 1975. Capillary pressures in stratigraphic traps. Bull. AAPG, 59: (6)939-956. Berggreen, R. G. 1985. Characterization of hydrogeology and groundwater contamination at a creosote wood treating plant in southern Illinois. In: Site Selection, Characterization, and Design Exploration, Proceedings of the 28th Annual Meeting of the Association of Engineering Geologists (AGE). Ebasco Strvices, Greensboro, NC. Berliner, R.W. 1950. Am. J. Psysiol. 160:325. Biggs, W. I., and Wool rich, W. R. 1925. Handbook of steam engineering. Eng. Press. Knoxville, TN. Bird, E. H. 1923. Some of the latest developments in gas purification. Chem. and Met. Eng. 29(0:16-18. Blanusa, M., Kostial, K., Matkovic, V., and Landeka, M. 1979. Cortical index of the femur in rats exposed to some toxic metals and ash from coal gasification. Arh. Hig. Rada Toksikol. 30 (suppl.):335-340. Boutwell, S. H., Brown, S. M., Roberts, B. R., and Atwood Anderson-Nichols & Co. 1985. Modeling remedial actions at uncontrolled hazardous waste sites. EPA-540/2-85/001. U.S. Environmental Protection Agency. Washington, DC. Bowen, H.J.M. 1966. Trace elements in biochemistry. Academic Press. New York, NY. Bretherick, L. 1981. Hazards in the chemical laboratory. 3rd. ed. The Royal Society of Chemistry. London, England, p. 137. Brown, J. A. 1926. The small gas plant and the large one. AGA Mon. 8(6):349-360. Brown, R. B. 1919. Disposal of waste from gas plants. AGA Mon. 1(4):189- 190. Brown. (Published yearly after 1889). Brown's directory of American gas companies. Moore Publishing Co. Inc. New York, NY. A 366
------- Bruusgaard, A. 1959. Occurrence of certain forms of cancer in gasworkers. (Norw.) T. Norsk. Laegefor. 79:755-756. Burnham, A. K., Calder, G. V., Fritz, J. S., Junk, G. A., Svec, H. J., and Willis, R. 1972. Identification and estimation of neutral organic contaminants in potable water. Anal. Chem,, 44(1):139-142. Burnham, A. K., Calder, G. V. , Fritz, J. S., Junk, G. A., Svec, H. J., and Vick, R. 1973. Trace organics in water: their isolation and identification. Journal of the American Water Works Association 65(11)722-725. Butlin, H.T. 1892. Cancer of the scrotum in chimney-sweeps and others. 1. Secondary cancer without primary cancer. II. Why foreign sweeps do not suffer from scrotal cancer; III. Tar and paraffin cancer. Br. Med. J. 1:1341-1346; 11:1-6, 66-71. Calle, E.E. 1985. Environmental health and safety assessments for direct coal liquefication: Vol. 7b. Guidance for the development of an occupational health surveillance system. ORNL/FETEP-8. Oak Ridge National Laboratory. Oak Ridge, TN. February, pp. 31-32. Campbell, J. H., Wang, F. T., Mead, S. W., and Busby, J. F. 1979. Groundwater quality near an underground coal gasification experiment. J. Hydrol. 44(304):241-266. Carlton, W.W. 1966. Experimental coal tar poisoning in the White Pekin Duck. Avian Dis. 10:484-502. Carswell, E. D. 1928. Tar emulsion. Proc. S. Gas Assn. pp. 56-62. Cecil, L. K. 1950. Underground disposal of process wastewater. Ind. Eng. Chem. 42:594. Centers for Disease Control. 1984. Report on Gas Works Park, Seattle, WA. June 28. Transmitted to Regional Administrator, U.S. EPA, Region 10, Seattle, WA, July 9, by V. R. Houk, M.D., Director, Center for Environmental Health. Chapman. 1921. Mech. Eng. 43:717. Cho, P., Walsh, P.J., and Copenhauer, E.D. 1983. Health and environmental effects document on direct coal 1iquefaction-1982: Volume 2. Appendices ORNL/TM-8624/V2. Oak Ridge National Laboratory. Oak Ridge, TN. September, pp. B-147 to -B-169. Cichowicz, N. L., Pease, R. W., Jr., Stoller, P. J,, and Yaffe, H. J. 1981. Use of remote sensing techniques in a systematic investigation of an uncontrolled hazardous waste site. U.S. Environmental Protection Agency. Cincinnati, OH. 367
------- Clean Harbors, Inc. 1985. Geohydrologic study report, Mendon Road Site, Attleboro, MA. File No. A-4193. Goldberg-Zoino and Associates, Inc. Newton Upper Falls, MA. May. Cochran and Hodge. 1985a. Guidance on remedial investigations under CERCLA. EPA-540/G-85/002. U.S. Environmental Protection Agency. Washington, DC. Cochran and Hodge. 1985b. Guidance on feasibility studies under CERCLA. EPA-540/G-85/003. U.S. Environmental Protection Agency. Washington, DC. Cole, D. W. 1972a. Letter to Mr. Mike Waller, Department of Parks and Recreation, Seattle, WA, January 4. Cole, D. W. 1972b. Letter to Mr. Mike Waller, Department of Parks and Recreation, Seattle, WA, March 30. Cole, D. W., and Machno, P. S. 1971. Myrtle Edwards Park—a study of the surface and subsurface soil materials. Submitted to the City of Seattle, Department of Parks and Recreation, December 22. Copenhauer, E.D., and Dudney, C.S. (Coord.) 1985. Environmental, health and safety assessments for direct coal liquefaction: Vol 7a. Potential health effects. ORNL/FETEP-7. Oak Ridge National Laboratory. Available from National Technical Information Service. Springfield, VA. 153p. Costigan, J. T. 1953. Experience with the centrifugal dehydration of gas tar, particularly from high Btu oil gas production. AGA Proc. pp. 765- 766. Costigan, J. T. 1954. Report on gas tar dehydration. AGA Proc. p. 590. Cruickshank, C.N.D., and Gourevitch, A. 1952. Skin cancer of the hand and forearm. Br. J. Ind. Med. 9:74-79. Cull inane, M. J., Jr. and Jones, L. W. 1984. Draft technical handbook for stabilization/solidificaiton of hazardous waste. U.S. Environmental Protection Agency; Cincinnati, OH. Cundall, K. N. 1927. Recovery of sulphur from gas. Chem. and Met. Eng. 34(3):143-147. D'Appolonia, D. J. 1980. Soil-bentonite slurry trench cutoffs. Journal of the Geotechnical Engineering Division, ASCE. GT-4:389-417. Oashiell, P. T. 1935a. Review of developments in heavy oil tar and emulsion handling. AGA Proc. pp. 800-810. Dashiell, P. T. 1935b. Review of developments in heavy oil tar and emulsion handling. Am. Gas J. (11):44. 368
------- Dashiell, P. T. 1944. Gas enriching oil from the viewpoint of the gas industry. AGA Proc. pp. 475-479. de Vries, W.M. 1928. Pitch cancer in the Netherlands. In: Report of the International Conference on Cancer. Fowler Wright Ltd. London, pp. 290-292. Dean, A.L. 1948. Epithelioma of scrotum. J. Urol. 60:508-518. Deelman, H.T. 1923. Some remarks on experimental tar (Fr.). Bull. Assoc. Fr. Etude Cancer. 12:24-30. Deniy, F., and Powell, A. R. 1933. Liquid purification of gas and some recent developments. AGA Proc. pp. 913-925. Oenig, F. 1945. Industrial coal carbonization. In: The chemistry of coal utilization, Vol. 1, Chapter 21, H. H. Lowry (ed.). John Wiley. New York. pp. 774-833. Dietz, D. N. 1971. Pollution of permeable strata by oil components. Water Pollution of Oil. London Institute of Petroleum. Dipple, A. 1976. Polynuclear aromatic carcinogens. In: Chemical Carcinogens (ACS Monograph No. 173), C.E. Searle (ed.). American Chemical Society. Washington, DC. pp. 245-314. Dobrin, M. B. 1960. Introduction to geophysical prospecting. McGraw-Hill. New York, NY. . Doerr, J. S., and Gibson, P. B. 1966. Commercial aspects of coal tars and pitches. In: Bituminous materials, asphalts, tars, and pitches, Vol. 3, Chapter 4. John Wiley. New York. pp. 117-138. Doig, A.T. 1970. Epithelioma of the scrotum in Scotland in 1967. Health Bull. 28:45-51. Doll, R. 1952. The causes of death among gas-workers with special reference to cancer of the lung. Br. J. Ind. Med. 9:180-185. Doll, R., Fisher, R.E.W., Gammon, E.J., Gunn, W., Hughes, G.O., Tyrer, F.H., and Wilson, W. 1965. Mortality of gasworkers with special reference to cancers of the lung and bladder, chronic bronchitis, and pneumoconiosis. Br. J. Ind. Med. 22:1-12. Doll, R., Vessey, M.P., Beasley, R.W.R., Buckley, A.R., Fear, E.C., Fisher, R.E.W., Gammon, E.J., Gunn, W., Hughes, G.O., Lee, K., and Norman-Smith, B. 1972. Mortality of gasworkers--final report of a prospective study. Br. J. Ind. Med. 29:394-406. Downing, R. C. 1932. Ten years of gas purification. Am. Gas J. (6):15-18 and (8):24-26. 369
------- Downing, R. C. 1934. Developments in condensing, scrubbing and purification of gas. Am. Gas J. (5):63-66. g Drew, K. 1984. Gasworks Park--summary of results. Prepared by Ecology and Environment, Inc., Seattle, WA, for U. S. EPA Region 10, J. E. Osborn, Regional Project Officer. TDD R10-8403-11, July 18. Dunkley, W. A., and Barnes, C. E. 1921. An American gas purification investigation. Gas J. (4):27. Durkin, W., Sun, N., Link, J., Paroly, W., and Schweitzer, R, 1978. Melanoma in a patient treated for psoriasis. South. Med. J. 71:732-733. Dutton, L. R. 1919. The disposal of waste water from water gas plants and streams adjacent to parks. AGA Monthly 1(4) :191-193. E. C. Jordan, Co. 1984. Site contamination audit construction procedures, Brattleboro-Hinsdale Bridge. Prepared for State of New Hampshire Department of Public Works and Highways. E. C. Jordan Co. Portland, ME. Ehrenfield, J., and Bass, J. 1983. Handbook for evaluating remedial action technology plans. EPA-600/2-83-076. U.S. Environmental Protection Agency. Cincinnati, OH. Ehrlich, G. G., Goerlitz, D. P., Godsy, E. M., and Hult, M. F. 1982. Degradation of phenolic contaminants in ground water by anaerobic bacteria: St. Louis Park, MM. Ground Water 20(6)703-710. A Epler, J.L., Larimer, F.W., Rao, T.K., Nix, C.E., and Ho, T. 1978. Energy- related pollutants in the environment: Use of short-term tests for mutagenicity in the isolation and identification of biohazards. Environ. Health Perspect. 27:11-20. ERT (Environmental Research & Technology). 1983. Recommended plan for a comprehensive solution of the polynuclear aromatic hydrocarbon contamination problem in the St. Louis Park area. Appendix I. Document P-B690-161. Concord, MA. ERT (Environmental Research & Technology) and Koppers Company, Inc. 1984. Handbook on manufactured gas plant sites. Prepared for Utility Solid Waste Activities Group, Superfund Committee, Washington, DC. Published by Edison Electric Institute, pp. 6-1 to 6-69. Fassett, D.W. Unpublished data. Eastman Kodak Co. Rochester, NY. Fetter, C. W. 1980. Applied hydrogeology. Charles E. Merrill Publishing Co. Columbus, OH. Fife, J.G. 196?. Carcinoma of the skin in machine tool setters. Br. J. Ind. Med. 19:123-125. 370
------- 4 Fischer, E. L. 1933. The manufacture of a high Btu gas in a standard carburetted water gas set to serve as an emergency substitute for natural gas or for peak loads. AGA Proc. pp. 789-794. Forstall, A. E. 1934. Reminiscences, changes in the use of town gas. Am. Gas J. (8):23-24. Franks, B. J., Goerlitz, D. F., and Baedeck, M. J. 1985. Defining extent of contamination using onsite analytical methods. In: Proc. of Conf. on Organics in Groundwater, Houston, TX. Freeze, R. A., and Cherry, J. A. 1979. Groundwater. Prentice-Hall, Inc. Englewood Cliffs, NO. Fuller, W. H. 1984. Cyanides in the environment with particular attention to the soil. In: Conference on Cyanide and the Environment. Colorado State University. Fort Collins, CO. p. 19-44. Fulweiler, W. H. 1921. Report on water-gas tar. AGA Mon. 3(10):558-562. Gas Engineers Handbook, 1st ed. 1934. McGraw Hill. New York. Gas Works Park, (no date). History of park development. Available through the site manager. U.S. EPA Region 10 Office. Seattle, WA. Glover, F. H. 1951. Report of subcommittee on tar dehydration. AGA Proc. pp. 265-266. Glowacki, W. L. 1945. Light oil from coke oven gas. In: The chemistry of coal utilization, Vol. 2, Chapter 28, H. H. Lowry (ed.). John Wiley. New York. pp. 1136-1231. Gollmar, H. A. 1945. Removal of sulfur compounds from coal gases. In: The chemistry of coal utilization, Vol. 2, Chapter 26, H. H. Lowry (ed.). John Wiley. New York. pp. 947-1007. Gosselin, R.E., Smith, R.P., and Hodge, H.C. n.d. Clinical toxicology of commercial products, 4th ed. Williams and Wilkins. Baltimore, MO. Graham, R., Hester, H.R., and Henderson, J.A. 1940. Coal-tar-pitch poisoning in pigs. J. Am. Vet. Med. Assoc. 96: 135-140. Green, R.M. 1910. Cancer of the scrotum. Boston Med. Surg. J. 163:792-797. Green, S. 1939. Light oil scrubbing for gum formers. AGA Proc. pp. 573- 585. Grigorev, Z.E. 1960. On the carcinogenic properties of the pechora coal tar. Prob. Oncol. 6:883-886. 371
------- Grimwood, J. B. 1896. Oil tar and Us uses. Amer. Gas Lt. 0. 65(8) :28?- 284. Haag, R. 1971. A report for substantiating the master plan for Myrtle 1 Edwards Park, City of Seattle. Prepared for Wes Uhlman, Mayor; Department of Parks and Recreation; Board of Park Commissioners; and Board of Public Works. April. Haga, I. 1913. Experimental studies on the occurence of atypical epithelial and mucous skin proliferations. (Ger.). Z. Krebsforsch. 12:525-576. Hall, E. I. 1947. Production of high Btu gas. AGA Men. (10):434-436. Hansen, E.S, 1983. Mortality from cancer and ischemic heart disease in Danish chimney sweeps: A five-year follow-up. Amer. J. Epid. 117:160-164. Hansen, P. 1916. Disposal of gas house wastes. Illinois Gas Association Proceedings 12:124-135. Harris, M. R., and Hansel, M. J. 1983. Removal of polynuclear aromatic hydrocarbons from contaminated groundwater. In: National Conference on Management of Uncontrolled Hazardous Waste Sites, October 31-November 2, 1983, Washington, DC. Hazardous Materials Control Research Institute. Silver Spring, MD. pp. 253-265. Harry Hendon and Associates, Inc. 1982. Stabilization of solid wastes disposal site for Alabama Gas Corporation. Birmingham, AL. Haslam, R. T., and Russell, R. P. 1926. Fuels and their combustion, 1st ed. McGraw Hill. New York. Hauschidt, C. J. 1922. Notes on dehydration of water gas tar emulsion. AGA Mon. 4(3):181-182. Heide, G., and Werner, H. 1984. Process of safely disposing of waste materials. U.S. Patent No. 4,456,400. Heller, I. 1930. Occupational cancers. Cancer caused by coal tar and coal tar products. J. Ind. Hyg. 12:169-196. Hem, J. D. 1970. Study and interpretation of the chemical characteristics of natural water. USGS Water Supply Paper 1473. 363 pp. Henry, S.A. 1937. The study of fatal cases of cancer of the scrotum from 1911 to 1935 in relation to occupation, with special reference to chimney sweeping and cotton mule skinning. Am. J. Cancer 31:28-57. Henry, S.A. 1947. Occupational cutaneous cancer attributable to certain chemicals in industry. Br. Med. Bui. 4:389-401. 372
------- Henry, S.A., and Irvine, E.D. 1936. Cancer of the scrotum in the Blackburn Registration District. 1837-1929. J. Hyg. 36:310-337. Henry, S.A., Kennaway, N.M., and Kennaway, E.L. 1931. The incidence of cancer of the bladder and prostate in certain occupations. J. Hyg. 31:125-137. Herbst, L. J. 1931. Purification of high sulphur gases. AGA proc. 13:1067- 1070. Hickok, E. A., Erdmann, J. B.( Simmonett, M. J., and Boyer, G. W. 1982. Groundwater contamination with creosote wastes. In: National Conference on Environmental Engineering, Minneapolis, MM. Environmental Engineering Division, American Society of Civil Engineers, pp. 430-437. Hieger, 1. 1929. The influence of dilution on the carcinogenic effect of tar. J. Pathol. Bacteriol. 32:419-423. Hill, J.O., Royer, R.E., Mitchell, C.E., and Dutcher, J.S. 1983. In vitro cytotoxicity to alveolar macrophages of tar from a low-Btu coal gasifier. Environ. Res. 31:484-492. Hill, W. H. 1945. Recovery of ammonia, cyanogen, pyridine and other nitrogenous compounds from industrial gases. In: Chem. of coal util., Vol. 2, Chapter 27, H. H. Lowry (ed.). John Wiley. New York. pp. 1008- 1135. Hobbs, C.H., McClellan, R.O., Clark, C.R., Henderson, R.F., Griffis, L.C., Hill, J.O., and Royer, R.E. 1979. Inhalation toxicology of primary effluents on potential health and environmental effects from fossil fuel conversion and use. In: Proceedings of the Symposium on Potential Health and Environmental Effects of Synthetic Fossil Fuel Technologies. (CONF-780903). Oak Ridge, TN. Oak Ridge National Laboratory, pp. 163-175. Hoogendoorn, D. 1984. Review of the development of remedial action techniques for soil contamination in the Netherlands. In: 5th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. Hughes, S., and Richards, W. 1885. Gas works: their construction and arrangement and the manufacture and distribution of coal gas, 7th ed. Crosby, Lockwood, and Co. London. Hughes, T.J., Wolff, T.J., Nichols, D., Sparacino, C., and Kolber, A.R. 1983. Synergism and antagonism in complex environmental mixtures: Unmasking of latent mutagenicity by chemical fractionatlon. In: Kolber, A., Wong, T.K., Grant, L.D., DeWoskin, R.S., and Hughes, T.J. (eds.). In vitro testing of environmental agents, Part B. Plenum. Oxford, pp. 115-144. 373
------- Hull, W. Q., and Kohlhoff, W. A. 1952. Oil gas manufacture. Ind. and Eng. Chem. 44(5):936-94S. Hult, M. F., and Schoenbe'rg, M. E. 1981. Preliminary evaluation of ground- water contamination by coal-tar derivatives, St. Louis Park area, MN. USG5 Open-File Report 81-72. 57 pp. Hult M. F., and Schoenberg, M. E. 1984. Preliminary evaluation of groundwater contamination by coal-tar derivatives, St. Louis Park area, MN. U.S. Geological Survey Water-Supply Paper 2211. 53 p. Hydrosample. 1984. Hazardous waste evaluation, Mendon Road Site, Attleboro, MA. Prepared for Curtois Sand and Gravel, Pawtucket, RI. November. IARC (International Agency for Research on Cancer). 1984. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 34, polynuclear aromatic compounds, part 3, industrial exposures in aluminum production, coal gasification, coke production, and iron and steel founding. World Health Organization. Lyon, France. June. 219 pp. IARC (International Agency for Research on Cancer). 1985. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 35, polynuclear aromatic compounds, part 4, bitumens, coal-tars and derived products, shale-oils and soots. World Health Organization. Lyon, France. January. 271 pp. Jarvis, H. 1980. Scrotal cancer in a pitch worker. J. Soc. Occup. Med. 30:61-62. Johnson, A. J. 1932. The refractory screen oil gas process. Gas Age Record 69(25):737-741. Johnson Division. 1975. Groundwater and wells. UOP, Inc. St. Paul, MN. Jones, H. E. 1928 Phenol recovery plant avoids waste pollution of streams. Chem. and Met. Eng. (4):215-218. Kennaway, E.L. 1925a. The anatomical distribution of the occupational cancers. J. Ind. Hyg.' 7:69-93. Kennaway, E.L. 1925b. Experiments on cancer producing substances. Br. Med. J. 11:1-4. Kennaway, E.L., and Kennaway, N.W. 1947. A further study of the incidence of cancer of the lung and larynx. Br. J. Cancer 1:260-298. Kennaway, N.W., and Kennaway, E.L. 1936. A study of the incidence of cancer of the lung and larynx. J. Hyg. 36:236-267. 374
------- Keys, W.S., and MacCary, L. M. 1971. Application of borehole geophysics to water-resources investigations. Collection of Environmental Data. Chapter El. U.S. Geological Survey. Reston, VA. Kohl, A. L., and Riesenfeld, F. C. 1985. Gas purification, 4th ed. Gulf Publishing. Houston, TX. Kosson, D. S., Dienemann, £. A., and Ahlert, R. C. 1986. Field studies of in-situ extraction and soil-based microbial treatment of an industrial sludge lagoon. Proceedings of the National Conference on Hazardous Wastes and Hazardous Materials. Hazardous Materials Control Research Institute. Washington, DC. p. 102-109. Kostial, K.. Blanusa, M.V., Rabar, I., Maljkovic, T., Kello, D., Landeka, M., Bunarevic, A., and Stara, J.F. 1981. Chronic studies in rats exposed to liquid effluent from coal gasification process. J. Appl. Toxicol. 1:3-10. (suppl):319-326. Kostial, K., Kello, D., Rabar, J., Maljkovic, T., and Blanusa, M. 1979. Influence of ash from coal gasification on the pharmacokinetics and toxicity of cadmium, manganese and mercury in suckling and adult rats. Arh. Hig. Rada Toksikol. 30 Kostial, K., Kello, D., Blanusa, M., Maljkovic, T., Rabar, I, Bunarevic, A., and Stara, J.F. 1980. Toxicologic studies of emissions from" coal gasification process. I. Subchronic feeding studies. J. Environ. Pathol. Toxicol. 4:437-448. Kostial, K., Rabar, I., Blanusa, M., Kello, D., Maljkovic, T., Landeka, M., Bunarevic, A., and Stara, J.F. 1982. Chronic and reproduction studies in rats exposed to gasifier ash leachates. Sci. Total Environ. 22:133-147. Kugel, W. F. 1947. Gas-making qualities of oils. AGA Mon. 29(6):297-300. Lafornara, J. P., Nadeau, R. J., Allen, H. !_., and Massey, T. I. 1982. Coal tar: pollutants of the past threaten the future. In: Proc. 1982 Hazardous Materials Spill Conf., Milwaukee, WI. Bureau of Explosives. Washington, DC. pp. 37-42. Lane, J. M. 1921. Method of cleaning accumulation of dirt from deep drip. AGA Mon. 3(4):245-246. Laudani. H. 1952. Report of tar dehydration subcommittee. AGA Proc. pp. 845-847. Lee, W.R., Alderson, M.R., and Downes, J.E. 1972. Scrotal cancer in the north-west of England, 1962-68. Br. J. Ind. Med. 29:188-195. Lenson, N. 1956. Multiple cutaneous carcinoma after creosote exposure. New Engl. J. Med. 254:520-523. 375
------- Leuders, C. J. 1942. Lynn Gas and Electric Company procedure of dehydrating Bunker "C" tar emulsion and coke oven tar emulsion. AGA Proc. pp. 310- 311. Liebe, G. 1892. Tar and paraffin cancer (Ger.) Schmidt's Jahrbuch. 236:71-80. Lowry, H. H. ed. 1945. Chemistry of Coal Utilization, Volumes 1 and 2, John Wiley and Sons, NY. Lowry, H. H. ed. 1963. Chemistry of Coal Utilization, Supplemental Volume, John Wiley and Sons, NY. Luke, D. 1954. Liver dystrophy associated with coal tar pitch poisoning in the pig. Vet. Rec. 66:643-645. Mackay and Shiu. 1977. J. Chem. Engr. Data 22(4) :399-402. Mackison, F.W., Stricoff, R.S., and Patridge, Jr. (eds.). Occupational health guidelines for chemical hazards. DHHS (NIOSH) Publication No. 81-123. 1981. National Institute for Occupational Safety and Health (NIOSH). Washington, DC. MacLean, C.W. 1969. Observations on coal tar poisoning in pigs. Vet. Rec. 84:594-598. Manouvriez, A. 1876. Diseases and hygiene of workers making tar and pitch briquettes (Fr.). Ann. Hyg. Publ. 45:459-482. Manz, A. 1980. Respiratory and urinary tract as indicator of occupational (coal tar) carcinomas in carbonization and pipe workers. In: Werner, W., and Schneider, H.W. (eds.). Air contamination by polycyclic aromatic hydrocarbons (Ger.) (VDI-Berichte No. 358). Vereins Deutscher Ingenieure. Dusseldorf. pp. 227-235. Marquard, F. F. 1928. The industrial viewpoint of the phenol waste disposal problem. Am. J. Pub. Health 18:1497-1500. Mattraw, H. C.( Jr., and Franks, B. J., (eds.). 1984. Movement and fate of creosote waste in ground water, Pensacola, Florida: U. S. Geological Survey toxic waste—ground water contamination program. U.S. Geological Survey Open-File Report 84-466. May, W. E., Wasik, S. P., and Freeman, D. H. 1978. Determinaton of the solubility behavior of some polvcylic aromatic hydrocarbons in water. Anal. Chem. 50(7):997-1000. Mayor's Committee on Gas Works Park. 1984. Draft report, Seattle, WA. June. McAuliffe. Nature 200:1092-1093. 376
------- McCarty, P. 1., Rittman, B. E., and Bouwer, I. J. 1984. Microbiological processes affecting chemical transformations in groundwater. In: Groundwater pollution microbiology, Britton, G., and Gerba, C. P., (eds.). John Wiley & Sons, Inc. Now York, NY. McGivney, S.E., and Shelton, S.P. 1984. Detection and treatment of ferrocyanide generated by the photographic and electroplating industries. Paper presented at the Conference on Cyanide and the Environment, Tuscon, A2. McKay, W. E. 1901. Tar and water from water-gas. Proc. Am. Gas L. Assn. 18(9):21-42. McManus, T. J. 1982. Risk assessment case studies: go or no go. Presented at Am. Soc. Civil Engrs. Nat. Conf. on Environ. July 14. Miller, D. G. 1983. In situ treatment alternatives and site assessment. In: Management of uncontrolled hazardous waste sites. Hazardous Material Control Research Institute (HCMCRI). Silver Spring, MD. pp. 221-225. Milne, J.E.H. 1970. Carcinoma of the scrotum. Med. J. Austr. 2:13-16. Minnesota Department of Health. 1938. Report on investigation of disposal of wastes at the Republic Creosoting Company, St. Louis Park, MN. 10 pp. Minnesota Department of Health. 1974. Report on investigation of phenol problem in private and municipal wells in St. Louis Park, MN, Hennepln County. 24 pp. Morgan, J. J. 1926. Manufactured gas, a textbook of American practice, 1st ed. J. J. Morgan. New York. Morgan, J. J., and Stolzenbach, C. F. 1934a. Heavy oil tar emulsions in the water gas process. AGA Mon. (7):245-248. Morgan, J. J., and Stolzenbach, C. F. 1934b. Heavy oil tar emulsions in the water gas process. AGA Mon. (8)-.277-280. Morgan, J. J. 1945. Water gas. In: Chemistry of coal utilization, Vol. 2, Chapter 37, H. H. Lowry (ed.). John Wiley. New York. pp. 1673-1749. Munro, N.B., Fry, R.J.M., Gammage, R.B., Haschek, W.M., Calle, E.E., Klein, J.A., and Schultz, T.W. 1983. Indirect coal liquefaction: A review of potential health effects and worker exposure during gasification and synthesis. ORNL-5938. Oak Ridge National Laboratory. Oak Ridge,'TN. Available from National Technical Information Service. Springfield, VA. January. 250 pp. Murphy, E. J. 1928. The treatment of water gas plant waste. Gas Age Record (6):817-818. 377
------- Murdock, W. J. 1926. Use of bituminous coal as generator fuel. AGA Proc. pp. 924-932. National Bureau of Standards (NBS). 1981. Certificate of Analysis for Standard Reference Material 1647. NBS. Washington^ DC. No author. 1896. Record of the proceedings of the special joint committee on the affairs of the Richmond City Gas Works. The Williams Printing Co. NWWA/API. 1984. Petroleum hydrocarbons and organic chemicals in groundwater. Proceedings of the NWWA/API Conference. Houston, TX. Odell, W. W. 1922. Water gas tar emulsions. AGA Hon. 4(3):136-140. O'Donovan, W.J. 1920. Epitheliomatous ulceration among tar workers. Br. J. Dermatol. Syphilis. 32:215-228. Ogawa, I., Junk, G. A., and Svec, H. J. 1981. Degradation of aromatic compounds in ground-water, and methods of sample preservation. Talanta 28:725-729. Oliver, T. 1908. Tar and asphalt workers' epithelioma and chimney-sweeps' cancer. Br. Med. J. 11:493-494. Orth, 0. G., Jr. 1984. Our latent environmental pollution. Letter to the City Desk, Seattle Times, Seattle, WA. April 4. Overcash, M. R., and Pal, D. 1979. Design of land treatment systems for industrial wastes — theory and practice. Ann Arbor Science. Ann Arbor, MI. Pacific Coast Gas Association. 1926. Proceedings of the P.C.G.A. 17:407. Page. 1881. Residual -products. Proc. Am. Gas L. Assn. 5(9):79-93. Parke, F. B. 1935a. Carburetted water gas from heavy oils with related tar and emulsion problems. AGA Proc. pp. 759-767. Parke, F. B. 1935b. Carburetted water gas from heavy oils with related tar and emulsion problems. Am. Gas J. (6):32-34. Parkin, G. F., and Calabria, C. R. 1985. Principles of bioreclamation of contaminated ground waters and leachates. 3rd Annual Symposium on International Industrial and Hazardous Waste. Alexandria, Egypt. Patch, I.L. 1954. Pitch and pulmonary carcinoma. Br. J. Tubercul. Dis. Chest. 48:145-150. 378
------- Patnode, T., Linkenheil, R., and Lynch, J. W. 1985. Closure and remedial action at a creosote impoundment. In: 6th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. Patty, F.A. 1981. Patty's industrial hygiene and toxicology, 3rd ed. John Wiley and Son, Inc. New York. Perkins Jordan Inc. 1984. Site assessment conducted for Massachusetts electric company, Spencer, MA. Reading, MA. Petrino, S. A. 1947. Chemical treatment of water gas tar emulsion. AGA Proc. pp. 608-614. Pew, A. E. 1940. Effect of marketing policies and refining methods on future availability of residue enrichment oils. AGA Proc. pp. 631-635. Pfannkuch, H. 0. 1982. Problems of monitoring network design to detect unanticipated contamination. Ground Water Monitoring Review 2(l):67-76. Powell, A. R. 1922. Cyanogen in illuminating gas and its removal. AGA Mon. 4:621-627. Powell, A. R. 1929. Report of subcommittee on disposal of waste from gas plants. AGA Proc. pp. 928-940. Powell, A. R. 1945a. Gas from coal carbonization-preparation and properties. In: The chemistry of coal utilization, Chapter 25, 2nd ed., H. H. Lowry (ed.). John Wiley. New York. pp. 921-946. Powell, A. R. 1945b. Removal of miscellaneous constituents from coal gas. In: The chemistry of coal utilization, Vol. 2, Chapter ^9, H. H. Lowry (ed.). John Wiley. New York. pp. 1232-1251. Quinn, E. J., Wasielewski, T. N., and Conway, H. L. 1985. Assessment of coal for constituents migration: Impacts on soils, groundwater, and surface water. Northeast Utilities Service Company. Hartford, CN. Rabar, I. 1980. The effect of ash from coal gasification on reproduction in rats (Abstract No. P. 259). Toxicol. Lett., Special Issue 1, 257. Rabar, I., Maljkovic, T., Kostial, K.( and Bunarevic, A. 1979. The Influence of ash from coal gasification on body weight in relation to age and sex. Arh. Hig. Rada Toksikol. 30 (suppl.)-.327-334. Washington, DC. Radian Corp. 1987. Survey of Town-Gas and By-Product Production in the U.S. (1800-1950), EPA Report No. EPA-600/7-85-004. 379
------- r Reilly, C.A., Jr., Peak, M.J., Matsushita, T., Kirchner, F.R., and Haugen, O.A. 1981a. Chemical and biological characterization of high-Btu coal gasification (The D.O., Gray, R.H., HYGAS Process) and Felix, W.D. IV. Biological Activity. In: Mahlum, (eds.). Coal conversion and the environment: chemical, biomedical and ecological considerations. 20th Annual Hanford Life Sciences Symposium, Richland, WA. October 1980. (CONF-801039). Reilly, C.A., Jr., Boparai, A.S., Bourne, S., Flotard, R.D., Haugan, D.A., Jones, R.E., Kirchner, F.R., Matsushita, T., Peak, M.J., Stamoudis, V.C., Stetter, J.R., and K.E. Wilzbach. 1981b. Health and environmental studies of coal gasification process streams and effluents. Presented at the EPA Environmental Aspects of Fuel Conversion Technology--)/!. A symposium on coal-based synfuels, Denver, Colorado, October 26-30, 1981. Reilly, C.A., Jr., and Peak, M.J. Washington, DC. Technical Information Center. U.S. Department of Energy, pp. 310-324. Reilly, C.A., Pahnke, V.A., Kirchner, F.R., Duggal, K., and Fritz, T.E. 1985a. Bladder tumors resulting from dermal exposure to aromatic amines. In: Proceedings of the 1st annual DOE Contractors Meeting—Health and environmental effects of complex chemical mixtures. CONF-8506155. U.S. Department of Energy, Office of Energy Research. Available from National Technical Information Service. Springfield, VA. August 1985. pp. 75-78. Reilly, C.A., Peraino, C., Haugen, D.A., Mahlum, D.D., arid Springer, D.I. 1985b. Promotion of preneoplastic changes in liver by coal-derived organic mixtures applied to skin. In: Proceedings of the 1st Annual DOE Contractors Meeting—Health and environmental effects of complex chemical mixtures. CONF-8506155. U.S. Department of Energy, Office of Energy Research. Available from National Technical Information Service. Springfield, VA. August 1985. pp. 79-82. Rhodes, E. 0. 1945. The chemical nature of coal tar. In: The chemistry of coal utilization, Vol. 2, Chapter 31, H. H. Lowry (ed.). John Wiley. New York. pp. 1287-1370. Rhodes, E. 0. 1966a. The history of coal tar and light oil. In: Bituminous materials, asphalts, tars, and pitches, Vol. 3, Chapter 1, A. J. Hoiberg (ed.). Interscience. pp. 1-32. Rhodes, E. 0. 1966b. Water-gas tars and oil gas tars. In: Bituminous materials, asphalts, tars, and pitches, Vol. 3, Chapter 2, A. J. Hoiberg (ed.). Interscience. pp. 33-56. Rittmann B. F., McCarthy, P. I., and Roberts, P. V. 1980. Trace organics biodegradation in aquifer recharge. Ground Water 18(3)236-243. Roberts, J. R., Cherry, J. A., and Schwartz, F. W. 1982. A case study of a j chemical spill: polychlorinated biphenyls (PCBs), 1. history, ™ distribution, and surface translocation. Water Res. Res. 18(3):525-534. 380
------- Ross, P. 1948. Occupational skin lesions due to pitch and tar. Br. Med. J. 11:369-374. Royer, C. 1984. Letter to the people of Seattle. Office of the Mayor, City of Seattle, Charles Royer, Mayor, July 25. Rover, R.E., Mitchell, C.E., Hanson, R.L., Dutcher, J.S., and Bechtold, W.E. 1983. Fractionation, chemical analysis, and mutagenicity testing of low- Btu coal gasifier tar. Environ. Res. 31:460-471. Rulkens, W. H., and Assink, J. W. 1984. Extraction as a method for cleaning contaminated soil: possibilities, problems, and research. In: 5th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. pp. 576-583. Ryan, J. R., and Smith, J. 1986. Land treatment of wood-preserving wastes. Proceedings of the National Conference on Hazardous Wastes and Hazardous Materials. Hazardous Materials Control Research Institute. Washington, DC. p. 80-86. Santodonato, J., Howard, P., and Basu, D. 1981. Health and ecological assessment of polynuclear aromatic hydrocarbons. J. Environ. Pathol. Toxicol. 5(1) :l-366. Schafer, J. M. 1984. Determining optimum pumping rates for creation of hydraulic barriers to ground water pollutant migration. DE84-013245. NTIS. Springfield, VA. Schamberg, J.F. 1910. Cancer in tar workers. J. Cutan. Dis. 28:644-662. Schmidt, J. 1943. A textbook of organic chemistry. D. Van Nostrand. New York, NY. Schoney, R., Warshawsky, 0., Hoi 1ingsworth, L., Hund, M., and Moore, G. 1981a. Mutagenicity of coal gasification and liquefaction products. Environ. Sci. Res. 22:461-475. Schoney, R., Warshawsky, D., Hoi 1ingsworth, L., Hund, M., and Moore, G. 1981b. Mutagenicity of products from coal gasification and liquefaction in the Salmonella/microsome assay. Environ. Mutagenesis 3:181-195. Schultz, T.W., Dumont, J.N., Clark, B.R., and Buchanan, M.V. 1982. Embryo- toxic and teratogenic effects of aqueous extracts of tar from a coal gasification electrostatic precipitator. Teratog. Carcinog. Mutagenesis 2:1-11. Schultz, T.W., Kyte, L.M., and Dumont, J.N. 1978. Structure-toxicfty correlations of organic contaminants in aqueous coal-conversion effluents. Arch. Environ. Contain. Toxicol. 7:457-463. 381
------- Schwartz, G. M. 1936. The geology of the Minneapolis-St. Paul metropolitan area. Minnesota Geological Survey Bulletin 27. 267 pp. Schwarz, F. P. 1977. Determination of temperature dependence of solubilities of polycyclic aromatic hydrocarbons in aqueous solutions by a fluorescence method. Journal of Chemical and Engineering Data 22(3)-.273-277. Schwarz, S. C.( and Keller, C. L. 1955. Current processing of tar emulsions at Portland Gas & Coke Company. AGA Proc. pp. 834-835. Schwille, F. 1967. Petroleum contamination of the subsoil—a hydrological problem. Joint Problems of Oil and Water Industries. Institute of Petroleum. London, pp. 23-54. Scott, A. 1922. On the occupation cancer of the paraffin and oil workers of the Scottish shale oil industry. Br. Med. J. IV:1108-1109. Seely, H. K. 1927. Report of the subcommittee on water gas tar emulsions, their prevention, treatment, and utilization. AGA Proc. 9:1078-1824. Seely, H. K. 1928. Report of the subcommittee on water gas tar production. •AGA Proc. pp. 1337-1348. Sexton, R.J. 1960. The hazards to health in the hydrogenation of coal: Part IV, the control program and the clinical effects. Archives of Environmental Health l(Sept.):42/208-65/231. Shambaugh, P. 1935. Tar cancer of the lip in fishermen. J. Am. Med. Assoc. 104:2326-2329. Shear, M.J. 1938. Studies in carcinogenesis. V. Methyl derivatives of l.,2-Benzanthracene. Am. J. Cancer 33:499-537. Shear, M.J. 1941. Studies in carcinogenesis. XVI. Production of subcutaneous tumors in mice by miscellaneous polycyclic compounds. S. Natl. Cancer Inst. 2:241-258. Shelton, F. H. 1897. The nuisance question in gas works. New Eng. A. Gas Eng. P. pp. 314-323. Simmons, G. A. 1924a. Other waste emulsions incidental to manufactured gas production. AGA Proc. 6:1431-1438. Simmons, G. A. 1924b. The dehydration of coal gas tar, water gas tar and other waste emulsions incidental to manufactured gas production. AGA Mon. 6(10):637-638. Sims, R. C. 1982. Land treatment of polynuclear aromatic compounds. Ph.D. dissertation. Dept. of Biol. Agr. Engineering. NC State University. Raleigh, NC. 382
------- Sims, R. C., and Overcash, M. R. 1983. Fate of polynuclear aromatic compounds in soil-plant systems. Residue Reviews 88:1-68. Sims, R. C., et al. 1984. Review of in-place treatment techniques for contaminated surface soils. Vols. 1 & 2. EPA-540/2-84-003 A and B. U.S. Environmental Protection Agency. Cincinnati, OH. Sittig, M. 1981. Handbook of toxic and hazardous chemicals. Noyes Data Corp. Siudyla, £. A. 1975. A hydrogeologic investigation of aromatic hydrocarbons in the aquifer supplying Ames, Iowa. Master's thesis, Iowa State University. Smith, F. A., and Eckle, T. F. 1966. Manufacture of coal tar and pitches. In: Bituminous materials, asphalts, tars, and pitches, Vol. 3, Chapter 3. John Wiley. New York. pp. 57-116. Smith, L.H., and Witschi, H.P. 1981a. Lung tumor assay with two UMD electrostatic precipitator tar samples. In: Life sciences synthetic fuels semiannual progress report for the period ending June 30, 1981. K. E. Cowser, (ed.). ORNL/TM-7926. Oak Ridge National Laboratory. Oak Ridge, TN. pp. 14-18. Smith, L.H., Haschek, W.M., and Witschi, H.P. 1981b. Acute toxicity studies of a gasifier sample--UMD 83, In: University of Minnesota-Duluth coal gasification project quarterly progress report for the period ending June 30, 1980. K.E. Cowser (ed.). ORNL/TM-7466. Oak Ridge National Laboratory. Oak Ridge, TN. pp. 8-12. Smith, W.E. 1952. Survey of some current British and European studies of occupational tumor problems. Arch. Ind. Hyg. Occup. Med. 5:242-263. Sollman, T. 1957. A manual of pharmacology. Saunders. Southam, A.H., and Wilson, S.R. 1922. Cancer of the scrotum: .The etiology, clinical features, and treatment of the disease. Br. Med. J. IV:971-973. Sperr, F. W., Jr. 1921. Disposal of waste from gas plants. Gas Age Record 48(11):566-571. Sperr, F. W. 1923. Progress in liquid purification. AGA Proc. pp. 1200- 1238. Spink, M.S., Baynes, A.H., and Tombleson, J.B.L. 1964. Skin carcinoma in the process of 'Stanford jointing1. Br. J. Ind. Med. 21:154-157. Spooner, P. A. 1984. Stabilization/solidification alternatives for remedial action. 6th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. p. 214-220. 383
------- Springer, O.L., Miller, R.A., Weimer, W.C., Ragan, H.A., Buschbom, R.L., and j Mahlum, D.D. 1984. Effects of inhalation exposure to SRC-II heavy and ; middle distillates. PNl-5273. U.S. Department of Energy. Available from National Technical Information Service. Springfield, VA. November. ! 51pp. Stanfill, D. F., and McMillan, K. S. 1985. Radar-mapping of gasoline and other hydrocarbons in the ground. 6th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. pp. 269-274. Steinbrueck, V. 1971. Gas generators and operating equipment. Registry of Historic Places, Inventory Nomination. College of Architecture and Urban Planning. Seattle, WA. ; Stotz, I., and Jamison, A. 1938. History of the gas industry. Stettiner \ Bros. New York. j Stover, E. L., and Chung, N. K. 1979. Petroleum processing and coal i conversion wastes, "j. WPCF 51(6):1393-1398. ! Sutton, C., and Calder, J. A. 1975. Solubility of alky! benzenes in distilled water and seawater at 25.0 C. Journal of Chemical and Engineering Data. 20(3):320-322. Teutschlaender, A. 1929. The 'Fohr-Kleinschmidt' pitch pulverization process^ as a process for prevention of pitch tumors in briquette factories. (Results of a visit to the Engelsburg mine near Bochum) (Ger.). Z. Krebsforsch. 30:231-240. Thompson S. N., Burgess, A. S., and O'Dea, D. 1983. Coal tar contamination and cleanup, Plattsburgh, NY. In: Proceedings of the National Conference on Management of Uncontrolled Hazardous Waste Sites, Washington, DC, October 31-November 2, 1983. Hazardous Materials Control Research Institute. Si Tver Spring, MD. pp. 331-337. Tillmanns, H. 1880. Oh tumours induced by tar, soot and tobacco (Ger.). Dtsch. Z. Chir. 13:519-53?. Timourian, H., Felton, J.S., Stuermer, D.H., Healy, S., Berry, P., Tompkins, M., Battaglia, G., Hatch, F.T., Thompson, L.H., 'Carrano, A.V., Minkler, J., and Salazar, E. 1982. Mutagenic and toxic activity of environmental effluents from underground coal gasification experiments. J. Toxicol. Environ. Health 9:975-995. Todd, D. K. 1980. Groundwater hydrology, 2nd ed. John Wiley and Sons. New York, NY. 384
------- TRW Energy Systems. 1976. Carcinogens relating to coal conversion processes. FE-2213-1. Energy Research and Development Administration. Available from National Technical Information Service. U.S. Department of Commerce. Springfield, VA. June 14. 129 pp. Troutman, D. E., Godsy, E. M., Goerlitz, D. F., and Ehrlich, G. G. 1984. Phenolic contamination in the sand and gravel aquifer from a surface impoundment of wood treatment wastes. Water-Resources Investigations Report No. 84-4230. Pensacola, FL. Tunis, A. B. 1933. Picture commemorates first display of gas in America. AGA Mon. 15(5) :171-174. U.S. EPA. 1978. Electrical resistivity evaluations at solid waste disposal facilities. SW-729, Office of Solid Waste, Washington, DC. U.S. EPA. 1978. Environmental assessment data base for coal liquification technology: Vol. 2--Synthoil, H-coal, and Exxon donor solvent process. EPA-600/7-78-184B. U.S. EPA. 1980a. The Carcinogen Assessment Group's list of carcinogens. July 14, 1980. U.S. EPA. 1980b. Ambient water quality criteria for acenaphthene. EPA-440/5-80-015. U.S. EPA. 1980c. Ambient water quality criteria for naphthalene. EPA-440/5-80-059. U.S. EPA. 1980d. Ambient water quality criteria for polynuclear aromatic hydrocarbons. EPA 450/5-80-069. U.S. EPA. 1980e. Ferric cyandide health and environmental effects profile *102. Office of Solid Waste, Washington, DC. U.S. EPA. 1982. Handbook for remedial action at waste disposal sites. EPA-625/6-82-006. OERR. Washington, DC. U.S. EPA. 1983a. Hazardous waste land treatment. SW-874. OSW. Washington, DC. U.S. EPA. 1983b. RCRA guidance document: land treatment units. Draft. OSW. Washington, DC. U.S. EPA. 1984a. Permit guidance manual on unsaturated zone monitoring for hazardous waste land treatment units. Draft for public comment. OSWER. Washington, DC. U.S. EPA. 1984b. Permit applicant's guidance manual for hazardous waste land treatment, storage, and disposal facilities. OSWER. Washington, DC. 385
------- U.S. EPA. 1984c. Permit guidance manual on hazardous waste land treatment demonstrations. Draft for public comment. EPA/530-SW-015. OSWER. Washington, DC. U.S. EPA. 1984d. Slurry trench construction for pollution migration control. EPA-590/2-84-001. OSWER. Washington, DC. U.S. Forest Products Laboratory. 1974. Characterization of wood-preserving creosote by physical and chemical methods of analysis. USDA Forest Research Paper FLP 195. Madison, WI. 32 pp. Umfleet, D. A., et al. 1984. Reclamation of PAH contaminated soils. Environmental Engineering Specialty Conference. ASCE. Los Angeles, CA. Unites, D. F., and Houseman, Jr., J. J. 1982. Field investigation and remedial action at sites contaminated with coal tars. In: Proc. 5th Annual Madison Conf. of Applied Res. and Practice on Municipal and Indus. Waste. Dept. of Engr. and Applied Sci., Univ. of Wisconsin Ext. Madison, WI. pp. 344-355. ,f Utermohle, C. E. 1948a. The production of high B.T.U. oil gas by the Hall process, AGA Proc. pp. 325-329. Utermohle, C. E. 1948b. Hall high B.T.U. oil gas process. AGA Mon. 30(ll):27-28. van Dam, J. 1967. The migration of hydrocarbons in a water-bearing stratum: joint problems of oil and water industries. Institute of Petroleum. London. van Zyl, D. 1984. Conference on cyanide and the environment. Colorado State University. Fort Collins, CO. Veley, M. A. 1885. The lime process for the purification of coal gas. J. Soc. Chem. Ind. (11):633-639. Villaume, J. F. 1982. The USA's first emergency Superfund site. In: Proc. 14th Mid-Atlantic Indus. Waste Conf. Allerman, J. E., and Kavanaugh, J. (eds.). Ann Arbor Science Pubs. Ann Arbor, MI. pp. 311-321. Villaume, J. F., Lowe, P. C., and Unites, D. F. 1983. Recovery of coal gasification wastes: An innovative approach. Proceedings of the Third International Symposium on Aquifer Restoration and Ground-Water Monitoring. NWWA. Worthington, OH. pp. 434-445. Villaume, J.F. 1984. Coal tar wastes: Their environmental fate and effects. In: Majumdar, S. K., and Miller, E.W. (eds.). In: Hazardous and Toxic wastes: Technology, Management and Health Effects, Chapter 25. The Pennsylvania Academy of Science. 386
------- Villaume, J. F. 1985. Investigations at sites contaminated with dense", nonaqueous phase liquids (NAPLS). Ground Water Monitoring Review 5(2):60-74. Vincent. 1907. Discussion of removal and disposal of tar. Prog. Age (7)-.387-388. Wagner, K., and Kosh, Z. 1984, In-situ treatment. 5th National Conference on Management of Uncontrolled Hazardous Waste Sites. Hazardous Materials Control Research Institute. Washington, DC. Walton, B.T. 1981. Chemical impurity produces extra compound eyes and heads in crickets. Science 212(Apr11):52-53. Wardale, N. H. 1930. Gas plant waste disposal. Proc. Pac. C. Gas Assn. 21:363-382. White, R. M., and Brandwein, 5. S. 1982. The application of geophysics to hazardous waste investigations. In: Proceedings of the American Defense Preparedness Association Symposium, Washington, DC. Willien, L. J. 1920. Disposal of gas plant wastes. Gas Age 46(11):349-352. Willien, L. J. 1923. Report of the committee on waste disposal from gas plants. AGA Proc. pp. 1239-1243. Wilson, P. J., Jr., and Wells, J. H. 1945. Ammoniacal liquor. In: The chemistry of coal utilization, Vol. 2, Chapter 32, H. H. Lowry (ed.). John Wiley. New York. pp. 1371-1481. Wilson, D.C., and Stevens, C. 1981. Problems arising from the redevelopment of gas works and similar sites. AER: Harwell Report R-10366. Department of Environment. London, England. Wilson, J. T., and McNabb, J. F. 1983. Biological transformation of organic pollutants in groundwater. Wilzbach, K.F., and Reilly, C.A., Jr. 1981. Chemical and toxicological characterization of HYGAS pilot plant process streams. In: Proceedings of the 1981 2nd International Gas Research Conference, Los Angeles, CA, September 28-October 1, 1981. Government Institute, Inc. Rockville, MD. pp. 1550-1561. Witschi, H.P., Haschek, W.M., and Smith, L.H. 1980. Level 2: Mammalian toxicology. In: University of Minnesota-Duluth coal gasification project quarterly report for the period ending December 31, 1979. K.E. Cowser (ed.). ORNL/TM-7268. Oak Ridge National Laboratory, Oak Ridge, TN, pp. 19-28. Woglom, W.H., and Herly, L. 1929. The carcinogenic activity of tar 1n various dilutions. J. Cancer Res. 13:367-372. 387
------- " I f ',' t Wood, E. C. 1962. Pollution of ground water by gas works waste. Proc. of the Society of Water Treatment and Exam. 11:32-33. Xanthakos, P. P. 1979. Slurry walls. McGraw-Hill. New York. 621pp. Yazicigil, H. 1977. Mathematical modelling and management of groundwater contaminated by aromatic hydrocarbons in the aquifer supplying Ames, Iowa. Master's thesis, Iowa State University. Yazicigil, H., and Sendlein, L. V. A. 1981. Management of groundwater contaminated by aromatic hydrocarbons in the aquifer supplying Ames, Iowa. Ground Water 19(6):648-665. Young, E. W. 1947. Mechanical dehydradation of coke oven tar emulsion. AGA Proc. pp. 698-700. Zohdy, A. A. R.( Eaton, G. P., and Mabey, 0. R. 1974. Applications of surface geophysics to ground-water investigations. In: Collection of environmental data, Chapter Dl. U.S. Geological Survey. 388
-------


                                         PB68-165790


                                      EPA/600/2-38/012
                                      February 1988
U.S. Production of Manufactured Gases:
 Assessment of Past Disposal Practices
                         by

     Scott M. Harkins, Robert S. Truesdale, Ronald Hill,
           Paula Hoffman, and Steven Winters

               Research Triangle Institute
                   P.O. Box 12194
       Research Triangle Park, North Carolina 27709
         EPA Contract No. 68-01-6826 D.O. #35
                 EPA Task Manager
                   Michael Black
             Land Pollution Control Division
          Office of Research and Development
    Hazardous Waste Engineering Research Laboratory
               Cincinnati, Ohio 45268
   HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
         OFFICE OF RESEARCH AND DEVELOPMENT
        U.S. ENVIRONMENTAL PROTECTION AGENCY
               CINCINNATI, OH 45268
              REPRODUCED BY
              UADEPARTMENTpF COMMERCE

                    WFORMATON SERVICE
                    SPRINGFIELD, VA. 22181

-------
                            _     TECHNICAL REPORT DATA
                            /prcose rttd Inttnictloru on the rrunt befon completing}
1. REPORT NO.
   EPA/600/2-88/012
                 3. RECIPIENT'S ACCESSIO
                             CCESSION N0_ -	
                            1657907g
4. TITLE AND SUBTITLE
 U.S. PRODUCTION OF MANUFACTURED GASES:
 OF PAST DISPOSAL PRACTICES
ASSESSMENT
                 5. REPORT DATE
                     February  1988
                 B. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
 Scott M. Harkins, Robert S. Truesdale, Ronald Hill,
 Paula Hoffman, and  Steven Winters
                 B. PERFORMING ORGANIZATION REPORT NO,
  PERFORMING ORGANIZATION NAME AND ADDRESS
 Research Triangle Institute
 P. 0. Box 12194
 Research Triangle Park, NC  27709
                                                           10. PROGRAM ELEMENT NO.
                 11. CONTRACT/GRANT NO.
                                                            68-01-6826 D.O. #35
13. SPONSORING AGENCY NAME AND ADDRESS
 Hazardous  Waste Engineering Research Laboratory
 Office of  Research and Development
 U.S. Environmental Protection Agency
 Cincinnati, OH  A5268	
                 13. TYPE OF REPORT AND PERIOD COVERED
                 14. SPONSORING AGENCY CODE
                  EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
    -, Former sites of gas manufacture present problems for remediation and reuse of the
 sites.  In sane cases, polluted groundwater and surface waters are near the sites.
 This study examines  the  history of the nanufactured-gas industry of the United States,
 its production processes,  disposal trends, waste toxicity, methods of site investi-
 gation, and the current  status of nanufactured-gas sites.  The report is Intended as
 a guide to those who are examining and evaluating nanufactured-gas sites for either
 environmental risks  or possible remediation.

      Six nanufactured-gas  sites and one spent oxide disposal area were visited during
 the project, and case studies were prepared for six former gas-manufacturing sites,
 two byproduct tar utilization facilities,  a creosoting plant and a coal tar processor.

      The current status  of nanufactured-gas sites in the United States was determined
 by contacting State  and  regional environmental officials and by discovering how their
 regions were treating nanufactured-gas sites. r.
7.
                               KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                             b.lOENTIFIERS/OPEN ENDEDTERMS
                                 cos AT i Field/Croup
B. DISTRIBUTION STATEMENT
 Release to Public
    19. SECURITY CLASS
      Unclassified
                                                                        21. NO. Or PAOES
    20. SiCURITY CLASS
      ^classified
                              22. PRICE
   P*m 2220-1 («••. 4.77)   mtviout BDITION n OBIOIKTC

-------
                        REGION V! LIBRARY
                        U S. ENVIRONMENTAL PROTECTION
                        AGENCY
                        1445 ROSS AVENUE  f
                        DALLAS, TEXAS 75202
PB88-165790
U.S. Production of Manufactured Gases
Assessment of Past Disposal Practices
Research Triangle  Inst.
Research Triangle  Park,  NC

Prepared for

Environmental Protection Agency, Cincinnati, OB
Feb 88
                                                                  •';•-.'
                                                                  -X'.

-------

-------

                                '•K:
                                i NOTICE
     This document has  been  reviewed  in  accordance  with U. S. Environmental
Protection Agency policy and approved for publication.  Mention of trade'&
names or commercial products does not constitute  endorsement or recommen-!
dation for use.
                                   11

-------

                                 FOREWORD
Today's tepidly developing and changing technologies and industrial    '^
products and practices frequently carry with them the increased generation
of solid and hazardous wastes.  These materials,  if improperly dealt  ^
with, can threaten both public health and the environment.   Abandoned  "
waste sites and accidental releases of toxic and  hazardous  substances to
the environment also have important environmental and public health   sV
implications.  The Hazardous Waste Engineering Research Laboratory
assists in providing an authoritative and defensible engineering basis
for assessing and solving these problems.  Its products support the   :-
policies, programs, and regulations of the Environmental Protection   ?§
Agency, the permitting and other responsibilities of State  and local   ••
governments, and the needs of both large and small businesses in handling
their wastes responsibly and economically.

This report reviews the history of the U. S. manufactured-gas industry,
the methods of production, wastes produced, disposal practices, potential
environmental effects of disposed wastes, and methods of site investigation
and remediation.  Several specific manufactured-gas sites are examined,
and a recent compilation of U. S. manufactured-gas sites is evaluated.?'
                                                                      >•*
For further information,, please contact the Land Pollution Control    ^
Division of the Hazardous Waste Engineering Research Laboratory.      ^
                        Thomas R.  Hauser,  Director
             Hazardous Waste Engineering Research  Laboratory
                                   111

-------
                                                                               5ft-;, -  3,-, -^rj
                                   CONTENTS
                                '.-fr
Chapter
         Notice 	'g  ii
         Foreword	• •	• •••:£  m
         Contents	    lv
         Figures	  viii
         Tables 	    xii
         Abbreviations and Symbols  	    xvi

         Executive Suntnary	  ES-1

         Introduction	:     1

   1.0   Historical Review of the Town Gas Industry	.;|     3
         1.1  Introduction	      3
         1.2  Town Gas Production	      5
              1.2.1  Producer Gas Production	      5
              1.2.2  Coal-Gas Production 	      8
                     1.2.2.1  Introduction 	:     8
                     1.2.2.2  Coal-Carbonization Retorts	    12
                     1.2.2.3  Byproduct Coke Ovens	    17
              1.2.3  Carbureted Water Gas	7   24
              1.2.4  Oil-Gas Production	f.   42
                     1.2.4.1  Introduction	.y   42
                     1.2.4.2  Pacific Coast Oil-Gas Processes	?)   43
                     1.2.4.3  High Btu Oil-Gas Processes	^   49
              1.2.5  Miscellaneous Gas Production Methods	;is   54
         1.3  Manufactured-Gas Cleaning and Purification Processes .....:    54
              1.3.1  Introduction	    54
              1.3.2  Condensers.	    59
              1.3.3  Tar Removal	    62
              1.3.4  Naphthalene and Light-Oil Scrubbers	.'   69
              1.3.5  Removal of Ammonia and Recovery	    78
              1.3.6  Phenol Removal and Recovery	.}    84
              1.3.7  Removal of Hydrogen Sulfide	t   88
                     1.3.7.1   Introduction	2*   88
                     1.3.7.2  Removal of Hydrogen Sulfide by Lime 	;*   88
                     1.3.7.3   Removal of Hydrogen Sulfide by Iron       ?j
                              Oxide	j*    90
                     1.3.7.4   Liquid Scrubbing for Hydrogen Sulfide    •£
                              Removal	    92
              1.3.8  Cyanide  Removal	'^    99
              1.3.9  Tar and  Light Oil Treatment	|f  102
              1.3.10 Gas Storage	.....$  105
              1.3.11 General  Purification Trains for Town Gases	ft  110
                                      iv

-------
.',•***
                                            CONTENTS  (continued)
                                                                IF"       Page

      1.4  Byproducts and Wastes from Town Gas Manufacture  	    118
           1.4,1   Introduction  	    118
           1.4.2   Description of Wastes	    119
                   1.4.2.1  Coal Tar, Water-Gas Tar, and  Oil-Gas
                           Tar	    119
                    .4.2.2  Oils 	    135
                    .4.2.3  Tar-Oil-Water Emulsions  	    136
                    .4.2.4  Waste Sludges	    139
                    .4.2.5  Ammonia  Recovery Wastes	    141
                    .4.2.6  Hydrogen Sulfide Removal Wastes  	    143
                    .4.2.7  Lampblack Wastes	    152
                   1.4.2.8  Ash, Clinker,  and  Coke  	;	    153
                   1.4.2.9  Firebrick and Building Materials  	    153
           1.4.3   Specific Articles On Waste  Disposal  	    154
      1.5  Production and Hlstorial Trends of the U.S. Town
           Gas  Industry	    159
           1.5.1   Introduction  	    159
           1.5.2   U.S. Gas Production Trends  	    159
           1.5.3   U.S. Gas Feedstock Trends 	    183
           1.5.4   Historical Events of the U.S. Gas Industry	    187
      1.6  Differences Between U.S. and  British Gas Industries	    18
      1.7  Conclusions from the Historical Review  	    196'

2.0   Investigation and Remediation of Town Gas Sites  	    203
      2.1  Contaminant Behavior and Fate	    203
           2.1.1   Byproduct Tars and Oils	    204
           2.1.2   Spent Oxides	    210
      2.2  Site Investigation 	?.....    211
           2.2.1   Introduction 	?.	    211
           2.2.2   Current Practices		    211
           2.2.3   Recommendations for Site Investigations	    216
                   2.2.3.1  Introduction  	    216
                   2.2.3.2  Information Collection and Review	    218
                   2.2.3.3  Field Investigation Plan Development,;	    222
      2.3  Site Remediation	K	    233
           2.3.1   Introduction 	$	    233
           2.3.2   Factors Affecting Site Remediation 	E....    234
           2.3.3   Remedial Action Alternatives 	f.....    237
                   2.3.3.1  Introduction  	.&....    237
                   2.3.3.2  Oils, Tars, and Lampblack 	M....    237
                   2.3.3.3  Spent Oxide Wastes 	.'$....    242
                   2.3.3.4  Contaminated Soil	';*.	    246
                   2.3.3.5  Contaminated Groundwater	i.....    255
           2.3.4   Conclusions 	.T.....    258
3.0   Site Investigations of Specific Town Gas Sites 	^.....    261
      3.1  Introduction	;f.....    261
                      -A                      :A '              I

-------
I _ ,^iiff
                                           CONTENTS (continued)                        W


               Chapter                                                                 v Page

                        3.2   Site  Visits Performed by RTI	    261
                             3.2.1  Introduction  	    261
                             3.2.2  Colonial  Gas  Company,  Lowell,  Massachusetts 	    262
                             3.2.3  Massachusetts Electric Co.,  Spencer,               «
                                    Massachusetts 	.:   263
                             3.2.4  Fulton Gas Works, Richmond,  Virginia 	.r  264
                             3.2.5  Mendon Road Spent Iron Oxide Disposal Site, near
                                    Attleboro, Massachusetts	    270
                             3.2.6  Pawtucket, Rhode Island 	;  271
                             3.2.7  Taunton,  Massachusetts	    271
                             3.2.8  Worchester, Massachusetts 	    271
                        3.3   Case  Studies of  Town Gas Sites 	    273
                             3.3.1  Introduction  	    273
                             3.3.2  Norwich,  Great Britain 	    273
                             3.3.3  Ames, Iowa 	    274
                                    3.3.3.1  Site History  	    275
                                    3.3.3.2  Extent of Contamination	    278
                                    3.3.3.3  Site Remediation	    282
                             3.3.4  Stroudsburg,  Pennsylvania	    283
                                    3.3.4.1  Site History  	    284
                                    3.3.4.2  Extent of Contamination	    286
                                    3.3.4.3  Site Remediation 	';   290
                                    3.3.4.4  New  Historical  Data on  Stroudsburg 	'   292
                             3.3.5  Plattsburgh,  New York  	'/  296
                                    3.3.5.1  Site History  	    297
                                    3.3.5.2  Extent of Contamination	    299
                                    3.3.5.3  Site Remediation 	.,   303
                             3.3.6  Seattle,  Washington 	.,   305
                                    3.3.6.1  Site History  	,;   305
                                    3.3.6.2  Extent of Contamination 	    310
;7   j                                 3.3.6.3  Site Remediation 	    311
t   '                          3.3.7  Brattleboro-Hinsdale Bridge, Brattleboro,          ;-.
t   i                                 Vermont 	."   315
t   |                                 3.3.7.1  Site History  	'..;   315
I   !                                 3.3.7.2  Extent of  Contamination	    315
F   '                                 3.3.7.3  Site Remediation	    320
r                             3.3.8  St.  Louis  Park,  Minnesota  	:   321
                                    3.3.8.1  Site History	.v   322
                                    3.3.8.2  Extent of  Contamination 	    324
                                    3.3.8.3  Site Remediation	    323
                             3.3.9  Pensacola,  Florida  	,  330
                                    3.3.9.1   Site History  	.»   331
                                    3.3.9.2 Methods of Investigation  	"   331
                                    3.3.9.3  Extent of  Contamination Findings	2|  332
                                    3.3.9.4  Site Remediation	|f  333
                       3.4  Conclusions	.£.	m  334


                                              tfe -.  vi

-------
                            CONTENTS  (continued)
                                                                          as-
                                                                          ?'•#•
                                                                          It-
Chapter

   4.0
State Status of Manufactured-Gas Sites	
4.1  Introduction	
4.2  Comparison of the Radian List and CERCLIS,
4.3  Examination of Manufactured-Gas Site
     Status 1n States	,
4.4  Evaluation of the Radian List of
     Manufactured-Gas SI tes	,
4.5  Conclusions	,
Page
>s-
§337
,337
               References.
  360
  361
 r-V
  363
                                      vii
                                                                          •'•&

-------

                                                FIGURE*

 i""-*         Number                                                                    Page
                1      Producer gas production 	      6
                2      Chapman gas producer 	     10
                3      Horizontal  retort with internal  producer gas'generator 	     11
 ;               4      Horizontal  retort 	     14
 j"*-            5      Charging machine	     15
 i
                6      Coke  pusher 	     16
 !               7      Continuous  vertical  retort 	     18
 i               8      Beehive  coke  oven  	     20
 I               9      Byproduct  coke  oven  	.'	     21
j^,*             10     Blue gas generator 	     25
 ;               11     Typical  three-shell water-gas set  	     28
 i               12     Blow and blow-run	     29
 i
                13     Up and down runs 	     30
                14     Air purge	     31
  fc,             15     Single-shell oil-gas apparatus 	     44
                16     Two-shell,  improved oil-gas apparatus 	     45
                17     Refractory screen oil-gas process 	     52
                18     Schematic diagram of Hall oil-gas process 	     55
                19    Direct  contact cooler 	     61
               20    P+A tar extractor 	?-.     63
               21     Cottrell  ESP 	"..     64
                                                  viii

-------
 I
1*
                                            FIGURES  (continued)


               Number                                                                 .    Page
                 	                         '                                        v-    	*—

                  22    Tar and water  separator  	     66

                  23    Naphthalene  scrubber	     70

                  24    Light oil  scrubber	     73

                  25    Indirect  process  for  ammonia recovery  	     80

                  26    Ammonia still  	     82

                  27    Semi direct process  for ammonia  recovery	     83

                  28    Benzene extraction  of phenols	     85

                  29    Koppers vapor  recirculation  process  for  phenol  removal	     87

                  30    Seaboard  process  for  H2S removal	     94

                  31    Thylox process  for  H2$ removal	     97

                  32    Cross section  of  single-lift gas  holder	    107

                  33    Cross section  of  multiple-lift  gas holder  	    108

                  34    Flowsheet  for  a coal-carbonization gas plant  	    Ill

                  35    Condensing and collecting system  of  a modern  byproduct
                        coke plant	    113

                  36    Flowsheet  for a carbureted water-gas plant	    114
                                                                                     •V
                  37    Flow diagram for gas manufacturing and byproducts of        •;]
                        Portland Gas and Coke Company Works  	..    116
                                                                                     --*•
                  38    Effects of carbonizing temperatures on yields of primary     •-';•
                        products from Pratt coal  	    128

                  39    Effects of carbonizing temperatures on composition of
                        tar from Pratt coal 	    129
                                                                                    y&
                  40    Total  U.S. manufactured-gas  sales, 1821-1956  	':.    163
                                                                                    '*iVv
                  41    Total  U.S. manufactured-gas  production,  by  type,             Tf
                        1919 to 1956	    164
                                                      ix
*.
•3?

-------
FIGURES (continued)
-------
FIGURES (continued)

Number Page

62 Gas Works Park (Seattle) 221

63 Monitoring of low relative density contaminant
with immiscible and soluble components 230
a*
64 Groundwater monitoring of multiple-density
components originating from single source area 231

65 Comparative groundwater monitoring of variable density
contaminants in uniform flow field 232

66 Ideal distribution of coal tar in the porous materials
at the Stroudsburg contamination site, as inferred from
capillary pressure theory 241

67 Treatment of soil by extraction with an aqueous solution 253

68 Fulton Gas Works (1876) 269

69 Stroudsburg site map with top-of-contamination (dash) and
groundwater (dot-dash) contours (in feet) shown 285

70 Top-of-sand contours (In feet) for the Stroudsburg coal-tar
4**- contamination site 285
! 71 Pittsburgh, New York, general site plan 298

' 72 Brattleboro-Hinsdale Bridge 319

. I 73 Location of former plant site, wetlands area, hydrologic
: | section, water table configurations, and location of key
• { wells at St. Louis Park, Minnesota. Generalized potentio-
;- ; metric surface, June 5, 1979, shown 323

74 Hydrogeologic and geophysical logs of Well W23 ("Hinkley"
, well on the site) 325

; ! 75 Radian 1 ist of manufactured-gas sites 338

:. | 76 CERCLIS waste sites 340
xi
-------
TABLES

Number Page
1 List of Manufactured-Gas Periodicals Reviewed 4
2 Composition and Characteristics of Producer Gas 7
3 Estimated Distribution of Gas Producers in the
Uni ted States 9
4 Composition and Total Heating Values of Typical
Coal Gases 23
5 Reactions During Blue Gas Manufacture 27
6 Typical Compositions of Blue Gas and Carbureted
Water Gas 33
7 Analyses of Fuels for Manufacture of Blue Gas and
Carbureted Water Gas 35
8 General Classes of Petroleum Products 37
9 Naphtha Distillation 39
an
10 Gas-Oil Distillation 40
11 Typical Carburetion 011s 41
12 Operating Cycles of Oil-Gas Producers 47
13 Comparison of Five Oil-Gas Plants ., 48
14 Distillation of 011s Commonly Used for 011 Gas £ 50
; 15 Results of Refractory Screen Oil-Gas Process \ 53
16 Hall Oil-Gas Process Operating Results "'"* 56
I 17 Miscellaneous Gas Production Processes 57
18 Temperature and Impurities in Raw Gases at Outlet of
Hydraulic Main or Washtox 58
xii
-------
TABLES (continued)
Number 5 T;> Page
19 Major Gas Purification Processes Used with Production ^
Processes '. ;¥ 60
20 Streams from Tar Separator £ 67
21 Other Methods of Tar-Water Separation ;* 68
22 Results of Naphthalene Scrubber at the Seaboard
Byproduct Coke Co., Kearny, New Jersey 71
23 Analysis of a Typical Crude Coke-Oven Light Oil '; 74
24 Chemicals Found in Light Oil from Coke Ovens \, 75
25 Distribution of Nitrogen in Coal-Carbonization "^
Products 79
26 Typical Concentrations of Hydrogen Sulfide in Town ' . ,
Gases 89
27 Operation of Seaboard Process .( 95
'-•' .if1
28 Operation of Nickel and Thylox Processes H 98
29 Concentration of HCN in Various Gases .5 100
30 Principal Components in Coal Tar Fractions .'I 103
31 Properties of Coal Tars 121
32 Comparative Analyses of Typical Dehydrated, Water-Gas •
Tars and Coal Tars Produced in 1921 to 1922 .; 122
33 Comparison of Some Pacific Coast Oil-Gas Tars .tP| 123
34 Comparative Analyses of Typical, Light, and Heavy f\
Water-Gas Tars and Coke-Oven Coal Tar 126
35 Approximate Quantities of Tar Produced in the ;
Manufacture of Town Gas 130
36 Uses of Carbureted Water-Gas Tars and Oil-Gas Tars 132
37 Analyses of Ammoniacal Liquors and the Still M
Wastes Therefrom ?'. .11 142
xiii
-------
• .»5S.S,V;
- - s»s*4wv
Number
38
39

41
42
43
44
45
46

52
53
54
55
56
TABLES (continued)
. f~. '.*-'.'.
An Analysis of Spent Oxide
Spent Oxide Compositions from 16 Gas Plants in
Illinois and Indiana ,Y
Estimated Generation of Spent Oxide Wastes from
Gas Production
Gas Plant Wastes, .£;• V:
Responses to Waste Disposal Survey
States Located Within Each Gas Production Region '-jr.,
Gas Heat Values Used to Convert Between ft3 and Therms 'I.
Significant Events of the Town Gas Industry ,
Summary of Typical Investigative Approaches for
Manufactured-Gass Sites .-*?.
Potential Treatment Technologies for Contaminated
Soils r. ..^f. ;..
;^. ' '• ';-i-'
Cost Estimates for Remedial Action Alternative at a
Creosote Impoundment
Volatile and Semfvolatile Organic AssaJjResults from
Test Pit and Brook .Samples "
Volatile and Semivolatlle Organic Assay'Results from
Boring Samples ;;. /.
Volatile and Semivolatile Organic Assay Results from
Monitoring Wei 1 s
Gas Production at Ames, Iowa .V
Neutral Compounds in a Contaminated Ames, Iowa, Well .E..
~$jf.' «'''-:4* . lig,,
Partial Analysis^of the Stroudsburg Coal Jar .......
Organic Contaminants in Shallow Groundwater
Production at Stroudsburg, Pennsylvania
xiv
Page
147

148

149
157
158
161
162
188
212

238

248

265

266
267
276
280
288
289
293
-------
TABLES (continued) ":f: S

*-X'
^
Number ;is :v ^ Page

57 Gas Production at Plattsburgh, New York 300
^

58 Gas Production at Seattle, Washington 307

59 Maximum Contaminant Levels: Gas Works Park,
Seattle, Washington 316

60 Gas Production at Brattleboro, Vermont 317

61 Maximum Contaminant Levels: Brattleboro, Vermont 320

62 Comparison of Radian Town Gas Sites to CERCLIS :•
for Alabama 341

63 Status of Manufactured-Gas Sites Within States 342

64 Gas Sit*s in California Compiled by Pacific Gas
and Electric Company 346

65 Delaware Gas Sites..' -. 349

66 Florida Gas Sites 350

67 Maryland Gas Sites ». 352

68 Michigan Gas Sites (Investigated by MichCon) .1. 353
'~\ ~*$r
69 New Jersey Gas Sites .'.. 354

70 New York State Gas Sites ;';*. A. 355
-* f
71 Pennsylvania Gas Sites 356

72 Virginia Gas Sites 359
XV
-------
ABBREVIATIONS AND SYMBOLS
AGA
bcf
Btu
CERCLA

COD
CP
CWG
DM30
EC50
EM
EP
GAC

GC/MS
g/kg
gpm
H202/UV
HCN
HPLC
IARC
LP
American Gas Association
Billion cubic feet
British thermal unit
Comprehensive Environmental Response,
Compensation, and Liability Act (1980)
Chemical oxygen demand
Candle power
Carbureted water gas
Dimethyl sulfoxide
Effective concentration affecting 50 percent of test
organisms
Electromagnetic
Extraction procedure
Granular-activated carbon
Gas chromatography/mass spectroscopy
Grams per kilogram
Gallons per minute
Hydrogen peroxide/ultraviolet
Hydrogen cyanide
High-performance liquid chromatograph
International Agency for Research on Cancer
Lethal concentration affecting 50 percent of test organisms
Liquid petroleum :?
xvi
.
8*
•"->£_;
--V-''
-------

-" ' ' ' • ---TL' ' --- ' — ' - ' — "•"" •*artMM*'"*'
;iji
ABBREVIATIONS AND SYMBOLS (continued)

mg -- Milligrams
mg/L -- Milligrams per liter
Mcf -- 1 million cubic feet (10^ ft3)
MSL — Mean seal level
NPL — National Priority List (Superfund)
03/UV — Ozone/ultraviolet
"** PAH -- Polynuclear aromatic hydrocarbon ;•/.
PCAP — Preliminary contamination assessment plan
PCB -- Polychlorlnated blphenyl
PCP — Pentachlorophenol
ppb -- Parts per billion ;.
ppm -- Parts per million ;
PVC — Polyvlnyl chloride
i
I RCRA -- Resource Conservation and Recovery Act (1976)
RTI — Research Triangle Institute
SCS ~ Soil Conservation Service
**•* SMR — Standardized mortality rate
TOC — Total organic carbon
fig/g — Mlcrograms per gram
USGS ~ U.S. Geological Survey
VOC — Volatile organic compound
*•"' ' -. V*- -
- WHO -- World Health Organization

^ xvii
-------
EXECUTIVE SUMMARY

Former sites of gas manufacture present problems for remediation and
reuse of the sites. In some cases, polluted groundwater and surface waters
are near the sites. This study examines the history of the manufactured-gas
industry of the United States, its production processes, disposal trends,
waste toxicity, methods of site investigation, and the current status of
manufactured-gas sites. The report is intended as a guide to those who are
examining and evaluating manufactured-gas sites for either environmental risks
or possible remediation.
The manufacture of gas for lighting and heating was performed in the
United States from 1816 into the 1960's. Three major processes were used to
manufacture gas: coal carbonization, carbureted water gas, and oil gas. Coal
carbonization consisted of heating bituminous coal in a sealed chamber, with
destructive distillation of gas from the coal and the formation of coke. The
gases were collected, cleaned, and distributed while coke was removed and sold
or used. The carbureted water-gas process used coke (or coal), steam, and
various oil products to produce a combustible product gas. Steam was fed
through a bed of incandescent coke, producing a gas containing hydrogen and
carbon monoxide. This gas (blue gas) then passed through two chambers
containing hot firebrick, where oil was sprayed into the gas and cracked into
gaseous hydrocarbons and tar. Oil gas cracked oil alone Into gaseous
hydrocarbons, tar, and carbon (lampblack). A variety of oil-based feedstocks
were used in the production of carbureted water gas and oil gas, including
naphtha, gas oil, fuel oil, and residuum oils.
The byproducts from the three processes were similar, but there were
important differences, which affect both the current character of wastes and
their toxicity. Tars produced from coal carbonization contained substantial
amounts of phenols and base nitrogen organics. The tars from carbureted water
gas and oil gas contain only trace amounts of these compounds because they
were not produced during the manufacture of gas. Coal carbonization also

ES-1
-------
produced substantial amounts of cyanide in the gas, which was removed during
gas cleaning and often appears in current wastes. Carbureted water gas and
oil gas produced only trace amounts of cyanide, and cyanide does not appear in
substantial quantities in wastes from these processes. Likewise, ammonia was
produced by coal carbonization, but it was not produced by oil or carbureted
water-gas manufacture. Wastes from the recovery of ammonia occur at plants
that coked coal to produce gas, but not at plants producing only carbureted
water gas or oil gas.
Gas production in the north central United States was principally coal
carbonization, oil gas was predominant on the West Coast, and carbureted water
gas was predominant in the South, the East Coast, and the Northeast. The
variation in the production processes used In various areas of the United
States reflects the relative cost of raw materials for production and markets
for byproducts in the regions. The types of production employed changed with
time, as did the materials used for gas production. This influenced both the
types of wastes produced and the disposal practices of the plants. Plant size
and access to markets were two major factors affecting the disposal practices
of manufactured-gas plants.
Tars and oils were produced as byproducts from all three production
processes. The tars and oils were generally recovered as byproducts from the
production of town gas, and they were usually separated from condensate water
by gravity separators. The tars could be either sold (as fuel or to tar
refiners), refined at the plant site, or burned in the boilers of the gas
plant. The recovered tars had a minimum value to the producing plants as fuel
because the use of tars as fuel replaced other fuels used for steam , •
production. Some tars were disposed very early in the production of coal-
carbonization gas, but recovered tars during this period were also frequently
burned in the coal-carbonization retorts. Smaller gas plants often produced
tars in insufficient quantities to justify their recovery, and these were
disposed with the waste condensate (this was particularly true of the
carbureted water-gas plants). Emulsions of tar and water occurred with the
production of carbureted water gas and oil gas, and because these were
difficult to separate, they were frequently disposed. The waste sludge from
ES-2
-------
I
) _ rsmv-'
the purification of light oils was generally disposed on the plant "dump,"
along with other off-spec or difficult to handle tars.
The tars produced for carbureted water gas were usually less viscous and
less dense than were the tars produced by coal carbonization. These tars are
more mobile in the environment than are most coal tars. The properties of
collected tars changed with respect to where the tar was collected within the
purification trains. The heavier tars condensed first within the gas
purification system, and these were the most viscous and the densest tars.
The tars that condensed later in the purification system were less viscous and
dense. Volatile organics (such as benzene and toluenes) were either scrubbed
from the gas as light oil or condensed in the gas holders or distribution
pipes as "drip oil." The variety of tars and oils produced within
manufactured-gas plants contributed to the wide range of organic contamination
generally present at gas sites.
Leakage of petroleum oils, tars, and aqueous condensates occurred
frequently from gas plants during plant operation. Early vessels used for the
underground storage of liquids were constructed of wood or brick. Several
historical references indicate that groundwater contamination was common near
gas plants, caused both by unintentional leakage from the plants and
intentional disposal practices.
The oils and tars from gas manufacture contain relatively high
concentrations of polycyclic aromatic hydrocarbons and are carcinogenic, with
numerous cases of skin cancer correlated with the occupational use of tars and
tar products. Phenols (from coal carbonization) are toxic to human, animal,
and plant life. Small concentrations of phenols cause taste and odor problems
in drinking water, imparting a medicinal taste to the water. Spent oxides
frequently develop low pH's and have relatively high concentrations of tars,
and the Iron cyanide complexes in spent oxide from coal carbonization appear
very stable and have relatively low toxicity.
The site investigation techniques applied to manufactured-gas sites are
not significantly different from those applied to other uncontrolled waste
sites and appear adequate for site assessment. Surface geophysical techniques
can be applied to help identify buried structures and the extent of possible
contamination. The location of underground structures at a site is
ES-3
-------
particularly Important because such structures frequently contain tars or oils
that could eventually leak or be released during future actions on the site.
Historical information on the operation, production, and layout of the gas
plants is frequently available and should be used wherever possible. Maps of
plant sites can be used to locate underground structures and possible dump
areas around the sites. The types of production employed by a plant determine
the nature and types of wastes produced, and the amount of gas produced
frequently affects the amount of waste remaining on a site. Many of the sites
examined to date are fairly stable (no wastes currently observed moving off
the site). These sites can often be adequately managed by taking no remedial
actions until the site is to be redeveloped.
Six manufactured-gas sites and one spent oxide disposal area were visited
during the project, and all showed visible contamination of soil by tars.
Ferrocyanides were visible form spent oxide at plants that produced coal gas,
but they were absent from those sites that produced carbureted water gas. The
characteristic odor of gas-manufacturing plants was observed at all the sites
examined. In addition to the visited sites, case studies were prepared for
six former gas-manufacturing sites, two byproduct tar utilization facilities,
a creosoting plant, and a coal-tar processor. These case studies were
prepared primarily from articles reported in the literature and illustrate
4's?-
current methods of site assessment and remediation. ??
The current status of manufactured-gas sites in the United States was
determined by contacting State and regional environmental officials and by
discovering how their regions were treating manufactured-gas sites. Many
States are examining manufactured-gas sites with other waste sites, and most
of these are conducting preliminary assessments of the sites. Where the
manufactured-gas sites have been ranked (by risk assessment), they have
,.;'
generally been ranked as posing a low hazard to both humans and the
environment. Groundwater contamination has been reported as several sites,
but it is not significant at many of the sites examined.
ES-4
-------
Wife
INTRODUCTION
After we had gone to the trouble of eliminating the oil and tar from
the stream, we met a difficulty not at all anticipated. Very near
our works and about ten years after they were installed, an arti-r
ficial ice plant was erected. The owners decided to dig artesian'
wells and found water of excellent quality, and ample quantity which
they used for three or four years with no evidence that we, their
neighbors, would cause them any trouble. In the early days of the
gas plant, the tar waste from the works had leaked through broken^
pipelines and from the wooden separator box used for waste disposal.
The tars seeped through the ordinary fissures of rock into the &?
ground around the well casing, and traces of oil began to appear In
the well water. Needless to say, there was very serious trouble for
a while and it is possible that other plants are storing up, /,
unawares, difficulties of the same kind (Dutton, 1919). •'

Between 1816 and the 1960's, combustible gas for heating, cooking, and
lighting was manufactured from coke, coal, and oil at 1,000 to 1,500 sites in

the United States. These facilities were called gas plants, gasworks, or town

gas plants. For most areas of the country, manufactured gas was the major gas

fuel available for use during this period. Some regional natural gas pipe-

lines were established before World War II, but it was only after the'war that

the technology was available for a national system of Interstate gas pipe-
lines. As natural gas was introduced into areas previously served by natural
gas, the gas companies stopped the gas-manufacturing operations and became

distributors of natural gas. Most companies maintained the manufacturing
••*.$>•
«*,,, facilities for several years after natural gas was available so that gas could
be manufactured to meet peak demand. With better storage of gas and the
installation of multiple pipelines serving regions, there was no longer any
need for manufactured gas, and the plants were demolished.

The old gas manufacturers frequently disposed solid and liquid wastes

onsite, making the current sites difficult to redevelop and posing potential

environmental problems from either groundwater or surface water contamination,

**** as evidenced by L.R. Dutton's testimony given at the start of the chapter.
-------
ELlii.-.-1-K.i^...
This report reviews the history of the U.S. manufactured-gas industry, the
methods of production, wastes produced, disposal practices, potential environ-
mental effects of disposed wastes, and methods of site Investigation and
remediation. Several specific manufactured-gas sites are examined, andTa
recent compilation of U.S. manufactured-gas sites is evaluated. •;-
Chapter 1 is a complete historical review of the U.S. manufactured-gas
industry, principally using information generated by the manufactured-gas
industry while it was in operation. The chapter reviews the production proc-
esses (1.2), gas purification methods (1.3), wastes produced and disposal
methods 1.4), trends of the gas industry (1.5), and a comparison of U.S,
practices and those used in Great Britain. -;& it
Chapter 2 describes the techniques previously used for site investiga-
tions (2.2) and site remediation (2.3). Chapter 3 reviews several specific
town gas sites, both those visited by the Research Triangle Institute (RTI)
(seven sites, Section 3.2), and sites reported in and reviewed through??
*- ' v£' '.*°v '' '•£,-
available literature (six gas-manufacturing sites and two tar-processing
plants, Section 3.3). Chapter 4 examines a recent compilation of town gas
sites and current handling of gas sites by individual ..States.
-------
1.0 HISTORICAL REVIEW OF THE TOWN GAS INDUSTRY |J
' kt.
^K
'\|£.
1.1 INTRODUCTION ff
This chapter 1s a review of the processes, wastes, geographic trends, and
historical trends of the U.S. town gas industry. The wastes produced,from
different production processes are frequently similar, but substantial differ-
ences in waste types, volumes, and disposal are dependent on the production
method employed. The chapter'ls divided into several sections: 1.2*•<•
-------
TABLE 1. LIST OF MANUFACTURED-GAS PERIODICALS REVIEWED
American Gas Association Proceedings
Pennsylvania Gas Association Proceedings
Proceedings of the Southern Gas Association
Proceedings of the Pacific Gas Association .,.';
Proceedings of the Illinois Gas Association '
Proceedings of the American Gas Institute
Proceedings of the American Gas Light Association Proceedings
Indiana Gas Association Proceedings
Proceedings of the New England Association of Gas Engineers
Gas Age (Gas Age Record)
American Gas Association Monthly V
American Gas Journal
Progressive Age "v
Brown's Directory of Gas Production Plants3 ;
3Data concerning several gas production sites were collected from Brown's
Directory.
-------
1.2 TOWN GAS PRODUCTION
1.2.1 Producer Gas Production "; |gr
"V"-1
Producer gas was not distributed to towns for lighting or heating, but
was used extensively as a fuel gas within gas-manufacturing plants. Produce
gas has a relatively low heating value and very few illuminantsT and it was
only used where the gas was burned near its production location! Producer g
was initially manufactured by burning coal or coke with insufficient air for
'vt-
complete combustion. This produces a flue gas high in carbon monoxide that
was combined with additional air to complete the combustion wherever the hea
• "vi;
is required. The early Siemens gas producer (1861) operated in];this manner.
Steam was later added to the air stream flowing into the coke bed to cool th
bed and to add additional CO and H2 to the producer gas using the two reac-
tions, H20 + C »' H2 + CO and 2^0 + C = C02 + 2H2 . Figure 1 is a diagram o
a producer gas bed and the relevant gas production reactions. The condition
and flows on this figure are only approximate because the actual numbers
depend very highly on the operation of the gas producer. i|| A
Producer gas used either coke, bituminous'cbal, anthracitejfcoal, or cok
coal mixtures for fuel. Producer gas composition varied with the fuel used,
rate of air feed, and amount of steam used. Gas produced from coke or anthr
cite would contain no tar materials whereas some tar would be evolved from
bituminous coal. If the gas were to be burned near the producer, these tars
could be burned with the producer gas. When the gas was transported a short
distance or was burned with orifice-type burners, coke or anthracite coal
would be used to avoid problems of tars condensing in the pipes'and burners.
Any cleanup of the gas prior to combustion was performed with dry scrubbers
(usually filled with woodchips). Additional cleaning was rarely performed
because it would require cooling the gas with loss of the heat and combustib
tars contained in the gas. It is possible to recover tar and ammonia from
producer gas, but this was not widely practiced in the United States (Morgan
1926). The tars (from bituminous fueled production) and the ash from the
producer would be the primary waste products from producer gas manufacturing
Because the gas was burned for industrial uses,^impurities in the gas (H2S,
HCN, C02) were not removed prior to combustion." Table 2 shows the approxi
composition and characteristics of producer gas from bituminousfcoal.
i-Jaffw,
,"•**>•„'tfr*•' .
•;^fe;'^
-------
T
Q

cent
FT pea HOW

CAS SPACE
DECREASE M HEAT/NG VALUE
DISTILLATION 20NE
VOLATILE MATTER OF FUEL
ADDED TOGAS
I
SECONDARY REDUCTION 2ONE
c*co,»tco -mjooaTu
CO+HtO*CO,+Ht -700BTU
PERLS
ATttSQ'F
Figure 1. Producer gas production.

Source: Gas Engineers Handbook, 1934.
IONS OF
PREHEAT
HEAT
ABSORBED

HEAT
t EVOLVED
LBAIR ANO
O.JU STEAM
PEJf LB COAL
«*--***•'
^Wr
-------
r
PAH compounds (mg/kg dry material)
Naphthalene 0.83
6enzo(a)anthracene 1.70
6enzo(a)pyrene 2.10
3,4-Benzofluoranthene 5.10
Benzo(k)fluoranthene 2.20
Chrysene 1.60
8enzo(ghi)perylene 2.00
Oibenzo(a,h)anthracene 0.75
Ideno(l,2,3-cd)pyrene 2.00
Pyrene 2.80
Fluoranthene 6.10
bis(2-ethylhexyl)phthalate 1.70
The high concentration of bound cyanides (and its blue color) identifies
the spent oxides as being produced as a waste from coal carbonization. Table
38 is an analysis of a spent iron oxide, listing the chemicals identified in
the spent oxide. This analysis should not be thought of as typical for a
spent oxide waste because of the very high variation of spent oxides from
plant to plant. This oxide was obviously not mixed with woodchips or other
fluff, and consequently it has a low organic matter content. The presence of
ferrocyanide compounds indicates that the oxide was ustcl to purify coal gas,
and the low tar content and high sulfur content of the oxide indicate a very
efficient tar-removal system was in place prior to the oxide purifiers. A
more typical spent oxide would have a larger tar content, a substantial amount
of organic matter (from woodchips), and a smaller amount of free sulfur.
Table 39 shows the average composition of spent oxides from eight water-gas
plants that operated in Illinois and Indiana in 1921. The tar content of
these oxides ranged between 0.6 and 19.0 percent of the dry spent oxides. The
conversion of the oxides used with CWG was much lower than was the predicted
use of oxides reported in the literature of the time. The predicted
conversions of spent oxide have sulfur concentrations of 50 to 60 percent.
The amount of oxide used by an individual plant to remove hydrogen sul-
fide was proportional to both the amount of gas produced by the plant and to
the hydrogen sulfide content of the gas purified. Table 40 shows the

146
-------
'•;-*, - :-.sL.-«:{
^: frv-i
•*':; ->7

w «•
kX
.S
iiii
•-i •
its!
!»« i •
*
UIM I*
vgh ••h
Buck

tlwt p*
«ich
Un,
leky

«t*>
» •
lkU «»
p*»r
c*k»
fM»l
y>
••
«=4
5l:J5
ii
1
^

|]
ii
; n .
i 'fi »

all L
it--JJ it
8
1
if i
*• "
m
i
]J
;
'
.

'$7
Os'
.*v
-------
.
Producer gas was manufactured for Industrial use and for use with1n''gas-
-"' ••%
manufacturing facilities. Because many installations were at industrial^
plants, there is little available data on the number of producer gas instal-
lations in the United States. An estimate of the number of producer gas!
facilities in the U.S. (about 1921) is in Table 3. This table does not il
• • ••'&''
include gas producers used with the production of town gas. There were many
sji.
different types of machinery for producer gas manufacturing. Production^.
•«'w&
equipment was classified by draft direction (up or down), production pressure
'*«5,.
(suction or positive), feed method (hand or mechanical), poking method (hand
• •• • "*?<•
or mechanical), ash removal (hand, Intermittent, or continuous), cleanliness
of gas produced, and equipment location (attached as part of combustion |;
'•;*: •
equipment or centrally located). Figure 2 is a diagram of a typical gas:;
producer, the Chapman. The body of the producer is stationary, and the bed is
. , v .r - -^-u
poked by a revolving agitator that floats on top of the coal bed. Air and
steam is fed to the bottom of the bed, ash continuously removed from the:!.
jfc
bottom of the bed, and coal continuously fed into the top of the bed. The
producer gas is removed through a pipe near the top of the apparatus. r
In contrast to centrally located producers (like the -Chapman), producer
gas installations at town gasworks were frequently an integral part of the
•:•• ' ' -''X-r •$!'.
machinery with produced coke-oven gas and coal gas. Figure 3 shows a hori-
••"?-•' 0-
zontal retort for coal-gas production with an attached producer gas generator.
• • $'
The producer gas is made in the chamber at the base of the apparatus and the
•';'•*
gas produced combined with secondary air and combusted to heat the six •
horizontal retorts. This apparatus is described further in the coal-gas{;
section (1.2.2.2). .; ' ;;
1.2.2 Coal-Gas Production ' v
—vii
1.2.2.1 Introduction— ||
The discovery that combustible gas could be produced from coal was first
^fl'1
described by Or. John Clayton, who between 1660 and 1670 heated coal and§
" f•"•
described the gas and tar produced. The first practical application of coal
gas was when William Murdock, a Scottish engineer, illuminated his home In
- J£-
1792 with gas from coal distilled in an iron retort. The basic method of£
producing coal gas has remained substantially the same ever since. A bitumin-
ous coal Is placed in a closed vessel that is heated. Tn¥:evolv1ng gaseffare

.-:- 8
-------
TABLE 3. ESTIMATED DISTRIBUTION OF GAS PRODUCERS IN THE
UNITED STATES
Industry
Number of :
producers ;
Steel
Glass
Ceramics and lime burning
Power generation
Metallurigical and other chemical fields
Total :
SOURCE: Chapman, 1921.
6,500
1,500
1,500
1,000
500
ii.ooo J;
-------
JrtVV^\V^X*"-^^^^^"^^fO'^^ * LjM

''I
£S3SSt>3SSi!&&&&g N lu 1 «T

Rgure 2. Chapman gas producer.

Source: Haslam and Russell, 1926.
£ 10
•.vElt. •
-------
§

§
CO
II
-------
I
I?
Si-
then removed and burned for heat or light. The coal remained 1n the vessel
until all of Its volatile materials evolved as gas, then the coke was removed
from the vessel. This section reviews the various apparatus and methods;that
••<-»•"- were used for the production of coal gas. It 1s divided Into two classes of
carbonizing apparatus—retorts and coke ovens. • ^'
1.2.2.2 Coal-Carbonization Retorts-
Coal-carbonization retorts were vessels 1n which bituminous coal was
placed and heated externally to destructively distill volatiles from the coal.
The major features common to coal retorts are (1) a closed vessel containing
coal, (2) a method of heating the vessel, (3) removal of volatlles from the
retort, and (4) methods of filling the retort with coal and removing coke.
Because the requirements for a retort were relatively simple, a wide variety
""* of retorts for coal carbonization were manufactured and used for gas produc-
tion. The major types were horizontal retorts, inclined retorts, and vertical
retorts.
The earliest retorts were essentially cast iron kettles with lids. The
kettles were filled with coal, covered, and heated by a coal fire. Gas,from
the coal was removed through a pipe, cooled, and distributed. Coke was;;
removed from the kettle by hand, after the kettle had cooled. The kettles
'I I were rapidly replaced by horizontal retorts constructed of cast iron. These
I \ retorts were cylindrical or half cylindrical tubes about 7 feet long with one
end sealed. The open end was used for charging with coal, and removing coke.
During gas production, the open end was sealed by a door, and the coal gas
••-•:•'•.' removed through a tube at the door. The retorts were heated by fires below
the tube or by producer gas to between 600 and 800 °C. The cast iron retorts
had a relatively short lifetime (6 to 8 months) and required frequent replace-
ment. Horizontal retorts of clay refractory materials replaced the cast Iron
retorts around 1850.
Horizontal retorts constructed of clay refractory were the major method
of producing coal gas through the start of World War I. They were similar in
**'" construction to the cast iron horizontal retorts, but were larger and carbon-
ized coal at higher temperatures than the cast iron retorts (above 900 *C).
They also had an overage life of 2.5 years, compared to 6 to 8 months for a
cast Iron retort (Hughes and Richards, 1885). A typical horizontal retort is
-.. .$?
'"/;'•
•': '&
12
-------
shown in Figure 4. It consists of six retorts (the MDH shaped objects,
approximately 16 to 26" wide x 18" high x 8 to 20" long) and a producer gas
furnace for heating the retorts. A set of retorts, and their heating
apparatus, 1s called a bench. Benches varied In the number of retorts per
bench but were usually fewer than 10 retorts. The producer gas furnace has
two doors; the small upper door is for charging with either coke from the
retorts or coal, while the lower door is for poking the bed and ash removal.
Primary air and steam is fed to the base of the producer bed. The producer
gas is then burned with secondary air around and between the retorts for heat.
The flue gas then exits through a stack at the rear of the retorts and sent to
a waste heat boiler or exhausted. Each retort has a door and a standpipe that
carries the tar and gas to a hydraulic main (essentially a water seal") above
the bench. Typical operation of the horizontal retorts after starting the
bench consisted of removing coke from the retorts and recharging them with
coal. Periodically, coal or coke was added to the producer below the retorts,
ash was removed from the product', carbon buildup on the inside of the retorts
was removed (scurfed), and the gas standpipes cleaned.
All of the retort operations were originally performed by hand, until
machines for charging coal and discharging coke from horizontal gas retorts
were developed. Figures 5 and 6 show machines for charging and pulling
horizontal gas retorts. These machines usually used doors on each end of the
retorts. :
Several other types of retorts, similar to horizontal retorts, were used
after 1900. They varied in the orientation of the retorts and were either
Inclined or vertical retorts. These two types were further divided into
intermittent retorts (charged and discharged as a batch process) or continuous
(with continuous feeding of coal and removal of coke). Inclined retorts have
the same design as horizontal retorts except the retorts are inclined at about
30°, with doors at each end of the retort. The original concept was to feed
the coal at the top of the retort and remove the coke from the bottom, with
gravity assisting the operation. In actual operation the coke frequently
jammed in the retorts and had to be removed by hand. It was also difficult to
heat the retorts evenly, and few Installations were made in the United States. .
Vertical retorts placed the retorts vertically, with coal fed to the top
of the retort and coke removed from the bottom. They came Into general Tuse
t£-
13
-------
•'5?
J*~Vj:

•'«££;•
«
OJ
ul
-------
i ...."
7 '?-'
JO IT. RETORT
OR BROtWCn CMARCINC MACHINE

," >,.sy
Figure 5. Charging machine.

Source: Morgan, 1926. -,%'{.
15
-------
Rgure 6. Coke pusher.

Source: Morgan, 1926.

16
".. i*\.;.'

•j^-
/•'ryr'4;
\.af'b^.3V,
•ifeew
If*"*"'"
aSl*
»-- 'trt!j?*s»BS
-------
after 1910 and were of two types, Intermittent or continuous. The .
Intermittent retorts were charged at Intervals, the coal coked for about,.12
hours, and then the coke was discharged by gravity Into coke cars below the
retorts. Continuous retorts fed coal and removed coke continuously, with
coking occurring as the coal progressed down the retort. Figure 7 is a .
sectional view of a continuous vertical retort. Coal was continuously fed to
the top of the retorts, was coked as it progressed through the retort, and was
then removed from the retort from the bottom. Producer gas was used to heat
the retorts.
Retorts could be heated by several methods. The most common method was
to use producer gas from coal or coke. The producer gas was manufactured in
either a central or attached apparatus. Combustion air was mixed with the
producer gas, and the gas was burned under the retorts to heat the retorts and
convert the coal to coke. Retorts could also be heated with the coal gas if
excess gas were available, but this was rarely done because the coal gas had a
much higher lighting and heating value than producer gas and was usually sold.
Early retorts were heated by surrounding the retort with a coal furnace. The
_jj
combustion of coal around the retort provided heat for the carbonization of
coal, but it was difficult to heat the retorts evenly or efficiently by'thls
method. Thus, it was rapidly replaced by the use of producer gas. 'f-
The raw coal tar condensed from the coal gas was also used to heat
retorts. The raw coal tar was either dripped into a combustion zone below the
retorts or burned with an atomizing burner similar to those used today with
fuel oil. This allowed some of the benches to be heated with tar instead of
coke or coal and converted what was frequently a waste into a fuel. Raw tar
could also be burned in the steam boiler of the plants. Consequently, the raw
tar always had value as a fuel substitute for coal or coke, in addition to its
chemical values. When raw tar could not be sold at a price greater than its
fuel value, it would be burned by the gas plant. •
The gas composition and wastes from retorts are very similar to those
produced by coke ovens. The gas compositions and byproducts are included with
those of coke ovens in Section 1.2.2.3. T
1.2.2.3 Byproduct Coke Ovens-- §
Byproduct coke ovens were first introduced in the United States 1n;.l892
and eventually displaced the use of retorts and beehive ovens for coke^produc-
17
-------
I
S-
8f

^r^^^';'^^i^'^^^^^^'f^^^^^
- ~&r

—Co*t t'Vxtw 0m*

Kntnuif MjehMH
StMt
Rgure 7. Contlnous vertical retort.

Source: Denig, 1945.
18
*
af>-

'•$•
~*t'
-------
4
,«**•
lion. Prior to the introduction of byproduct coke ovens, most coke was
manufactured in beehive coke ovens that produced coke from bituminous coal,
with no collected byproducts. They did not produce gas for distribution but
were the oldest form of oven for the carbonization cf coal.
The name "beehive" comes from the oven's shape, which is similar to the
old basket beehives. Figure 8 1s a diagram of the beehive coke oven. The
oven was charged with coal through the hole in the top, and the coal was coked
by admitting air through openings in the side door. Volatiles from the coal
were burned within the chamber, providing heat for devolatllizing the layer of
coal on the bottom of the beehive oven. Charges of coal in the beehive ovens
were typically 5 or 6 tons, with coking occurring oven 2 to 3 days. After
coking, the coke was removed from the oven and quenched with water. Beehive
ovens lose all of the volatiles of the coal, either to the air or by combus-
tion, and were inefficient compared to coking methods that recovered these
components. The major advantage to the beehive oven was its ability to pro-
'luce high-grade coke with a minimum of capital Investment. Waste heat ovens
were similar to the beehive ovens but attempted to better utilize the waste
gases rrom the coking chamber, which were collected and burned under the oven
with air for additional heat. The only waste produced by the beehive and
waste heat ovens is the coke quench water, which may have been contaminated
with some of the organics remaining 1n the coked coal.
The substantial waste of heat, combustible gases, tars, ammonia, and
volatiles from the operation of beehive coke ovens led to the development of
coking processes that would produce a high-grade coke, conserve heat, and
recover marketable byproducts. Byproduct coke ovens are basically large hori-
zontal retorts, but with large rectangular coking chambers and more mechanized
movement of coal and coke. Figure 9 shows a typical byproduct coke oven. The
ovens are rectangular chambers that are approximately 40 feet long by 10 to
12 feet and 12 to 20 inches wide. They are charged from the top with coal,
and heated by combustion in flues along the sides of each oven. After the
coal is coked, doors at each end of the oven are opened, and the coke 1s
mechanically pushed from the chamber and quenched with a water spray.
Byproduct coke ovens were constructed for the economical production of
metallurgical coke and recovery of byproducts from the coking process. Exten-
19
-------
I
'; •%.*::T**j*-«'
Sell* m Tetl
4048
Figure 8. Beehive coke oven

Source: Denig, 1945.
20
•!^<
-------

Rgure 9. Byproduct coke oven.
Source: Morgan, 192&
I
21
-------
slve recovery and recycling of waste heat was practiced to reduce fuel con-
sumption for heating the coke ovens. Coke ovens were produced 1n many models
with variations in oven size, flue orientation (horizontal or vertical),
method of heat recovery (recuperative or regenerative), and type of gas used
to heat the ovens.
Byproduct coke ovens could be heated by 35 to 40 percent of the coal gas
produced within their ovens. This left approximately 60 percent of the coal
gas as a surplus that could be sold and distributed to industrial users or
consumers. The coke-oven gas had a heating value of about 560 Btu/ft^ (after
being stripped of light oils) and was readily marketable as a fuel gas. Many
coke ovens produced lower Btu gases (producer gas or blue gas) to heat the
coke ovens, freeing a larger portion of the coal gas for sale. This allowed
coke-oven facilities more flexibility in the quantity of gas they could sell.
In periods of low gas demand, coal gas would be burned to heat the ovens, but
in periods of higher gas demand all coal gas would be sold and the ovens
heated with producer gas or blue gas.
Table 4 shows the gas composition of coal gas produced from byproduct
coke ovens, horizontal retorts, and vertical retorts. These gases are all pro-
duced by the carbonization of bituminous coal and are very similar 1n compo-
sition and heating value. The cleanup processes, byproducts, and wastes from
these coal-carbonization processes are also very similar and are discussed 1n
Sections 1.3 (cleanup processes) and 1.4 (wastes and byproducts). The raw
coal gas was cleaned to remove tar, ammonia, cyanide, and hydrogen sulfide.
The byproducts from these cleanup processes were either sold, used, or
disposed. y
Because many products besides gas were produced from coal carbonization,
there was a substantial overlap between coke-manufacturing companies selling
gas as a byproduct and gas production companies selling coke, ammonia, and tar
as byproducts. Some gas distribution companies purchased coke-oven gas for
distribution but did not manufacture the gas. The distinction between coke
companies and gas companies Is not important from a process standpoint, but it
is an important consideration when determining who will pay for site remedia-
tion (e.g., gas companies were absorbed by current gas distribution companies,
while coke producers remained as a separate Industry). %
22
-------
•"•'••"^^'"•^^^"^

<
**
_J
u
a
••^
£

&
M
jtj
5
g
g
e
1
£
fr*
M

^O
f^
•
•»

m
«>
^
3
a
c
&
o
I.
4»
2

e
c
01
£

1
•
c
•
X
4>
u

2
•
1
si
fg
• 9
o|
1
X
o
1
Is
— c
— •
»HC
e
IA IO «
P4 **4 CD
« « IA

oj M «r
• • •
M n m

a IA a
«•« N W
IA IA IA

a
•* i i
1 1

«H a «o
• o •
m IH M
m ct tn
IA a m
• • •
u> o •>
M M M
a a a
-
«o « r-
• • •
<•> IA «l

a IA a
• • •
CM •-> t*

- 5 i
• • CM >
y <» o 4» o
— b H (.
.1! .11 .1
S S 3
U U 0 |

•
^b
s
01
r4
g
?
O
a

••
5
M.
f
23
-------

' ' — *— — — ••••••••^^^••••^••••••••jmMHBLjBMHa - _ - r^-kt'J^..i,L _
1.2.3 Carbureted Water Gas
Blue gas is prepared by passing steam through a bed of incandescent car-
bon. The steam reacts with the carbon to produce a fuel gas composed pri-
marily of carbon monoxide and hydrogen. This gas is also known as water gas
or blue water gas. When liquid hydrocarbons are thermally cracked into the
water gas, a fuel gas known as carbureted water gas (CWG) is produced. Blue
gas was sometimes produced as an Industrial fuel but was not distributed to
consumers because of its low fuel value (about 300 Btu/ft^) and lack of illum-
inants (bright-burning hydrocarbons). The shortcomings of blue gas were over-
come by the thermal cracking of liquid hydrocarbons into the gas to produce
carbureted water gas. This both increased the heating value of the gas and
its illuminating power. CWG was a very good fuel gas and was widely produced
and distributed to consumers.
The discovery of blue gas 1s attributed to Fontana In 1780. He passed
steam over Incandescent carbon and produced a flammable gas. Blue gas was
only rarely produced until Lowe's Invention of carbureted water gas in 1875.
Liquid hydrocarbons were sprayed into the blue gas (carbureted) and thermally
cracked to form gases and tars. Carbureted water gas became the predominant
form of gas production In the United States and was produced until the demise
of the manufactured-gas industry. The production of carbureted water gas was
economically possible because of the growth of the U.S. petroleum industry
after the 1880's. The petroleum industry supplied the inexpensive hydrocarbon
feedstocks required for the production of carbureted water gas. The availa-
bility of cheap petroleum-based feedstocks for gas production created a gas
Industry based on oil instead of coal. The gas industries of Great Britain
and Europe did not have cheap oil products and subsequently did not adopt oil-
gas and carbureted water-gas production to the same extent as did those in the
United States. v-
Figure 10 is a diagram of a blue-gas generator. Blue gas is produced in
a cyclical manner: (1) air is blown through the bed, burning coke and heating
the bed; (2) the air is cut off, and steam is blown through the bed, producing
24
-------
^*$^-#8#*iigi^^
Source: Morgan. 1945.
A • Generate*
B - GasofMakeor
hydrogen pipe.
C - Stack.
0 - Wash-box, or seal
separatee
E - Hot main connection.
F » Blast connection.
G • Steam connection.
H • Exptostondooc
Coaling door.
Cllnkering doors.
Bottom gas off-take.
Heat valves.
Dust catcher.
Stack valve.
Seal pot or drain tank.
Controls.
Instrument board.
Rgure 10. Blue gas generator.
; 25
-------
*
blue gas and cooling the bed; then (3) the cycle 1s repeated. This 1s the
simplest cycle that can be used to produce blue gas, .and variations of this
cycle were employed to Improve the production of gas. During the "runs" (with
steam), carbon monoxide and hydrogen are produced, principally from the water
gas shift reaction:
H20 + C • CO + H2 .
This reaction Is endothermlc and rapidly cools the coke bed. When the bed has
cooled, the steam is stopped and air is blown through the bed ("blow") to
reheat the coke. The cyclical process is made more heat efficient by recover-
ing heat in the flue gases during the blow and by preheating the air used In
the blow. During the blow, the coke bed tends to form carbon monoxide from
incomplete combustion. This gas was similar to producer gas and could fre-
quently be burned when additional combustion air was added. A complete set of
the reactions occurring during the blow-and-make periods of blue-gas pro-
duction is given In Table 5.
A blue-gas producer Is the front third of apparatus used to produce car-
bureted water gas. Figure 11 shows a three-shell water-gas set. The first
shell is a blue-gas generator, and the second shell (carburetor) and a third
shell (superheater) are attached to it. The carburetor and superheater are
checkerbrlcked with firebricks. The bricks are arranged so that a large sur-
face area of the bricks Is exposed to gases flowing through the shell, but
with a relatively low pressure drop.
This apparatus was also operated in a cyclical manner, with alternate
blows to heat the coke bed and the checkerbrick, followed by runs In which
blue gas was produced and hydrocarbons cracked Into the gas from oils sprayed
onto the hot firebrick of the carburetor. The blow and run parts of the cycle
are described below and Illustrated In Figures 12, 13, and 14.
Blow; A1r is blown through the coke bed to heat the bed. Air
enters from the bottom of the bed and flows upward through the coke.
Air 1s admitted to the top of the carburetor, then 1t burns carbon
monoxide 1n the gas from the generator, supplying additional heat
for the checkerbricks. The gases flow downward through the carbure-
tor, then upward through the superheater, exiting from the top of
the superheater and flowing to a waste heat boiler/ If
26
-------
",if
-;$?
W(».
§—
•
•3T
•>1C
•"jt«,
:;a

TABLE 5. REACTIONS DURING BLUE GAS MANUFACTURE
Blow: f
: 02 + C = C02
;', 02 + 2C * 2CO
C02 + C = 2CO
2CO + 02 - 2C02

Make:
. H20 + C = H2 + CO
^2H20'+ C = 2H2 H
ti H20 +;CO = H2 * C02
'% C « 2CO
27
M1.'.
• »r
••'$,,
•*
,V^fe»'
' Sj«*;';
'irtfe'
f
-------
:-
MM«BM*mnrfMIAMI in* *»••. — -i_.iTi-i~r f-'-f"-]-•••fliASM i Til r m I"«M«*
MMMto*
CLCVATIOM • StCTION *-A
Rgure 11. Typical three-shell water-gas set.

Source: Morgan, 1945.

a
'd
28
-------
ii*^^'ug;^^afMy»*'^^'*^-J!°^t»»iM^>^"^wai'».i«^t»ey«>'yi«»i
Combustion Products to
Waste Heat Boiler
Superheater
Blow, To Heat Apparatus
:«&..
Air
Coke
Bed
Checker-
brick
CO Rich Gas
to Washbox
Checker-
brick
Blow-Run
Figure 12. Blow and blow-run.
i
:M
m
'". •

-------
X-
it»
Steam
Up Run

Product Gas
to Washbox
,.3jf
Steam
Down Run
Figure 13. Up and down runs.

' -' "3t-.: 30

Product Gas
to Washbox
-------
fl

Air
Product Gas
to Washbox
Figure 14. Air purge.
31
•YsW-i
•^
-------
Blow-run; This part of the cycle collects the carbon-monoxide-rich
gas from the generator bed and adds it to the product gas. The air
flow is the same as that during the blow except no air is added at
the top of the carburetor and the gases are routed through the wash-
box to the gas mains.

Up-run: During the up-run, steam is admitted to the base of the
generator, flows upward through the bed of incandescent coke (form-
ing blue gas), out the top of the generator to the top of the carbu-
retor (where oil is sprayed into the gas and onto the checkerbrick,
cracking the hydrocarbons), down through the carburetor, upward
through the superheater (where additional cracking of the hydrocar-
bons occur), and out through the top of the superheater and washbox
to the gas mains. During the up-run, the bottom of the coke bed
cools faster than does the top.

Down-run; The down-run (or back-run) is identical to the up-run
except that steam is introduced at the top of the generator bed,
flows down through the bed, and then to the top of the carburetor.
The top of the bed is cooled during the down-run, maintaining a hot
area in the center of the bed. More efficient operation of the
generator is obtained with split runs (up and down) than if the
entire run were performed 1n the same direction.

Air purge; The air purge actually starts the blow, but gas from the
superheater is sent to the gas mains. This purges the apparatus of
higher Btu gases and recovers them as part of the product gas.

Table 6 shows some typical compositions of blue gas and ca^b'jreted water

gas. The carbureted water-gas process was used to produce gases of widely
varying Btu contents. This was accomplished by varying the amount of oil

cracked into the blue gas. The specific heating value of carbureted water gas
produced by individual companies was determined by economic considerations,
but it was usually set between 500 and 600 Btu/ft^. Higher Btu-carbureted
water gas could be mixed with lower grades of gas (producer, blue gas, or coal
gas) to produce a mixed gas for distribution. This had the net effect of
increasing gas production capacity without increasing the number of water- gas
sets used to produce the gas. The highest Btu-carbureted water gas could be
mixed with natural gas without reducing the heating value of natural gas (both
natural gas and high Btu-carbureted water gas have heating values of about
1,000 Btu/ft^). The higher heating value comes from increased use of
carburetion oils, increasing the cost of the gas.

A variety of feedstocks were used in the production of carbureted water
gas, and these raw materials varied both with time and location of individual

32
-------
I

f-
(X
I

'tt

I- • i
I

fc
I
sooMO>f-i
Uk n <4
asi-ieDe>« t to
t-« N C4 M 1
S
mm **
o* f» i i 10 i
M tH I I I
aaa^-rc^ar^ci ft «

amaanoi^a^c! i i «n a
«r (4 14 <4 «•> d I I tH
10
r~ at

h. I V «•«
• • o.
4> 4> *. & O *>
^ •-s. o
• .
a, •
~ee
i &J
3=
oaou.MUMou
a.ZM
10

o>
3
M

L
u
I.

O
4*
' Ct

is
• a. •
m q _
i« *> a

x f —

2 So
e •<> •

.8 *"
S C4>

•S 15
4* •«-
e e
3 - - o

a -5 = 3S
o-o.
J .5
33
:;fe;
•*£_

-------

gas plants. Two types of raw material are required for carbureted water-gas
production: (1) a solid carbon material for the generator and (2) a liquid
hydrocarbon for the carburetor. Because several petroleum fractions and
sources of carbon could be used, the specific feedstocks employed at Individ-
ual plants were selected based on economic factors.
The original carbon materials used In CWG production were anthracite coal
or coke from bituminous coal. Both were considered Ideal generator fuels
because they had very high carbon and low volatile contents. Consequently,
they were very clean fuels to use for blue gas, producing hot fires and little
smoke during blows. Increasing prices for anthracite coal after the turn of
the century and shortages of coke during World War I encouraged modifications
of the standard carbureted water-gas process to allow the use of bituminous
coals in the generator. Because bituminous coal was cheaper than coke, many
plants replaced the use of coke or anthracite by bituminous coals after the
war. The conversion to bituminous coals was not universal because some plants
had coking facilities onsite and some difficulties occurred with the conver-
sion. The use of bituminous coals reduced the gas production capacity of
carbureted water-gas apparatus, entrainment of coal from the generator into
the carburetor occurred, and smoke was frequently produced during the air
blows of the gas cycles. Some of the problems were reduced by modifying the
operation of the sets, primarily through the "pier" process and the use of
reversed air blasts through the carburetor and superheater during blows.
Table 7 presents an analysis of fuels frequently used 1n the generator
for the production of blue gas and water gas. The use of raw bituminous coal
instead of coke or anthracite Introduced some coal constituents into the tars
and waste liquids of CWG plants. Coke and anthracite coals have very low
volatile contents, and tar acids (phenols), tar bases, and cyanides were pro-
duced in only trace amounts from CWG when these generator fuels were used.
When bituminous coal was used, the coal actually coked within the generator,
releasing coal gas and volatile constituents into the product gas. About
58 percent of the coal gas from the bituminous coal was added to the carbu-
reted water gas, while the remainder was burned during the blows (Mufdock,
•vV,
1926). About 8 percent of the final product gas was from coal gas,'and the
••?>-:
amount of tar acids, tar bases, and cyanides produced would also be about
34
-------
off
4ft

S
^
III
i
u
§

s
CD
&

i
<
K
£
UL
£
&
«/»
i
•
h>
H

•
•
•
.a
>
L
T>
e
«
•
ft
^
^
•
5

— •
•Jl
^
F
ev
o
|"gc
""a

3.1
o&o
L
3
M

.c
<

TJO
•f
K t»
~
~ u
3 2

3
• v •.-•S.r
^F-- v: «- •.-;;>
•:p:> ''''W ' '• ''
\*:.' " "'3'

'
2 «
«O II 1 1 1 1 1 1 O
Mr4 II 1 1 1 1 1 1 *»

I : : i i s i ! s i
B « a
ei

1 MB O r« 3" B 01 B 1
I «r m ^iBBiAOtai
B«0 r«BB*4tDKI
rt M «N « l«» •< tH T

2 22 22** 2?S
Ot A 00 h»ttv4 4OIAB

8fl f*» f^ *^* *^ ^ Cft P^
B«l ft o» T «•» u> r- i-i
Ot AB BOB h> <-i
«•> r< T KOtNBOtCBO
«Kh>WIA
N nv ft 9 ft ea «0 * «
« IAIO iH »<
e — o je •
5 c « Jl ° 8 - .
s £ .sg : . s u
i 2 , c . g S w 8
a.- ^ Slf-5x2S.
.8 . "-8 S 8 • * 5 8
30 . . — o •*— e»
4>4>a 4» • ei • —
.. c — — 4»*> u • 9 * *
•t n u 1 r i . t |
5i ££?ta3ii§~
4 ^4£ ^«lii£
',-';.i'^^^p. • ' -
_'J 'f'^yf'
.'"*f?^l"-
.'•^s§
•ip»4j:-i"
•/"'.•,
t
2* i X*H
9 • 1
«— • •
" *S.

i i \r i

s ; s
**• " **•
*","-'• **
•.-',•.,•:,
» ft ''• IA
IA m m
B 11 B

ci ft- n
ft M «
14 B 0>

f« il 14
t-i r« N
ot n r-
IA IA m
r» co •»
r- B v
h- jo m

5 S S
i-« f* ca

^J
Elkhom gaa coal
Franklin .County, IL
Wtatco* County, WA,
Sub-bltuMinous co

}

§
uT

(
•

m
B

f.
i

s
ot

a
HI

Portland iMpblock
brlquottos
35
-------
8 percent (per volume of gas manufactured) of that produced from coal carboni-
zation.
Any liquid hydrocarbon that could be thermally cracked Into gaseous,
liquid, and solid products could be used 1n the production of carbureted water
gas. This included many of the distilled fractions of petroleum oils, but 1n
practice gas-manufacturing companies used Inexpensive oil fractions that had
only limited alternative markets. As the petroleum Industry changed between
1880 and the 1950's, the gas industry modified its use of petroleum products.
Table 8 shows the general classes of products distilled from petroleum. The
fractions are listed in the order of distillation temperatures, with the lower
boiling fractions at the top and the higher boiling fractions toward the bot-
tom of the table. Although any of the fractions could be used in the produc-
tion of carbureted water gas, three fractions were principally used. These
were naphtha, gas oil, and residual fuel oil. Crude oil and "topped" crude
oils were also used to a more limited extent.
These fractions each have different distributions of hydrocarbons, and
the specific composition of any carburetion oil was dependent on both the
source of the original crude oil and Us processing during distillation. The
carburetion oils differed substantially 1n their compositions, which 1n turn
influenced the amount of byproducts from the process and the character of the
byproducts.
The early carbureted water-gas processes used naphtha fractions of
petroleum as the carburetion oil. Naphtha was rich in short-chain alphatic
and light aromatic hydrocarbons. It vaporized readily in the carburetor and
superheater, with almost all of the naphtha cracking to gaseous hydrocarbons.
Tar produced from carbureted water gas using naphtha was 1.7 to 3.5 percent of
the original naphtha (McKay, 1901). The early oil refiners produced
principally lamp oil (kerosene) and lubricating oils. The naphtha fraction
(during this period) was the liquid hydrocarbon fraction that boiled at
temperatures above gaseous hydrocarbons and below the kerosene fraction.
There was little demand for the naphtha fractions until the Invention and use
of internal combustion engines. The gas Industry used naphtha for the
carburetion of water gas and the enrichment of coal gas from about 1880
through World War I, when other uses of naphtha Increased the price of this
petroleum fraction. As the price of naphtha increased, gas manufacturers
36
-------
TABLE 8. GENERAL CLASSES OF PETROLEUM PRODUCTS

Fm QA* - -.

, '

HYDROCARBON •

1 BASIS

Nininui -
----- 1".~-^.
-| . — -.
n« UK MUM
, INTERUtDIATC

[ OISTILLAT13 1

1C i ii *
MM«4wpnt P

Bsrastf
f TBrmnnu.
HurrOiu
CRUOC

rent OLEUM
HKAVV

DISTILLATfM
LMMCATIM «U
sas—
bS
•mm fr~*« on
ts-'a^-,-^
HFSIDITJ*
KnxlH-M. Fl n. lln J J
M.VDGM
Source: Biggs and Woolrich.1925.
VA.

•,*<*
-------
-.I1?

VJ;
--""3'
. - •#- t • "• -.-».** -,«&• -*.,. ,-
••«»'• '-& - •;- •<*> ^-;' ' ,
•v* ••.?•• *Mv" ':
'.vii* • , it' ' •: • •:•-• it-
>,.<-,- - ,» . 1- 'i1 . ,;,,5T
•5 7 -;«.r
"4 •••-".?:, ".??
switched to other oils, and the use of naphtha ended altogether about 1930
(Oashiell, 1944). Table 9 shows a typical distillation curve for a naphtha
-• - •" ?•;, *">£•
fraction, used for carbureted water- gas production in 1897. *?
Although naphtha was the preferred fraction for the production of carbu-
reted water gas, a fraction boiling between kerosene and lubricating oils was
increasingly used after about 1895. This fraction came to be known as gas oil
and was a more viscous and heavier petroleum fraction than was naphtha. It
also produced more tars in cracking, 12 to 18 percent by volume of the origi-
nal carburetion oils (McKay, 1901). Table 10 presents a distillation curve
for a typical gas oil used for the production of carbureted water gas 1n 1897.
This fraction was the predominant carburetion oil until Increased demand for
gasoline and the Invention of catalytic cracking of the gas-oil fraction Into
gasoline and residual fuel oil (the heavy residue left from the cracking proc-
ess). The use of gas oil as a cracking stock for gasoline meant that the
price and availability of gas oil was linked to the price and demand of gaso-
line. Gas-o1l supplies became more expensive and less available as the demand
for gasoline increased. The gas Industry began to switch from gas oil to fuel
oils around 1930. The great variability of oils used for the manufacture of
carbureted water gas is shown In Table 11. Each oil sample was analyzed and
divided into four constituents: aromatics, olefins, parafins, and naphthenes.
A similar study of 50 gas-making oils showed the following ranges of proper-
ties (Kugel, 1947): f
Specific gravity (60 *F) 1.049 to 0.754
Viscosity (100 *F) 27 to 288 S.S.
Flash point Below 62 to 75 °F ;J
Pour point Trace to 14.0 percent
Sulfur 0 to 3.7 percent. |v
As the price of fuel oils Increased during the late 1940's, some facili-
ties switched to heavier fuels oils, such as residual oils with h1gh*carbon
contents. With fuel oils and heavy residual oils, the tar byproducts from the
carbureted water-gas process Increased to up to 25 percent of the original oil
38
--4V
-------
••• i
' "i
TABLE 9. NAPHTHA DISTILLATION
•-*-. Naphtha

Specific gravity » 0.6930, or 72° Baume at 60 *F (color white)
Fraction (°F)
100-150
150-200
200-250
250-300
Above 300
Residue

Vol. (t)
10.90
54.09
28.00
4.20
1.75
Wt. (%)
10.34
53.69
29.07
V 4'45
" 1.91
Sp. Gr.
0.6579
0.6885
0.7196
0.7370
0.7560
Beaume
: 83°
73.5
64.5
60
55
Color
White
White
White
White
Clear red-
brown
None appreciable ;

98.94
•
99.46

•.&•
••»:
''-.•I'.!
-uv'; :
SOURCE: McKay, 1901.
39
•f«f*

.£:

-------
TABLE 10. GAS-OIL DISTILLATION
-•?*• •
>'•"•'
Specific gravity •

Fraction (*F) Vol. (I)

180-300
300-350
350-400
400-450
450-500
500-550
550-600
600-650
650-700
Above 700
Residue
,. .
0.8462, or 35' 0.3 Baume at 60 «F

• (1) ;• sp. Gr. Beaume
4.40
4.55
3.50
6.30
6.95
10.45
16.35
21.95
18.35
7.50
3.83
4.10
3.24
5.95
6.73
10.27
16.46
22.17
18.99
7.98
0.30
0.7369
0.7639
0.7823
0.8001
0.8108
0.8320
0.8491
0.8625
0.8764
0.9009
solid
|>3.5
^ •*
;v49
45
40.75
38.5
35
32.25
29.75
25.5
100.30
SOURCE: McKay, 1901.
100.02
40
J .
Nearly white :"
Nearly white
Nearly white
Slightly yellow
Pale :yellow
Pale yellow
Pale yellow
Yellow
Dark
Black'

•S3
:l%
.••$3
'' ,i-M
-------
OT
o
o
g
EC
3
m
cc
s

*
%
&*
MS
8 I
s = : 2 82288 s 2
S i S 2 S9S55 2 i
?o 99 9SS9S " 9
1 I I IP|| 3. I
85 5S 333S2 = S
si s a 2 8
- •.
S S 893
no -
«« »«N »
MM wr-tn M
«••
xa'a?1
•'-;/•'•.,, S C

33 SS2 S
oo '£ '
zz z'

MI r«*
*i
r«f
iiif r
Na
wli
nker
1 P* *
* aSS 2
S
Dashie
-------
f
I

fed to the process. Larger plants, which consumed large quantities of oil,
changed to less expensive oil types faster than did the smaller plants.
Changes in oil type were accompanied by changes in the production apparatus
and operating procedures of the plant, and these costs were better absorbed by
larger production plants. - «
J, ' - \ " ' " '
The major byproduct from the production of carbureted water gas was the
uncracked portion of the liquid hydrocarbons fed to the carburetor. This tar
was produced in varying amounts from the process, and both the amount of tar
produced and its characteristics were dependent on the original hydrocarbon
feed material and the operation of the gas apparatus. These tars contain many
of the compounds that are present in coal tar, but they contain no tar acids
(phenolic compounds) and only traces of coal nitrogen compounds. The use of
bituminous coals as a generator fuel increased the amount of these compounds
in the water-gas tar, but they are -still present in relatively small, amounts
when compared to coal tars. Because of the generally small nitrogen content
of coke and petroleum products, very small amounts of ammonia and cyanide
appeared in the gas from carbureted water-gas operations, and this is reflect-
ed by low concentrations of these compounds in byproducts. :
1.2.4 Oil -Gas Production
1.2.4.1 Introduction—
The production of carbureted water gas required only two raw materials,
carbon and an oil. Transporting coal or coke to certain areas of the United
States (mainly the Pacific Coast) was expensive for the gas companies. States
along the Pacific Coast had ample supplies of inexpensive oil products after
1890, but coal materials had to be transported from the East. This led to
modifications of the water-gas process that eventually eliminated the need for
coal or coke in the generator. The production of gas from oil was Invented in
England in 1815, and gas from whale oil was produced in some U.S. cities in
the early 1800's (see Section 1.2.5). It was L. P. Lowe, the son ofithe .
•>• . . j^
inventor of carbureted water gas, who invented an oil -gas process using
•!•> •
refractory material in 1889. Ten years passed before the first "modern" oil-
gas plant was constructed in California, and it was 1902 when an oil-gas plant
was installed in Oakland, California, for lighting purposes. S._
42
-------
vim
T},

."•&.
- :In addition to the product!orfrprbcesses used on^thejpacific Coast/^sev-
• '"^fej'' -•* i"--<4&; '-" -'" 'ASaBSf1' ,™t- ~:
eral other oil-gas processes ''we^^Sedlwr thetproductio^gf gas. When
natural gas became available tb^^ufactured-gas areas";|fpMne.carbureted!
gas facilities were converted tbWlgh Btii oil-gas production. The high"
™ , :- -,, -. ..uftJs£ta.'c •*,.-Bac. j*,-,^j»" . . ** . • „.- ,,— i^iik.iijawia-.^ , . - ••$
, and it could be!
oil gas had a heating value veryfclose to that of natura
' ' -'** -"*J
either mixed with natural gas'td^pe^^ads'^substltirtig^for^naturaTigas
when natural gas supplies were interrupted. Other relatively minor oiffgas
•- "'-'iSS>~-s-' "«•'• fr-*1*- ' ••'"" "••'•-y- ... _ ,^_ _ . . L.-«3rfifc.i_»aw.,., . .'.;.'
^sections; the first covers the
describes the high Btu oil-gas
>s,v^
rrt m * \R • '•" •- -a^TT*
•oIKgas prpce
ThY other'"in
v . . < • ' >' .I,!*
it - -."-.' «4tW -S-SSS
us gas produc
es.r'and.the second
oil-gas
methods are Included with the m
• ;... V"A?ft
of the section.
IvS"^;
1.2.4.2 Pacific CoastOil -GaY^lScesseT--
Jhe major oil-gas process used Jn-fthe United States
oil-gas process that was based on',tne;gasification of
ing them through a chamber of he^ed^checkerbricks.^'TKe
'•V-^SSSfeS?^^!^-11 • •- '-'"iiCV 'Vt?";
with alternate heating and gas-manufacturing parts of
. •;•;„,.. ,, •* -^v :
for the Pacific Coast oi1-gas process"were of;two
ictlonl
methods at^tlhTSndl
apparatus and the two-shell
Figure 15
ratusf
.<&*&ii''?$fi$&&:" ' .;•1.vvitAv >, , • A.^AiBgg^j, •'•.• ,t.--•
for heating (and oil and steam for gas manufacture) at^the;top of
tus, with stack and product'ga^inoyed:>rom the basejsft^e apparatus-
type of process was used in
'cost^"''r;*""7
;The two-shell apparatus
better utilization of
shell designs.
, ,
plants because it
enerators
,„• - ••wi-l^
nfthe generator
This design"pVrWIs'moife:efficient
:'-"* -i'i'
by using primary air and seconda
. » , ._',;^- * . T > * . -v,-^ t- . ,, a -i,-4,tt*;®(flri
,the,!generato
'•> •, SW2.** r.^l ' , f ,.rfJ
-------
and Oil
5 pray
Fire BrFck
Stack
Valv*
Combustion
Chamber
flake OH and
Steam Sprays
Asbestos
Lining —v
Checker*
drick
Chacher
fir/c/f
TTTTI
Gaa
Off-fake
*-Co/Tibastion Chambers'
15. Single-shell oil-gas apparatus.
Source: Morgan. 192&
-------
— OIL GAS PROCESS COMPANY -
itXH
-------

Table 12 shows representative operating cycles for each type of oil-gas

apparatus. Each part of the cycle is described below.

Blow with air: Air is blown through the apparatus to burn off car-
bon on the checkerbrick, heating the apparatus. The combustion
products are vented to the stack or a waste heat boiler.

Heat with air and oil: Air and oil are turned on and sprayed into
the generator. The combustion of the oil heats the checkerbrick to
a temperature of about 1,600 °F.

Hake with steam: The air and and oil are shut off and steam is fed
to the apparatus. The steam cools the bricks somewhat as it reacts
with carbon on the bricks to form blue gas. The blue gas is sent
through the washbox and mixed with product gas. The cooling of the
bricks is desirable because high temperatures cause excessive carbon
formation, and the highest brick temperatures occur at the combus-
tion chamber. This part of the cycle is used with the single-shell
heat and make-down apparatus and the two-shell apparatus. It was"
generally not employed with the straight shot apparatus. .

Make with oil and steam; After the apparatus is heated, the high-
pressure steam is used to atomize the oil. The stack valve is shut,
and the valve to the gas take off 1s opened so that the product
gases will be collected. The atomized oil and steam becomes a mix-
ture of gaseous hydrocarbons, fixed gases, tar, and lampblack. The
gas mixture leaves the generator and flows through a water-sealed
washer. . v

Steam purge; After the make cycle, the apparatus 1s purged with .
steam to remove combustible gases from the apparatus. The gases
from the purge are mixed with the product gas. The purge is neces-
sary to prevent the formation of flammable mixture within the appa-
ratus when air is admitted as the next step of the cycle.

Table 13 shows a comparison of operating data for five oil-gas facili-
ties. The two-shell apparatus (Jones) results are from two plants, and the
single-shell straight shot apparatus results are from three plants. All of

the plants were located in California. The two-shell process produced sub-
stantially less lampblack than did the straight shot process (12.5 vs. 21.2
Ibs/Mcf gas), while the straight shot process produced less tar (4.3 vs. 1.8
Ibs/Mcf gas). .. v,.

The major byproducts from the oil-gas process are lampblack, tar,; and

light oil. As in the carbureted water-gas process, only very small amounts of

ammonia, cyanides, tar bases, and tar adds (phenols, creysols) are produced.

The major difference between the byproducts from oil-gas manufacture and those

\ 46 .'.-. .:-
-------
TABLE 12. OPERATING CYCLES OF OIL-GAS PRODUCERS
Heat up
make down
Heat dowtif:
make down^
Two--
Blow with air (min)
Heat with oil + air (min)

Total heating (min)
s

10
Make with steam (min)
Make with oil + steam (min)
Purge with steam (min)

Total making period (min)
Ov
16
8_
''•TV-
,'. 242 -
i , si? i&v'
2 -''s^y?
6 • V::'v
2* .- '
"'•; ' '
"f£$$*'i
u>m
i
7
2

10
<*$••
';v
•H
is™.
, Ajv
SOURCE: Morgan, 1926.
•\£j
'I®
-------
I
TABLE 13. COMPARISON OF FIVE OIL-GAS PLANTS

Weight in Pound*'
Ilema
W«Uht of Materlali late Generator
Make period
Make oil
Steam

Touli
lllul period
Air

Hot period
Air
ou..
Steam

W«i(b< of Material* Ml of Generator
Make period

Tat- .

TouU
t»Mt ptfiod

Tout*
KM. period

TotaU

Difftranc* (In eluding carbon dcpnuftcd en
brick*. If .8. n*phth*l«i«. and IOMM)

Jonea
Potrcra
S0.4M
24.80
«. 40
3.30
S4.M
111.00
l.M
113.50
97.00
«.M
8.M
tot.n
308. 24
M.n
12.00
4.00
22.40
T4.IT
112.00
l.M
113.50
W.OO
13.05
107.05
9H.72
1.1.54
4.4
San Joe«
51.03
23.50
3.50
5.30
83.33
44.00
1.20
45. JO
120.00
T.30
1.00
141.30
2M.S3
37.00
13.00
4.50
IT. 50
T2.1A
43.00
1.20
43.20
122.30
12.33
134. £3
240. W
!«.«
7.3
8ln«le-ititU up blut
SaaU
nari»n
M.Sl
14.20
2.00
1.40
80.11
88.00
0.80
88.80
140.00
I.OS
4.00
152.08
330.M
».77
23.00
2.50
12.80
67.07
88.00
0.80
M.80
139.00
11.18
140.79
194.83
28.3(1
ft.2
Southern
California
Cae
Company
57.94
12.20
7.60
5.30
83.04
48.00
0.80
48.HO
119.00
S.23
3.70
128.93
25S.77
31.80
19.00
1.50
14.70
87.00
48.00
0.80
48.80
114.00
7.83
121.83
238.23
20.54
7.9
La Anceloi
Qae and
Beetrid
Corp.
59.08
10.50
8.00
3.80
79.38
78.00
0.90
78.90
153.00
8.07
3.00
104.07 ,
320.33
33.25
21.70
1.50
9.00 ,
45.45
f
77.00
0.90
77.90
149.40
10.35
159.75
103.10
17.23
5.4
' From Final Report of th« Joint Committee on Efficiency ana] Eeonomr of Oae of the R. R.
Cummiaeion of the Hlala of California.
Source: Gas Engineers Handbook, 1934.
48
-------
from water-gas manufacture is the large amount of lampblack (petroleum coke)
produced from the oil-gas process. This lampblack was deposited 1n the wash-
boxes or scrubbers of the plant and was disposed by burning, brlquettlng and
sale, or dumping. The generators also required frequent rebuilding when they
became clogged with carbon.
The fuels used for the Pacific Coast oil-gas process came principally
from the oil fields in southern California or from the processing of the Cali-
fornia crudes. This crude oil had an asphaltic base Instead of the paraf-
finic-based crudes of Pennsylvania. The raw crude oil was used directly for
oil-gas production until about 1919, when "topped" crudes or residual oils
started replacing the raw crude oil. Topped crude oils were those in which
the more volatile and valuable fractions were distilled from the crude,
leaving a residual fuel of higher boiling components and a high carbon con-
tent. Rather than continuing to distill the residue to heavy asphalts and
coke, the refiners sold the residue to gas companies, which used it for the
manufacture of oil gas. Table 14 shows the distillation curves for a typical
California crude and a refinery residuum. The crude oil would have been a
much better feedstock for the manufacture of oil gas, 1n that the lower boil-
ing components would be readily cracked Into the gas, while the residuum would
produce much larger quantities of lampblack and require more oil to produce
gas comparable to that produced from the crude oil. Because the residuum was
less expensive than the crude oil, gas manufacturers preferred the use of
residuum oi1.
1.2.4.3 High Btu Oil-Gas Processes—
The introduction of natural gas to areas previously served by manufac-
tured gas brought substantial changes in the operations of the manufactured-
gas companies. Initially, the gas pipelines Installed in manufactured-gas
regions were for base capacity. The purchasing gas company was required to
buy a fixed amount of natural gas from the pipeline, with financial penalties
for using more gas than originally contracted. Consequently, natural gas was
purchased for the base load of the gas company, i.e., the amount of gas used
everyday by the gas consumers. The gas company then had to provide whatever
additional gas was required to meet peak demands of the population they
served. This meant either storing large quantities of natural gas to smooth
49

-------
I
TABLE 14. DISTILLATION OF OILS COMMONLY USED FOR OIL GAS
Temperature (8C)
Up to 150
150-200
200-250
250-300
Total to 300
300-332
300-350
350-400
400-407
Residue
Gravity, Baume at 60° F
Percent
Crude oil
8.5
13.5
15.0
20.0
57.0
35.7
—
*•••
•••v.
7.3
18.4
of distillate by volume
'•Sgf Residuum %
;•.."•" -- ;v';
""•:,, 1.3 :%
'•$$$' • 5.7 'if i;
.""' 11.9 '^
-•-,-. 18'9
'•rtii " '^'
15.1 "}'•
• »..l- 35.2 ,- .
•r;:>;. 22.4 ,^
';4,' 8.4 ;i
18.0
SOURCE: Morgan, 1926.
.

,1
50
«i
•:$ff'
*..
.%, -
-------
r
II
out the peaks and valleys of demand or manufacturing whatever gas was required
in excess of the purchased natural gas.
When gas companies switched from manufactured gas (Btu content of approx-
imately 550 Btu/ft3) to natural gas (Btu content of about 1,000 Btu/ft3), vir-
tually every gas appliance had to be readjusted for the higher Btu fuel. Only
two types of gas had heating values in the same range of natural gas and could
be successfully mixed with natural gas in peak demand periods—a high Btu oil
gas (approximately 1,000 Btu/ft3) and LP gas. LP gas was the the distilled
petroleum fraction that is a gas at atmospheric pressures and temperatures,
but it could be stored as a liquid under pressure and was vaporized into the
gas distribution system when needed. It contributed no byproducts or wastes at
sites where the process was used.
Existing apparatus for the production of carbureted water gas were fre-
quently converted for the production of high Btu oil gas. This allowed the
gas companies to produce a manufactured gas for mixing with natural gas during
peak loads and a plant that could provide manufactured gas whenever the nat-
ural gas supplies were interrupted. Because the conversions did not involve
the purchase of additional equipment, it was a cost-effective method of pro-
viding gas for peak loads.
The simplest conversion of carbureted water-gas apparatus for the produc-
tion of high Btu oil gas was the refractory screen oil process. This conver-
sion consisted of replacing the coke in the generator of the water-gas appara-
tus with a high-temperature refractory brick and adding additional oil sprays
and oil-handling equipment. Figure 17 is a diagram of the converted appara-
tus. The apparatus is operated in a manner similar to the Pacific Coast
processes, with a 3- to 6-tninute cycle. This process was successfully demon-
strated with a wide variety of hydrocarbon feedstocks with up to 16 percent
carbon and between 10 and 39° A.P.I. (Johnson, 1932), Table 15 shows the
results of the process when using fuel oil and gas oil. All of the tars pro-
duced by the process could be recycled back into the process, reducing the
overall fuel requirements. The refractory screen oil-gas process involved a
minimum modification of existing carbureted water-gas apparatus and could
produce high Btu oil gas for peak loads at relatively low costs to the gas
companies.
51
-------
M
M
0)
a
M
IB
* 8
c *-
e> ._•
§
o
U
£
at
DC
£
I
52
-------
TABLE 15. RESULTS OF REFRACTORY SCREEN OIL-GAS PROCESS
2.4
24.2
0.4
6.6
32.0
27.0
4.4
3.0
2.6
20.1
0.4
7.7
36.2
25.2
3.8
4.0
2.9
15.9
0.4
8.8
40.4
23.7
3.4
4.5
Heating Value- -B.t.u. per cu. ft. 1100 1000 900 800
Specific Gravity 0.740 0.682 0.631 0.579
Gas Analyses
Carbon Dioxide—Per cent... 2.0
Illuminaats *. .28.6
Oxygen 0.4
Carbon Monoxide 5.4
Hydrogen 27.7
Methane 28.5
Ethane 4.9
Nitrogen 2.5
Naphthalene—Grs. per 100 cu.
ft 1.39
Hydrogen sulphide — Gra. per
100 cu. ft 80
Organic sulphur—Grs. per 100
cu. ft 2.9
Oil requirements—No tar return
Fuel Oil—Gals, per MCF.... 4.63 5.17 5.72 6.13
Gas Oil—Gala, per MCP.... 7.60 *6.25 4.95 3.60
Total Oil—Gals, per MCF.... 12.23 11.42 10.67 9.78
Oil requirements—With tar return
Total-Oil—Gals, per MCF... 10.3 9.7 9.2 8.5
Steam—Ibs. per MCF 32.0 36.7 40.7 44.5
Tar—Gals, per MCF 1.95 1.74 1.54 1.30
Overall Thermal Efficiency 75.5 72.6 69.2 67.0
Basis of Figures:
Fuel oil—12-18 deg. A.P.I. Cracked Mid-Continent Resid-
uum—0.7 per cent sulphur
Gas oil—34-38 deg. A.P.I. Pennsylvania gas oil
Source: Johnson, 1932.
53
-------
Another adaptation of carbureted water-gas apparatus was the Hall high
Btu oil-gas process. It utilized the carburetors and superheaters of two
adjacent carbureted water-gas sets to form a single oil-gas set. Figure 18
shows the configuration of the equipment of the Hall process. The apparatus
was operated with a fairly complex cycle that captured more of the heat
created during blows, with resulting increases in thermal efficiency and
reduced fuel consumption. Table 16 shows the operating results of the process
for five different oils. The light oil recovered from the process was
approximately 0.35 gallons/MCF, with its characteristics comparable to that
produced using carbureted water gas (Utermohle, 1948afb and Utermohl, 1948b).
1.2.5 Miscellaneous Gas Production Methods
Besides the three major types of gas production processes (coal, carbu-
reted water gas, and oil), there were several minor processes that were com-
monly used, principally by small manufacturers. These processes are listed
with their uses and waste products in Table 17. These processes were typi-
cally employed for the lighting of small towns, hotels, or factories. Because
they were, in general, small producers who used processes with minimal wastes,
sites using exclusively these processes will probably pose only minimal
hazards. The production of rosin gas or whale oil gas was primarily used
prior to the discovery of bituminous coal in the United States in 1840.
1.3 MANUFACTURED-GAS CLEANING AND PURIFICATION PROCESSES
1.3.1 Introduction
The raw gas from manufactured-gas processes contained many components
that were removed prior to gas distribution. Components that would condense
within the distribution system, corrode pipes, or produce noxious gases when
burned were removed by various processes. Cleaning and purification processes
removed undesirable materials from the raw gas. These processes were employed
sequentially, with the gas flowing through the entire purification train prior
to distribution.
The processes employed to clean the gas were dependent on the method of
gas production and sometimes on the specific raw materials used in gas produc-
tion. Table 18 shows the general temperatures and Impurities in manufactured
gases as they enter the purification train. The specific concentrations of

54
-------
Figure 18. Schematic diagram of Hall oil-gas process.

Source: Utermohle, 1948a.
55
-------
TABLE 16, HALL OIL-GAS PROCESS OPERATING RESULTS

Oil Analysis
Conradson carbon, %
Ash content, %
Enriching value, M Btu per gallon
(Dick Method) (avg. of 1400.
1500, 1600' F results)
.20
.02

103.6
3.16
.01

98.8
6.02
.04

102.5
13.03
.16

93.8
12.56
.16

95.2
Operating Results
Btu of gas per cubic foot
Specific gravky of gas
Gals. Heat oil per MCFt
Gals. Make oil per MCFf
Gals. Total oil per MCFf
Gals. Tar per MCFt
Gas made per day, MCFf
Thermal efficiency, %
1046
.855
.99
11.50
12.49
2.56
5,952
79.2
1006
.866
.80
11.46
12.26
2.73
5,088
80.9
1047
.834
.95
10.67
11.62
2.30
5,088
81.8
966
.867
.28
11.69
11.97
2.63
3,576
80.6
974
.833
69
13.31
14.00
•
3,504

'Million cubic feet.
56
-------

in
UJ
in
in
UJ
g
^

i
M
t-

in
3

g
5
5
^j

a
M
a

m
h.

UJ
2
K

n
3
•o

M
9
ft*

e
0

«M
n
0
o
u

M
•0
o

•0
0
4^
i

C
o

"J3
u
3
•o
0
u
a.

0
a
x

N
U

1
c u "o
— 3 a
•o c
c o u
— U 3
. ax
• M •
U • X
O)
TJ -D
3*3
O •»> O
* u. o
A •
u » -o
n *> .- c
*) u x *i
O 1- O
« -*> a
4> 9 W
Mi. • —
• c -o
* « 0
X — U
< •»* Ji O

M
c
o
£,
t.
A
U
O
U
x
J£
^4
*> I

o> -
— o
-IU

_
'o

« W
• *»
X U
* 0

o u
* TJ
X «
O X
L
X C
a. —

W
^
a

_
IM
o
J»
•
M
•o n
— 01
3
O 1.
* •

U v-
* n
j3
•o u
— 0
O 0
n u

- t
XX

A
e c c
x — o
.—
X C *>
u — o
• n 3
> E -o
• 0
< u u

„"
c
o
X

'a o

*o u
«i -o
'n -u
o x
u
X C
a. —

M
M
01
C
M
a:
•a
*-*
3
0 •
* -

• 3
X
a 0 0
x e u
« .ua
O A A O
U « • *>
X • g —
*> e E • 0
— » •- N
•O — — O 3
0 O •>- 0 •
— ii a a
x a M M
3 u n e • jt
X O C — 4> U
n • o -o ! X-
— X 3 0 X
4> 19 e x «
u a o i. *> •
— • u 3 — —
< c am * «-
n
a

JQ
1

C
o
¥1
O

c
o

0)
X

«••

•o

4J
•
U •
•o 0
X 3
X T3

JJ A
^M ^

CM
CM
U
M
s
M

^
§ «
*u S"
— uo
ssr

?•!"

^•J
O ^9 +
3 —
e uo
u. • •
(L 00

A
e

^.
8

B
e
o
jo
u

u
o
u
•o
X
X

4>
X
o>
J

•*•
e _,,
§ o
— "•
e u
u
4> X

a—
u

01
c e
— o
JC U
u *~
u e
o •-

^
e

M
• «•
a •
01
Ss
•I5

•
^^
CO

•«.

*
v4
^^
ff
a

f>
•
a

u
•

i
o

^
01
c
o
.0
u
n
JdT
•o
X -
X N
X

N -
oo
wu

X
•J
4)
U
a

*••
*o
n
"S 3
41
a. u

n
4
at

«
•
57
-------
r
TABLE IS. TEMPERATURE AND IMPURITIES IN RAW GASES AT OUTLET OF
HYDRAULIC MAIN OR WASHBOX

Coal
gas

Blue
gas
Carbureted
blue
gas
Pacific
Coast oil
gas
Temperature °F
Impurities
Percent by volume
Water vapor
Ammonia
Tar and oi1 vapors
Parts per million (ppm)
Cyanogen
Naphthalene
Hydrogen sulfide
Organic sulfur
140-190
19-30
1-2
2-3.5
160-200
32-78
160-200
32-78
1-2
150-200
25-78
1-1.5
1,007-1,410 b b ..
3,700-9,300 — 1,490-4,660 2,790-11,200
8,000-12,800 1,500-3,200 1,920-4,800 3,200-4,800
594-850 b 170-510 340-510
aSmall amounts with bituminous coal.
definite figures available but amounts are small.
58
-------
the impurities were dependent on the raw materials used to manufacture the gas
(e.g., sulfur content of coal or oil) and the operation of the gas production
process. Table 19 shows the types of gas purification processes and whether
they were used with specific gas production processes. This section is
divided into descriptions of specific purification processes followed by
descriptions of general purification systems for coal gas and carbureted water
gas or oil gas.
1.3.2 Condensers
After the raw gas leaves the production apparatus, it passes through a
water-sealed hydraulic main or a washbox where the gas is initially cooled and
some of the heavy tars are condensed and removed. The purpose of condensing
the gas is to cool it to ambient temperature and remove all constituents that
are not gases. The condenser causes water vapor and tars to condense from the
gas and form a liquid, which is then removed from the condenser. Air condens-
ers (condensers that transferred heat from the product gas to air) were the
first type employed for the cooling of gas. It was originally believed that
slow cooling of the gas allowed more of the illuminants to be retained in the
gas and hence be distributed. These condensers were frequently lengths of
pipe that zig-zagged across the wall of the retort house.
Water-cooled condensers replaced the air-cooled versions about 1900.
These condensers were basically shell and tube construction, with cooling
water passing through the shell and the gas flowing through the the tubes.
The heat from the gas was transferred from the gas through the tubes and to
the water.
Direct cooling (or scrubbing) of the gas by direct contact with recircu-
lated condensate began about 1907 and spread rapidly to both carbureted water-
gas plants and coal-carbonization plants. It is also the method currently
used for cooling of coke-oven gas. In direct cooling of the gas, it is con-
tacted with cooled recycled water. The water is heated as it absorbs heat
from the gas, and additional condensed water vapor and tars are removed in the
water. The tars are then separated from the condensate water, the water is
cooled, and then reused in the gas cooler. The direct cooling of the gas is
usually accomplished in a counter-current packed scrubber, as shown in Fig-
ure 19.

59
-------
.

CO
UJ
CO
CO
UJ
u
o
cr
Q.
z
O
t
^3
a
o
a:
CL

I
1-
5

CO
CO
UJ
CO
a:

o
O
a.
z
g
§
(E
E
ID
a.
CO
a
(E
o

5
a!
T™

00 .
pi
*

X
b
1 -X •
a u >
6 • o
• — u
b
, - "2
•Q e Z e
-o o o £
•< — z «
*> b

x
— b
o •
c >
• o
j: u
a. •
b

X tJ 1 b ^»
O — M a • r-
— b 3 3 C *» M
X O O" b — «•• Ol
JC — O M m *•
— N- ^ n ^^
"
§• b ^*
C T3 • 9
• O — .*> Ol
b b X ••• OB
»-* O • **
in —
z
b.^
• OS

— "• CO
_IJ —
I •"
X
•— II
e >
o e
I u
E «

X — 3 O
a— b e
— O 0 —
-i • JJ
— b
b > -a >
• gee
»- 1 • u
b b

n
b b
c • >
• * E
T: b •
0 O b

H
fl
C
U
o
b
a
e
o

^
4
u

•»
9
a.
«

eeee ee. aeatee vt

w w os ee ee oe oe

vim ee eteeeeee

V> 1 OE «/»«/>

33 CB 3333

31 ae i/» «i «/>

vi 3 o: eece.cz ac

33 or BOOM

33 K £333

33 fl: 3333

C
>
o
• 3 — ~
it e jt OB — — o
MOO Ji — O
• — 0 b J O •
o *> «x — TJ
o •*> *> a « • a
b M u • • • a b
o — • 3 « » c a«» u

e o b o a -9 999 at
e xob • e c e c
u u a: CD u b3333
a a 3
o • — «
60
-------
F
FP.OM
MAIN
LIQUOC
UIQUOK PUMP VMCTICM PIFK
COOLIMCi COItW
eoouco urauoie KCTURKI Pipe
eiiceukATm« PUMP
GA./« OUTUftT
UQUOR OI/CHARQC *O CQIU/*
VAUVCJ*
Figure 19. Direct contact cooler.
Source: Morgan, 1926.
61
-------
1.3.3 Tar Removal
Tar is a complex mixture of carbon and hydrocarbons that forms when
either petroleum is thermally cracked or coal is carbonized. When raw manu-
factured gas is cooled, the tar condenses from the gas and usually separates
from the condensed water. The distinguishing feature of tars (in the manufac-
tured-gas industry) is that they have a specific gravity greater than 1.0 and
sink when placed into water. Organic hydrocarbons that have specific gravi-
ties less than 1.0 and float on water are considered oils. Tars were con-
densed and recovered with condensate at several locations within the purifica-
tion train. The heaviest tars condensed in the washbox or hydraulic main.
The lighter tars were condensed with water from the gas either in indirect or
direct condensers. Tar fog (aerosols of tar remaining in the gas after scrub-
bing) are removed with either a P. and A. (Pelouze and Audouin) tar extractor
or an electrostatic precipitator (ESP). Figure 20 shows a P. and A. tar
extractor. The gas flows through a pipe in the center of the apparatus, then
through several concentric perforated inverted bells. As the gas flows
through the perforations in the first bell, the tar aerosols impact on the
metal of the second bell, removing the tar from the gas. The counterweight
attached to the bells allows the bells to move up and down within the appara-
tus, exposing more perforations when the gas flow is high and avoiding exces-
sive pressure drops across the extractor. ESP's were introduced about 1924
for the removal of tar fog from gas (Downing, 1934). Figure 21 is a Cottrell
ESP. It consists of a steel shell containing vertical tubes. A charged wire
runs down the center of each tube. As the gas flows through the tubes, the
tar aerosols become charged and impact on the tube walls, removing the tar
from the gas. The ESP's were very efficient for the removal of the tar fog,
and they were installed on many of the larger coke ovens and carbureted water-
gas plants.
A common method for the removal of tar aerosols was the use of shavings
scrubbers. These were basically towers or boxes that were filled with wood
shavings (or sometimes other materials, such as oyster shells, coke, or slag).
62
-------
If
Figure 20. P+A tar extractor.

Source: Morgan, 1926.
63
-------
U'-M*
•'D-i
Figure 21. CottrellESP.

Source: Morgan, 1926.
64
-------
fT
I!
The gas would flow through the woodchips, and entrained tar would hit the wood
and be removed from the gas. The tar-contaminated shavings would be periodic-
ally removed and discarded or burned. The shavings scrubbers were used prin-
cipally after condensers or scrubbers and removed entrained tar aerosols. The
removed tar was prevented from entering the iron oxide boxes, extending the
useful life of the iron oxide. Small carbureted water-gas plants and small
oil-gas plants were most likely to use shavings scrubbers because their gas
production was small and the tar loadings were relatively low.
The tar was usually separated from the condensates by gravity in tar
separators similar to that shown in Figure 22. The tar/condensate mixture
flows into the separator and separates into three distinct layers by gravity.
An oil layer of lighter hydrocarbons floats to the top of the liquid and is
retained by oil skimmers. The tar sinks to the bottom of the tank and is
removed. Water is the middle layer, and it flows through the gaps in the
baffles and exits through the water outlet. The tar separator produces three
distinct products, which receive different treatments, depending on the pro-
duction process. Table 20 lists what was generally done with these three
products. Because carbureted water gas and oil gas produce very little
ammonia or phenolic compounds, these were not recovered from oil-gas and car-
bureted water-gas condensates. The oils from the separator were frequently
not recovered, particularly with oil and carbureted water gas. In these
ca'ses, such oil would be disposed with the condensate.
This type of tar separator had one major problem: The tar, oil, and
condensate had to separate relatively rapidly and form the three distinct
layers. This usually did not present a problem for coal-gas plants, but car-
bureted water-gas and oil-gas plants frequently formed oil/tar/water emul-
sions. These emulsions were relatively stable and were difficult to separate.
An emulsion would quickly fill the tar separator, with little or no separation
..f the tar. The emulsion would then flow out of the separator through both
the tar outlet and condensate outlet. In addition to gravity-based tar sepa-
rators, several other methods were employed for the separation of condensates,
oil, and tar. These are listed in Table 21.
In situations where the entire 011/tar/water mixture was disposed instead
of recovered, the mixture separated into the three fractions after disposal.
65
-------

I
a
a.
a»
n •
•a

a
o
CM

8
66
-------
TABLE 20. STREAMS FROM TAR SEPARATOR
Stream Treatment
Tar Burned as fuel, sold to refiners, distilled on site,
mixed with carburetion stocks, or disposed

Oil Recovered and mixed with light oils, mixed with
carburetion stocks, or disposed with condensate water

Water (condensate) Disposed into stream, treated for recovery of phenols
and ammonia (coal gas only), flowed through coke beds
prior to disposal, used as coke quench water, recycled
to cooler-scrubbers
67
-------
TABLE 21. OTHER METHODS OF TAR-WATER SEPARATION
Method
Description
Steam still
Centrifugal process
Warner tar
dehydration system

R.S. de-emulsifying
System
Steam Is used to distill water from the tar. High
cost due to high steam consumption but will handle
emulsions containing any concentration of water.

Water and tar are separated by density in a
centrifuge. Fairly low cost of operation but requires
frequent cleaning of tars from the equipment.

A modification of the steam still in which water is
distilled from the tar.

Tar-water emulsion is placed Into a tank, 30 Ib of
soda ash is added (for 5,000 gal tank), and- the
emulsion 1s heated to 312 *F under pressure. Most
emulsions then separate in 30 min - 18 hr. Water is
flashed from the tank to cool it to 212 °F.
SOURCE: Seely, 1928.
68
-------
In rivers or streams, the oil and water would be carried downstream, with some
of the oil depositing on the banks of the stream. The tar fraction would sink
to the bottom of the stream and was usually stopped by fine silts along river
bottoms. In- the ground, however, the mixture separates so that oils float on
the groundwater surface, the water soluble components dissolve in the ground-
water, and the tar layer sinks through the groundwater until stopped by a low
permeability layer of ground.
1.3.4 Naphthalene and Light-Oil Scrubbers
After the tars were removed from raw manufactured gases, naphthalene and
light oil were frequently removed from the gas. Naphthalene is a fairly vola-
tile PAH compound, which frequently was not completely removed with the tar.
Naphthalene would crystallize within the gas distribution system, plugging
orifices and reducing flow through pipes. It would often drop out of the gas
as the gas passed through iron oxide purifiers, decreasing the life of the
oxide. The naphthalene could be easily removed from the gas by scrubbing with
a relatively small amount of a petroleum oil. The naphthalene-enriched oil
could then be either distilled for the recovery of naphthalene or used in the
carburetion of water gas or the production of oil gas. Figure 23 shows a
naphtho.ene scrubber that consists of two stages: The first stage scrubs the
gas with a recirculated oil, and the second stage uses a small amount of fresh
oil for the scrubbing. The use of two stages allows most of the naphthalene
to be removed in the first stage, with almost complete removal of the naphtha-
lene in the second stage. Used oil from the second stage is added to the
recirculating oil of the first stage. Some of the recirculating oil is con-
tinuously removed. The naphthalene-containing oil from the process was never
considered a waste product, in that the fuel value of the original oil was
enhanced by the naphthalene, and the oil could be either sold or used at the
plant. The naphthalene could be recovered from the oil (if profitable under
market conditions) by distilling the naphthalene-containing oil. Recovered
oil could then be reused in the process.
Any fluid petroleum oil could be used to scrub naphthalene from the gas,
and the most common oils were gas oil and fuel oil. Because the naphthalene
had a large affinity for the oil, relatively low oil fl.owrates were used for
the removal of naphthalene. Table 22 shows typical operating results for a
naphthalene scrubber.
69
-------
Figure 23. Naphthalene scrubber.
Source: Gas Engineers Handbook, 1934.
70
-------
.
1ABLE 22. RESULTS OF NAPHTHALENE SCRUBBER AT SEABOARD BYPRODUCT
COKE CO., KEARNY, N£W JERSEY
Inlet naphthalene (ppm)
Max 577
Min 298
Average 436

Naphthalene in outlet gar (ppm)
Average 69

Oil consumption 17.5 gal/106 ft3

Spent oil
Specific gravity (22 °C) 0.875
Light oil (to 200 8C) 20.1%

SOURCE: Gas Engineers Handbook, 1934.
71
-------
Light oil consisted of the light aromatic compounds contained in the gas.
They were primarily benzenes, xylenes, and related compounds. These compounds
were originally considered beneficial in the gas because they burned with a
brighter flame than did other gas constituents. With the invention of the gas
mantle and the switch from light to heating standards for gas, the illuminants
were no longer necessary for the gas quality. During World War I, the demand
for benzene and xylene chemicals increased greatly, and many gas plants began
to recover the light oils from the gas. The method of removing light oils
from the gas is very similar to that for the removal of naphthalene, except
that the light oils were always recovered. (The recovery of light oils was a
purely economic decision when the recovered oils were worth more than their
heating value in the gas. When not recovered, the light oils enriched the
distributed gas and caused no problems in the distribution system.) Figure 24
shows a representative light-oil scrubber. The entering gas is scrubbed
counter-currently, first by recirculated oil, then by fresh oil. Spent oil is
removed from the recirculating oil and distilled to produce the light oil and
regenerated scrubbing oil. A variety of oils was used in the scrubbing of
light oil, including gas oil, green oil, fuel oils, tetralin, and lighter tar
fractions.
Light oil contains a variety of intermediate boiling hydrocarbons.
Table 23 shows a typical analysis of a coke-oven light oil, divided into
distillation fractions. Table 24 is a list of compounds commonly found in
light oil from coke ovens. Constituents of light oil from oil gas or coke
oven gas would have a subset of these constituents, excluding the phenols and
base nitrogen compounds. Light oil was used as a feedstock for the production
of benzene, toulene, xylsne, and other organic chemicals, or it was mixed with
gasolene to increase its octane. A complete history of light-oil recovery was
prepared by Glowacki (1945).
Light oils were recovered at most coal-carbonization plants, large carbu-
reted water-gas plants, and large oil-gas plants. Small gas production plants
would usually not recover the light oils (they did not produce enough to make
their recovery profitable). When the light oils were not recovered, they
passed through additional gas purifiers, then into the distribution system,
and were ultimately burned with the product gas.
72
-------
CITIZENS Woo
OUT OIL
I
Figure 24. Light oil scrubber.

Source: Green, 1939.

73
-------
TABLE 23. ANALYSIS OF A TYPICAL
CRUDE COKE-OVEN LIGHT OIL

Percentages
by Volume
I. Furcrunninns.
Cyclopentadienc 0.5
Carbon disutfide 0.5
AmyIcnes and unidentified 1.0

II. Crude benzol.
Bvnzene 57.0
Thiuphcne 0.2
Saturated nonaromatic hydro-
carbons, unidentified 0.2
l;n5Rturates, unidentified 3.0

HI. Crude toluol.
Toluene 13.0
Saturated nonaromatie hydro-
carbons, unidentified 0.1
Vnsaturatea, unidentified 1.0

IV. Crude light solvent.
Xylenea 5.0
Ethyl benzene 0.4
Styrene 0.8
Saturated nonaromatie hydro-
carbons 0.3
U maturates, unidentified 1.0

V. Crude heavy solvent.
Coumarone, indent, dicyclo-
pentadiene 5.0
Polyalky! benzenes, hydrin-
dene, etc. 4.0
Naphthalene 1.0
I'nidentined "heavy oils" 1.0

VI. Wash oil 5.0'
Total 100.0
• Thr amount at wnih oil pr»*ent deprndu
cririil.T uiHin th» pcrferawne* »nU dnlcn of (he
ilHwntallutlna •ppiralu* t* wll it upon tb«
iinturc of iht w**h ofl caplortd.

Source: Glowacki, 1945.
74
-------
TABLE 24. CHEMICALS FOUND IN
LIGHT OIL FROM COKE OVENS
Aromatic hiftlroearbont
benzene
toluene
o-xylen«
m-xylene
p-xylene
ethyl beniene
hydrindene

isopropyl beniene
o-ethyl toluene
m-ethyl toluene
p-ethyl toluene
n-propyl beniene
meaitylena
paaudoaumene
hcinimcllitenc
naphthalene
(1.2-dihydronafihlhalene) t
(I.4-
C»H4(CH,)i
C«H«
-------
TABLE 24. (con.)
.-trnrmifi'f H]/tlrntartx>n* tnl/i UN-
natiiralnl *\tlr chain*
styrcne
indcne

(2-methyl indcne) t
(3-mclhyl indent') t

\eutral Ofyirn rompoiim/x
acetone
methyUthyl Icetnne
roumarone

aciftophenone t
(2-iiiethyl eoumarone) t
(3-methyl eoumarone) t
(5-rncthyl coumarone) t
(6-mcthyl eoumarone) t
(7-methyl eoumarone) t

\ttilral ami aeitiic nitrogen
compound*

hydrogen cyanide
aectonilrile
benionitrile

I'henoU

phenol
o-cresol
m-creaol
p-creaol
2.3-dimrthyl phenol
2,4-dimeihyl phenol
2,5-dimcthyl phenol
2,8-dimelhyl phenol
3,-t-dimethyl phenol
3.5-dimethyl phenol
o-ethyl phenol
m-ethyl phenol
p-«thyl phenol

ftattc niVrofrn com pound t

pyrrole

pyridine
aniline
2-methyl pyridine
3-methyl pyridine
4-mcthyl pyridine
<»-loluidine
2.3-
-------
TABLE 24. (con.)
2.4,5-irimethyl pyridine
2,4,'i-uitnet.hyl pyriiiine
(2,3.4-lrimethyl pyridina) t
(2,3,5-trimelhyl pyridint) t
(2,3,8-trimethyl pyridin*) f

-------
17
1.3.5 Removal of Ammonia and Recovery
The production of ammonia, cyanides, and phenolic compounds occurred with
gas produced by coal carbonization. These compounds were produced in trace
amounts by carbureted water gas and oil gas and were not removed or recovered
from these processes. Prior to the Haber process for the synthetic production
of ammonia, coal carbonization was the principal source of fixed nitrogen.
The removal of ammonia from the gas was always accomplished by scrubbing the
gas with water, condensate, or sulfuric acid. Ammonia has a very high affin-
ity for both water and acid solutions and is readily removed by aqueous
scrubbing.
During coal carbonization, a portion of the nitrogen in the coal is con-
verted to ammonia, and other nitrogen forms cyanides, organic nitrogen com-
pounds, or remains in the coke. Table 25 shows the average distribution of
nitrogen compounds from high-temperature carbonization of coal. Approximately
18 percent of the nitrogen in coal is converted to ammonia during carboniza-
tion. This is about 1.1 percent by volume of the raw coal gas.
There were three basic processes for the removal of ammonia from coal
gas. These were the direct method, the indirect method, and the semidirect
method. They differ primarily in the treatment of condensate containing the
ammonia and are described in detail in several commonly available references
(Wilson and Wells, 1945; Kohl and Riesenfeld, 1985; Hill, 1945). In the
direct method, the raw coal gas was scrubbed directly with a solution of sul-
furic acid. The ammonia was absorbed into the solution, reacted with the
sulfuric acid, and the resulting ammonium sulfate precipitated. This method
was the simplest method of removing ammonia as a product from the gas, but the
resulting ammonium sulfate was of poor quality and generally contained sub-
stantial impurities. An additional drawback to the process was the degrada-
tion of the coal tar from contact with the sulfuric acid.
The indirect process, as shown in Figure 25, removes ammonia from the
coal gas by first absorbing the NH3 into water, then releasing the ammonia as
a gas in an ammonia sti'ii. The raw coal gas first contacts recirculated
78
-------
TABLE 25. DISTRIBUTION OF NITROGEN IN COAL CARBONIZATION PRODUCTS

% of nitrogen
originally In coal

Ammonia 18.0
Cyanide 1.2
In tar 3.3
Free in gas 27.5
In coke 50.0
-------
Primiry Coolers
Direct Type
Coaling Wjtef

Flushing Liquor'
Circulating Tank
(jquor
Collecting Tank
Figure 25. Indirect process for ammonia recovery.

Source: Wilson and Wells. 1945.
80
-------
I!
flushing liquor in the hydraulic main. The gas is cooled to a certain extent,
and the heavy tars condense. The fixed ammonia compounds (those that do not
release ammonia when the solution is boiled, such as (NH4)2S04 and NfyCl, are
dissolved into the flushing liquor. This gas is then further cooled either by
direct or indirect condensation, with most of the tar and water being con-
densed from the gas. The condensate, which has a high ammonia concentration,
is separated from the tar in gravity separators. The remaining tar aerosols
in the gas are removed by an ESP, and the remaining ammonia in the gas is
removed by scrubbing with water.
The condensate and ammonia scrubber water are mixed and fed to an ammonia
still thai uses lime and heat to decompose ammonia salts and free the ammonia
as a gas. Figure 26 is a diagram of the ammonia stills that were generally
used. The ammonia still is constructed of a fixed still, volatile still, and
lime keg. The volatile still removes all of the free ammonia and other vola-
tile compounds from the crude ammonia liquor. The fixed still decomposes
fixed ammonia salts in the liquor and liberates the ammonia gas. Lime water
is fed to the lime keg while ammonia still waste is removed from the base of
the fixed still. The free ammonia and steam that exit the top of the volatile
still were scrubbed either with water (to reabsorb the ammonia as an aqueous
ammonia product) or with sulfuric acid (to produce ammonium sulfate).
The semidirect process (Figure 27), patented by the Koppers Company in
1909, was a variation of the indirect process. The processes were identical
except the indirect process did not use water scrubbers to remove the final
amounts of ammonia from the gas. Instead, the coal gas (after complete tar
removal) was bubbled through sulfuric acid with the ammonia from the lime
still. This reduced the amount of crude ammonia liquor that was processed
through the ammonia still and allowed for better heat utilization in the satu-
rator. There were also reductions in capital and operating costs with the
semidirect process, with only marginal effects on the quality of ammonium
sulfate product.
The lime still would have been effective at removing volatile organics
that were dissolved in the liquor, but tar acids (principally phenols) were
retained in the still waste and frequently constituted a major disposal prob-
lem for the gas plants. The phenols have a very low taste threshold in water,
81
-------
.Ammonia Vtpor lnk.1
Ammonu Liquor Supply
Vim to Atmmpncft -^
LIMC rccolax
Supply 10 Litn« L(|—.
Ovtrtlc* Return to Mnw-^.
Supply Iram Mii
0*»Mt ifTuier Bypni

-Cendtmita Ritura w SIM
Lxjuor Supply toO«pn«na4itintTo««r
Liquor Return from 0*pnenalirin| Tomr
LIMC MlXCft PUMIi
Figure 26. Ammonia still.

Source: Hill, 1945.
82
-------
Spray*
Hot Gases
Itftmi

V. Wlltf Ifit In
in. Vtpor «ntf Slum
+ NHj
and Tar
shine Liouor
r
Son
n*c
n
i— ffY-i—
T1 ' 'T*
! Indirect !
Primary Cooler
i Tubular 1
;J4,|,
Ti'Tt-*1'?'
-Coolmg Water
+ and Tar
Ophtegmator,
i i i | To Gil
Mam
Before
Saturator
Cooling
Water
Still
Waste
Circulating
Pump
^flushing Liquor
Circulating Tank
reed Pump*
Milk-of-Lime
Feed
Am mom*
Still
Figure 27. Semidirect process for ammonia recovery.

Source: Wilson and Wells, 1945.
83
-------
particularly when the water is chlorinated. The removal and treatment of
phenols in the still waste is discussed in Section 1.3.5.
Sometimes the gas plant would sell the ammonia liquor directly to a chem-
ical company for the production of ammonium sulfate. The ammonia liquor could
be used directly as scrubbing liquor to absorb $63 produced by burning sulfur.
This is essentially the process for producing sulfuric acid, except that by
using ammonia liquor, ammonium sulfate can be produced directly as a product.
1.3.6 Phenol Removal and Recovery
Phenol was produced in the carbonization of coal. As an acidic compound,
it was readily absorbed in the condensate and ammonia liquor during the puri-
fication of the coal gas. The phenol remained in the ammonia still waste and
had to be removed from this waste stream before disposing of the water. The
phenolic compounds were very noticeable in water, imparting a medicine taste
to it. This occurred even at low concentrations and was exacerbated when the
water was chlorinated. There were several methods that were commonly used for
the removal of phenol from the ammonia still waste.
-.3
The simplest method of disposing of phenol containing liquid wastes was
to discharge the water directly into the city sewer system (if one were avail-
able). The phenol in the wastes was rapidly degraded by organisms in the
sewage and by the activated sludge method of sewage disposal.
A common method of disposal was to use the water to quench coke as it was
removed from the ovens. This method substantially reduced the volume of the
wastes, but it degraded the value of the coke, greatly increased the corrosion
of steel in the coke-quenching area, and evaporated phenols into the air.
These evaporated phenols generally killed any remaining plant life around the
coke plant and may have been washed into surface water.
If recovery of the phenols were desired, the phenol was extracted from
the raw ammonia liquor by washing the liquor with benzene or light oil, then
recovering the phenol from the benzene by washing it with a solution of sodium
hydroxide. This process is shown in Figure 28. The process uses benzene or
light oil, which continuously absorbs phenol in one tower, while the solution
is continuously regenerated by contact with a sodium hydroxide solution in a
second tower. The sodium phenolate was then usually converted to raw phenols
by "springing" the solution with carbon dioxide. The process actually removed
84
-------
1
Phenol
absorbers-:
Jrr-
seat—>\JT
. Wj Litpor hncs

• - B«nzol lints

Caustic and sodium~
pnvnolatt and
phenol lints
Crude
•I I
Phenol storage
KflSSSWyifSS
•Stnzol/y
'circulat-.
ing tank'-
' ~**J****S

To cfefffieno/ofact liquor storage
(,/ C/fC — |
liquor m s—-j
Figure 28. Benzene extraction of phenols.

Source: Jones, 192&
85..
-------
a variety of tar acid compounds from the liquor, although the recovered prod-
uct was primarily phenol. The recovered tar acids from one plant were anal-
yzed (dry basis) as 57 percent phenol, 13 percent o-cresol, 8 percent
m-cresol, 10 percent p-cresol, and 10 percent higher tar acids (Wilson and
Wells, 1945). This process generally removed about 75 percent of the phenols
contained in the ammonia liquor, but higher removal efficiencies were obtained
when the phenols were separated from the benzene by distillation instead of
extraction with caustic.
A second common method of recovering ohenols was the Koppers vapor recir-
culation process. In this process (shown in Figure 29), ammonia liquor was
removed from the base of the free still (after removal of the free ammonia,
but before the fixed ammonia salts are decomposed) and was stripped by steam.
The steam-stripped ammonia liquor was then returned to the lime keg section of
the ammonia still for the decomposition of fixed ammonia salts. The steam and
phenols were then scrubbed by a solution of sodium hydroxide, removing the
phenols as sodium phenolate. The sodium phenolate could then be sprung as
phenol-using carbon dioxide. This process had higher removal efficiencies
than did extraction of phenols, and it generally gave about 97 percent
removal. Inlet concentrations of phenol were about 2.5 g/L.
Wilson and Wells (1945) mention the disposal of waste ammoniacal liquors
into the ground but advise:
Discharge into an opening, such as a disused well, Is dangerous,
because the final fate of the liquor is unknown. It may be grad-
ually dissipated and purified as it seeps through the soil. On the
other hand, it may find its way into some water-bearing strata or
percolate unchanged through the layers of soil to drain into a
stream. In such a case, the pollution would not appear immediately,
but when it did, deposits of the material in the contaminated soil
would cause the trouble to persist over a long periou of time.
The ammoniacal liquors could also be discharged directly in a stream or
bay, if the water were not used for drinking purposes. This would have been
more common along coastal areas, where the discharges could flow directly into
the ocean, and complaints would be minimal. Evaporation of the liquors by
flue gas or steam was also suggested as a method for disposal, but it was not
generally employed.
86
-------
W«ak Ammonia
Liquor to Still
Steam
Oephenolized
Still Waste
UquMi. Wtak Liquor
Vapors.
Sodium Httnoltta
Ptwnol

S(**/n
S^luraled Vioon
Producli of combustion lor C0>
Liquor
Pump
-Ammonia Still-
Sodium
Phenolate
Springing'
Tank
Blower for Gas
Containing CO?
• Dephenotizmg Plant -
• Springing Plant •
Figure 29. Koppers vapor recirculation process for phenol removal.

Source: Wilson and Wells, 1945.
87
-------
1.3.7 Removal of Hydrogen Sulfide
1.3.7.1 Introduction--
The need to remove hydrogen sulfide from town gases was recognized very
early in the industry. If left in the gas, the H2$ would cause corrosion in
the distribution system and appliances, be a nuisance to the consumer, and be
an odor problem with even small leaks of gas. Hydrogen sulfide was produced
by all major gas production methods, so its removal was universal within the
industry. The concentration of hydrogen sulfide in the raw gas (and hence the
amount of H2S to be removed) was proportional to the original sulfur concen-
tration in the gas feedstocks. For coal carbonization, the sulfur concentra-
tion in the original coal determined the gas H2S concentration; for carbureted
water gas and oil gas, the sulfur concentration of the oil used was the
primary variable. Table 26 shows typical concentrations of hydrogen sulfide
in town gases, although these numbers would vary considerably, depending on
the sulfur concentration of the feedstocks used to produce the town gases.
The sulfur removed from the gas could either be recovered as a salable
byproduct, discharged as H2S to the air, or discarded as waste. Lime was the
original material used for the purification of gas until the process was
widely replaced by iron oxides after about 1885. Iron oxides were universally
used for the removal of hydrogen sulfide from coal gas, water gas, and oil gas
until about 1927, when several liquid purification processes for hydrogen
sulfide removal became available (primarily the Seaboard and Thylox proc-
esses) .
1.3.7.2 Hydrogen Sulfide Removal by Lime--
Hydrated lime was one of the earliest techniques used to remove H2S, C02,
and other impurities from coal gas. This lime was produced by calcining lime-
stone, then slaking the lime with water to form calcium hydroxide. The rele-
vant reactions for the purification of coal gas with hydrated lime are:
Ca(OH)2 + H2S = CaS + 2H20
Ca(OH)2 + C02 « CaC03 + H20.
The lime also removed some cyanides (which reacted with iron impurities
in the lime to form ferrocyanides) and some tar materials. Stolchiometrlc-
ally, each mole of lime could remove one mole of C02 or H2$. Actual

88
-------
r
TABLE 26. TYPICAL CONCENTRATIONS OF HYDROGEN SULFIDE IN
TOWN GASES

Gas HpS concentration (ppm)

Coal gas 3,200-7,990
Carbureted water gas 800-2,400
Pacific Coast oil gas 3,200-4,000

SOURCE: Morgan, 1926.
89
-------
conversion of the lime to sulfide was usually about 40 percent, so that large
quantities of lime were required for purification of the gas (Veley, 1885).
The spent lime could not be regenerated, and it usually had a foul odor from
the tars and a blue color from the ferric cyanides. After disposal, the CaS
would slowly combine with C02 to rerelease H2S by the reaction:
CaS + C02 + H20 = CaCOs + H2S.
Although some spent lime was sold or given away for agricultural pur-
poses, much of it was discarded. Because it could only be used once for puri-
fication, it was a costly purification method to use. The discovery and use
of the iron oxide process for removing H2S around 1885 replaced almost all the
use of lime for gas purification. The iron oxide process did not remove C02
from the gas, and C02 gave a gas with poor lighting and burning properties.
Some lime was frequently used in a bed directly after the Iron oxide purifiers
to remove C02 from the gas. This use of lime involved much smaller quantities
of lime than were previously employed at operating gas plants.
1.3.7.3 Removal of Hydrogen Sulfide by Iron Oxide--
Iron oxide removed H2S from the gas, was regenerated with oxygen from
air, then reused to remove more H2S. The iron oxide could be regenerated
until it was between 40 and 50 percent sulfur by weight, at which time It was
generally discarded. This regeneration allowed iron oxide to remove much more
H2S than did lime and substantially reduced the cost of gas manufacture.
The relevant reactions for the removal of hydrogen sulfide and regenera-
tion of the spent oxide are below:
H2S REMOVAL
(1) Fe203 * 3H2S = Fe2S3+ 3H20
(2) Fe203 - 3H2S = 2FeS + S + 3H20
REGENERATION
(3) 2Fe2S3 + 302 » 2Fe203 + 6S
(4) 4FeS + 302 - 2Fe203 + 4S
DEACTIVATION
(5) FeS + S « FeS2.

90
-------
r
|
Reaction (1) was the most desirable for gas purification, and it occurred
under slightly alkaline conditions. Reaction (2) occurred under slightly
acidic conditions. The formation of ferrous sulfide (FeS) was undesirable
because it combines with free sulfur to form FeS2 (reaction 5), which cannot
be regenerated. During revivification, some sulfuric acid is formed by the
reactions:
(6) FeS + 202 = FeS04
(7) FeS04 + H20 = H2S04 + FeO.
Some hydrated lime or soda ash (Na2C03) was added to the iron oxide to keep it
in an alkaline state. Some ammonia was usually present (or added) to the gas
passing through the iron oxide to keep the oxide alkaline and to promote the
removal of cyanide from the gas. A small concentration of ammonia apparently
promoted the removal of cyanides as ferrocyanide while a high ammonia concen-
tration caused the cyanides to be removed as thiocyanates.
The iron oxide used for the removal of hydrogen sulfide was of three
major types: rusted iron borings, bog ore, and precipitated iron oxides.
Each of these materials was usually mixed with a fluffing material to provide
for better gas flow through the iron oxide (after 1930, however, some plants
stopped adding fluff material to the iron oxide). The fluffing material was
primarily woodchips, but blast furnace slag and corn cobs were also used. The
iron borings were usually added to the woodchips, then sprayed with water and
exposed to air to rust the borings. Salt or ferrous sulfate was often added
to the water to promote the rusting. Most plants used the rusted iron bor-
ings, but some used bog ore (naturally precipitated iron oxide) during World
War I and World War II, and some plants switched to precipitated iron oxides
after they were introduced about 1930.
The oxides were placed into boxes, and the town gas flowed through the
box. Several oxide boxes were connected in series, and the order in which the
gas contacted the boxes rotated so that gas contacted the most fouled oxide
first and boxes of fresh oxide last. This permitted maximum utilization of
the oxides, while removing the H2S concentration in the product gas to very
low levels. The oxides that contacted the gas first were periodically dis-
carded, the box refilled with fresh oxide, and the box added as the last oxide
to purify the gas.
91
-------
f
Originally, the oxides were revived by physically removing them from the
box, exposing them to air, and then replacing the material into the box. This
was very labor intensive, and because the regeneration of the oxide was exo-
thermic, considerable care was required to prevent the oxides from becoming
deactivated or igniting the tars and bulk material with the oxide. This was
replaced by the practice of reviving the oxide continually while the oxide was
removing I^S. A small amount of air (approximately 2 percent) was added to
the gas prior to the gas entering the iron oxide purifiers. The oxygen con-
tinuously regenerated the oxide in the boxes and greatly reduced the labor
required for the purification. The major disadvantage of this method was that
the nitrogen added to the gas with the air reduced the heating value of the
gas. The oxide was sometimes revived by switching the box out of the combust-
ible gas and blowing air through the oxide.
1.3.7.4 Liquid Scrubbing for Hydrogen Sulfide Removal--
Lime water was the original method of removing impurities from coal car-
bonization gases. It was principally used in Great Britain, but its use was
fairly rapidly replaced by use of hydrated lime 1n beds. The basic process
was to use a solution of hydrated lime in water (milk of lime) and bubble the
raw coal gas through the liquid. Lime removed the hydrogen sulfide as CaS,
carbon dioxide as CaC03, and other impurities by their solubility in water.
Tars and oils were also condensed into the lime water. The contaminated lime
water was generally run directly into the nearest river, much to the displeas-
ure of those downstream. The CaS reacted with carbon dioxide and water to
rerelease hydrogen sulfide while the oils and phenols contaminated the water
and killed fish.
: i Lime water was not used at a significant level in the United States
because, by the time gas was produced, beds of hydrated lime were used instead
of the lime solutions.
1.3.7.4.1 Seaboard process—The first major liquid purification process
for the removal of hydrogen sulfide was the Seaboard process, which was named
^ for the plant in New Jersey where it was developed. This process used a
solution of sodium carbonate to scrub fyS from the coal gas and release the
•
H2$ into the air when the solution was regenerated. This process was invented
in 1920 and installed in 6 plants (with 12 under construction) by 1923 (Bird,

1 02
l
!

ll
-------
1923). It was used on coal carbonization plants, carbureted water-gas plants,
and oi1-gas plants.
figure 30 is a diagram of the Seaboard process. The process used either
two packed columns or a single packed column divided into two sections. In
this figure, the gas is scrubbed in the upper half of the column by a solution
of sodium carbonate (1 to 3 percent). The solution is introduced at the top
of the column and flows down the packing in the column. The gas enters the
middle of the column and flows out through the top of the column. As it
progresses through the column, the hydrogen sulfide and cyanide gases are
absorbed into the solution. The solution then flows to the top of the bottom
column. There it flows over another set of packing and contacts air (blown
into the base of the column and removed from the top of the lower column).
The air strips the HjS from the solution, reviving the solution (actifica-
tion). The reactivated solution is then removed from the base of the column
and returned to the absorber (the upper column). The solution is continuously
recycled, but it must be replenished periodically by adding fresh solution.
The cyanide in the gas is removed as sodium chiocyanate, which cannot be
regenerated to sodium carbonate. Sodium thiosulfate and sodium sulfate were
also formed by side reactions in the scrubber liquid.
The actifier air contains the H2S that was originally in the product gas.
This stream was usually just vented to the environment, although sometimes it
was used as boiler air so that the H£S would be oxidized to SO? and reduce
odor problems created by H^S. Table 27 shows some typical operating param-
eters for the Seaboard process. The removal efficiency of the Seaboard proc-
ess was generally between 70 and 95 percent. The remaining hydrogen sulfide
in the gas was removed by a bed of iron oxide that immediately followed the
Seaboard process. The Seaboard process was extremely efficient at removing
hydrogen cyanide, so that no cyanide would be removed in the iron oxide that
was used with the Seaboard process.
The Seaboard process greatly reduced the amount of iron oxide purifica-
tion required to remove hydrogen sulfide from town gas. Because it discharged
all of the sulfur it removed to the atmosphere, processes were developed that
were similar to the Seaboard process, but that recovered the sulfur as a
byproduct.
93
-------
f
Figure 30. Seaboard process for H^& removal.

Source: Morgan, 1926.
94
-------
TABLE 27. OPERATION OF SEABOARD PROCESS
Plant
B
Gas purified (10^ ft3/day) 5,317 2,557 353
Inlet H2S (ppm) 2,760 6,950 7,100
Outlet H2S (ppm) 145 304 17
H2S removed/day (Ib) 10,250 15,166 29,920
Na2C03 used/day (Ib) 1,000 2,005 149
•V removal of H2S 94.7 95.6 99.8

SOURCE: Herbst, 1931.
95
-------
f:
1.3.7.4.2 Thylox, Nickel, and Ferrox processes—The Thylox process was
developed shortly after the Seaboard process, and It recovered the sulfur.
Two other processes, the Nickel process and the Ferrox process, used the same
apparatus as the Thylox process, but they used different scrubber solutions.
Figure 31 is a diagram of the equipment used for the three processes. The gas
is scrubbed counter-currently with the absorber solution in the absorber. For
the Thylox process, this solution was a mixture of arsenic trioxide and sodium
carbonate. The Ferrox process used an iron compound suspended in soda ash,
and the Nickel process used a solution of a nickel salt in soda ash (Downing,
1934). The foul solutions were then pumped with compressed air Into the
thionizer, where the oxygen in the air oxidized the fyS to sulfur crystals.
Table 28 lists some typical operating data for the Thylox and Nickel proc-
esses. The arsenic, iron, and nickel act as a catalyst for the oxidation.
The sulfur slurry is then drawn from the top of the thionizer, and the sulfur
is recovered by filtration. The Arsenic and Ferrox processes could be used
with either carbureted water gas, oil gas, or coal-carbonization gas. The
nickel catalyst in the Nickel process was poisoned rapidly by cyanide, and the
process could be used only on gases that had low cyanide concentrations. This
limited the process to use only with oil and water gas.
The Thylox, Nickel, and Ferrox processes were all very efficient in the
removal of cyanide, as was the Seaboard process. Cyanide was converted to
thiocyanates in all four processes. Each liquid process also required the
periodic replacement of the scrubber solutions. This was accomplished
either through normal fluid losses of the system (carryover to the iron oxide
beds, spills, evaporation, and liquid loss with the filtered sulfur product),
the continuous withdrawal and replacement of spent solution, or the periodic
draining and fluid replacement of all the scrubber liquid.
The three sulfur recovery processes were fairly efficient in the removal
of hydrogen sulfide (about 98 percent) but were generally followed by an iron
oxide bed to remove the last traces of the H2S. The spent iron oxide from
this type of operation would be expected to contain some of the scrubber solu-
tion that would be carried over from the liquid purification processes. The
arsenic or nickel salts could occur in the spent oxides.
96
L
-------
r
Level Regulator
'*— Skimming Tank and
Sludge Trough
Soda and Thyfox
Mixing Tank
Sulfur Sludge Tank
Gas Outlet

Gas Inlet
Solution Circulation Pump -" "fail Solution / '-Air Compressor
*• fir Cleaner
Filter
Sulfur Cake
Hopper
Autoclave
Sulfur Pans
Figure 31. Thytox process for r^S removal.

Source: Gollmar, 1945.
97
-------
TABLE 28. OPERATION OF NICKEL AND THYLOX PROCESSES

Inlet H2$ (ppm)
Outlet H2S (ppm)
H2S removal efficiency ft)
HCN inlet (ppm)
HCN outlet (ppm)
HCN removal efficiency ft)
Na2C03 consumption
(lb/106 ft3)
A$203 consumption
(lb/10& ft3)
Thy 1 ox
Avg. of 3
coal-gas
plants
4,794
85.2
98.2
322
0
100
0.07
0.022
Total gas volume purified
(105 ft3/day) 3,000-8,000
Nickel salt consumed
(lb/106 ft3)
--
process
Coal and water-
gas plant
4,315
112
97.4
81
0
100
0.06
0.024
14,000
--
Nickel process
Avg. of 4
oil-gas plants
—
--
70-100
—
--
—
0.102
0
._
0.023
SOURCES: Gas Engineers Handbook, 1934; Cundall, 1P27.
98
-------
1.3.8 Cyanide Removal
Cyanide was also an impurity in gas produced by coal carbonization, but
it was produced in only trace quantities by carbureted water gas and oil gas.
The recovery of cyanides for sale was only profitable at the larger coal-gas
plants and only prior to the Haber process for ammonia production (cyanide can
be produced from ammonia and coke). Table 29 shows representative concen-
tration of cyanide in coal gas, carbureted water gas, and oil gas. The cya-
nide in coa] gas was either recovered as a product or was removed with hydro-
gen sulfide. Because both hydrogen sulfide and cyanide are acid gases, proc-
esses that removed hydrogen sulfide generally removed cyanide as well.
The concentration of cyanogen in coal gas was generally between 0.12 and
0.20 percent (Hi]I, 1945). Because cyanide was rarely recovered from coal
gas, the recovery processes will be described in only general details. Addi-
tional details of specific processes may be found in articles by Hill (1945)
and Powell (1922). The Bueb process used a scrubbing solution of ferrous
sulfate in ammonia liquor. Hydrogen sulfide in the gas reacted with the fer-
rous sulfate to form ferrous sulfide. This in turn reacted with cyanide to
form ammonium-iron-cyanide complexes. The discharge from this process is a
light-colored mud, which turns blue on exposure to air. It has a cyanogen
content, (as Prussian blue) of 13.5 percent and an ammonia content of 6 to
7 percent (Hill, 1945). This product is then boiled and filtered, producing
an ammonium sulfate solution and a filter cake of about 30 percent Prussian
blue. The blue mud product can then be converted to calcium ferrocyanide by
boiling with lime (driving off the ammonia), or potassium ferrocyanide by
adding KC1 to a solution of the calcium ferrocyanide.
In the Foul is process, a water-ferrous carbonate slurry (from sodium
carbonate and ferrous chloride) is contacted with the coal gas. The cyanide
reacts with the ferrous carbonate to yield a product of sodium ferrocyanide.
The Burkheiser purification process used a slurry of iron oxide in water to
simultaneously remove both HCN and H2S. Dissolved ammonia keeps the liquid
alkaline and helps remove the cyanide as thiocyanide compounds.
Cyanide was generally not recovered from the coal gas, but was instead
removed with the hydrogen sulfide. The removal of hydrogen cyanide by iron
oxide purification, the Seaboard process, the Thylox process, and lime purifi-

99
-------
L
TABLE 29. CONCENTRATION OF HCN IN VARIOUS GASES
HCN concentration 1n
Gas raw gas (ppm)

Vertical retort 886
Coke oven 516-947
Carbureted water gas Trace to 26 ppm
Oil gas a

aNot listed but known to be comparable to carbureted water gas.
100
-------
cation are described below. The purification processes themselves are
described in Section 1.3.7.
The earliest method of removing hydrogen sulfide was to run the raw coal
gas from the condensers directly through a bed of hydrated lime. The lime
removed the hydrogen sulfide, and the cyanides in the gas would be removed by
iron impurities in the lime. This caused the formation of Prussian blue in
the lime and "gave rise to the technical term blue billy" (Veley, 1885) for
the spent lime wastes.
If cyanide were not removed by a specific process before iron oxide puri-
fication, then the iron oxide would remove the cyanide. Hill gave the follow-
ing possible reactions for *..'ie removal of cyanide with iron oxide:
Fe(OH)2 + 2HCN = Fe(CN)2 + 2H20 and/or
FeS + 2HCN = Fe(CN)2 + H2S.
The ferrous cyanide then combines with ammonium cyanide to form complex
compounds such as (NH^FefCNJs and (NH4)2Fe2(CN)6- The final form of the
cyanide is as complex fern'-, ferro-, and ferri-ferro ammonium cyanide com-
plexes. These chemicals are best identified by their intense blue color. A
large amount of ammonia in the gas, or strong fixed alkali in the oxide,
caused the cyanide to be removed as thiocyanates (either sodium, potassium, or
ammonium thiocyanate). The cyanides were generally disposed with the spent
oxides, although several methods for the removal and recovery of ferrocyanides
and ferricyanides from the spent oxide were developed. These methods usually
removed the sulfur from the spent oxide, then treated the remaining mass with
strong alkalies.
The Seaboard process, which removed H2S by absorption into a solution of
Na2CC>3, was a very efficient process for removing HCN. The HCN was originally
absorbed as sodium cyanide, which is then converted to sodium thiocyanate.
Each mole of cyanide removed requires a mole of sodium carbonate, and the
thiocyanate could be recovered as a byproduct. This was not generally done,
however, because it was usually discarded with spent scrubber solution.
The Thylox process used a solution of sodium carbonate and arsenic triox-
ide solution to remove hydrogen sulfide and recover it as sulfur. The process
also removed cyanide as thiocyanate in a manner similar to the Seaboard
101
-------
process. The thiocyanates would accumulate in the solution and were removed
with a side stream of scrubber liquid.
1.3.9 Tar and Light Oil Treatment
Many gas production plants did not refine or process their byproduct
tars; instead, they sold them to processors, sold them as fuel, used them
onsite, or discarded them. It is beyond the scope of this study to review tar
processing in detail, but several aspects of tar treatment should be mentioned
because they could occur at many of the gas sites. Table 30 is a list of the
tar fractions and major components present in coal tar. The component list
for tar from water-gas and oil-gas processes would be similar to this, except
that there would be no tar acids, tar bases, or nitrogen heterocyclics as
major components. Rhodes (1945) prepared a list of about 350 chemicals that
were identified in coal tar, and estimates of the actual number of compounds
run to 5,000 (Smith and Eckle, 1966). The chemicals contained in water-gas
tars and oil-gas tars would be a subset of this list, with many of the tar
acids and tar bases being present in coal tar appearing only as trace
constituents in water-gas and oil tars.
Raw tars generally did not have very much product value. They could be
burned in the plant boilers for steam production, burned under the benches
used for coal carbonization, sold as boiler fuel to a local company, or dis-
carded. Tars were a resource to most companies, a byproduct that was sold and
produced income. Near the beginning of the industry, tars were disposed
because uses had not yet been developed for them; later, tar/water emulsions
were disposed when they could not be separated. Small plants that did not
produce sufficient tar for recovery or use would discard it rather than spend
money to prevent its release.
Tars were distilled into fractions that could be marketed as products,
and the fractions were frequently treated with acid and caustic washes to
improve the tar quality and remove undesirable components. The gas purifica-
tion system separated the recovered hydrocarbons into two fractions—the tar
and light oil. The tar condensed with water or was rei.ioved with an ESP. The
light oil was scrubbed out of the gas after the ammonia was removed.
The crude light oils (either recovered by the process described 1n Sec-
tion 1.3.3 or distilled as the highest boiling fraction of the tar) were

102
-------
TABLE 30. PRINCIPAL COMPONENTS IN COAL TAR FRACTIONS
Tar fraction
Boiling range*
Major components
Light oi
Middle oi
Methylnaphthalene
Light creosote
Middle creosote
Heavy creosote
To 210
210-230
230-270
270-315
315-355

Above 355
SOURCE: Smith and Eckle, 1966.

aAs determined by ASTM test 020-56.
Benzene
Toluene
Xylene
Tar acids
Tar bases
Solvent naphtha
Tar acids
Tar bases
Naphthalene
Mixed metnylnaphthalenes
Acenaphthene
Diphenylene oxide
Fluorene
Phenanthrene
Anthracene
Carbazole
Chrysene
Fluoranthene
Pyrene
103
-------
usually treated with sulfuric acid prior to additional refining. The light
oil was charged to an agitator (5,000 to 13,000 gallons) to which strong
sulfuric acid was slowly added (66 deg Baume). It was frequently added in
small amounts, followed by removal of the acid and the sludge it contained.
The total acid consumed was about 0.4 pounds of 66 deg Baume sulfuric acid per
gallon light oil treated. The acid layer was removed after 6 to 8 hours of
treatment in the agitator, and the remaining acidity of the oil *as neutral-
iced 0> acJQing 0.06 to 0.12 pounds of sodium hydroxide per gallon of oil.
Several beneficial reactions occurred during the acid treatment of light oil.
These included o>idation and/or removal of sulfur compounds, the removal of
nitrogen bases into the acid, the polymerization of unsaturated organic com-
pounds, the sulfonation of aromatic compounds, the oxidation of unstable
hydrocarbons, and the polymerization of certain aromatic hydrocarbons
(Glowacki, 1945).
The acid sludge waste is a waste product from plants that produced the
light oils. Although the volume produced by the midsized plants was not par-
ticularly large, its acid character and high concentration of tar bases is
cause for concern. This sludge was sometimes treated for the recovery of the
unused sulfuric acid, but it was frequently just dumped or poured somewhere
and burned. It was not burned in boilers because of the high sulfur content
(placing sulfuric acid into boilers is usually not a recommended practice
because of the resulting corrosion). The acid sludge from light oils
recovered from oil or carbureted water gas would be of substantially different
character from that of coal-carbonization plants. The nitrogen bases would be
present in the acid sludge from coal carbonization, but they would be absent
from acid sludge produced from oil-gas and carbureted water-gas production.
The basic technique for separating the tar and light oil into marketable
fractions was distillation. The distillation could be performed either con-
tinuously or by batch distillation. In both types of distillation, the oil or
tar was separated into fractions with similar boiling points. The batch still
was first charged with tar, and the still was heated slowly. The lower boil-
ing fractions of the tar vaporized preferentially at lower temperatures, and
these components were condensed and recovered as a liquid. Condensed frac-
tions of the tar were removed at various times (corresponding to different
104
-------
still pot temperatures). Each of the collected fractions (they were recovered
as the fractions described in Table 30) had compositions and properties that
were generally more desirable than those of the original raw tar.
In continuous distillation, the tar is fed continuously to a distillation
column containing multiple fractionation trays. The bottoms of the column are
continuously boiled, producing vapor that flows up through the trays. The
vapors from the top of the column are condensed and a portion of the conden-
sate returned to the top of the column. This liquid (and the feed) flow down
the column from tray to tray. A temperature profile exists within the column,
and the liquid composition existing on each tray is different, with the higher
trays having a higher concentration of volatile components and the lower trays
containing more compounds that boil at high temperatures. Various fractions
of the tar can be removed at several of the trays. Because it operates in a
continuous manner, continuous distillation was usually'employed at the larger
tar-processing facilities. Batch distillation was used early in the industry
and at smaller processors.
Tar-processing operation sites would have had much more handling and
treatment of the tars than did plants that merely recovered tar and sold it to
tar processors. In many cases, a tar processor was located adjacent to the
gas plant and could receive the tar byproducts directly from the gas produc-
tion plant.
1.3.10 Gas Storage
This section describes how gas was stored at town gas facilities. Tanks
that were used for the storage of product gas were also frequently used for
the storage of tars and waste condensates at gas production plants. Because
these tanks frequently leaked, they were a significant source of contamina-
tion.
The operating basis for the early gas holders was originally discovered
by the French chemist Lavoisier in 1781. His lab-scale gas holder consisted
of an inverted cylindrical bucket in a tub of water. The bucket was suspended
from a cord attached to its bottom, where the cord was run through a pulley
and attached to a counterweight. When gas was placed into the holder, the
bucket rose. The water in the tub formed a seal around the bucket. When gas
was removed from the bucket, it dropped farther into the water. This arrange-

105
-------
merit allowed gases to be collected and removed for experimentation. The
earliest gas holders used by the manufactured-gas industry were of this same
basic design but larger.
Figure 32 shows a diagram of an early single-lift gas holder. The water-
holding portion of the gas holder was usually placed underground or partly
underground. This allowed the earth to support the walls of the water-holding
tank and reduced construction costs. The plant operators soon discovered that
tars could be stored in the gas holder instead of water. This reduced
corrosion of the tanks and allowed the gas holder to serve as a tar tank in
addition to its use for gas storage. Even when tar was not stored in the
tanks, the water contained in the tanks became fouled by water-soluble and
organic compounds in the gas.
The early gas holders used masonry tanks for the water and iron plates
for the bell itself. Alrich (1934) describes the early masonry tanks:
The important consideration of holder tanks in the earlier years of
our Industry was the necessity for water tightness; not only did
foul water leaking from the tanks contaminate the water in wells
upon which even populous communities relied for their supplies, but
the holders [were] frequently located closely adjacent to dwellings,
[and] the buildings were rendered uninhabitable by the foul water
entering through cellars.
He also states that the soils in England were much better suited to the con-
struction of watertight masonry tanks, and when the same designs were applied
in the United States they leaked rather badly.
Many plants also lost substantial quantities of condensate water through
leakage. Because this water was generally recycled to the scrubbers, the loss
of water had to be made up from other external sources. "The question
frequently arose, 'Why does one gas plant have an excess of water and another
plant apparently have none?1 Upon investigating this question we found that
in every case where a gas plant had no excess water there was a pit holder or
some other leaking underground structure through which excess water was
undoubtedly leaking into the ground" (Bains, 1921).
The single-lift gas holder had one obvious problem, the depth of water in
the tank had to be the same as the height of the bell. To increase the size
of the gas holder without increasing the size of the tank, the telescopic or
multiple-lift gas holder was used. Figure 33 is a diagram of a multiple

106
-------
Figure 32. Cross section of single-lift gas holder.
Source: Morgan, 1926.
107
-------
Figure 33. Cross section of multiple-
lift gas holder.
Source: Morgan, 1926.
108
-------
(four-lift) gas holder. The top section (A) would fill with gas first, and
its base would reach the water level in the tank. The top section would then
form a seal with a second lift (B), and together sections A and B would hold
the gas. Subsequent sections would automatically be picked up by the gas
holder as it filled, and the sections would each collapse into the tank as gas
was withdrawn. This allowed greatly increased storage capacity over single-
lift gas holders. There could in principle be any number of lifts, but in
general fewer than five were used. The raised area of the concrete tank
(dumpling) for Figures 32 and 33 allowed the tanks to be constructed with less
excavation of the plant site and the tank to operate with less water. More
concrete is required for this construction than for flat-bottomed tank
construction. The tank bottom was usually flat for the early gas holders or
smaller gas holders.
By 1926, the use of brick to construct the water-holding tank was obso-
lete (Morgan, 1926). Tanks during this period were constructed of steel
plates, and the water tanks could be either below ground, semiburled, or above
ground. For very large tanks, buried or semiburied concrete constructioi. was
used. Small gas tanks were typically constructed above ground, with the
entire tank structure resting on a concrete slab. Any leakage from this type
of tank would be readily visible to the operators.
Waterless gas holders were used at some plants after about 1925. These
were cylindrical tanks that contained a free-floating piston that would move
up and down within the tank as the volume of gas stored changed. The piston
was usually sealed around the edges of the tank by a tar seal (a seal applied
by some mechanical means wjth a layer of tar above the mechanical seal). This
tar would slowly leak down the inside walls of the tank, collect at the bottom
of the tank, and be pumped back to the floating piston seal. Waterless gas
holders were generally used for very large (500,000 to 15,000,000 ft3) tanks,
and the water-sealed gas holders were used for smaller tanks. The tar used to
form the seal was generally produced somewhere within the plant.
The gas holders previously described held gas at constant pressures
slightly greater than atmospheric pressure. The volumes of the tanks were
required to change as the amount of gas stored changed. High-pressure gas
storage tanks were installed at some plants during the 1920's, but they were
10Q
-------
not in common use until after World War II. These steel tanks store gas under
high pressures so that larger volumes of gas can be stored in smaller-sized
tanks. With high-pressure storage, the pressure of the gas in the tank can be
changed as the amount of gas stored is varied, rather than having the tank
volume change.
1.3.11 General Purification Trains for Town Gases
The processes for the production of town gases are described in Sec-
tion 1.2, and Sections 1.3.1 through 1.3.9 describe the various methods of
purifying the raw gases prior to distribution. This section integrates the
production and purification processes by examining several complete town gas
production facilities. These descriptions are not intended to be representa-
tive of all of the plants using a given production process, but they will help
to give readers generic descriptions of town gas plants.
Figure 34 shows a material flowsheet for a typical coal gas plant. This
flowsheet indicates a plant in which bituminous coal is carbonized to produce
coke, tar, and ammonia. Bituminous coal is first crushed and fed to the coal-
carbonizing apparatus (Section 1.2.2). The coal is carbonized to produce coke
and gas (containing tars and other byproducts). The coke is used to manufac-
ture producer gas (Section 1.2.1) to heat the coal-carbonization apparatus; it
can also be sold or used to produce carbureted water gas (Section 1.2.3).
Coke breeze (coke of small particle size) is used in the boiler room. The raw
gas is scrubbed with weak ammonia liquor in the hydraulic and foul main, then
it is cooled in the primary condenser (Section 1.3.2), blown through the
exhauster, and tars are removed by the tar extractor (Section 1.3.3). The
collected tars and condensate are combined and fed to a tar-liquid separator.
Weak ammonia liquor and tar are separated, and the tar is either processed
further or sold as raw coal tar. The ammonia is then scrubbed from the gas
(Section 1.3.5), and ^S is removed by liquid or iron oxide purifiers
(Section 1.3.7). The purified gas is then metered, stored, and distributed to
consumers.
Figure 34 shows a flowsheet that would be typical of small-to-midsized
coal-carbonization plants whose primary purpose was the production of fuel
gas. No recovery of light oils was performed, but the organics that condensed
in the storage and distribution system were recovered as drip oils. The light

110
-------
' POWER
PLANT
LJ J
M *
U
LJ
K
D

COKE SCREENS
LJ
i
<
(
^

, COKE
^

g PRODUCER $
'
MARKET
ee
O •

CARBURETTED
BLUE GAS
PLANT

COAL
<
PILE

SCALES AND
CRUSHERS
<
CARBOI
APPAF
s&
*&* ,
f HYDR>
MA
r
SIZING
IATUS
1 J*
t^ULIC
IN
FOUL MAIN f
+
PRIMARY
CONDENSER
OR WASHER
COOLER
| EXHAUSTER
TAR
EXTRACTOR
FRESH
WATCR
SECONDARY
CONDENSER
OR WASHER
COOLER
AMMONIA
SCRUBBER
WEAK AMMONIA LIQUOR
AMMONIA
LIQUOR
IRON OXIDE
PURIFIERS
STRONG
LIQUOR WELL
AMMONIA
PLANT
DRIP OIL
TANK
DRIP
OIL
ce
o
STATION
HOLDER
DISTRIBUTION
MAINS
Figure 34. Flowsheet for a coal carbonization gas plant.

Source: Morgan, 1926.
Ill
-------
organics that did not condense would just enrich the fuel value of the gas
when the gas was burned. Phenols would be in the ammonia liquor, and the
ammonia liquor could either be sold in this form or the ammonia could be
recovered onsite (Section 1.3.5). The ammonia recovery in this figure is the
indirect process. Cyanide was removed in the iron oxide or liquid scrubbers,
and it was not recovered as a product.
Figure 35 is a flowsheet similar to the one shown in Figure 34, but it is
for a modern (1945) byproduct coke-oven plant. It is indentical to Figure 34
except that (1) phenol is shown recovered from the ammonia liquor (Section
1.3.6), (2) ammonia is recovered by the semidirect process (Section 1.3.5),
(3) light oil is recovered, and (4) liquid purification (Section 1.3.7.4) is
employed for the removal of H2S and HCN, with the recovery of both sulfur and
thiocyanates. This flowsheet would be typical of large byproduct coke ovens.
The products of the process are coke, gas, tar, sodium phenolates, ammonium
sulfate, light oils, sulfur, and ammonium thiocyanate. Although some plants
would recover all the byproducts as indicated by this figure, there would be
many variations of this basic design. As an example, some plants would not
recover light oils, use iron oxide purifiers for h^S removal, or use the
Seaboard process for ^S removal and not recover sulfur. Moreover, some
plants would not recover thiocyanates as a product, would not recover phenols
(they would dispose of them instead), or would not recover ammonia.
Figure 36 shows a material flowsheet for the production of carbureted
water gas, which is described fully in Section 1.2.3. The generator contains
a carbon fuel (either coke, anthracite coal, bituminous coke, or petroleum
coke briquets). Air to the generator, superheater, and carburetor is supplied
by a blower. Carburetion oil is pumped from storage, preheated, and sprayed
into the carburetor. The carburetion oil could be naptha, gas oil, fuel oil,
or heavy residual oils. Waste heat produced during the blows is passed
through a waste heat boiler, which produces the steam sprayed through the
generator. Raw gas is passed through a washbox and condenser (Section 1.3.2).
Because the production of gas is not continuous, a relief holder (Section
1.3.10) is used to dampen the gas flowrate changes and provide a relatively
constant flow through the exhauster, tar extractor (Section 1.3.3), purifiers
(Section 1.3.7.3), and finally to the metering and distribution system. Tars
112
-------
OOOLEK
MAIN ,—v tXKtUSTW ftOOTIR
MOTOR no.
ICNZCNC.TaLUCNC
CMC
SOLVENT NtfMTHA SUlSUft TMlOCYAMTC 6AS
Figure 35. Condensing and collecting system of a modern byproduct coke plant.

Source: Rhodes, 1945.
113
-------
r .i.-
FUEL STORAGE
GENERATOR

OIL STORAGEI * *
WASTE HEAT
BOILER

STACK
WATER
BLOWER
CARBURETTER
—-/SUPER-HEATER

TAR SEPARATOR
EXHAUSTER

TAR
EXTRACTOR
STATION METER
STORAGE HOLDER
MAINS
PURIFIERS
FUEL Off GAS
AIR ox WASTE GASC5 • -
WATER on TAR
STEAM * x-
Figure 36. Flowsheet for a carbureted water-gas plant.

Source: Morgan, 1926.
114
-------
and condensate are collected from the washbox, condenser, relief holder, and
tar extractor. The tar and condensate are then separated in the tar separator
(Section 1.3.3). The product tar was frequently sold as a boiler fuel, burned
in the plant boilers, or remixed with the carburetion oils.
This flowsheet is much simpler than are those for the coal-gas produc-
tion. No ammonia is produced or recovered, no phenols are produced or recov-
ered, and no cyanide is produced or recovered from the carbureted water-gas
process. In fact, some small amounts of phenols, ammonia, and cyanides were
produced by the process, but they were not in recoverable quantities and were
in much smaller concentrations than were those in gas from coal carbonization.
The purifiers generally used iron oxides, although liquid purification could
be employed. Sulfur recovery was practical at some of the larger plants or at
those that used carburetion oils containing a high concentration of sulfur.
Light oils were not. recovered in this flowsheet, but the organics condensing
as liquids in the relief holder are collected and recycled to the tar separa-
tor. The tar extractor was frequently a tower packed with wood shavings in
which entrained tar aerosols would either condense or be removed by impact
with the shavings. The tars conti-insing in the tar extractor would drip to the
base of the tower, then they would be removed and mixed with the other plant
tars. The wood shavings required periodic replacement because heavier tars
would eventually build up on the shavings and plug the shavings scrubber.
This process was much better suited than coal carbonization for use in small
gas plants. Less labor was required to produce gas, the gas was of generally
high quality, and there were fewer byproducts (no ammonia, phenols, cyanides,
and organic nitrogen compounds) to recover or dispose.
Very small gas plants producing carbureted water gas might only operate
the gas production equipment during part of the day and rely on the gas stor-
age holder to supply gas when gas was not being produced. The larger plants,
however, usually operated several separate units (similar to that shown in
Figure 36) to produce the quantities of gas required. Individual units (or
sets) would be started up or taken out of production depending on gas demand.
Figure 37 shows a flow diagram for a typical oil-gas production plant,
the Portland Gas and.Coke Company works. This diagram does not show the steam
115
-------
o

2
7>

•o
£
a
a
o a.
So
01-0
O>T)
a c
I
01
116
-------
and air Inputs that are required for the gas generators, but it does adequate-
ly show the gas-cleaning and byproduct recovery operations. Oil, steam, and
air are used to produce gas in the generator (Section 1.2.4.2). The raw gas
is initially cooled in the washboxes. Most of the lampblack that is produced
by the process deposits in the washbox, along with the heaviest of the tars.
The gas is then scrubbed at lower temperatures in the tar scrubber to remove
tars (the scrubbing liquid is not shown on the figure, but it would usually be
recycled water and condensate from the tar separators). The gas is temporar-
ily stored in a relief holder, and it is then scrubbed to remove more tars
prior to being exhausted into the iron oxide purifiers. After this step, the
light oils are scrubbed from the gas, which is followed by the storage and
distribution of the purified gas. Steps including the wash-oil to the light-
oil scrubbers and the recycled condensate to the second condenser-scrubber are
not shown.
The lampblack-heavy tar-water mixture is fed to a thickener to remove
some of the water from the mixture. The thickened sludge is then dewatered,
dried, packaged, and sold. The lampblack product could be sold as fuel, bri-
queted (for use in water-gas generators or sold as fuel), or burned in the
boilers of the plant. The lampblack could also be slightly dewatered prior to
burning in the plant boilers. Sometimes the lampblack was not recovered at
all; instead, it was merely routed from the washboxes to an appropriate
lagoon. The raw tar and condensates were separated in gravity tar separators
(Section 1.3.3), which was followed by the dewatering of the collected tars.
The product tars were then distilled into marketable fractions and sold. Wash
oil containing light oils and naphthalene was regenerated by distilling the
light oil and naphthalene from the oil. The light oil and naphthalene were
then separated in a second still. The recovered light oil was then acid-
washed and distilled into marketable fractions. The acid washing of the light
oil produced a waste acid sludge, but this sludge would be substantially
different from the acid sludge produced from the acid washing of light oil
from coal carbonization (Section 1.3.4).
The oil-gas plants along the West Coast frequently operated as extensions
of oil refineries. The petroleum refiners would sell residual oils with high
carbon contents to the gas companies, and the gas companies would use it to
117
-------
produce a wide array of petroleum products in addition to gas. The light oils
and tars that were cracked from the oils had sufficient market value that the
plants recovered and sold them, rather than recycle the light oils into the
gas generation oils. The units were operated essentially as thermal crackers
of petroleum, producing lampblack, tars, light oils, and gas as products.
The production of oil gas was not accompanied by the production of signi-
ficant amounts of phenols, cyanides, ammonia, and base nitrogen organics.
These constituents would not be recovered at plants that produced exclusively
oil gas, and they would be present only in trace amounts in any wastes from
the process. The amount of hydrogen sulfide produced in the gas was propor-
tional to the sulfur content of the oils used in the generators. Sulfur
recovery processes could be used to remove hydrogen sulfide from oil gas.
Many possible variations are possible for this flow diagram. Smaller
plants whose primary purpose was the production of gas would probably not
recover the light oils from the gas. Most of the light oils would remain in
the gas and enrich the heating value of the distributed gas. Many plants
would sell the raw tars to distillers, rather than distill it onsite. The
recovery of lampblack could vary and would range between disposing of the
washbox sludge and condensate to complete recovery and use of the lampblack.
If the plant also produced gas by coal carbonization (e.g., as was done at the
one in Seattle, Washington), the lampblack sludge could be mixed with
bituminous coal prior to coking. The tar in the lampblack would be added to
the recovered coal tars, and the carbon would be added to the coke produced.
1.4 BYPRODUCTS AND WASTES,FROM TOWN GAS PRODUCTION
1.4.1 Introduction
Each of the three processes for the production of town gas also produced
nongas materials that were not directly related to the production and distri-
bution of combustible gas to consumers. These materials could frequently be
recovered, recycled, or sold but were also disposed at some production plants.
The only difference between byproducts and wastes is that, 1f a material could
be sold or given away, it was considered a byproduct, but if the material were
discarded it was considered a waste. This distinction between byproducts and
wastes is somewhat unimportant for the types of waste disposed on or near gas
118
-------
sites because some byproducts would spill or leak at the site and off-spec
byproducts may be disposed. It is important for the quantity of wastes
disposed, however, because some materials were always disposed while others
were frequently recovered.
Several factors affected whether a given byproduct was recovered or dis-
posed. If there were no market or use for a material, it was considered a
waste for disposal. Sometimes these wastes did have a value as fill (such as
spent oxide, ash, lampblack, clinker, and broken firebrick) and were used as
fill around the plant or given away as fill. Some potential wastes such as
ammonia, phenol, lampblack, and tars could be recovered and sold, but they
were often not recovered because the price for the material did not justify
its recovery. Any material that was recovered at a gas site was a potential
waste because some of the products would not meet marketable standards.
1.4.2 Description of Wastes
1.4.2.1 Coal Tar, Water-Gas Tar, and Oil-Gas Tar--
When most people think of tar, they generally remember the -tars that they
have seen. These are principally either road or roofing tar, which is usually
a solid but pliable material that softens as its temperature is increased.
The prospect of this tar flowing through the ground or contaminating water is
remote, even to the casual observer. However, the raw tars produced by town
gas processes were frequently liquids at ambient temperatures with viscosities
sometimes not too different from water. Tars were considered to be any
organic liquid that was more dense than water (density > 1 g/cm^). The tars
would sink to the bottom of the tar separators, with the water forming a sepa-
rate layer above it. The tars that collected in this manner generally had
organic compounds normally associated with light oils, but they were dissolved
in the heavier tar layer. The range of tars produced for the manufacture of
town gases was considerable, ranging from tars that were slightly more dense
and viscous than water, to tars that were solid at ambient temperatures and
required heating before they could flow. Raw tar properties varied substan-
tially within individual production processes because the heavier tars usually
condensed in the washbox and lighter tars in the condensers.
Tar was usually defined as a nonaqueous viscous liquid of very complex
composition produced by the destructive distillation or partial combustion of

119
-------
organic matter. The tars produced by town gas processes fit into three gen-
eral categories, depending on the production process. Coal tars were tarry
liquids produced by the partial combustion or destructive distillation of
coal. They were usually further classified by the specific process that pro-
duced the tar, but they were divided into two major classes: high-temperature
tars and low-temperature tars. Coal tars contained principally aromatic
hydrocarbons: benzene, naphthalene, anthracene, and related compounds. They
also contained phenolics and tar bases.
Oil tars were tarry fluids produced by the destructive distillation or
thermal cracking of petroleum oils. The tars produced by the major oil-gas
processes were high-temperature oil tars. They were composed principally of
aromatic hydrocarbons; benzene, toluene, naphthalene, phenanthrene, and methyl
anthracene were reported components. Other complex aromatic hydrocarbons are
also present. "No true anthracene has been identified 1n any of the American
oil tars. They are further characterized by the almost entire absence of tar
acids and tar bases, and this seems to constitute the chief difference between
this type of tar and high temperature coal tar" (Bateman, 1922).
Water-gas tar is the tar produced from the oil that is cracked from
petroleum oils in the carburetor of carbureted water-gas (CWG) machines.
Water-gas tar is very similar to tar produced by oil-gas manufacture. It also
is very similar to coal tar but "could be distinguished from coal tar only by
its lack of phenolics and tar bases" (Bateman, 1922).
Table 31 shows typical analyses for various types of raw coal tars. The
tars from horizontal retorts, vertical retorts, inclined retorts, and coke-
oven coal tars are listed. Table 32 compares the properties of two CWG tars
to three types of coal tars. Table 33 lists the properties of three oil tars
produced by the Pacific Coast oil-gas process. The properties listed in these
tables are those of the raw tars produced by the processes. These properties
reflect the mixing of the tars that condensed in various parts of the
purification train. The properties of the tar condensing in various parts of
the purification train would be substantially different. Tars condensing in
the hotter portions of the purification train (e.g., the washboxes) would be
higher boiling and more viscous than would be tars condensing in the cooler
sections of the purification system (e.g., secondary scrubbers or the tar
extractor).
120
-------
E .1
v ••
n
o
h
V)
IT
O
u.
O
—-MTI(~C
'vl Itlt'iilii
i-r e»o« S
J« Z ? .

S g |7'
b.

,V
5 I
<• s r—rt"T 5—SSSct—JM
* •- B L 1 V L L I i I -I i 1 1
j. X

i
Sg

8J
^i
It
-3
II
kta,
Cl
01

§
§
121
-------
(9

9 -
Cl

3 b
o e-. « '_2 «s CJ i-

S — -r i- I o «o — a) to o 6 — ySboorsOoe
•i ci -« e-i to
10 c -» ?i O >o O
«f i- ri «s • tr. •
" 5 ^ e P S
ooSir. — o ri d <-rs
— o c ri ee
— v: rj
— rf si e
S»o«sc»v:ri
P-
§0 O

_ *l O C7 I-

— O O « —
m s o
Q y si
s o ?i — w
1 "?

ec
rs ^
•s ^ ^
•- •«

>. ^ ?
£
t
•~U. f~ % V. U- C
'/. V. S •/. % i. — K M U.
•o
o
E
2

if
122
-------
fl
TABLE 33. COMPARISON OF SOME PACIFIC COAST
OIL-GAS TARS
Kiittiplu
Six-rifle uravily at GOT
Insoluble in CSi, per cent hy wt
Specific viscosity Englcr, 50 cc at 40"C (104"F)
Float test, sec at 32'C
Softening point (ring and ball), *C
Distillation: per cent by wt.
To 210"C (4IO°F)
To 235'C (455'F)
To 270°C (518'F)
To 315'C (599°F)
To 355°C (671°F)

"K
SlM'cifir urnvity totnl diatillntc at 60*K
Sulplionation rcxidun, total dixtillHtc (per cent vol.) . . .
1
1.200
12.5
13.2

5.8
16.6
26.1
33.0
41.3
58.7
105.5
222
1.071
2.55
2
1.297
24.2
247
2.7
15.5
20.6
24.2
31.0
60.0
140.0
284
1.115
0.30
3
1.334
30.7
33.8
1.2
4.4
8.0
13.4
23.2
76.8
137.0
279
1.120
trorc
4
1.317
28.7
32.fi
2.7
10.8
14.5
18.4
27.6
72.4
148.5
299
1.110
1.6-2
•Sample 1: Mcd.-tcmp. fuel oil-Ran lar.
Snmpl
-------
r
In general, the raw CWG tars were less dense and less viscous than were
tar.> produced by coal carbonization. The low-temperature coal tar in Table 31
has a lower density and viscosity than do the coal tars, and this reflects the
lower temperature of carbonization. Low-temperature coal carbonization was
not employed in the United States to any great extent, however. The lower
viscosity of the CWG tars means that they are generally more mobile and flow-
able than are the raw coal-gas tars. They generally also have a much lower
carbon content than do the coal tars. The specific properties of the CWG tars
depended substantially on how the plant operated the CWG apparatus. Because
gas production occurred in cycles, the carburetor and superheater started out
very hot when the oil was first injected into them to produce gas. They
cooled relatively rapidly, requiring that the production of gas be stopped and
the apparatus reheated. Hence, the cyclical nature of the process actually
alternated between heating the apparatus and cooling the apparatus while pro-
ducing gas. When the gas production part of the cycle began, the apparatus
was at its highest temperature. The high temperature tended to overcrack the
oil, producing very heavy tars, carbon, and gas. As the apparatus cooled, the
lower temperatures tended to undercrack the oils, merely vaporizing the oils.
An apparatus that was operated at higher temperatures produced tars that were
higher boiling, denser, more viscous, and had higher carbon contents than did
an apparatus operated at lower temperatures. Apparatus operated at low
temperatures produced tars that more resembled the original feed oils.
The oil-gas tars (Tables 32 and 33) highly resembled CWG tars because
both were produced principally by the thermal cracking of petroleum products.
The discussion above regarding the properties of CWG tars as related to the
operation of the gas-manufacturing apparatus applies to the production of oil-
gas tars as wel1.
The tars produced by oil-gas and CWG production are very similar, and it
would be very difficult to distinguish between the two. The oil-gas tars,
however, would generally have higher carbon contents than would CWG tars. The
petroleum-based tars (CWG and oil tars) can be distinguished from coal tars by
the presence of phenols and nitrogen-containing organics in the coal tars.
The amount of tar produced by CWG or oil-gas production depended on the
oil used for the gas manufacture. The first carburetion oils used were the
124
-------
11
naphtha fraction of petroleum. This was the fraction that was between gaseous
hydrocarbons and kerosene. It made an excellent carburetlon oil and produced
only a small amount of tar (which was probably not worth recovering). The tar
produced when using naphtha was only 1.7 to 3.5 percent of the original carbu-
retion oil (McKay, 1901). No analyses of the tar from naphtha were dis-
covered, but it would be very similar to that of the CWG tar 1n Table 32. The
tar from naphtha would probably be slightly less dense, less viscous, and
contain more lower-boiling hydrocarbons than would be tar from gas oils, and
it would be fairly mobile. After World War I, the increased'demand for
gasoline (produced from the naphtha fractions of petroleum) led gas producers
to switch from naphtha to gas oils. The gas-oil fraction of petroleum was
between kerosene and lubricating oils. The gas oils produced more tars than
did the use of naphtha, CWG tars (produced using gas oils) were between 12.3
and 18.3 percent of the original oil volume (McKay, 1901). The use of gas
oils became less attractive after catalytic cracking of the gas-oil fraction
into heavy fuel oils and gasoline was adopted by petroleum refineries. This
alternative use of the gas oils competed with the gas Industries' use of the
oils, increasing prices and causing some shortages of gas oil. The industry
subsequently switched from gas oil to heavy fuel oils for the manufacture of
CWG. The tars produced from the use of gas oils became know as light water-
gas tars, and those from heavy fuel oils or residuum oils were called heavy
water-gas tars. Table 34 is a comparison of light water-gas tars, heavy
water-gas tar, and coke-oven tar. The heavy water-gas tar was denser and more
viscous than was the light water-gas tar. It had much more carbon than did
the light CWG tar and fewer low-boiling organics. The use of heavy fuel oils
for carburetion also increased the amount of tar formed from the production of
CWG to up to 25 percent of the oil fed to the process.
Odell (1922) described water-gas tars as follows:
In the carburetion of water-gas the aim is always to convert as much
of the oil used as possible Into fixed gases; the conversion, how-
ever, is never 100 per cent complete, but invariably appreciable
amounts of tarry condensable matter or carbon, or both, form 1n the
checkered chambers. This condensable matter, which 1s water gas
tar, may be composed of substances resulting chiefly from the crack-
ing of oil, or it may consist, in part of some of the Ingredients of
the original oil which resisted cracking. As produced by the var-
ious plants water gas tars are not uniform in character, but may
very materially differ 1n their chemical and physical properties.

125
-------
TABLE 34. COMPARATIVE ANALYSES OF TYPICAL, LIGHT, AND
HEAVY WATER-GAS TARS AND COKE-OVEN COAL TAR
Property
Water, per cent by volume
Specific gravity at 15.5/15.3'C
Specific viscosity. Englcr, 50 cc at 40'C
Float test at 32°C, seconds
CSi insoluble, per cent by weight
Distillation, Engler, per cent by weight
to 170°C
to 235'C
U»300*C
to 355*C
Residue at 300'C, per cent by weight
Residue at 355*C. per cent by weight
S.I', of residue at 300'C, R & B
S.I', of residue nt 355'C, 11 & B
Sulformtion index, 0 to 300'C
Sulfonnlion index. :t()0* to 3SS*G'
T.-ir nci'ds, per cent liy volume
IjRlll
wiitcr-
R:IS tnr
0.5
l.OSd
2.0
—
1.1

1.0
12.1
44. C
07.6
fti.4
31.0
34 'C
UU'O
1.2
0.2
Xonc
Heavy
wntcr-
RIIS Uir
1.1
1.212
—
74
8.9

0.2
4.2
1C. 4
31. 8
83.1
07.0
ore
we
1.0
0.8
None
Coke-oven
coal tar
_
1.19S
—
3*
7.8

7.3
—
21.5
35.8
—
—
44*C
7fC
Tmcc
Trace
1.53
Source: Rhodes, 196Sb.
126
-------
The tar may be brown in color, thin or watery 1n consistency,
contain a large percentage of light oils, and have a specific
gravity but slightly greate7- than 1.00, or it may be a black liquid
of the consistency of molasses, containing a much smaller percentage
of light oil and a specific gravity as high as 1.15. Furthermore,
the percentages of free carbon and naphthalene are different in the
various tars, varying from almost zero to a relatively high
percentage.
The amount and character of coal tar produced by coal carbonization
varies substantially with the temperature at which the coal is carbonized. As
the carbonization temperature is increased, the amount of gas produced
increases because more of the tars are converted to coke and gas. Figure 38
shows how the yields of gas, light oil, liquor and ammonia, tar, and coke
change as the carbonizing temperature is increased. The carbonization temper-
ature also affects the composition of the tar produced by the process. Fig-
ure 39 shows the effect of temperatures on the tar composition. As the amount
of pitch residue in the tar is increased, the tar density and viscosity also
increase. The carbonization temperature for coke ovens was about 850 to
900 °C, and the horizontal retorts operated at higher temperatures of 1,000 to
1,100 °C. Figure 39 shows that the tar produced from byproduct coke ovens
would have more tar acids and less pitch than would tars produced from hori-
zontal retorts. Rhodes (1945) gives a complete account of the effects of
coal-carbonization temperature on the byproducts produced during the coking of
coal.
Table 35 shows the amount of tar produced by each of the gas production
processes. The estimates of the amount of tar produced for each process
should be considered approximate but representative of the amount of tar
produced by each process. This table shows that the production of CWG by
naphtha produced very little tars, while oil-gas production was a very large
tar producer. The large amount of tar produced by the oil-gas process
reflects the use of high-boiling residuum oils for the production of oil gas.
When heavy fuel oil was used for CWG production, the tar production increased
relative to the tar production when gas oil was used.
The coal, CWG, and oil tars each had substantial uses that generally
justified their recovery and use. It is not within the scope of this study to
review the multitude of uses for tar products, but some mention of the major
127
-------
r
15

10
c 0
£
•o"
S 5
80
75
70
Gas
Ammonia and Liquor
. i i
500 600 700 800 900 1000 1100
Carbonizing Temperature. *C
Figure 38. Effects of carbonizing
temperatures on yields of
primary products from Pratt
coal.
Source: Rhodes, 1945.

Note: Yields of primary products plotted
individually.
128
-------
70
65
60
55
50
45
40
§15
£20
c 15
o
510
10
5
2.5
0.
Pitch
Residue above 350 * C
Paraffins and Naphthenes
J I
Olefins
Tar Bases
T
T
500 600 700 800 900 1000 1100
Carbonizing Temperature, *C
Figure 39. Effects of carbonizing
temperatures on
composition of tar from Pratt
coal. source: Rhodes, 1945.
Note: Concentrations of tar constituents
plotted Individually.
129
-------

<
u.
c
UJ
c
p
(J
£
3
4
2

u
3
H~
z
M
£
a
a.
CC
u.
o

UJ
1-
l-l
t-

H
^™
X
O
tt
a.
a.

.
m

(^
J
0
^
K-

C
o
•K
4
u
;
•^
e
L
O
L
1-

n
c
o
••
<™
Sj1
•vo
tl
c —
O —

^
a
O

1
*»
*»!
••* ^D
•»» o
tl U
C f>
o

"• c
o

_
n
o
cS
5 —
-J C
>>. o
« M
• ^
0 •
— o
i

H
K
U
£
a.

J>
^xo o
n oo M
» ^fH r4|
III III -S>
Q Q Si S « Q MS
CD CM N UJ P- LO S
r. co co * oo
«•»

nun
CM U) CJ
>-< r4 CM III II
III III II
00 •* 11

^^
E
n •
t! • 2 ^^
C O •*> » 4>
• *> i. n
> • O K « •
o u *> s • o
c • -« a u
o • • u —
- je • - L. .- u
*> o a « «i • o •
• U • 3 J *- — —
N — at o «• — •-••
-•»>• a * • o j: j:
c u *>— c a • » n
J D C • — TJ <• — V- 0. 1 1
ji -a o o *> • j: — «->•••
t- o N — c .*> *; o x — —
• U .- *> O *JC > • 01 Ji
u o. i. u o i. a. n • • c 3
>. o • --^ 3 • • • n— o
— CD X > ^ZUX MO
• u —
o • .-
W 00

(
^
ro

f-l

Jt
j

^J
e

• • *

:t-< M T
SO) O>
e Af-i fi
a • •

HI jfc ^ A
J??
^ a
U "O
130
-------
uses of tars is clearly appropriate. Rhodes (1945) divides the products from
coal tar into two classes: the principal crudes and chemicals. The principal
crudes are products produced directly from the distillation of raw coal tars.
They include wood-preserving oils, road tars, industrial pitches, and pitch
coke. The major basic chemicals produced from coal tar are naphthalene, tar
acids (phenolics), and tar bases (nitrogen-containing organics). Raw coal tar
also contains some light oils (about 2 percent) with lower-boiling organics
such as benzene, toluene, and xylenes. The manufacture of coal-tar products
has been reviewed by Smith and Eckle (1966), and the commercial aspects of
coal-tar pitch have been studied by Doerr and Gibson (1966). Rhodes (1966a)
examined the history of coal-tar and light-oil use, and he also examined the
uses of coal and water-gas tars. The uses of heavy CWG tar and oil tar and
the uses of light CWG tar are listed in Table 36. The principal use of CWG
tars was as a fuel. The CWG tars could be burned in the plant boilers,
replacing the coal that would normally have to be consumed. The tars
therefore had a minimum value as fuel to the plant and would be burned if they
could not be sold for a price that exceeded the fuel value of the tars.
The tars produced by town gas processes were generally recovered as a
byproduct of the plant operations. There were several reasons why tars were
disposed rather than recovered at gas production plants:
• Early plant operators disposed of tars rather than make
attempts at recovery.
• The production of off-spec tars that could not be sold
occurred, and these tars were either burned or discarded.
• Small gas plants were likely to dispose of tars in that they
frequently did not produce enough tar to make recovery practi-
cal .
• Tars (particularly water-gas and oil-gas tars) frequently
formed emulsions when the tars condensed with the steam. These
emulsions could usually be broken, but when several attempts to
break the emulsion into separate tar and water fractions
failed, it was disposed. In some cases, plants would not even
attempt to sep rate the emulsions. Instead, if the tar did not
separate from the water in the gravity separator, it was dis-
posed. (Emulsions are discussed in Section 1.4.2.3).
131
-------
Ul
u

u.
o

3
«n

CM
c
o
* w •
« c
u « —
O II
-O 4» O

e «'t-

• u
TJ • •
• « —
c • «
c «. t.
a • «. an
ja -o o jj o> e
•^ —««- e o c o
o — o a —
— X •-•».*> 4)
• it *> a • •
a M jj <• c o L.
«• • — > • u •
a a i. 6 « a
— T> • -o a «
• mncnoec— n

u • • u t. • —
•a *>— a— i»— •
« L c. JD *J • u
3T3««-aace«
•0*4)^06 c
c o «— o a • c —
w o: 3B u. r
Q.
u
a.
a
o
« o
c x »
o u
— *) L.
« — o
u a«»- •
• • ^ c
a.->e t •
u « o -o • — —
• — O O 4* • O
-U-O I. OX
X 4» • *»
— -o — o u ox x

3 O O ••• •" I. H *~
u.o:ua.uiuz_i
a
i
i.
L O
3
X M
M C

3» *
I ^N
b —
• XI.
!>; H- —
• ^ o
132
-------
Prior to the discovery that coal tar could be used to preserve wood 1n
1838, there were very few uses for the raw coal tars except as a fuel. Coal
tar was not distilled in the United States prior to the early 1860's; hence,
there was really no market for the raw tars. Many of the early plants dis-
posed of the tar with the condensates from coal carbonization. This was done
by whatever methods were most convenient for the plant, which generally meant
running the wastes into the nearest body of water. Because most of the early
plants were located along the coast, this was often done without causing
noticeable problems. If the wastes could not be discharged to water, a pit or
lagoon would often serve as a substitute.
Raw coal tar could be burned at the plant rather than be merely disposed.
Undoubtedly, some plants did recover and burn much of the tar they produced
during this early period. Hughes and Richards (1885) states that:
When there is not a sale for the tar, or when there is a great
demand for the coke, tar may be employed advantageously for heating
the retorts, thus entirely replacing coke for that purpose...In a
works having only six benches, or settings, the yield of tar would
be sufficient to heat one of them.
Although Hughes' book was published in England, it was a seventh edition (the
first edition being published around 1850), and it was probably common know-
ledge that the raw coal tar could be burned to heat some of the horizontal
retorts. In any event, the early production of town gas was principally for
the lighting of streets and shops, and then only during a certain portion of
the evening. The gas produced was too expensive for people to use in their
homes, and the amount of gas produced was relatively small until after 1865,
when people started to use gas lights in their homes (Stotz and Jamison,
1938). It is probably impossible to reliably estimate either the total amount
of tar produced during the early years of the industry or to determine how
much of the tar was burned as fuel. The amount of tar disposed by methods in
which it could be a hazard today would also be very difficult to estimate
because much of the early waste tars were dispersed.
Some tars that were disposed by the plants early in their operation would
not continue to be disposed during later operation. In 1896, Grimwood
described the recovery of CWG tars:
133
-------
Aside from its commercial value and supposing it to be deficient in
all the compounds which make coal tar a marketable residue, there
arises the question of how to dispose of 1t without cost and annoy-
ance to the neighborhood...In 1891 and 1892 all, or nearly all, our
tar was a waste product; now we have a good and sufficient market
for what we do not use...The raw tar Is of no value except as
fuel...Mixed with anthracite screening and coke breeze 1t makes a
very fierce fire and serves admirably as a boiler fuel--a use to
which I believe it is universally put.
Plants that recovered tars either for sale as raw tars or refined onsite
into products would often produce tars not of marketable quality. Sometimes
these tars could be mixed with better tars to produce an acceptable product,
they could be burned or mixed with the coal prior to carbonization, or they
could be discarded. The most common tar product likely to be discarded is the
coal-tar pitch remaining in the still after the lighter fractions have been
distilled from the raw tar. This tar had to be handled hot, in that it would
solidify at ambient temperatures. Burning the tar meant that it had to be fed
to a fire somehow, and the equipment for burning this heavy a tar would not be
commonly available. Holding the tar for any length of time meant either
heating it continuously or letting ft solidify and then remeltfng it at a
later time. Hence, the most expedient way of dealing with off-spec heavy tars
was merely to add them to the waste dump.
Small gas plants had substantially less Incentive to recover tars than
did the larger gas plants. First, the small gas plants generally produced
much less volume of tars than did larger plants. The least expensive way of
dealing with the tars and condensate was to run them into a stream, or along
the railroad tracks, or into a lagoon or pit. Because the volumes were gener-
ally small, this method of disposal would not create immediate problems.
Because CWG or oil gas was generally less expensive to produce at small
plants, the disposal of waste condensates by this method was more common at
these small CWG plants. Vincent (1907), in a discussion on the removal and
disposal of tar, stated:
I have noticed in a rather superficial Investigation that probably
the large majority of quite small gas companies are allowing the tar
to run to waste, generally creating a nuisance in the community and
also wasting a very valuable product...Tar can be burned under the
boilers with equipment any ordinary workman can make: and while they
cannot make enough to run the plant, the whole year around, they can
make so much of it that it will ultimately reduce the cost.
134
-------
J.A. Brown (1926) discusses the economics of removal and disposal:
In the small plant the expenditure of every dollar 1s of such impor-
tance in the monthly cost sheet that extraordinary caution Is taken
and resourcefulness exercised to avoid the expenditure. The small
gas plant has nearly the same equipment as the large one, differing
mainly in size. Any lack, loss, or failure to function results in
much large loss in efficiency in the small gas operation. Any addi-
tional labor or repair expense so looms up :n the cost of gas in the
holder that the small plant operator is particularly skilled at
avoiding such expense.
Small gas plants were also more likely to use CWG using naphtha as a car-
buretion oil than were larger plants. They were slower in converting from one
oil feedstock to another because of the high capital cost of the conversion
relative to the quantities of carburetion oil they consumed. Naphtha produced
only a small amount of tar, and disposal of both the condensates and the tar
were very likely.
1.4.2.2 Oils-
1.4.2.2.1 Carburetion oils—The carburetion oils used in the production
of oil gas and CWG were not intentionally disposed, but it was normal for some
of the feed oils to be spilled while transferring the oil or to leak from
storage tanks. These oils ranged from low-boiling naphtha fractions to
higher-boiling, high-carbon residuum oils. These oils would be, in general,
much more mobile in groundwater than would be the tars produced from the oils.
It is possible that at some gas production plants the major contamination
could come from either an old spill of the oils or a steady unnoticed leak
from oil storage tanks. -The carburetion oils are described with their
respective processes in Section 1.2.
1.4.2.2.2 Light oi I—Light oils were recovered from oil gas, CWG, and
coal-carbonization gases (the process is described in Section 1.3.4). Light
oils were scrubbed out of the gas by a relatively heavy oil, then the light
oil was separated from the heavier oil by distillation. Light oils would not
be disposed as a waste, but leakage and spills of the scrubbing oils, or dis-
tilled light oils, could create local areas of contamination at gas plants.
The composition of light oils is described in Section 1.3.4. They are com-
posed principally of light aromatic hydrocarbons (benzenes, toluenes, and
xylenes).

135
-------
1.4.2.2.3 Drip oils—Drip oils were any hydrocarbons that condensed as a
liquid in the gas holder, meter, or mains. After the gas was produced and
cleaned, some hydrocarbons remained in the gas. As the gas cooled further to
ambient temperatures, some of the heavier organics condensed out as a liquid.
This organic condensate (or drip oil) was collected in special tanks at the
low end of the gas mains. The drip oil was collected and either mixed with
the raw tar or recovered light oils. It had a composition similar to the
recovered light oil. Because the drip oils were collected in a separate tank
(usually underground), some of the drip-oils could leak from the tanks and
into the surrounding soil. Some of the drip-oil tanks remaining at gas sites
may also be intact and may possibly contain the drip oils. Drip oils were not
considered wastes because they could always be added to either the raw tar,
light oils, or carburetion oils.
1.4.2.3 Tar-Oil-Water Emulsions--
The difficulties of recovering tars from tar-oil-water emulsions were one
of the major headaches that plagued the operators of CWG and oil-gas plants.
A tar with a high water content could not be sold (buyers specified low water
contents for purchased tars), could not be burned (a water content below
25 percent is required for the combustion of the tar), and could not be dis-
posed offsite (local sewer authorities would not permit the disposal of the
emulsion in the sewer system, and the emulsions would contaminate a large
•
amount of water if dumped into a river or lake). The emulsions were, in
short, a problem nobody particularly wanted to deal with. As a relatively
difficult material to separate, some of the gas plants disposed of the emul-
sions, rather than spend the time and effort to break the emulsion. The emul-
sions are tars that were very likely to be either stored for long periods of
time or discarded.
Emulsions were not usually a problem in the production of coal-carboniza-
tion gas. The coal tar separated relatively cleanly from the condensates, and
each could be recovered using only gravity separators. The formation of emul-
sions was a problem occurring primarily in the production of CWG and oil gas.
There was really no problem with the formation of emulsions when paraffinic-
based oils were used as the carburetion oils. The tars produced almost always
separated cleanly. The emulsion problem began when the manufacturers of CWG

136
-------
had to switch from paraffink-based petroleum oils (produced 1n the East) to
asphaltic-based oils (produced in the Texas Gulf region and California). The
emulsion problems became even more acute when the oil feedstocks were switched
from gas oils to heavy fuel oils. As described by Bennett (1935):
Since 1903 gas oils of asphaltic base have been used. Lately heavy
fuel oils have attained a wide-spread use as enriching material.
Their use has invariably resulted in permanent emulsions which do
not respond readily to the ordinary method of separation, i.e.,
settlement. The reason for this disturbing condition can be found
in a brief examination of the petroleum industry. Pennsylvania
crude oils (paraffine base) present no dehydration problem to the
oil producer nor to the tar producer since emulsions in the field
are unusual. As the field progress westward crude-oil emulsions
steadily become worse and the ratio of asphalt to paraffine base
oils becomes greater. The California fields In general produce the
most stable emulsions and contain the highest quantity of asphalt
bases. The carbureted water gas industry's shift, since 1903, has
been a change from gas oils, principally of paraffine base, to oils
which contain and produce more asphaltic constituents.
Numerous papers deal specifically with the problems of the formation and
separation of tar-oil-water emulsions (Barlow and Kennedy, 1922; Hauschidt,
1922; Odell, 1922; Simmons, 1924b; Seely, 1927; Seely, 1928; CarsweTl, 1928;
Morgan and Stolzenbach, 1934a; Morgan, 1934b; Parke, 1934b; Parke, 1935a;
Oashiell, 1935; Bennett, 1935; Leuders, 1942; Petrino, 1947; Young, 1947;
Hall, 1947; Glover, 1951; Laudani, 1952; Costigan, 1953; Costigan, 1954;
Schwarz and Keller, 1955). The volume of information specifically addressing
the problem of emulsions indicates both the size of the problem to the
industry and the efforts expended to solve the problem.
According to Odell (1922), the emulsions are formed when the raw gas is
cooled, and the water, tars, and carbon are removed simultaneously. The pres-
ence of uncracked oils in the tar and finely divided carbon made the emulsion
more stable. Rapid cooling of the gas created emulsions because the droplet
size of tar and condensate is very small, creating a more stable emulsion.
The practice of dumping all of the plant tar and oils into a common receiver
also assisted in the formation of emulsions. When the tars collected from
different parts of the purification train were collected and treated sepa-
rately, emulsion problems decreased.
137
-------
A poll of 50 large CWG producers (Seeley, 1927) revealed that most of the

plants had experienced emulsion problems at one time or another. Sixty-eight

percent of the CWG manufacturers using coke as generator fuel reported emul-

sion problems, and 100 percent of the plants using bituminous coal in their

generators reported problems with emulsions. All of the plants using oil with

greater than 1.5 percent carbon reported emulsion problem, while only 80 per-

cent of the plants using oil with less than 1.5 percent carbon reported emul-

sion problems. Of the 78 percent of the total plants reporting emulsion prob-

lems, only 32 percent had overcome the problems, while the other 68 percent

still had emulsion problems. The most common solution to the problems was to

raise the superheater temperature or change the grade of carburetion oil used.

The scope of the emulsion problems faced by individual plants can best be

understood by examining the amount of emulsion produced by the plants.
According to Morgan and Stolzenbach (1934):

Carbureted blue gas plants using heavy oils produce on the average
two to four gallons of emulsions per thousand cubic feet of gas.
These emulsions contain about 60 percent of water. A medium sized
plant producing 10 million cubic feet of carbureted water gas per
day will produce from 20,000 to 40,000 gallons of tar emulsion. The
usual ways of disposing of this tar are as fuel under boilers in the
plant, and by sale to tar refiners. In either case, the tar emul-
sions must be dehydrated to a greater or lesser extent before dis-
posal can be made of it. On account of the low value of the tar for
either purpose, the dehydration process must be one that can be
operated at low cost...Attempts to dehydrate the emulsions by the
methods which have been developed in connection [with] ordinary
water gas tar emulsions or oil field emulsions have not been suc-
cessful in the case of most types of heavy oil tar emulsions. The
heavy oil tar emulsions are better stabilized, and appear to be
quite different from the types of emulsions which have previously
been studied.

A plant producing emulsions would quickly find all of its liquid storage tanks

filled, with nowhere else to store the emulsion. When all of the tanks at the
plant had been filled, the plant operators were faced with either dumping the
emulsions into pits or lagoons at the plant site or stopping gas production

while they dealt with the emulsion problem. Very few plants would have shut

down. Some of the heavier tars from the washboxes separate from the tar-water

emulsion, reducing the higher-boiling organic content of the emulsion.
138
-------
The tars contained in the emulsions would have essentially the same com-
position of the tars described in Section 1.4.2.1, except that some of the
heaviest tar components would separate and be removed. Eventually, the
disposed emulsions would separate into tar, water, and oil fractions. The tar
fraction would sink in water, and the oil fraction would float on the surface
of water.
Several methods were commonly used in the separation of tar-oil-water
emulsions. The method that always worked for the separation of tar from water
is the steam still. Water is simply distilled from the tar, leaving behind a
dehydrated tar product. This method had two major drawbacks. First, it was
relatively expensive, in that about 1.1 pounds of steam was required for each
pound of water evaporated. Second, the temperatures involved caused substan-
tial crosslinking of tar constituents, degrading the chemical value of the
tars. Centrifuges were frequently used to separate tar from water. The spin-
ning centrifuge basket separated the tar and water by density differences.
The operation of the centrifuge was relatively expensive because it required
frequent cleaning. The Warner tar dehydrator was essentially a steam still
that heated the emulsion to 240 °F to cause a separation of the tar and water.
The R.S. Dehydrator treated the emulsion with heat, pressure, and chemical
reaction to separate the emulsions. A tank was filled with emulsion, and soda
ash was added to the tank. The emulsion was heated to a steam pressure of
65 psig, then a valve to the tank was opened and part of the water in the
emulsion flashed to steam and was withdrawn. The tar layer then usually sepa-
rated and produced a tar with a water content of 10 to 12 percent water. In
actual practice, the plants would try one or two methods of separating the tar
from the emulsion, but they would probably dispose of t'"1-: '.^c if their normal
methods of tar handling were ineffective.
1.4.2.4 Waste Sludges —
1.4.2.4.1 Hater purification s!udges--0ne method used to purify waste
condensate at many plants was to treat the aqueous waste stream with lime (or
soda ash) and copperas (ferric sulfate) prior to discharging of the water.
This process added 5 pounds of lime and^4 pounds of ferrous sulfate to the
effluent water. The solids in the waste coagulated as small particles and
settled as a sludge with about 10 percent solids and 90 percent oil and water.
139
-------
This process produced about 1 ton of sludge per day when 72,000 gallons of

water were treated per day at the Brooklyn Union Gas Company. Approximately

one-third of the operating costs was for the removal of the sludge from the

plant site (Murphy, 1928). This sludge would apparently be very similar to

currently produced API separator sludge. The water purification sludge could

be mixed with coke and burned in the plant's boilers, or it could be disposed

at or near the gas production site.

1.4.2.4.2 Acid sludge from light oil agitators—Light oil from either

the distillation of tar or scrubbing the gas was frequently treated with sul-

furic acid to remove basic compounds and to improve the quality of the light

oil. The recovery of light oil and its treatment by sulfuric acid is dis-
cussed in Section 1.3.9. Consequently, this section will deal principally

with the characteristics and disposal of the sludge.

According to Powell (1929):

In plants that recover and purify light oil, the acid sludge result-
ing from the sulfuric acid treatment constitutes a waste disposal
problem. Willien (1920) has described the usual method of disposing
of the material. It is placed in an acid resisting vessel which is
heated with direct steam. The light oil given off during this heat-
ing may or may not be recovered by a condenser. The heating causes
the solid matter to separate at the top as a spongy, carbonaceous
material. The amount of this solid material produced in a medium
sized plant is not large, and it may usually be discarded on the
dump or burned out in the yard. Because of its sulfur content it is
better not to burn it under boilers.

The acid layer forms under the solid matter and is withdrawn. The recov-

ered acid can be slowly fed to the saturators for the recovery of ammonia as

ammonium sulfate; however, because the recovered acid is almost black, it
should be added slowly. Glowacki (1945) describes the waste and its treatment
as follows:

The acid sludge drained from the agitator during the washing process
is an intimate mixture of unused oil, entrained light oil, and reac-
tion products: "resins." In modern practice, the material is taken
to some convenient spot and burned. In the past, fairly elaborate
acid reclaiming plants have been devised and built; in general the
value of the reclaimed material failed to justify the labor, mainte-
nance, and investment costs of the reclamation equipment. A few
plants can still be found at American Installations.
140
-------
Either the waste acid sludge, or the carbonaceous material from the
reclamation of the acid would be wastes for disposal. The burning of the
waste would consist of digging a small trench in the ground at the disposal
area, filling the trench with the waste, and then burning the waste in the
trench. Only a portion of the waste would actually burn, and the residue
would remain in the waste disposal area. Whether the acid was worth recover-
ing from the waste depended primarily on the cost of sulfuric acid. Because
the sulfuric acid recovered was a very low grade, its recovery would have been
practiced primarily at larger plants.
The waste itself would be very acidic, and the base nitrogen compounds in
the light oils from coal-carbonization plants would be extracted into the
waste and generally disposed with the waste.
1.4.2.4.3 Tar decanter settlings and saturator sludge--Two types of
solid or semi sol id, black, and pitchy sludges were produced in the tar decan-
ters and the saturators (used for ammonium sulfate manufacture). The tar-
decanter settlings are the solid materials that come from the tar and and
flushing liquors. They consist primarily of coal and solid matter carried
with the gas into the washboxes. The saturator sludge is a hard pitchy mate-
rial that forms in the saturators used for the production of ammonium sulfate.
"Its exact nature is not known, but it is supposed to be formed by the action
of sulfuric acid on the small quantities of tar that are carried by the gas
into the saturator" (Powell. 1929). This sludge was probably produced by the
acid-catalyzed polymerization of unsaturated hydrocarbons In the saturator.
"Fortunately, the quantities of these solid, pitchy wastes are not large. It
is usual practice to discard them on the dump or in an excavation" (Powell,
1929). The tar-decanter sludge would be produced by coal-gas, CWG, and oil-
gas plants. The saturator sludge would only be produced by coal-gas plants
using the semidirect process for the production of ammonium sulfate.
1.4.2.5 Ammonia Recovery Wastes--
Ammonia recovery was practiced only at coal-carbonization plants; when
ammonia stills were used 10 release fixed ammonia salts, ammonia still waste
was produced. The recovery of ammonia is described in Section 1.3.5, and the
removal of phenolic compounds from the ammonia still wastes is examined In
Section 1.3.6. Table 37 shows the composition of the original ammonia liquor

141
-------
TABLE 37. ANALYSES OF AMMON1ACAL LIQUORS
AND THE STILL WASTES THEREFROM
ANAUYRKS OH AMMONIACAL Liq irons AND TtiK STILL WA.HTKS TIIKRKKROM
Ty|x- <>f oarUmiziiii! "jx'nilion Coke Ovctw "
IMaiit A 11
\Vpnk lii|in>r ciini|iu."iliiiM, gram.* |>or liter
N'H.1. total 6.54 7.1X5
free 3.35
fixed 3.19
Siilfide it HjS 0.138
<'.trlniimt« ON COj 0.81
(/'yanidc on HCN
Thicwuirntc iw HiS^j
ThiocynnaUM :w flCNS
I'lionols ILS CiHtOH
Oxygen nliwirption (4-hr tdt), ppm

Itcciivcry n( fixed ammonia

>ViiKto oiim|M)«ilii)ii, RraniH [XT liter
NIIi. tuLal O.CM1 0.0034
free 0.0034
fixed
AlkHlini'.y a* CuO 1 ..r>7 1.44
Sulfidc iw H-.S - 0.075
Carbonate ivs COj 0.37
Thinsulfatc ax HiSjOi
ThiiieyHimte as HCNS
1'hpii.iln iw C«H»OH
Oxygen ul)sorj)ti(>n (4-hr test), ppm
Vertical Itotorls
A " H "
14.5
9.0
5.5
1.9

0.12
0.40
1.22
3.55

Limed
3.6
0.21
0.54
1.8
5.800
13.3
10.0
3.3
1.5

0.05
1.78
2.03
3.5
16.500

Unlimcd
3.1
0.1
3.0

Trace

1.60
1.83
1.7
10,080
Source: Wilson and Wells, 1945.
142
-------
and the compositions of the still wastes from two coke-oven plants and two
vertical retort plants. The ammonia still waste was first treated for the
removal of phenols (by extraction, if the Koppers vapor recirculation process
was not used; see Section 1.3.6). "The waste is generally discharged Into
baffled sumps. Here solid matter settles out and the liquid cools. Accumula-
tions of sediment are removed from the sumps by bypassing them periodically
for cleaning" (Wilson and Wells, 1945). "The quantity of lime settlings is
not large with good operating conditions, and the material is usually disposed
of on the dump" (Powell, 1929).
The wastewater from the separation tanks could be either recycled as
scrubber water or discharged into the nearest stream. Its discharge generally
created only minor problems if the phenols were removed to adequate levels.
Wastewater containing phenols could be run directlv to the city sewers or used
in coke-quenching operations (as described in Section 1.3.5).
The amount of ammonia still wastes that were produced varied with the
ammonia recovery process employed by the plant. Coal-carbonization plants
using the direct process of ammonia recovery produced between 20 gallons
(Marquard, 1928) and 30 gallons (Powell, 1929) of waste per ton of coal car-
bonized. The indirect ammonia recovery plants produced about 90 gallons of
waste per ton of coal carbonized (Marquard, 1928; Powell, 1929). The semi-
direct process for the production of ammonia would produce some intermediate
volume of ammonia still wastes.
1.4.2.6 Hydrogen Sulfide Removal Wastes--
1.4.2.6.1 Spent lime—The-disposal of spent lime was a substantial
problem for the early gas plants. The spent lime contained a relatively high
concentration of CaS, which upon exposure to the atmosphere slowly reacted
with water and carbon dioxide to form CaC03 and H2S. The spent lime also
contained substantial amounts of tar from the gas, and the tar was also very
odorous. "The residue from dry lime purification is under certain conditions
readily disposed of, being valuable in many cases for agricultural purposes."
(Hughes and Richards, 1885). The spent lime for gas plants would have been
either disposed near the plant or sold as an agricultural lime. The spent
lime, once dispersed, would release hydrogen sulfide and then perform as nor-
mal lime in the soil. The tars would be sufficiently dilute to biodegrade,

143
-------
and any other constituents in the spent lime would be diluted below noticeable
levels. The amount of slaked lime required to purify coal gas was about 1
bushel for every 5,000 to 9,000 ft3 gas (i.e., 4,020 to 7,230 ft3 gas/ft3
slaked lime) (AGLA, 1875).
1.4.2.6.2 Spent oxide—The spent oxide from removal of hydrogen sulfide
from town gases is a waste generally found at any previous gas site. The use
of iron oxide quickly replaced lime for ^S removal, and it was the dominant
method of hydrogen sulfide removal until the demise of the industry. The use
of liquid purification was employed at some of the larger plants after about
1925, but iron oxide was still used at smaller works. The use of iron oxide
purification, the types of oxide used, methods of regeneration, and the
fillers mixed with the iron oxides are discussed in Section 1.3.7.3.
The composition of spent iron oxide varied substantially among town gas
production plants. According to Auebach (1897):
The gas purifying material...varies in the most extraordinary way.
from one works to another; the water varies from 2 to 40, the sul-
phur from 10 to 55, the sulphocyanides from 0 to 16, the ammonf-a
from 0 to 8, and the Prussian blue from 0 to 15 per cent; the colors
vary from yellow to black with all shades of blue, some are dry
powders, some are wet masses, and some are half sawdust and chips;
and the value varies accordingly.
Water-gas processes produce only small amounts of ammonia and cyanides, so the
spent oxide from water-gas production contains only small amounts of sulpho-
cyanides, ammonia, and Prussian blue. The spent oxides from coal-carboniza-
tion plants would contain substantial amounts of both sulfocyanides and ferri-
ferrocyanide's.
Spent oxide wastes were universally disposed in the United States,
although sulfur and Prussian blue were recovered from spent oxides in the
United Kingdom. Spent oxides were usually disposed by using the material as
fill, either around the plant, at the local dump, or on private property.
Downing (1932) stated that:
The disposal cf spent oxide is a vexatious problem for many gas
plants. Because of a possibility of fires starting through heat
generated by revivification, it is necessary to hold the spent mate-
rial at the plant until this danger is past. As soon as city
authorities learn of this menace the material Is prohibited at pub-
lic dumps. Continuous storage on gas works land eventually becomes

144
-------
impossible. The material makes excellent filling for roads or pri-
vate property when properly handled. It should be covered with
ashes or dirt immediately to prevent the access of air and conse-
quent combustion.

Consequently, spent iron oxide wastes are a major waste material remaining on

and around the manufacturing sites of manufactured gas. Morgan (1926)

described the utilization of spent oxides in the United States:

In England and on-the European Continent, considerable work has been
done on the utilization of spent oxide. When cyanogen is not
removed from the gas previous to the purifiers, the spent oxide
contains considerable ferrocyanide which was formerly recovered in
Europe, but which it does not pay to recover in this country. In
Europe, also, large quantities of spent oxide are used for the manu-
facture of sulfuric acid. In one sulfuric acid plant it is claimed
that the burnt oxide regenerated by a special process produces a
purifying material of good mechanical condition and special activ-
ity. At present, however, in the United States there is a plentiful
supply of cheap brimstone for the manufacture of sulfuric acid, and
the spent oxide has no market value.

The spent iron oxide wastes contain tar, some volatile organics, iron
oxide, Fe2S3, FeS, FeSz, sulfur, fluff materials (usually woodchips), ferric

ferrocyanide (Prussian blue) Fe4[Fe(CN)6J3, and thiocyanates. The cyanide

compounds would be almost absent from oxides from CWG, but they would be in

oxides from coal gas or mixed coal/water-gas operations. Spent oxide wastes

degrade somewhat after disposal. The FeS oxidizes to form sulfuric acid,

which helps to rust and dissolve the remaining iron oxides in the waste.

Depending on the amount of tar in the waste, the woodchips may or may not be

broken down by the acid. The highly acidic conditions do not appear to decom-

pose the ferric-ferrocyanide compounds. When the iron oxides dissolve away,

the ferric-ferrocyanide compounds become small unattached particles that can

migrate short distances from the waste to stain wood, rocks, soil, and other
materials not originally in the spent oxide. This bright blue color is char-
acteristic of cyanide-containing wastes from coal-gas processes, but it is

usually absent from water-gas spent oxide waste. A recent analysis of dis-

posed spent oxide wastes in Massachusetts reveals that:

pH 1.7-3.8

Cyanide, total 7,500. ppm

Cyanide, water soluble 0.7 ppm

145
-------
TABLE 38. AN ANALYSIS OF
SPENT OXIOE
AN ANALYSIS OK SPENT OXIDE 4IT
Free sulfur
Moisture
Ferric monohydrate
Ferrous monohydrate
Bo.«ic ferric sulfatc
Ferric ammonium ferrocynnidc
Ferrosoferric ammonium ferrocyanide
Ferric pyridic ferrocyanide
Organic matter peat fiber
Tar
Silica
Naphthalene
Pyridine sulfate
Ammonium sulfate
Calcium sulfate
Ferrous sulfatc
Ammonium thiocyanate
Sulfur otherwise combined
Organic matter soluble in alkalies
(humus)
Combined water and loss (by difference)
Percent
44.70
17.88
5.26
6.2.5
.2/5
.80
.50
.20
.68
.21
1.05
0.72
0.77
.06
.12
.02
.30
1,
3.
2.
1.
4.
I.
2.
0.
0.
1.
1.33

1.54
2.36
••^•m^^w
100.00
Source: Hill, 1945.
147
-------
9
4

9
-o •
c a
u*-
c*> a
x •
a u

u
•o

*> HI

L CM
3 I C
Uli
W
O
O
u
X
o
0)
m

ai
s
ta
eo -" i
CM r» m m • Q
u> « r- s CM
O
oa
r» O> • f»
rt CM f> Q ^*
to m
• *»
•o *-
tl *> K
H- «-* e
•i <-N»V. o
• n M c —
. *a •- • o +>
4) • CT— U
*• x • *» 3

.a-o *> u o

K L CO.
—'TOM •
S U as

x — U'H **.
O K • -D TJ •
O "D • I.
*> — — CM* • O
C *> **• X Q> Ol
• CO— • CO)
O. • i. *> L. — O

n c o.— > • o '^
• C • • .0
o e a n — & —

>< ct • o *» • c
" u c ct *> 3
-3 — — •••H4>
«•- m t. 9V 3 •

$•; i o > S*:*2
Q W K I- <
-------

^B
^
g
ot
o.

^^
3

I
ox
IX

C/)
Ul
H

Ul
o

X
o
I™
s
(L
M
l^
/•^

I
M
H

ff
^j
2>
§
i
§
Ul

•
s

111
_J
ffi
^M

«""
>*.
^x

L.
• f rt
•o a M
X T3 O>
0 •
*x*>
C I. s.
• e
e. cto
«/» • a

n
e
o

^rf*

c
O
••
41

U M
*> «
C M x->
• UK
U W^
o e
o —
to
X

O) >i S
m r- M
l 1 I
u> o> u>
••4 • T+
m

r- N o>
r- N n
Q 8 <9
l 1 l
fH r* r*
c*) Q) C1)
• n •
Q • Q
S

VO CM O5
r- M m

19 5> S
1 1 l
a u> a
m i* m
a
a • a

M
• o>
a

i. •-
' *
• *>
> »
^ o
• u
H 4*
• • U
a u —
S «fc
_ JJ ._
3 i S
u u a.
£

•
n
£
A
•••

a
•c
«•>' II

•V. *>"
•A ••
•H A
C
a •
« -o

n jt
• *3
O. J3

JC £
U *>

"o '»
o
J. *J
*» 0)

*
« •
wo >>
0).- ^)
t-l X
o <-
• 3
c c «-
• 0 —
a u 3
<-— n
Jts §
lx *
*> c"S •
!•;
o
U — l.
9 4>

^3 rf ^

r4Mn
(fl ^'^x>-'
,
* ••

M* ft
g
€ 3
O M
u »
bu *c
• ^
149
-------
approximate weight and volume of oxides produced from the three major gas
production processes. The assumptions and locations of the data used to gen-
erate this information are on the table. This table Indicates that the
production of CWG generally produced less iron oxide waste than did either oil
gas or coal-carbonization gas. The estimates in Table 40 are useful for rough
estimates of the amount of spent iron oxide-generated gas production.
1.4.2.6.3 Liquid scrubbing wastes—The solutions used for the liquid
scrubbing of hydrogen sulfide from town gases could not be used indefinitely.
The solutions generally became deactivated by side reactions that produced
inert salts. The products of these side reactions had to be removed and
either recovered or discarded. The four significant liquid purification proc-
esses (Seaboard, Nickel, Thylox, and Ferrox processes) are described in Sec-
tion 1.3.7.4.
The Seaboard process uses a solution of 3 to 3.5 percent NagCC^ to absorb
hydrogen sulfide from manufactured gas. The solution is regenerated by blow-
ing air through the H2$ rich solution, releasing the hydrogen sulfide to the
air. The use of air to strip the H2S from the gas also oxidizes some of the
absorbed H£S to sodium thiosulfate (Na2$205). This occurs with about 5 per-
cent of the H2$, which is absorbed by the process. The cyanide in the gas is
also absorbed and oxidized to form sodium thiocyanate (NaSCN). This process
required between 20 and 60 pounds of sodium carbonate for every 106 ft^ of gas
purifif^. The sodium thiocyantes v.ere sometimes recovered from the scrubbing
liquid, -jut they were Dually disposed rather than recovered. The thiocyan-
ates would be formed when the process was applied to coal-gas production, but
they would have been formed only in small amounts when the process was applied
to oil gas or CWG (because of the small amount of cyanide in these pr>r--.s).
The Thvlox, Ferrox, and Nickel processes each used solutions of soduim
carbonate for the removal of hydrogen sulfide. The metals added to the solu-
*ions (arsenic, iron, and nickel) served as catalysts in the regeneration of
the solutions. In the presence of the catalysis, the ^S is oxidized to f -e
sulfur and water. Cyanides were removed and oxidized to sodium thiocyanate by
both the Ferrox and Thylox processes. In the Nickel process, the cyanide
reacted with the nickel catalyst, deactivating it. This process was not used
-------
r
for the removal of ^S from coal-carbonization gases because of the cyanide
present in the gas. It was used with either oil gas or CWG.
The Ferrox process used a solution of 3.0 percent sodium carbonate and
0.5 percent ferric hydroxide. The sulfur produced by the process entrapped
both the ferric hydroxide and sodium carbonate in the product sulfur [the
product sulfur had 20 to 40 percent (dry basis) total impurities attributable
to these compounds]. This reduced the marketablity of the product sulfur and
also required relatively large amounts of makeup sodium carbonate (about 350
lb;106 ft3 gas treated) and ferric hydroxide (about 280 lb/106 ft3 gas
treated) (Kohl and Riesenfeld, 1985).
The Thylo* process used a solution of sodium carbonate and arsenic triox-
ide to absorb hydrogen sulfide from the gas and recover it as sulfur crystals.
The sulfur produced was of a high grade and was usually marketed for agricul-
tural purposes. Sodium carbonate consumption from the process was 60 to 120
lb/106 ft3 treated, and arsenic trioxide consumption was 15 to 27 lb/106 ft3
gas treated (Gas Engineers Handbook, 1934). The process required that a small
portion of the recycling solution be continously withdrawn from the system to
prevent the accumulation of sodium thiosulfate and thicyanate salts. Ihe
arsenic in this purge stream could be recovered by acidifying the solution and
recovering the arsenic as arsenic sulfide crystals. The recovered arsenic
could then be returned to the scrubber with additional sodium carbonate solu-
tion. Because the recovery and recycling of the arsenic was an economic deci-
sion, some plants may have disposed of the purge stream rather than attempt
recovery of the arseii .. "If feasible, the solution removed can merely be
discarded, or, if necessary, it can be acidified and filtered to remove its
arsenic as arsenic sulfide before being discarded" (Gollmar, 1945). Some of
the arsenic also remained in the recovered sulfur product, but at levels too
low to cause problems when the sulfur was used for agricultural purposes.
The Nickel process used a colloidal solution of nickel sulfide and sodium
carbonate to scrub hydrogen sulfide from gas and recover the sulfur. Like the
Thylox and Seaboard processes, sodium thiosulfate and sodium thiocyanide accu-
mulate in the solution. The consumption of the nickel sulfide was 23 lb/106
ft3 gas oil or.CWG treated, and sodium carbonate consumption was 51 to 120
lb/106 ft3 gas treated -Cundall; 1927).
-------
The disposal of waste scrubber solutions was generally performed by dis-
charging the liquid wherever it was practical. No references were found as to
the disposal practices in articles that reviewed the operations of the proc-
esses, but a survey of gas manufacturers did report disposal practices for
scrubber liquids. This survey (in 1930) sent questionnaires to 100 large gas
companies with production of greater than 500 x 10^ ftVyear. Of the 57 com-
panies that responded, 12 used some type of liquid purification, 5 Indicated
that they discharged their waste liqours to ponds, sand flats, or cinder fills
and the remaining 7 said they discharged to either city sewers or to river
tidewater (Wardale, 1930). This survey indicates that some plants using
liquid purification could have substantial contamination from arsenic or
nickel if they disposed these scrubber solutions onslte.
1.4.2.7 Lampblack Wastes—
The production of gas by the Pacific Coast oil-gas process was accom-
panied by the generation of large amounts of lampblack. The feedstocks for
the production of oil gas were asphaltlc-based oils and had high carbon-to-
hydrogen ratios. When these oils were thermally cracked for the production of
oil gas, much of the original carbon in the oils formed elemental carbon.
This carbon (lampblack) usually washed out in the washboxes, where the heav-
iest tars also condensed. The material recovered 1n the washbox was a sludge
with large amounts of free carbon, some heavy tars, and water from the wash-
box.
Morgan (1926) states that:
From 12 to 24 pounds of lampblack are formed per 1000 cubic feet of
gas made, and practically all of this is thrown out 1n the wash-box.
The water from the wash-box containing this lampblack in suspension
passes off through large overflow pipes. In smaller plants this
water suspension of lampblack flows into small settling pits, from
which after settling the clear water 1s drawn off. The lampblack 1s
then mixed with tar and used for boiler fuel. !n larger plants the
lampblack in the overflow may be separated from the water by an
Oliver continuous rotary filter. It may then be brlqueted with a
small amount of tar and sold as a superior boiler fuel.
The briqueted lampblack was sometimes used as generator fuel for the produc-
tion of CWG. Although the lampblack had value as a fuel, many small plants
would dispose of the lampblack rather than recover it, and large plants might
152
-------
produce so much lampblack that they disposed of the material they could not
use. The organic tars removed from the washbox with -w lampblack would be
relatively heavy tars, with the composition dependent on ctie temperature of
the washbox.
1.4.2.8 Ash, Clinker, and Coke--
Ash, slag, and coke were wastes produced in the production of town gases.
The ash was produced from boilers, CWG generators, and producer gas genera-
tors. The coke or coal placed into producer gas generators or CWG generators
could not be combusted completely to ash. The requirement that air and steam
be able to flow through the coke beds meant that the ash had to be removed
with a substantial amount of unburned coke remaining in the ash. The ash was
then usually run through a coarse grate to remove any large pieces of coke
(which were recycled to the generators), and the material falling through the
grate was discarded. This both recovered usable coke and decreased the carbon
content of the ash, making the ash irore suitable as a fill material. The ash
produced by CWG and producer gas had substantial amounts of unburned coke.
Ash from the boilers, however, was combusted much more completely. Within the
generator bed, some large agglomerations of ash would form. These were called
"clinkers" and were removed from the generators at regular intervals. "Water
gas generator clinker, and boiler house and producer gas ash are normally
disposed of b> using (them] for fill and grading purposes" (Powell, 1929).
Although ash was apparently used in Europe for the manufacture of brick or
cement, this was not done in rhe United States because of the relative cheap-
ness of other raw materials.
I-he amount of ash produced by gas manufacturers was directly proportional
to the ash content of the coals and coke used for gas production. The ash
produced ot oil-gas plants ivDuld be a petroleum ash and would have a different
composition than the codl ash.
i.4.2.9 Firebrick awl Building Materials--
The apparatus fur the production of oil gas and CWG was lined with fire-
bricks that were alternately heated and cooled during the manufacture of gas.
Coal-carbonization apparatus used firebricks for linings and heat exchangers.
The apparatus periodical)., had to be relined with new firebricks because of
the wear associated with gas manufacture. Broken firebricks were used as fill

153
-------
material wherever needed around the plant or they were added to the dump.
Buildings were sometimes removed during plant operations, and the final clear-
ing of the site occurred after manufactured gas was replaced by natural gas.
These building materials were also used to fill areas on or near the site.
1.4.3 Specific Articles on Waste Disposal
During the literature review, several articles that specifically
addressed the waste disposal practices of the industry were discovered. These
articles, which take a fairly wide view of waste disposal practices, are
described in this section.
Shelter (1897) reviews "The Nuisance Question in Gas Works." He
describes the sources of odor, noise, smoke, and offensive drainage from gas-
works. Methods of reducing the problems created by operating a gas plant are
described, as are methods of improving the plant appearance. Shelton states
that:
Offensive refuse drainage may come from: 1) unintercepted scrubber
water or condenser water saturated or laden with ammonia, tar or
oily scum; 2) tar or oil wasted; 3) the rain washings of spent lime
or old oxide; 4) general gas works and surface drainage; 5) drip
water not properly disposed of.
Hansen (1916) describes the "Disposal of Gas House Wastes" in which he
describes the objectionable effects of gas house waste disposal and describes
methods for preventing these effects. Because Hansen's work was presented to
a group of gas plant operators, it was not especially well-received according
to the reviewer comments. Hansen states that:
Wastes vary greatly in quantity and character due to variable recov-
ery of useful constituents and to the use of variable quantities of
water. Generally speaking, the quantity of wastes per million cubic
feet of gas manufactured is greater and more offensive in the
smaller plants than in the larger ones because of the smaller recov-
ery of marketable products and greater waste due to leaky tanks and
defective apparatus.
He lists several caser of stream and water pollution attributable to gas plant
wastes and how the disposed wastes give fish gassy odors and impart medicinal
tastes to water. At Centralia, Illinois, according to Hansen:
Much complaint was made of tarry wastes adhering to the legs of
cattle, and to injury of soil and crops by tarry deposits...Another

154
-------
bad effect of gas house wastes which has here and there given rise
to more or less serious trouble is the pollution of the soil, which
in turn gives rise to gassy tastes in well waters and to gassy odors
in cellars. A striking example of this occurred at Joliet, where
one of the public water supply wells^was affected with a gassy taste
which could be explained on no other'basis than contamination from a
gas plant near by...At the town of Carthage, in southern Ohio...pol-
lution was occasioned by coal tar wastes used at a tar paper fac-
tory. These wastes were permitted to flow into a pit at least 2,000
feet from the affected wells.
t
An estimate that 1,600,000 gallons of tar and oily wastes exist underground at
Lowell, Massachusetts, is presented. When some contaminated areas were tapped
by excavations, the wastes "flowed out in springs." Methods of removing oils
and tars from aqueous wastes by coagulation with ferrous sulfate and lime are
described, as is the use of sand and coke filters.
Brown (1919) describes how the chlorination of water containing trace
amounts of gas plant waste produces objectionable tastes in the water. The
levels of organic material themselves did not produce objectionable tastes,
but the tastes became noticeable after chlorination.
The American Gas Association (AGA) had a standing committee on waste
disposal from gas plants during the 1920's. Their articles (as reported in
tl •? annual proceedings of the AGA) detail the wastes produced and the normal
methods, of waste disposal.
Willien (1920) documented the injurious effects attributed to the waste
from gas plants and described the types of waste produced from coal-gas and
CWG plants. The effects of gas plant wastes included driving away fish and
contaminating oyster beds, damaging paint on pleasure boats, objectio- ,-ible
odors, pollution of wells, deposits in sewer systems, and pollution of drink-
ing water. According to Willien, "Pollution of wells...is caused by the seep-
age of gas plant waste through the ground and contaminating the ground water.
This may result from a crack in a tar well or holder pit through which tar
leaks, or from leak\ tar, oil, and ammonia pipes."
Sperr (1921) describes methods of tar separation that can be applied to
aqueous gas plant wastes. Typical systems for the gravity separation of tar
from water are described, as is the use of centrifuges for the dehydration of
tar emulsions.
Willien (1923) describes the formation, treatment, and storage of tar
emulsions and tars.
155
-------
Powell (1929) classifies and describes the wastes produced by gas manu-
facturing. Dividing the wastes into two classes (solid and liquid wastes), he
describes the wastes and the usual methods of disposal. "It must be real-
ized," states Powell, "that gas plant wastes are really by-products whose
value is too low to make direct sale feasible." Table 41 1'sts the wastes as
1 is ted by Powel1.
The only survey of waste disposal methods was published by Wardale in
1930. A survey was sent to 100 gas companies in the United States with gas
production greater than 500 x 10^ ft^/year. Answers were received from 57
companies. 10 of which were no longer producing gas (they had converted to
natural gas). Table 42 summarizes the questions and answers most related to
waste disposal. Although this survey was not comprehensive of the entire
industry (smaller gas plants were not even contacted), it is the only reported
survey of gas plant disposal practices.
One possible method of waste disposal that was originally thought ic be
commonly used by plants was the disposal of waste by injecting it into wells.
Only two references to the use of wells for the disposal of wastes were uncov-
ered during this investigation. The first is an article listed in a biblio-
graphy on plant waste disposal. The bibliography was published in the 1955
AGA proceedings, and the referenced article was titled "Underground Disposal
of Process Waste Water," by l.K. Cecil (1950). A summary of Cecil's article
states: "Underground disposal of brines and chemical waste.; water. Acidizing
the injection well semi-annually maintains disposal capacity. Cooling tower
blowdown containing chromates is similarly handled." The second reference is
by Wilson and Wells (1945), who state that:
Disposal of ammoniacal liquors or waste by discharge into the ground
is seldom possible except in very small carbonizing operations.
Discharge into an opening, such as a disused well, is dangerous,
because the ultimate fate of the liquor is unknown. It may be grad-
ually dissipated and purified as it seeps through the soil. On the
other hand, it may find its way into some water bearing strata or
percolate unchanged through the layers of soil to drain into a
stream. In such a case the pollution would not appear immediately,
but when it did, deposits of the material in the contaminated soil
would cai'se the trouble to persist over a long period of time.
156
-------
TABLE 41. GAS PLANT WASTES
1. Solid wastes

1. Ash and clinker
2. Spent oxide
3. Tar decanter settlings and saturator sludge
4. Lime settlings

II. Liquid wastes

1. Phenol-bearing wastes
a. Ammonia still waste
b. Other phenol-bearing wastes

2. Wastewater not containing phenol
a. Coke quenching water
b. Producer gas cooler water
c. Water-gas tar separator overflow

3. Acid sludge from light oil agitators

4. Tar emulsions

SOURCE: Powell, 1929.
157
-------
r
TABLE 42. RESPONSES TO WASTE DISPOSAL SURVEY
Do you produce ammonia? What disposal is made of still wastes?

22 - Ammonia plants
5 - Settling basins or coke filters
1 - Phenol removal equipment
13 - Discharge untreated weak liquor or still waste
1 - Sells weak ammonia liquor

What is done with spent oxide from purifiers?

24 - Use as fill onsite or given away as fill
1 - Sold for sulfuric acid manufacture
1 - Dumps it at sea
13 - Haul it to city dump
1 - Dumps it into river at flood level
1 - Gives it to stable for horse bedding
Several - Mention need to cover or mix spent oxide with dirt

Do you use liquid purification? How are waste liquors disposed of?

12 - Use liquid purification
5 - Discharge to ponds, sand flats, cinder fills
7 - Discharge to city sewers or river tidewater

What disposition is made of wastes containing oil?

i 3 - Pump into relief holder
^ 8 - Use baffle separators and coke filters
i 15 - Use separators or settling basins, remove oil by skimming, burning
; it in boiler, or mixing it with tar
; 4 - Run wastes to sewers or creeks without treatment
r

What other wastes do you dispose of besides waters from scrubbers, washboxes,
purifiers, and sanitary and surface water drains?

10 - Ammonia still waste or weak ammonia solution
1 - Shavings from tar scrubber, which are burned after dark
2 - Coke quench water
1 - Water-softening residue

What methods of treatment before discharge to sewers?

18 - Baffled separators
13 - Baffled separators and coke or cinder filters
1 - Oliver-Borden filters
1 - Ferrous scHate and soda ash treatment before coke filters
1 - Recirculates water to washboxes
6 - Discharge without treatment

SOURCE: Wardale, 1930.

158
-------
r
1.5 PRODUCTION AND HISTORICAL TRENDS OF THE U.S. TOWN GAS INDUSTRY
1.5.1 Introduction
This section examines the historical trends of the U.S. town gas indus-
try. By studying the production trends for various parts of the country, the
predominant methods of gas production can be identified, the amount of gas
produced can be examined, and the approximate time that the manufacture was
replaced by natural gas use can be determined. The gas production processes,
feedstocks, and innovations in the industry affected both the quantities and
disposal practices for wastes. By tracing the changes that occurred in the
industry, additional insight to the problems of current gas sites can be
acquired.
Most of the statistical data on gas production, employment, and feedstock
use were collected during the operation of the manufactured-gas industry by
AGA. RTI's effort to collect and examine this data is probably the first time
the industry has been examined since the late 1950's.
Section 1.5.2 reviews the historical production trends within the U.S.
and individual regions. Section 1.5.3 shows how the feedstocks for gas manu-
facture changed with time. Section 1.5.4 plots the historically significant
events of the industry.
1.5.2 U.S. Gas Production Trends
The production trends of the U.S. manufactured-gas Industry show the
amounts of yas produced, the types of gas manufactured, and when the manufac-
ture of gas stopped. The types of wastes from gas production varied with the
manufacturing processes (coal gas, CWG, and oil gas), and the amounts of waste
produced are approximately proportional to the the amount of gas manufactured.
The gas production within a region can be used to estimate (in a qualitative
manner) the waste types that would be found at former gas-manufacturing sites.
The gas production trends can be studied for either the entire country or
for separate regions. Examining the entire United States allows overall
trends to be studied, whereas regional trends are more relevant for applica-
tions to looil trends. Statistical data were compiled from the information
collected and reported annually by tie AGA. The original data were collected
on a State-by-State basis, with regional totals. The regions used by the AGA,
-------
and the States within each region, are listed in Table 43. Attempting to
compile and analyze the statistics on a State-by-State basis is feasible, but
it was not performed on this project because of the substantial effort re-
quired. Most of the earlier data on gas production were reported on an Mcf
basis (10& ft3), and data after 1945 were in millions of therms (1 therm =
100,000 Btu). Table 44 shows the gas heat values and conversion factors used
for each type of gas.
Figure 40 shows the total U.S. manufactured-gas sales between 1821 and
1956. This figure includes manufactured gas that was mixed with natural gas
and distributed as a mixed gas product. This figure indicates that U.S. gas
production was relatively small before 1900, increased rapidly to 400 billion
cubic feet (bcf) between 1900 and the beginning of the Great Depression
(1929), then fell about 25 percent during the Depression but recovered during
World War II. The production of gas peaked shortly after World War II, before
declining about 50 percent between 1947 and 1956. The apparent drop in gas
production in 1920 did not actually occur. The data prior to 1920 came from a
source (Fulweiler, 1921) different from the information between 1920 and 1956
(AGA, 1961).
Figure 41 shows how the manufactured gas was produced between 1919 and
1956. This figure does not include gas manufactured for mixing with natural
gas, and the production of retort gas was included with coke-oven gas prior to
1928. This plot shows several interesting trends. There was a steady rise in
purchases of coke-oven gas between 1920 and 1930, reflecting increased produc-
tion of metallurgical coke by byproduct coke ovens during the period. There
was a steady decline in retort gas production by gas companies during the
period, displaying a tendency of smaller coal-carbonization plants (using
retorts) to switch to other forms of manufactured gas as existing retorts wore
out. The large drop in oil-gas production in 1.928 occurred because much of
California switched to natural gas that year. The production of coke-oven
gas, oil-gas, and coke-oven gas purchases remained relatively constant between
1930 and 1950, and CWG production showed a substantial decline and increase
during the same period. This shows that CWG production was more sensitive to
gas demand than was coal-gas production. In relative amounts of gas produced,
this figure indicates that the production of CWG was approximately equal to
160
-------
TABLE 43. STATES LOCATED WITHIN EACH GAS PRODUCTION REGION
New England States
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont

Middle Atlantic States
New Jersey
New York
Pennsylvania

East North Central States
II linois
Indiana
Michigan
Ohio
Wi scons in

West North Central States
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota

South Atlantic States
Delaware
District of Columbia
Florida
Georgia
Maryland
North Carolina
South Carolina
V i rg i n i a
West Virginia
East South Central States
Alabama
Kentucky
Mississippi
Tennessee

West South Central States
Arkansas
Louisiana
Oklahoma
Texas

Mountain States
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming

Pacific Coast States
California
Oregon
Washington
1G1
-------
TABLE 44. GAS HEAT VALUES USED TO CONVERT BETWEEN FT3 AND THERMS3
Gas type Btu/ft3 106 therm/109 ft3
Coke-oven gas 540 5.4
Retort gas 520 5.2
Carbureted water gas 600 6.0
Oil gas 600 6.0
al therm = 100,000 Btu.
16?
-------
•ft
CO
in
01
CM
as
M
M

-------
I I i i I l I I I I I I I I t I I i
0(30000000
(0
<0
in
o>
T-
o
«•
c>
5
0)
a

3?
o
T3
2
a
M
O)
3
a
5
.2
3
en
6OL) uojionpojd

164
-------
the production and purchases of coal gas. The production of CWG was much more
suited to smaller plants, so the relative number of plants producing the two
gases was greatly different. In 1920, the AGA reported the following distri-
bution of gas-manufacturiiig plants (as compiled from Brown's Directory):
Coal-gas plants 189
CWG plants 429
Combined gas plants
Coal and CWG 150
CWG and oil 3
Coke oven and CWG 9
Coal, CWG, and coke oven 5
Coal and oil 3
Coal, CWG, and purchased 5
CWG and purchased 12
CWG and natural 4
Oil and natural 3
Reformed natural gas 4
Type not listed 1
Purchased, no mfg. 99
Total manufactured-gas plants 987
Byproduct coke ovens 82
This distribution shows that 43 percent of the U.S. gas plants in 1920
produced exclusively CWG and 62 percent of the plants produced at least some
CWG. The 82 byproduct coke ovens sold gas to companies for distribution.
Figure 41 also shows a decrease in all types of gas production, beginning
about 1950. The decrease in the coke-oven gas produced in 1928 is an artifact
of the way the data were collected. Retort coal gas was Included with the
produced coke-oven gas prior to 192S, but it was collected separately after
1923.
Ihe U.S. manufactured-gas production for each region is shown in Figures
42a and 42b. These figures do not include gas that was manufactured and mixed
with natural gas. Hence, whenever a company acquired natural gas, but still
produced gas for peak loads, its production was excluded from the data. The
if.r,
-------

1 1 1
r rO CJ «-
1 1
O O>
r-' O
1
03
d
i
r> •
d
i
CD
d
!
m
d
•$•
d
i
K>
d
i
(N
d
i
d
R
0>
o
d
at
in
O)
T«

d
PJ
o>

"5>

of
o
"53
M
M
•a
2
a

u
ra
(sujjeq) 6ot) peonpaid SBQ
166
-------
2 I
1
TSO
Z3II
< no +

O
O)
o
CO
O
in
i r

O Q
8
o>
en
in
en

en
m
01
T"

"5
(A

5?
O)
n

(A
CM
cn
E
o
CN
i r

o o
(suujein 9oi)
167
-------
data do include coke-oven gas purchased for resale by gas companies selling
manufactured gas. These figures show the relative amount of manufactured gas
produced by various regions in the United States. The Middle Atlantic region,
composed of New York, New Jersey, and Pennsylvania, produced about twice the
amount of gas of any other region during this period. The New England, South
Atlantic, and East North Central regions produced comparable amounts of gas
during this period, and each of the other regions produced gas oc levels
smaller than 1/10 the production of the Middle Atlantic states. There is a
large change in scale between Figures 42a and 42b, which allov/s the gas
production in the smaller gas production regions to be examined. The V/est
South Central region produced no gas during this period because of the avail-
ability of natural gas in the region. The figures indicate the introduction
of natural gas to the regions by the resulting drops in records of manu-
factured-gas sales. The start of the production declines for the regions is
listed below:
Region Yejir
West South Central ?
Mountain 1948
South Atlantic 1951
Middle Alantic 1951
East North Central 1952
•New England 1952
East South Central 1955
Pacific 1956
The employment trends of the gas industry tracked the production trends.
Figure 43 shows employment in the U.S. gas industry, divided by the type of
gas sold by companies. It shows a dip in all employment during the Great
Depression, with increases in employment during World War II and until 1950.
Between 1950 and 1955, employment in companies selling manufactured gas drop-
ped sharply, and employment in, companies selling mixed gas increased during
the period, prior to decreasing after 1955. This indicates that manufactured-
gas companies switched to distributing mixed gases after natural gas pipelines
were installed in their regions. The employment in'Companies producing or
distributing natural gas increased steadily after World War II.
168
-------
tn
.O co
a O)
.S 0>c

15 T3 -D m « C

= SS"-32

sis
OOOOOOOOOOO
OOOOOOOOOO
m
to
at

o>
a
M
n
o>
s
1
a
s,
ai
C
(spuBsnoqi) lueiuAoiduig
169
-------
Figures 44 through 52 show regional gas production by gas type for the
nine U.S. regions between 1928 and 1956. They are ordered by the total amount
of gas produced within each region, so that Figure 44 is for the Kiddle

Atlantic (with the largest gas production) and Figure 52 is for the West South
Central States (which had very little gas production). These figures include

gas that was manufactured and mixed with natural gas for distribution.

Specific features of the figures are described below:

• Figure 44: Middle Atlantic States--CWG was the major gas type
manufactured in this region. The rate of CWG production
doubled between 1935 and 1952, and production of other gas
types remained relatively constant during the period. Rela-
tively little retort coal gas and oil gas was produced, and the
production of coke-oven gas was equally divided between that
produced by gas companies and that purchased from coke com-
panies. Natural gas became available in the area after 1951,
resulting in the decline of coke-oven gas produced by gas com-
panies and CWG production. The gas companies continued to
purchase coke-oven gas during this period. A comparison of
this figure and Figure 42a shows that the CWG was mixed with
natural gas for distribution by companies in this region (gas
manufactured and mixed with natural gas is not shown in Fig-
ure 42a).

• Figure 45: New Fr.gland States--CWG and coke-oven gas were the
major production processes. CWG production increased during
and after V/orld War II, and coke-oven gas production and pur-
chases remained relatively constant. Natural gas was intro-
duced to the region in 1952, resulting in declines in all gas-
manufacturing production. Oil-gas production increased between
1945 and 1952 and fell to zero later. This indicates that gas
utilities in the region either converted CWG apparatus to oil
gas or installed oil-gas apparatus. High Btu oil gas had a
heating value close to natural gas and was used to supplement
natural gas for peak loads.

• Figure 46: South Atlantic States--CWG was the major gas pro-
duced in this region. Some coke-oven gas was purchased, and a
small amount of retort gas and oil gas was produced, but the
toul of the gas from these sources was less than half of the
CWG production. Gas production dropped steadily after 1945,
but some increase in oil-gas production is observed. The oil-
gas production would principally be from converted CWG appara-
tus and used for gas production during peak loads.

• Figure 47; East North Central States — Purchases of coke-oven
gas exceeded the other types of gas production between 1929 and
1948. The production of coal gas (both produced and purchased)
-------
6oi) uoipnpojd

171
-------
(O
Ift

O)
T»

«
en

•o

t
u

0
J
•M
O

•§

2
a

«
CB
O
I
(stujeqi 9oi) uoj
172
-------
o
« o o » =
OOOffO
a o < •••
O O O O O O
O CD to ^r CN o
CM — — — — —
(suuein 9oi) uoi.pnpojj

173
-------
•S £ J2 2 -
-------
to
in
a

-------
II

O
(suueiii 9oi) uoi^inpojd

176
-------
to
to
CD

5)
m
CJ
I

at
U
£

O
V)

Tn
a
LU
O
s

-------
CO
in
o>

-------
\i
t
f-
J.
i
•

m-
I

m-
8
O)

^
A
l|
S a
H
II
Ja i
a o
OO
a o

D
D

?
M
5
|
a
M
a
a
e
S
•^
O
<

S
at
g £•
Sg
11
o §
"Z =
CCO
•*• x

i I
O O
T n
i
i
E
E

E
t
(
I
t

c
•f
-i;

i M
1-
1
h
1

h
1
h-
1
h

1
*.
1

o o o
CN •-
§
a>

8
*-

10
in
o
C4

O
States
he West South Cen
_o

TJ
S
a
«

si
s

en
(suuegj 9oi)
179
-------
was about twice the CWG production in the region. Sharp drops
in CWG production occurred after 1949, and coke-oven gas
production and purchases dropped at slower rates. A steady
decline in retort gas production occurred between 1928 and
1950.

• Figure 48: West North Central States--CWG production was the
major gas produced by gas companies in this region, but CWG
production dropped sharply after 1930 and showed a small
increase after World War II. Coke-oven gas purchases dropped
after 1948, and coke-oven gas production dropped to zero in
1950. Only a small amount of oil gas was produced in the
region.

• Figure 49: Pacific Coast States--Gas production in States
bordering the Pacific Ocean was principally by the oil-gas
process during this period. This figure is somewhat misleading
in that by this period California was producing and dis-
tributing natural gas, and Oregon and Washington continued to
manufacture gas. Some CWG was produced and very little coal-
carbonization gas was produced in this region. The oil-gas
production shows a very rapid decline at the end of World
War II (1945). This is because the oil-gas plants were oper-
ated at relatively high levels during the war so that by-
products needed for the war effort could be produced. Gas was
still being produced at substantial levels through 1956.

* Figure 50: East South Central States—This is the only region
examined where coke-oven gas purchases were the major source of
manufactured gas. The purchases of coke-oven gas dwarfed the
gas production by gas distributors, although CWG was produced
for several years after World War II. CWG production declined
sharply in J950, and coke-oven gas was still purchased (prob-
ably for mixing with natural gas) through about 1955. No oil
gas or coke-oven gas was produced by gas companies during this
period.

• Figure 51; Mountain States—This region had very low levels of
gas production. Retort gas and CWG were produced in 1928 but
declined sharply after 1928. Oil gas and purchases of coke-
oven g<-s predominated between 1931 and 1948. Gas production
essentially stopped in 1949.

• Figure 52: West South Central States—There was no significant
gas production in this region after 1929. There would be some
gas production before this period, however.

Figures 53 and 54 show some early information on gas production in

Massachusetts (Grimwood, 1896). Figure 53 shows the amounts of coal gas, CWG,

and oil gas produced between 1886 and 1900. This figure clearly shows the

180
-------
: I!
in
s

03
o
o
01

to
co
oo
M
o
n
a

I
i
M
3
n
in
S

o
il
181
-------
zoo

D + O
o
o
0»
(O

CO
I

o
a
M
M
a
O

I
w

o
M
in
en
00
co
CM «-
(SUOHB6 goi) HO

182
-------
increase in CWG production after 1890 while coal gas production fell, due
principally to the replacement of coal-gas retorts by CWG apparatus. Simi-
larly, Figure 54 shows the use of enriching oils increasing with increased CWG
production. Naphtha was the major carburetion oil used during this period,
but gas oil and crude oil were also used. The changes in gas-oil use and
naphtha use between 1896 and 1899 are exactly opposite. When gas-oil use
increased, naphtha use decreased; likewise, when naphtha use increased, gas-
oil use decreased. This indicates that either of the two feedstocks could be
used, with the amounts of each purchased dependent on price and availability.
The regional gas production shown in this section shows clear patterns of
variation with respect to the production methods employed in the various U.S.
regions and in the relative amounts of gas produced within the regions.
1.5.3 U.S. Gas Feedstock Trends
Just as there were trends with respect to the types of gas produced,
there were also variations of the types and amounts of raw materials used in
the production of gas. Two major types of feedstocks were used in the produc-
tion of town gas—solid carbon-based fuel and liquid oils. Figure 55 shows
the use of solid fuel for gas manufacture between 1919 and 1965. Two types of
coal (anthracite and bituminous) and coke produced from bituminous coal were
used in the manufacture of gas. Anthracite coal was used as both generator
•fuel (for CWG and producer gas) and as boiler fuel. The use of anthracite
declined before 1930 because reduced supplies of anthracite increased costs of
the fuel. Coke was used primarily in the gas generators of CWG apparatus, and
some of the coke was used for producer gas and as boiler fuel. The rise in
coke use prior to 1930 is from the increased production of CWG. Coko was
produced from bituminous coal in either retorts or coke ovens. Figure 55 also
shows the characteristic drop in fuel use during the Great Depression and
increasing fuel purchases during World War 11. The decline in solid-fuel
purchases after 1950 parallels that of the gas-manufacturing trends.
Figure 56 shows the total oil used in gas manufacturing between 1919 and
1965. Oils were used primarily for the carburetion of CWG and for the produc-
tion of oil gas, but they were also used as boiler fuels by the gas producers.
Figure 57 shows the types of oils used between 1945 and 1952. The major trend,
shown in this figure is the substantially increased use of other heavy oils

183
-------
in

oi
i—

a
3

8»
a
en
•o
o
M

"3
tn

u»
in

I
8MOQ W |BOO SUOJL gOl
184
-------
j

1
1 o
!- «
1 0
1 *~
i
1
i
1
i
j
!
in
(O
o>
T-
o
o>
0>
»"
di
c
•••
b.
2
u
JB
r
i
i-S

i — i — i — i — i — r- ill'11
i i i
M
a
a
e
in
o>
OZ
cM^^oioor-cqin^'OeNj-o
,-• -' d d d o o o o o o
HO JO SUOliBQ 6Ql

185
-------
o
I!O jo
186
-------
between 1945 and 1950. Because the production of CWG also increased during
the same period, most of this increased production used other heavy oils
(which were principally the heavy residuum oils that remained after the
catalytic cracking of gas oils). The other use of other heavy oils increased
as the use of lighter Bunker "C" oils decreased during the period, indicating
that gas manufacturers switched from the C oils to heavier oils. Because
there were more tars and lampblack created and more emulsion problems
associated with the use of the residuum oils, this change in oil feedstocks
increased the amount of waste produced by the industry.
1.5.4 Historical Events_of the U.S. Gas Industry
Table 45 is a listing of the significant events in the manufactured-gas
industry. This listing includes many of the developments in gas production,
purification, markets, and feedstock usage that affected the types and charac-
ter of waste produced by the town gas industry.
1.6 DIFFERENCES BETWEEN THE U.S. AND BRITISH GAS INDUSTRIES
The redevelopment of gas production sites has occurred much more fre-
quently in Great Britain than it has in the United States. The Harwell report
on the problems arising from the redevelopment of gas sites (Wilson and
Stevens, 1981) was published several years before a somewhat similar work was
published in the United States (Handbook on Manufactured Gas Sites, Environ-
mental Research and Technology [ERT], 1984). There is a tendency to apply the
information from the British work on site redevelopment directly to U.S.
sites. This section outlines the major differences between the U.S. and
British gas industries, and it relates those differences to current waste
problems at U.S. sites.
In the United States, the availability of petroleum and petroleum distil-
lates encouraged their use for the production and enrichment of town gas.
British gas was primarily coal gas and coke-oven gas, reflecting the abundance
of coal in the United Kingdom and the absence of significant oil resources.
Because the tars produced from oil-gas and CWG production are generally less
viscous than coal tars, the problems of tar migration from the U.S. facilities
are probably greater than are the tar migration problems associated with the
U.K. coal-gas plants.

187
-------
TABLE 45. SIGNIFICANT EVENTS OF THE TOWN GAS INDUSTRY
Year
Event
Reference3
1806 A home and street lighted by manufactured gas in
Newport, RI
1809 Milk of lime used for H2S removal in Britain
1812 Company chartered to light London streets
1315 English patent for oil-gas production issued
1815 English patent for oil-gas process issued
1816 First U.S. coal-gas company incorporated
1816 Coal-gas plant installed in Baltimore, MD
1316 First public display of gas lighting in Baltimore, MD
1820 First coal-tar distillation plant started in England
1322 Coal-gas plant installed in Boston, MA
1325 Coal-gas plant installed in New York, NY
Tunis,
1933;
Morgan,
1926

Powel1,
1945o and
1945b

Rhodes,
1966a

Rhodes,
1966b

Hull and
Kohloff,
1952;
Rhodes,
1966b

Hull and
Kohloff,
1952

Rhodes,
1966a;
Morgan,
1926

Tunis,
1933

Rhodes,
1966a

Rhodes,
1966a;
Morgan,
1926

Rhodes,
1966a;
Morgan,
1926
(continued)
188
-------
Year
TABLE 45 (continued)
Event
Reference3
1829 Water-sealed gas holder used in England; masonry tanks Alrich,
were used to hold the water 1934
1838 First timber treated with coal tar in England
1838 Heavy oil (creosote) first used to preserve wood from
decay and marine worms
1847 First benzene recovered from coal tar in England
1849 Iron oxide process for H2S removal patented
Rhodes,
1966a

Stover
and
Chung,
1979

Rhodes,
1966a

Powell,
1945a and
1945b
Around
1850
Before
1850
1850

1856
Horizontal firebrick retorts were commonly used for
coal-gas production
Cast iron retorts used for coal-gas manufacture,
600-800 *C

Clay retorts used for coal-gas production instead of
cast iron

Dye from light-oil fraction of coal tar discovered;
analine dyes follow this discovery
1856 First coke ovens with byproduct recovery installed in
France

1857 Dye manufactured from coal-tar products in England
1860 British "Sulfur Act of 1860" limited sulfur in
gas to 22 grains per 100 cubic feet
Rhodes,
1966a
Rhodes,
1966a

Morgan,
1926

Stover
and Chung,
1979

Morgan,
1926

Rhodes,
1966a

Powell,
1945a and
1945b
1859-
1900 Air-cooled condensers used to cool manufactured gas
Downing,
1934
(continued)
189
-------
IABLL 4b (continued)
Year
Early
1860's
1861
1865
1869
1870
1870
1870
1872
1876
1877
1880
1882
1884
Event
First U.S. coal tar distilled In Boston, MA
Three-lift holder tank introduced in England
Phenol recovered from coal-gas liquids for antiseptic
purposes in England
Dyes manufactured from coal tar
Fontana identified Blue Gas by passing steam over
incandescent carbon
Water gas (blue gas) discovered; 330 Btu/ft^, very
poor luminosity
Iron oxide purification introduced to U.S.
T.S.C. Lowe invents carbureted water gas; it has higher
heating value and luminosity than does coal gas
First iron gas holder tank installed in U.S.
Antiseptic and deodorizing solutions produced from
tar-acid oils in England
Indigo produced from coal tar
A considerable percentage of the gas output of the
country was carbureted water gas
Use of down stream for carbureted water-gas production
introduced
Reference3
Lane,
1921
Alrich,
1934
Rhodes,
1966a
Stover
and Chung,
1979
Morgan,
1945
Rhodes,
1966b;
Morgan,
1926
Powell,
1945a and
1945b
Rhodes,
1966b;
Morgan,
1926
Alrich,
1934
Rhodes,
1966a
Stover
and Chung,
1979
Morgan,
1926
Morgan,
1926
(continued)
190
-------
TABLE 45 (continued)
Year
Before
1885
1886
1887
1888
1889
1885 -
1890
1892
1894
1894
1900
Before
1900
After
1900
Event
Lime used as purifying agent to remove C02, H2S, and
organic sulfur
Mantles introduced for gas lighting
First U.S. tar distillation plant installed in
Philadelphia, PA
First steel gas holder tank installed in U.S.
L.P. Lowe patents oil-gas process in the U.S.
Development of rusted iron borings (iron oxide) process
for H2S removal
First U.S. byproduct coke oven installed in Syracuse, NY
(10 years after England and Germany)
First three-lift holder tank installed in U.S.
Byproduct coke plant erected in Johnstown, PA
Pacific Coast oil -gas process developed
Tar removal by bubbling gas through strong ammonia
solution (Livisey washer)
Water-cooled condensers used to cool manufactured gas
Reference3
Downing,
1934
Forstall ,
1934
Rhodes,
1966a
Alrich,
1934
Rhodes,
1966b;
Morgan,
1926;
Hull and
Kohloff,
1952
Downing,
1934
Rhodes,
1966a;
Morgan,
1926
Alrich,
1934
Lane, 1921
Hull and
Kohlhoff,
1952
Downing,
1934
Down i ng ,
1934 i
(continued)
191
-------
TABLE 45 (continued)
Year
Before
1900

Around
1900

Early
1900's

Early
1900 's

1901

1902

1903

1905

1907

1910

Event

Luminous flame burners used for lighting

Vertical retorts used to produce coal gas

Light-oil recovery scrubbers introduced

Direct-contact washer-cooled with P and A tar extractor
introduced for tar removal
Steel gas holder tanks preferred to brick holder tanks;
steel tanks were now cheaper
First use of crude oil in a carbureted water-gas plant
in California
First oil-gas plant installed in Oakland, CA

First oil-gas plant in U.S. installed 1n Oakland, CA;
uses the Pacific Coast oil-gas process
Carbureted water-gas industry begins change from
paraf fink-based oils to asphaltic-based oils
Lime scrubbing replaced by Iron Oxide Purification
in Britian
Centrifuges introduced for separation of emulsions

Washer-cooler introduced; contacted gas directly with
recirculated condensate from gas
Turbo exhauster; used to increase the pressure of
manufactured gas flowing to scrubbers
Aluminia from bauxite used for 83$ removal; this process
was not used very much
First use of water-gas tar to preserve railraod ties;
tar mixed with ZnCl prior to wood treatment
Reference3

Forstall ,
1934

Morgan,
1934

Downing,
1934

Downing,
1934
Alrich,
1934
Morgan,
1926
Morgan,
1926
Rhodes,
1966b
Fischer,
1933
Powell ,
1945
Fischer,
1933
Downing,
1934
Downing,
1934
Downing,
1934
Fulweiler,
1921
(continued)
192
-------
TABLE 45 (continued)
Year
1912

1915

1916

1919-
1920

1920

1920-
1929

1921

1925

1929-
1932

Around
1930

1930

Event
Refiners start cracking petroleum oils to increase the
production of gasoline
World War I spurs development of tar recovery and use in
the U.S.; demand for tar products increases
Water purification process using lime and copperas
(FeS04) followed by coke filter described
Dry-gas holders introduced

Production and prices of coal-tar chemicals dropped
after World War I
Out of the 917 gas plants 1n the U.S., 596 of them are
carbureted water gas

Growing use of phenolic and alkyd resins promotes the
recovery of naphthalene and phenol
Seaboard process for h^S removal introduced

Seaboard liquid process for H2S and HCN removal
developed by the Koppers Co.
Nickel process for H2$ removal and sulfur recovery
invented
....
Great Depression cuts deeply into prices and production
of tar-based chemicals

Horizontal and vertical retorts abandoned or replaced
by oil gas, water gas, or natural gas

Use of heavy fuel oils for oil and carbureted water gas
begins
High surface area iron oxide sponges introduced; they
had double the S removal of homemade FeO
Reference3
Rhodes,
1966b
Rhodes,
1966a
Hansen,
1916
Alrich,
1934

Rhodes,
1966a
Rhodes,
1966b

Rhodes,
1966a
Denig and
Powel 1 ,
1933
Sperr,
1923
Cundall ,
1927

Rhodes,
1966a

Rhodes,
1966a
Downing,
1934
(continued)
193
-------
TABLE 45 (continued)
Year
Event
Reference3
Early
1930's
Early
1930's
1932-
1945
Electrostatic precipitation for tar removal introduced
Tetralin (tetrahydronaphthalene) used to remove
naphthalenes from gas
World War II greatly increased demand and production
of tar-based chemicals
1933 Seaboard H2S removal process installed at 30 plants
1938 Catalytic cracking of crude-oil residuals by refineries
produces high yields of gasoline and gas oil

1949 Federal Power Commission allows certain pipelines that
previously transported oil to carry natural gas
Downing,
1934
Downing,
1934
Rhodes,
1966a

Denig and
Powell,
1933

Pew, 1940
Rhodes,
1966b
194
-------
Land area for the production of gas was generally more available for the
U.S. plants. There was more area for onsite disposal of waste products and
less need to use underground structures for storage (and placing other struc-
tures directly over underground structures).
British town gas sites closed when North Sea natural gas became available
(1967 through 1974). U.S. plants had closed much earlier when pipeline nat-
ural gas from western fields became available (1945 through 1955). Because
the U.K. plants closed later, during a period of increased environmental con-
sciousness, they were generally better decommissioned than were the U.S.
plants.
Britain, a relatively small country, was more homogenous in the produc-
tion techniques and purification processes employed. In the United States,
different production processes were employed In various areas of the country
to take advantage of local resources and markets. Markets for byproducts were
frequently more accessible in Great Britain than they were in the United
States. This meant that the recovery of byproducts was practiced more
extensively in the United Kingdom than it was in the United States. Products
discarded for economic reasons in the United States would frequently be
recovered in the United Kingdom.
Sale and recovery of sulfur from spent oxide was practiced (and profit-
able) in Great Britain. Spent oxide was viewed as a usable byproduct from the
manufacture of gas. The sale and recovery of spent oxide was employed at very
few U.S. plants, and spent oxide was universally viewed as a waste for dispo-
sal. Because spent oxide was utilized in Great Britain, gas plants disposed
less of it and had much less incentive to switch to liquid purification proc-
esses for H2S removal. -The quantities of spent oxide wastes disposed in the
United States were consequently a larger percentage of the spent oxides pro-
duced than were those disposed In the United Kingdom.
Tars and oils recovered from town gas production were more valuable in
Great Britain than they were in the United States (due to higher petroleum
prices in Great Britain). Disposal of tars and oils was much less likely in
Great Britain than it was In the United States. Because coal tar was
generally regarded as more valuable than CWG tars or oil tars, more of the
tars produced in the United Kingdom would have been recovered.
105
-------
The United States was much slower than was Great Britain in distilling
coal tar and recovering coal-tar byproducts. The United States did not start
recovering coal-tar chemicals on a large scale until World War I. This was
due in part to the importation of coal-tar chemicals from Germany and Europe
and also to the use of CWG in the United States. Because CWG tars did not
contain many of the most valuable chemicals in coal tar (e.g., anthracene,
used in the production of dyes), there was less incentive to process the tars
for recovery.
1.7 CONCLUSIONS FROM THE HISTORICAL REVIEW
Three major processes were used for the production of town gas in the
United States. These were (1) coal carbonization, (2) carbureted water gas
(CWG), and (3) oil gas. In general, all three processes were employed in all
areas of the United States, but each process became predominant in specific
geographical ereas in the United States. Gas plants along the West Coast
started as coal-gas plants, switched to CWG, then converted to oil-gas
production. Plants along the East Coast were generally CWG, with some coal-
gas production, and coal-gas production was predominant in the Middle States.
Because the gas purification processes, byproducts, and wastes from the gas
production varied with each production method, it is important to understand
the specific production methods and associated byproduct recovery operations
of individual gas sites.
The feedstocks used in gas production changed during the operation of gas
plants. The coal used for coal carbonization did not change substantially
over time, but the carbon and hydrocarbons used for CWG production and oil-gas
production changed substantially over time, which had a significant effect on
the wastes produced. CWG production originally used coke or anthracite coal
in the generator and low-boiling naphtha fractions «s hydrocarbon feedstock;
Later, bituminous coal often was used directly in the generator, and the
hydrocarbon feed was switched first to gas-oil fractions, and later to heavy
fuel oils and residual oils. Oil gas originally utilized either gas-oil frac-
tions of petroleum or crude oil, but later switched to heavier fuel oils and
residual oils. The choice of feedstocks was determined by the prevalent eco-
nomics of the oil industry during the production of town gas. The conversion
from lower-boiling petroleum fractions (naphtha and gas oil) to heavier oils

196
-------

'If
(fuel oil and residual oil) was accompanied by increases in the tars produced
by the processes and the increased formation of tar-water emulsions. For oil-
gas production, the amount of lampblack produced per 10& ft^ gas manufactured
increased with the conversion to feedstocks with higher carbon contents. The
emulsions that formed were often difficult to separate, and they were often
discarded when separation attempts failed.
Coal carbonization produced a fuel gas containing substantial amounts of
ammonia, cyanide, phenolic compounds, and hydrogen sulfide. The presence of.
these chemicals determined the cleanup processes for their removal from the
gas and any recovery processes. They also appeared in the wastes from coal
carbonization. In contrast, both CWG and oil gas contained only small amounts
of nitrogen compounds (ammonia and cyanide) and only trace quantities of
phenols. All three processes produced gas containing hydrogen sulfide.
Ammonia and phenol were not produced, removed, or recovered from CWG and oil
gas, but they were from coal-carbonization gases. This relatively simple
correlation explains much of the variation seen currently at sites. The
absence of phenols in tars from Stroudsburg, Pennsylvania (oil and CWG), and
Ames, Iowa (CWG), are two more prominent examples. Iron oxide was used almost
universally to remove hydrogen sulfide from town gases. The iron oxide also
reacted with hydrogen cyanide in the gas to produce blue iron cyanide
complexes. These ferriferrocyanides are relatively stable, and they persist
at gas sites that produced coal gas and disposed spent oxides onsite (an
almost universal practice). They are the most visible waste at plants that
produced coal gas, but they are absent from plants that produced only oil gas
or CWG.
The removal of hydrogen sulfide was required for all three gas production
processes, with the amount of hydrogen sulfide removal required being depen-
dent on the coal sulfur concentration for coal-carbonization gases or the
sulfur concentration in oil for oil gas and CWG. Between 1816 and 1855, lime
was used for the removal of hydrogen sulfide and other impurities from town
gas. Lime use was characterized by low conversion of the lime to CaS, diffi-
cult disposal problems, and high cost. The use of lime was essentially
replaced by iron oxide purification after 1890. Both the lime and spent i
oxide were considered wastes; although there were many attempts to use them
107
-------
for some productive purpose, they were universally disposed. Lime use
occurred primarily during a period when the cost of town gas was very high,
and it was used principally to light only streets and shops in cities. With
the introduction of iron oxide purification, gas prices dropped and gas became
a larger consumer item. Spent lime wastes were not a significant problem at
most U.S. sites because of the low gas production rates during the time that
lime was used. Spent lime was also used for agricultural purposes, which
reduced the amounts of spent lime that had to be discarded. Because lime was
also used in the recovery of ammonia from coal gas, spent lime sludges from
ammonia recovery are possible at most coal-gas plants that recovered ammonia
(but it would be present in much smaller quantities than if used for hydrogen
sulfide removal). Spent iron oxides, however, are the predominant waste from
the removal of hydrogen sulfide.
Spent iron oxides were universally regarded as wastes, and they were
often used as a general fill material around gas plants. They constitute a
major discarded waste that can be located on most sites. Unfortunately, there
is wide variation in the composition of spent oxide wastes, which hinders
characterization efforts. Organic hydrocarbon content, sulfur content, cyan-
ide content, and mixtures with woodchips are all variables affecting the cur-
rent composition of spent oxide wastes.
Alternatives to the use of iron oxide for hydrogen sulfide removal were
introduced after 1921. The Seaboard process used a solution of sodium carbon-
ate to scrub hydrogen sulfide from the gas. Solutions were regenerated by
blowing air through the.scrubbing liquid, rereleasing the hydrogen sulfide to
the atmosphere. A process usi-ng a solution of arsenic salts to remove hydro-
gen sulfide and recover it as a sulfur was introduced around 1925. This
process would be accompanied by possible arsenic contamination of sites,
especially if spent solutions were disposed. This process was frequently used
upstream of iron oxide beds (the arsenic process would remove most of the
hydrogen sulfide, and the iron oxide would reduce the hydrogen sulfide content
of the gas to very low concentrations). The spent oxide waste from this type
of operation would have potential arsenic contamination resulting from
carryover of the scrubber solution.
The composition and characteristics of coal- and water-gas tars varied
substantially among plants. Water-gas tars and oil-gas tars tend to be very
198
-------
similar In composition and properties because both are essentially produced by
the thermal cracking of petroleum fractions. They tend to be less viscous
than are coal gas tars, and they contain only trace amounts of phenolic and
base nitrogen compounds.
The formation of tar-water emulsions was a major problem of the industry,
and it frequently resulted in the disposal of these oily materials when the
emulsions could not be broken. Water and tar are condensed simultaneously in
the purification of town gas. The resulting mixture of tar, oils, and water
would usually separate into layers, and the tar and oil could be recovered.
When emulsions formed, the tar would not separate from the water, and the
gravity separators frequently used for the separation would not function.
Emulsions were rarely formed from production of coal gas, but were a frequent
problem for both carbureted water-gas production and oil-gas production.
Emulsions could generally be separated by mechanical and thermal methods, but
occasionally emulsions would form that defied all attempts at separation.
These emulsions were disposed by any means available, including the use of
open, unlined lagoons, direct discharge to bodies of water (where feasible),
or into any convenient unused well. Lagoons were frequently used for storage
of emulsions. This allowed additional time for the emulsions to separate by
gravity or for alternative batch methods of separation to be used. The plant
at Plattsbugh, New York, utilized lagoons for the storage and disposal of tar-
water emulsions.
The formation of emulsions became more prevalent when oil and CWG pro-
ducers switched from lower-boiling petroleum fractions to heavier and higher
carbon-content residual oils.
Tars and oils were generally recovered from the production of town gases.
Although early plants disposed essentially all of their tars and waste conden-
sates (usually to the nearest body of water), they rapidly discovered that
this waste was worth recovering. Coal tars could be separated by gravity from
the condensate and oils. These tars could then be either burned (as fuel in
the retorts or boilers), refined and sold, or sold as a raw byproduct. Water-
gas tars were recovered and sold as a liquid fuel, burned in the plant's own
steam boiler, or recycled back into the hydrocarbons used for cracking into
the gas. All tars had a minimum value to the plant as fuel because the tars
199
-------
could replace a portion of the coal that would normally be burned at the
plant.
Several specific practices contributed to the contamination of gas pro-
duction sites by tars and oils. Many of the original gas holders for plants
were partly buried below ground and frequently filled with coal tar. They
were usually not well sealed at the base, and some of the tar contained in
them leaked into the ground. Tar wells (tar storage tanks) and tar separators
were frequently constructed underground of masonry or cement, and they often
leaked. Some storage tanks were constructed of wood. Wastes were usually
disposed either at the plant site or adjacent to the plant. These practices
indicate that any former gas site will probably have some tar and oil contami-
nation, with the extent of contamination being dependent on the specific prac-
tices of the plant.
Most of the byproducts from town gas production could be considered
either products or wastes, depending on the prevailing price that could be
obtained for the byproduct. Spent iron oxide was always considered a waste,
in spite of continuing attempts to develop uses for the material. Recovered
tars could be sold, but they had a minimum fuel value that determined their
value as a fuel. Plant size and access to markets were two of the primary
factors that influenced the waste disposal practices of gas production plants.
Smaller plants did not have the same economy of scale as did the larger
plants, and frequently they did not recover materials that the larger plants
recovered extensively. This was particularly true of small water-gas and oil
plants, which sometimes let the tars and condensates flow to waste rather than
attempt to recover any o-f the tar. Transportation costs of shipping tars or
ammonia liquors to appropriate markets frequently prevented the sale of by-
products that might have b^en worth recovering.
There is a substantial tendency to apply the work done in the United
Kingdon with old town gas sites to U.S. plants. There are, however, several
substantial differences between plants in these two countries. First, the
United States had abundant petroleum resources, which made the use of CWG and
oil gas practical. The United Kingdom had only limited petroleum resources
and produced coal gas almost entirely. Coal tars and tar products also com-
manded a higher price in the United Kingdom than they did in the United
200
-------
States, thereby encouraging United Kingdom plants to recover these byproducts.
The market for spent oxides in the United Kingdom was well developed (it was
used for the manufacture of sulfuric acid); low sulfur prices in the United
States prevented the development of any markets for spent oxides. Similarly,
liquid-scrubbing methods for the removal of hydrogen sulfide from gas were
developed in the United States, but the United Kingdom plants continued to use
iron oxides because they could market the spent oxides. Gas plants in the
United Kingdom also were generally placed on smaller sites than were those in
the United States. Consequently, wastes from U.K. plants would be more likely
to be hauled away to disposal sites, rather than discarded onsite.
After the first natural gas pipelines were installed in an area formerly
served by manufactured gas, the natural gas was generally used to meet base-
line demand, and the manufactured-gas plant was modified to produce gas for
mixing with the natural gas to meet peak demands. As larger pipelines were
installed for natural gas delivery and better storage methods for natural gas
became available, the need for a standby gas production facility evaporated.
The manufacturing plants were generally idle for several years before they
were decommissioned. The most frequent reason for decommissioning the plants
was to remove structures from the site and reduce the site valuation for tax
purposes. The purpose of site decommissioning was to remove surface struc-
tures from the site. Gas storage tanks were cut off at ground level, and the
tanks were filled with debris from the plant site. Underground tanks and
structures were rarely removed, and some tanks and tar separators were left
filled with tar or liquid wastes. Many gas companies still own the original
sites used for the manufacture of gas, in that it is generally much cheaper to
keep the site as unused-land than it would be to clean the site for sale.
During the literature review, RTI discovered that the literature describ-
ing the operations of gas plants is very substantial. This is not surprising
in that the manufacture of town gas was once a large industry. Several refer-
ences were discovered that deal specifically with the waste disposal practices
and problems of the U.S. industry. These articles indicate that groundwater
contamination in areas around gas sites was common while the plants were in
operation and that contamination of downstream water supplies was also a com-
mon problem.
201
-------
2.0 INVESTIGATION AND REMEDIATION OF TOWN GAS SITES

The investigation and remediation of abandoned town gas sites is a large
task, considering the large number of sites that have been discovered and the
even larger number that remain undiscovered. Contacts made with State and
Federal agencies during the course of this project indicated that, of the
sites that have been discovered, only a few have progressed beyond preliminary
assessments, and fewer still have had remedial actions implemented to address
contamination. Thus, site investigation activities and remedial action activ-
ities at town gas sites should increase markedly over the next few years.
As with any uncontrolled site contaminated with potentially hazardous
chemicals, site investigation activities should focus on determining threats
to human health and the environment posed by the site and on generating the
information necessary to evaluate and select remedial alternatives. Selection
of remedial alternatives should concentrate on cost-effective alternatives
that effectively mitigate the threat, with an emphasis on treatment or des-
truction alternatives that eliminate the hazardous nature of the wastes. This
chapter discusses the behavior of contaminants commonly occurring at abandoned
town gas sites, reviews current practices in investigating and remediating
these sites, and presents recommended practices based on this review. The
case studies, presented in Chapter 3, provide background information support--
ing the information presented in this chapter.
2.1 CONTAMINANT BEHAVIOR AND FATE
The most commonly occurring and environmentally significant contaminants
at abandoned town gas sites are byproduct tars and oils and spent oxide
wastes. Significant aspects of the behavior of these contaminants in the
subsurface environment are discussed in the following sections.
Preceding page blank 203
-------
?.1.1 Byproduct lars and Oils
Byproduct tars and oils represent multiple-density contaminants at gas- (
works sites. For the purpose of this discussion, byproduct oils are defined
as liquid hydrocarbon from gas manufacture with densities less than water;
byproduct tars are defined as liquid hydrocarbons with densities greater than
water. These substances are of concern environmentally because of their
potential to contain high concentrations of carcinogenic compounds, such as
PAH's and nitrogen heterocyclics. From the standpoint of groundwater contami-
nation, the byproduct oils are of most concern because of their higher solu-
bilities and tendency to float on the watertable, where soluble components may
be leached out by infiltration. The byproduct tars are also of concern, how-
ever, because of their potential to flow in density currents through subsur-
face fractures and coarse-grained deposits. A discussion of the hydrogeologic
behavior of these immiscible, variable density contaminants adapted from Alex-
ander (1984) follows.
Byproduct tars and oils from gas manufacture a e immiscible fluids and as
such do not readily mix with groundwater. The flow of immiscible fluids is
more complex than is the flow of soluble contaminants. An immiscible fluid
that is more dense (e.g., tar) than water will migrate according to the com-
bined effects of relative density and the fluid-fluid and fluid-solid inter-
facial pressures. Because of the density contrast, the fluid will generally
sink within the groundwater. Lighter hydrocarbons, such as byproduct oil,
will generally "float" on the water table or on the tension-saturated zone.
The existence of capillary pressure in a two-phase flow system means that the
migration of an immiscible fluid is not entirely dependent on the flow of
groundwater and, as a result, can migrate in an opposite direction of the
dominant flow system. It is not uncommon in spills of low-density fluids, for
example, for the fluid to migrate "upgradient" of the groundwater flow system
within the capillary fringe. The theoretical aspects of multiple-phase flow
of hydrocarbons in the subsurface are discussed in detail by van Dam (1967).
One of the biggest problems associated with the release of the lighter
hydrocarbons into the subsurface is that their relative solubility increases
204
-------
the volume of groundwater that is contaminated. Rainwater that percolates
through the "pancake" of light hydrocarbons typically formed over the ground-

water body eventually weakens the concentration of oil causing dissolved com-

ponents of the oil to enter and be transported by the flow of groundwater

(Dietz, 1971).
An example showing the soluble component of an immiscible contaminant is
provided in Figure 58 with the following designation of zones (from Pfannkuch,

1982):

• Zone I is the above-ground and surface zone where leaked or projec-
ted oil runs off ard collects in surface depressions, thus forming
the area from which infiltration takes place. The configuration of
this area depends on the local topography, the amount spilled, and
the conditions of release or eruption.

• Zone II is the soil profile. From Zone I the oil starts infiltra-
ting into the subsurface via the organic soil layer, if such a layer
is present. This zone is characterized by its high organic content
and high moisture content due to soil structure. If the soil is
oleophilic, it has a much higher oil retention capacity than do the
underlying nonorganic deposits.

• Zone III is the vadose or unsat'irated zone. This is the most impor-
tant zone for oil retention. Water saturates the pore space only
partially and ranges in value from zero to field capacity. Oil, as
the nonwetting phase, moves downward under the forces of gravity.
At first it moves as a more or less continuous phase or "oil body,"
displacing excess water from the larger pores. When all oil has
infiltrated from the surface, the "oil body" will move downward by
translation, but small amounts of oil will be left behind the trail-
ing end, trapped as insular disconnected droplets. The oil body
continues to move in a disintegrated fashion until all of the oil is
trapped in the pore spaces of the vadose zone if its total retention
capacity exceeds the infiltrated spill volume. Any oil in excess of
this total retention capacity reaches the groundwater body and
spreads on the water table through the capillary fringe.

• * Zone IV is the capillary fringe that is partially watersaturated,
directly connected with the groundwater body vertically, but contin-
uous laterally. When excess oil reaches this zone, it will spread
laterally under its own hydrostatic pressure and form a lens on the
water table. The spreading will halt when the hydrostatic forces in
the oil phase are counterbalanced by the capillary forces at the
outer edges of the spreading oil lens. This movement is governed by
the phenomena of relative permeabilities and multiphase flow in
porous media.
205
-------
SURFACE
Vwr: VVATIR TABU GROUNDWATER FLOW
CF : CAPILLARY FRINGE
I Surface zone
II Soil profile
III Vadose zone (unsaturated)
IV Capillary fringe
V Groundwater body
Source: Pfannduch, 1982.
Figure 58. Subsurface propagation of a nonmiscible containment.
206
-------
If the porous medium is homogeneous and Isotropic and the water
table is horizontal, then the oil lens would be perfectly circular
around the center of infiltration. In most realistic cases, the
water table has a slope that gives rise to an elliptically elongated
lens extending in the direction of the flow. The shape of this lens
depends on the water-table gradient, groundwater flow velocities,
the capillary properties of the multiphase flow system, and the
shape and orientation of the original infiltration area.

• Zone V is the groundwater body. Most hydrocarbon compounds in a
spill are lighter than water and therefore tend to float on the
water table. Under the hydrostatic head of the continuous oil
column, an actual depression and penetration of the groundwater body
below the water table occurs. This inverted mound will dissipate as
the overlying oil body spreads laterally. The penetration and sub-
sequent retraction may result in leaving trapped insular oil behind
in the groundwater body. The most important feature of Zone V in
the emplacement stage is the formation of an interface between the
bottom of the oil lens and the free-flowing groundwater. It is at
this interface that small but significant amounts of hydrocarbon
compounds go into solution with the water and are spread by convec-
tive and dispersive transport mechanisms.

Model experiments have been useful for studying the mechanism of low-
density oil spread in porous media above the water table (Schwille, 1967).
The seepage and spreading of heating oil in layers of varying hydraulic con-
ductivity and hydraulic gradients are shown in Figure 59. The oil seeps

downward under the influence of gravity, and its geometry is influenced by the
rate of infiltration, the hydraulic conductivity, capillarity, and the hydrau-
lic gradient.

Multiple discharges of different kinds of chemicals can lead to a complex
pattern of contaminant plumes (Figure 60). In this example, the heavy petro-
leum product that is denser than water is flowing down the slope of the con-

fining bed in an opposite direction to the flow of dissolved and low-density
products. Migration of heavy coal-tar derivatives through density currents is
illustrated by a case described by Berggreen (1985), in which creosote has
migrated along slickensides (fractures) in a low-permeability clay to bedrock
at a depth of 120 feet. Byproduct tar migration through density currents is
illustrated by the Brattleboro, Vermont, and St. Louis Park, Minnesota, case
studies in Chapter 3.
207
-------
I 3 t oil

oo
BOO
— z_-
«s«*
na°sT
fioSJ
i i 1
iii
S 3 =
1 1 1 1
s ss a
sa
goo
51 all
100cm
Source: Schwille, 1967.

Figure 59. Seepage and spreading of heating oil in porous media
above the water table.
208
-------
SOURCE OF PRODUCT
( CnqUr dinilly than woltf)
SOURCE OF PRODUCT
( UttMr <«n«itj than wttr )
OIReCTION Of
GROUND-WATER FLOW
CONFINING BED
Source: Miller, 1983.
Figure 60. Effects of variable density migration in the subsurface.
209
-------
2.1.2 Spent Oxides

Spent oxides are extremely heterogeneous and variable in nature, as dis™
cussed in Chapter 1. The most significant contaminants in spent oxide wastes

are sulfuric acid, arsenic, and complexed iron cyanides. These complexed

cyanides occur in the form of ferric ferrocyanide, imparting a blue color to

the spent oxide wastes.

There has been considerable research on the fate and transport of cyanide
compounds in the environment by the mining and mineral-processing industry,
which uses cyanides to leach metal-containing ores. A recent symposium (van

Zyl, 1984) summarized the state of knowledge on this subject, but it also
pointed out many gaps in the knowledge necessary to predict environmental

impacts accurately. Many of these gaps concerned iron cyanide complexes.
Conclusions from this symposium of relevance to this study are:

• Low levels of free cyanides do not persist to soils because of bio-
logical and chemical degradation. Biological degradation in soil is
inhibited by concentrations of 2 ppm free cyanide under anaerooic
conditions and 200 ppm free cyanide under aerobic conditions.

• Ferro- and ferricyanide complexes in solution are photodecomposed to
free cyanide. Their toxicity in water is related to the degree of^j
decomposition. ~

• When KCN in municipal landfill leachate is passed through saturated,
anaerobic soil, Prussian blue (ferric ferrocyanide) precipitates and
accumulates in the uppermost soil layers. This suggests that Prus-
sian blue is quite immobile in soil.

Tree cyanide migration in saturated, anaerobic soils increases with
increasing CaC03 content and decreases with increasing concentra-
tions of Mn and hydrous iron oxides.

• Complexed iron cyanide (Fe(CN)g~3) migration in saturated, anaerobic
soils is retarded by high free Fe03 and increases with increasing pH
and CaCO] content. At low pH, iron cyanide mobility decreases with
increasing clay content.

This information suggests that complex iron cyanides are relatively immobile
in a municipal landfill environment and that chemical treatments may be devel-
oped for complexed iron cyanides that will limit releases of free cyanides in

the soil environment to levels that can be biologically degraded.
210
-------
2.2 SITE INVESTIGATION
2.2.1 Introduction
Our review of case studies (Chapter 3) and assessment of past disposal
practices (Chapter 1) have indicated that appropriate procedures for conduct-
ing hydrogeological investigations of town gas facilities are not signifi-
cantly different from those used for investigating uncontrolled chemical and
industrial waste sites. The primary difference is that town gas sites gener-
ally tend to be older, and less background information is available about past
site activities. In many cases, the present-day site has been cleared, and
little or no evidence of past site activities is visible at the ground sur-
face. As a result, research into historical records often is necessary to
determine the physical layout and operating history of the plant. As with any
investigation of an industrial site, it is extremely important to utilize
process information to help determine what contaminants may be present at the
site and where these materials may be located.
2.2.2 Current Practices
Most investigations of manufactured-gas plant sites rely on conventional
site investigation methods that are not significantly different from contami-
nation investigations of other industrial sites. These methods include sur-
face water sampling, shallow soil and groundwater sampling (from borings and
test pits), and, when necessitated by the results of these sampling activi-
ties, more extensive groundwater monitoring. In many instances, these methods
appear adequate for an initial understanding of the potential for adverse
impacts on human health and the environment. A typical approach used in the
investigation of manufactured-gas plant sites is summarized in Table 46.
Actual case studies are presented in Chapter 3 of this report.
It is apparent from RTI's review of relevant case studies (Chapter 3)
that other potentially useful (and often cost-effective) alternative tech-
niques of investigation, such as geophysics and soil-gas sampling, have not
been extensively employed at manufactured-gas sites to date. However, based
on limited use at manufactured-gas sites and more extensive utilization at
industrial waste sites, these techniques show potential utility for screening
sites to optimize sampling and analysis plans.

211
-------

VI
M
01

TURED-C
u
a
IE
o
u.
w
i
£
0.
<
Ul
H
5
H
W>
Ul
>
2

a.
u.
o

;
Ul
2)
^
H

g|
.•
*
0
U

u
'c

o
a
k
3
a.

^
0
5
j

i
o •fl
3 C 4>
k * «l —

CD • — t
o -a a « a >
uk c 13 n f a k
o — c c a 3 3
- o c • o • m/i
• • k •- k
•ok t • « « 01 k —
k O « • k O t *
o • c t a t o o u
o o o o •> o • —
tc u— E x » a
k * o 41 o. a
k « 0 U C *• —
M3 t • — t • OO
•III tkXU4l t
• c x a a t — «u
c •- o t k n c
._ — — k o •
• •a a* a t -o — v>
3C E « O k — O •
x- * to a o o • D
O. O £ u w
ax k « o 73 o
cm t— x c — "
— 41 6 -o >o • —
•O k • •- • —
3- o t a » nxt
— M <~ 0 • X 0 *
u c « E c a
c • •o* o • aia
— — c * at « k cc
a • c— a ~—
• -o — k o c k
•1C * C C • 41 k O
TJ O £ • OlXO • O
k — .- M S X U "
04i on o a c c
o * • i t — u o g
tU« OKI • — O t
kO4l C« U-O • •
— k at oc — • k ^
X O — U — • k 0 t 11
exa o t « o o
a— k o a t • * t
a — •- -o 4> TJ x-o E
o — — »c — k — ecc
0 0 • O OO O • 3 k
^30 » — a o •
o -o c t • Ok o oaa
c x — u at
(•O >•- t X * * k
t t k*. tt t tO«
.- C 41 t — — > — — -X *>
cc o 13 >-. at tc • cc. K • v>

n —

k t OX
t k — 41 •
!" i§ ^
0— X 0 •
«• • U k

•ol "S^ ^
c.2 t!f °
O TJ t k 41
0.0 C
41 • • •«. t
• 41 3 !
• nun a
em o o

t -O 0 t> M >
•o c out
• O k • TI ' ~
t 41 a u
X H 0 t •
4l C 4l t k X C

41 4i .- • c c a
— t • ~ •- 4* 4*1:
— a. on • •
— o e o—i
u — • — • •
• 4l • k»- • *
•- C 01 3 — • t
• 41 H II
>- a v- e 01 i- •
u
k
5
N

X
O

ac -o
C V. k
^ O O • •- X
t ... t 3 k
k x *» e t> o- o
3 a o «i~1 — * t
?C 4l * 4l • U k
•- x it • e e o
k — 413 X C - — 0 X
f— u u — e — •
— 1 • E 0 4» 4) —
• *. t k — «O
• M30 t -1 X t -0
O- 1 41 41 • 41 4) C
II X "»s C « t > O •
Sot— -ok e k
« u « -o t no. -o
t ex ••« — —
41 • U k X • CL«- • t
It • 4lk *-4lk OH-
t-rf • 3 - 3 a X 0 •-
— • nil O" k 4i— u
41 41 «~ CO- « W • k
01 • • • k •. C «>>4l O
> C • X O • t C •-
— •- k X— • TJ «- 11 — O
u a 4i — c — -o o -o t •
t fi c — • at* cxx —
• ••- • Et— • *> o —
tun 3 • • c— • c je
41 t *- 0*01 « o t X - 1 u
xk o e — o— o • at •
• «»3 t— «^« a— -okx
— TJ * >— O — t £ "kt U
4>4l| t — O. k 0 • ••£
• CO 41 • 6 k t • II JIM* 4J
• • o x e * • o— t e
a -o k — tit .a t k x •« x •» i
— a • x EEO.U • c s
•a c t t 3 o <• — — • a.
c o-o • k-o c dk • — « —
• 4) e a 3 c a •— » 3
• • E — • t — a. k o 41 — o-
t • x o o e x — • o t e t t
o wo o o e o o a »- ^ t c
— c c — — i-o J c e
*E o o • o • •— a *
•» • a »- o 3 k o a
C CO*— 4) • > — O— 4l 41 •
o c t 41 c— t o t xk c e t e«
o • t a • E — o too ••
41 o a c E • c s • f — — i —
a • c .. a.— o •• *• c • OJ.41W 0.4^
- * - E o o « lota o t c o ~
k — k — k -o k > a — 30— >
u t a t t o t x k •— t —
O - *
C X • 4>
o •« u -o c
u • e t • k t
• • k • • 1
•k C 3 4> M *
o o • o t • k
— > o-o x —
t 4l .- • 3
— • 4i a — T
•a — —
41 — t •
HO »2S (2"S. »2x
.
T
e
•* •
0 —

e 41
H
•- •
Ji

t *Q 1
,: X C «
ft »
4> C « *
.. « M
•a ii k •
t -o 3
* « t -a •
x c to
k o .- o —
0 — 0 II

•0—3 a •
CO X
f5 • O

k— t O C

O— 3 X
• 41-0 41-O
» C t — •
• 0 * C
»Tj o ~
• ™ a O4»
»- « . c 3
k e x • o
300 -0
* •- — k n
o » o e
•^ M U O
o o t o —
M U • *l
j »S ,c S
0— k ""£
«— o *~
> a u • u
k fi U • — •
• • • c a a
•in • En
k 6 .^ O. II 9
a o o
k c o • e
Kii fi
• n •» n
C— 4l * O «l •
o a c t • c
— E f • • •
o • i « — x—
u • • • • 4i a
t <• x—
— J3 t t ——0
c» o • • x a 6

i
c
o
u
X 0

— t
3 at
V •
k
t -o c
o co
• ••-
t 41
41 •

3 • 41 e
o • * •
k • to
oik e
jc • to
u x u
• • o
.OX k
0 • •
^ C **
-" S?
SS o!
•DO • •
— •• ^ %•
*» u
O — O 3

H
« k
S3
A

* •
— 3
K
^

e
o
c
o
u

-
212
-------
f!

!
e
e
o
u

Ul
5

«
i

':
i
4

t
t
'a
x
h-

0
O
a
k
3
a.

^
0

4*
=

VI *

C *•" M
0 0. -0-
— — E c a
4>— 4 4
u'» k * e
o o-o • •
— • 0 k I
— 3 «
*O 4 C • ^ ft
t 3 O O • •
4J -O — X U M
U •- 4» JS OH
• fl <• V k •
a • > 4 a
M k U JS •
3 • • 4*
M • •» - X —
k • k 4* M
• • k •
xx GL a x t
4» I 3 4> X
•04 —4)
CMC t
O 4 4 "fl C
• e ••-
•O k - • U
• 4 ex e-o
M O 4 4
• k — C» T) C
J3 • 4> C k —
X • — 0 -
^ «•* N il U **
... 0 .-3 03
k k 4 O
a k • •>
o 4> x en
c — no. k
— • k • • 3

O, O X — •
:.: HI: I1
-O *> .- 3 ft •
• •- k O" O
• II Q.— • C
•• 4» e o e x •
• a. u 3
4> k - o. • *• •
e c • 4» on
?k O 4
• — • a 4
Ok* U 4* •— * 4*
— Ok • — X—
• 0.3 — e — —
> u — O — 4 •
• 0-U Okk C3
o o o u— TJ