I.C. 9027. 1985 
 
(J.^ ,o«., -^^^ 
 
 
 o V 
 
 '>-0^ 
 
 
 
 . »» A 
 
 <-. "'..s 
 
 
 . »* A 
 
 • , 1 ■• ^^ 
 
 
 
 ^oV" 
 
 
 '^oV^ 
 
 ,-^°^ 
 
 ^^-^t. 
 
 
 <^^ . o ~ o 
 
 ^oV 
 
 
 '^o^ 
 
 o^'^^*. 
 
 •^o^ 
 
 .^^ 
 
 
 Xp <^^ *j(C^ 
 
 
 vV^. 
 
 
 ,-1°^ 
 
 A 
 
 V" ^l.^rl'* «^ aO 
 
 ^"-^t. 
 
 .* ^^ 
 
 
 ,4o«. «^ 
 
 *«^ 
 
 
 
 ^^ * 
 
 
 
 
 
 
 
 v-^^ 
 
 /.v:«k- °o >*,-^i.\ o°*.c^>o y.-^i.V co*.^^-> /.-ji^;:.**. 
 
 ^t^ 
 
 4 o 
 
 <i> «<> <" 
 
 
 '^'O^ 
 
 <ci^ « 
 
 'bv" 
 
 ^o-n^. 
 

 
 
 
 
 V'-*'"•^<^" %-:^?^•>" v---'\'^'' ■%''-''^->*' V•-<^••^^* ^' 
 
 
 
 'bV 
 
 >i9 -^^ ' " 
 
 ^^^\ ^'^^S J'''\ "^^S ^'h^^\/'^^/ ./\ 
 
 '^oy 
 
 
 
 
 
 
 'C *'T.»* ,0 
 
 •^^a^ 
 
 *w* .-^^^ \/ .-j^^--. %<^^* .-life'- \/ yMM'_ %.*^ /^■•. *-^-„./ •• 
 
 ,* ... 
 
 
 
 
 '^K' 
 
 '^^r 
 
 -J>c,- 
 
 
 
 
 ^^ .0-., ^^ 
 
 
 
 ^ A 
 
 '<"^ 
 
 \V ft • • 
 
 
 4 CU 
 
 
 ^^-^^^ V 
 
 
 '^V 
 
 
4 
 
Bureau of Mines Information Circular/1985 
 
 Control of Acid Mine Drainage 
 
 Proceedings of a Technology Transfer 
 Seminar 
 
 By Staff, Bureau of Mines 
 
 If- 
 
 5bv.>c 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
 C 
 
 x 
 m 
 
 > 
 c 
 
 9c 
 
 j^S^^^ I 
 
 ^(NES 75TH AV'*'^ 
 

 Information Circular 9027 
 
 Control of Acid Mine Drainage 
 
 Proceedings of a Technology Transfer 
 Seminar 
 
 By Staff, Bureau of Mines 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
 Donald Paul Hodel, Secretary 
 
 BUREAU OF MINES 
 Robert C. Norton, Director 
 
1 
 
 
 1 
 
CONTENTS 
 
 Abstract 
 
 Introduction , 
 
 Prediction of Acid Drainage Potential in Advance of Mining, by Patricia M. 
 
 Erickson, Richard W. Hammack, and Robert L, P. Kleinmann , 
 
 Hydrologic Aspects of Acid Mine Drainage Control, by Kenneth J. Ladwig , 
 
 Oxygen Content of Unsaturated Coal Mine Waste, by Patricia M. Erickson , 
 
 Control of Acid Mine Drainage by Application of Bactericidal Materials, by 
 
 Patricia M. Erickson, Robert L. P. Kleinmann, and Steven J. Onysko , 
 
 Alkaline Injection: An Overview of Recent Work, by Kenneth J. Ladwig, 
 
 Patricia M. Erickson, and Robert L. P. Kleinmann , 
 
 Comparative Tests To Remove I-langanese From Acid Mine Drainage, by George R. 
 
 Watzlaf , 
 
 Treatment of Acid Mine Water by Wetlands, by Robert L. P. Kleinmann , 
 
 In-Line Aeration and Treatment of Acid Mine Drainage: Performance and Prelimi- 
 nary Design Criteria, by Terry Ackman and Robert L. P. Kleinmann 
 
 Page 
 
 1 
 
 2 
 
 3 
 
 12 
 
 19 
 
 25 
 
 35 
 
 41 
 
 48 
 
 53 
 
UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT 
 
 With Factors for Conversion to U.S. Customary Units 
 and the International System of Units (SI) ' 
 
 Abbreviation 
 
 Unit of measure 
 
 To convert to — 
 
 Multiply by- 
 
 - 
 
 or unit 
 
 
 
 
 
 acre 
 
 acre 
 
 hectares 
 
 0.405 
 
 
 cm 
 
 centimeter 
 
 inches 
 
 0.3937 
 
 
 ft 
 
 foot 
 
 meters 
 
 0.3048 
 
 
 ft2 
 
 square foot 
 
 square centimeter 
 
 929.0 
 
 
 ft3 
 
 cubic foot 
 
 cubic meters 
 
 0.028 
 
 
 ft/s 
 
 foot per second 
 
 centimeters per second 
 
 30.48 
 
 
 g 
 
 gram 
 
 ounces 
 
 0.0353 
 
 
 gal 
 
 gallon 
 
 liters 
 
 3.785 
 
 
 gal/h 
 
 gallon per hour 
 
 liters per hour 
 
 3.785 
 
 
 gal/min 
 
 gallon per minute 
 
 liters per minute 
 
 3.785 
 
 
 h 
 
 hour 
 
 NAp 
 
 
 
 ha 
 
 hectare 
 
 acres 
 
 2.471 
 
 
 in 
 
 inch 
 
 centimeters 
 
 2.54 
 
 
 kg 
 
 kilogram 
 
 pounds 
 
 2.205 
 
 
 L 
 
 liter 
 
 cubic inches 
 
 61.025 
 
 
 lb 
 
 pound 
 
 kilograms 
 
 0.4536 
 
 
 Ib/min 
 
 pound per minute 
 
 kilograms per minute 
 
 0.4536 
 
 
 L/min 
 
 liter per minute 
 
 gallons per minute 
 
 0.2642 
 
 
 m 
 
 meter 
 
 feet 
 
 3.28 
 
 
 m3 
 
 cubic meter 
 
 cubic yards 
 
 1.308 
 
 
 mVs 
 
 cubic meter per second 
 
 gallons per second 
 
 264.2 
 
 
 mile 
 
 mile 
 
 kilometers 
 
 1.609 
 
 
 mg 
 
 milligram 
 
 grains 
 
 0.0154 
 
 
 mg/L 
 
 milligram per liter 
 
 NAp 
 
 
 
 mg/(L-h) 
 
 milligram per liter per hour 
 
 NAp 
 
 
 
 min 
 
 minute 
 
 NAp 
 
 
 
 mL 
 
 milliliter 
 
 cubic inches 
 
 0.061 
 
 
 mm 
 
 millimeter 
 
 inches 
 
 0.0394 
 
 
 mmho/m 
 
 millimho per meter 
 
 NAp 
 
 
 
 ym 
 
 micrometer 
 
 inches 
 
 3.94 X 10- 
 
 5 
 
 pet 
 
 percent 
 
 NAp 
 
 
 
 psi 
 
 pound per square inch 
 
 grams per square 
 centimeter 
 
 70.307 
 
 
 std ft^/min 
 
 standard cubic foot per 
 minute 
 
 NAp 
 
 
 
 ton 
 
 ton 
 
 metric tons 
 
 0.907 
 
 
 yr 
 
 year 
 
 NAp 
 
 
 
 NAp Not applicable. 
 
 ^ Owing to the preference of individual authors, U.S. customary and 
 both been used in this report. Conversion factors are provided for the 
 the reader. 
 
 SI units have 
 assistance of 
 
CONTROL OF ACID MINE DRAINAGE 
 
 Proceedings of o Technology Transfer Seminar 
 
 By Staff, Bureau of Mines 
 
 ABSTRACT 
 
 Acid mine drainage can be controlled by water treatment, retardation 
 of the pyrite oxidation reaction system, or enhanced prediction that 
 allows preventive action to be taken. The Bureau of Mines is conduct- 
 ing research in each of these areas; the results of this research are 
 summarized in the eight papers that comprise this volume. Field work, 
 to evaluate overburden analysis, alkaline injection, and bactericidal 
 control of acid formation is described, along with two new inexpensive 
 methods to treat acid mine water. These papers were prepared for an 
 acid mine drainage technology transfer meeting held in Pittsburgh, PA, 
 on April 3 and 4, 1985. 
 
INTRODUCTION 
 
 Acid drainage from coal mines is one 
 of the most persistent industrial pollu- 
 tion problems in the United States. Over 
 5,000 miles of streams and rivers are 
 adversely affected, primarily by under- 
 ground mines that have been abandoned 
 for decades. Meanwhile, at active mining 
 operations and at sites where mining oc- 
 curred after 1977, discharge water must 
 be treated to meet fairly stringent regu- 
 latory limits — at a cost to the industry 
 of over $1 million per day. 
 
 The Bureau of Mines has a special re- 
 sponsibility to facilitate integration of 
 mining and mineral processing with envi- 
 ronmental safeguards. This responsibil- 
 ity is twofold: the development of tech- 
 niques to reduce or eliminate environmen- 
 tal degradation, and the Improvement of 
 existing pollution control processes to 
 make them more efficient and more cost 
 effective. Research in acid mine drain- 
 age exemplifies the Bureau's concern for 
 these environmental aspects of mining. 
 
 This collection of papers summarizes 
 much of the Bureau's recent research on 
 acid mine drainage and will give the 
 reader a sense of where the research is 
 headed, in addition to providing details 
 regarding new technology and recently ac- 
 quired knowledge. The research papers 
 address four basic objectives: 
 
 1. Improved prediction of acid poten- 
 tial. — The Bureau is attempting to ad- 
 dress three fundamental problems associ- 
 ated with premine prediction of acid mine 
 water: (1) the lack of field verifica- 
 tion for currently available techniques 
 using overburden analysis, (2) difficul- 
 ties encountered when one attempts to in- 
 corporate pyrite reactivity and kinetics 
 into premine prediction, and (3) incor- 
 porating the effectiveness of reclamation 
 measures in predicting eventual acid pro- 
 duction. Available techniques are being 
 evaluated at sites where the extent of 
 acid production from reclaimed spoil can 
 be monitored. This will enable the Bu- 
 reau to evaluate their applicability and 
 
 the effect of potentially mltigrating 
 measures taken by the mining companies. 
 This research should lead to better per- 
 mitting by State agencies, improved mine 
 planning, and improved reclamation. 
 
 2. Improved mine planning. — Improved 
 prediction will allow an awareness of the 
 potential problem, but the effect that 
 mining methods and procedural changes 
 will have on the extent of the problem 
 must still be systematically determined. 
 Fundamental aspects of such factors as 
 hydrology and oxygen diffusion must be 
 understood before modified reclamation 
 plans and new closure methodology can be 
 developed to prevent acid mine drainage 
 in the future. 
 
 3. At-source control of acid forma- 
 tion. — Current Bureau research indicates 
 that it is possible to reduce acid loads 
 under certain conditions using long-term 
 inhibition of bacterial catalysis at or 
 near the surface, or chemical treatment 
 to reduce pyrite reactivity. The lat- 
 ter will most likely require the estab- 
 lishment of a near-neutral pH regime or 
 low-Eh environment. Reduced acid loads, 
 although less desirable than total 
 prevention, are now achievable. Water 
 treatment costs and reclamation costs can 
 both be reduced if acid production is 
 decreased. 
 
 4. Improved water treatment. — The Bu- 
 reau has developed two low-cost alterna- 
 tives to conventional mine water treat- 
 ment facilities. For low flows of acid 
 water, a low-maintenance system, consist- 
 ing of a Sphagnum moss wetland to remove 
 iron, followed by limestone neutraliza- 
 tion, has been demonstrated to be effec- 
 tive in a pilot-scale test; full-scale 
 tests are in progress. For higher flows, 
 the Bureau has developed a pipeline neu- 
 tralization and aeration system that can 
 be scaled up or down to meet most treat- 
 ment needs; the entire system costs only 
 a few thousand dollars and appears to be 
 more efficient than a conventional treat- 
 ment facility. 
 
PREDICTION OF ACID DRAINAGE POTENTIAL IN ADVANCE OF MINING 
 
 By Patricia M. Erickson,^ Richard W. Haramack.,2 
 and Robert L. P. Kleinmann^ 
 
 INTRODUCTION 
 
 Surface coal mine operators are re- 
 quired by law to identify the potential 
 for acidic drainage prior to opening a 
 new mine (9^).'^ In many cases, particu- 
 larly in the Appalachian region, the per- 
 mit application must contain the results 
 of overburden analyses intended to quan- 
 tify the acidic or alkaline weathering 
 products of the affected strata. These 
 data serve two purposes: to provide the 
 regulatory agency with a means to esti- 
 mate the hydrologic consequences of the 
 proposed mine, and to allow the proposed 
 operator to plan the mine with regard for 
 probable water treatment requirements. 
 Until the Bureau's current project, there 
 has been no systematic field evaluation 
 of these analytical techniques. 
 
 The acid-base account is the most com- 
 monly used overburden analysis technique 
 (6) . The method is based on measuring 
 the total sulfur content of each litho- 
 logic unit and converting that value to 
 an acid potential based on the stoichio- 
 metry of complete pyrite oxidation. Sim- 
 ilarly, the neutralization potential is 
 determined for each lithology by its 
 ability to neutralize strong acid. The 
 two values, acid and alkaline potential, 
 respectively, are represented as calcium 
 carbonate equivalents for calculation of 
 a net excess or deficiency of neutral- 
 izers. A deficiency greater than 5 tons 
 CaC03 per 1,000 tons of rock is generally 
 considered a potential source of acid 
 mine drainage (11). 
 
 ^Supervisory physical scientist. 
 
 ^Geologist. 
 
 ■^Researcn supervisor. 
 Pittsburgh Research Center, Bureau of 
 Mines, Pittsburgh, PA. 
 
 '^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 The acid-base account uses rapid and 
 simple analytical techniques; it is, 
 therefore, inexpensive. The results, 
 however, indicate only the total acid and 
 alkaline loads that could be produced if 
 all the pyrite and carbonates reacted. 
 The major flaw in interpreting these data 
 for water quality is that reaction kinet- 
 ics are ignored. Complete oxidation of 
 pyrite may take decades, even if all the 
 pyrite is reactive; acidity in solution 
 is determined by the rate of oxidation 
 and flushing. 
 
 In contrast, calcium carbonate dis- 
 solves rapidly to an equilibrium value of 
 approximately 60 mg/L alkalinity at at- 
 mospheric CO2 partial pressure (_5) . More 
 carbonate mineral dissolves to achieve 
 the same equilibrium concentration if 
 acidity is present; higher concentrations 
 can be dissolved at high carbon dioxide 
 partial pressures (]_) • Because solution 
 acidity and alkalinity are controlled 
 largely by the kinetics and thermodynam- 
 ics of many reactions, use of total mass 
 balance data to predict water quality is 
 suspect. The developers of the acid-base 
 account technique did not intend it to 
 predict drainage quality, but rather to 
 identify strata that may produce acid. 
 
 Other overburden analytical mathods can 
 be classified as simulated weathering 
 techniques. These have in common that 
 the strata, either individually or as a 
 thickness-weighted composite, are sub- 
 jected to oxidizing conditions to accel- 
 erate chemical weathering of the materi- 
 als. Chemical composition of drainage 
 obtained from periodically leaching the 
 sample is classified as acidic or alka- 
 line and is presumed to qualitatively 
 predict the nature of postmining drainage 
 at the proposed mine. 
 
Oxidative processes can be accelerated 
 by heat, addition of chemical oxidants, 
 reduced particle size of the solid phase, 
 inoculation with bacteria, and other 
 means ; innumerable protocols could be de- 
 vised for weathering tests. Two methods 
 that have been used for overburden anal- 
 ysis utilize crushed core samples sub- 
 jected to humidified air streams; the 
 techniques differ in that one utilizes 
 individual weathering tests for each 
 lithology (2) and the other utilizes a 
 composite sample assembled according to 
 the backfilling plan (8^) . 
 
 Weathering tests may provide a real- 
 istic estimate of postmining drainage 
 quality if they duplicate the kinetics 
 of relevant reactions under field condi- 
 tions. These tests require longer pe- 
 riods of time, from several weeks to 
 months, and are more expensive to use 
 than the acid-base account technique. 
 The accuracy of any simulated weathering 
 technique must be verified to determine 
 its predictive capability. 
 
 Prediction of acid drainage potential 
 from overburden analyses and other pre- 
 mining data relies on interpretation by 
 
 mine operators , consultants , and regu- 
 latory personnel. The data merely indi- 
 cate the maximum acid and alkaline loads 
 (acid-base account) or drainage quality 
 under a given set of conditions (weather- 
 ing tests). Effects of mining-related 
 factors such as mining method, use of se- 
 lective handling, and ameliorant applica- 
 tions are not considered in overburden 
 analysis. There is no consensus on a 
 method to combine the lithologic data, 
 mining plans, and reclamation plans into 
 a predictive scheme. 
 
 The Bureau is currently conducting 
 contract and in-house research to im- 
 prove acid drainage prediction. The con- 
 tract research consists of two phases: 
 (1) field evaluation of three overburden 
 analytical techniques at 30 mine sites 
 and (2) design of an empirical predictive 
 scheme that encompasses overburden data 
 and site-specific factors that could in- 
 fluence actual drainage quality. The 
 in-house research is oriented toward de- 
 veloping an alternative overburden analy- 
 sis method that takes into account pyrite 
 reactivity. Both projects are discussed 
 in following sections. 
 
 CONTRACT RESEARCH PROGRAM—PREMINING PREDICTION OF ACID DRAINAGE POTENTIAL 
 
 The Bureau awarded a research contract 
 to Engineers International, Inc., in 1982 
 to improve the state of premining pre- 
 diction. Phase 1, nearly complete now, 
 addresses the validity of using available 
 overburden analysis techniques to predict 
 postmining drainage quality. Phase 2 
 will focus on the development of an em- 
 pirical predictive scheme encompassing 
 mining-related variables. 
 
 FIELD EVALUATION OF OVERBURDEN ANALYSIS 
 
 The objective of this phase of the re- 
 search is to determine the utility of 
 three overburden analysis techniques for 
 predicting drainage quality after mining. 
 To accomplish this goal within a reason- 
 able time, the program plan called for 
 the collection of actual postmining data 
 
 and the equivalent of premining overbur- 
 den data from 30 reclaimed mines. The 
 validity of this phase depends mainly on 
 obtaining overburden samples that rep- 
 resent the overburden in the reclaimed 
 section. At nine sites, cores sampled 
 less than 1 yr ago were available for 
 analysis. At the remaining sites, chan- 
 nel samples from an active highwall adja- 
 cent to each reclaimed mine section (used 
 as the postmining water data source) were 
 collected for overburden analysis. This 
 method was chosen for two reasons: 
 (1) The cost was much lower than the cost 
 of drilling cores on adjacent unmined 
 land, and (2) visual observation could be 
 checked against company records to verify 
 continuity of the overburden lithology. 
 Fresh material was exposed on the high- 
 wall before sampling. 
 
Site-selection criteria were designed 
 to ensure that the predictive capabili- 
 ties of the overburden analysis methods 
 would be evaluated. Historical records, 
 provided by State regulatory agencies and 
 the coal companies , were used to elimi- 
 nate mine sites having significant net 
 acid or alkaline potential. It was felt 
 that any overburden analysis technique 
 can adequately predict an acid or alka- 
 line discharge when the carbonate or py- 
 rite is totally absent, respectively; the 
 target sites were those that are present- 
 ly difficult to assess. 
 
 In some cases, disagreement between 
 overburden analysis results at the time 
 of permitting and actual drainage qual- 
 ity was used to select a site; in other 
 cases, professional judgment had to be 
 used. Sites at which nonstandard prac- 
 tices might be the significant determi- 
 nant in postmining drainage quality were 
 avoided. These included backfills con- 
 taining acid drainage treatment sludge, 
 fly ash or preparation plant refuse, and 
 sites treated with ameliorative chemicals 
 other than agricultural limestone and 
 fertilizer. 
 
 Samples were subjected to laboratory 
 analyses. Acid-base accounting was used 
 on the samples from all 30 sites; weath- 
 ering tests published by Caruccio (2^) and 
 Sturey (8^) were performed on samples from 
 16 and 5 sites, respectively. Cold alka- 
 linity determinations were also made on 
 Caruccio weathering test samples. Table 
 1 illustrates other types of information 
 obtained from adjacent areas and avail- 
 able records added to the premining data 
 set. 
 
 The most critical postmining data in- 
 volved the quality of water issuing from 
 the reclaimed mine section. To charac- 
 terize the drainage, a field monitoring 
 program was instituted at each site to 
 measure the volume and quality of dis- 
 charges at least eight times during a 
 1-yr period. Analyses are indicated in 
 table 2. Where possible, data collected 
 by the mine operator were also used. 
 
 TABLE 1, - Premining equivalent data 
 collected to supplement overburden 
 sampling and analysis 
 
 Information 
 
 Local geology, 
 hydrology , 
 and mining 
 history. 
 
 Surface and 
 ground water 
 quality and 
 quantity. 
 
 Climatic data. 
 
 Sources 
 
 Permit applications. 
 State and Federal 
 agencies . 
 
 Current project, 
 historical records. 
 
 Government records, 
 mining company records, 
 
 TABLE 2. - Analyses performed on 
 premining and postmining water 
 samples 
 
 Field measurements Laboratory analyses 
 
 SUITE 1 
 
 pH... 
 Acidi 
 Disso 
 Speci 
 Tempe 
 
 ty, alkalinity. 
 Ived oxygen. . . . 
 fie conductance 
 rature 
 
 Iron 
 Sulfate 
 
 SUITE 2 
 
 
 pH 
 
 
 Iron 
 
 Acidity, alkalinity. 
 
 
 Sulfate 
 
 Dissolved oxygen.... 
 
 
 Calcium 
 
 Specific conductance 
 
 
 Magnesium 
 
 Temperature 
 
 
 Manganese 
 
 
 SUITE 3 
 
 
 pH 
 
 
 Iron 
 
 Acidity, alkalinity. 
 
 
 Sulfate 
 
 Dissolved oxygen.... 
 
 
 Calcium 
 
 Specific conductance 
 
 
 Magnesium 
 
 Temperature. ........ 
 
 
 Manganese 
 Aluminum 
 
 
 
 Table 3 summarizes ancillary postmining 
 data, collected primarily for use in 
 phase 2. 
 
 Phase 1 data collection is now com- 
 plete, and statistical analysis is in 
 progress. Table 4 summarizes the ranges 
 of values for acid-base account parame- 
 ters observed in samples from 30 sites. 
 The most acidic thickness-weighted value 
 for a single overburden column was a 
 
TABLE 3. - Supplementary postmining data 
 
 Mining Maps. 
 
 Drilling logs. 
 
 Mining method and equipment. 
 
 Materials handling. 
 
 Reclamation Backfilling plans and maps. 
 Materials handling. 
 Equipment. 
 
 Chronological records. 
 Topsoil storage. 
 Soil amendments. 
 Vegetation. 
 
 TABLE 4. - Ranges of values of acid-base 
 account parameters for individual 
 lithologies 
 
 TABLE 5. - Average leachate quality 
 range for Caruccio weathering 
 tests on individual lithologies 
 
 Parameter 
 
 Range 
 
 pH, paste.... 3.0- 7.9 
 
 Sulfur, pet: 
 
 Total <.05- 8.3 
 
 Pyritic <.05- 7.2 
 
 Sulfate <.05- .73 
 
 Organic <.05- .31 
 
 Neutralization potential 
 
 per 1,000 tons CaC03 -2.7 -940 
 
 deficiency of 1,300 tons as calcium car- 
 bonate. At the other extreme was a West 
 Virginia site having an excess alkalinity 
 of 1,200 tons as calcium carbonate. In- 
 terestingly, one of several toe-of -spoil 
 seeps at the latter site is acidic. 
 
 The Caruccio weathering test was per- 
 formed on overburden samples from 16 
 sites, having acid-base account results 
 Indicating overall neutrality (6 sites), 
 acidity (5 sites), or alkalinity (5 
 sites). Ranges of cumulative leachate 
 quality for individual lithologies are 
 shown in table 5. 
 
 The quality of surface runoff and spoil 
 seepage are the dependent variables for 
 statistical analysis in phase 1. Three 
 sets of primary independent variables 
 were derived from the three overburden 
 analysis methods used in the study. An- 
 cillary data (tables 1 and 3) will be 
 used in this analysis only as needed to 
 
 Parameter 
 
 Acidity, as CaC03 : 
 
 Hot total 
 
 Mineral 
 
 Alkalinity 
 
 Sulfate 
 
 Concentration 
 range, mg/L 
 
 <l-53,000 
 
 0-13,000 
 
 <1- 380 
 
 <l-33,000 
 
 classify sites in the case of bimodal 
 distributions. For example, sites may be 
 classified by degree of vegetative cover, 
 time since mining, or other factors. 
 Significant correlations between depen- 
 dent and independent variables will be 
 identified by simple linear regression 
 analysis, factor analysis, and multivari- 
 ate regression analysis. The phase 1 re- 
 search product will be a set of equations 
 that relate observed drainage quality to 
 overburden analysis data. 
 
 EMPIRICAL PREDICTIVE METHOD 
 
 Phase 2 of the contract research 
 will focus on the design and testing of a 
 method to predict postmining drainage 
 quality. This phase of the research will 
 make extensive use of the data developed 
 during phase 1. Our working hypothesis 
 is that the mining-related factors play 
 a critical role in the observed drainage 
 quality. Therefore, we expect that the 
 phase 1 results will indicate moderate 
 correlation coefficients between overbur- 
 den analysis data and postmining drainage 
 quality. The objective of phase 2, then, 
 is to improve the prediction by including 
 the nonnumerical (categorical) factors 
 shown in tables 1 and 3. 
 
 The output of the research may come in 
 different forms. For example, there may 
 be a nonlinear equation in which non- 
 numeric factors have been assigned nu- 
 merical rankings. Alternatively, the 
 predictive method may be based on an n- 
 dimensional decision surface, as in pat- 
 tern recognition. 
 
IN-HOUSE RESEARCH: PYRITE REACTIVITY ANALYSIS 
 
 BACKGROUND 
 
 Acid-base accounting methods depend 
 upon sulfur analysis to accurately quan- 
 tify the potential acidity of overburden 
 materials. Potential acidity is general- 
 ly considered to be a function of the 
 total sulfur content, although only py- 
 ritic sulfur contributes significantly 
 to acid production. Pyritic sulfur must 
 be distinguished from non-acid-producing 
 sulfur forms, in cases where (1) the ma- 
 terial is weathered and much of the orig- 
 inal pyritic sulfur has been oxidized to 
 sulfate sulfur or (2) the material is 
 carbonaceous and a significant proportion 
 of the total sulfur is bonded to organic 
 molecules . 
 
 Pyrite occurs as different forms or 
 morphologies in coal and overburden mate- 
 rials. Many authors have provided petro- 
 graphic descriptions of various pyrite 
 morphologies; descriptions given by King 
 (_3) have been adopted for this study be- 
 cause they are usable and encompass all 
 pyrite morphologies. According to King, 
 pyrite occurs in five basic morphologies: 
 (1) spherical aggregates of euhedral py- 
 rite crystals (framboids), (2) isolated 
 euhedral pyrite crystals, (3) nonspheri- 
 cal aggregates of euhedral pyrite crys- 
 tals, (4) irregularly shaped massive py- 
 rite, and (5) fracture-filling massive 
 pyrite. 
 
 Differences in pyrite reactivity corre- 
 lative to different pyrite morphologies 
 were first noted by Caruccio (J^) . Caruc- 
 cio indicated that framboidal pyrite was 
 the most reactive pyrite form and related 
 the percentage of framboidal pyrite to 
 the acid-producing potential of selected 
 samples. Most researchers agree that, in 
 general, the smaller the grain of pyrite, 
 the more reactive the pyrite and the 
 greater the potential acidity. 
 
 The importance of reactive pyrite forms 
 in the generation of acid mine drainage 
 was generally dismissed when it was real- 
 ized that these fonns could not account 
 
 for all of the acidity observed. Howev- 
 er, the oxidation of reactive forms under 
 ambient conditions may establish a chem- 
 ical environment that favors bacterial 
 catalysis and permits the less reactive 
 pyrite to react. Research into this pos- 
 sible triggering mechanism may identify 
 parameters that are inherently more accu- 
 rate in predicting potential acidity than 
 total sulfur or pyritic sulfur content 
 alone. Quantification of reactive pyrite 
 forms would not be able to predict total 
 acidity but would allow for more accurate 
 assessment of pyrite oxidation rates, and 
 thus possibly distinguish "go" or "no go" 
 acid generation situations. Initial Bu- 
 reau research into pyrite reactivity was 
 based on the hypothesis that different 
 pyrite morphologies and grain sizes would 
 thermally decompose and oxidize at dif- 
 ferent temperatures corresponding to the 
 relative stability of each form; more re- 
 active forms would be expected to react 
 at lower temperatures because of lower 
 activation energies. 
 
 Previous studies have used thermogravi- 
 metric (TG) and differential thermal 
 analysis techniques (DTA) to investigate 
 the thermal behavior of museum-grade py- 
 rite and marcasite. Warne (10) used DTA 
 thermograms to identify pyrite and mar- 
 casite in coal, carbonate, and clay ma- 
 trices. He found that minimum pyrite 
 concentrations of 0.5 to 1.0 pet could be 
 detected in a coal matrix by DTA despite 
 kaolinite and ankerite interferences. 
 However, DTA thermograms of pyrite and 
 marcasite were so similar that they could 
 not be differentiated. Luganov (4^) ob- 
 served the thermal behavior of pyrite 
 under inert atmospheres. They found that 
 DTA's of pyrite displayed an exothermic 
 effect at 380° C followed by endother- 
 mic effects at 480° to 500° C, 550° to 
 570° C. The exothermic effect at 380° C 
 was attributed to partial oxidation of 
 pyrite by oxygen adsorbed on the surface. 
 Subsequent endothermic effects at 480° 
 to 500° C and 550° to 570° C were thought 
 to represent the reaction of exothermic 
 products with pyrite. DTA's of pyrite 
 
treated with acid and ethanol to dissolve 
 ferric oxides and remove adsorbed oxygen 
 displayed only an endothermic effect at 
 680° C. 
 
 EXPERIMENTAL WORK 
 
 A modified evolved-gas analysis tech- 
 nique was used to examine the thermal be- 
 havior of sulfur species. This technique 
 employs a resistance furnace for the pro- 
 grammed heating of coal and overburden 
 samples in an oxygen atmosphere. The 
 evolution of sulfur dioxide and sulfur 
 trioxide gases was measured by an infra- 
 red detector and recorded simultaneously 
 with the sample temperature on a two- 
 channel recorder. A Leco SC-32 Sulfur 
 Analyzer^ was used to ignite samples and 
 monitor the evolution of sulfur oxides. 
 
 Initial tests of the evolved-gas analy- 
 sis technique were made to determine if 
 pyritic and sulfate sulfur could be 
 thermally distinguished. Figure 1 is a 
 thermogram of a sample containing 0.480 g 
 Fe2(S04)3*H20 and 0.020 g FeS2 (pyrite). 
 Results of this test indicate that pyrit- 
 ic sulfur and sulfate sulfur can be tem- 
 porally differentiated or time-resolved 
 at an isothermal furnace temperature of 
 767° C. Hydrated sulfate salts of cal- 
 cium, magnesium, manganese, and ferrous 
 
 iron were also tested; the sulfate inter- 
 ference in all cases constituted less 
 than 1 pet of the total pyrite response. 
 
 Standards of museum-grade pyrite were 
 prepared in a silica gel matrix and 
 tested at a furnace temperature of 
 767° C. Time-resolved, evolved-gas ther- 
 mograms of pyrite standards are shown in 
 figure 2. A plot of peak area for char- 
 acteristic pyrite peaks versus the con- 
 centration of pyrite standards (fig. 3) 
 yields a linear relationship. This indi- 
 cates that the Leco SC-32 Sulfur Analyzer 
 may be useful for quantification of py- 
 ritic and sulfate sulfur species in sam- 
 ples of low carbon content. The effect 
 of carbonaceous material on pyrite ther- 
 mograms is shown in figure 4. The large 
 exothermic effect resulting from the com- 
 bustion of carbonaceous materials effec- 
 tively masks all pyrite peaks. Interac- 
 tion of pyrite with the organic matrix 
 and pyrolysis products may result in the 
 shifting of characteristic pyrite peaks. 
 
 The relative reactivity of framboidal 
 and isolated euhedral pyrite morpholo- 
 gies was compared by preparing a 3-pct-S 
 
 ^Reference to specific products does 
 not imply endorsement by the Bureau of 
 Mines . 
 
 UJ 
 
 o 
 
 z 
 < 
 
 CD 
 
 o 
 
 CO 
 CD 
 
 < 
 
 UJ 
 
 > 
 
 _l 
 
 LJ 
 
 T 
 
 _Pyrite_ 
 response 
 
 H h 
 
 
 
 "1 1 
 
 Furnace temp- 767 **C 
 
 Sulfate response 
 
 4 6 
 
 TIME, min 
 
 FIGURE 1. - Time-resolved, evolved-gas analysis of a sample containing 0.20 g pyrite and 0.480 g 
 Fe2(S04)3.H20. 
 
1.0 
 
 LU 
 O 
 
 < 
 CD 
 
 cr 
 o 
 
 CO 
 CD 
 
 < 
 
 LU 
 
 > 
 
 < 
 
 -I 
 LU 
 
 .5 - 
 
 ' KEY ' 
 0.75 pet sulfur (pyrite) 
 0.50 pet sulfur (pyrite) 
 0.25 pet sulfur (pyrite) 
 Furnace temp= 767° C 
 
 I 2 
 
 TIME, min 
 
 FIGURE 2. - Time-resolved, evolved-gas analy- 
 sis of pyrite standords. 
 
 o 
 
 CL 
 
 LlI 
 
 cr 
 
 >■ 
 
 Q. 
 
 CO 
 < 
 
 a: 
 
 Z) 
 CO 
 
 2 3 
 
 PEAK AREA 
 
 FIGURE 3. - Plot of cumulative pyrite peck 
 area versus sulfur content. 
 
 
 
 
 1 1 1 1 I > 
 
 
 800 
 
 
 
 3 
 
 
 
 
 700 
 
 
 liJ 
 
 
 
 / \^ 
 
 
 
 
 
 13 
 
 
 
 / N. 
 
 
 
 
 Z 
 
 
 
 / N. 
 
 
 600 
 
 III 
 
 < 
 
 
 
 / \^ 
 
 
 
 rr 
 
 m 
 
 
 
 / ^N. 
 
 
 
 — ) 
 
 o 
 
 CO 
 
 ;d 
 
 
 / t<EY ^-- 
 
 
 500 
 
 
 III 
 
 
 III 
 
 < 
 
 
 _ 
 
 i'', / Absorbance 
 
 
 400 
 
 a. 
 
 LiJ 
 
 
 
 I 1 / r^^ Furnace temp^ 400° C 
 
 
 
 hi 
 
 > 
 
 
 
 \r ~"-. 
 
 
 300 
 
 H 
 
 _i 
 u 
 q: 
 
 1 
 
 \^ 
 
 1 -■--, 
 
 
 200 
 
 UJ 
 _l 
 CL 
 
 < 
 
 
 
 / 
 
 1 '^^^ 
 
 1 
 
 I 1 1 1 1 Y 
 
 
 100 
 
 en 
 
 
 
 1 
 
 
 
 
 
 
 
 1 2 3 4 5 6 
 
 7" 
 
 
 
 
 
 TIME, min 
 
 
 
 
 FIGURE 4. - Evolved-gas analysis of carbona- 
 ceous sample. 
 
 4.5 
 4.0 
 3.5 
 3.0 
 2.5 
 2.0 
 1.5 
 
 UJ 
 
 o 
 
 z 
 < 
 
 CD 
 CC 
 
 o 
 1) 
 
 CD 
 < 
 
 UJ 
 
 > 
 
 UJ 
 
 a: 
 
 
 1 1 1 1 
 KEY 
 
 1 
 
 
 _ 
 
 — Isolated euhedral pyrife 
 
 1 
 
 
 — 
 
 — Framboidal pyrife 
 
 : 
 
 
 — 
 
 — Heating curve 
 
 
 
 — 
 
 -- Exothermal effects 
 
 
 
 
 Furnace temp^ 538° C 
 
 £ 
 
 
 
 1 
 
 i 
 :: 
 
 
 
 i! 
 j! 
 i! 
 
 l^ 
 
 . 
 
 - 
 
 / i ! 
 / i 1 
 
 > • 
 
 - 
 
 
 i i 
 
 • 
 
 - 
 
 
 
 
 - 
 
 o 
 
 o 
 UJ 
 
 vc 
 q: 
 
 UJ 
 CL 
 
 s 
 
 UJ 
 
 600 
 500 
 400 
 300 
 200 
 100 
 
 12 3 4 5 6 7 
 
 TIME, min 
 
 FIGURE 5. - Evolved-gas analysis of isolated 
 euhedral and framboidal pyrite morphologies. 
 
 standard of each form in a silica gel ma- 
 trix (100- to 200-mesh). Framboidal py- 
 rite was supplied by Dr. Alfred Stiller 
 of West Virginia University, who con- 
 firmed its purity by Mossbauer spectro- 
 scopy. Framboidal and isolated euhedral 
 pyrite standards were run individually at 
 a furnace temperature of 538° C and at 
 purge and lance flows of 4 and 1 L/min, 
 respectively. The superimposition of the 
 two thermograms (fig. 5) illustrates the 
 
10 
 
 difference in thermal reactivity between 
 framboidal pyrite and the more stable 
 isolated euhedral pyrite. At this time, 
 no thermograms have been run on samples 
 containing both framboidal and isolated 
 euhedral pyrite. Therefore, the amount 
 of interaction between pyrite forms, if 
 any, and the characteristics of the re- 
 sulting thermogram cannot be predicted. 
 
 Although Bureau of Mines research into 
 pyrite reactivity is still in preliminary 
 stages, it can be concluded that — 
 
 1. Pyritic and sulfate sulfur in non- 
 carbonaceous materials can be differenti- 
 ated using a Leco SC-32 Sulfur Analyzer. 
 
 2. Pyritic sulfur can be quantitative- 
 ly determined in noncarbonaceous matrices 
 using evolved-gas analysis techniques. 
 
 3. Framboidal and isolated euhedral 
 pyrite morphologies differ significantly 
 in thermal reactivity. 
 
 4. Carbonaceous materials seriously 
 interfere with the evaluation of sulfur 
 species using evolved-gas analysis. 
 
 Future Bureau of Mines research 
 pyrite reactivity will include — 
 
 into 
 
 1. Investigations of the thermal reac- 
 tivity of other pyrite morphologies. 
 
 2. The evaluation of evolved-gas anal- 
 ysis as a quantitative technique for de- 
 termining pyritic and sulfate sulfur in 
 noncarbonaceous materials. 
 
 3. The development of a technique for 
 performing routine evolved-gas analy- 
 sis of sulfur species in carbonaceous 
 materials. 
 
 4. Correlation with contract research. 
 
 REFERENCES 
 
 1. Caruccio, F. T. The Quantifica- 
 tion of Reactive Pyrite by Grain Size 
 Distribution. Paper in Preprints, Third 
 Symposium on Coal Mine Drainage Re- 
 search, Pittsburgh, PA, 1970, pp. 123- 
 131. 
 
 2. Caruccio, F. T. , and G. Geidel. 
 Estimating the Minimum Acid Load That Can 
 Be Expected From a Coal Strip Mine. Pa- 
 per in Proceedings, 1981 Symposium on 
 Surface Mining Hydrology, Sedimentology , 
 and Reclamation, Lexington, KY, Dec. 7- 
 11, 1981, ed. by D. H. Graves. Univ. KY, 
 1981, pp. 117-122. 
 
 3. King, H. M. , and J. J. Renton. The 
 Mode of Occurrence and Distribution of 
 Sulfur in West Virginia Coals. Paper in 
 Carboniferous Coal Guidebook, WV Geol. 
 and Econ. Surv. Bull., v. 37, 1979, 
 pp. 278-301. 
 
 4. Luganov, V. A,, and V. I. Shabalin. 
 Behavior of Pyrite During Heating. Can. 
 Metall. Q. , v. 21, 1982, pp. 157-162. 
 
 5. Plummer, N. C, , and F. T. MacKen- 
 zie. Predicting Mineral Solubility From 
 Rate Data. Am. J. Sci., v. 274, 1974, 
 pp. 61-83. 
 
 6, Sobek, A. A., W. A. Schuller, J. R. 
 Freeman, and R. M. Smith. Field and Lab- 
 oratory Methods Applicable to Overburdens 
 and Minesoils. EPA 600/2-78-054, 1978, 
 203 pp. 
 
 7. Stumm, W. , 
 Aquatic Chemistry, 
 pp. 249-257. 
 
 and J. J. Morgan. 
 Wiley, 2d ed. , 1981, 
 
 8. Sturey, C. S., J. R. Freeman, J. W. 
 Sturm, and T. A. Keeney. Overburden 
 Analyses by Acid-Base Accounting and Sim- 
 ulated Weathering Studies as a Means of 
 Determining the Probable Hydrological 
 Consequences of Mining and Reclamation. 
 Paper in Proceedings, 1982 Symposium on 
 Surface Mining, Hydrology, Sedimentology, 
 and Reclamation, Lexington, KY, Dec. 6- 
 10, 1982, ed. by D. H. Graves. Univ. KY, 
 1982, pp. 163-179. 
 
11 
 
 9. U.S. Code of Federal Regulations. 
 Title 30 — Minerals Resources; Chapter 
 VII — Office of Surface Mining, Department 
 of the Interior; subchapter G — Surface 
 Coal Mining and Reclamation Operation 
 Permits; July 1, 1984. 
 
 10. Warne, S. S. J. Identification 
 and Evaluation of Materials in Coal by 
 
 Differential Thermal Analysis. 
 Fuel, May 1965, pp. 207-215. 
 
 J. Inst. 
 
 11. West Virginia Acid Mine Drainage 
 Task Force. Suggested Guidelines For 
 Method of Operation in Surface Mining of 
 Areas With Potentially Acid Producing 
 Materials. 1979, 20 pp. 
 
12 
 
 HYDROLOGIC ASPECTS OF ACID MINE DEIAINAGE CONTROL 
 By Kenneth J. Ladwig^ 
 
 INTRODUCTION 
 
 Water is obviously a principal com- 
 ponent of the acid mine drainage (AMD) 
 problem, functioning as a reactant in py- 
 rite oxidation, as a reaction medium, and 
 as a transport medium for oxidation prod- 
 ucts. The role of water as a transport 
 medium is the focus of one segment of the 
 Bureau of Mines AMD program. 
 
 Describing the contaminant transport 
 process serves two basic purposes. The 
 first is to develop site-specific charac- 
 terizations of the hydrology, Including 
 defining recharge areas and flow paths, 
 estimating rates and volumes of mine wa- 
 ter flow, delineating lateral variations 
 in water quality, and determining contam- 
 inant loads at the discharge. The site- 
 specific data are critical to the success 
 of any abatement procedure, regardless of 
 the technical approach chosen. Efficient 
 and cost-effective abatement requires 
 knowledge of sources of spoil water re- 
 charge, zones of acid production, and 
 movement of water through the acid-pro- 
 producing zones. 
 
 The second purpose is to examine in 
 greater detail the interaction between 
 acid production and hydrologic transport. 
 While field studies are by nature site 
 specific, data obtained from several 
 mines will be used to develop a more gen- 
 eralized conceptual understanding of the 
 transport process. The conceptual model 
 will then serve as the basis for Improved 
 reclamation and abatement technology. Of 
 central Importance in this phase of the 
 study are (1) the Interaction of the mine 
 water with the other components Involved 
 in acid generation and (2) the hydrochem- 
 ical evolution of the mine water. 
 
 ' Hydrologist, Pittsburgh Research Cen- 
 ter, Bureau of Mines, Pittsburgh, PA. 
 
 We Investigated the transport process 
 at both underground and surface coal 
 mines, with most of the underground mine 
 work being done in the northern anthra- 
 cite field of eastern Pennsylvania. The 
 purpose of this work is to describe the 
 hydrogeochemical processes occurring in a 
 flooded mine complex. The initial phase 
 of this work was reported in RI 8837 
 (4). 2 
 
 The surface mine work was done prin- 
 cipally at reclaimed surface mines in 
 Pennsylvania and West Virginia. Why re- 
 claimed sites? The fact that many re- 
 claimed mines in these States are still 
 producing considerable volumes of AMD at- 
 tests to the shortfalls of past and cur- 
 rent reclamation practices. By monitor- 
 ing these sites, we can examine what went 
 wrong, determine what steps might be 
 taken to deal with the current problem, 
 and develop methods for avoiding similar 
 problems in the future. 
 
 Described in the following sections are 
 results of a case study conducted at a 
 reclaimed surface mine in West Virginia 
 and a summary of the underground mine 
 study in eastern Pennsylvania. The em- 
 phasis is on developing a practical moni- 
 toring program and then intergrating the 
 site hydrology with the AMD abatement 
 plan. While it is unlikely that simple 
 hydrologic modification alone will elimi- 
 nate the problem, a thorough knowledge of 
 site-specific hydrology is fundamental to 
 the development and execution of a suc- 
 cessful abatement plan. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this report. 
 
13 
 
 SURFACE MINE CASE STUDY 
 
 SITE DESCRIPTION 
 
 METHODS 
 
 A small abandoned mine site in Upshur 
 County, WV , was monitored to evaluate the 
 use of bactericidal treatment to control 
 AMD. The Lower Kittanning seam was mined 
 from the U-shaped, 6-ha site in the late 
 1970's. Although the site was completely 
 revegetated, including the highwall, the 
 area was not regraded to approximate 
 original contour. Average spoil thick- 
 ness was about 7 m. Present topography 
 consists of a 12-m slope at the highwall, 
 a relatively flat bench over the mined 
 area, and a 12-m outslope leading to a 
 toe-of-spoil seep (fig. 1). 
 
 The methods used are standard proce- 
 dures for surface and ground water moni- 
 toring. Relative to perpetual water 
 treatment and AMD abatement costs, the 
 methods are not expensive, nor are they 
 technically complex. As will be illus- 
 trated, monitoring can yield valuable in- 
 formation on acid production and movement 
 at a surface mine site. Some type of 
 spoil water monitoring is highly recom- 
 mended prior to initiating abatement 
 plans. 
 
 Limit of disturbed area 
 
 Seal*, m 
 
 FIGURE 1. - Map of surface mine study site in Upshur County, WV, showing surface features, well 
 locations, and results of an electromagnetic induction survey. 
 
14 
 
 Following initial site reconnaissance 
 to locate all seepage points and describe 
 surface features , a series of electromag- 
 netic induction (EM) surveys were used to 
 describe subsurface features. EM can be 
 used at surface mines to help determine 
 spoil thickness and variations in thick- 
 ness across a site, and to locate wet 
 zones, mining relicts (highwalls, side- 
 walls, unmined blocks), mine floor struc- 
 tures, and zones of acid-producing mate- 
 rial (3). While they do not eliminate 
 the need for monitoring wells, EM surveys 
 help identify potential trouble areas. 
 Detailed surveying at the Upshur site 
 took just under 2 days to complete. 
 
 Monitoring wells were installed to de- 
 fine spoil water flow conditions and 
 spoil water quality. Spoil borings were 
 drilled to the mine underclay, and wells 
 were constructed from 2-in polyvinyl 
 chloride pipe, slotted along the lower 
 10 ft. The borings were backfilled and 
 the wells were sampled using standard 
 procedures (_5) . 
 
 Spoil samples were collected from sev- 
 eral depths during drilling of the moni- 
 toring wells. The samples were used to 
 characterize the distribution of materi- 
 als present on the site and to help re- 
 construct the backfilling sequence. All 
 samples were visually classified in the 
 field. Selected samples were subjected 
 to laboratory tests, including leaching 
 tests using the method described by Car- 
 uccio (2) . 
 
 Seepage discharge was monitored for 
 flow rate and water quality. Both sam- 
 pling and flow monitoring were done as 
 near to the point of seepage as possible 
 to minimize mixing with surface runoff. 
 As a compromise between cost, accuracy, 
 and maintenance requirements , flow gaging 
 was done with a simple V-notch weir con- 
 structed from plywood and stainless steel 
 il) • The weir was inexpensive and 
 
 reliable. 
 
 RESULTS 
 
 The EM surveys revealed an area of high 
 apparent conductivity (greater than 14 
 
 mmho/m) on the northwestern part of the 
 site (fig. 1). Progressively lower con- 
 ductivities were observed in the direc- 
 tion of the seep. Although the cause of 
 the high conductivity was not immediately 
 known, the area enclosed by the 14-mmho/m 
 contour on figure 1 was targeted as a 
 possible trouble spot. A more detailed 
 description of the geophysical survey is 
 given elsewhere (_3, site SMI). 
 
 Following the geophysical survey, a 
 series of spoil borings were drilled. 
 Spoil samples collected during drill- 
 ing showed the material in the area of 
 high conductivity (wells 4 and 6) con- 
 tained significant proportions of a fine- 
 grained, black material. In fact, the 
 entire thickness of spoil at well site 4 
 was comprised of the black material. 
 Holes drilled outside the high-conduc- 
 tivity zone (wells 1-3, 5, 7-9) contained 
 predominantly weathered sandstone. 
 
 The Kittanning coals in the study area 
 are "dirty" seams , and the black material 
 found at well sites 4 and 6 was believed 
 to be coal cleanings or shaly partings. 
 Laboratory tests on the spoil material 
 showed the mean sulfur content of the 
 black material (1.24 pet) was consider- 
 ably higher than that of the sandstone 
 spoil (0.12 pet). Samples of the under- 
 clay were also analyzed and found to have 
 a sulfur content similar to that of the 
 shaly material (1.20 pet). Of 29 spoil 
 samples analyzed, 6 had negative neutral- 
 ization potential (4 samples from wells 4 
 and 6, and 2 outslope samples). These 
 data again point to the area inside the 
 14-mmho/m contour in figure 1 as a pri- 
 mary trouble spot. 
 
 Final confirmation was provided by mon- 
 itor well water samples. The poorest wa- 
 ter quality on the site was found in well 
 4 (fig. 2). Mean sulfate and acidity 
 concentrations at well 4 were about twice 
 as high as the average concentrations for 
 the spoil and seep. Mean iron concentra- 
 tions at well 4 were more than twice the 
 mean spoil concentration and more than 
 six times the mean seep concentration. 
 
15 
 
 E 
 
 O 
 
 I- 
 < 
 CO 
 
 III 
 
 o 
 
 z 
 
 o 
 o 
 
 ^00 
 
 1.000- 
 
 800- 
 
 600- 
 
 400- 
 
 200- 
 
 
 • 
 • 
 
 9 
 
 O 
 
 9 
 
 Q. 
 
 • 
 
 I 
 
 SULFATE 
 
 ACIDITY 
 
 IRON 
 
 FIGURE 2. - Mean sulfate, acidity, and iron for well 7, well 4, the seep, and averaged for all of the 
 wells dri lied into spoi I. 
 
 Conversely, good-quality water was 
 found on the southern part of the site, 
 particularly near well 7 (fig. 2). The 
 well 7 area receives direct inflow of 
 highwall seepage, as well as infiltra- 
 tion recharge through inert sand soil. 
 As a result, there is much less contami- 
 nation evident. As this recharge con- 
 tinues to migrate through the spoil, the 
 water leaches some contaminants and mixes 
 with water of poorer quality prior to 
 discharge. 
 
 The water quality at the seep lies be- 
 tween that found in the well 4 area and 
 the well 7 area (fig. 2). Flow at the 
 
 seep is perennial and anomalouly high for 
 such a small site. Total seepage dis- 
 charge for the 1983 calendar year was 10 
 million gal, or about 50 pet of the total 
 precipitation for the same period. 
 
 Principally two factors contribute to 
 the high volume of discharge. One is the 
 uncontrolled highwall seepage into the 
 spoil on the southern part of the site. 
 The water level in well 7 is the highest 
 on the site at all times of the year, in- 
 dicating this is a perennial source of 
 recharge. The second factor is the ab- 
 sence of adequate surface water diver- 
 sions on top of the highwall and on the 
 
16 
 
 loining bench. Surface water from a small 
 recharge area above the site flows onto 
 the highwall and down a channel on the 
 highwall slope. Flows in the channel as 
 high as 15 gal/min have been observed 
 following a rainstorm. All of the chan- 
 nel flow infiltrates directly into the 
 spoil before reaching the bottom of the 
 slope. The mining bench itself is graded 
 back toward the highwall, further stimu- 
 lating ponding and infiltration at the 
 base of the highwall slope. 
 
 DISCUSSION OF RESULTS 
 
 The hydrologic study at the Upshur site 
 suggests at least two avenues for site 
 improvement. The first is to attempt to 
 abate acid production at the source. The 
 primary source of acid production at the 
 site appears to be relatively well de- 
 fined. Abatement procedures targeted 
 directly at the acid-producing area may 
 be the most cost-effective means of ob- 
 taining a significant reduction in seep 
 contamination. 
 
 Application of an organic compound is 
 currently being tested at the site to 
 inhibit AMD production. The bacteria- 
 inhibiting compound, potassium benzoate, 
 has been applied at the surface on the 
 northeastern part of the site. (The use 
 of organic compounds such as benzoate to 
 inhibit bacterial catalysis is described 
 elsewhere in these proceedings.) The 
 
 effect of the application is being moni- 
 tored in lysimeters and wells and at the 
 seep. 
 
 The second approach is a simple reduc- 
 tion in recharge to the site. For exam- 
 ple, subsurface drains to remove clean 
 highwall seepage prior to flow through 
 the spoil and minimal regarding to pro- 
 mote runoff rather than infiltration 
 would greatly decrease the total volume 
 of water discharged at the seep. Instal- 
 lation of these controls would reduce 
 mean flow by an estimated 50 to 75 pet 
 and very likely change the character of 
 the seep from perennial to intermittent. 
 Although contaminant concentrations at 
 the seep might increase following flow 
 reduction measures , we expect the reduced 
 volume would more than offset the in- 
 creased concentration, resulting in a net 
 decrease in contaminant load. 
 
 The case study presented here illus- 
 trates the use of relatively inexpensive 
 ground water monitoring for targeting AMD 
 abatement measures. Data obtained from 
 such studies are an integral part of the 
 Bureau of Mines research on improving ex- 
 isting abatement technology. As an end 
 product, this work, in conjunction with 
 research on overburden analysis, pyrite 
 reactivity, and spoil air, will be used 
 to develop predictive methods to avoid 
 the pitfalls associated with current min- 
 ing and reclamation practice. 
 
 UNDERGROUND MINES 
 
 To study the AMD problem at underground 
 mines, the Bureau initiated a field in- 
 vestigation of the mine water system in 
 the Wyoming Basin of the Northern Antra- 
 cite Field. The purpose of the study was 
 to evaluate the effect of mine flooding 
 on AMD formation. Specific project goals 
 included identification of sites where 
 pyrite oxidation may still be occurring 
 and mapping patterns of contaminant flow. 
 
 Between 1980 and 1982, nine abandoned 
 mine shafts were monitored for vertical 
 variations in the chemical composition of 
 the mine water system. Each shaft inter- 
 sected several coal seams. Monitoring 
 
 included the collection of shaft water 
 samples, downhole Eh and pH measurement, 
 fluid resistivity logging, spontaneous 
 potential logging, and fluid tempera- 
 ture logging. In addition to the shaft 
 logging, the four major outfalls in the 
 Wyoming Basin were monitored on a week- 
 ly basis from October 1982 through Sep- 
 tember 1983. These data were compared 
 with available historical data for the 
 outfalls. 
 
 Water quality at the outfalls in the 
 Wyoming Basin has exhibited marked im- 
 provement since inundation of the mine 
 complex. For example, between 1968 and 
 
17 
 
 1980 sulfate concentrations decreased by 
 49 pet at the Buttonwood Outfall (fig. 
 3). At all of the outfalls, pH has in- 
 creased to near neutral and net acidity 
 has decreased. 
 
 Weekly monitoring indicated water qual- 
 ity was similar at three of the four out- 
 falls (Buttonwood, South Wilkes Barre, 
 and Askam) , despite large differences in 
 respective recharge areas and predicted 
 residence times (table 1). The similar- 
 ity may reflect a long-term trend toward 
 uniformity coupled with the general im- 
 provement in water quality. The Nanti- 
 coke Outfall, which exhibits sulfate con- 
 centrations 25 to 35 pet higher than the 
 other three outfalls, discharges the 
 "youngest," or most recently formed, mine 
 pool. If a trend toward uniformity does 
 exist, the Nanticoke Outfall water qual- 
 ity may be expected to improve more rap- 
 idly than water quality at the other 
 outfalls. 
 
 TABLE 1. - Mean pH, sulfate, and flow 
 for the four outfalls in the Wyoming 
 Basin for the period October 1982 
 through September 1983 
 
 Outfall 
 
 pH 
 
 Sulfate, 
 mg/L 
 
 Flow, 
 gal/min 
 
 South Wilkes Barre 
 
 Buttonwood 
 
 Askam 
 
 5.9 
 5.9 
 5.9 
 6.0 
 
 1,200 
 1,020 
 1,130 
 1,640 
 
 25,380 
 5,690 
 5,650 
 
 Nanticoke 
 
 2,900 
 
 No significant seasonal trends in con- 
 taminant levels were observed, despite 
 order of magnitude variations in flow. 
 The absence of seasonal trends again im- 
 plies a uniform source. Thorough mixing 
 of the surface water recharge with the 
 bulk mine pool apparently occurs prior to 
 outfall discharge. 
 
 The shaft monitoring revealed marked 
 changes in water quality with depth with- 
 in the basin. In five of the nine shafts 
 studied, water was layered into two major 
 zones separated by sharp changes in Eh, 
 pH, and water quality parameters. An ex- 
 ample of the vertical change in pH and 
 sulfate is shown in figure 4. 
 
 The stratification appears to be re- 
 lated to discharge elevations at the time 
 of inundation, as well as to present 
 flow conditions. In each case, the sharp 
 change in water quality occurred just 
 above or below seams with mined barrier 
 pillars. Relative positions of mined 
 barrier pillars, outfall installations, 
 and natural structural features combine 
 to create an environment more favorable 
 to flushing in the shallower parts of the 
 mine system. As a result, the least con- 
 taminated water was found in the upper 
 zones of the system, while the poorest 
 quality was observed in flow-restricted, 
 deeper zones. 
 
 , 4,000 
 
 3,000 - 
 
 o 
 
 z 
 o 
 o 
 
 CO 
 
 2,000 
 
 1,000 - 
 
 s 
 
 1 
 
 ^ 
 
 - 
 
 ^ 
 
 t 
 
 s 
 
 1 
 
 KEY 
 • Mean 
 I Range 
 
 H-t- 
 
 1 
 
 
 1968 1970 
 
 1972 1974 1976 
 
 1978 
 
 1980 
 
 tr 
 
 FIGURE 3. - Mean and range of sulfate concen- 
 ations at the Buttonwood Outfall. 
 
 Surface, 183 m 
 600f— F"™ai 
 
 500- 
 
 400 
 
 300 
 
 y 200 
 
 u 
 
 100 
 
 -100 
 
 (602 ft) 
 Top of rock 
 Hillmcn 
 
 Diamond 
 
 Lance 
 
 Top Pitlston 
 Bottom Pittston 
 
 Ross 
 
 ^•Present bottom 
 
 Red Ash 
 
 KEY 
 
 « Allreodings 6/26/81 
 ■■■■ Coalbed not mined from shaft 
 ezzz2 Coalbed mined from shaft I 
 
 \ 
 
 150 
 
 100 e 
 
 50 LJ 
 
 800 IPOO 1,200 1,400 1,600 5 6 7 
 SULFATE CONCENTRATION, mq/L pH 
 
 FIGURE 4. - Vertical profile of pH and sulfate 
 in Gaylord shaft, Wyoming Basin. 
 
18 
 
 The improvement in water quality ap- 
 pears to indicate a decrease or cessation 
 of pyrite oxidation, along with neutrali- 
 zation and flushing of preexisting con- 
 taminants. The rate of flushing and min- 
 imum contamination levels attainable are 
 difficult to quantify, Pyrite oxidation 
 is still occurring at the surface in old 
 refuse piles and strip pits , and these 
 oxidation products are continuously 
 washed into the subsurface flow system. 
 The recharging pollutants are probably 
 confined to small, near-surface flow sys- 
 tems and may tend to control the minimum 
 contamination levels attained at the dis- 
 charge points. 
 
 In addition to the surface contami- 
 nants , the reservoir of oxidation prod- 
 ucts in the flooded mine complex will 
 continue to discharge for many years. 
 Stimulation of flow from the deep zones 
 by the addition of fully penetrating dis- 
 charge structures may increase the rate 
 of flushing but would aggravate the pol- 
 lutant load on the surface streams if 
 the discharge is left untreated. The 
 construction of additional outfalls would 
 also lower water levels, increasing the 
 unflooded volume of the mine complex and 
 possibly renewing pyrite oxidation in 
 these areas. 
 
 REFERENCES 
 
 1. Ackers, P., W. R, White, J. A. 
 Perkins, and A. J. M. Harrison. Weirs 
 and Flumes for Flow Measurement. Wiley, 
 1978, 327 pp. 
 
 2. Caruccio, F. T. , and G. Geidal. 
 Estimating the Minimum Acid Load That Can 
 Be Expected From a Coal Strip Mine. Pa- 
 per in Proceedings, 1981 Symposium on 
 Surface Mining, Sedimentology and Recla- 
 mation, Lexington, KY, Dec. 7-11, 1981, 
 ed. by D. H. Graves. Univ. KY, 1981, 
 pp. 117-122. 
 
 3. Ladwig, K, J. Use of Surface Geo- 
 physics To Determine Flow Patterns in 
 Surface Mine Spoil. Paper in Surface and 
 
 Borehole Geophysical Methods in Ground 
 Water Investigations (San Antonio, TX, 
 Feb. 6-9, 1984). National Water Well 
 Association, Worthington, OH, 1984, 
 pp. 455-471. 
 
 4. Ladwig, K. J., P. M, Erickson, 
 R. L, P. Kleinmann, and E. T. Posluszny. 
 Stratification in Water Quality in Inun- 
 dated Anthracite Mines, Eastern Pennsyl- 
 vania. BuMines RI 8837, 1984, 35 pp. 
 
 5. Scalf, M. R. , J. F. McNabb, W. J. 
 Dunlap, R. L. Cosby, and J. Fryberger. 
 Manual of Ground-Water Sampling Proce- 
 dures. Natl. Water Well Assoc, 1981, 
 93 pp. 
 
19 
 
 OXYGEN CONTENT OF UNSATURATED COAL MINE WASTE 
 By Patricia M, Ericksonl 
 
 INTRODUCTION 
 
 Acid mine drainage (AMD) results from 
 the oxidation of pyrite in the presence 
 of oxygen, water, and iron-oxidizing bac- 
 teria. Any of these three components 
 acting on the pyrite provides a potential 
 control point for reducing AMD formation. 
 
 The purpose of this project is to deter- 
 mine the oxygen availability in coal 
 refuse and spoil to improve our under- 
 standing of its potential to control acid 
 production. 
 
 BACKGROUND 
 
 The overall rate of acid production 
 is controlled by the rate-limiting step 
 in the chemical reactions of pyrite. 
 The rate dependence of pyrite oxidation 
 has been investigated in the labor.atory. 
 Under a variety of conditions near at- 
 mospheric pressure, the pyrite oxida- 
 tion rate was shown to depend on oxygen 
 partial pressure at values less than 
 2.0 m),2 10 (O, or 20 pet ( n_) . The 
 actual rate dependence under field con- 
 ditions is critical to the design of 
 abatement strategies. If field acid pro- 
 duction rates are a function of oxygen 
 availability at all partial pressures, 
 then even a limited reduction in atmos- 
 pheric diffusion into the pyritic mate- 
 rial will reduce the acid load. Alterna- 
 tively, if oxygen is rate limiting only 
 at low partial pressures, rigorous exclu- 
 sion of oxygen would be required to af- 
 fect acid production. 
 
 Few reports are available on the oxygen 
 status of coal refuse and spoil. Hons 
 measured pore gas composition as part of 
 a lignite waste revegetation study (8^). 
 Jaynes (9-10) monitored oxygen and carbon 
 dioxide within a backfilled surface coal 
 mine and developed an acid production 
 
 model. Other models have been presented 
 by Colvin (4_) and Brown (J^) • Further 
 work has been reported on metal mining 
 waste products and solution mining sites 
 (^-_3, 7^), Oxygen profiled tend to fall 
 into two categories. Compacted materials 
 tend to show decreased partial pressures 
 of oxygen with increasing depth. Oxygen 
 profiles of less compacted materials, 
 such as heap leaching systems and coarse 
 waste disposal sites, appear to be con- 
 sistent with air convection through ex- 
 posed faces. Actual field data on acid 
 production rates and pore gas composition 
 are necessary to calibrate available mod- 
 els or formulate new models and to eluci- 
 date the probable effects of proposed 
 acid abatement strategies. 
 
 To date. Bureau of Mines work has fo- 
 cused on characterizing gas composition 
 profiles in coal mine refuse and spoil. 
 Water quality data are also being col- 
 lected for investigation of possible cor- 
 relation between acid production and oxy- 
 gen availability. Preliminary findings 
 were reported earlier (5) , Only the oxy- 
 gen content of the pore gas is discussed 
 in this paper. 
 
 OXYGEN IN COAL REFUSE 
 
 Four inactive coal refuse disposal 
 areas, ranging in approximate age from 2 
 
 ^Supervisory physical scientist, Pitts- 
 burgh Research Center, Bureau of Mines, 
 Pittsburgh, PA. 
 
 to 12 yr, were Included in the study. 
 Soil gas probes ( 13) , installed to depths 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
20 
 
 of 15 to 90 cm, were sampled periodi- 
 cally. Samples were analyzed by gas 
 chromatography for atmospheric gases and 
 low-molecular-weight hydrocarbons . 
 
 Table 1 illustrates the range of oxygen 
 concentrations found in pore gas at the 
 four sites. Atmospheric oxygen levels 
 (~21 pet) were generally found only in 
 the uppermost 30 cm of the refuse. Val- 
 ues less than 2 pet were observed at 
 depths as shallow as 15 cm below the sur- 
 face. Based on the lowest literature 
 values (11), oxygen concentrations were 
 sufficiently low at times to restrict py- 
 rite oxidation. 
 
 Long-term monitoring was conducted only 
 at one site (Morgan County, OH). Figure 
 1 illustrates the average pore gas oxygen 
 
 profile generated from 12 sets of gas 
 samples. This profile appears to be con- 
 sistent with a model of the system based 
 on oxygen diffusion from the atmosphere 
 and oxygen consumption within the refuse 
 by pyrite oxidation. Near-surface pyrit- 
 ic material is undergoing active oxida- 
 tion, evidenced by runoff acidity in the 
 range of 10,000 to 20,000 mg/L. 
 
 Gas composition in the refuse showed 
 a strong seasonal dependence. Figure 2 
 shows average oxygen profiles for samples 
 taken in the summer and winter seasons. 
 There was very little overlap between the 
 data sets. During the summer, the oxygen 
 profile was sharp and showed the greatest 
 change within the uppermost 15 cm of ref- 
 use. Oxygen was found at significant 
 
 TABLE 1. - Range of pore gas content in coarse coal refuse 
 
 
 Oxygen, moisture-free vol pet 
 
 Depth, cm 
 
 Allegheny County, PA 
 
 Wise County, VA 
 
 Morgan County, OH 
 
 
 Site A 
 
 Site B 
 
 
 15 
 
 20 
 
 20.4-20.7 
 
 18.9-19.9 
 
 ND 
 
 6.5-14.9 
 
 ND 
 
 ND 
 
 ND 
 
 20.0-20.4 
 
 ND 
 
 5.7-17.2 
 
 .4- 3.5 
 
 ND 
 
 20.2-20.8 
 
 ND 
 
 9.1-20.7 
 
 ND 
 
 ND 
 
 .2- 9.6 
 
 0.3-21.8 
 ND 
 
 30 
 
 .1-20.8 
 
 35 
 
 ND 
 
 66 
 
 ND 
 
 90 
 
 ND 
 
 OXYGEN, av pet 
 10 15 
 
 20 
 
 25 
 
 FIGURE 1. - O2 for four sets of gas probes 
 at the Morgan County, OH, site. 
 
 E 
 o 
 
 I- 
 
 Q. 
 UJ 
 Q 
 
 OXYGEN, pet 
 
 
 
 
 5 
 
 10 15 
 
 20 
 
 25 
 
 20 
 
 1 
 
 1 1 
 
 ^ 
 
 - 
 
 40 
 
 - 
 
 
 KEY \ 
 
 
 - 
 
 60 
 
 
 
 Summer \ 
 Winter \ 
 
 
 
 80 
 
 < 
 
 \ 1 
 
 , \ 
 
 1 
 
 
 FIGURE 2. - Refuse was more oxygenated at all 
 probe depths in winter (December through February) 
 than in summer (June through August). 
 
21 
 
 concentrations at greater depths in the 
 winter months. One would expect that the 
 oxidation reactions occur at lower rates 
 during the cold season and that the 
 
 decreased oxygen consumption explains the 
 higher oxygen content of the pore gas at 
 depths up to 80 cm. 
 
 SPOIL OXYGEN CONTENT 
 
 We are currently monitoring pore gas 
 composition in spoil at three regraded 
 surface mines. Multiport gas-sampling 
 wells (5) were installed at two or more 
 locations on each site to monitor pore 
 gas composition in the unsaturated zone. 
 The oxygen profiles were distinctly dif- 
 ferent from those in coal refuse at the 
 Morgan County site and did not show site- 
 to-site consistency. Oxygen concentra- 
 tions of 5 to 20 pet occurred to depths 
 of several meters and, in some cases, 
 throughout the unsaturated zones. 
 
 Figure 3 shows the average oxygen pro- 
 file obtained from seven gas-sampling 
 wells at an unvegetated site in Clarion 
 County, PA. Overall, the oxygen content 
 of the gas decreased with depth. Howev- 
 er, the zones of greatest change in oxy- 
 gen content differed for individual wells 
 on the 11-acre site (fig. 4). Seasonal 
 trends have not been examined yet. 
 
 Figure 5 shows the average oxygen pro- 
 files for three wells at a recently re- 
 vegetated site in Upshur County, WV. 
 This isolated ridge was reclaimed accord- 
 ing to state-of-the-art guidelines, in- 
 cluding the selective placement of toxic 
 spoil above a nonreactive base pad and 
 below a compacted clay cap (6^) . Gas- 
 sampling wells were placed parallel to 
 the ridge axis at approximately equal 
 spacing. Sampling ports were located as 
 follows: in the soil, in the top, mid- 
 dle, and bottom of the acidic spoil zone, 
 and in the base pad. Two of the wells 
 showed decreasing oxygen with increasing 
 depth, while the third well showed a peak 
 oxygen content in the middle of the 
 acidic spoil zone (fig. 5). The greatest 
 change in oxygen for a given change in 
 depth occurred in the acidic spoil zone 
 of wells 847 and 846 and between the soil 
 and acidic spoil zones for well 793. The 
 high oxygen values at the 5.9-m depth in 
 
 OXYGEN, av pet 
 10 15 
 
 OXYGEN, avpct 
 
 20 
 
 25 
 
 FIGURE 3. - Oj profile from seven gas-sampling 
 wells at the Clarion County, PA, site. 
 
 2- 
 
 CL 
 LlI 
 Q 
 
 6- 
 
 8 
 
 
 5 
 
 10 15 20 
 
 25 
 
 - 
 
 I 
 
 c 
 
 1 1 1/ 
 y^ KEY 
 
 - 
 
 c/ 
 
 1 ^ 
 
 '^ o Well 1 
 o Well 11 
 
 1 1 1 
 
 
 FIGURE 4. - Oj profiles for two of seven gos- 
 sampling wells at the Clarion County, PA, site. 
 The change in Oj pet varied for the same depth 
 interval. 
 
22 
 
 
 
 X 
 
 I- 
 
 Q. 
 UJ 
 Q 
 
 10 
 
 15 
 
 OXYGEN, av pet 
 5 10 15 20 
 
 25 
 
 KEY 
 
 Well 847 
 Well 793 
 Well 846 
 
 FIGURE 5. - Oj profiles for three gas-sampling 
 wells at the reclamation study site in Upshur 
 County, WV. 
 
 OXYGEN, av pet 
 10 15 
 
 20 25 
 
 FIGURE 7. - Oj profiles in revegetated spoil 
 at the abandoned site in Upshur County, WV. 
 
 ASONDJ FMAMJ J 
 1983 1984 
 
 TIME, months 
 
 FIGURE 6. - Variation in oxygen content 2 ft 
 below the compacted clay layer at the reclaimed 
 study site in Upshur County, WV. 
 
 well 847 may be due to the location of 
 the well on the exposed end of the mined 
 ridge. Well 847 is surrounded by steep 
 slopes on three sides as compared to 
 slopes on two sides for the other wells. 
 
 The clay cap was placed over the acidic 
 spoil to minimize rainfall infiltration 
 (6^). We also thought it might act as 
 a diffusion barrier. Figure 6 summa- 
 rizes the preliminary data available from 
 the air well ports located 0.6 m beneath 
 the clay cap. There appeared to be a 
 distinct seasonal increase in oxygen dur- 
 ing the spring and summer. The changing 
 oxygen levels indicate that the clay lay- 
 er is not a good diffusion barrier. 
 
 The third site, also located in Upshur 
 County, WV, is an abandoned, revegetated 
 surface mine. A three-port gas well (No. 
 8) was installed adjacent to buried py- 
 ritic material, which has been identified 
 as the major source of acid drainage on 
 the site. A four-port well (No. 5) was 
 placed in the outslope area, which is 
 composed of rocky spoil. Figure 7 shows 
 the average profiles for both wells. 
 Well 8 data indicated a steep decrease in 
 oxygen content within the uppermost 1.8 m 
 of spoil. Well 5 showed a different type 
 of oxygen profile, not consistent with 
 vertical downward diffusion. The loca- 
 tion of well 5 on the outslope may ac- 
 count for this observation; diffusion 
 
and/or convection may occur along the 
 outslope face or toe. Similar profiles 
 
 23 
 
 have been observed in coarse mining waste 
 rock (7). 
 
 DISCUSSION 
 
 Pore gas oxygen content in coal refuse 
 generally decreased to only a few percent 
 within 1 ra below the surface at four un- 
 vegetated sites. According to most lit- 
 erature reports, a lower limit of 1 or 2 
 pet oxygen could be used to define the 
 pyrite-oxidizing zone. In that case, the 
 bulk refuse should not be contributing 
 much to the acid load at these sites. 
 However, in the winter months oxygen lev- 
 els greater than 15 pet were observed at 
 greater depths. In the absence of con- 
 sumption in the shallow zone, oxygen 
 would be available to oxidize pyrite at 
 greater depths. Cover with a nonpyritic 
 coal refuse would probably not reduce the 
 acid load. Coal mine spoil would proba- 
 bly be an ineffective cover material for 
 the same reason: At the three sites we 
 studied, the spoil pore gas usually con- 
 tained sufficient oxygen to support py- 
 rite oxidation. 
 
 Coal mine spoil oxygen profiles showed 
 great variety among sites and laterally 
 on a single site. In 10 of 12 wells at 
 3 sites, oxygen usually decreased with 
 depth. These profiles are consistent 
 with oxygen diffusion from the atmosphere 
 downward through the spoil. The notable 
 exceptions were the outslope area at the 
 abandoned site and the exposed end of the 
 ridge at the recently reclaimed site. 
 Profiles from these two areas were simi- 
 lar to profiles observed in coarse waste 
 subject to air convection through exposed 
 slopes ( 7_) . 
 
 Gas composition monitoring can provide 
 useful information about the location of 
 pyrite oxidation zones. The steep gradi- 
 ent observed in the summer in coal refuse 
 apparently is indicative of a zone of ac- 
 tive oxidation. Similar zones in mine 
 spoil, in the absence of a change in dif- 
 fusion coefficient with depth, may also 
 be indicators of pyrite oxidation. For 
 
 example, well 8 at the abandoned West 
 Virginia site showed a steep gradient and 
 is known to be adjacent to a mass of 
 buried pyritic material currently produc- 
 ing acid. Identification of acid source 
 areas will allow application of remedial 
 treatments to selected zones, thereby re- 
 ducing cost. 
 
 The seasonal trends in oxygen profiles 
 are not consistent. We observed peaks in 
 oxygen concentrations during the winter 
 in coal refuse and during the summer in 
 one spoil site. The winter peaks suggest 
 that the refuse is more oxygenated when 
 chemical activity decreases due to lower 
 temperatures. We do not know why peak 
 oxygen levels were observed in the summer 
 at the spoil site. 
 
 The results reported in this paper sug- 
 gest that inert cover materials may not 
 be useful as diffusion barriers to reduce 
 pyrite oxidation. Covering the pyritic 
 refuse or spoil with an oxygen-consuming 
 layer is probably a better control strat- 
 egy. Vegetation, soil containing an ac- 
 tive microbial population, and decaying 
 organic matter are candidate cover mate- 
 rials. We are planning to conduct tests 
 this year to evaluate the effects of veg- 
 etation and mulch on pore gas profiles in 
 coal refuse. Previously reported revege- 
 tation studies have generally neglected 
 measuring oxygen availability; instead, 
 the plant growth and water quality were 
 usually monitored. Measurements of gas 
 diffusion rates are also needed to deter- 
 mine the flux of oxygen through the waste 
 materials. 
 
 Future work will also include appli- 
 cation of available computer models to 
 fit the field data. The best-fit model 
 will then be used to assess the proba- 
 ble impacts of proposed acid abatement 
 techniques , 
 
24 
 
 REFERENCES 
 
 1. Brown, W. E. The Control of Acid 
 Mine Drainage Using an Oxygen Diffusion 
 Barrier. M.S. Thesis, OH State Univ., 
 1970, 86 pp. 
 
 2. Cathles , L. M. Predictive Capabil- 
 ities of a Finite Difference Model of 
 Copper Leaching in Low Grade Industrial 
 Sulfide Waste Dumps. Math. Geol, , v. 11, 
 1979, pp. 175-191. 
 
 3. Cathles, L. M. , and J. A. Apps. A 
 Model of the Dump Leaching Process That 
 Incorporates Oxygen Balance, Heat Balance 
 and Air Convection. Metall. Trans. B, v. 
 6B, 1975, pp. 617-624. 
 
 4. Colvin, S. L. Oxygen Diffusion in 
 Strip-Mined Soils. M.S. Thesis, lA State 
 Univ., Ames, lA, 1977, 72 pp. 
 
 5. Erickson, P. M. , R. L. P. Klein- 
 mann, W. R. Homeister, and R. C. Briggs. 
 Reclamation and the Control of Acid Mine 
 Drainage. Paper in Proceedings, Sympo- 
 sium on the Reclamation, Treatment and 
 Utilisation of Coal Mining Wastes. Na- 
 tional Coal Board (United Kingdom), 1984, 
 pp. 30.1-30.18. 
 
 6. Geidel, G. , and F. T. Caruccio. A 
 Field Evaluation of the Selective Place- 
 ment of Acidic Material Within the Back- 
 fill of a Reclaimed Coal Mine. Paper in 
 Proceedings, 1984 Symposium on Surface 
 Mining, Hydrology, Sedimentology , and 
 Reclamation, Lexington, KY, Dec. 2-7, 
 1984, ed. by D. H, Graves. Univ. KY, 
 1984, pp. 127-131. 
 
 7. Harries, J. R. , and I. A. M. Rit- 
 chie. The Effect of Rehabilitation on 
 
 the Oxygen Concentrations in Waste Rock 
 Dumps Containing Pyritic Material. Paper 
 in Proceedings, 1984 Symposium on Surface 
 Mining, Hydrology, Sedimentology, and 
 Reclamation, Lexington, KY, Dec. 2-7, 
 1984, ed. by D. H. Graves. Univ. KY, 
 1984, pp. 463-466. 
 
 8. Hons , F. M. Chemical and Physical 
 Properties of Lignite Spoil Materials and 
 Their Influence Upon Successful Reclama- 
 tion. Ph.D. Thesis, TX A&M Univ. , 1978, 
 137 pp. 
 
 9. Jaynes , D. B. , A. S. Rogowski , and 
 H. B. Pionke. Acid Mine Drainage From 
 Reclaimed Coal Strip Mines. I. Model 
 Description. Water Resour. Res., v. 20, 
 1984, pp. 233-242. 
 
 10. Jaynes, D. B. , A. S. Rogowski, 
 H. B. Pionke, and E. L. Jacoby, Jr. At- 
 mosphere and Temperature Changes Within 
 Reclaimed Coal Strip Mine. Soil Sci., v. 
 136, 1983, pp. 164-177. 
 
 11. Morth, A. H. , E. E. Smith, and 
 K. S. Shumate. Pyritic Systems: A Math- 
 ematical Model. EPA-R2-72-002, 1972, 
 169 pp. 
 
 12. NUS Corporation. The Effects of 
 Various Gas Atmospheres on the Oxidation 
 of Coal Mine Pyrites. EPA 14010 ECC, 
 1971, 140 pp. 
 
 13. Staley, T. E. A Point-Source 
 Method for Sampling Soil Atmospheres. 
 Trans. Am. Soc. Agri. Eng. , v. 23, 1980, 
 pp. 578-580, 584. 
 
25 
 
 CONTROL OF ACID MINE DRAINAGE BY APPLICATION OF BACTERICIDAL MATERIALS 
 
 By Patricia M. Erickson,'' Robert L. P. Kleinmann,2 
 and Steven J, Onysko-^ 
 
 INTRODUCTION 
 
 The kinetics of acid formation are de- 
 pendent on the availability of oxygen, 
 the surface area of pyrite exposed, the 
 activity of iron-oxidizing bacteria, and 
 the chemical characteristics of the in- 
 fluent water. The principal iron-oxidiz- 
 ing bacterium involved in accelerating 
 pyrite oxidation is Thiobacillus f erro- 
 oxidans (9^, _1_5).^ The Bureau of Mines 
 has previously reported the results of 
 
 full-scale field tests that showed how 
 anionic surfactants (cleansing deter- 
 gents) can be used to reduce the activity 
 of T. ferrooxidans (12-13) and thereby 
 abate acid formation. After a brief dis- 
 cussion of the literature, this paper 
 will review the surfactant solution tech- 
 nique and report progress on two alterna- 
 tive procedures. 
 
 ACKNOWLEDGMENTS 
 
 The controlled release surfactant 
 formulations were provided by BFGood- 
 rich and Granger Technologies, Inc. The 
 
 assistance of both companies is grate- 
 fully acknowledged. 
 
 BACKGROUND INFORMATION 
 
 The possible involvement of bacteria in 
 the formation of acid drainage was first 
 reported in 1919 by Parr and Powell, who 
 determined that coal inoculated with an 
 unsterilized ferrous sulfate solution 
 produced drainage with higher concen- 
 trations of sulfate than did sterile 
 controls (J^) . The possibility of reduc- 
 ing acid drainage by bacterial inhibition 
 was first considered in 1953 but was re- 
 jected as impractical due to probable 
 rapid repopulation (14) . Later labora- 
 tory studies demonstrated the vulnerabil- 
 ity of T^ ferrooxidans in coal and coal 
 refuse to anionic surfactants and conse- 
 quent acidity reductions of 65 to 80 pet 
 (7). 
 
 Supervisory physical scientist. 
 
 ^Research supervisor. 
 
 ^Civil engineer. 
 Pittsburgh Research Center, Bureau of 
 Mines, Pittsburgh, PA. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 Full-scale tests at active and inactive 
 coal refuse areas demonstrated that sodi- 
 um lauryl sulfate (SLS) surfactant appli- 
 cation could effectively reduce acid pro- 
 duction and thereby lower water treatment 
 costs. Sufficient surfactant was applied 
 by hydroseeder to the coal refuse to sat- 
 urate the adsorptive capacity of the top 
 1 ft of refuse, based on a laboratory de- 
 termination (13). The 1-ft-thick treat- 
 ment zone was selected for several rea- 
 sons: (1) Oxidation was assumed to occur 
 largely in a near-surface oxygenated zone 
 (3, 6), (2) desorption and downward mi- 
 gration would result in treatment at 
 greater depth, and (3) it was preferred 
 to undertreat rather than overtreat, to 
 prevent significant surfactant concentra- 
 tions off the site. 
 
 Water quality improved at the test 
 sites in 1 to 3 months. Acidity, sul- 
 fate, and manganese decreased 60 to 90 
 pet; iron decreased 90 to 95 pet (fig. 
 1). After about 4 months, contaminant 
 concentrations slowly climbed back to 
 
26 
 
 E 
 
 o 
 
 01 
 
 UJ 
 
 o 
 
 5,000 
 
 4,000 
 
 3,000 
 
 2,000 
 
 8 1,000 
 
 
 
 KEY 
 
 -X Acidity 
 
 A A Sulfate 
 
 o o Iron 
 
 
 .-A 
 — X 
 
 ^t=^-^» ^^-:---.v-4>-K>-^i-o-<> 
 
 40 60 80 100 120 140 
 
 TIME AFTER TREATMENT, days 
 
 160 180 
 
 FIGURE 1. - Improvement in drainage quality following surfactant solution application at a site in 
 West Virginia. 
 
 previous levels. Effluent surfactant 
 concentrations were negligible. 
 
 As a result of these tests, the mining 
 industry has begun to apply surfactants 
 to coal refuse, coal stockpile areas, un- 
 reclaimed mine spoil, and waste sulfide 
 rock, with mixed results. One coal com- 
 pany that applied an anionic surfactant 
 two or three times a year to a developing 
 coal refuse pile has had no acid problem 
 over a 5-yr period despite the fact that 
 coal refuse at the plant typically pro- 
 duces acidic drainage within 6 months. 
 At the other extreme are sites where the 
 technique produced no apparent effect or 
 only a short-term improvement in water 
 quality. Some of these failures can be 
 explained simply, such as when the dosage 
 rate or site conditions were obviously 
 inappropriate. At other sites it may 
 never be known why the technique failed 
 to reduce acid production. 
 
 A previous report describes in some de- 
 tail when and how the surfactant should 
 
 be applied (13) . It is worthwhile to re- 
 state the three most significant points: 
 
 1. Determine beforehand if the tech- 
 nique is potentially cost effective for 
 the site. Assume a material cost of 
 $600/ acre annually plus the cost of three 
 applications by watering truck or hydro- 
 seeder, a 60-pct decrease in neutraliza- 
 tion costs, and a 90-pct decrease in 
 sludge accumulation; if the calculated 
 annual savings are not significantly 
 greater than the assumed costs , the tech- 
 nique is probably not appropriate. 
 
 2. The surfactant must reach and ad- 
 sorb to the pyritic material. If the 
 site is covered with topsoil, a surfac- 
 tant application will not reach the py- 
 ritic material and will therefore accom- 
 plish nothing. If an adsorption test 
 indicates that the pyritic material has 
 low adsorptive capacity, the surfactant 
 will wash away rapidly, providing only 
 brief abatement. 
 
27 
 
 3. Owing to slow hydrologic flow- 
 through time or pooled acid water on the 
 old mine floor or in a refuse area, the 
 effect of surfactant treatment may be de- 
 layed, masked, or made insignificant. In 
 the case of slow flow-through time (as 
 much as a year at some sites), improve- 
 ment in water quality at the discharge 
 point cannot occur faster than water 
 flows through the material. If a signif- 
 icant pool of acid water exists, years of 
 continued application of surfactant could 
 be required before an increase in water 
 quality is observed, unless the acid pool 
 is first neutralized or drained. 
 
 Application of anionic surfactant solu- 
 tion, although effective in reducing wa- 
 ter treatment costs, cannot be regarded 
 as a long-term control measure. Two mod- 
 ifications of the basic approach are be- 
 ing considered by the Bureau of Mines: 
 
 1. The surfactant can be rendered less 
 soluble. This has been accomplished us- 
 ing slow-release technology developed for 
 more conventional biocides (2^, 10) . Con- 
 trolled release of surfactant over a pe- 
 riod of many years may be possible. 
 
 2. Other environmentally safe chemi- 
 cals have been identified that inhibit T. 
 f errooxidans and that react with acid 
 
 to form slightly soluble 
 
 Thus, these chemicals may 
 
 material in 
 
 mine drainage 
 
 precipitates . 
 
 form their own slow-release 
 
 the acid-producing environment . 
 
 The remainder of this paper will summa- 
 rize the results of laboratory and pilot- 
 scale experiments and introduce full- 
 scale field tests that are in progress. 
 
 SLOW RELEASE OF SURFACTANTS 
 
 This approach has been under investiga- 
 tion since surfactants were first con- 
 sidered for field use (7). Early surfac- 
 tant-rubber formulations reduced acid 
 formation by over 95 pet in a pilot-scale 
 field test but were effective for less 
 than 1 yr (8^) . Subsequent research has 
 been directed towards extending the re- 
 lease lifetime of the material and field 
 tests of the resultant formulations. 
 
 LABORATORY TESTS 
 
 and surface area, 
 dissolution. 
 
 influenced the rate of 
 
 Laboratory and pilo 
 been conducted on more 
 manufactured for the 
 and Granger Technologi 
 manufactured prior to 
 SLS as the active ingr 
 sitions are propriet 
 are referred to in thi 
 betic code. 
 
 t-scale tests have 
 than 20 materials 
 tests by BFGoodrich 
 es . The materials, 
 1982, all contained 
 edient. The compo- 
 ary , and materials 
 s report by alpha- 
 
 Laboratory tests were conducted ini- 
 tially to determine which variables most 
 strongly influenced the SLS release rates 
 (_H_) . Every parameter investigated, in- 
 cluding nature of matrix, SLS loading, 
 
 Figure 2 shows release curves for five 
 formulations for illustrative purpose. 
 The data were obtained by periodically 
 rinsing a 5-g sample with 100 mL deion- 
 ized water. Leachates were combined to 
 400 to 500 mL total volumes and analyzed 
 for anionic surfactants by the methylene 
 blue method (J_) . The percent SLS remain- 
 ing in the matrix was calculated from the 
 nominal SLS content of the sample and the 
 cumulative mass of SLS extracted. Nomi- 
 nal SLS contents, ranging from 20 to 65 
 pet of the total sample weight, were nor- 
 malized to 100 pet for comparison. 
 
 All formulations exhibited an initial 
 rapid release of detergent followed by 
 a slower dissolution phase. The first 
 phase was more pronounced in samples hav- 
 ing a larger fraction of SLS at or near 
 the pellet surface. For example, samples 
 D and E are cylindrical pellets of the 
 same formulation having diameters of 4.6 
 and 3.2 mm, respectively. Approximately 
 three times more detergent was dissolved 
 
28 
 
 < 
 
 Q 
 CO 
 UJ 
 
 cr 
 
 60 
 
 KEY 
 
 X Formulation A 
 D Formulation B 
 ^ Formulation C 
 O Formulation D 
 A Formulation E 
 
 
 
 I 2 3 
 
 CUMULATIVE LEACH VOLUME, L 
 
 FIGURE 2. - SLS release curves for five control led 
 release formulations subjected to intermittent leach 
 ing in the laboratory. 
 
 from sample E than from sample D in the 
 first liter of leach water (fig. 2), 
 
 Under the experimental conditions 4,000 
 mL is approximately equivalent to 40 in 
 of precipitation. Cumulative extracted 
 SLS at this point ranged from 4 to 38 pet 
 of the total surfactant content of the 
 samples. These values cannot be extrapo- 
 lated to an expected lifetime, however, 
 because release rates were decreasing 
 over time. In some formulations much of 
 the detergent appeared to be unavailable 
 (curves D and E of figure 2). 
 
 While the laboratory results confirmed 
 that surfactant loading, pellet geometry, 
 and matrix type affected SLS release 
 rates, no empirical equations could be 
 developed to predict release curves for 
 new formulations. Outdoor evaluation of 
 potential materials was considered pre- 
 ferable to continued laboratory testing. 
 
 PILOT- SCALE TESTS 
 
 Pilot-scale testing was conducted out- 
 doors to determine release rates under 
 field conditions. The test area con- 
 sisted of two small coal refuse piles, 
 
 each about 7 ft wide, 12 ft long, and 1.5 
 ft high at the lengthwise crest. Garden 
 edging was used to divide each slope into 
 six test plots. A rain gage was placed 
 about 15 ft from the refuse piles. 
 
 Approximately 250 to 500 g of pellets 
 were spread by hand on each of 21 test 
 plots during February 1983. The coal 
 refuse contained about 5 pet sulfur and 
 produced drainage acidity on the order 
 of 10^ mg/L prior to the controlled re- 
 lease application. No attempts were made 
 to monitor drainage quality during the 
 experiment. 
 
 Periodically, a selected number of pel- 
 lets were removed at random from each 
 plot and residual SLS content was deter- 
 mined. In one method, the samples were 
 dried to constant weight at room tempera- 
 ture, and SLS release was calculated by 
 weight loss: 
 
 SLS release = nominal weight 
 
 - actual weight 
 
 This method is based on the assumption 
 that all weight loss resulted from SLS 
 dissolution. Nominal weights were deter- 
 mined as the mean weight of 10 repli- 
 cate samples of fresh pellets of the same 
 formulation. 
 
 The second method involved aqueous ex- 
 traction of residual SLS from air-dried 
 samples. The pellets were placed in a 
 minimum of 500 mL deionized water and al- 
 lowed to equilibrate for several days. 
 The extracts were analyzed for anionic 
 surfactants, and the extracted pellets 
 were air-dried for determination of ma- 
 trix weight. This method was based on 
 the assumption that all residual SLS 
 could be extracted into deionized water. 
 Values were calculated from actual dry 
 matrix weight and nominal dry matrix 
 weight. A typical release curve is shown 
 in figure 3 for a formulation nominally 
 containing 50 pet SLS by weight. Three 
 calculation methods used to determine 
 residual SLS content usually yielded re- 
 sults that agreed to within 10 pet. This 
 
29 
 
 2 4 6 8 10 12 
 CUMULATIVE PRECIPITATION, in of rain 
 
 FIGURE 3. - SLS release curve from the outdoor 
 pilot-scale test. Formulation was approximately 
 50 wt pet SLS. Multiple data points were calcu- 
 lated using weight loss and extraction data. 
 
 large variation is due to the indirect 
 measurements mentioned previously. The 
 curves generally followed the same pat- 
 tern observed in the laboratory study 
 (fig. 2), although SLS release was much 
 more rapid in the field. 
 
 Table 1 shows the residual SLS content 
 for all the formulations after 11 in of 
 precipitation. Essentially all the sur- 
 factant dissolved from nine of the sam- 
 ples within the 4-month period repre- 
 sented by the tabulated results. Seven 
 of these samples were composed of early 
 matrix formulations. At the other ex- 
 treme, two samples released essentially 
 none of the surfactant during the pilot- 
 scale test. Several formulations exhib- 
 ited release rates (residual SLS 65 to 90 
 pet) that might provide the desired re- 
 lease lifetime of several years. 
 
 Negative numbers on table 1 resulted 
 when some of the weight loss assumed to 
 be SLS dissolution was actually loss of 
 matrix. Some of the thinner rubber ma- 
 trices underwent significant degradation 
 that produced visible shrinkage of the 
 pellets. All samples tested in the 
 
 TABLE 1. - Residual SLS after exposure 
 of formulations to 11 in of precip- 
 itation on coal refuse test piles 
 
 Plot 
 
 No 
 
 SLS CO 
 
 ntent , 
 
 pet 
 
 of 
 
 init 
 
 ial' 
 
 1... 
 
 -2 
 
 12... 
 
 -15 
 
 2... 
 
 12 
 
 13... 
 
 18 
 
 3... 
 
 27 
 
 14... 
 
 20 
 
 4... 
 
 -14 
 
 15... 
 
 -7 
 
 5... 
 
 
 
 16... 
 
 98 
 
 6.. 
 
 -16 
 
 17... 
 
 65 
 
 7.. 
 
 -8 
 
 18.. 
 
 20 
 
 8.. 
 
 -5 
 
 19.. 
 
 47 
 
 9.. 
 
 16 
 
 20.. 
 
 125 
 
 10.. 
 
 -18 
 
 21.. 
 
 83 
 
 11.. 
 
 90 
 
 
 
 Plot SLS content, 
 No 
 
 pet of 
 initial 1 
 
 to 
 
 Initial SLS content, ranging from 20 
 65 pet, was normalized ' to 100 pet. 
 
 laboratory and in the pilot-scale test 
 exhibited much higher SLS dissolution 
 rates in the latter ease. Exposure to 
 ultraviolet light and moist, acidic ref- 
 use probably contributed to faster re- 
 lease through degradation of the ma- 
 trices. Burial of the controlled release 
 pellets beneath a soil cover should re- 
 tard release rates by reducing degrada- 
 tion and limiting contact with rainfall. 
 
 FIELD PROJECTS 
 
 The Bureau is participating in one 
 field trial of the controlled release 
 concept (_5 ) . The site is a 15-acre iso- 
 lated ridge in Upshur County, WV, which 
 was mined and reclaimed in three sec- 
 tions. State-of-the-art reclamation 
 techniques, including a clay cap emplaced 
 over the toxic material, were used (19). 
 Surfactant solution and a controlled re- 
 lease surfactant formulation were applied 
 to one section below the clay layer. 
 Since completion of reclamation during 
 spring 1983, seeps and surface runoff 
 have been monitored. To date, the post- 
 mining hydrology has not developed suf- 
 ficiently to allow characterization of 
 drainage quality from the various 
 sections . 
 
30 
 
 Selection of the controlled release 
 material was based on early laboratory 
 data; we now know that the surfactant is 
 released from the matrix in less than 1 
 yr when the pellets are applied to the 
 surface of acidic material. Exposed to 
 no sunlight and less water under the clay 
 cap, detergent release should be signifi- 
 cantly slowed. 
 
 Both Goodrich and Granger are now de- 
 veloping new formulations to optimize 
 surfactant release rates. The former 
 company is currently conducting field 
 tests of 1984 formulations that we have 
 not tested (4). In the oldest test, the 
 controlled release pellets were applied 
 during summer to a portion of a coal ref- 
 use site prior to application of seed and 
 soil to the entire site. At the end of 
 the first growing season, there was good 
 vegetation cover on the treated refuse, 
 compared with extensive acid burnout 
 areas on the untreated portion. 
 
 precipitates when added to synthetic AMD 
 in the pH range of 4 to 5. The precipi- 
 tates probably consist of ferric or fer- 
 rous salts of the organic acids. Further 
 testing was encouraged by the fact that 
 these organic acids are used as food and 
 beverage preservatives and hence should 
 be environmentally safe. 
 
 Laboratory tests of bacterial inhibi- 
 tion have previously been reported (18) . 
 In solution cultures of a pure strain of 
 T. f errooxidans , bacterial activity was 
 monitored as the utilization of ferrous 
 iron in the medium. The bacteria derive 
 energy from oxidation of ferrous iron. 
 Figure 4 illustrates the results in unin- 
 hibited bacteria culture, in sterile me- 
 dium, and in two bacterial cultures con- 
 taining benzoic acid. We found that 10 
 mg/L of either benzoic or sorbic acid was 
 sufficient to decrease the rate of fer- 
 rous iron oxidation to that of sterile 
 controls. 
 
 ORGANIC ACID INHIBITORS 
 
 PILOT-SCALE TESTS 
 
 We began to investigate another alter- 
 native for control of T. f errooxidans 
 when the limitations of the solution sur- 
 factant technique became apparent. For 
 materials having low affinity for sur- 
 factant and sites having high water 
 flow rates, a less soluble inhibitor was 
 needed. The concept was to identify 
 organic compounds with the following 
 properties: 
 
 1 . Toxic to T^ ferrooxidans but innoc- 
 uous to other organisms . 
 
 2. Sparingly soluble in AMD or neu- 
 tralized mine drainage. 
 
 3. Actively bactericidal once redis- 
 solved or in response to acid production. 
 
 Preliminary experiments consisted of a 
 survey of 25 organic compounds , which 
 might be inhibitors and which might pre- 
 cipitate as sparingly soluble compounds 
 in AMD, These experiments yielded two 
 candidate compounds: sodium benzoate and 
 potassium sorbate. We found that O.l-pct 
 solutions of either salt formed organic 
 
 Bactericidal effectiveness of potas- 
 sium sorbate, sodium benzoate, and SLS 
 was investigated for reducing acid pro- 
 duction from fresh and weathered refuse; 
 
 CJ> 
 
 O 
 (f) 
 
 3 
 O 
 
 cc 
 
 DC 
 UJ 
 
 10 
 
 12 3 4 5 6 
 
 TIME, weeks 
 
 FIGURE 4. - Ferrous iron oxidation by T. ferro- 
 oxidans , as function of added benzoic acid. The 
 sterile culture indicates the rate of abiotic 
 oxidation. 
 
31 
 
 preliminary results have been published 
 previously (17) . 
 
 Drums filled with 200 kg of fresh coal 
 refuse were leached weekly by saturating 
 the material for 24 h with tap water. 
 The drained leachate was analyzed for pH, 
 acidity, total dissolved iron, and sul- 
 fate. In the first week of the experi- 
 ment, 24 L of inhibitor solution replaced 
 the water in six of the drums. The three 
 inhibitors were each tested at concentra- 
 tions of 500 and 5,000 mg/L (equivalent 
 to 60 and 600 mg chemical per kilogram 
 of refuse) . Ten drums of refuse were 
 "treated" with tap water and used for ex- 
 perimental control. 
 
 The low doses of treatment chemicals 
 were marginally effective, delaying acid 
 production 1.5 to 5 weeks after leachate 
 from the control barrels became acidic 
 (fig. 5). High treatment doses of 5,000 
 mg/L were effective for 8 to 10 weeks 
 (fig. 6). Potassium sorbate yielded the 
 best results in both treatment series. 
 
 At low dosage rate, sorbate was least 
 expensive on the basis of cost per week 
 of delayed acidification. However, at 
 the high dosage rate, the duration of the 
 treatments were more similar and the 
 
 chemical of choice would probably depend 
 on cost per pound. Approximate bulk 
 prices are $0.90/lb for sodium benzoate, 
 $1.67/lb for SLS, and $3.52/lb for potas- 
 sium sorbate. Field trials will be re- 
 quired before an accurate cost analysis 
 can be made. The longevity of SLS treat- 
 ment under field conditions is about 
 twice as great as in the high-dosage 
 pilot-scale test; the experimental condi- 
 tions of extremely high leaching rates 
 probably underestimate the duration of 
 all three inhibitors. 
 
 After 22 weeks of weathering, 9 of the 
 10 control barrels were treated with the 
 chemical inhibitors to determine their 
 effectiveness in the highly acidic envi- 
 ronment of aged refuse. Drainage acidity 
 levels were approximately 8,000 to 14,000 
 mg/L at the start of this experiment. 
 During the 22-week leaching program, 
 drainage from the untreated barrel re- 
 tained as a control became 70 pet less 
 contaminated. The easily oxldizable py- 
 rite may have been consumed during the 
 initial 22 weeks of weathering; cumula- 
 tive sulfate load data indicated approxi- 
 mately 10 pet of the total pyrite had 
 been oxidized before treatments were ap- 
 plied to the aged refuse. 
 
 "5 
 
 c 
 o 
 
 o 
 o 
 
 E 
 
 jj 
 
 O 
 O 
 
 o 
 
 9 
 
 o 
 
 < 
 
 10 
 
 KEY 
 
 Control 
 
 Sodium lauryl sulfate 
 
 Potassium sorbate 
 
 Sodium benzoate 
 
 
 
 25 
 
 5 10 15 
 
 TIME, weeks 
 
 FIGURE 5, - Acidity levels in leachate from cool 
 refuse treated witfi 500 mg/L of chemical inhibitor. 
 
 D 
 
 c 
 o 
 
 o 
 o 
 
 E 
 
 o 
 o 
 
 CO 
 
 o 
 
 en 
 
 t 
 g 
 
 < 
 
 10 
 
 KEY 
 Control 
 
 Sodium lauryl sulfate 
 Potassium sorbate 
 Sodium benzoate 
 
 re 
 
 fu 
 
 5 10 15 
 
 TIME, weeks 
 
 CURE 6.- Leachate acidity from fresh cool 
 se treated with 5,000 mg/L of chemical inhibitor. 
 
32 
 
 15 
 
 KEY 
 
 Control 
 
 Sodium lauryl sulfate 
 
 Potossium sorbate 
 
 - Sodium benzoate 
 
 20 
 
 25 
 
 5 10 15 
 TIME, weeks 
 
 FIGURE 7. • Effect of low doses of treatment 
 chemicals on weathered coal refuse leachate 
 compared with leachate from untreated refuse. 
 
 Superimposed on the trend of decreasing 
 contaminant concentrations, additional 
 improvements in drainage quality were ob- 
 served (figs. 7-8). At the low dosage 
 rate of 500 mg/L, only potassium sorbate 
 produced significantly better drainage 
 than did the control. All three chemi- 
 cals were effective at the 5,000-mg/L 
 dosage rate. Cumulative acid loads (fig. 
 9) were 17, 29, and 38 pet lower for so- 
 dium benzoate, potassium sorbate, and SLS 
 treatments, respectively, at the end of 
 22 weeks than in the control drainage. 
 Seven weeks after treatment, when the in- 
 hibitors were most effective, cumulative 
 acid loads were 45 to 62 pet lower in the 
 high treatment dose leachates than in the 
 control leachate. 
 
 A field test is now in progress at a 
 revegetated mine site in West Virginia. 
 Dry potassium benzoate powder was applied 
 to the surface on 2 acres overlying the 
 major acid-producing zone. Water quality 
 is being monitored in the vadose zone, in 
 
 o 
 c 
 o 
 J2 
 
 15 
 
 KEY 
 
 Control 
 
 Sodium lauryl sulfate 
 
 Potassium sorbate 
 
 — Sodium benzoate 
 
 ^n; 
 
 20 
 
 5 10 15 
 
 TIME, weeks 
 
 FIGURE 8. - High doses of three treatment 
 chemicals reduced acidity of weathered coal 
 refuse leachate, compared to leachate of un- 
 treated coal refuse. 
 
 25 
 
 10 15 
 
 TIME, weeks 
 
 20 
 
 25 
 
 FIGURE 9. - Cumulative acidity produced by weath- 
 ered cool refuse after application of high doses of in- 
 hibitory chemicals. Curve symbols as in figure 8. 
 
 the saturated zone, and at the discharge 
 seep. 
 
 SUMMARY 
 
 Preliminary experiments were conducted 
 on two alternatives to the surfactant so- 
 lution technique for controlling acid 
 
 drainage. Controlled release of sur- 
 factants appears to be a feasible means 
 of extending the bactericide lifetime. 
 
Further work, in the form of field tests, 
 is needed to determine the cost effec- 
 tiveness of this method. 
 
 The organic Inhibitors , benzoate and 
 sorbate, were of the same general order 
 of effectiveness as surfactant solution 
 in pilot-scale tests. There may be some 
 
 33 
 
 cost advantage in using benzoate; a field 
 test of this compound is in progress. 
 Under moderately acidic conditions where 
 adsorption is unlikely, such as in some 
 underground mines, the metal-organic salt 
 precipitate may have further advan- 
 tages in extending the duration of acid 
 control. 
 
 REFERENCES 
 
 1. American Public Health Association. 
 Standard Methods for the Examination of 
 Water and Waste Water. 13th ed., 1971, 
 874 pp. 
 
 2. Cardarelli, N. Controlled Release 
 Pesticide Formulations. CRC Press, 
 Cleveland, OH, 1976, 224 pp. 
 
 3. Erickson, P. M. , R. L. P. Klein- 
 mann, and P. S. A. Campion. Reducing 
 Oxidation of Pyrite Through Selective 
 Reclamation Practices. Paper in Proceed- 
 ings, 1982 Symposium on Surface Mining 
 Hydrology, Sedimentology , and Reclama- 
 tion, Lexington, KY, Dec. 6-10, 1982, 
 ed. by D. H. Graves. Univ. KY, 1982, 
 pp. 97-102. 
 
 4. Fox, L. A., and V. Rastogi. Devel- 
 opments in Controlled Release Technology 
 and Its Application in Acid Mine Drain- 
 age. Paper in Proceedings, 1983 Symposi- 
 um on Surface Mining, Hydrology, Sedi- 
 mentology, and Reclamation, Lexington, 
 KY, Nov. 27-Dec. 2, 1983, ed. by D. H. 
 Graves. Univ. KY, 1983, pp. 447-455. 
 
 5. Geidel, G. , and F. T. Caruccio. A 
 Field Evaluation of the Selective Place- 
 ment of Acidic Material Within the Back- 
 fill of a Reclaimed Coal Mine. Paper in 
 Proceedings, 1984 Symposium on Surface 
 Mining, Hydrology, Sedimentology, and 
 Reclamation, Lexington, KY, Dec. 2-7, 
 1984, ed. by D. H. Graves. Univ. KY, 
 1984, pp. 127-131. 
 
 6. Good, D. M. , V. T. Ricca, and K. S. 
 Shumate. The Relation of Refuse Pile Hy- 
 drology to Acid Production. Paper in 
 Preprints of Papers Presented Before the 
 Third Symposium on Coal Mine Drainage 
 
 Research (Pittsburgh, PA, May 19-20, 
 1970). Mellon Inst., Pittsburgh, PA, 
 1970, pp. 145-151. 
 
 7. Kleinmann, R. L. P. The Biogeo- 
 
 chemistry of Acid Mine Drainage and a 
 
 Method To Control Acid Formation. Ph.D. 
 
 Thesis, Princeton Univ., Princeton, NJ, 
 1979, 104 pp. 
 
 8. 
 
 Bactericidal Control of 
 
 Acid Problems in Surface Mines and Coal 
 Refuse. Paper in Proceedings, 1980 Sym- 
 posium on Surface Mining, Hydrology, Sed- 
 imentology, and Reclamation, Lexington, 
 KY, Dec. 1-5, 1980, ed. by D. H. Graves. 
 Univ. KY, 1980, pp. 333-337. 
 
 9. Kleinmann, R. L. P., and D. A. 
 Crerar. Thiobacillus f errooxidans and 
 the Formation of Acidity in Simulated 
 Coal Mine Environments. Geomicrobiol. 
 J., V. 1, 1979, pp. 373-388. 
 
 10. Kleinmann, R. L. P., D. A. Crerar, 
 and R. R. Pacelli. Biogeochemistry of 
 Acid Mine Drainage and a Method to Con- 
 trol Acid Formation. Min. Eng., v. 33, 
 
 1980, pp. 300-306. 
 
 11. Kleinmann, R. L. P., and P. M. 
 Erickson. Field Evaluation of a Bacteri- 
 cidal Treatment To Control Acid Drainage. 
 Paper in Proceedings, 1981 Symposium on 
 Surface Mining Hydrology, Sedimentology, 
 and Reclamation, Lexington, KY, Dec. 7- 
 11, 1981, ed. by D. H. Graves. Univ. KY, 
 
 1981, pp. 325-329. 
 
 12. 
 
 Full-Scale Field Trials of 
 
 a Bactericidal Treatment To Control Acid 
 Mine Drainage. Paper in Proceedings, 
 1982 Symposium on Surface Mining 
 
34 
 
 Hydrology, Sedimentology , and Reclama- 
 tion, Lexington, KY, Dec. 6-10, 1982, ed. 
 by D. H. Graves. Univ KY, 1982, pp. 617- 
 622. 
 
 13. Kleinmann, R. L. P. , and P. M. 
 Erickson. Control of Acid Mine Drainage 
 From Coal Refuse Using Anionic Surfac- 
 tants. BuMines RI 8847, 1983, 16 pp. 
 
 14. Leathen, W. W. The Influence of 
 Bacteria on the Formation of Acid Mine 
 Drainage. Abstracted in Coal and the En- 
 vironmental Abstract Series: Mine Drain- 
 age Bibliography, ed. by V. Gleason and 
 H. H. Russell. Bituminous Coal Research, 
 Monroeville, PA, 1976, 288 pp. 
 
 15. Leathen, W. W. , S. Braley, Sr. , 
 and L. D. Mclntyre. The Role of Bacteria 
 in the Formation of Acid From Certain 
 Sulfuritic Constituents Associated With 
 Bituminous Coal. Part 2. Ferrous Iron 
 Oxidizing Bacteria. Appl. Microbiol. , v. 
 1, 1953, pp. 65-68. 
 
 16. Lorenz , W. C. , and R. W. Stephan. 
 Factors That Affect the Formation of Coal 
 Mine Drainage Pollution in Appalachia. 
 
 Attachment C. Appendix C, Acid Mine 
 Drainage in Appalachia. Appalachian Re- 
 gional Committee, Washington, DC, 1969, 
 21 pp. 
 
 17. Onysko, S. J., P. M, Erickson. 
 R. L. P. Kleinmann, and M. Hood. Control 
 of Acid Drainage From Fresh Coal Refuse: 
 Food Preservatives as Economical Alterna- 
 tives to Detergents. Paper in Proceed- 
 ings, 1984 Symposium on Surface Mining, 
 Hydrology, Sedimentology, and Reclama- 
 tion, Lexington, KY, Dec. 2-7, 1984, ed. 
 by D. H. Graves. Univ. KY, 1984, pp. 35- 
 42. 
 
 18. Onysko, S. J. , R. L. P. Kleinmann, 
 and P. M. Erickson. Ferrous Iron Oxida- 
 tion by Thiobacillus f errooxidans : Inhi- 
 bition With Benzoic Acid, Sorbic Acid, 
 and Sodium Lauryl Sulfate. Appl. and En- 
 viron. Microbiol., v. 28, No. 1, July 
 1984, pp. 229-231. 
 
 19. West Virginia Acid Mine Drainage 
 Task Force. Suggested Guidelines for 
 Method of Operation in Surface Mining of 
 Areas With Potentially Acid-Producing Ma- 
 terials. 1979, 20 pp. 
 
35 
 
 ALKALINE INJECTION: AN OVERVIEW OF RECENT WORK 
 
 By Kenneth J. Ladwig, ^ Patricia M, Erick.son,2 and Robert L. P. Klei 
 
 nmann- 
 
 INTRODUCTION 
 
 Injection of alkaline fluid into sur- 
 face mine spoil to control acid mine 
 drainage (AMD) is a procedure generating 
 considerable interest in Pennsylvania. 
 At least six different mine companies or 
 contractors have attempted some form of 
 injection in the last 2 yr, and many more 
 are considering its use. This paper 
 gives a brief overview of the current 
 status of alkaline injection and of the 
 Bureau of Mines injection research. 
 
 Introduction of alkalinity is the stan- 
 dard method of mitigating acid dis- 
 charges. Surface alkaline loading prior 
 to flow through the spoil has been used 
 to slow down the acid-production process 
 ii_, ^) • ^ More commonly, alkalinity is 
 added to the discharge to neutralize ex- 
 isting acidity with conventional water 
 treatment (4). 
 
 decreasing sludge storage 
 requirements. 
 
 and removal 
 
 2. The metal precipitates may coat py- 
 rite surfaces, "armoring" them from fur- 
 ther chemical weathering, 
 
 3. The alkaline environment within the 
 spoil would be less favorable to contin- 
 ued pyrite oxidation. 
 
 4. The high-pH environment would limit 
 metal leaching within spoil. 
 
 5. Spoil water that "leaks" through 
 the mine floor discharges to the ground 
 water system untreated. Alkalinity in- 
 troduced into the spoil water reservoir 
 may offer at least partial treatment of 
 the leakage and decrease overall ground 
 water degradation. 
 
 The premise of alkaline injection is 
 the in-place neutralization of acid water 
 stored in the spoil. In this respect, 
 alkaline injection is not much different 
 than conventional water treatment. Alka- 
 line materials that have been used for 
 injection are sodium hydroxide, hydrated 
 lime, and sodium carbonate, all of which 
 are commonly used in AMD water treatment. 
 Some of the potential advantages of in- 
 jection over conventional water treatment 
 follow: 
 
 1. Raising the pH of the spoil water 
 may result in the precipitation and fil- 
 tering of some metals prior to discharge, 
 
 6. Treatment by alkaline injection 
 could be done on an intermittent basis, 
 lowering labor costs. 
 
 While the premise of alkaline injection 
 is straightforward, implementation is 
 not. The extent to which any of the 
 above listed advantages are realized is 
 not known. Of the six attempts with 
 which we are familiar, none have yet sub- 
 stantially improved spoil seep water 
 quality. Unfortunately, documentation of 
 these injections was generally incom- 
 plete. For this reason, the Bureau ini- 
 tiated a study to evaluate the technical 
 merit of the injection approach. 
 
 ^Hydrologist. 
 
 ^Supervisory physical science, 
 
 •^Research supervisor. 
 Pittsburgh Research Center, Bureau of 
 Mines, Pittsburgh, PA, 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 Described in the following section are 
 two injection programs for which a rea- 
 sonable amount of documentation was 
 available. At the Fayette site, the Bu- 
 reau monitored the results of an injec- 
 tion performed by Kaiser Refractories, 
 Much of the data were generously supplied 
 by Bernard Leber of Kaiser Aluminum and 
 Chemical Corp, and Mike Popchak of the 
 
36 
 
 Kaiser Refractories division. Descrip- 
 tions and data for the Clearfield site 
 were kindly provided by Jim McNeil of 
 
 Al Hamilton Contracting Co. Bureau re- 
 search on alkaline injection is described 
 in the final section. 
 
 EXAMPLES OF ALKALINE INJECTIONS 
 
 FAYETTE COUNTY, PA 
 
 A lime slurry injection project was 
 conducted at a surface mine in Fayette 
 County in 1983. The lU-ha Fayette site 
 was mined at various times between the 
 mid-1950 's and late 1970 's. At present, 
 about one-half of the site has been re- 
 vegetated (fig. 1). Spoil thickness 
 ranges from 15 to 30 m. 
 
 The site has several toe-of-spoil seep- 
 age areas, one of which (seepage area A, 
 figure 1) discharges directly onto the 
 flood plain of a small, perennial stream. 
 Seepage area A has two major seeps, 16 
 and 18. The proportion of flow from each 
 of these seeps varied throughout the 
 study, but in 1983, flow at seep 18 was 
 generally much higher than at seep 16. 
 Owing to steep topography and space limi- 
 tations, installation of conventional 
 
 treatment facilities and settling ponds 
 between seep area A and the stream was 
 considered impractical. 
 
 In February 1983, Kaiser initiated an 
 injection program in seepage area A. 
 Fifteen injection wells were installed 
 approximately 90 m up gradient from the 
 seepage area (fig. 1). The wells were 
 drilled an average of 18 m to the mine 
 floor. Two-inch-diameter polyvinyl chlo- 
 ride well pipe was placed in each hole 
 and cemented at the surface. The lower 
 16 m of the pipes were perforated with 
 0.32-cm holes. Water levels were 3 to 
 5 m above the mine floor. 
 
 Between February and October 1983, 
 119 tons of hydrated lime were pumped 
 into the wells in slurry form. The 
 slurry concentration ranged from 4 pet 
 lime during the early stages of injection 
 
 Scale, m 
 
 FIGURE 1. - Map of surface mine study site in Fayette County, PA. 
 
37 
 
 to 0.4 pet during the latter stages. In- 
 jections were performed weekly at a rate 
 of 7 to 42 tons of lime per week. 
 
 The transit time from the injection 
 wells to seepage area A was estimated by 
 analyzing the movement of a sodium tracer 
 in the lime slurry (fig. 2). The first 
 arrival of the injected sodium at seepage 
 area A was about 4 months after the be- 
 ginning of the injection. Peak concen- 
 trations occurred 7 to 9 months after the 
 beginning of the injection. Concentra- 
 tions began tailing off following the 
 cessation of injection in October. Peak 
 sodium concentrations at the seeps indi- 
 cated a 1:3 ratio of injected water to 
 spoil water. 
 
 Trends in pH and acidity for seeps 16 
 and 18 are shown in figures 3-6. Although 
 very few pre-injection data were avail- 
 able, there did appear to be a modest im- 
 provement at seep 16 beginning about 7 
 months after the initial injection. The 
 pH at seep 16 increased by 1/4 to 1/2 of 
 a pH unit, while the acidity decreased by 
 30 to 40 pet. However, no significant 
 changes in water quality were observed at 
 seep 18. As seep 18 comprises a larger 
 percentage of the total flow from seepage 
 area A, the overall impact of the injec- 
 tion was minimal. 
 
 1983 
 
 FIGURE 2. 
 and 17. 
 
 19&4 
 
 SAMPLE DATE 
 
 1985 
 
 - Sodium concentrations at seeps 16 
 
 Assuming complete reaction, 119 tons of 
 lime is capable of neutralizing 36 mil- 
 lion gal of water with an average acidity 
 of 1,000 mg/L. Because the total dis- 
 charge from seepage area A in 1983 was 
 less than 20 million gal, the high lime 
 dosage should have had a profound impact 
 on seep quality. 
 
 The explanation of the poor results may 
 lie in the inefficient mixing of the lime 
 with the spoil water. The solubility of 
 lime in deionized water is 1,600 mg/L at 
 20<^ C (1). Saturation with respect to 
 lime would produce a solution of 0.16 pet 
 dissolved lime. Because the lime slurry 
 was mixed at concentrations of 0.4 to 
 4 pet, 60 to 95 pet of the lime was in 
 suspension rather than solution. At a 
 treatment plant using mixers to induce 
 turbulent flow, much of ' the suspended 
 lime might eventually contact acidic 
 water and participate in the neutraliza- 
 tion reactions. However, the injected 
 fluid was not subjected to continuous 
 turbulent flow and very likely did not 
 mix efficiently with the spoil water. In 
 the absence of turbulence and mixing, 
 suspended lime will settle rapidly. En- 
 hanced solution will occur only along the 
 slurry-spoil water contact surface, con- 
 siderably slowing the rate of lime con- 
 sumption. As a conservative estimate, 
 over 50 pet of the 119 tons of lime at 
 the Fayette site may have settled out of 
 suspension shortly after injection. 
 
 The lime remaining in solution mixed 
 with the spoil water at a 1:3 dilution 
 rate at the peak injection period, as 
 previously determined from the sodium 
 data. Assuming an initial concentration 
 of 1,600 mg/L dissolved lime in the in- 
 jection fluid, the maximum lime concen- 
 tration following 1:3 mixing with the 
 spoil water is 400 mg/L. This amount of 
 lime is capable of neutralizing only 
 490 mg/L acidity. 
 
 These calculations are intended to il- 
 lustrate in a general sense the controls 
 placed on the system by the solubility of 
 lime and the low velocity of ground water 
 flow. Although these numbers are in 
 
38 
 
 1983 
 
 1985 
 
 1984 
 
 SAMPLE DATE 
 
 FIGURE 3. - Seep 16 pH. Lime injection began 
 in February 1983. 
 
 1J500 
 
 1,400 
 
 1^00 
 
 800 
 
 400 
 
 1983 
 
 1984 
 
 SAMPLE DATE 
 
 1985 
 
 FIGURE 4. - Seep 16 acidity. 
 
 1983 
 
 1984 
 SAMPLE DATE 
 
 1985 
 
 FIGURE 5. - Seep 18 pH. Lime injection began 
 in February 1983. 
 
 1400 
 
 1^00- 
 
 600 
 
 1983 
 
 1984 
 
 SAMPLE DATE 
 
 1985 
 
 FIGURE 6. - Seep 18 acidity. 
 
 reasonable agreement with the data from 
 seep 16, they do not explain why no 
 change was observed at seep 18. Ongoing 
 work at the site is designed to better 
 describe the hydrologic differences be- 
 tween the two seeps. 
 
 CLEARFIELD COUNTY, PA 
 
 A hydrated lime injection program is 
 also being conducted by a mine company 
 at a site in Clearfield County, PA. The 
 
 approach taken at the Clearfield site was 
 similar to that described for the Fayette 
 site. In June 1982, 22 injection wells 
 were drilled an average of 15 m to the 
 mine floor. The wells were located about 
 75 m upgradient of the toe-of-spoil seep. 
 Approximately 3,000 gal of 4-pct lime 
 slurry are pumped into each well on a 
 monthly basis from April through Novem- 
 ber. Due to cold temperatures, no injec- 
 tions occur between December and March. 
 
39 
 
 Preliminary data indicate that toe-of- 
 spoil seeps at the Clearfield site have 
 not exhibited appreciable improvement 
 since injection began. However, dovm- 
 stream monitoring does indicate improve- 
 ment in the receiving stream. In the 
 last 2 yr, there have been reductions in 
 acidity and iron concentrations at the 
 downstream monitoring station. 
 
 The lack of improvement at the toe- 
 of-spoil seeps may again be a result of 
 the low solubility of lime, as described 
 for the Fayette site. Why then did the 
 downstream water quality improve? 
 
 Owing to other modifications in the 
 watershed contemporaneous with the 
 
 injection, it is not possible at this 
 time to attribute the downstream water 
 quality improvement solely to the injec- 
 tion program. If the improvement is re- 
 lated to the injection, it may reflect an 
 improvement in ground water quality below 
 the mine floor. Stored spoil water leak- 
 ing through the mine floor may contain 
 a high alkaline load following contact 
 with the settled lime. The mine water 
 recharges the underlying ground water 
 system and eventually discharges by dif- 
 fuse seepage to the receiving stream, 
 resulting in improved water quality down- 
 stream from the site. While this is 
 purely conjecture at this time, the down- 
 stream water quality improvement cer- 
 tainly merits further study. 
 
 BUREAU OF MINES INJECTION PROJECT 
 
 The widespread interest in injection 
 technology, along with the limited suc- 
 cess to date, prompted a Bureau of Mines 
 study of the injection approach. While 
 alkaline injection is not a cure-all for 
 AMD problems at surface mines , the se- 
 lective use of injection in combination 
 with other abatement procedures may offer 
 several benefits. Possibly the most 
 valuable potential benefit is the renova- 
 tion of contaminated ground water below 
 the mine floor, a problem not currently 
 addressed by any other treatment 
 technology. 
 
 Critical to the success of alkaline in- 
 jection is good mixing of the alkaline 
 fluid and the contaminated spoil water. 
 This requires detailed understanding of 
 site hydrology and acid-producing charac- 
 teristics, including source material, 
 flow paths, flow rates, flow volumes, and 
 spoil-water chemistry. We believe that 
 inadequate mixing, largely due to the low 
 solubility of lime and low flow veloc- 
 ities, was one of the primary shortcom- 
 ings of the previous injection attempts. 
 Our approach will differ from these at- 
 tempts in two ways. 
 
 First, sodium carbonate solution will 
 replace lime slurry as the alkaline 
 
 fluid. Sodium carbonate is about 100 
 times more soluble than lime (2^) , allow- 
 ing mobility of a greater fraction of the 
 alkaline load. The concentration of the 
 sodium carbonate solution will be select- 
 ed to maximize alkaline loading with min- 
 imal density contrasts between the in- 
 jected fluid and the spoil water. 
 
 Second, injection wells will be situ- 
 ated at least 300 m upgradient from the 
 seep to enhance dispersion of the inject- 
 ed fluid. Dispersion in porous media is 
 directly related to distance along the 
 flow path ( O • Placing the injection 
 wells on the upgradient end of the site 
 will allow for maximum mechanical disper- 
 sion of the alkaline fluid. This will 
 also minimize the possibility of the al- 
 kaline fluid migrating directly to the 
 seep as an unreacted plume. Monitoring 
 wells will be sampled to track the prog- 
 ress of the injected fluid in the 
 spoil. 
 
 Bureau work to date has consisted of 
 pilot-scale testing of lime and sodium 
 carbonate, and preliminary site evalua- 
 tion for a full-scale field test. In the 
 pilot-scale tests, sodium carbonate was 
 considerably more mobile than lime. The 
 full-scale field test began in spring of 
 
40 
 
 1985 at the Fayette County site described 
 earlier. 
 
 In addition to the field test, we hope 
 to conduct laboratory column studies to 
 simulate and study in detail the reac- 
 tions between the injected fluid and 
 
 spoil water. In particular, we are in- 
 terested in observing the reaction prod- 
 ucts — gaseous, aqueous, and solid — and 
 evaluating their effect on the metal ion 
 chemistry and pyrite oxidation system. 
 This work is tentatively scheduled to be- 
 gin by mid-1985. 
 
 REFERENCES 
 
 1. Caruccio, F. T. , and G. Geidel. 
 Induced Alkaline Recharge Zones to Miti- 
 gate Acidic Seeps. Paper in Proceedings, 
 1984 Symposium on Surface Mining, Hydrol- 
 ogy, Sedimentology , and Reclamation, Lex- 
 ington, KY, Dec. 2-7, 1984, ed. by D. H. 
 Grover. Univ. of KY, 1984, pp. 43-48. 
 
 2. Freeze, R. A., and J. A. Cherry. 
 Groundwater. Prentice-Hall, 1979, pp. 
 388-413. 
 
 3. Neast, R. C. (ed.). Handbook of 
 Chemistry and Physics. CRC Press, 53d 
 ed., 1972-73, pp. B-77 and B-137. 
 
 4. U.S. Environmental Protection Agen- 
 cy. Design Manual: Neutralization of 
 Acid Mine Drainage. EPA-6001Z-83-001, 
 1983, 231 pp. 
 
 5. Waddell, R. K. , R. R. Parizek, and 
 D. R. Buss. The Application of Limestone 
 and Lime Dust in the Abatement of Acidic 
 Drainage in Centre County, PA. PA Dep. 
 Trans. Office Res. and Spec. Studies, 
 Project 73-9, Final Report — Executive 
 Summary, 1980, 79 pp. 
 
41 
 
 COMPARATIVE TESTS TO ElEMOVE MANGANESE FROM ACID MINE DRAINAGE 
 
 By George R. Watzlaf ' 
 
 INTRODUCTION 
 
 The Surface Mining Control and Recla- 
 mation Act of 1977 mandates that mine 
 drainage discharge water meet qual- 
 ity standards for pH, iron, manganese, 
 and total suspended solids (11).^ These 
 standards are shown in table 1. Typical 
 treatment of acid mine drainage involves 
 addition of an alkaline material (such 
 as lime or sodium hydroxide) , natural or 
 mechanical aeration, and settling. When 
 mine drainage is neutralized to a pH near 
 7, the ferrous iron oxidizes and forms an 
 iron sludge, Fe(0H)3. This treatment 
 satisfies the effluent standards for pH 
 and iron, but may not remove much manga- 
 nese from the water. 
 
 Typical acid mine drainage contains 1 
 to 8 mg/L manganese, but concentrations 
 of 50 to 100 mg/L are not uncommon {6_, 
 9). At present, most mine operators with 
 manganese problems are using excess alka- 
 linity to raise pH of the mine water to 
 
 ^Mining engineer. Pittsburgh Research 
 Center, Bureau of Mines, Pittsburgh, PA. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 about 10.0 to precipitate manganese. 
 During the precipitation of manganese, as 
 Mn02, acid is produced and the pH of the 
 water decreases. Whether or not the pH 
 will fall below 9.0 depends on individual 
 mine water characteristics. If the pH 
 remains above 9.0 the mine operator has 
 two options: apply to State autorities 
 for a variance to discharge high-pH wa- 
 ter, or reacidify the high-pH water. An 
 alternative to excess alkalinity is the 
 use of chemical oxidants such as chlorine 
 gas, hypochlorite salts (sodium and cal- 
 cium), ozone, potassium piermanganate, or 
 hydrogen peroxide (^-^> 9^). These oxi- 
 dants can oxidize soluble manganese 
 to insoluble Mn02 ^t pH values within the 
 regulatory criteria. 
 
 Based on a review of the literature on 
 manganese removal, three chemical treat- 
 ments were selected for field testing: 
 excess alkalinity, sodium hypochlorite, 
 and potassium permanganate. The selec- 
 tion of these treatment methods was based 
 on ease of use, costs (capital and oper- 
 ating), availability, effectiveness, and 
 likelihood of acceptance by the mining 
 industry. 
 
 TABLE 1. - Effluent limitations 
 
 
 Maximum 
 allowable 
 
 Av 
 
 erage of daily values 
 for 30 consecutive 
 discharge days 
 
 Iron, total 
 
 
 • .me /L . . 
 
 7.0 
 
 4.0 
 
 70.0 
 
 6.0-9.0 
 
 
 
 3.5 
 
 Mflnganp^Pj ^n^fl^ , , 
 
 
 . .m2 /L. . 
 
 2.0 
 
 Total suspended so 
 pH 
 
 lids. 
 
 . .mg/L. . 
 
 35.0 
 6.0-9.0 
 
42 
 
 EXTENT OF MANGANESE IN ACID MINE DRAINAGE 
 
 Mine drainage discharges that require 
 treatment for removal of manganese are 
 widespread throughout the Eastern U.S. 
 coalfields. A study of mine discharges 
 containing manganese was conducted by 
 Pennsylvania's Department of Environmen- 
 tal Resources for Jefferson, Clearfield, 
 Clinton, Venango, and Clarion Counties. 
 This study found manganese concentrations 
 averaging 75 mg/L and ranging from 20 to 
 170 mg/L. In southwestern Pennsylvania, 
 mining of the Waynesburg seam can result 
 in manganese concentrations over 200 
 mg/L. In eastern Kentucky manganese con- 
 centrations of 10 to 100 mg/L are common. 
 Mine discharges in southern Illinois typ- 
 ically have 5 to 10 mg/L of manganese, 
 but values in excess of 300 mg/L have 
 been reported (_5) . 
 
 Sodium hypochlorite solution is being 
 used to remove manganese from mine drain- 
 age in eastern Kentucky. Two sites were 
 visited used a combination of sodium hy- 
 pochlorite and sodium hydroxide. At one 
 site, the operator initially used calcium 
 hypochlorite briquettes but had dif- 
 ficulty controlling the quantity of chem- 
 ical added to the system. At the other 
 treatment facility, a flocculant was 
 
 required to achieve adequate settling of 
 the precipitated sludge. 
 
 Potassium permanganate has been used to 
 treat mine drainage at a site in Pennsyl- 
 vania. Granular potassium permanganate 
 was added directly to the sodium hydrox- 
 ide solution. The operator at this site 
 achieved good results for a few months 
 but had difficulty maintaining the proper 
 dosage. The operator is now using excess 
 alkalinity to remove manganese. 
 
 Bureau personnel visited two sites in 
 Pennsylvania that use the excess alkalin- 
 ity method for removal of manganese. One 
 operator uses sodium hydroxide in liquid 
 form, and the other adds hydrated lime 
 via a flash mixer. At both sites the op- 
 erators add alkali to raise pH to about 
 10.0 and discharge at a pH near 9.0. 
 
 With adequate control, all three treat- 
 ment methods can be effective in reducing 
 manganese concentrations below effluent 
 limitations. To directly compare the 
 costs and effectiveness of these methods, 
 they must be used to treat the same mine 
 water. 
 
 CHEMISTRY OF MANGANESE REMOVAL 
 
 The chemistry for removing manganese 
 with pH adjustment is very similar to 
 that of iron removal, but oxidation of 
 Mn^"*" to Mn^"*" requires higher pH values 
 than are required for iron oxidation. To 
 remove manganese, the following two re- 
 actions are promoted: 
 
 Mn2+ + 1/2 O2 + 2H+ ^ Mn"*-^ + H2O (1) 
 
 Mn"^"^ + 2H2O ->■ Mn02 + 4H+ (2) 
 
 The rate of manganese oxidation is pH 
 dependent and extremely slow at pH val- 
 ues less than 8.0. Reduction of man- 
 ganese concentrations below 2 mg/L 
 can occur at pH 8.4 (8^), but most mine 
 drainages require pH values over 9.5 (_5, 
 7-9). 
 
 To remove manganese at near-neutral pH, 
 a chemical oxidant must be used. Oxi- 
 dizing agents commonly used in water 
 treatment include chlorine, sodium hypo- 
 chlorite, calcium hypochlorite, potassium 
 permanganate, hydrogen peroxide, and 
 ozone. The following reactions show how 
 sodium hypochlorite (NaOCl) and potassium 
 permanganate (KMn04) each oxidize dis- 
 solved manganese (Mn^"*") and convert it to 
 manganese oxide (Mn02): 
 
 NaOCl + Mn2+ + H2O 
 
 ->■ Mn02 + Na+ + CI" + 2H"^ (3) 
 2KMn04 + 3 Mn2+ + 2H2O 
 
 -J- 5 Mn02 + 2K+ + 4H+ (4) 
 
43 
 
 Because these chemicals will oxidize 
 both Fe'^2 3nd Mn"^^ in acid mine drainage, 
 it is reasonable to first oxidize Fe^"*", 
 by increasing pH and aerating, before 
 adding an oxidizing agent. This reduces 
 the requirement for the chemical oxidant 
 and lowers costs. Also, some manganese 
 is removed by coprecipitation with iron 
 even at near-neutral pH, by sorbtion to 
 Fe(0H)3 {]_) , further reducing the chemi- 
 cal oxidant requirement. 
 
 FIELD 
 
 Regardless of the method used, control- 
 ling the addition of chemical treatment 
 is very important. Many variables influ- 
 ence the removal of iron and manganese, 
 and experimentation with different chemi- 
 cal dosages may be required to achieve 
 optimal results. As the quantity and 
 quality of AMD change, the dosage of the 
 treatment chemical must change to ensure 
 effective manganese removal. 
 
 TESTS 
 
 The purpose of the field tests was to 
 determine the most economic chemical 
 treatment that would successfully reduce 
 manganese concentrations below 2 rag/L. 
 All testing was conducted at the same 
 surface mine site in southwestern Penn- 
 sylvania. Based on ease of use, costs 
 (capital and operating), availability, 
 effectiveness, and likelihood of accept- 
 ance by the mining industry, sodium hypo- 
 chlorite, potassium permanganate, and ex- 
 cess alkalinity were chosen for field 
 testing. 
 
 The field site is an active surface 
 mine with over half of the site mined and 
 reclaimed. Existing treatment consists 
 of sodium hydroxide (NaOH) addition with 
 two settling ponds connected in series 
 (fig. 1). Raw water contains concentra- 
 tions of manganese consistently over 100 
 mg/L. Flow at this site is seasonal and 
 averaged 40 gal/min during the testing 
 period. 
 
 Three series of tests were conducted. 
 In all tests, the raw water was first 
 treated with NaOH for pH adjustment 
 and some iron oxidation before adding 
 any other chemicals. Series 1 used chem- 
 ical dosages based on reaction stoichio- 
 metry for complete removal of manganese 
 and iron. Series 2 used varying amounts 
 of each chemical treatment to determine 
 the minimum dosage required to reduce 
 manganese below 2 mg/L. The water in 
 series 2, which was first neutralized 
 with NaOH, still contained high fer- 
 rous iron concentrations. Therefore, 
 series 3 tests repeated the procedure of 
 series 2, but used additional aeration to 
 reduce iron levels before further chemi- 
 cal treatment. Reduced iron levels were 
 
 Seep 
 
 Raw 
 water 
 
 i Pond 2- 
 
 Final 
 effluent 
 
 FIGURE 1. - Water treatment 
 system at field site. 
 
44 
 
 expected to lower the chemical require- 
 ments for sodium hypochlorite and potas- 
 sium permanganate. 
 
 All costs presented in this study were 
 based on bulk purchases of each chemical, 
 including delivery. The costs for 20- 
 pct sodium hydroxide and 15-pct sodium 
 hypochlorite solutions were $0.28/ gal 
 and $0.80/gal, respectively. The cost 
 for granular potassium permanganate was 
 $1.34/lb. 
 
 SERIES 1 
 
 In these tests, theoretically calcu- 
 lated dosages of the three chemical 
 treatments were used to determine if they 
 would effectively reduce manganese con- 
 centrations below 2 mg/L. Raw water was 
 first treated with NaOH to raise pH to 
 8.8 with some of the precipitated solids 
 settling in pond 1 (fig. 1). The quality 
 of the raw water and water after the ini- 
 tial NaOH treatment and settling is shown 
 in table 2. 
 
 TABLE 2. - Water quality for test 
 series 1 
 
 After 
 
 
 initial 
 
 Raw 
 
 NaOH 
 
 water 
 
 treatment 
 
 pH 5.3 
 
 Acidity (as CaC03) 
 
 mg/L.. 53.0 
 Alkalinity (as CaC03) 
 
 mg/L.. 
 
 Fe + 2 mg/L.. 230 
 
 Total Fe mg/L.. 230 
 
 Mn mg/L.. 120 
 
 8.8 
 
 82 
 
 92 
 
 160 
 97 
 
 Sodium hypochlorite, potassium perman- 
 ganate, and sodium hydroxide were added 
 as 10-, 3-, and 20-pct solutions, respec- 
 tively, at point B (fig. 1). Each chem- 
 ical was gravity-fed from 55-gal drums 
 through plastic tubing. Dosage was reg- 
 ulated with a polyvinyl chloride needle 
 valve. Samples were collected in a 
 5-gal container at point C (fig. 1). 
 This container was then partially sub- 
 mersed in pond 2 (to maintain pond tem- 
 perature) and left to settle for 23 h. 
 After the settling period, samples of 
 the supernatant liquid were taken and 
 analyzed. 
 
 Series 1 consisted of six tests: two 
 controls, two excess alkalinity, one so- 
 dium hypochlorite, and one potassium per- 
 manganate. Table 3 shows the results of 
 these tests. Iron was reduced below ef- 
 fluent standards in all six tests. The 
 two controls did not reduce manganese be- 
 low effluent limitations, but some man- 
 ganese was removed, probably by sorbtion 
 to Fe(0H)3. Also, some manganese oxida- 
 tion may have occurred since pH in these 
 tests was 8.4 and 8.6. The tests involv- 
 ing additional chemical treatment all 
 reduced manganese concentrations below 
 2 mg/L. 
 
 The cost of each chemical treatment 
 (table 3) indicates that excess alkalin- 
 ity was the most cost-effective method in 
 this series. However, the dosages of 
 NaOCl and KMn04 used in these tests may 
 have been greater than the actual minimum 
 effective dosage. Series 2 tests were 
 performed to determine these minimum 
 chemical requirements. 
 
 TABLE 3. - Results of test series 1 
 
 Test 
 
 Water quality after 
 23 h of settling 
 
 Chemical cost per 
 
 
 pH 
 
 Total Fe, 
 mg/L 
 
 Mn, 
 mg/L 
 
 1,000 gal of water 
 
 Control 1 
 
 8.6 
 8.4 
 11.0 
 9.4 
 8.1 
 7.0 
 
 0.1 
 1.2 
 .6 
 .6 
 .6 
 .8 
 
 16 
 
 27 
 
 .6 
 
 1.1 
 
 .7 
 
 1.2 
 
 $1.06 
 
 Control 2 
 
 .80 
 
 Excess alkalinity 1 (to pH 11.3) 
 Excess alkalinity 2 (to pH 10.3) 
 Sodium hvDOchlorite. •.••.••..... 
 
 1.55 
 1.31 
 2.28 
 
 Potassium permanganate 
 
 4.49 
 
45 
 
 SERIES 2 
 
 This series of tests consisted of try- 
 ing several dosages of the three treat- 
 ment chemicals. As in series 1, raw 
 water was first treated with NaOH to 
 raise pH. After the raw water was treat- 
 ed with NaOH, ferrous iron concentrations 
 remained high. This was caused by inade- 
 quate aeration and the short detention 
 time of pond 1. The quality of the water 
 used for this series of tests is shown in 
 table 4. 
 
 TABLE 4. - Water quality for test 
 series 2 after initial NaOH 
 treatment 
 
 pH 9.0 
 
 Alkalinity (as CaC03) mg/L. . 110 
 
 Fe"'^ mg/L.. 88 
 
 Total Fe mg/L. . 140 
 
 Mn mg/L. . 78 
 
 20 
 
 ^ 15 
 
 - 
 
 
 ^ 
 
 Q 
 
 o 
 
 \ 
 
 z 
 
 \ 
 
 o 
 
 ) 
 
 ^ 10 
 
 9 A 
 
 UJ 
 
 
 A 
 
 en 
 
 
 
 LU 
 
 
 
 2 
 
 
 
 g 
 
 \ 
 
 
 
 KEY 
 o Potassium permanganate 
 A Sodium hypochlorite 
 D Excess alkalinity 
 
 2 3 4 
 TOTAL CHEMICAL COSTS 
 
 PER 1,000 GAL OF WATER, dollars 
 
 FIGURE 2. - Te^t series 2: Costs of chemical 
 treatments versus manganese cbncentrations after 
 23 h of settling. 
 
 Twenty-three 400-mL samples were col- 
 lected at point C (fig. 1). Three sam- 
 ples were used as controls. The remain- 
 ing 20 were treated as follows: 6 dif- 
 ferent dosages of NaOCl, 6 different 
 dosages of KMn04, and 8 different dosages 
 of NaOH to raise pH between 9.4 and 10.5. 
 These samples were left to settle for 23 
 h, after which the supernatant liquid was 
 analyzed. 
 
 The results of these tests are sum- 
 marized in figure 2. This graph plots 
 total chemical cost versus the concentra- 
 tion of manganese remaining in solution 
 after 23 h of settling. Included in each 
 chemical cost is the cost for the initial 
 NaOH treatment ($0.83/1,000 gal). 
 
 As in series 1 tests, excess alkalinity 
 proved to be the most cost-effective 
 method. Ferrous iron concentrations of 
 88 mg/L may have caused an increase in 
 demand for NaOCl and KMn04. It was de- 
 cided to try another series of tests to 
 determine the effects of lower ferrous 
 iron concentrations. 
 
 Series 3 
 
 In these tests, raw water was collected 
 at the seep (point /. of figure 1). NaOH 
 was added to the raw water to raise pH to 
 7.5. This water was then aerated by 
 
 100 
 
 £ 
 (J 
 o 
 
 80 
 
 60 
 
 UJ 
 
 UJ 40 
 
 < 
 
 20- 
 
 
 
 KEY 
 ° Potassium permanganate 
 A Sodium hypochlorite 
 □ Excess alkalinity 
 
 2 3 4 5 
 TOTAL CHEMICAL COSTS 
 
 PER 1,000 GAL OF WATER, dollars 
 
 FIGURE 3. - Test series 3: Costs of chemical 
 treatments versus manganese concentrations after 
 23 h of settling. 
 
 pouring it from one bucket to another, 
 causing iron to oxidize and precipitate 
 and pH to decrease. The procedure of 
 neutralization and aeration was repeated 
 until pH stabilized at 7.5. Analysis 
 showed that ferrous iron concentrations 
 were reduced to approximately 1 mg/L (ta- 
 ble 5). Treatment chemicals were then 
 added to this water, which was low in 
 ferrous iron. 
 
46 
 
 TABLE 5. - Water quality for test 
 series 3 after initial NaOH 
 treatment and induced aeration 
 
 pH 7.5 
 
 Alkalinity (as CaC03) mg/L. . 21 
 
 Acidity (as CaC03) mg/L.. 5.0 
 
 Fe+2 mg/L. . 0. 9 
 
 Total Fe mg/L.. 3.7 
 
 Mn mg/L. . 95 
 
 Twenty 400-mL samples were collect- 
 ed and treated as follows: one con- 
 trol, seven NaOCl-treated samples, six 
 
 KMn04-treated samples, and six excess 
 NaOH samples with pH raised to between 
 9.2 and 10.5. The samples were left to 
 settle for 23 h. The supernatant liquid 
 was sampled and analyzed. 
 
 The results of these tests are shown in 
 figure 3. Again the cost of the initial 
 NaOH ($0.36/1,000 gal) was added to each 
 cost. As in the first two series of 
 tests, the most cost-effective method was 
 excess alkalinity. The removal of fer- 
 rous iron did not reduce the chemical re- 
 quirements for the NaOCl and KMn04. 
 
 DISCUSSION AND SUMMARY 
 
 Excess alkalinity was the least expen- 
 sive method to remove manganese from acid 
 mine drainage. Any alkaline material 
 capable of raising pH above 10 can effec- 
 tively remove manganese. One drawback of 
 the excess alkalinity method is that the 
 final effluent may not meet effluent lim- 
 itations (pH less than 9.0). The mine 
 operator must get a variance in order to 
 discharge high-pH water. If a variance 
 to discharge high-pH water is not grant- 
 ed, the operator has to either add acid 
 to lower pH or use an oxidizer such as 
 NaOCl or KMn04. 
 
 Sodium hypochlorite was more expensive 
 than excess alkalinity but less expensive 
 than potassium permanganate. Sodium hy- 
 pochlorite is commercially sold as a 15- 
 pct-available-chlorine solution. This 
 solution can be easily introduced into 
 the treatment system. A disadvantage of 
 sodium hypochlorite is that it loses 
 potency with age. The 15-pct-available- 
 chlorine is guaranteed 10 pet by time of 
 delivery, and additional storage can lead 
 to further reduction in strength. An- 
 other disadvantage is the possibility of 
 residual chlorine in the effluent, which 
 may be regulated by State agencies. 
 
 Potassium permanganate was the most ex- 
 pensive of the three chemical treatments. 
 An advantage of potassium permanganate is 
 that it acts as a color indicator for 
 correct dosage. KMn04 is sold in nugget 
 or granular form, and if KMn04 is to be 
 added as a solution, the diluting and 
 
 mixing must be done on site. It is im- 
 portant not to add too much KMn04, since 
 an excess will increase manganese concen- 
 trations. This effect is shown in fig- 
 ures 2 and 3, where manganese concentra- 
 tions increase when excess permanganate 
 is added. 
 
 In all three treatments, controlling 
 chemical dosage is very important. In 
 addition to wasting money, adding too 
 much chemical can have other deleterious 
 effects. In the case of excess alkalin- 
 ity, an overdose can result in very high 
 pH values. An overdose of NaOCl can re- 
 sult in residual chlorine. An overdose 
 of KMn04 will result in more, not less. 
 
 CO CK 
 O LlI 
 
 o 
 
 X < 
 O CD 
 
 _l o 
 < o 
 
 LlI 
 Q. 
 
 I - 
 
 Initial NaOH. 
 
 UJ O 
 
 Initial NaOH: 
 
 ;— \ 
 
 'ON 
 
 o o 
 
 UJ o 
 
 nitial NaOH: 
 
 SERIES 1 
 
 SERIES 2 
 
 SERIES 3 
 
 FIGURE 4. - Chemical costs to reduce manganese 
 concentrations below 2 mg/L. 
 
47 
 
 manganese in solution. On the other 
 hand, using too little of any of the 
 three chemicals will result in discharg- 
 ing water that exceeds effluent limita- 
 tions for manganese. 
 
 Cost comparison of the three treatments 
 in each series of tests to reduce man- 
 ganese concentrations below 2 mg/L is 
 shown in figure 4. Excess alkalinity was 
 the least expensive method of manganese 
 removal, costing an average of $1 per 
 1,000 gal of water treated. Although 
 
 these chemical costs were less than half 
 of those for both sodium hypochlorite and 
 potassium permanganate, this method is 
 still quite expensive. At this site, the 
 chemical costs of the excess alkalinity 
 method to remove manganese were approxi- 
 mately twice the costs to treat the AMD 
 for neutralization and iron removal. 
 Elsewhere in these proceedings other AMD 
 treatment and abatement methods are pre- 
 sented. The in-line system, in particu- 
 lar, has shown the potential to be an in- 
 expensive method to remove manganese. 
 
 REFERENCES 
 
 1. Clark, J. W. , W. Viessman, Jr., and 
 M. J. Hammer. Water Supply and Pollution 
 Control. Harper and Row, 1977, pp. 444- 
 447. 
 
 2. Environmental Protection Agency. 
 Innovative and Alternative Technology 
 Assessment Manual. 1978, 443 pp. 
 
 3. . Onsite Wastewater Treatment 
 
 and Disposal Systems. Design Manual, 
 1980, 392 pp. 
 
 4. Evangelow, V. P. Controlling Iron 
 and Manganese in Sediment Ponds. Recla- 
 mation News and Views (Univ. KY) , v. 2, 
 No. 1, 1984, pp. 1-6. 
 
 5. Hood, W. C. , and S. M. Stepusin. 
 Manganese Content of Some Southern Illi- 
 nois Shales and Its Relation to Acid Mine 
 Drainage Problems. Abstract in Program 
 and Abstracts, Clay Mineral Conference. 
 Clay Miner. Soc. Axinu. Meeting, Cleve- 
 land, OH, Oct. 5-10, 1974, p. 35. 
 
 6. Kim, A. G. , B. S. Heisey, R. L. P. 
 Kleinmann and M, Deul. Acid Mine Drain- 
 age: Control and Abatement Research. 
 BuiMines IC 8905, 1982, 22 pp. 
 
 7. Marshall, K. C. Biogeocheraistry 
 of Manganese Minerals. Ch. in Biogeo- 
 chemlcal Cycling of Mineral-Forming Ele- 
 ments, ed, by P. A. Trudinger and D, J. 
 Swaine. Elsevier, 1979, pp. 253-292. 
 
 8. Nicholas, G. D. , and E. G. Foree. 
 Chemical Treatment of Mine Drainage for 
 Removal of Manganese to Permissible Lim- 
 its. Paper in Proceedings, 1979 Sym- 
 posium on Surface Mining, Hydrology, Sed- 
 imentology, and Reclamation, Lexington, 
 KY, Dec. 4-7, 1979, ed. by S. B. Carpen- 
 ter. Univ. KY, 1979, pp. 181-187. 
 
 9. Patterson, J, W. Wastewater 
 Treatment Technology. Ann Arbor Science, 
 1975, 265 pp. 
 
 10. Rozelle, R. B., and H. A. Swain, 
 Jr. Removal of Manganese From Mine 
 Drainage by Ozone and Chlorine. EPA 
 Technol. Ser. EPA 670/2-75-006, 1975, 
 47 pp. 
 
 11. U.S. Code of Federal Regulations. 
 Title 30 — Mineral Resources; Chapter 
 VII — Office of Surface Mining Reclamation 
 and Enforcement, Department of the Inte- 
 rior; Subchapter B — General Performance 
 Standards; Part 715 — General Performance 
 Standards. July 1, 1981. 
 
48 
 
 TREATMENT OF ACID MINE WATER BY WETLANDS 
 By Robert L, P. Kleinmann^ 
 
 INTRODUCTION 
 
 Wetlands are a potential natural treat- 
 ment system for small flows of acid mine 
 water. Previous studies of mine water 
 flowing through bogs dominated by Sphag- 
 ntjm moss indicate that such a wetland 
 removes the iron and reduces acidity, 
 without harm to the moss. A group from 
 Wright State University studied a site in 
 the Powelson Wildlife area in Ohio where 
 Sphagnum recurvum was found growing in pH 
 2.5 water. Iron, magnesium, sulfate, 
 calcixim, and manganese all decreased, 
 while pH increased from 2.5 to 4.6 as the 
 water flowed through the bog. A natural 
 outcrop of limestone located at the down- 
 stream end provided sufficient neutrali- 
 zation to raise the effluent pH to be- 
 tween 6 and 7 (4_) . ^ 
 
 A similar study was conducted by a West 
 Virginia University group at Tub Run Bog 
 in northern West Virginia (_5 ) . They 
 found that acid drainage flowing into the 
 wetland area rapidly improved in quality. 
 In 20 to 50 m, pH rose from 3.05-3.55 to 
 5.45-6.05, while only 10 to 20 m of flow 
 through the bog was needed to reduce 
 sulfate concentrations from 210-275 mg/L 
 to 5-15 mg/L and iron from 26-73 mg/L to 
 less than 2 mg/L. Overall, they found 
 that the water quality of the bog ef- 
 fluent was equal or superior to that 
 of nearby streams unaffected by mine 
 drainage. 
 
 In laboratory experiments it has been 
 shown that 1 kg (wet weight) of S. 
 recurvum can remove up to 92 pet of the 
 influent 50 mg/L of iron in 16.5 L of pH 
 3.8 synthetic mine water solution (3) by 
 cation exchange. In a natural wetland, 
 bacterial oxidation and sulfate reduction 
 in the organic-rich bottom waters add to 
 the iron removal capability. It has also 
 been demonstrated in the laboratory that 
 S. recurvum can tolerate acid mine drain- 
 age with iron concentrations as high as 
 500 mg/L for 4 weeks. Although the moss 
 was stressed, iron removal by cation ex- 
 change continued. In the field, higher 
 evapotranspiration rates and less ideal 
 conditions result in a long-term thresh- 
 old of less than 150 mg/L. 
 
 Such field observations and laboratory 
 studies suggest that a Sphagnum- dominated 
 biological treatment system is feasible. 
 Since discharge from such a biological 
 treatment system will not meet Federal 
 and State pH limitations (pH 6-9) for 
 mine water discharges, it was decided to 
 incorporate a passive limestone neutrali- 
 zation step down-gradient of the moss to 
 raise the pH to at least 6.0. Normally, 
 limestone in mine water would be rendered 
 useless by Fe(0H)3 precipitation, but 
 efficient iron removal by the wetland 
 would eliminate this problem. 
 
 PILOT-SCALE EVALUATION OF THE BOG-LIMESTONE SYSTEM 
 
 The Bureau of Mines decided that a 
 pilot-scale field test was needed to de- 
 termine if a bog system could be con- 
 structed to treat acid mine water. In 
 
 'Research supervisor, Pittsburgh Re- 
 search Center, Bureau of Mines, Pitts- 
 burgh, PA. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 September 1981, a contract was initiated 
 with Peer Consultants and Wright State 
 University (subcontractor) to construct a 
 pilot-scale test facility at an actual 
 mine drainage site. 
 
 A six-section plexiglass tank was con- 
 structed and mounted on a steel flat-bed 
 trailer. Live Sphagnum moss was harvest- 
 ed from the previously studied bog in the 
 Powelson Wildlife area and transplanted 
 
49 
 
 into the plexiglass chamber, which was 
 then towed to an acid mine drainage site 
 in the Zaleski State Forest in south- 
 eastern Ohio. A sketch of the portable 
 bog system (fig. 1) shows how water flows 
 through the divided chambers sequential- 
 ly. The first five sections were packed 
 loosely with Sphagnum moss (a mixture of 
 S. recurvum v. brevifolium and S. fimbri- 
 atum ) , while the last section was packed 
 with coarsely crushed limestone. Water 
 samples were collected at the intake, at 
 the end of the Sphagnum moss, and at the 
 outlet. The limestone was also analyzed 
 periodically. 
 
 The acid source water for the bog sys- 
 tem was an adjacent stream badly contam- 
 inated by acid mine drainage. Water was 
 already being pumped from the stream by 
 the U.S. Geological Survey sampling sta- 
 tion at the site. A portion of this 
 pumped water was used for our project. 
 Flow rates through the bog during the 
 initial 8 weeks of the test (June- July 
 1982) ranged from 1.4 to 19.8 gal/h owing 
 to problems with the pumping equipment 
 and inundation of the bog by heavy rain- 
 fall. This was subsequently stabilized 
 by increasing the diameter of the inlet 
 tube and inclining the inlet side of the 
 trailer 1.8 in above the outlet side, 
 simulating the natural gradient observed 
 at the bog in the Powelson Wildlife area. 
 A flow rate of approximately 2 gal/h was 
 used during August and September 1982; 
 after September, flow was Increased to 
 approximately 18 gal/h, and then to about 
 25 gal/h during 1983. 
 
 Although at times under stress due to 
 inundation, the Sphagnum moss remained 
 
 Inlet 
 
 Outlet 
 
 -3.66 m (12')- 
 
 ^ 
 
 Sphagnum moss 
 
 i?)- 
 
 Sphagnum moss 
 
 ==) 
 
 <= 
 
 Sphagnum moss 
 
 S-^: 
 
 1.93 m 
 (6' 4") 
 
 FIGURE 1. - Flow path of acid mine water 
 through the bog-limestone system. 
 
 viable throughout the test. Iron was re- 
 moved from the acid water by the moss so 
 that only minor amounts of visible ferric 
 hydroxide coating occurred on the lime- 
 stone. Chemical analysis (table 1) con- 
 firmed that some coating occurred, but 
 the effect on neutralization was insig- 
 nificant. Aluminum concentrations, which 
 are not significantly affected by the 
 Sphagnum moss, may prove to be a problem 
 if it turns out that aluminum hydroxide 
 floe armors the limestone. 
 
 Dissolved oxygen concentrations indi- 
 cate that anaerobic conditions did not 
 occur, even at the bottom of the moss 
 mat. Sulfate concentrations were not af- 
 fected by flow through the bog system, 
 and tests for hydrogen sulfide confirm 
 that little if any sulfate reduction was 
 occurring, presumably owing to the rela- 
 tively shallow depth (6 in) of the porta- 
 ble bog. Sulfate reduction is an im- 
 portant aspect of acid drainage treatment 
 
 TABLE 1 
 
 Results of analysis of limestone samples 
 
 Length of 
 
 Concentration, mg/L 
 
 exposure to 
 AMD, weeks 
 
 Iron 
 
 Manganese 
 
 Aluminum 
 
 Calcium 
 
 Magnesium 
 
 Unexposed 
 
 1 
 
 1,200 
 1,244 
 1,520 
 1,538 
 1,560 
 1,751 
 1,784 
 1,824 
 
 87.8 
 76.0 
 79.7 
 83.2 
 
 103 
 
 116 
 
 120 
 
 128 
 
 1,080 
 722 
 643 
 1,050 
 1,533 
 1,739 
 1,526 
 4,055 
 
 194,000 
 204,000 
 207,000 
 189,700 
 200,630 
 202,130 
 204,100 
 201,113 
 
 98,500 
 129.000 
 
 3 
 
 5 
 
 128,000 
 142,650 
 
 13 
 
 16 
 
 19 
 
 108,770 
 104,950 
 103,248 
 
 23 
 
 101,128 
 
50 
 
 by a natural bog (_5) ; its general absence 
 in our pilot-scale test implies that 
 our iron removal rates are probably 
 conservative. 
 
 Figure 2 shows the effect of the Sphag- 
 num moss on ferrous iron concentrations 
 after the flooding problem was corrected. 
 Ferrous iron oxidation averaged 61 pet 
 and peaked at 97 pet. Total iron con- 
 centrations, which include suspended 
 Fe(0H)3 floe, were very erratic, with 
 influent concentrations ranging from 15.9 
 to 640 mg/L within a week's time. These 
 fluctuations reflect resuspension of 
 Fe(0H)3 floe from the stream bottom dur- 
 ing storms; our small bog did not have 
 the detention time to filter out this 
 floe well, although presumably a larger 
 bog would. The Sphagnum bed typical- 
 ly removed 50 to 70 pet of the total 
 iron. 
 
 70 
 
 KEY 
 Inlet 
 
 ^ ^ End of Sphagnum 
 
 o o Outlet 
 
 November December 
 
 FIGURE 2.- Effect of the Sphagnum moss and lime- 
 stone on Fe 2+ concentrations in acid mine water. 
 
 700 
 
 600 
 
 500 
 
 £ 
 
 to 
 
 < 
 
 >- 
 Q 
 
 o 400 
 
 o 
 
 o 
 
 o 
 
 300- 
 
 200- 
 
 100- 
 
 
 
 1 1 11 1 1 
 KEY r\ 
 
 
 • • iniei / \ 
 
 o— o Outlet / \ 
 
 
 Bog flooded* / V * \ 
 
 — 
 
 'A 
 
 n 
 
 / 
 
 / Flow rate y \ \ 
 
 ~~ 
 
 1 
 
 V 
 
 j p^s increased — ^ 
 
 II i< 
 
 Inlet line \ W 
 clogged \ \* /'^ 
 
 1 II 
 
 
 July August September October November December 
 
 FIGURE 3. - Effect of the bog-limestone system on the titratable acidity of acid mine water. 
 
 June 
 14-30 
 
51 
 
 Acidity was not significantly affected 
 by flow through the Sphagnum mat , but de- 
 creased 43 to 90 pet as the water passed 
 through the coarsely crushed limestone 
 (fig. 3). The 90-pct reduction in acid- 
 ity was observed when the initial acidity 
 of the influent water exceeded 605 mg/L 
 (as CaC03); the 43-pct reduction was ob- 
 served when acidity at the inlet was less 
 than 150 mg/L. 
 
 Generally, pH increased as acidity de- 
 creased. Adsorption of the H"^ ion, al- 
 though known to be significant in a 
 natural bog (3), did not occur enough in 
 our small system to raise the pH as it 
 flowed through the Sphagnum moss. How- 
 ever, as the water flowed through the 
 limestone bed, pH increased an average of 
 1.4 and as much as 2.5 units. 
 
 A reduction of 88 pet in the ferrous 
 iron concentration in the water was 
 achieved in the moss bed. Initially it 
 was observed that virtually all of the 
 Fe^"*" reduction occurred as the water 
 passed through the first two chambers 
 containing 24 linear feet (16.5 ft^) of 
 the moss. During the final month of 
 sampling, after the monitoring sites in 
 the portable bog had been changed, this 
 reduction in Fe^"*" was found to actually 
 occur after the water has passed through 
 only one chamber of 12 linear feet (8.3 
 ft-') of moss. For the entire bog system, 
 at an average flow of 22 gal/h, levels of 
 Fe^"*" were reduced by 15 mg/L on average 
 at a rate of 5.5 mg/(L»h) or 1.8 mg/L 
 per cubic foot of moss. The removal rate 
 in the first, chamber was of course much 
 higher. 
 
 FULL-SCALE FIELD EVALUATION OF TREATMENT BY WETLANDS 
 
 The Bureau of Mines is now involved in 
 field evaluation of the wetland approach 
 at mine sites in Pennsylvania and West 
 Virginia. The wetlands have been con- 
 structed by the respective mining com- 
 panies for water treatment; the Bureau is 
 facilitating monitoring and evaluation of 
 the sites so that others can learn from 
 these efforts. Four wetland areas con- 
 structed during 1984 and two volunteer 
 wetland areas on mined lands are current- 
 ly being monitored; two additional sites 
 are planned for 1985. 
 
 At the volunteer wetland areas, C&K 
 Coal Co. is attempting to enhance already 
 established Typha bogs and to divert ad- 
 ditional mine water to the wetland areas 
 for treatment. At the better studied of 
 the two areas, flows range from 30 to 40 
 gal/min, with an influent pH of 5.5 to 
 5.8. Influent iron concentration aver- 
 ages 20 to 25 mg/L; manganese ranges from 
 30 to 40 mg/L. The velocity of the water 
 in the wetland ranges from 0.1 to 1.0 
 ft/s (as measured in less vegetated ar- 
 eas) over a 150-ft width with a total 
 length of about 85 ft. Effluent water 
 has less than 1 rag/L of iron, less than 2 
 mg/L manganese, and a near-neutral pH, 
 Manganese removal is attributed to bac- 
 terial activity (1-2). 
 
 With an understanding of wetlands 
 gained from the pilot-scale test and ob- 
 servation of the volunteer wetland areas, 
 wetland treatment systems have been con- 
 structed of Sphagnum alone, and of Sphag- 
 num and Typha together. The vegetation 
 was transplanted from nearby wetlands by 
 personnel of Brehm Laboratory of Wright 
 State University and by Ben Pesavento, of 
 Environment Analytic, who are also re- 
 sponsible for monthly monitoring and sam- 
 ple collection. These initial wetland 
 areas range in size from 750 to 8,500 
 ft^, of which 40 to 60 pet is actual wet- 
 ted area, and treat flows of 2-8 gal/min. 
 Preliminary results are shown in table 2 
 for the three wetland areas constructed 
 at least 2 months ago. In addition to 
 cation exchange, oxidation, and removal 
 as iron sulfides , these results may par- 
 tially reflect dilution of iron and man- 
 ganese in the bog by ground water. 
 
 It appears that wetlands can be con- 
 structed in acid mine water discharges 
 and that they will improve drainage 
 quality. They require continuous flow, 
 without a lot of variation; long-terra 
 maintenance requirements have yet to be 
 determined. They appear to be most ap- 
 propriate for relatively small flows 
 (less than 10 gal/min) owing to the large 
 
52 
 
 TABLE 2. - Performance of wetlands 2 months after construction 
 or augmentation, milligrams per liter 
 
 Mine site 
 
 Iron 
 
 Manganese 
 
 
 Influent 
 
 Effluent 
 
 Influent 
 
 Effluent 
 
 Mine 1................. 
 
 24 
 
 8.7 
 24 
 
 0.5 
 
 1.2 
 
 .6 
 
 43.8 
 24.5 
 16 
 
 16.1 
 
 Mine 2 
 
 15.5 
 
 Mine 3 
 
 3.8 
 
 surface area requirement — we like to al- 
 low 200 ft^ of wetted area per gallon per 
 minute of flow. However, only space lim- 
 its the extension of this system to 
 greater flows. An attempt will be made 
 
 to treat acid flows of 50 to 100 gal/min 
 in larger wetland systems, starting with 
 partial treatment in 1985 and, if suc- 
 cessful, followed by full-scale tests in 
 1986. 
 
 REFERENCES 
 
 1. Emerson, S, , S. Kalhorn, L. Jacobs, 
 B. M Tebo, K. H. Nealson, and R. A. Ros- 
 son. Environmental Oxidation Rate of 
 Manganese (II): Bacterial Catalysis. 
 Geochim. et Cosmochim. Acta, v. 46, 1982, 
 pp. 1073-1079. 
 
 2. Gregory, E., and J. T. Staley. 
 Widespread Distribution of Ability To 
 Oxidize Manganese Among Freshwater Bac- 
 teria. App. and Environ. Microbiol. , v. 
 44, No. 2, 1982, pp. 509-511. 
 
 3. Harris, R. L. , T. 0. Tiernan, 
 J. Hinders, J. G. Solch, B. E. Huntsman, 
 and M. L. Taylor. Treatment of Mine 
 Drainage From Abandoned Mines by Biologi- 
 cal Iron Oxidation and Limestone Neutral- 
 ization. Peer Consultants report pre- 
 pared for Bureau of Mines under contract 
 
 J0113033, 1984, 113 pp.; available from 
 Robert Kleinmann, BuMines , Pittsburgh, 
 PA. 
 
 4. Huntsman, B. E., J. G. Solch, and 
 M. D. Porter. Utilization of Sphagnum 
 Species Dominated Bog for Coal Acid Mine 
 Drainage Abatement. Geol. Soc. America 
 (91st Ann, Meeting) Abstracts, Toronto, 
 Ontario, Canada, 1978, pp. 322. 
 
 5. Wieder, R. K. , G. E, Lang, and 
 A. E. Whitehouse. Modification of Acid 
 Mine Drainage in a Fresh Water Wetland. 
 Paper in Proceedings , Acid Mine Drainage 
 Research and Development , 3d WV Surface 
 Mine Drainage Task Force Sjmiposium, WV 
 Surface Mine Drainage Task Force, 
 Charleston, WV, 1982, pp. 38-62. 
 
53 
 
 IN-LINE AERATION AND TREATMENT OF ACID MINE DRAINAGE: PERFORMANCE 
 AND PRELIMINARY DESIGN CRITERIA 
 
 Bv Terry Ackman^ and Robert L. P. Kleinniann^ 
 
 INTRODUCTION 
 
 It is estimated that the U.S. coal 
 mining industry spends over $1 million 
 per day treating acidic mine water so 
 that it can be legally discharged ( 4_) . ^ 
 This figure includes the amortized cost 
 of the large water treatment plants (a 
 conventional lime neutralization facility 
 typically costs over $1 million to 
 construct), treatment chemicals (lime, 
 soda ash, sodium hydroxide, flocculant, 
 etc.), maintenance, electric power, and 
 labor. 
 
 Although expensive, conventional acid 
 mine drainage (AMD) treatment is a simple 
 process. The water is neutralized, typi- 
 cally to a pH of 8 to 9, and then aerated 
 to oxidize the iron to the Fe-^"*" state, 
 causing precipitation of Fe(0H)3 (Yellow- 
 boy) sludge. The water is then separated 
 from the sludge in a series of settling 
 basins or ponds and discharged. 
 
 Above a pH of 3.5, the rate of iron ox- 
 idation is controlled by dissolved oxygen 
 (DO) and pH. Fully aerated mine water 
 contains about 8 mg/L DO, which is con- 
 sumed at the rate of 1 mg/L for every 7 
 mg/L Fe"*"^ oxidized; consequently, the 
 DO initially present can only oxidize 
 50 to 60 mg/L Fe^"^ ^]J • ^^ °^^ assumes, 
 though, that DO is not depleted but in- 
 stead is maintained at a constant level 
 by continuous aeration, the effect of pH 
 on the rate of iron oxidation can be 
 calculated. Table 1 illustrates the 
 effect of pH on the required aeration 
 time for an initial Fe^"*" concentration of 
 100 mg/L. Inspection of the reaction 
 times listed in table 1 reveals why pH is 
 
 Mining engineer. 
 
 ^Research supervisor. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 raised to 7,5 or above at most treatment 
 plants to quickly oxidize the ferrous 
 iron, 
 
 TABLE 1. - Time required to oxidize 97 
 pet of 100 mg/L Fe^"*" at various con- 
 stant pH's, and constant oxygen sat- 
 uration (8 mg/L DO) 
 
 pH 
 
 Time, h 
 
 4,5 3.5 
 
 5 3.5 
 
 5.5 3.5 
 
 6 3.5 
 
 6.5 3.5 
 
 7 3,5 
 
 7.5 3,5 
 
 8 3,5 
 
 8,5 3,5 
 
 10^ 
 10^ 
 
 102 
 
 IQl 
 
 10-' 
 10-2 
 10-3 
 10-^ 
 
 For replenishment of DO in mine water, 
 settling ponds or lagoons are constructed 
 wide and shallow to maximize diffusion of 
 oxygen into the water and thereby in- 
 crease oxygen transfer from the atmos- 
 phere. However, oxygen diffusion is rel- 
 atively slow (9^) , so that at many sites 
 supplementary aeration sources are neces- 
 sary (8^), For example, oxygen transfer 
 can be increased by increasing turbu- 
 lence. This is typically accomplished by 
 incorporating a series of open-channel 
 drops in the flow path of the water. 
 Mechanical aerators can also be used to 
 continuously introduce bubbles of air in- 
 to the water. This continuous replenish- 
 ment of DO is effective in maintaining a 
 rapid reaction rate, but it also has dis- 
 advantages: a separate aeration tank or 
 basin is required; there are high initial 
 capital costs; and there are operating 
 costs associated with power consumption 
 and maintenance, especially where gypsum 
 precipitation is a problem. 
 
54 
 
 This report describes a Bureau of 
 Mines-designed treatment system that has 
 been tested at mine sites in Pennsylvania 
 and West Virginia. The in-line aeration 
 and treatment system (ILS) functions in 
 existing AMD pipelines, using energy pro- 
 vided by existing mine water discharge 
 
 pumps. It appears to be a low-cost al- 
 ternative to conventional treatment 
 plants and, in fact, appears to accel- 
 erate iron oxidation rates. The system 
 has no moving parts and thus has the ad- 
 vantages of low maintenance and operating 
 costs. 
 
 UNIT DESCRIPTION 
 
 The ILS consists of two off-the-shelf 
 components: a jet pump O) and a static 
 mixer. Both components can be described 
 as aeration and mixing devices. Jet 
 pvimps are simply nozzles that entrain air 
 by Venturi action (fig. 1). The jet pump 
 used is made of polyvinyl chloride (PVC). 
 Water enters under pressure and is con- 
 verted by the jet pump into a high- 
 velocity stream. This stream then passes 
 through a suction chamber, which is open 
 to the atmosphere. If the system is be- 
 ing used for neutralization as well as 
 
 aeration, the suction chamber also serves 
 as the injection point for the neutraliz- 
 ing material. Multiple jet pump units 
 may be placed in parallel as long as wa- 
 ter pressures of at least 20 psi are 
 maintained. 
 
 After passing through the jet pump, the 
 flow enters the static mixer (fig. 2). 
 The static mixer consists of 1-ft sec- 
 tions of pipe made of copolymer poly- 
 propylene resins, laminated together end 
 to end with fiberglass. Inside each 
 
 Suction chamber 
 
 Nozzle 
 
 Parallel section 
 
 \ \ V \ \ \ \ X \ v ^ \ \^ 
 
 T 
 
 \\\\\s \\xs 
 
 Diffuser 
 
 Suction 
 
 FIGURE 1. - Jet pump diagra 
 
 m. 
 
 Static mixer 
 
 Flow 
 
 •. • • . 
 
 FIGURE 2. - Diagram of the static mixer. Air bubbles are reduced in size by the turbulence, 
 significantly increasing interfacial contact. 
 
55 
 
 section is a 1-ft helix that forces the 
 water to follow a spiral path. Static 
 mixers are used routinely in sewage and 
 industrial waste water treatment plants 
 as vertical airlift aeration and mixing 
 units, but that design was modified some- 
 what for this horizontal application: 
 
 each helical unit was rotationally offset 
 90° from its neighbor, thereby interrupt- 
 ing the corkscrew every foot and enhanc- 
 ing the mixing action. Eight 1-ft sec- 
 tions were used, which provided the 
 contact time of a normal 32-ft pipe be- 
 cause of the induced spiral flow. 
 
 PERFORMANCE CHARACTERISTICS 
 
 AERATION OF NEAR-NEUTRAL MINE WATERS 
 
 The ILS was first tested as an aeration 
 unit at a mine site in Greene County, PA. 
 Influent Fe^"^ levels were erratic but 
 often exceeded 100 mg/L at near-neutral 
 pH. As an alternative to mechanical 
 aeration, the ILS was installed at the 
 end of the discharge pipe from the under- 
 ground mine. 
 
 Monitoring the discharge from the site 
 began on the fourth day after installa- 
 tion of the ILS. Ferrous iron concentra- 
 tions dropped from 10 to 20 mg/L before 
 installation of the ILS to 0.2 to 0.9 
 mg/L. Total iron concentrations fell 
 from over 20 mg/L to less than 2 mg/L. 
 
 conducted at actual mine sites using 
 sodium hydroxide (NaOH) , quick lime 
 (CaO), or hydrated lime (Ca(0H)2), with 
 the latter two added as slurries. The 
 effluent pH was easily adjusted in each 
 case, and the violent mixing action of 
 the ILS minimized excessive lime use. 
 
 Tables 2-4 allow the comparison of ac- 
 tual NaOH or lime consvimption with theo- 
 retical "best case" neutralization. The 
 theoretical values are derived assuming 
 optimal efficiency (90 pet for CaO, 95 
 pet for Ca(0H)2, and 99 pet for NaOH) and 
 a pH endpoint of 8.3 ( 5^) ; our experience 
 indicates that conventional treatment 
 plants use 25 pet more lime than these 
 calculated values. 
 
 Subsequent aeration tests were conduct- 
 ed with more acidic water. Iron oxida- 
 tion continued to be impressive despite 
 an influent pH of 4.6 to 5.6. Figure 3 
 is a graph of average Fe^"*" values for all 
 samples of pH 5.5 ±0.2. Although very 
 little iron oxidation occurred in the 
 ILS, the discharge from the first pond 
 (24-h detention time) averaged only 6 
 mg/L Fe^"*". This represents not only much 
 greater iron oxidation than without the 
 ILS at this pH, but also a much faster 
 rate than expected in oxygen-saturated 
 water (table 1). A more detailed analy- 
 sis of this topic may be found in RI 8868 
 (2). 
 
 SIMULTANEOUS NEUTRALIZATION 
 AND AERATION 
 
 The suction port of the jet pumps can 
 be used for addition of neutralizing 
 chemicals without significantly interfer- 
 ing with air intake. Field tests were 
 
 150 r 
 
 d) 100 
 
 E 
 
 z 
 
 O 
 
 tr 
 
 CO 
 
 Z) 
 
 O 
 tr 
 
 LU 
 
 LL 
 
 50 
 
 IW^ 
 
 RAW 
 WATER 
 
 AFTER 
 POND 
 
 FIGURE 3. - Effect of the ILS as an aeration sys- 
 tem on average Fe 2' concentration at pH 5,5 :t0.2 at 
 the Greene County, PA, site. Pond has a 24-h de- 
 tention time. 
 
56 
 
 TABLE 2. - NaOH use at site 2--Braxton County, WV 
 
 Test 
 
 Raw 
 
 Net acidity 
 
 Flow, 
 
 Na in 
 
 Na in 
 
 Treated 
 
 Theoretical 
 
 Actual 
 
 run 
 
 pH 
 
 of raw 
 
 gal/mln 
 
 raw water. 
 
 treated 
 
 pH 
 
 NaOH use. 
 
 NaOH use, 
 
 
 
 water, mg/L 
 
 
 lag/L 
 
 water, mg/L 
 
 
 Ib/min 
 
 Ib/min 
 
 SINGLE TREATMENT 
 
 1... 
 
 3.2 
 
 3,784 
 
 521 
 
 22 
 
 1,243 
 
 5.1 
 
 13.4 
 
 5.3 
 
 2... 
 
 3.3 
 
 3,951 
 
 469 
 
 22 
 
 1,216 
 
 5.3 
 
 12.6 
 
 4.7 
 
 3... 
 
 3.2 
 
 3,784 
 
 543 
 
 23 
 
 1,000 
 
 5.2 
 
 14.0 
 
 4.4 
 
 4... 
 
 3.2 
 
 4,022 
 
 533 
 
 23 
 
 1,176 
 
 5.0 
 
 14.6 
 
 5.1 
 
 5... 
 
 2.7 
 
 3,689 
 
 385 
 
 27 
 
 1,634 
 
 6.6 
 
 9.7 
 
 5.2 
 
 6... 
 
 2.6 
 
 3,689 
 
 533 
 
 23 
 
 1,094 
 
 4.9 
 
 13.4 
 
 4.8 
 
 7... 
 
 2.5 
 
 3,713 
 
 530 
 
 23 
 
 1,209 
 
 4.9 
 
 12.9 
 
 5.2 
 
 8... 
 
 2.8 
 
 3,677 
 
 261 
 
 35 
 
 1,779 
 
 6.8 
 
 6.5 
 
 3.8 
 
 9... 
 
 2.9 
 
 3,641 
 
 345 
 
 35 
 
 3,558 
 
 12.8 
 
 8.5 
 
 10.1 
 
 
 DOUBLE TREATMENT 
 
 10.. 
 
 4.8 
 
 75 
 
 543 
 
 1,860 
 
 1,865 
 
 8.4 
 
 0.3 
 
 0.03 
 
 11.. 
 
 4.6 
 
 89 
 
 475 
 
 1,831 
 
 2,193 
 
 11.3 
 
 .3 
 
 1.4 
 
 12.. 
 
 4.6 
 
 87 
 
 340 
 
 1,865 
 
 2,021 
 
 9.9 
 
 .2 
 
 .4 
 
 13.. 
 
 4.6 
 
 95 
 
 523 
 
 1,728 
 
 2,175 
 
 10.7 
 
 .4 
 
 1.9 
 
 14.. 
 
 4.6 
 
 71 
 
 475 
 
 1,514 
 
 2,153 
 
 10.6 
 
 .2 
 
 2.5 
 
 15.. 
 
 4.3 
 
 68 
 
 337 
 
 1,888 
 
 1,872 
 
 8.6 
 
 .1 
 
 .04 
 
 TABLE 3. - Lime use at site 3 Armstrong Country, PA 
 
 Sam- 
 
 Raw 
 
 Net acidity 
 
 Flow, 
 
 Ca in 
 
 Ca in 
 
 Treated 
 
 Theoretical 
 
 Actual 
 
 ple 
 
 pH 
 
 of raw 
 
 gal/min 
 
 raw water. 
 
 treated 
 
 pH 
 
 lime use. 
 
 lime use. 
 
 
 
 water, mg/L 
 
 
 mg/L 
 
 water, mg/L 
 
 
 Ib/min 
 
 Ib/min 
 
 1... 
 
 2.7 
 
 830 
 
 363 
 
 284 
 
 1,078 
 
 11.7 
 
 1.7 
 
 4.4 
 
 2... 
 
 3.0 
 
 753 
 
 363 
 
 271 
 
 268 
 
 3.1 
 
 1.5 
 
 .0 
 
 3... 
 
 3.0 
 
 830 
 
 363 
 
 277 
 
 618 
 
 7.3 
 
 1.7 
 
 1.9 
 
 4... 
 
 3.0 
 
 791 
 
 363 
 
 279 
 
 671 
 
 8.8 
 
 1.6 
 
 2.2 
 
 5... 
 
 2.9 
 
 830 
 
 363 
 
 287 
 
 608 
 
 5.7 
 
 1.7 
 
 1.8 
 
 6... 
 
 N/A 
 
 791 
 
 363 
 
 280 
 
 692 
 
 8.8 
 
 1.6 
 
 2.3 
 
 7... 
 
 2.9 
 
 830 
 
 363 
 
 286 
 
 617 
 
 6.9 
 
 1.7 
 
 1.8 
 
 TABLE 4. - Lime use at site 4 Westmoreland County, PA 
 
 Sam- 
 ple 
 
 Raw 
 pH 
 
 Net acidity 
 
 of raw 
 water, mg/L 
 
 Flow, 
 gal/min 
 
 Ca in 
 
 raw water, 
 
 mg/L 
 
 Ca in 
 
 treated 
 
 water, mg/L 
 
 Treated 
 pH 
 
 Theoretical 
 
 lime use, 
 
 Ib/min 
 
 Actual 
 
 lime use, 
 
 Ib/min 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8l 
 
 9 
 
 Plant^ 
 
 5.6 
 5.7 
 5.4 
 5.4 
 5.6 
 5.5 
 5.5 
 5.4 
 5.4 
 
 4.8 
 
 973 
 
 877 
 1,010 
 1,040 
 
 942 
 1,012 
 1,062 
 
 986 
 1,018 
 
 1,280 
 
 469 
 457 
 457 
 469 
 542 
 485 
 485 
 485 
 485 
 
 1,450 
 
 445 
 454 
 424 
 419 
 451 
 425 
 421 
 420 
 405 
 
 421 
 
 1,015 
 1,020 
 1,057 
 1,164 
 
 749.8 
 
 901 
 
 909 
 1,018 
 
 948 
 
 1,081 
 
 8.4 
 7.7 
 7.0 
 6.9 
 6.6 
 7.0 
 6.9 
 7.1 
 7.0 
 
 8.2 
 
 2.6 
 2.3 
 2.6 
 2.8 
 2.9 
 2.8 
 2.9 
 2.7 
 2.8 
 
 10.5 
 
 4.1 
 4.0 
 4.5 
 5.4 
 2.5 
 3.5 
 3.3 
 4.5 
 4.1 
 
 ^14. 7 
 419.1 
 
 ^Fe and Mn in filtered samples were within effluent standards, 
 ^Normal plant operation. 
 ^Measured by chemical analysis. 
 '^Physically measured dry feed. 
 
57 
 
 Table 2 represents a two-stage process, 
 usig NaOH to treat mine water with high 
 acidity and high iron. Samples 1 through 
 9 represent a single treatment pass 
 through the ILS from pond 1 to pond 2 
 (initially empty before the test). Sam- 
 ples 10 through 15 represent water pumped 
 from pond 2 through the ILS to pond 3 36 
 h after the first treatment. Effluent 
 water from the two-stage treatment met 
 effluent standards. Actual NaOH usage 
 was calculated from the difference in 
 sodium concentrations in unfiltered, 
 acidified samples of treated and raw 
 water. Theoretical NaOH requirement was 
 calculated by Lovell's equations (_5 ) . 
 NaOH use was approximately half of that 
 theoretically required. However, iron 
 was precipitated as both Fe(0H)3 and 
 Fe(0H)2 in the first step of the treat- 
 ment. As explained later, Fe(0H)2 will 
 eventually oxidize, adding acidity to the 
 pond water. 
 
 Table 3 summarizes the results of a 
 field test using Ca(0H)2. This operation 
 did not allow a quantitative comparison 
 with actual consumption of lime by the 
 conventional water treatment plant, but 
 the plant operator felt that lime usage 
 was reduced enough to design an ILS to 
 replace the existing system. Except at 
 high pH (sample 1), the ILS values met 
 discharge criteria and approached the 
 theoretical optimal values for lime con- 
 sumption. As discussed later, both iron 
 and manganese were reduced to effluent 
 levels at a discharge pH as low as 6.9, 
 indicating that greater potential cost 
 savings can be obtained. 
 
 Table 4 presents the results of a field 
 test using CaO slurry to neutralize water 
 being pumped from an underground mine 
 pool. Owing to the high levels of dis- 
 solved iron (over 500 mg/L) , the ILS unit 
 could not oxidize all of the iron in a 
 single pass; as at the NaOH site (table 
 2) , some of the iron precipitated as 
 Fe(0H)2. Water sample 8, which met dis- 
 charge standards after filtration, can be 
 used for comparing actual costs with 
 those for operation of the conventional 
 treatment plant (table 4, bottom row). 
 Since flow through the ILS is one-third 
 
 that of normal plant operation, the ob- 
 served lime use of 4.5 Ib/min at pH 7.1 
 must be scaled up to 13.5 Ib/min. This 
 is within 1 pet of the amount of lime 
 consumed in neutralizing acidity during 
 operation of the conventional treatment 
 plant (as calculated from chemical analy- 
 sis) but is 30 pet more efficient than 
 actual lime use, as measured during nor- 
 mal plant operation. Analysis of the 
 sludge during operation of the conven- 
 tional treatment plant confirms that a 
 lot of unreacted lime is being wasted, 
 especially in the aeration basin, owing 
 to insufficient mixing action, 
 
 IRON OXIDATION 
 
 During field testing of the ILS, it be- 
 came apparent that iron oxidation was 
 proceeding much faster thdn anticipated. 
 At low pH (4.6 to 5.5), iron oxidation 
 was accelerated by a factor of 10 to 400; 
 at near-neutral pH (6.9 to 7.5), iron ox- 
 idation was accelerated by as much as 
 1,000 (2^). Figure 3 illustrates iron ox- 
 idation at the Greene County, PA, test 
 site; 98.7 pet of the 190 mg/L Fe^"^ in 
 the influent water was oxidized in the 
 4-8 transit time in the ILS. Most of 
 this oxidation apparently occurred in the 
 jet pump section of the ILS since water 
 samples collected between the jet pump 
 and the static mixer had an average pH of 
 6.7 and an Fe^"*" concentration of only 4.8 
 mg/L. To obtain such rapid iron oxida- 
 tion in a conventional water treatment 
 system, the pH would have to be raised to 
 at least 8.5. 
 
 However, the iron oxidation capacity of 
 the existing ILS design is limited. As 
 influent Fe^"*" concentrations approach 300 
 mg/L, the efficiency of the system de- 
 creases. Tables 5 and 6 document field 
 trials with average influent Fe^"*" concen- 
 trations of 965 and 527 mg/L, respective- 
 ly. The amount of Fe^''" oxidized during 
 transit through the ILS ranged between 
 283 and 479 mg/L using NaOH (table 5) and 
 between 163 and 345 rag/L using CaO (table 
 6). The amount of Fe^"*" oxidized was cal- 
 culated as the difference between Fe^"*" 
 concentrations in acidified, unfiltered 
 samples of raw and treated water. The 
 
58 
 
 
 
 
 
 0^ ^ 
 
 ^ o o> in CM 
 
 O 1-H 
 
 1—1 
 
 vO 
 
 
 1—1 
 
 
 
 
 
 
 
 
 
 
 
 
 (3 
 
 
 ^ CT> 
 
 CM O . —H .— I 
 
 o • 
 
 o 
 
 . 
 
 
 CM 
 
 
 
 
 
 
 
 
 
 
 3 
 
 in 
 
 
 t— ( 1—1 
 
 t-H O 
 
 • 
 
 00 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 .—1 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O O 
 
 00 O CM 00 CM 
 
 o o 
 
 o 
 
 vD 
 
 
 <f 
 
 O 
 
 
 
 
 
 
 c 
 
 
 
 . O . 1-H 1-H 
 
 o 
 
 
 • 
 
 
 
 • 
 
 
 « — 1 
 
 o 
 
 CTi CM 
 
 CM r^ in i-H in CO 
 
 CM 
 
 o in 
 
 
 
 4-) 
 
 3 
 
 <f 
 
 
 CO -— 1 CM 
 
 1—1 
 
 
 o 
 
 
 
 CO 
 
 
 
 CM 
 
 CO CM 
 
 as t . r^ . 
 
 1-H 
 
 • v£) 
 
 
 
 (3 
 0) 
 
 Pi 
 
 t— 1 
 
 
 
 
 
 f— 1 
 
 
 
 •H 
 CO 
 
 
 (3 
 3 
 
 CM 
 
 in 
 
 O -H 1-H 
 
 o 
 
 r^ in 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1— H 
 
 
 
 
 
 
 
 
 o> m 
 
 CO o r^ o CO 
 
 o ^ 
 
 t— 1 
 
 r^ 
 
 
 CO 
 
 >% 
 
 
 Pi 
 
 
 
 
 
 
 
 
 4J 
 
 CO 
 
 3 
 
 <r) 
 
 -— " 00 
 
 CM O • CM .-< 
 ^ CM 
 
 o . 
 
 o 
 
 • 
 
 o 
 
 
 CM 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CU 
 
 Pi 
 
 f—< 
 
 
 
 
 o 
 
 1—1 
 
 
 
 C 
 
 
 CTN 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 
 
 in 
 
 <f in 
 
 o >* CO 00 r^ in 
 o • . 00 • 
 
 1-H 
 
 rH CM 
 
 . in 
 
 
 
 
 
 o <j- 
 
 -^ o vD in CO 
 
 o ^ 
 
 r— 1 
 
 C3> 
 
 
 <* 
 
 
 C 
 
 
 
 
 c 
 
 
 Csl 00 
 
 CM O . — 1 
 
 o . 
 
 o 
 
 . 
 
 
 CM 
 
 (U 
 
 
 3 
 
 CM 
 
 in 
 
 — < O CM —1 
 
 . 
 
 r-« in 
 
 
 
 CU 
 
 3 
 
 OJ 
 
 
 -• o 
 
 -H O 
 
 • 
 
 a\ 
 
 
 
 )-l 
 
 
 Pi 
 
 
 
 1-H 
 
 o 
 
 
 
 
 
 Pi 
 
 r-H 
 
 
 
 
 o 
 
 
 
 
 o 
 
 M-l 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 s 
 
 
 
 o o 
 
 00 O — < CO <i- 
 
 o o 
 
 o 
 
 CO 
 
 
 <r 
 
 <U 
 
 
 00 
 
 
 
 
 
 
 
 
 o 
 
 (3 
 
 
 
 . O . CM -H 
 
 o 
 
 
 . 
 
 
 
 43 
 
 
 
 in 
 
 CM CM 
 
 CM CM in -H 00 -H 
 
 as 
 
 C3> <r 
 
 
 
 o 
 
 3 
 
 1— 1 
 
 
 CO ^^ CO 
 
 1—1 
 
 
 1— < 
 
 
 
 
 
 (3 
 
 <!■ 
 
 ^ ^H 
 
 Os .1-H . \D . 
 
 1-H 
 
 . m 
 
 
 
 
 Pi 
 
 i-H 
 
 
 
 
 
 1—1 
 
 
 
 bO 
 
 c 
 
 •H 
 
 
 ^ 
 
 CO 
 
 in 
 
 CM O CM 
 
 o 
 
 \£) in 
 
 
 
 
 
 
 cNi <r 
 
 -^ O -d- 00 »x> 
 
 o ^ 
 
 1—1 
 
 <f 
 
 
 ■vT 
 
 
 
 
 
 
 
 
 
 (3 
 
 
 —1 CO 
 
 CO o . in CM 
 
 o • 
 
 o 
 
 • 
 
 
 CO 
 
 I-) 
 
 
 ^~~ 
 
 
 
 
 
 
 
 
 
 3 
 
 O 
 
 
 —1 in 
 
 — 1 o 
 
 . 
 
 00 
 
 
 
 4J 
 
 
 ^>- 
 
 
 
 
 
 
 
 
 
 Pi 
 
 I— 1 
 
 
 ^H 
 
 
 o 
 
 
 
 
 4J 
 0) 
 
 
 (3 
 
 CO 
 
 CO <!■ 
 <r cTi 
 
 CTn — 1 vO CM CM <r 
 00 ... o . 
 
 CO 
 
 1-H 
 
 O CO 
 
 . in 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 in CM 
 
 O O CO v£> CM 
 
 O CTv 
 
 CM 
 
 00 
 
 
 o 
 
 CO 
 
 <: 
 
 3 
 
 CM 
 
 <)■ 
 
 O O C3^ -H 
 
 . 
 
 r^ m 
 
 
 
 
 i3 
 
 
 cj <r 
 
 o o • r^ 
 
 o . 
 
 . 
 
 • 
 
 
 o 
 
 
 pj 
 
 Pi 
 
 
 
 1 1 
 
 o 
 
 
 
 > 
 
 
 ^ 
 
 CTi 
 
 <r 
 
 o -H <r 
 
 1—1 1—1 
 
 o 
 
 CM 
 
 
 o 
 
 <u 
 
 
 
 
 
 
 
 
 
 ^s 
 
 
 
 
 " 
 
 
 
 I— 1 
 
 
 ^ 
 
 4-1 
 
 «^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r-i 
 
 
 
 
 
 1-H 
 
 CO 
 
 4-1 
 
 vO 
 
 00 
 
 o r- 
 
 CM CM —< CM in in 
 
 CO 
 
 vO <i- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 >^ 
 
 
 
 
 O 00 
 
 O O v£) ^ 1^ 
 
 in CO 
 
 1-H 
 
 00 
 
 
 o 
 
 3 
 
 (3 
 
 C 
 
 o 
 
 <i- r-- 
 
 r^* . . 1 — 1 . 
 
 1-H 
 
 . CM 
 
 
 4-) 
 
 
 C 
 
 
 00 CO 
 
 O O • 1 vO 
 
 CTi . 
 
 . 
 
 • 
 
 
 o 
 
 d 
 
 3 
 
 3 
 
 CM 
 
 CO 
 
 O CM ^ 
 
 . 
 
 vo m 
 
 
 (3 
 
 
 3 
 
 00 
 
 CO 
 
 O — 1 O 
 
 1— H 
 
 o 
 
 vO 
 
 
 o 
 
 •H 
 
 o 
 
 Pi 
 
 
 
 
 o 
 
 
 
 S 
 
 
 Pi 
 
 
 
 1 
 
 
 
 
 
 i—H 
 
 CO 
 
 u 
 
 
 
 
 
 
 
 
 O 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 -d 
 
 m 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ in 
 
 <r CM CO in o 
 
 <t CM 
 
 CO 
 
 CJ^ 
 
 
 o 
 
 4J 
 
 c 
 
 
 CO 
 
 1-1 in 
 
 CJ^ CO cyi r- v£> O 
 
 CJN 
 
 ON CM 
 
 
 c 
 
 
 (3 
 
 
 in <r 
 
 CM 00 • • ^ 
 
 1—1 . 
 
 . 
 
 • 
 
 
 o 
 
 
 CO 
 
 C 
 
 vO 
 
 CO r-« 
 
 00 • . vO . 
 
 o 
 
 . CO 
 
 
 o 
 
 
 3 
 
 t^ 
 
 <t 
 
 00 o o 
 
 CO 
 
 o 
 
 -* 
 
 
 o 
 
 CO 
 
 i-f 
 
 3 
 
 1—1 
 
 <f 
 
 r-> as T-{ 
 
 . 
 
 o in 
 
 
 4J 
 
 4J 
 
 Pi 
 
 
 
 
 
 
 
 
 
 >^ 
 
 (U 
 
 Pd 
 
 
 
 
 o 
 
 
 
 XI 
 CO 
 
 (3 
 
 5J 
 
 
 
 
 
 
 
 
 
 1—1 
 
 CO 
 X) 
 
 ^1 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ' 
 
 
 
 
 
 
 
 
 
 M 
 
 
 
 
 
 vo r-- 
 
 CO CO CM C^^ CM 
 
 CO vO 
 
 CM 
 
 <J\ 
 
 
 o 
 
 
 d 
 
 -* 
 
 
 
 
 
 
 
 CQ 
 
 4J 
 
 (3 
 
 
 vO CO 
 
 cvj r^ • 1 
 
 . 
 
 . 
 
 9 
 
 
 o 
 
 1-H 
 
 4-) 
 
 
 CO 
 
 vO CTN 
 
 CM CM ^ CJN 00 CM 
 
 1— 1 
 
 O CM 
 
 
 I 
 
 CO 
 
 3 
 
 vO 
 
 CO 
 
 r^ ^^ 
 
 CM 
 
 o 
 
 <3- 
 
 
 o 
 
 1-H 
 
 CO 
 
 C 
 
 as 
 
 CO 00 
 
 as . . r^ . 
 
 1-H 
 
 • CO 
 
 
 1 
 
 (U 
 
 Pi 
 
 
 
 1 
 
 
 
 
 
 ^ 
 
 
 CU 
 
 3 
 
 1-H 
 
 <f 
 
 O O -H 
 
 . 
 
 r^ m 
 
 
 CM 
 
 u 
 
 
 
 
 
 
 
 
 
 1-H 
 
 
 1 
 
 Pi 
 
 
 
 1-H 
 
 o 
 
 
 
 
 
 
 CTs 00 
 
 o o o o r^ 
 
 <f -* 
 
 CM 
 
 ^O 
 
 
 o 
 
 
 cu 
 
 
 
 
 
 
 
 
 u 
 
 d) 
 
 (3 
 
 
 r- <t 
 
 O O vO 
 
 a\ • 
 
 • 
 
 • 
 
 
 o 
 
 •H 
 
 <!• 
 
 CO 
 
 
 
 
 
 
 
 •H 
 
 rH 
 
 3 
 
 m 
 
 <f 
 
 O -H 
 
 CM 
 
 o 
 
 vC 
 
 
 o 
 
 CO 
 
 
 
 00 
 
 r-. o 
 
 CTn vO -vf CO o <* 
 
 CM 
 
 r^ vo 
 
 
 CO 
 
 W) 
 
 Pi 
 
 
 
 •» 
 
 
 
 
 
 .^ 
 
 4-1 
 
 Q) 
 
 (3 
 
 CM 
 
 <r 00 
 
 0^ . . ON . 
 
 —1 
 
 • 00 
 
 
 
 (3 
 
 
 
 
 1—1 
 
 
 
 
 
 1-H 
 
 0) 
 
 4J 
 
 3 
 
 CM 
 
 <3- 
 
 O CM 1-H 
 
 • 
 
 r-< o- 
 
 
 (3 
 
 •H 
 
 
 
 
 
 
 
 
 
 
 M 
 
 •H 
 
 Pi 
 
 
 
 1-H 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 W 
 
 (3 
 
 
 CO r-( 
 
 00 CO 
 
 ^H ^f CM CM CM 
 00 vD . 1 
 
 in o 
 
 CM 
 
 • 
 
 O 
 
 
 CO 
 
 CU 
 
 CO 
 
 
 
 
 
 
 
 
 •t-l 
 
 
 
 
 
 
 
 
 *-) 
 
 
 3 
 
 <d- 
 
 CVJ 
 
 in 1-1 
 
 CM CM 
 
 o 
 
 in 
 
 
 o 
 
 >-i 
 
 (3 
 
 CM 
 
 
 
 
 
 
 
 CO 
 -C3 
 
 
 Pi 
 
 
 
 1 
 
 
 
 
 
 as 
 
 CU 
 
 O 
 •H 
 
 C 
 
 CO 
 
 <r CM 
 
 <!■ -vJ- 
 
 >;f O r^ CJN r^ in 
 . . . 00 . 
 
 <r 
 
 <3- ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 ^—4 
 
 . in 
 
 
 •H 
 
 
 C 
 
 
 CM r^ 
 
 cvj vo «* i-< <r 
 
 P-. <f 
 
 CM 
 
 CM 
 
 
 o 
 
 
 4J 
 
 3 
 
 CM 
 
 in 
 
 00 1-H CM ^ 
 
 • 
 
 00 in 
 
 
 ^ 
 
 
 3 
 
 CO 
 
 <t CO 
 
 ^ v£) . 1 
 
 1 
 
 . 
 
 • 
 
 
 CO 
 
 CO 
 
 CO 
 
 Pi 
 
 
 
 0^ rH 
 
 o 
 
 
 
 
 pi 
 
 
 CO 
 
 kO O 
 
 CM 
 
 o 
 
 in 
 
 
 CJ^ 
 
 01 
 
 ■d 
 
 
 
 
 
 
 
 
 1 
 
 
 "c" 
 
 
 o> o 
 
 0^ CO CM CO — < 
 
 CM CO 
 
 CM 
 
 en 
 
 
 o 
 
 I— 1 
 Cu 
 
 •H 
 
 
 . 
 
 . . 
 
 
 • 
 
 • • 
 
 . 
 
 
 
 3 
 
 04 
 
 >X) <f 
 
 O 00 
 
 . 
 
 • 
 
 • 
 
 
 CO 
 
 d 
 
 O 
 
 
 . 
 
 . . 
 
 ...... 
 
 • 
 
 • • 
 
 CM 
 
 . 
 
 
 Pi 
 
 
 CO 
 
 r-- 
 
 CVJ 
 
 o 
 
 in 
 
 
 a\ 
 
 CO 
 
 
 
 i-J 
 
 4-1 hJ 
 
 +J hJ 4-1 i-J 4-) C 
 
 (3 
 
 • i-J 
 
 ^^ 
 
 lO 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 1 
 
 
 
 o -^ 
 
 O ~~-. O --^ O •!-{ 
 
 1 
 
 
 X 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "H" 
 
 
 CM t^ 
 
 in vO 1—1 CM CM 
 
 en <r 
 
 CM 
 
 1 — 1 
 
 
 o 
 
 
 
 
 bO 
 
 O, bO 
 
 a bO Ou bO cx d 
 
 • bO 
 
 o 
 
 w 
 
 
 3 
 
 .— 1 
 
 <!■ CO 
 
 — 1 vO 1 
 
 1 • 
 
 « 
 
 • 
 
 
 CO 
 
 • •N 
 
 . 
 
 
 d 
 
 • d 
 
 . d • d • ^ — 
 
 
 • d 
 
 N.-' 
 
 k3 
 
 
 Pi 
 
 
 CO 
 
 vO 
 
 CSJ 
 
 O 
 
 in 
 
 
 as 
 
 CM 
 
 vD 
 
 
 
 
 ..... ^0 
 
 43 
 
 
 (U 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 /--N 
 
 
 
 
 
 ..... 4J 
 
 iH 
 
 
 fa 
 
 
 
 
 
 
 
 
 
 
 X 
 
 W 
 
 
 
 
 ..... M-l 
 
 
 
 
 H 
 
 
 
 
 . • . « . 
 
 • • 
 
 • • 
 
 • 
 • 
 
 • 
 • 
 
 
 
 o 
 
 
 
 
 
 -d 
 
 
 
 -d 
 (3 
 
 
 
 
 ^ u 
 
 i_q +j tJ 4J hJ 
 
 +J (3 
 
 c 
 
 • 
 
 
 hJ 
 
 (U 
 
 <3 
 
 
 
 
 4-1 
 
 
 • c 
 
 • CO 
 
 
 
 
 ^~. o 
 
 •---. o ^- o ^ — 
 
 O -H 
 
 •H 
 
 • 
 
 
 
 fa 
 
 H 
 
 
 
 
 CO 
 
 
 • o 
 
 to 
 
 
 
 
 bo a 
 
 bO Ou bO P- bO 
 
 p- d s 
 
 • 
 
 
 bO 
 
 
 
 
 
 
 
 
 • -H 
 
 ,^> rO 
 
 
 
 
 e : 
 
 m 
 
 Itered) . . . .m 
 
 Itered) 
 
 ered) m 
 
 • ^0 
 
 • w 
 
 • U-l 
 . 
 
 . t3 
 
 • 4-) 
 /-N CO 
 t3 • 
 
 Q) • 
 S-i • 
 QJ . 
 
 lO 
 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 
 o 
 
 •H 
 4-1 
 
 CO 
 U 
 4J 
 (3 
 d) 
 O 
 
 d 
 
 c 
 
 • CO 
 
 ^^ ^0 
 
 O X 
 w o 
 
 (U ^^ 
 fa 0) 
 
 fa 
 
 CO 
 
 
 
 n3 
 (U 
 
 -d CM 
 
 CU -d 
 
 (unfiltered) 
 (unflltered) 
 (filtered)., 
 (filtered).. 
 
 
 • 4-) 
 
 • CO 
 
 • M 
 
 • +J 
 
 • c 
 
 . <u 
 
 • O 
 
 • c 
 
 • o 
 
 • o 
 
 • + 
 
 zed as Fe(OH 
 ed as Fe(OH) 
 
 
 
 
 
 . . .H -H +J 
 
 4-1 • 
 
 
 • 
 
 c 
 
 
 CO CO 
 
 
 
 N 
 
 N (U 
 
 (U n3 
 
 ■d 
 
 Xoi 
 
 •rH > 
 
 
 
 
 
 • . iw M-l tH 
 
 rH • 
 
 
 • 
 
 o 
 
 
 CO 
 
 
 
 •H 
 
 •H > 
 
 > -d 13 -d -d (u 
 
 (U 
 
 Pu 0) 
 
 •d O 
 
 
 
 
 •— . 
 
 • • (3 (3 'H 
 
 •H • 
 
 
 • 
 
 o 
 
 
 13 
 
 
 
 -d 
 
 -d o 
 
 O CU (U (U (U d 
 
 d 
 
 fa 
 
 •H d 
 
 
 
 
 t3 -TJiN • 3 3 M-l 
 
 14-1 • 
 
 
 • 
 
 
 
 CU -d 
 
 
 
 •H 
 
 •H d 
 
 d > > > > 3 
 
 5 
 
 4J 
 
 X! 0) 
 
 
 
 
 CU (U 
 
 •"O ^O ^'^ ^-' ^—^ 
 
 ^^ . 
 
 
 • 
 
 ^- 
 
 
 N CU 
 
 
 
 X! 
 
 X <U 
 
 a; o o o o CO 
 
 CO 
 
 C .H 
 
 O M 
 
 
 
 
 N N 
 
 <U (U 
 
 Xt 
 
 -d 
 
 X CM 
 
 
 •H > 
 
 
 
 O 
 
 O M 
 
 )-i d d d d c 
 
 (3 
 
 (U CO 
 
 
 
 
 
 •H 'H 
 
 > > X) n3 X) 
 
 X) <U 
 
 (U 
 
 a 
 
 (U 
 
 
 -d O 
 
 
 
 
 
 (U <U (U (U o 
 
 O 
 
 3 -H 
 
 + + 
 
 
 
 
 Tj -d 
 
 O O <U (U 0) 
 
 0) d 
 
 d 
 
 
 fa 
 
 
 •H d 
 
 
 
 + 
 
 + + 
 
 + Sh iH M >-i O 
 
 o 
 
 iH 4J 
 
 CM CM 
 
 
 
 
 •H -H 
 
 e > > > 
 
 > 3 
 
 3 
 
 +j 
 
 
 
 X! (U 
 
 
 
 CM 
 
 N (M 
 
 N 
 
 
 M-l •^^ 
 
 CU <u 
 
 
 
 
 XI X! 
 
 (U 0) g o o 
 
 O CD 
 
 CO 
 
 c 
 
 iH 
 
 
 O M 
 
 
 
 0) 
 
 <U (U 
 
 <U d C (3 (3 CM 
 
 CM 
 
 M-l (3 
 
 fa fa 
 
 
 
 
 o o 
 
 M M d d d 
 
 d c 
 
 C 
 
 (U 
 
 CO 
 
 
 
 
 
 fa 
 
 fa fa 
 
 fa a s s s o 
 
 o 
 
 W M 
 
 — CM 
 
 
 
 
 
 0) Q) 0) 
 
 CU o 
 
 o 
 
 3 
 
 •H 
 
 
 + + 
 
 
 
 
 
 
 
 
 
 
 + + 
 
 + + S-l M S-I 
 
 ^ o 
 
 o 
 
 rH 
 
 4-) 
 
 
 CM (M 
 
 
 
 
 
 
 
 
 
 
 CN Cvl 
 
 CN CM 
 
 
 
 M-l 
 
 •H 
 
 
 CU 0) 
 
 
 
 
 
 
 
 
 
 
 (U <U 
 
 0) <u rt fl c 
 
 (3 CM 
 
 CM M-l 
 
 C 
 
 
 fa fa 
 
 
 
 
 
 
 
 
 
 
 
 Cu fe 
 
 CJ4 fe g S S 
 
 S O 
 
 o 
 
 w 
 
 M 
 
 
 '- CM 
 
 
 
 
 
 
 
 
 
59 
 
 amount of Fe^"'" removed was calculated in 
 the same way from unacidified, filtered 
 treated water. Filtering was assumed to 
 approximate settling of iron hydroxides. 
 
 At these high levels of influent Fe^"*", 
 additional quantities of Fe^"*" were re- 
 moved as Fe(0H)2, which produces a green 
 sludge in the ILS effluent at a pH as low 
 as 4.9. Fe(0H)2 is unstable below a pH 
 of about 7; its formation in the ILS sug- 
 gests that a transient pH of at least 8 
 (6) exists at the point of alkaline in- 
 jection and that the dissolution of oxy- 
 gen cannot match the chemical oxygen de- 
 mand represented by the Fe^"*" at that pH. 
 The formation of Fe(0H)2 has both advan- 
 tages and disadvantages: the precipita- 
 tion of both sludge forms reduces dis- 
 solved iron concentrations , allowing 
 discharge of the effluent water, but the 
 Fe(0H)2 sludge will gradually oxidize to 
 Fe(0H)3, lowering the pH of the sludge 
 and, through diffusion, the effluent wa- 
 ter. The difference on tables 5-7 be- 
 tween iron oxidized and iron removed re- 
 flects formation and precipitation of 
 Fe(0H)2. 
 
 The apparent high transient pH in the 
 jet pump may partially explain the ex- 
 tremely high rate of iron oxidation in 
 the ILS. Table 1 indicates that a 1-s pH 
 of 8.5 to 9.0 would be sufficient if DO 
 is continuously replenished; an instan- 
 taneous pH over 10 would reduce the 
 
 required reaction time to milliseconds in 
 whatever fraction of the fluid is at a 
 high pH. It is not known whether the 
 apparent limitation on iron oxidation is 
 due to limited air intake and the rate of 
 oxygen dissolution (from the bubbles into 
 the water) , or to limited catalysis by 
 some other mechanism. 
 
 Apparent oxidation of dissolved iron in 
 the pond, after ILS treatment, continues 
 to be rapid for several days (J^) . This 
 apparent effect is caused by fine-ground 
 suspended Fe(0H)2 or Fe(0H)3 particles 
 that are analyzed as dissolved iron 
 in unfiltered samples. These particles 
 slowly settle in the pond, mimicking oxi- 
 dation and hydrolysis; filtration with a 
 0.45 ym filter confirms this explanation. 
 
 REMOVAL OF MANGANESE 
 
 Manganese, when present in mine water 
 at concentrations greater than 4 mg/L, 
 can significantly add to the costs of wa- 
 ter treatment. In a conventional treat- 
 ment plant, the pH must be raised to 
 above 10 (typically 10.5) for rapid oxi- 
 dation and removal of manganese; this 
 adds greatly to the costs of neutraliza- 
 tion, produces an effluent that is unac- 
 ceptably alkaline, and can cause redisso- 
 lution of iron. Three of the four field 
 sites where the ILS was tested had manga- 
 nese problems. Manganese averaged 68 mg/ 
 L at the West Virginia site (table 5), 
 
 TABLE 7. - Oxidation site 3 — Armstrong County, PA 
 
 Run 1 
 
 Run 2 
 
 Run 3 
 
 Run 4 
 
 Run 5 
 
 Run 6 
 
 Run 7 
 
 Fe^"^ oxidized^ mg/L. . 
 
 Fe^"^ oxidized pet. . 
 
 Fe^* removed-^ mg/L. . 
 
 Fe 
 
 2 + 
 
 removed pet . . 
 
 lin. removed (unfiltered) rag/L.. 
 
 Mn removed (unfiltered) pet.. 
 
 Mn removed (filtered) mg/L.. 
 
 Mn removed (filtered) pet.. 
 
 O2 consumed std ft-^/min,. 
 
 O2 consumed Ib/min. . 
 
 Effluent pH 
 
 Initial Fe^"^ concentration. .. .mg/L. . 
 
 6 
 
 7 
 
 0.3 
 
 0.4 
 
 0.2 
 
 2 
 
 -0.2 
 
 2 
 
 0.02 
 
 0.002 
 
 3.1 
 
 80.5 
 
 75 
 
 98 
 
 75 
 
 99 
 
 2.1 
 
 21 
 
 9 
 
 95 
 
 0.4 
 
 0.03 
 
 11.7 
 
 75.9 
 
 67 
 
 94 
 
 70 
 
 99 
 
 0.1 
 
 1 
 
 9 
 
 91 
 
 0.3 
 
 0.02 
 
 7.3 
 
 71.3 
 
 74 
 
 98 
 
 74 
 
 99 
 
 0.4 
 
 4 
 
 9 
 
 90 
 
 0.3 
 
 0.03 
 
 8.8 
 
 74.7 
 
 77 
 
 97 
 
 78 
 
 98 
 
 -0.3 
 
 -3 
 
 2 
 
 24 
 
 0.4 
 
 0.03 
 
 5.7 
 
 79.5 
 
 76 
 
 99 
 
 77 
 
 99 
 
 -0.6 
 
 -7 
 
 9 
 
 99 
 
 0.4 
 
 0.03 
 
 8.8 
 
 77.3 
 
 78 
 
 99 
 
 78 
 
 99 
 
 -0.4 
 
 -4 
 
 6 
 
 69 
 
 0.4 
 
 0.03 
 
 6.9 
 
 79.5 
 
 Untreated sample, run through ILS without neutralization. 
 ^Fe^* oxidized as Fe(0H)3. 
 ^Fe^* removed as Fe(0H)3 and Fe(0H)2. 
 
60 
 
 14 mg/L at the Westmoreland County, PA, 
 site (table 6), and about 10 mg/L at the 
 Armstrong County, PA, site (table 7). At 
 all three sites, manganese was reduced to 
 within effluent limits after passage 
 through the ILS. 
 
 At the West Virginia site, the first 
 treatment step raised the pH to 4.9 to 
 6.8 and had little apparent effect on 
 manganese concentrations in unfiltered 
 samples. However, 6 h after discharge to 
 the pond, and despite the low pH, man- 
 ganese concentrations had fallen to 13.5 
 mg/L, an 80-pct reduction. Sealed water 
 samples kept in the laboratory showed 
 similar declines in manganese concentra- 
 tions, with no detectable dissolved 
 manganese present after 11 days of 
 storage. 
 
 At the Westmoreland County site, water 
 treated to a pH of 7.1 or greater met 
 effluent standards for manganese after 
 filtration. Filtered samples that had 
 been treated to a pH of 6.9 to 7.0 ap- 
 proached manganese effluent standards 
 
 (3.1-5.6); it was not possible to collect 
 samples after settling. 
 
 At the Armstrong County site, a similar 
 pattern was observed. Fe^"*" was reduced 
 to below effluent standards at a pH of 
 5.7 and up, but manganese exceeded ef- 
 fluent limits under pH 7.3. ILS test 
 runs at pH 7.3 or above met effluent 
 standards. 
 
 It appears that manganese was precipi- 
 tating as very small particles during 
 neutralization and aeration in the ILS. 
 Filtration removed most of these parti- 
 cles; settling removed all of them. It 
 is possible that the transient pH in the 
 jet pump was high enough to allow for 
 rapid formation of Mn02. Alternatively, 
 it is possible that the manganese is be- 
 ing removed from solution as a coprecipi- 
 tate on particles of iron hydroxide as 
 they form and swirl in the ILS. It has 
 previously been shown that adsorption of 
 manganese by Fe(0H)3 increases rapidly 
 above pH 8, rising from 0.15 to over 0.6 
 mol Mn2+ per mol Fe^+ at pH 8.6 (7). 
 
 PRELIMINARY DESIGN SPECIFICATIONS 
 
 Three parameters should be considered 
 in the design of the ILS: available wa- 
 ter pressure, flow, and influent Fe^"*" 
 concentration. Table 8 partially sum- 
 marizes flow capacity of the existing ILS 
 design at various water pressures. In 
 general, adding jet pumps (in parallel) 
 increases capacity. The ILS that was 
 designed and tested by the Bureau had 
 valves on two of the three jet pumps. 
 This allowed for variable flow rates and 
 is a potentially useful feature on sites 
 where surface runoff during storm events 
 determines treatment requirements. 
 
 Additional helixors influence flow ca- 
 pacity and may increase oxygen transfer. 
 
 TABLE 8. - Flow rates for 
 
 Increased water pressure increases flow 
 capacity. If flows are above 500 gal/ 
 min, larger capacity jet pumps can be 
 substituted or additional jet pumps can 
 be placed in parallel. 
 
 If Fe^"*" levels are above 300 mg/L, a 
 two-stage treatment process may be neces- 
 sary. This can actually be quite effi- 
 cient, as shown in tables 2 and 5, and 
 does not necessarily require a second ILS 
 unit. For example, a surface mine can, 
 by installing valves in multiple suction 
 and discharge lines , pump the once- 
 treated water from the first-stage set- 
 tling pond through the same ILS to 
 another settling pond. Similarly, an 
 
 the ILS, gallons per minute 
 
 Test design 
 
 20 psi 
 
 30 psi 
 
 40 psi 
 
 50 psi 
 
 64 psi 
 
 3 jet pumps in parallel with 1 helixor. . 
 3 jet pumps in parallel with 2 helixors 
 in series .,, 
 
 NA 
 
 329 
 
 NA 
 
 411 
 317 
 261 
 
 469 
 457 
 310 
 
 521 
 542 
 344 
 
 NA 
 NA 
 
 2 jet pumps in parallel with 2 helixors 
 in series 
 
 363 
 
 NA Not available. 
 
61 
 
 underground mine that is intermittently 
 discharging through an ILS unit into a 
 settling pond can pump from the pond 
 through the same ILS to a second settling 
 pond while the underground pump is off. 
 However, if there is continuous flow, 
 then two ILS units and a second pump are 
 necessary. Our tests of the two-step 
 process indicate that a 50-pct reduction 
 in neutralization costs is possible with 
 such a system (table 2). 
 
 Actual oxygen consumption rates, as 
 calculated from the amount of iron oxi- 
 dized during passage through the ILS, are 
 shown on tables 5-7. Air transfer tables 
 provided by the jet pump manufacturer do 
 not appear to correlate with observed 
 oxygen consumption. For example, in our 
 ILS design, sample 2 (table 5) consumed 
 2.3 std ft^/min O2 operating at 40 psi, 
 with three jet pumps in parallel and with 
 10 psi back, pressure; the manufacturer's 
 tables predict 1.5 std ft-^/min per jet or 
 4.5 std ft^/min of air intake for the 
 three-jet pump system. Sample 1, operat- 
 ing at 50 psi, consumed 2.4 std ft-^/min 
 O2; the same tables predict std ft^/ 
 mln of air intake under these operating 
 conditions. Actual air flow measure- 
 ments are needed so that oxygen transfer 
 can be quantified. 
 
 Another aspect of system design is 
 cost. The 3-in PVC jet pumps and static 
 mixers cost about $900 and $2,500 each, 
 respectively. Associated PVC plumbing 
 costs about $500. For our tests we pur- 
 chased a hydraulic pump with a diesel 
 power unit capable of providing pressures 
 up to 50 psi with three jet pumps, but 
 any heavy-duty pump should serve. The 
 total cost is, of course, much less than 
 for construction of a conventional em- 
 placed treatment system. Operating costs 
 should also be low owing to the efficient 
 mixing action, efficient iron oxidation, 
 and lack of moving parts. 
 
 There are other advantages. The system 
 is small, and if desired, portable. It 
 requires no electrical power, although it 
 does add slightly to the Ipad on the mine 
 water discharge pump (approximately 10 
 pet). The basic design is simple and 
 easily modified to cover a wide range of 
 flow and pressure conditions and can 
 operate continuously or intermittently. 
 It can also be dismantled easily for use 
 elsewhere if water treatment is no long- 
 er required. Finally, although settling 
 ponds are required, they do not serve as 
 aeration basins and therefore do not re- 
 quire as large an area as would be the 
 case for a conventional treatment plant. 
 
 REFERENCES 
 
 1. Ackman, T. E. , and R. L. P. Klein- 
 mann. In-Line Aeration and Treatment of 
 Acid Mine Drainage. Paper in Proceed- 
 ings, 1984 Symposium on Surface Mining, 
 Hydrology, Sedimentology , and Reclama- 
 tion, Dec. 3-7, 1984, Lexington, KY, ed. 
 by D. H. Graves. Univ. KY, Lexington, 
 KY, pp. 29-34. 
 
 2. . In-Line Aeration and Treat- 
 ment of Acid Mine Drainage, BuMines RI 
 8868, 1984, 9 pp. 
 
 3. Gosline, J. E. , and M. P. O'Brien. 
 The Water Jet Pump. Univ. CA, Publ. 
 Eng., V. 3, No. 3, 1942, pp. 167-190. 
 
 4. Kim, A. G. , B. S. Heisey, R. L. P. 
 Kleinmann, and M. Deul . Acid Mine 
 Drainage: Control and Abatement Re- 
 search. BuMines IC 8905, 1982, 22 pp. 
 
 5. Lovell, H. L, The Reagents. Ch. 3 
 in Fundamentals of Water Pollution Con- 
 trol in Coal Mining. PA State Univ., 
 State College, PA, 1982, 360 pp. 
 
 6. Snoeyink, V, L, , and D, Jenkins, 
 Water Chemistry, Wiley, 1980, 463 pp. 
 
 7. Sturam, W. , and J. J. Morgan, 
 Aquatic Chemistry, Wiley-Interscience, 
 2d ed., 1981, 780 pp, 
 
 8. U.S. Environmental Protection Agen- 
 cy. Neutralization of Acid Mine Drain- 
 age. EPA-600/2-83-001, 1983, 231 pp. 
 
 9. Weber, W, J, Physiochemical Pro- 
 cesses. Wiley-Interscience, 1972, 640 
 pp. 
 
 ifU.S. CPO: 1983- 50>fl19 20,018 
 
 INT.-BU.OF MINES, PGH., PA. 27998 
 
y(8s 
 
D DD 
 
 n 
 o 
 
 a 
 n 
 
 > 
 
 ::j^ 
 
 3 
 
 
 
 /D 
 
 Q_ 
 
 
 
 Q 
 
 
 
 
 
 Q. 
 
 3 
 
 -•• 
 
 3 
 
 
 -♦■ 
 c 
 
 3 
 
 -+• 
 
 (/> 
 
 
 
 a 
 
 i 
 
 Q 
 
 (A 
 
 c 
 
 
 
 
 (/) 
 
 n 
 
 
 
 (/) 
 
 
 _. 
 
 3- 
 
 3 
 
 T> 
 
 3- 
 
 Ul 
 
 3 
 
 Q 
 
 Q 
 
 
 -!■ 
 
 
 
 o 
 
 3 
 
 
 a 
 
 
 
 3 
 
 
 (O 
 
 ^7 
 
 
 
 -^ 
 
 Q. 
 
 n' 
 
 m 
 
 3 
 
 VI 
 
 Q 
 
 <t 
 
 Q 
 
 • 
 
 IQ 
 
 <D 
 
 n 
 
 "^ 
 
 
 
 
 
 n 
 
 
 n 
 
 
 
 
 
 
 a. 
 
 
 -•• 
 • 
 
 <5 
 
 3 
 
 
 < 
 
 -«■ 
 3- 
 
 
 
 u> 
 
 
 < 
 
 
 
 
 (h 
 
 
 n> 
 
 (/) 
 
 
 I- 
 
 
 CD 
 
 C/3 
 CD 
 
 a O — 
 Doc 
 
 T3§> 
 
 m ™ C 
 2 CD 
 
 ^ m O 
 
 < m Z 
 > Z m 
 z cco 
 ^ m 
 
 O 
 
 m 
 
 > 
 
 H 
 
 li 
 ^5 
 
 _ m 
 
 H 
 
 m 
 O 
 
 JD 
 
 m 
 O 
 c 
 > 
 
 o 
 
 TJ 
 TJ 
 O 
 3J 
 
 H 
 C 
 Z 
 
 H 
 -< 
 
 m 
 
 O 
 
 -< 
 m 
 33 
 
 - m "1 
 
 2 2 > 
 
 7^ -• ^ 
 
 ± O D 
 
 01 "n -n 
 
 H m 
 
 I m 
 

 .-^^^ 
 
 ^oV" 
 
 > 
 
 
 
 .^^ "^^ 
 
 
 
 
 _. . / \ , 
 
 °- /--^i-X y-'M^---^ /'-m^-^-K, «'■ ' 
 
 .0 iu 
 
 ^^^^J^t/^-^ ** /.C^.'^^v 
 
 A*?^ c " « * <^. 
 
 4P^ 
 
 
 o. * 
 
 
 
 . --...* /Jfe'-. %../ .'i««i". -^,.<-* /^fe; \„c/ :M£^ \,^' :^: \>/ 
 
 V<b~ 
 
 vV^. 
 
 VV 
 
 ^'^^'^p 
 
 "jxc,-' 
 
 tA' ^ 
 
 iVA. 
 
 ,^' V^ -y^W* > 
 
 
 
 ■Iv"^ »>^ 
 
 » Ay '^ 
 
 co^c;^-*°o y..^:-.\ <-"-i^'> *■«* -^iX 
 
 
 '*\<^' ../<'/* '"^•■^*' J.(>' ... ^^. '"•*" a\' -^- "'^- ••• 
 
 o. -o.\* ^^ <^.'*'T;t^' ,0^ '<3. 
 
 * A 
 
 ■^ 
 
 'bV^ 
 
 ^^0^ 
 
 'bV^ 
 
 ^^-^^^ 
 
 ,-io^ \^^^i ^^^. r'^^^m: ^^^^ - 
 
 4 W ■rr> 
 
 0' 
 
 
 ,'o ^. 
 
 rr<^ .G 
 
 C,vP 
 
 
 .^^r 
 
 
 /.•i;^. °- y.-^i-^. c°*..i;^.>o ./..jmi'/-?.. 
 
 
 
 ^R<, « 
 
 '^'' "^ 'JW^-^* <S^ '^4^ 
 
 
 
 
 0^ e'^ll^J*. 
 
 
 
 
 ■^ LIB "^ (V" o « o . «4> 
 
 •~ O, 'o , » * A 
 
 47 %. * 
 
 
 
 W^' 
 
 
 
 
 'y c ° "" * <^^ 
 

 • O ^'a * » 1 1 • O.^ O^ * o « o ' 
 
 .^•^ 
 
 
 • JT 
 
I