.......... ......... ......I. . ..v.. . e..— .7? ... . ... . ...... L........:...._...:... . .-....I... ...... ........., .....:.........._......... 1...... . .................§....... ... .. .... .... ......1. ... .. .. . . .. . . ... .. 5.... ......Z... ......{ern m.............._............... ... . . I... .. ... .. .... ..........._. .. .... . .....f..._ (A . ... .. .... ... ....... .. .. .....I. ._ ..1.... . ... .. . ......i... ...... ... ........_.. ...}... ......t... ... ... .... ... 1:31.... .... ...... .....C... 11...... .... ...: ......h... . 311...... . . ...... ...._........ ....,..;..... ...; . .. ......l... .............. 1:143... ... .. .. ........ ... ....1..._.....; .. . ... .... ......:.2 .. .. .. .... _X....... _. .. . 2..... ..,._.... . ..1. . 1.7.5:... _ 2.3.... ...: .. . I... .........I.... . ......L :7... .21.......... .. .2..-.» ......1. .v.. 2... -..... .......... .... .... .1.......J..... ...... .. 33...... . ......H....3 .233-.. , . . 1.... (1...... . : .2.... L ... ........... 1......2 ... ............. .....1 11.33... .......... .. .. ............ .. . ......m... ...... .. ....x... ......A.... .. . . ‘ ...‘.. . . r. . 1» .\.~... \. ~ ~ ~ I ‘ ~ . . . . ...... .:2...::.:... . 12.2.2.2...)1 .... 2... .1 . . ......E... ...... ......;......... ......«I. . ..‘ . ............ .... ......... .... .. . . -..... .7, .....1... ... . 1 .... -... .. ....3... ....a ..1. . .. 2.. I ..........1. L... ... ..n. . ... ...... . . . .. . .................L . I . . .....i . . .. .. . ...»...2... .. . . . .. ...: . .. I .......... . 1 .23.... ..............I .. . .. .....1. ...:z.....«. ...: . . .. . 1..... 31.31.21... .. . .. .....................2. . :2:.......:.......:. 3.13.... .. . . .. . ......713. .................... . . . 2:21...71.741.12.311: . .. . 31.2.... . . «... 1 ~ .... .2.. I 34 . .....:.3 .............1 ...1 1.. ..........J. .. , .1... . 4.1....» 24.. . i .. ... . . . ...... . . . . . . 37;: .27... .. . .. .. .. 1.11:...) . .... ............?.. . . .. . . . ... 1. .. ‘ ... . E .. ... .. ...... .... .2.. .. . 5.1.21.1»... .. 3... ... .7... . 7213.....1 ..........._.... ...»w...... .. ...... . . ...s‘..,...: . . .....x.... .... -17.... 1.1.1.... . .3}... . . .. l ..v 11...... . .. .3....... ...... . . . 34..-»... 1.1:. . . . .. . . ......; 17;... . . . ....>..\L.~.. . . .. .... . .. . _. . ......1...............Z ...... .11... . . . 11-22.... .. ......2213 ...1 . . . . 3 . . . ... . ......m.....\... ... . .... ...f . 1 . ...... .. ........... . ........1.. 1.1.. .... . . .3... .... . . . .. . ...... .s. .. .. . . .....TT . . . . . .. . 12.1,; .. 31...»... ...: . ... .. . .. .. 1.3.)... .. . w.. . ...: ......1». .........y.......... . 5...... . . .. . . .1. . .....T.. . .. . . ......1... I .. . . . . .... 1..... . . . . . .. ...... . . . .. . . .. .. ... ... . . . . y. . . . . .. . 71.... 3:15.... . .. . . . . . .. .11.... .. . ... .... .. . . . .. 1... s... 1. . . .1 . . . ... . _. . .1 . . 1 .. ...... .. .. .3 . 1.3.... 4 . . . I . x ............. . . .. .. .... . . . . . . . . . ...... 1. . _ . .. , .....~ .. . . . .. ... . . ..\.1.I..x.... . . 1...: ... . ... . 1. .. ..u!.......1... . . . ... . .IA...v....v1 . . ... I ....A: . 1 I11, J": LIBRARY Michigan State University This is to certify that the thesis entitled PATHWAYS AND INTERACTIONS OF COPPER WITH AQUATIC SEDIMENTS presented by Jeffrey Thomas Cline has been accepted towards fulfillment of the requirements for Ph. D. Geology degree in 929224 - ajor professor Date May 6, 197A 0-7639 ABSTRACT PATHWAYS AND INTERACTIONS OF COPPER WITH AQUATIC SEDIMENTS By Jeffrey Thomas Cline The reactions of Cu++ with bottom sediments and sediment pore waters were studied to delineate the path- ways of capper within sediments. Organic rich sediment from Burke Lake. Michigan. was utilized in controlled laboratory experiments to develop a schematic model for Cu++-organic sediment reactions. The model was tested on sediments and water from Houghton Lake. Michigan. a copper contaminated lake. The reactions of Cu++ and for comparison Co++ with the colloidal organics in sediment pore water were investigated. without disturbance of the system. by means of a computer-centered spectrophotometer- Spectrofluorimeter combination instrument. The changes in fluorescence. absorbance and light scatter emissions of the organics due to their reaction with Cu++ and Co++ show that these metal ions react with the pore water organics. Cupric ion reacts strongly with the organics in a non-linear step function. The Cu++-organic reaction is <3 /Q>d:s\~ Jeffrey Thomas Cline also pH dependent. reversible and the complex formed is stable under dilute conditions. The Cu++-organic reaction is likely a chelation. Cobaltic ion reacts more weakly and is probably surface adsorbed. Cupric ions are more efficient than Co++ in the flocculation of pore water organics. A low specific gravity organic-metal precipitate results which. under turbulent conditions. in an aquatic system. could cause copper mobilization and distribution in the sediment. The uptake of Cu...... by bottom sediments is rapid until saturation, as predicted by ligand complex theory. but continued diffusion dependent reaction occurs at a slow rate ranging from 0.28 to 0.18/day. The Cu++-sediment reaction has high stability constants ranging from a log (Q) of 2 to 8 depending on conditions. The reaction reaches an adsorption endpoint at about 8070 Cu/gm of Burke Lake sediment. After initial Cu++-sediment reaction as much as 65% of the Cu++ is exchangeable as organically complexed Cu++ and surface adsorbed Cu++ while 35% is in a non-exchangeable form. Then. with burial the Cu can accumulate upward in the sediment as a result of an upward migration by diffusion, on bubble surfaces. or with sediment pore water. A schematic model quantitatively and qualitatively presents the trends of these laboratory results. A test of the schematic model in a natural aquatic system, Houghton Lake. shows that Cu++ is (1) highly associated with low specific gravity organic floccules and other particulate Jeffrey Thomas Cline matter in the water, (2) quantitatively removed from solution by sediment reaction. (3) associated with bottom sediment organic matter and (4) the most highly concentrated in the upper centimeters of sediment. ACKNOWLEDGMENTS Special thanks are extended to Dr. Jack Holland for his suggestions, technical assistance and use of his computer-centered fluorimeter. Thanks are due to Drs. F. D'Itri, B. Knesek, M. Mortland and S. Upchurch for review of this manuscript. Richard Chambers is acknowledged for needed encouragement. Carol Cline is thanked for typing and especially for her patience. Gratitude is also extended to the Lake Survey Center (NOAA) for the preparation of the figures. PATHWAYS AND INTERACTIONS OF COPPER WITH AQUATIC SEDIMENTS By Jeffrey Thomas Cline A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geology 197A TABLE OF CONTENTS Page LIST OF FIGURES O O O O O O O O O O O O O O O 0 iv LIST OF TABLES . . . . . . . . . . . . . . . . vi INTRODUCTION 0 O O O O O O O O O 0 O O O O O O 1 DESCRIPTION OF HOUGHTON LAKE AND BURKE LAKE . . 8 METHODS I Tests on Metal-Organic Pore Water Inter- actions . . . . . . . . . . . . . . . . . 12 Fluorescence. Absorbance and Light Scatter of Pore Water Interactions O O O O O O O O O O O 0 12 Specific Gravity of Organics Flocs . 15 Order of Reactivity of Hg++, Cu++, and Co++ with Organic Compounds . . . 16 Test of Chelation of Heavy Metals by Organics 0 O O O O O O O O O 0 O O 17 METHODS II Tests on Sediment-Metal Interactions . . . 18 Rates of Reaction . . . . . . . . . . 18 Quantitative Aspects of Uptake . . . 18 Pathways of Cu++ Within the Sediment 20 RESULTS AND DISCUSSION Heavy Metal-Organic Interactions in Sediment Pore Water . . . . . . . . . . . . . . . . 22 ii Solubility of the Organics . . . . . Specific Gravity of Organic Floc . . Order of Reactivity of Hg++, Cu++ Co++ with Organic Molecules . . . . Test of Chelation of Cu++ by Natural Organics . . . . . . . . . . . . . . Extent of Organic-Metal Reaction in Pore Water Solution . . . . . . . . Formation of Metal-Organic Flocs . . Sediment-Heavy Metal Interactions . . . . TESTS OF THE LABORATORY DERIVED RESULTS IN NATURAL Rates of Reaction . . . . . . . . . Quantitative Aspects of Uptake . . . Endpoints of Heavy Metal Reaction with Sediments . . . . . . . . . . . Pathways of Cu++ Within the Sediment ENVIRONMENTS Houghton Lake Copper Study . . . . . . . Related Studies from Literature . . . . . SUMMARY AND CONCLUSIONS . . . . . . . . . . . APPENDIX . LIST OF REFERENCES . . . . . . . . . . . . . . iii , and Page 22 27 28 33 35 48 52 52 58 68 69 82 9O 94 103 1&2 LIST OF FIGURES Figures 1. lC)- 1].- 12. Generalized model of heavy metal (M) phases and pathways in a natural fresh water environment. Map of lower Michigan showing the location of the two sample sites. Sample light scatter output. Dependence of organic solubility on pH as shown by light scatter. Effect of Cu++ concentration on organic solubility-light scatter. Competition between Hg++, Co++, and Cu++ for adsorption sites on organics in pore water. Equilibrium of Cu-organic chelate shown by pH adjustments. Example fluorograms of an organic-rich, pore-water solution containing an vincreasing concentration of Cu+ . Chan e of fluorescent (a) and absorbant (b intensity of dissolved organics in pore water at specified wavelengths (A) with an increase in Cu++ concentra- tion. Standard organic fluorescence curves at specified wave lengths (A). Effect of Co++ concentration on or anic fluorescence (a) and absorbance %b). Comparison of Co++ and Cu++ effect on organic solubility in pore water using light scatter. iv Page 11 24 25 26 32 34 36 40 4M 47 “9 Figures 13. 14. 15. 16. 17. 18. 190 20. 21. Al. A2. Time dependence of Cu++ and Co++ reaction with organic rich bottom sediment. Freundlich Isotherms for the reaction of Cu++ with pore water from Burke Lake sediment. Freundlich Isotherms comparing the fit of data and the stability of reaction of Cu++ (o) and Cot+ (o) with Burke Lake sediment. Freundlich Isotherms comparing the reactions of Cu++ (———) and Co++ (- - -) with Burke Lake sediment after two hours (0) and after 11 days (0). End points for Cu++ reaction with Burke Lake sediment through time. Concentration of Cu++ necessary in Burke Lake sediment before toxic concentrations remain in solution as a function of reaction time. Comparison of Cu phases within a sediment column and with time of cores at months 1 and 2. Diagram of Houghton Lake showing the eight coring locations (x) and the four water sample sites (0) tested for anaerobic mobilization of Cu++ Model of heavy metal (M)-phase relation- Ships and preferred pathways through the subaquatic environment. Rates of diffusion dependent Cu++ reaction with sediment. Rates of diffusion dependent Co++ reaction with sediment. Page 5h 61 63 67 7O 74 80 87 100 130 132 LIST OF TABLES Tables 1. Al. A2. A30 Au. A5. Description of Burke Lake and Houghton Lake sediments and corresponding pore waters. I 0 o o o o o o o O o o o O o O o 0 Rates of diffusion dependent Cu++ and Co++ reaction with sediment . . . . . . . . . . Equilibrium constants (log Q) for Cu++ and Co++ interactions with bottom sediments from Burke Lake. . . . . . . . . . . . . . Change of phase of Cu in Houghton Lake sediment with time. . . . . . . . . . . . . Copper concentrations, percent organic carbon (0.0.) and size with decreasing depth in eight cores taken in Houghton Lake 0 O O O O O O O O O O O O O O O O O 0 Data presented in Figure A concerning the dependence of organic solubility in pore water on pH of the solution . . . . . . . . Effect of Cu++ concentration on organic solubility-light scatter as shown in Figure 50 o o o o o o o o o o o o o o o o o The data for competition experiment shown in Figure 6 are presented . . . . . . . . . Data for Figure 7 showing the equilibrium between the Cu-organic chelate with pH adjustments. . . . . . . . . . . . . . . . Data for Figure 9 - (a) and (b) giving the change in fluorescent (a) and absorbant (b) intensity measured in cm at various wave- lengths (X) in mu as Cu concentration Changes 0 O C O O O O O O O O O O C O O O 0 vi Page 10 56 65 75 84 103 104 105 110 113 Tables A6. 137. A13. A59. A10. 4A1]_. A122. A13. Aln. Data for Figure 10 showing the intensity of fluorescence of diluted pore water is measured in cm at three wavelengths . . Data for Figure 11 (a) and (b) giving effect of Co++ on organic fluorescence (a) and absorbance (b) intensity in cm at two wavelengths (A) in mu. . . . . . . Data for Figure 12 comparing Co++ and Cu++ affects on the light scatter of a pore water solution (in cm) . . . . . . . . . . Data for Figure 13 showing the time dependence of Cu++ and Co + reaction with organic rich sediments from Burke Lake. All data from the experiment are given though only the 100 mg M++ data are presented in Figure 13. . . . . . . . Data and figures nsed to obtain the rate constants K (day’ ) in Table 2 of the Co++ and Cu...... reactions with Burke Lake sediment................. Data for Figure In glying two Freundlich Isotherms for one Cu ~0rganic pore water reaction . . . . . . . . . . . . . . . . . Data for Figure 15 yielding Freundlich Isotherms of Cu++ and Co++ reaction with Burke Lake sediment. . . . . . . . . . . . Data for Figure 16 yielding Freundlich Isotherms for Cu++ and Co+ reactions with Burke Lake sediment through time . . Data for Table 4 showing the change of phase of Cu++ as pore water Cu. KCl extractable Cu. EDTA extractable Cu and precipitated Cu. . . . . . . . . . . . . . vii Page 115 116 117 118 126 130 131 135 137 INTRODUCTION Heavy metals have been proven toxic to man and aquatic organisms in small but excess quantities (Federal Water Pollution Control Administration. 1970). Many of these. such as the cations c0pper. mercury. cadmium. zinc. chromium. selenium. and iron and the anionic radicals borate and arsenate have been found in concentrations above background levels in the aquatic environment (Upchurch. 1973). These findings have led to studies of metal distribution (Kennedy gt al,. 1971: Walters gt g;,. 1972) and correlations between increased levels of metal toxicants and disruption of the food chain (D'Itri. 1972: Massaro and Giblin.1972: Thommes gt a;.. 1972). Thus. the pathways of heavy metals through the aquatic and subaquatic environment (sediment) are of current concern. This study will present by means of a schematic model the quantitative and qualitative interaction of capper with sediment and sediment water. Heavy metals occur naturally in solution in minute quantities. except near ore deposits. and are usually in the form of complexes or simple catibns as aquo complexes. The metal ions may be complexed by anions such as carbon- ates. chlorides. sulfates. phosphates. and nitrates 2 (Carrels and Christ. 1965). The minute concentrations of the heavy metals in aquatic environments cannot be reconciled by these inorganic complexes alone. Organic anions. adsorbates. and complexing agents interfere with inorganic complexes and establish more stable. heavy metal-organic equilibria (Oden. 1922; Krauskopf. 1956; Kitano §3_g1.. 1970: Faust and Hunter. 1971). About ten percent of these organic complexing agents are small-chained hydrocarbons and identifiable as amino acids. urea. phenols. alcohols. esters, amino sugars. lipids. fatty acids and proteins (Degens. 1970). The other ninety percent of the organic compounds are referred to as humic substances or humic "acids“. All of these natural organic compounds are distributed widely over the earth's surface in soils, lakes, rivers. and in the ocean. Humic substances are classified by solubility and the two classes most often referred to are: humic acid which is soluble in base. insoluble in acid. and fulvic acid which is soluble in both acid and base (Kononova. 1966). The humic substances have very high molecular weights ranging from 2000 to 300.000 (Schnitzer. 1966: Eglington. and murphy. 1970). The molecules consist of long carbon chains: complex aromatic structures (Christman and Minear. 1971); and oxygen. nitrogen. sulfur and phosphorus functional groups (Saxby. 1969). such as the aromatic amino (-N3) and carboxylate (-C00-) groups (Martell. 1971). In surface sediment the humic substances have an average CsOstfl ratio by weight of 51:05:133 (Schnitzer. 1971). In deeper sediments the percentage of carbon and hydrogen increases while oxygen and especially nitrogen become depleted due to biological activity (Hood. 1970). Humic substances and less complex organic compounds such as amino acids. proteins. carbohydrates and lipids. often occur in combination as polymers or are joined by means of oxygen and nitrogen functional groups in a Six membered ring called a clathrate compound (Degens. 1970). These large organic molecules often display colloidal properties in solution (Oden. 1922) such that changes in ionic strength. presence of heavy-metal cations or changes in pH can cause precipitation of a low specific gravity floc (Breger. 1970). All of these various organic substances often directly influence the chemistry of dissolved substances in water. especially heavy-metal cations. Malcolm 33 a1. (1970) theorized that fulvic acid is an important scavenger of heavy-metal ions such as cobalt and iron. and that the acid should be a major factor in the physica- chemical transformations of these ions in the natural environment. Heavy metals can be complexed by both natural organic compounds such as humic substances (Kraynov gt_§;,. 1966; Schnitzer and Skinner. 1966; Ellis and Knesek, 1971: Christman and Minear. 1971). and by synthetic, organic compounds (NTA, EDTA) (Knesek, h unpublished; Childs. 1971). Schnitzer (1971) studied the chemical structure. reactions and reactivities of soil humic substances and found that they form complexes with metal ions that are both water soluble and insoluble. In solution these metals become bound to or complexed by natural. organic compounds through low energy bonds (adsorption). high energy bonds (chelation). and very high energy. carbon-metal covalent bonds. Much research has been done to establish the para- meters for complex stability of natural organic compounds found in soils (Kononova. 1966: Schnitzer. 1971). Few studies (Krauskopf. 1956: Schindler. gt a1.. 1972: Chau. V.K.. personal communication. 1973) have character- ized organic heavy metal interactions in bottom sediments. interstitial waters and surface waters. As a result. little is known about the consequences of metal inter- actions on man and his environment. Prior to investigations by Birge and Juday (1934). it was assumed that lakes. streams and rivers were inorganic solutions as long as they were not contaminated by organic refuse or industrial waste (Ruttner. 1962). The organic matter in these waters was considered to consist of organisms and organic detritus and the plants of the associated aquatic communities were supported by assimi- lation of inorganic nutrients. Birge and Juday (1928) showed that lake waters contain small amounts of particulate organic matter and considerably larger amounts of dissolved organic matter. Soil scientists. limnologists and geochemists (Birge and Juday. 1928; KrauskOpf. 1956: Hutchinson. 1957: Scheffer and Ulrich. 1960; Ruttner. 1962; Shapiro. 196A: Kononova, 1966; Robson and Colombo, 1967; Eglington and Murphy. 1969: Breger. 1970: Bremmer gt a;,. 1970: Schnitzer. 1971) have found that this dissolved organic matter is important in the chemistry of waters. soils and sediments. An application of known parameters and an evaluation of the role played by organic complexes in trace metal equilibria in natural waters and sediments is lacking. Figure 1 shows the possible forms (phases) and pathways of a heavy metal in a water and sediment system. In this study a selected. organic-rich sedimentary environment was chosen and its sediment used as a heavy metal reactant within the laboratory. Copper was chosen as a typical heavy metal reactant. with Co and sometimes Hg used in comparison. Some of the mechanisms of organic compound- heavy metal reactions are delineated by utilizing the fluorescent, absorbent. and light scatter properties of natural organics. The stabilities of the metal-organic complexes are investigated and the kinetics of some of the reactions of Cu and Co with sedimentary components are discussed. The laboratory data and resulting hypothe- sis show trends that are tested by observations in a natural environment that has been Cu contaminated. Finally. a schematic model is proposed to show the general stoiciometry and direction of Cu pathways between given phases in the sediment. The results discussed are not to be interpreted as well defined numbers for all Cu++- sediment reactions but should show reaction trends under .a variety of similar conditions. 1| Air Air WON” Water ‘7 ionic volatile organic M+ soluble organic , METAL SOURCES fig M++ Remove! of M “3. RH: MtMo’M—grganic. 05 inorganic complexFM °" ”cm" —morgamc MO . particulate bound colloidal bound 4) Water _ Water Sediment Sediment v surface adsorbed .- Z w M++ complexed soluble 5 Mi inorganic precipiles o . . w as organic complexed DONICUlO m NW . & soluble ion 0 volatile inorganic I volatile organic .— O. I“ ° (r V Permanent BuHal Figure 1. Generalized model of heavy metal phases and pathways in a natural fresh water environment. DESCRIPTION OF HOUGHTON LAKE AND BURKE LAKE \ Burke Lake sediment and Houghton Lake water and sediment are the systems studied. Burke Lake sediment is used as a source for controlled laboratory experiments because it is free of an excess of heavy metals and also is highly reactive. Organic rich Houghton Lake sediments are used to test the model based on the Burke Lake sediments. Specific chemical and physical descriptions of the studied sediment from the two lakes are given in Table 1. Burke Lake is a small. eutrophic. kettle lake located in a wooded area of central Michigan. 13 km north- east of Lansing, Michigan (Figure 2). It is about 2.000 m2 (2 acres) in area and 11 m deep. There are no houses. farms or industries on the lake and it is fed by ground water. The water visibility ranges from 1 m to 5 m depending on the time of year. Around its edges is an extremely organic-rich and biologically-active. black sediment more than a meter thick. The sediment contains little clay. silt. or sand. Houghton Lake is a copper contaminated lake located in central Michigan (Figure 2). It is a glacial lake in its final stages of in filling. The surface area cf the lake is 82 km2 (32 sq. mi.) and the average depth is 3 meters (Novi. J.. personal communication. 1972). The lake is highly productive. It is surrounded on three sides by swamps which contribute organic matter to the lake resulting in the high color of its water and a low average secchi disc reading of 0.5 to l m. A black organic-rich sediment covers most of the bottom but sand and some marl clay sediments are also present. 10 TABIJZ 1. Description of Burke Lake and Houghton Lake sediments and corresponding pore water. All heavy metal concentrations in pore water are mg/l and in sediments mg/kg dry weight. The grain size was measured on a Coulter Counter (error = 1%) after sonic disagregation. Burke Lake Houghton Lakg pgre water sediments pore water sediments organic carbon 92 mg/l 31 to #09; 72 mg/l 12 to 31% original pH 7.6 --- 7.7 ...... original Cu 0.0 0.0 1.0 30 to 90 original Co 0.0 0.0 0.0 5.0 9‘ H20 in wet sed. --- 72 to 80 --- 70 to 75 avera e grain --- Li --- 5 size Imicrons) original Eh negative negative 11 STRAITS OF MACKINAC 0600 o “as °o HOUGHTON LAKE {is . MICHIGAN ST. CLAIR RIVER OBURKELAKE F1QHPPZ. Map of lower Michigan showing the locations of the two sample sites. In heavy print are Burke Lake. near Lansing (increased 531 in scale) and Houghton . Lake. in north central Michigan. METHODS - I Tests on Metal-Organic Pore Water Interactions Fluorescencg. Absorbance and Light Scatter of Pore Watgr Intgractions The fluorescing properties of natural organic compounds were utilized in studying metal-organic reactions. In general. fluorescence may be expected from organic compounds having a conjugated system. such as aromatic compounds. and not expected from aliphatic compounds (Guilbamflt. 1967). Thus. natural organic compounds such as tryptophan and tyrosin have functionally active amino groups which can react with a cation. The reaction will cause the amino acids to lose their fluorescing ability (Guilbaurh.l967). These compounds and the proteins they are associated with are typical indicator molecules in the reactions of this study. The specific chemical properties of the molecules in the pore water solutions were studied by means of a °OMPuter centered spectrophotometer-spectrofluorimeter combination instrument (Holland. 1971). On this instrument absorbance and fluorescence measurements are made simultaneously and light scatter measurements are made 12 13 individually. The simultaneous measurement technique makes valid the interrelationships between absorption and fluorescent emission. The computer collects the data from the energy outputs resulting from absorbance. fluorescence and light scatter of the molecules. corrects the data for chemical and instrumental factors that may adversely affect the output energies. and produces a quantitative measurement that is directly proportional to the concentration of the excited molecules (Holland. 1971). This system as a whole has an accuracy of t 0.002 absorbance units over the absorbance range of 0 to 2.0 units. This same accuracy is given for fluorescence and light scatter. Organic-rich pore water from Burke Lake and Houghton Lake sediments were used as a source of naturally fluorescing organic compounds. The concentration of organic carbon in the pore water was 92 ugC/ml for Burke Lake and 72 ugC/ml for Houghton Lake. The pH of the pore water was 7.6 for Burke Lake and 7.7 for Houghton Lake. The pH was adjusted in a set of samples to test for PH effects on the pore water organics. Twenty-five m1 of Burke Lake pore water was pipetted into each of five 50 m1'V01umetric flasks. The pH of two aliquots was adjusted down to 2.1 and 5.5 with dilute HCl. The pH of two Other solutions was adjusted up to 8.9 and 10.0 with 1N NaOH, All five were then diluted into volume. 14 The heavy metal ions Cu++ and Co++ were added to another set of flasks to study the metal-organic reactions. Twenty-five m1 aliquotes of the pore water were pipetted into 50 m1 volumetric flasks. To a series of the flasks. Cu++ was added in concentrations of 1, 2. 3, a, 5, 10. 15, and 20 ug/ml. Cobalt ion was added in the same concentra- tions to another series ofthe flasks containing an equal quantity of pore water from the same bulk solution. No metal ions were added to one set of the flasks but the total amount of pore water in the flasks was 5. 10. 15. 20 and 25 ml diluted to 50 ml. These latter samples were the standard organic solutions. The pH of the solu- tions containing the metal ions was adjusted to 7.3 t 0.1 and the samples were diluted to volume. A last set of samples was prepared which contained Cu++ and Co++ in distilled water. These were used as 'background samples and therefore no results from these solutions will be presented. Nitrogen was bubbled through each solution to purge :it of oxygen. The temperature of each sample was 22°C. {Then a 1 ml portion of each solution was individually excited by a source of energy through wavelengths in the tzltraviolet spectrum. Light scattering. a function of size and quantity (of particles. was used on each set of solutions. Exci- tation was set at 330 mm and the emission scanned from 300 mu to 360 mu. The computer collected the data. 15 corrected it (Holland. 1971) and yielded an instant output in quanta of energy (Figure 3). The light scatter spectra of these solutions are shown to have a maximum intensity at 330 mu (Figure 3). The fluorescent and absorbant intensity of each set of solutions was then measured. The intensities of the fluorescence and absorbance emissions of each sample were measured simultaneously throughout each scan. The emission energy for each was set at 380 mu while the excitation spectra was scanned through wavelengths ranging from 250 mu to 360 mu. The computer yielded absorption and corrected fluorescence curves on a single output. The error as grand variation mean for the absorbance :and.f1uorescence experiments is 3.3% (Appendix. Table A6). Specifig Gravity of Orggnic Flocs The specific gravity of the organic floc (precipi- “hated colloid) was measured using Burke Lake pore water :1n.order to determine its potential sedimentary behavior. ESodium hydroxide was added to a pore water solution (causing flocculation of the colloidal and soluble organics. lifter precipitation. the water was decanted and the preci- Ioitate and remaining solution poured into two 50 ml Ioicnometers. The picnometers were placed in a 21° constant itemperature bath for 15 minutes. They were then weighed Vvith the concentrated floc, refilled with distilled water. again placed in the 21° bath and weighed again. The 16 organic floc was dried and weighed. In a similar experi- + ment Cu + at a concentration of 25 ug/ml was used as a flocculation agent. Order of Reactivity of Hgf+._Cu++. and Co++ with Organic -_—7 Compoundg The significance and intensity of the Cu++-organic reaction was better understood by a competition experiment using a supposedly strong reactor (Hg++) and a weaker :reactor (Co++). ~Twenty-five ml of Burke Lake pore *water was pipetted into each of six, 50 ml. volumetric flasks. The three metals HgI+. Cu++. and Co++ were individually added in equal quantities such that each :flask contained an increased concentration of each metal. 'their concentrations ranging from 1 ug/ml to 20 ug/ml. JEach.f1ask contained a set concentration of 46 ug/ml organic (carbon. The solutions were adjusted to a pH of 7.1 and Ixrought to volume. After two hours the samples were laillipore filtered. The small pore size of the filter (0.45m) facilitated removal of all the particulate and finest of the colloidal organic compounds. This method Should have removed almost all of the nonionic heavy hmetals (Riley. 1938; Chau, Y.K., personal communication. 1973). The filtrate was analyzed by atomic absorption spectrophotometry for metal species remaining in solution. In a similar run, Cu++ alone was added to the flasks and 17 the same procedure of filtration and analysis was done. The error in this experiment is developed in Appendix. Table A3. Test of Chelation of Heavy Metals by Organics Cupric ion was added to a Burke Lake pore water solution and the pH adjusted. The pH was adjusted in four of five test solutions such that two solutions had a pH above and two solutions had a pH below 7.3 t 0.2. the pH of the original solution. Each solution had a copper concentration of 4 ug Cu++/ml. Two hours later. half of each solution was Millipore filtered (0.45 u) to extract particulate and high molecular weight organics. A clear filtrate resulted from the yellow-colored solution. The copper remaining in solution was considered unreacted. free Cu++ and its concentration was found by atomic absorption analysis. The quantity of Cu++ adsorbed by the organics filtered from solution was calculated by difference. Then. the remaining solutions containing the Cu++-pore water mixture. were readjusted back to a pH of 7.3 t 0.4. Each solution was again Millipore filtered and the Cu remaining in solution measured. The precision of this experiment is developed in Appendix, Table A4. METHODS - II Tests on Sediment-Metal Interactions Rates of Reaction The rates of copper and cobalt uptake were inves- tigated using Burke Lake sediment as a reactant. The reaction temperature was ambient (22°C to 2400), sedi- ment pH was 7.4 t 0.2. mean grain size was 4 u. percent carbon was 35 and sediment water content was 73 percent. Cupric ion was added in quantities of 0.5. 2. 5. 10. 25, 50. 75. and 100 mg to a series of solutions containing 25 m1 of wet sediment. The solutions were all diluted to 125 ml. To a similar series of sediment solutions. Co++ was added in quantities equal to the Cu++ additions. After two hours and every day for 11 days, the concentra- tion of heavy metal in the solution overlying the sediment was measured by atomic absorption analysis. Then. each sediment was dried and weighed. The precision of the methods used and the fit of the resulting data to a linear curve is developed and discussed in Appendix. Tables A9 and A10. Quantitative Aspects of Uptake To find the extent of the organic-heavy metal reaction.in.the general model (Figure 1) in terms of log Q, 18 l9 ++ . Cu . and as a comparison. 00'”, were allowed to react *with Burke Lake sediment and pore water. Cupric ion. ranging in concentration from 1 to 20 ug/ml. was pipetted in increasing quantities into a series of 50 m1 volumetric flasks containing 25 ml of Burke Lake pore water. The ‘pH of'the solutions was adjusted to 7.3 t 0.1. The pore water solution was analyzed for organic carbon according to the methods of Van Hall. gt ai” (1963). The pore water contained 46 ug/ml of organic carbon. After two hours and a probable reaction steady state for metal- pore water reactions. each solution was Millipore filtered (0.1k5 u) and the copper remaining in solution measured by atomic absorption analysis. Burke Lake sediment was also used as an adsorbant. Cupric ion was added in duplicate in concentrations of 2.172. 10.86. 27.15. 54.30. and 5.430 mg to a series of flasks; containing 25 m1 of wet sediment. Each was brought to a total volume of 100 ml. After an initial shaking the mixtures were allowed to stand at room temperature (22-24530) for five days. The solutions were then analyzed for capper. Cobalt was tested for log Q in the same manner? as copper. The concentrations of Co were 2. 5. 10. 20. and 50 mg. In one last test. the time dependency of the log Q Values of the Cu++ and Co++ reactions were investigated. For this test the results of the rates of reaction test 20 given above were used. In Appendix. Table A11. A12. and A13. the error of the method and the fit of the data to the results is developed. Pathways of Cu++ Within the Sediment To test for the various forms (phases) of Cu in 'the sediment Cu++ was added to wet Houghton Lake sedi- ments (Table l) in a concentration of 443 rug/500 m1 sediment (6.569g dry weight) and extracted after two hours and again after nine days. The sediment contained 60 ngCu/gm as background but this was considered negligible in affecting" the general trends. since the added Cu was 6.74 mg Cu/gm of sediment. The Cu++ was intimately mixed into the sediment with a magnetic stirrer. A fraction was extracted. split into four nearly equal 20 m1 portions and centrifuged. A resulting supernatant pore water was decarrted and collected. The sediment was then washed four ‘times with a 5 percent KCl solution followed by a single washing with distilled H20 to remove surface adsoriaed Cu++ (Mortland. personal communication, 1971). The washings were collected in four volumetric flasks and diluted to volume. Four washings with 0.01M Na-EDTA were then completed and collected in volumetric flasks. Na-ED‘I‘A is a very strong chelating agent of Cu'" and ‘Will extract Cu++ from weaker complexes such as natural metal~organic complexes (Knesek, personal communication. 1971). The samples were then dryed at 110°C. and weighed. 21 Finally. 5 g of KMnoz, and 5 m1 of sulfuric acid was added to each sample. and the mixture was digested for 30 minutes to solubilize the remaining Cu++. The percent variation between the four samples in the three extractions is (1) KCl = i- 2%. (2) EDTA = t 5%. and (3) KMnOu, = t 4%. The error and confidence limits of the various extractions is developed in Appendix, Table A14. RESULTS AND DISCUSSION Heavy Metal-Organic Interactions in Sediment Pore Water Solubility of the Orgnics The sediment organic matter-heavy metal reactions were studied by investigations on the reactions of the organic compounds in pore water. The pore water organic compounds were utilized to study the organic-metal inter- action because they are: (l) representative of the ‘sediment organic matter. (2) are highly reactive chemically. and (3) have fluorescent. absorbant and light scatter ‘properties. Some of the properties of the pore water organics were studied by utilizing their changing response to ultraviolet light with changing chemistry. Light scattering which is a function of the size and quantity of particles. was used in determining the solubility of the organics in the pore water solution containing a variety of electrolytes. First. the effect of pH on organic solubility was tested (Methods I. Light Scatter of Pore Water Interactions). Organic-rich pore water (92 ug C/ml) was extracted from Burke Lake sediment by centrifugation. The pH of the resulting series of pore water samples was adjusted in the range 2 to 10 and a light scatter spectrum having a peak intensity (c) at 330 mu was obtained from each sample 22 23 The relationship of light scatter peak height at 330 mu to pH change suggests that solubility of organics in solution is inversely related to an increase in pH (Figure 4). The organic molecules and colloidal surfaces are negatively charged (Neihof and Loeb. 1972) and thus. OH- effectively flocculates the organic molecules. As flocculation begins many nuclei form. and molecules begin to floc together. The increase in size and number of particles increases the light scatter. The effect of increased concentration of divalent heavy metals on organic molecule solubility was then tested using Cu++. Light scatter showed that as the concentration of Cu...... increased. the flocculation of the organics increased (Figure 5). The flocculation increased until an endpoint was reached and no further change in uI+ concentration. light scatter occurred with increasing C Thus. either the affects of increased Cu++ concentration increased ionic strength. or both caused flocculation of pore water organics. The increase in ionic strength was only about 10'5 so it is suggested that increased Cu++ concentration is the major factor causing flocculation. The results shown in Figures 5 and 6 are similar to those relationships found by Krauskopf (1967) for the effect of electrolytes on colloids. These are: (l) diva- lent and trivalent ions are more effective in coagulating colloids than are univalent ions. (2) H+ and OH- although univalent are especially effective as coagulants. and 24 Gem pagoda Housman unmfia camsmm .m madman 3:: .1523 m><>> 0mm _ _ _ oom (17:) man wninvno salivos BAIIV‘IBH 25 12— RELATIVE SCATTER PEAK HEIGHT (cm) ((5) O) l ¥ 1 l l l l 0 2 4 6 8 10 12 pH Figure 4. Dependence of organic solubility on the pH of the solution as shown by light scatter. The dotted line represents the data linearly. (Data in Table A1, Appendix) RELATIVE SCATTER PEAK HEIGHT (cm) (Cb) 26 END POINT 4— 2... I I I I I o 4 8 12 16 2o COPPER CONCENTRATlON (ug/mn ‘ ++ . . . . . Flgure 5. Effect of Cu concentration on organic solub111ty—11ght scatter. The data are derived from light scatter outpng at 330 mu. A reaction end point occurs at 13 ug/ml of Cu in the organic rich solution. (Data in Appendix, Table A2) 27 (3) some electrolytes (some divalent transition metals) are more effective as coagulants than other equally charged electrolytes. Spegiiic Gravity of Orggnic Floc The specific gravity of the flocculated organics ‘was investigated to find what effect this might have on the sedimentation or mobility of a metal contained in the floc. The specific gravity of the organic floc is an average 156. When the flocculating agent is Cu++ at 20 ng/ml of pore water, the specific gravity of the resulting floc is an average 2.1. Cline. at El. (1973) have shown that the heavy metal, Hg++, is also incorporated by the organics in pore water and a heavy metal-organic floc results which has a low specific gravity. A heavy metal-organic interaction could result in heavy metal mobilization as part of a low specific gravity organic floc. Even as a precipitate, natural organics can be quite mobile in the aquatic environment. Clays are the last inorganic sediment to settle from a dynamic environment due to their small size. platy shape. surface charge. and a specific gravity ranging from 2.1 to 2.9 (Mason and Berry, 1968). The specific gravity of the organics is less than that of clays which means that they and associated substances such as heavy metals will be 28 easily and quickly mobilized in a system with minimal turbulence and demobilized only in very quiet aqueous environments. ++ Order of Reactivity of Hg++, Cu , and Co++ Wi_h Organic Molecules The interaction of a heavy metal with an organic molecule depends not only on the chemistry of the organic compound but to a large extent on whether a heavy metal can be complexed. The extent of reaction of different metal ions with organic complexing agents is related to the size. configuration, and charge of the ions. The relative reactivities of the heavy metals are listed with soil organics (Kononova. 1966: Schnitzer and Skinner, 1966: Broadbent, 1957) and with natural and synthetic chelating agents (Knesek, unpublished: Weast, 1968). To understand the significance and intensity of the Cu++-organic reaction. it is necessary to compare its reaction with natural organic compounds to that of other heavy metal cations. The reactivity of Cu"+ with natural organics as compared to the reactivity of Hg++ and Co++, was deter- imined by means of a competition experiment using Burke Lake sediment as a source of natural organics. When only Cu++ ‘was added to the pore water solutions (Figure 6), it was adsorbed (adsorption here is any metal-organic bond) in greater concentration than in the solution which ++ contained Hg++, Co , and Cu++ (Figure 6). The range of 29 the data points in Figure 6 are t 0.1 ug/ml or within the circles around each point. (See Appendix. Table 3A). ++ ++. Cu ) compete for available These metals (Hg++, Co sites. supressing the adsorption of each other in varying degrees. For anunknown reason the metals do not suppress adsorption of each other to the extent that the total adsorption of Hg++. Cu++. and Co++ in the mixture adds up to the adsorption of Cu++ when it alone is in solution. Each type of metal ion may be adsorbed by sites specific to only that ion and not the others. Cobalt does not success- fully compete with either Cu++ or Hg++ for organic adsorption sites. but Cu++ and Hg++ compete with each other. Mercury (Hgf+) suppresses Cu++ uptake and is adsorbed by the organics in approximately equal concentrations to Cu++. Copper likely suppresses Hg?"+ uptake also. At about 15 to 20 ug/ml of HgT+, Cu++. and Co++ the total metal ion concentration in solution is a high #5 to 60 ug/ml and inorganic precipitation and/or inorganic colloid formation may cause the apparent changes in adsorption character. Inorganic colloids may be filtered by Millipore filtration (O.fl5 u) and may cause the apparent increase 0 ++ o o o o + in Co adsorption and continuing increase in Cu+ adsorption while Hg++ drops off. The following order of reactivity of the three heavy metal ions results: ++ Hgf+. Cu++ :>’Co 30 The above order of reactivity is in general agree- ment with that of the same ions with EDTA (Stary. 1964). At 25°C EDTA has a log Ks with Hg++ of 21.8, with Co++ 16.6 and with Cu++ 18.8. Cobaltic ion is much less reactive than Cu++ with the natural organics and is less reactive than would be predicted from the EDTA stability series. Mercuric ion and Cu...... are not only more strongly chelated than cobalt by given organic sites. but they also have more types of sites available to them for reaction. As an example of this. Broadbent (1957) showed by chroma- tography that the effluent of an acid washed column containing Cu++~ and Ca++-organic complexes show four elution peaks with copper and two with calcium as acidity is increased. He concluded that there is a retention of copper by functional groups which do not combine with calcium. Cobalt cannot react with as many functional groups as metals higher in the stability series because it is not of favorable size for maximum chelate substitution. Charge is important. but in this case all three ions are equally charged. 0f the three metal ions 4. then. Co++ is an example of a weak reactor and Cu+ and ++ strong reactors with organic ligands. It follows that ++ Hg any study on Cu in a natural environment must account for its strong reaction with natural organic ligands. 31 Figure 6. Competition between Hg++, Co++ and Cu++ for adsorption sites on organics in pore water. A comparison is made of the mixed metal ion (Hg++ + Cu++ + Co++) adsorption and the non-competitive adsorption of Cu++ onto the organics. The closer a metal is to being totally adsorbed by the organics, the closer it will come to the "Total Adsorption of M*+" line. The range in determinations is shown by the bars on the data points (See data in Appendix - Table 3A). 2 OJ 20- 18— 16- 14- P 8 _ _ 2 0 1 1 crime 838% ++ 2 no zoEEEmozoo 6— CONCENTRATION OF M++ ADDED (ug/ml) 33 Test of Chelation of Cu++ By Natural Organics The bonding or association between a heavy metal and an organic molecule is often due to chelation of the cation. Cupric ion was allowed to react at varying values of pH with Burke Lake pore water to test for chelation (Methods I. Test of Chelation of Cu++ by Natural Organics). It was found that the percentage of ++ that becomes bound to the filterable substances. total'Cu such as the larger particulate and colloidal organic material. increases in direct response to a pH increase (Figure z,curve a). At a pH greater than 7.5 a copper hydroxide precipitate may also form but the increasing reactivity of the organics likely minimizes this reaction (Krauskopf. 1956). Thus. the reaction between the natural organic compounds and heavy metals is pH dependent. To find if the uptake of Cu++ by the organic mole- cules is reversible. each pH adjusted solution was brought back to the original pH of 7.3. The quantity of copper :remaining in solution or the percent of copper inall of ‘the solutions is approximately the same (Figure 7. curve b). Tflie metal reaction with the organics is thus reversible. Thierefore. the natural organic compound-heavy metal :reaction in the aquatic environment and in the model fulfills the theoretical requirements of complex compounds (Martell. 1971; Stevenson and Ardakani. 1971) proving that natural organic molecules complex and may chelate the heavy metals . 100 h- 80 F- 60 —- 40-— % COMPLEXED COPPER 20'— Figure 7. 34 I I l l I I 2 '4 6 8 10 12 14 pH Equilibrium of Cu—organic chelate sh wn by pH adjustments. Curve (a) is a plot of percent of Cu uptake by organics versus pH of the solution while curve (b) is a plot of the percent Cu uptake of each sample after readjusting to a pH of 7.39 i 0.03. (Data in Appendix, Table A4) 35 The natural metal-organic complexes can usually act as metal chelates. depending on their stability. A test on a complex to find if it is a chelate is to find if a dilution will cause release of metal ion from the complex (Martell. 1971). Most natural waters are deficient in heavy metals even though their sediments may be enriched with heavy metals. Thus the complex or chelate in the sediments is stable even in dilute solutions of heavy metals such as natural waters. As the Cu” concentration decreases in an organic-rich solution (Figure 6. Cu++ alone). inie percent of Cu++ in solution does not increase. Thus. (tilution does not cause a release of metal ion from ruatural organic complexes indicating that the complexes are chelates. Extent of Organic-Metal Reaction in Porejlater Solution The heavy metal-organic interactions in solution were studied by means of the fluorescent and absorbant properties of organic compounds found in pore water (Methods I). Example fluorograms and absorbance spectra of Burke Lake pore-water solutions containing increasing quantities of Cu++ are shown in Figure 8. The quantity of organic molecules, pH, temperature and machine Operating conditions (Holland. 1971) are the same for each example. The emission energy for each is set at 380 mu. Each solution is individually excited by a source of energy through wavelengths ranging from 250 to 360 mu. The intensity 36 1 ugnnICu+i'Added EsugnnICu+*'Added FF I III‘T I I I I T fl I I I T I I I 250 360 250 360 10 ug/ml Cu++ Added I I I T I I I I I I 250 .360 250 360 Flgure 8. Example fluorograms of an organic rich pore water solution containing a constant carbon content and an increasing concentration of Cu++. Fluorescence (FL) and absorbance (Ab) are simultaneously measured 'and recorded as shown. The vertical scale is in relative quantum units (5) while the horizontal scale is in wavelength (mu) of emission. 37 of absorbance and corrected fluorescence emissions (Holland. 1971) or quanta are simultaneously measured throughout the scan. Figure 8 shows that each entire fluorescence spectrum decreases in intensity while the entire absorbance spectrum increases as copper concentrations are increased. A heavy metal causes a decrease of organic fluorescence in three ways: (1) the metal ion can cause the fluorescing organic to precipitate from solution: (2) when the metal ion reacts with the fluorescing molecule it stabilizes the resonating electrons of the fluorescing molecule: and (3) it can. without reacting. dampen fluorescence by its closeness to the electrons of the fluorescing molecule (Holland. personal communication. 1972). Absorbance often acts in an opposite manner to fluorescence. Every molecule must absorb before it can fluoresce. Non-fluoresc- ing species however. can also absorb. Often. when a ligand makes bonds with a cation. absorbance will increase. Thus as reactions such as the cation-organic ligand reaction occur with fluorescing molecules. fluorescence will decrease while absorbance usually increases. If a cation is only surface adsorbed and no reaction occurs. absorbance will not increase (Holland. personal communication, 1972). But, when no bonds are made. fluorescence will still decrease due to close association of the metal ion to the resonating electrons of the organic molecules. 38 In comparing the four spectra in Figure 8. the affect of cOpper concentrations on the intensity of fhuorescence is observed. At low concentrations. Cu++ is surrounded by a maximum number of ligand functional groups. After the more highly reactive sites are occupied. the Cu++ must fill less desirable sites where it is surrounded by waters of hydration (Martell. 1971). Then,at a cation/ligand ratio at which the cation is only weakly complexed. it should no longer affect the fluorescence of the molecules and an endpoint is reached. Each spectrum in Figure 8 is a result of a mixture of fluorescing species. many of which are in the same molecule. At every wavelength in the entire spectrum the intensity of fluorescence decreases as the fluorescing species react or as concentration decreases due to preci- pitation. The change in fluorescence is slightly different at each wavelength because the character of the fluorescing molecules varies. Thus. to observe heavy metal organic reactions in the following experiments. changes in fluores- cence are measured at a series of specific wavelengths. Figure 9 relates the intensity of fluorescence (Figure 9. a) and absorbance (Figure 9. b) at specific wavelengths versus the total concentration of copper. The initial rapid decrease in fluorescence indicates that Cu++ is interacting with a maximum number of organic molecules. The inflection at h to 5 ug Cu++/m1 suggests a reaction endpoint of copper with highly reactive. organic functional 1|" I Figure 9. 39 Change of fluorescent (a) and absorbent (b) intensity of dissolved organics in pore water at specified wavelengths (A) with an increase in Cu++ concentration. E.P. = endpoint. I.P. = inflection point. (Data in Appendix, Table A5). 140 ON seems so no 29232528 m: m: w o d _ .4 Va 4 NH _ ON a _ _ a —1 EV VH ) W3 NI ALISNELNI lNBBBOSBV I ,0 (‘l . gamma :0 no 292255200 v ON wH m: 3 NH OH m w A a q % _ T j — 35mm u < 1E 0mm H 4 0 1E mam ...I. K O W AS A: NM -LNBOSBHOITH )WC) NI ALISNHLNI ( L 41 group. The decrease in slope after 5 ug Cu++/ml implies that capper is not incorporated as energetically because the most stable bond sites are occupied. At concentrations greater than 5 ug/ml. the Cu++ is not surrounded by a maximum number of functional groups and thus it cannot affect the organic fluorescence to as great an extent. The final change in slape at approximately 15 ug Cu++/ml is likely due to saturation of most of the stable bond sites of the organic molecules. The Cu...... no longer affects the fluorescence of the organic compounds after this concentration and an endpoint is reached. The endpoint on the absorbance curve is not as apparent because with increased Cu++ concentration. some residual reaction causing non-fluorescent related absorption even though the organic-metal reactions are essentially complete as shown by fluorescence. To calculate the quantity of organics that had reacted with the Cu++. comparisons were made between results from Figure 9 and a standard organic curve (Figure 10) (Methods I). The fluorescent intensity of the standard organic solutions was plotted for three wavelengths versus the known percentage of pore water or concentration of organic carbon in solution. All of the metal containing solutions originally contained 50 percent pore water or 46 ug C/liter. As the metal ions react with the organics. the fluorescence of the solutions decreases to some intensity. Comparison of this new intensity to the standard plot at a specific wavelength. will give a 42 measure of the percent unreacted pore water or percent unreacted carbon remaining in solution. The percent pore water or quantity of organic carbon that has reacted at an inflection point or an endpoint can then be calculated. In the Cu++-organic reactions (Figure 9) an inflec- tion point occurs at 4 - 5 ug Cu++/m1 in solution. A comparison of the intensity of fluorescence at this wavelength. to the intensity of fluorescence of the standard curves show that approximately 29 percent of the organics from the original 50 percent pore water solution are still fluorescing. Thus 29/50ths or 58 percent of the organics in solution no longer fluoresce or fluoresce at a lower wavelength. The quantity of fluorescing organic carbon dropped from 46 ng c/hl to 27 ug c/hl or 19 ug c/hi. At the second inflection point or apparent endpoint. 30 ug C/ml had stopped fluorescing or fluoresce.at a lower wavelength. Only 5 ug Cu++/ml reacted with 19 ng/ml of organic carbon while it took three times as much Cu++ (15 ug/ml) to react with only 11 ug/ml more of the organic carbon. Probably. 66 percent of the organic molecules in Burke Lake pore water had functional groups capable of reacting with Cu...... but #2 percent of the organics had functional groups strongly reactable with Cu++. Figure 9 also gives information on the absorbance change of the copper-pore water solutions. The absorbance data show a possible reaction step function as Cu++ Figure 10. 43 Standard organic fluorescent curves at specified wavelengths (A). These lines are used to calculate the amount of organics still fluorescing in the heavy- metal. organic reactions. The peak height is measured on the fluorograms of the metal containing solutions. The percent pore water still fluorescing is found from these curves. (Data in Appendix. Table A6). LILI ORGANIC CARBON (ug/gm) 9 19 28 37 46 I I I I 3M: 255m}; 14 12 /\= 270mg 10 Q A: 330mg FLUORESCENT INTENSITY IN CM (6p) g l I I O 10 20 30 4o 50 % PORE WATER IN SOLUTION 45 concentration increases. This step function reaction. especially the fluorescence step function. gives further evidence that Cu++ is being chelated by natural organics. Stevenson and Ardakani (1971) postulate that chelate reactions occur in a stepwise manner. Cupric ions appear to react with the natural organic compounds in a step- wise manner. This, together with former evidence (reversibility and pH dependency of the metal-organic reaction) reinforces the hypothesis that Cu.” is not only complexed but chelated. to a large degree. by natural organics in solution. The fluorescence and absorbance changes of the pore water solutions to which Co++ are added (Figure 11) are different than Cu++ induced changes. A weak inflection in the fluorescence curve (Figure 11. curve a) occurs at 3 - 5 ug/ml of added 00“". a broad endpoint at approxi- mately 15 ng/ml of added Cd” and a decrease of slope occurs after each inflection. Unlike the copper-organic reaction only 6 ug/ml of organic carbon (13 percent) had reacted with 00'” at the inflection point and 27 ug/ml of Organic carbon (29 percent) had reacted at the endpoint. In comparison to Cu++ then. 00‘” reacts with fewer Organic molecules and/or fewer functional sites. With an inc:lr‘ease in 00'” concentration. fluorescence was slightly depressed but absorbance showed no change. This indicates that either weak bonds or no bonds are formed between the Co“ . ++ and the organic ligands. Co depresses the Figure 11. #6 Effect of Co++ concentration on or anic fluorescence (a) and absorbance (b?. A fluorescence endpoint (E.P.) appears at 15 ug/ml of Co+ while an inflection point (I.P.) in the slope of the curve occurs at 3 to 5 ug/ml Co++ added. (Data in Appendix, Table A7). RELATIVE FLUORESCENT INTENSITY-CM ((1:1) RELATIVE ABSORBENT INTENSITY-CM (ab) 12 p-i 0A 10 a? (b) 43 A = 270mg —O /\ : 330mg I I I I I I 2 4 6 8 10 12 14 I6 CONCENTRATION OF COBALT (ug/ml) .— _ I 18 I 20 '48 fluorescence of the organic molecules weakly due to its closeness or physical adsorption to them and not reaction with them. Cobalt ion has a much lower tendency toward covalent bonding character than Cu++. It is hypothesized that the association of Co++ with natural organic mole- cules in solution is by surface adsorption. Formation of Metal-Organic Flocs The effect of Cu++ and in comparison. Co++- organic reactions on the solubility of the organic mole- cules in pore water was investigated by means of light dispersion (Holland. 1971). Colloids. like those in pore water solutions. disperse light to a greater or lesser degree depending on the size and number of the colloidal and flocculated particles. The effect of the heavy metals Cu++ and 00‘” on the solubility. size and/or number of colloidal particles is demonstrated by the results of light scatter spectra shown in Figure 12. As Cu” concentration increases an increase in light scatter results. This effect of Cu” on the organics in solution is due to increased particle size (flocculation). an increase in the number of particles in solution. organic- metal colloid formation. and/or precipitation of suspended Organics in solution. An endpoint in flocculation occurs at 13 us Cu++/ml (Figure 12). This agrees closely With the absorption and fluorescence data (Figure 9) in Which an endpoint occurred at 15 ug Cu++/ml. The chelation 49 CU++ A (f) V 2 0 Z (75 Z :1 z CO++ m E < D (’3 DJ 2 .— <1: _1 DJ E 0 I I I I I 4 8 12 16 20 CONCENTRATION OF M++ (ug/ml) Flgure 12. Comparison of Co++ and Cu++ effect on organic ~solubility in pore water using light scatter. Increased light scatter of solutions caused by.formation of colloids or floccules yielding . increased scatter. (Data in Appendix, Table A8). 50 of Cu++ by the organic molecules thus causes flocculation and will result in eventual precipitation of a metal- organic complex. Cobalt interaction with the organics (Figure 12) has no apparent affect on organic particle size and number at less than 20 ug/ml in these pore 4.. does not cause water solutions. Therefore. the Co+ flocculation and/or precipitation of an organic heavy- metal floc as readily as Cu++. It does not strongly attract organic ligands and will not readily cause a flocculation. These data support the assumption that the less reactive Co++ is only surface adsorbed onto the organics while Cu++. with its strong covalent bonding character. is chelated by the same organics and flocculates them. Since Co++ is not incorporated in the same manner as Cu++. it does not cause organic flocculation with the possible exception of very high concentrations where ionic strength is the governing factor. It should be noted that the highest concentration of Cu++ used was only 3.2 x 10-6! and Co++ was 3.3 x 10.6m. The increase in ionic strength due to these metals was negligible. Thus. if a heavy-metal interaction with soluble organic molecules is a strong chemical interaction as with Cu++. and not merely a weaker interaction like physical adsorption as with Co++. the metal ion can cause a precipitation of a metal organic floc at very dilute concentrations (3.2 x 10’624). 51 -The consequence of the Cu++ organic reaction is a control on the dispersion of the heavy metal depending on the energy of the system. If heavy metal-organic inter- action occurs in a turbulent system such as a river. it could have a mobilizing affect on the heavy metal due to the relatively low specific gravity of the organic preci~ pitate or due to the solubility of some organic-metal complexes. Conversely. in a non-turbulent situation the reaction could have a demobilizing affect on the heavy metals. In any case. Cu++ readily reacts with organic molecules in solution. causes a flocculation of a metal organic complex. and eventually becomes incorporated in the bottom sediments. Sediment-Heavy Metal Interactions Rates of Rgaction The time dependency of Co++ and Cu++ uptake into . seadiment is demonstrated (Figure 13) by the two solutions mo st highly concentrated in their heavy metal (Methods 11). Tfiie initial metal-sediment reaction rate could not be measured. During the first two hours approximately 85 percent of the metals are removed from solutions by crielation. adsorption. or inorganic precipitation. For tlie next 11 days each metal is slowly removed from solution in a probable diffusion dependent rate. Copper appears to react with and diffuse into the sediment at a faster rate than 00‘”. Cotton and Wilkinson (1968) have found that the rate of reaction for a given ion has little or no dependence on the identity of the ligands. The rate of reaction. then for Cu++ with the natural organic ligands would be approximately 108/sec and for Co“. los/sec (Cotton and Wilkinson. 1968) which is the same rate of substitution of various aquo ions. To determine the rate constants of Cu++ and Co++ uptake and diffusion into the sediment (Table 2) and the order of the reactions. first and second order rate equations were employed. In a first order reaction 52 Figure 13. 53 Time dependence of Cu++ and Co++ reaction with organic rich bottom sediments. Up to 84.5 percent of the metal ions react immediately. Aftgr 11 days 94 percent of the Co and 99.5 percent of the Cu++ is no longer in solution. (Data in Appendix, Table A9). 54 + + +0 +U C C a fi \ . m m .T w H. c c_ \ .T _ \ .T \K .7 \ .1 :1 \ T \ I \ I I IIrIIILIIILIIITIIIwIIIuIIIlmIIIImvN 1 _ m w w m m w w m m m is 20:38: a as ++:o TIME (Days) 55 the rate depends on a single species reacting to form a product (Castellan. 1966) or on one species of an interacting pair which is much more concentrated than the other. Thus. the products are not involved in the rate law. If a is the original concentration of the reactants (A) and c the reactant concentration at time t. the first order rate reaction (Denbigh. 1968) is: 2.303 log (c/a) = - Kt The concentration of A decreasesexponentially with time. A plot of log (c/a) versus t will be linear if the reaction is first order and the slope of the line will equal K. the rate constant. The second order rate reaction (Castellan. 1966) is: 1/b = 1/2 + Kt A plot of l/c versus t should yield a straight line if the reaction depends on the concentration of both species and is second order. Both equations are related since the first equation is the limit of the second equation. In other words. in a second order reaction involving the concentration of two species. if one of the species becomes very dilute in comparison to the other. the rate will appear to be first order (Castellan. 1966: Denbigh. 1968). ‘When diffusion dependent uptake of Cu++ is plotted versus time. the 0.5 mg Cu”. 5 mg Cu” and 25 mg Cu++ solutions are first order and yield first order rate constants (Table 2). A first order plot of data from the 56 TABLE 2. Rates of diffusion dependent Cu++ and Co++ reaction with sedimept. First and second order rate constants (day‘ ) are shown (See Appendix - Figures Al and A2 for the rate plots used to find the K's). Cu ngl/ZS m1 sediment Order K (day'l) 0.5 1 0.262 50 l 00259 25.0 1 0.252 100.0 2 .....250 1 Av. K = 0.242 5 Co (mgflj ml: sediment Order K (day’l) 5.0 1 0.201 25.0 1 0.104 50.0 1 0.128 100.0 1 9_,_i_i__2_ Av. K 0.136 57 100 mg solution is non-linear. but a second order plot (l/c versus time) yields a second order rate constant. The first order rate equation therefore applies only to dilute concentrations of Cu++. though not enough data was taken to be certain. The rate of the diffusion dependent uptake of Co++ is only half as great as that of Cu++ (Table 2), but is a relatively constant first order reaction. Cupric ion maintains a nearly constant ++ follows the second rate of uptake but unlike Co++. Cu order rate law at high concentrations because the copper- organic diffusion reaction begins to depend not only on organic molecule concentration but also on copper concen- tration as predicted by theory (Castellan. 1966). The results in Table 2 show general trends and are not to be considered as repeatable in detail in another natural environment. In summary. 011++ diffuses approximately twice as rapidly as Co++ into Burke Lake sediments but both react rapidly at first as would be predicted (Cotton and Wilkinson. 1968). In the general model in Figure 1. then. this would mean that a high percentage of the heavy metals are immediately taken up by sediments but complete adsorption or uptake may be diffusion dependent and take several days. 58 Quantitative Asnects of Uptake The metal uptake by bottom sediment is rapid and likely irreversible due to the strong organic and inorganic reactions with the sediment. The extent of the heavy metal-sediment reaction is important for a quantitative estiJnate of the pathways of a heavy metal. Thus. the stability constants of the heavy metal-organic reactions wilil. be derived and the end point of reaction between sediments and a heavy metal enriched solution shown. To describe the adsorption of ions from a liquid onto a solid surface. the equation developed by Freundlich (1926) is log (X/M) = (l/h) log C + log Q where X/M = the quantity (meg/100 gm) of ions adsorbed per unit weight of adsorber (sediment in this case). C = The equilibrium concentration (meq/l) of the adsorbate after adsorption has occurred (Cu++, Co++ in this case) Q = The equilibrium constant n = A constant Aiplxot of log (X/M) versus log C should yield a straight line with an intercept of log (X/M) equal to 10s Q Where log C = 0. A high value for Q would result from a high) value of product or metal-complex versus the uncom- Plexed metal left in solution. 59 The result of a plot of log (X/M) versus log C for the Burke Lake pore water - Cu++ mixture is shown in Figure 14. The adsorption of Cu++ by pore water organics is a linear plot in the range of concentrations used and thus follows Freundlich theory. The total organic plot was derived for comparison of results to literature by using an assumed value of 640 for molecular weight of fulvic acid (Schnitzer. 1971) instead of 12. the molecu- lar weight of carbon. The data do not intercept log C = 0 but the calculated log Q values of 5.10 for a copper- organic carbon complex and 5.38 for a hypothetical fulvic acid complex of COpper were derived by least squares derivation (Appendix. Table A11. A12. and A13). This high stability constant shows that the organic-Cu complex is quite stable in natural pore water solutions and would be shown as irreversible in the generalized model (Figure 1). When Cu++ was allowed to react with Burke Lake sediment aslightly different. but analagous result to the pore water reaction was attained (Figure 15). The Freundlich Isotherm is not linear through the range of Cu++ concentrations employed. A point of saturation or Cu—sediment equilibrium may have been reached where the inflection point occurs. Riley (1939) obtained a similar nonlinearity. The log Q value calculated by linear regression was 4.29 while the graphically derived log Q would be 1.55. Since the sediment apparently became saturated before the concentration of Cu++ reached Figure 14. 6O Freundlich Isotherms for the reaction of Cu++ with pore water from Burke Lake sediment. In one case the weight of carbon that reacted with a given concentration of Cu++ was used in the calculation of log Q. In the second case the weight of carbon in solution was multiplied by 1.92 and the ratio (X/M) calculated. (Data in Appendix, Table All). 61 —1.0 1.1 KI I I I I I 3.- N. O. 03. o s. N. o (T) m N (\I N (W (\I (JBQJOSDP )0 LIIfi,:’pcq108pL3 ++ w saIOLu) (4,“)901 LOG C (moles of M++ not adsorbedjhter) Figure 15. 62 Freundlich Isotherms comparing the fit of data and the stability of reaction of Cu++ (o) and Co++ (o) with Burke Lake sediment. The dotted line connects the data but is represented by only one point and is therefore not necessarily a valid interpretation. (Data in Appendix. Table All). 63 mg wd v.0 me:_.\.UmQEOQO so: ++S_ _o 920:: Q 00.. 0.0| Vol wdl _ _ _ mgl cal ++zu AS OI o was ONI vNI ') 9m D (W saIOLu) ( (Jaqmspe IO (LIST poqrospw * 64 the intercept (log C = 0). the extrapolated log Q value of 4.29 is thought to be more realistic. When Co++ was tested with the Burke Lake sediment under the same conditions as Cu++. it followed the Freundlich Isotherm closely (Figure 15). Graphically it intercepts log C = 0 at approximately the same 10g (X/M) value as Cu++ but it is not as highly adsorbed as Cu++ initially. Cobalt concentrations do not reach a point where the adsorbant becomes saturated. Since the Freundlich Isotherm was derived for gas adsorption in a single layer onto a solid. then an aqueous ion such as Co++ that is only adsorbed onto a solid surface would closely approximate the isotherm. On the other hand. if the molecules also chemically react with the surface. an endpoint in the apparent high adsorption should occur as shown by the change of lepe of the Cu++ curves. Thus. the hypothesis derived earlier that Co++ is associated with natural organ- ics by surface adsorption while Cu++ reacts with the organic molecules is further reinforced. Assuming that the log Q values for Co++ and Cu"-+ sediment interaction are time dependent. calculations for log Q were made for solutions with reaction times ranging from 2 hours to 11 days (Figure 16) (Methods II). Table 3 summarizes these results plus those log Q values attained in the two previous experiments. The log Q values increase with time and reach an apparent maximum after 11 days. The maximum log Q of 8.99 for Cu++ compares favorably with 65 TABLE 3. Equilibrium constants (log Q) for Cu++ and Co++ interactions with bottom sediments from Burke Lake. The calculated values of log Q were obtained by linear regression on the Cu data before the break in the curves. The graphic stability constants are extracted directly from the figures. The contact time of the metal ions with the sediment are given. (Data in Appendix, Tables A11. A12. A13) Eigggg Adsorbant 1gp Lo values 16 pore water Cu++ 5.10*** 17 sediment Cu++ 4.29** 17 sediment Cu++ 1.55* 17 sediment Co++ 1.78% 18 sediment (2 hours) Co++ 1.63* 18 sediment (11 days) Co++ 1.91* 18 sediment (2 hours) Cu++ 4.59.W 18 sediment (11 days) Cu++ 8.99** * l. The value of Log Q is obtained by interpolation from the given figure ** 2. The Log Q value is calculated before any breaks in the curve by regression analysis. *** 3. The Log Q of the pore water solution is calculated by regression analysis using the weight of organic carbon for the adsorbant weight. Figure 16. 66 Freundlich Isotherms comparing the reactions of Cu‘"+ ( ) and Co++ (- -) with Burke Lake sediment after two hours (0) and fifter 11 days (0). The log Q values for Cu+ were calculated using the slope of the solid lines before the break instead of where they intersected LOG C = 0. (Data in Appendix. Tables A9, A13). LOG(fi) (moIes M ++ adsorbed/gm oI adsorber) .0 Q 9 o .— N I: m '00 '0 .0 o I o I\> OcI: an? Ool: Qme Oclm 30:3 \> Del» 3056 _ II_.III_ _ .IIIILIIILI 0.0 0.» cm #0 IN.O ITO IH.N IO.m I05 r00 0 “305m 0* _<_++ son momowcmaizowv 68 a log K value of 8.69 obtained by Schnitzer and Skinner (1966) who used fulvic acid extracts from soil as the adsorbent. The log Q values of Cu++ and Cd” are much less if the heavy metal has been in contact with the sediment for a short time and a steady state is not reached. Brief and variable metal-sediment contact time is prtfbably more the case in nature because fresh solutions arui fine grained sediments are continually passing over the; sediment. Pore water—heavy metal contact is not so 1variable and the log Q of 5.1 for Cu++-pore water is a viable estimate to be used in a model which represents a variety of non-equilibrium situations in natural aqueous systems. Cu'H' was found to have higher stability constants thazi Co++, as expected. but still. the constants for both arer sufficiently high to show that both will nearly quantitatively react with the bottom sediments and remain there. In the model then, Cu++ and Co++ have high enough stability constants with the bottom sediments that the overall trend is a quantitative uptake by the sediments. Endpoints of Heavy Metal Reaction with Sediments An endpoint of metal—sediment adsorption is deter- mined for periods of metal-sediment contact (Figure 17). The endpoints are shown at 35 mg Cu added to 100 ml solution plus 25 g sediment or a concentration of 8070 ug Cu/gm dry sediment. (Methods II). With time the endpoint Figure 17. End points for Cu++ reaction with Burke Lake sediment through time. The breaks in the curves are the approximate endpoints. At 0.5 mg Cu in solution. the Cu is toxic to small aquatic organisms. (Data in Appendix, Table A9). MG Cu IN SOLUTION mwooxoo Ln 'ow'xI'ooLo'o .09 .08 .07 .06 .05 .04 .03 .02 01 0 Toxic Cu I + 20 7O “—p I I 40 60 MG Cu ADDED 80 100 2 days 11 days 71 remains the same but the total Cu++ in solution decreases. If at least 5 ug Cu/gm remain in solution. then. the solution is toxic to most aquatic organisms (Riley. 1.939: Federal Water Pollution Control Association. 1968; Upchurch. 1973). As shown in Figure 17. if less than 40 mg Cu are in the sediment or 12,900 ug Cu/gm dry sedi- ment is present the solution is immediately detoxified. More adsorption of Cu” through time causes detoxification of solutions containing greater and greater concentrations of copper as will be demonstrated in detail later. The endpoints in Figure 17 show a trend which can be extra- polated in general to other environments but the actual numbers are good only under the given set of conditions (Methods II). In general. Figure 1? demonstrates that sediments can quantitatively remove high concentrations Of Cu++ from solution. Then even after apparent sediment saturation with Cu++. the Cu will be adsorbed even though the Cu++ concentration in solution is very low. When Cu++ was let stand for nine days with Houghton Lake sediments in the same manner as with the Burke Lake Sadiments. an endpoint occured at 7.500 ug Cu/gm in the sediments (Cline. In Press). This comparatively smaller adsorption capacity of Houghton Lake sediment is likely a result of the larger average grain size. the greater quantity of inert sand. and the smaller percentage of Organics (Table l) . 72 Figure 18 shows the quantity of Cu that must be in Burke Lake sediment as residence time increases. before toxic levels (5 ug/ml) of Cu are reached in solution. In a dynamic environment (2 hours residence time) containing organic rich sediments. more than 12.000 ug Cu/gm can accumulate in the sediments before toxic amounts of Cu++ will remain in solution (Figure 18). In a stagnant situation with greater than 4 days residence time. up to 34.000 ug Cu/gm can be bound in the sediment without toxic levels of Cu++ escaping back into solution. Mackenthun and Cooley (1952) exposed a variety of bottom- dwelling animals to bottom muds from Lake Monona. Wisconsin containing known quantities of precipitated CuSOu. Their tests indicate that the toxic effects of Cu become apparent around 9000 ug Cu/gm sediment. This figure compares. in general. with endpoints for copper adsorption in the dynamic. non-equilibrium environment of the Burke Lake sediment (8070 ug/ml) and the Houghton Lake sediment endpoint of 7.500 ug Cu/gm. Thus. in the generalized model (Figure 1). heavy metals react rapidly and quanti- tatively with bottom sediments resulting in detoxified aquatic solutions. Figure 18. 73 Concentration of Cu++ necessary in Burke Lake sediment before toxic concentrations remain in solution as a function of reaction time. Toxic Cu++ is considered to be 5 ng/ml in solution. MG (Eu/“GM SEDIIVIENT 36,000 34 32 30.000 28 26 24 I\) I\) 20,000 >—-‘ (I) 01 y—A A 12 10,000 2,000 74 9 Sediment Saturation TOXIC Cu+3L In Water p. — Cu++ Below TOXIC Levels In Water L. I. L. I I I I I I L I I I I O I 2 3 4 5 6 7 8 9 10 11 TI ME—DAYS (Reaction tIme) 75 Pathways of Cu“ Within the Sediment After only two hours of Cu++-sediment contact. the Cu++ was approximately 88 percent readily exchangeable as KCl exchangeable. surface adsorbed: EDTA extractable organic complexes: and pore water phases of Cu (Table 4) (Methods II). The KCl extractions removed 27 percent of the sediment bound Cu++ showing that a high percentage was only sufrace adsorbed onto organic and inorganic sites. The EDTA extraction removed 60 percent of the total sedi- nmnt bound Cu++. showing that Cu++ was complexed by organic ligands in the sediment. These data indicate in general that the initial Cu++-sediment reactions are adsorption and complexation. These percentages are trends and can only approximate results under different conditions. TABLE 4. Change of Phase of Cu in Houghton Lake Sediment with Time. (See Appendix. Table A14S for raw data and estimate of error. % Cu++ after % Cu++ after Phases 2 hours 9 days Pore water 2.8 0.2 EDTA extractable-organic complex 60.0 33.0 KCl exchangeable-surface adsorbed 27.0 5.0 KMnOu digestible-precipitated Cu 11.8 61.0 76 The original quantity of Cu++ bound in pore water. as surface adsorbed. complexed. and precipitated Cu++. continuously changes with time (Table 4). The precipitated, KMnOu digestible Cu++ increased from 9.8 percent to 66.9 percent of the total sediment Cu...... in nine days while the other forms of Cu++ all decreased in quantity. The difference in each of the four extracted forms of Cu++ is a difference in the amount of energy of the Cu++ sediment bonds. In pore water. Cu++ is in a soluble and/Or colloidal form as a result of sediment exchange and organic complexes that are soluble or colloidal. The Cu++ defined as surface adsorbed is the weakly bound Cu++g complexed Cu...... is fairly strongly bound. Thus. the forms of Cu++ change from initially weaker bonds of adsorption and chelation to strong bonds of precipitates through time. After the heavy metal is taken into the sediments. as shown in the generalized model (Figure l). the percentage of it that becomes tightly bound increases with time and/Or burial. Through time the changing chemistry of the sedi- ment due to bacterial action can cause a phase shift of heavy metals and possibly a cycling within the sediment column. To test for cycling of a heavy metal in the sediment. Cline and Upchurch (1973) mixed Cu++ (as CuClz) with organic rich sediment from Burke Lake. homogenized the solution. poured it into glass tubes, corked them on 77 the bottom. inserted them into the sediment for 1 week to 3 months. pulled them as a function of time, and analyzed them with depth. Cline and Upchurch (1973) showed that there was a rapid upward migration of Cu. After one week they found little difference in the Cu concentration throughout the sediment column. but by the end of four weeks, the Cu was highly concentrated in the tap of the core as compared to the bottom. It was suggested that the observed migration could be a result of compaction and dewatering. directional ion migration (diffusion). or decay of organics by bacteria and release Of complexed copper. Three different mechanisms for the rapid migration of cOpper in the tubes were suggested: (1) In tests on sea foam it was shown by Wallace (personal communication. 1973). that bubble surfaces are extremely reactive. Trace constituents such as heavy metals. phosphates and nitrates are concentrated many fold by the bubble surfaces. Therefore. it seems that the gases released by bacteria to form bubbles in sediment could act as a transport mechanism for heavy metals within the sediment. (2) There are bacteria that methylate mercury and cause its mobilization (Wood gt_gl.. 1968). No other heavy metals form a dimethyl or methyl gas but maybe gaseous organic chelates are formed. (3) Szalay and Szilazi (1968) found that Se and As migrate as hydrogen-metal gases in the sediment and undergo ion exchange with the iron in humates in the 78 upper few cm of sediment. become sulfide precipitates and remain in the top layer of sediments. A similar type of gaseous migration may happen with other heavy metals although no other metals are known to form the hydrogen- metal gas. To determine the phases in which the Cu is found and possibly identify the transport mechanism. Cline and Upchurch (1973) chemically partitioned cores taken after one and two months. Figure 19 compares percent Cu in the KCl exchangeable adsorbed form. the EDTA extractable complexed form. and the precipitated form-KMnOu digestion (Cline and Upchurch. 1973). The-quantity of Cu in pore water is insignificant and not included. Month 1 core shows an enrichment of the KMnOu digested phase of Cu in the upper 10 cm of the core. This indicates that as the copper migrates upward. it is reincorporated into the sediment (likely as sulfide. oxide. organic complex or carbonate) when it reaches the biologically active oxidized portion of the sediment. A comparison of the extractions on the core from :month 2 to the core from month 1 (Figure 19) show that the residual. KMnOh extracted. form or Cu increases at the expense of the less strongly bound adsorbed and complexed Cu with time. Thus. as found previously. the percentage of total Cu in the sediment that becomes ‘tightly bound increases. In the sediment of the generalized ‘mOdel. the results of the Cu interactions and cycling Figure 19. 79 Comparison Of Cu phases within a sediment column and with time of cores at months 1 and 2. The time comparison shows the Change in percent of Cu phase through 60 cm of core. (From Cline and Upchurch. 1973). 80 KMnO4 Month 2 I ........ M o n th KCL EDTA KCL 9O 70 0 0 5 mm(n »m(w NX —X VK mx<[ M (M (M M++/M+v—‘M++ Precipitate ) (Soluble Inorganic & Organic IIOOICI (Organic) :7 9 10cm - Permanent.” Eiiiriél 101 If a metal ion reacts with dissolved substances or mobile particulate matter. it will eventually precipitate and indirectly become part of the bottom sediments depending on the turbulence of the system. A heavy metal may leave the system by gaseous diffusion (i.e. Hg (CH3)2) or by incorporation into biota. When the heavy metal is initially incorporated by the sediment. a high percent of it is only weakly bound. With time and burial. the metal phases tend to change to a more tightly bound form. At the same time bacteria can cause a change in the environment of the metal species causing a drop in pH and Eh. The heavy metal may be mobilized and can be transported upward on bubble interfaces. by diffusion. or as a soluble organic complex in pore water. The metal will accumulate in the biologically active oxygenated portion of the sediment and become tightly bound there with time. Thus. the various heavy metal phases may cycle within the sediment along pathways governed by the micro-environments in the sediment. The direction of cycling apparently causes a heavy metal to concentrate within the upper sediment strata, even though losses of the metal occur due to permanent burial. solubilization. downward diffu- sion. and removal of metal containing biota. The model (Figure 21) places strong emphasis on organic-metal interactions. The organic reactions do not exclude inorganic reaction. but demonstrate the importance and often the predominance of organic-heavy 102 metal interaction in most environments. The organic based model was shown to be operational with Cu in natural situations and is therefore a practical tool for predicting the reaction pathways in the aquatic environment of Cu and probably most heavy metals. APPENDIX 103 TABLE A1. Data presented in Figure 4 concerning the dependence of organic solubility in pore water on pH of the solution. The relative intensity of light scatter in cm measured at 330 mu are shown at the given pH's. Relative Scattgp Intensipy (cm) pfl_ 2.0 2.2 6.5 5.2 6.6 7.5 7.5 9.0 11.1 10.4 The light scatter spectra were so reproducible that no difference between two spectra Of the same metal-organic concentration could be measured. The instrumentation has an accuracy of t 0.002 absorbance over the absorbance range 0-2.0 (Holland. 1971). This same accuracy is given for light scatter and fluorescent analyses. The error of the method will be developed further in Table A6. 104 TABLE A2. Effect of Cu++ concentration on organic solubility-light scatter as shown in Figure 5. The relative scatter intensity is measured in cm at 330 mu for each concen- tration of Cu in ug/ml. The peak heights are measured to the nearest cm. Relative ScattepHeight (cm) Cu ConcI (pglml) 5.7 0 6.5 3 7.3 5 9.0 10 10.4 15 10.4 20 105 TABLE A3. The data for the competition experiment shown in Figure 6 are presented below. The standards for atomic absorption analysis are given together with the relative percent of absorbance (Ab) and resulting concentrations of each metal in each sample after filtration. The percent absorbances for Cu++ and Co++ are given without decimals for ease of sample handling since the readings are all relative. The standard deviations (s) and confidence limits (a = 0.10) are given for the standards and the samples. 1 - 2 The standard deviation is given by S = L§§%l- (Krumbein and Graybill, 1965) for sample size less than 30: I is the mean of a set of samples Y and N is the number of samples. The confidence limits (t) for the samples and the standards is given by the formula I t ta g where ta will always represent the 90% confidence limit for this data. The confidence limits and resulting percent error is always calculated on absorbance. Co Standards Concentration Ap t S T t t E31 1.00 86. 99. 96 5.10 94 2.94 2.00 180. 178, 187 4.72 182 2.73 3.00 258, 265, 274 8.02 266 4.63 4.00 377. 382. 390 6.56 383 3.78 Average % S = 3.1 TABLE A3 Continued Concentration (umel) 1.00 2.00 3.00 4.00 1.00 2. 00 3. 00 4. 00 5.00 Conc. Add d @2111 1.00 3.00 5.00 10.0 15.0 20.0 Ab 236, 106 Cu Standards t 8 Ab "‘ DE). 68, 68, 65 1.9 128, 136, 137 4.9 206. 210, 200 5.0 260. 265, 269 4.5 Average % S Hg_Standards 306’ 309 0.2]. 10.1. 9.7. 9.5 0.31 14.9. 16.0, 16.0 0.64 21.9. 22.0, 22.8 0.49 28.1. 27.7, 28.0 0.17 Average % S dill Axsrase Ab Remaining Co Conc1 Cu (pggml) [1.3 -- 0050 85' 95 54 -- 0.83 240, 243 71 -- 1.10 195, 206 153 "' 2024 919 91 210 -- 3.00 132, 140 241 -- 3.46 145, 140 Ab dilII 1821 23.0, 23.4 1/100. 1/2 23 9. 24.4 1/100 23. 6, 22.8 l/lOO. 1/2 18.0. 17.5 1/100, 1/5 19.2. 19.1 1/100. 1/10 16. 5, 16.0 1/100, 1/20 NNH N \‘INMOW O O O \O \owmoow dil. fig 1/10 0.21 0.179 0.369 0.283 0.098 Average Remaining CO Concl;ll 1.00 2.78 4.60 10.0 15.0 16.4 Avepage Remaining Concj Hg OWHOOO O O O O O OUI\1\O-(=‘\O CDCDOU-(C'KAJ 107 TABLE A3 Continued Concentration of Adsorbed Metal Concp:§egin (uglmll Conc. (uglml) End Conc. m Adsorbed 92 fig 92. 92. fig 92 l 0.95 0.93 1.00 0.05 0.07 0.00 3 0.83 0.44 2.78 2.17 2.56 0.22 5 1.10 0.95 4.60 3.90 4.05 0.40 10 2.24 1.70 10.0 7.76 8.30 0.00 15 3.00 3.58 15.0 12.0 11.4 0.00 20 3.46 6.08 16.4 16.54 13.9 3.60 Standard deviations and confidence limits of relative percent absorbance of samples are given for the absorption. Conc. M++ t 5 (Cu) i 3 (Co) * s (Hg) i t t t t t Added (Ab) C(Ab) (Cu) (Co) (Hg) Tfié7fil) m““' '“ 1.00 2.12 7.07 0.282 9.46 24 1.2 3.00 0.707 2.12 0.353 3.15 9.2 1.5 5.00 2.12 7.77 0.565 9.46 27 2.4 10.0 0.707 0.00 0.353 3.15 0 1.5 15.0 2.82 5.65 0.707 12.58 20 3.11 20.0 3.53 3.53 0.353 15.75 15 1.5 Then the total percent error (100 x confidence limit/mean) for the metals is the addition of percent error in measuring standards and percent error in measuring samples. lCOnc. Cu Z error Conc. Cu 5 error error S d Standards Added Samples ( t ) to ta t ml ml 1.00 4.3 1.00 9.9 14 2.00 5.3 3.00 5.7 9.6 3.00 3.5 5.00 12 16 4.00 2Lfl_ 10.0 2.0 5.9 15.0 6.0 8.9 Average = 3.9 20.0 6.5 10 % error --- Average = 10 % error 108 TABLE A3 Continued gong. Cg % error ConcI Co % error error St Standards Added Sam les t: Total It) @61le11 1.00 10.5 1.00 26 30 2.00 1.9 3.00 3.7 8.0 3.00 2.4 5.00 13 17 4.00 2.4 10.0 0 4.3 15.0 14 18 Average = 4.3 20.0 10 14 % error Average = 10 % error gong. Hg % error Conc..Hg Z error error St Standards Added Samples gt) total t) 5722mm E2311 1.00 8.2 1.00 5.1 10 2.00 4.6 3.00 6.4 11 3.00 8.3 5.00 10 15 4,00 3.2 10.00 8.5 13 5.00 .21 15.0 16 21 20.0 9.2 ;4 Average = 5.0 % error Average = 14 % error The confidence limits and the resulting percent error derived for Cu, Co, and Hg are extremely high. These high results are due to having a small analyses number of two. The Student's t test and the standard deviation do not adequately describe the data at this low sample level. The analyses were not done for the purposes of statistical manipulation and therefore the sample size does not readily yeild applicable error analyses. The percent error of 10% for Cu++ analyses yeilds a range in concentra- tion of 1.00 t 0.1 ug/ml to 20 t 2 ug/ml. These ranges of ++ . . - - Cu pre01s1on are unreasonable Since Cline (In press) 109 TABLE A3 Continued has shown that the total accuracy in a similar Atomic Absorption study (that was statistically oriented) was t l to 2%. Since the error in accuracy in this experiment is greater than the error in precision in Cline's (In press) work, it is suggested that the above tests of error, as suggested by the reviewer. do not adequately describe the data. The total experimental accuracy is estimated to be less than t 0.1 ug/ml. the confidence limit shown by the circles around the data points in Figure 6. The standard deviations and variability of Co, Cu. and Hg standards is essentially the same in the following data. It will therefore not be shown again. 110 TABLE A4. Data for Figure 7 showing the equilibrium between the Cu-organic chelate with pH adjustments. Four ug Cu/ml is in each solution with a constant 46 ug C/ml. Shown are the results of filtering the solution after pH adjustment from 7.36 to between 2.63 to 10.92 and again filtering the solutions after readjusting the pH's back to 7.36 t 0.04. Cu Standards Conc ml Absorbance Standard deviation 1.00 2.00 3.00 4.00 71. 140, 201. 247. 66. 137. 195. 259. 65 136 203 250, 263 was» Vtmu U'tt-‘ON 8 6 Average % S 2.7 The correlation (r) within the samples and the resulting sample variability are given below. Data for First Filtration - Fi - Curv pg Absorban e Conc Cu m1 5 Cu Adsorbed (Ab) 2.63 240. 236 3.62 9.50 5.16 157. 161 2.40 40.0 7.36 77. 77 1.15 71.3 9.00 37. 39 0.57 85.8 10.92 l4. 14 0.20 95.0 Cu Standards C n ml Y (Ab) + t (Ab) % error 1.00 67 50"" 800 2.00 137 3.5 2.5 3.00 200 7.0 3.5 4.00 254 12.6 4.9 111 TABLE A4 Continued Cu Samples pfl_ Y (Ab) t t (Ab) t S (Ab) % error (Ab) rr Tota Ab 2.63 238 4.7 2.82 2.0 6.7 5.16 159 4.7 2.82 2.9 7.6 7.36 77 0.0 0.00 0.0 4.7 9.00 38 2.4 1.41 6.3 11 10.92 14 0.0 0.00 0.0 4.2 Average % error = Data for Filtration After Adjusting;pH to 2.4 Fi re - curve b pH original pH back Absorbance Cu Cones % Cu Adsorbgd 1“» ~ 2.63 7.32 62. 59 0.97 75.8 5.16 7.42 74, 76 1.10 72.5 7.36 7.36 77. 77 1.13 71.8 9.00 7.36 50. 55 1.18 70.5 10.72 7.34 38. 39 0.99 75.2 Standard Deviation and Error of the pH Adjusted Solutions - Curve b 25 original it t S (Ab) t t (Ab) % error % total w 2.63 60 2.12 3.6 6.0 11 5.16 75 1.41 2.4 3.2 7.9 7.36 77 0.00 0.0 0.0 4.7 9.00 52 3.58 5.9 11 16 10.72 38 0.707 1.2 3.1 2.8 Average % total error 9.5 The error in this experiment is 4.7% for the standards and a total of 6.9 and 9.5% for the total run. Since the original concentration of Cu was 4.00 ug Cu/ml. the error 112 TABLE A4 Continued is 4.0 t 0.27 ug/ml and 4.0 t 0.38 ug/ml in the two sample sets. The accuracy of the experiment is dependent on filtration error and pipetting error. It is estimated that an accuracy of 0.05 ng Cu/ml for the experiments is much less than “precision". 113 TABLE A5. Data for Figure 9 - (a) and (b) giving the change in fluorescent (a) and absorbant (b) intensity measured in cm at various wavelengths (L) in mu as Cu concentration changes. Eluorescgnce (a) Conc. Cu (Hg/m1) Intensity in cm .(= 255 A: 220 A: O 12.7 8.2 5.0 1 11.8 7.8 4.8 2 10.7 7.1 4.3 3 9.8 6.7 3.9 4 8.6 5.8 3.6 5 9.2 5.8 3.5 10 7.6 4.7 2.8 15 6.0 3.9 2.1 20 6.6 3.8 2.0 Absorbance (b) _an. Cu (mg/m1) Intensity in cm X: 2 0 _L:_33g O 7.5 4.5 1 7.8 4.7 2 7.9 4.7 3 8.2 5.0 4 8.8 5.2 5 8.8 5.2 10 10.8 6.1 15 12.5 6.7 20 13.5 7.7 114 TABLE A5 Continued The above results were repeatable in duplicate runs such that no difference could be measured. The measurement itself is good to the nearest 0.5 mm. 115 TABLE A6. Data for Figure 10 showing the intensity of fluorescence of diluted pore water solutions is measured in cm at three wavelengths (A) in mu. 2 Pore Water Organic Carbon FluorescentI Intensity (cm) in_Sq1utign_ (gglml) A: 255 X: 2201-.- 330 10 9.2 3.9 2.5 1.4 20 18.0 6.9 4.1 2.6 40 35.0 11.2 8.1 5.0 50 46.0 14.4 10.2 6.4 The variation of the data points from the theoretical linear line fit. Line Mean Variation Stan%a§d Deviation of Line S ofgling K= 255 4.6% 2.2% A: 270 309% 5'5% x: 330 .1212 J1 Grand Variation 3.3% Grand Mean =-2.9% Mean The variation and standard deviation of the data points about the various lines are shown. Pipetting error or variation in pipetting is about 1 drop/25 ml or about 0.5%. The grand variation mean of 3.3% is attributed to machine variation and 0.5% attributed to sample handling. The same variation, 3.3%, can be expected in all of the fluorescence, absorbance and light scatter spectra. 116 TABLE A7. Data for Figure 11 (a) and (b) giving effect of Co++ on organic fluorescence (a) and absorbance (b) intensity in cm at two wavelengths (A) in mu. Conc. Co Fluorescent Intensity Absorbant Intensity m1 (cm) (cm) A: 220 A: 330 L: 220 A: 330 0 10.2 6.4 8.5 4.6 1 9.8 5.8 --- --- 2 9.2 5.6 --- --- 3 8.8 5.3 --- --- 4 8.8 5.3 --- --- 5 8.8 5.1 8.3 4.4 10 707 I406 802 “'03 15 7.1 4.2 --- --- 20 7.2 4.1 8.2 4.4 The results are reproducible within the limits of the intensity measurement or to the nearest tenth of a cm. The variation of 3.3% can be expected for these results. TABLE A8. Conc. Cu (uggml) 10 15 20 The light scatter spectra the measurement range of t 0.5 cm. 117 Data for Figure 12 comparing Co++ and Cu ++ affects on the light scatter of a pore water solution. Light Scatter Intensity (cm) 6.0 7.3 9.1 10.4 10.5 Conc. Co (Eggmlfi O 2 10 20 3.3% is applicable to the above results. Li ht Scatter 5.7 5.8 5.9 6.0 were reproducible within The variation of 118 +l3 showing the time dependence All data from the Conc mg in (EEZm12801p- 0.42 0.67 1.20 1.30 1.45 12.5 33.5 tion 0.0525 0.0837 0.150 0.162 0.181 1.56 4.19 12400 150 5 Conc mg in 11311.... m1 Solu- 0.18 0.40 0.53 0.51 0.50 1.00 7.50 tion 0.022 0.050 0.066 0.063 0.062 0.12 0.64 TABLE A9. Data {or Figur? of Cu+ and Co$+ reaction with organic rich sediments from Burke Lake. experiment are given though only the 100 mg M data are presented in Figure 13. Time = 2 hours for Cu Standards Samples Cu Absorbance ConcI Cu Dilutipn Ap, (uglml) 1.:— Added 0.500 39. 41 0.500 --- 35. 37 1.00 74, 82 2.00 --- 54, 58 2.00 166, 168 5.00 --- 99, 103 3.00 222, 234 10.0 --- 105. 111 25.0 --- 118, 119 50.0 1/50 19. 20 75.0 1/50 54. 56 100.0 1/50 193. 199 Time = 2 days for Cu Standards Samples %3 Ap, Conc. Cu Dilution gp m Added 0.500 1+5, 52 00500 --- 9’ 8 1.00 11.5, 104 2.00 """"‘ (+0, 35 2.00 188’ 180 5000 --- 51, 51 3.00 280, 291 10.00 -—- 48, 40 25.0 --- 44, 43 50.0 --- 94. 102 75.0 1/25 25. 27 100.0 1/50 73. 76 39.35 4. 88 119 TABLE A9 Continued Time = 3 days for Cu Standards Samplgg do Ab. Concr Cu 2.123192. Ab ml Added 0.500 32. 33 0.500 --- 9. 10 1.00 68. 72 2.00 --- 28. 30 2.00 135. 130 5.00 --- 50. 47 3.00 176. 185 10.0 --- 58. 55 25.0 --- 60, 64 50.0 --- 61. 59 75.0 ---1 9. 146 100.0 1/25 48. 48 Time = 4 days for Cu Standards Samples Conc. AD QQQ£1_QE,2112212fl Ah Cu Ad ed m1 1:2:— 0.500 39. 38 0.500 --- 17. 17 1.00 78. 77 2.00 --- 31, 32 2.00 150, 149 5.00 2-; 69, 67 3.00 210, 219 10.0 --- 76, 79 25.0 '“"' 68' 72 50.0 --- 7o! 71 7500 "" 1190, 144 100.0 1/50 41. 43 Time = 5 days for Cu Standards Samples Cppc. 5p, Conc, Cu Dilution A_ Cu dd d (figAml) 0.500 39. 40 0.500 --- 14, 14 1.00 714’, 78 2.00 ""'"' 22, 21 2.00 150, 156 5.00 --- 40, 42 3000 212, 218 10.0 --- L70, “’3 2500 "" 429 “’9 50.0 "- “7, “'9 75.0 --- 105. 109 100 --- 41 x 10 onc 1:27:30 0.20 0.45 0.80 0.92 0.95 0.81 2.05 20.0 Concil 0.20 0.30 0.82 0.96 0.87 0.87 1.94 15 Concél 0.10 0.21 0.50 0.50 0.55 0.56 1.42 10.0 mg in Solution 0.024 0.054 0.096 0.11 0.11 0.097 0.25 2.40 mg in Solution 0.024 0.036 0.098 0.12 0.11 0.11 0.23 1.80 mg in Solution 0.012 0.025 0.059 0.059 0.066 0.066 0.167 1.18 120 TABLE A9 Continued Time = 6 days for Cu Standards Sapplgg Conc. Ap Conc._Cu Dilution Ap Conc mg i Cu Added ZEEZEll Solution (uglml) 0.500 38, 36 0.500 "'"' 10 0010 00012 1.00 75, 55 2.00 --- 28, 27 0.38 0.024 2.00 149, 154 5.00 --- 31, 32 0.42 0.049 3.00 285, 279 10.0 -—- 36, 39 0.52 0.060 25.0 --- 40, 45 0.57 0.066 50.0 “-'" 71! 79 1.00 0.116 75.0 -—- 131. 140 1.88 0.218 100 --- 64 x 10 8.80 1.02 Time = 8 days for Cu Standard§_ Samples Conc. Ap_ Conc, Cu 5p Conc :mg i C Added IIEZEll Solution (figgmlz 0.50 23, 25 0.5000 4 0.10 0.011 1.00 55, 54 2.00 9 0.20 0.023 2.00 109. 105 5.00 15, 14 0.30 0.034 3.00 148, 161 10.0 15. 15 0.30 0.034 4.00 197. 205 25.0 15, 16 0.30 0.034 50.0 32 0.62 0.071 75.0 61 1.13 0.127 100 41 x 10 7.80 0.889 Time = 9 days for Cu Standards Samples ggngp 5p, Conc,Cu 5p Conc Cu mg in Cu Added I:EZi:17 Solption (uEZml) 0.50 34. 37 0.50 0 0.00 0.000 1.00 68. 73 2.00 6 0.10 0.011 2.00 144. 138 5.00 8 0.12 0.014 .3.00 200, 202 10.0 12. 11 0.18 0.020 14.00 259. 266 25.0 12, 12 0.18 0.020 50.0 29. 31 0.42 0.047 75.0 67, 62 0.80 0.090 100 190 x 2 5.20 0.582 121 TABLE A9 Continued Time = 11 days for Cu Standards Samples 92 Ag Conc.Cu gp Conc Cu mg Cu in Conc Added Ijngjlf Solution (:mel) 0.50 34, 33 0.50 0 0.00 0.000 2.00 138, 137 5.00 10, 9 0.17 0.018 3.00 200, 211 10.0 10, 10 0.17 0.018 4.00 259. 270 25.0 11. 12 0.18 0.019 50.0 29. 31 0.45 0.048 75.0 65, 59 0.82 0.101 100 167 x 2 5.13 0.513 Time = 2 hours for Co Standargs Samples 99_ Ap Conc,Co Dilution gp ConcI mg C9 in Conc Added Co Solution E2511 ml 0.50 24, 26 0.50 --- 7 0.20 0.025 1.00 59. 49 2.00 --- 14, 13 0.27 0.034 3.00 164,.157 5.00 --- 42, 44 0.84 0.105 5.00 245, 243 10.00 --- 95. 105 1.95 0.243 25.00 1/2 179. 189 7.53 0.941 50.00 1/50 39, 42 41.0 5.13 75.00 1/50 87, 78 80.5 10.1 100 1/50 159. 140 144 18.1 Time = 2.5 days for Co Standards Samples 99 Ap Conc, Dilution 5p Conc. mg Co in Conc Co Added Co Solution 1222811 1:21 1221011 0050 31, 28 0050 "'"' O 0.00 O 1.00 52! 62 2.00 --- 19, 20 0.35 0:82? 3.00 149. 165 5.00 -~— 31. 37 0.60 0 091 5.00 250, 259 10.0 --- 48, an 0.80 0:116 25-0 '"‘ 140. 150 2.87 0.416 50-0 1/70 94. 103 15, 3,29 75-0 5/50 199. 217 41.0 5.95 100 1/50 109. 108 102 14.8 122 TABLE A9 Continued Time = 6 days for Co Standards Samples Q9, Ap_ Conc. Co Dilution gp Conc Added @211 0.50 31. 33 0.500 --- 0 1.00 66, 63 2000 --- 18, 19 2.00 128, 138 5.00 --- 19, 19 3.00 192, 200 10.0 ~-- 21, 27 4.00 258, 258 25.0 --- 141, 151 75.0 1/10 224, 234 100 1/50 112, 113 Time = 11 days fgpro Standggds Samples ConcI Ap_ Conc, Dilutigg Ap_ Co 99 (uglml) Added 0.5 “’7' “'9 00500 "' O 1.0 85, 90 2.00 --- 0 2.0 179. 175 5.00 --- 8 3.0 258, 272 10.0 ~-- 25, 26 4.0 312, 328 25.0 ~-- 188, 192 50.0 1/20 49, 52 75.0 1/20 126. 139 100 1/20 196. 213 Gone. ml 0.00 0.32 0.33 0.40 2.30 1.50 3.58 90.0 Concl Co (uglml) 0.00 0.00 0.10 0.27 2.15 0.56 1.50 2.33 mg Co in Solution 0.000 0.039 0.039 0.049 0.281 1.83 4.37 10.98 mg 1 Solution 0.000 0.000 0.012 0.032 0.258 1.30 3.48 5.40 123 TABLE A9 Continued The weights of sediment or adsorber in each reaction flask are given below. Co thfof App. Wt. Beaker Wt. Sediment m 0.5 106.522 ' 103.787 2.735 2 100.259 97.128 3.131 5 107.707 104.686 3.021 10 102.484 99.150 3.334 25 104.436 100.836 3.600 50 102.758 99.848 2.910 75 100.298 97.385 2.913 100 110.410 107.450 2.960 Cu Wt._of AppL Wt. Beaker Wt. Sediment (mg) 0.5 106.913 104.519 2.394 2 103.238 100.331 2.907 5 103.300 100.402 2.898 10 100.727 97.846 2.881 25 99.641 96.530 3.111 50 99.593 96.400 3.193 75 106.907 103.805 3.102 100 107.715 104.779 2.936 Confidence Limits, Standard Deviation, and Percent Error for Cu after 2 Hours and 2 Days Standards 2 Hours C02? Op '1'? (Ab) iS‘ (Ab) t (Ab) 2. Error 0.50 40 1.4 6.3 15 1.00 78 4.2 18 23 2.00 164 1.4 6.3 3.7 3.00 228 7.1 31 14 Average % Error 2 14% 124 TABLE A9 Continued Standards 2 days Conc: Cu Y (Ab) :8 (Ab) t (Ab) % error 0.50 43.5 2.12 3.1 7.1 1.00 111 3.56 16 14 2.00 184 4.24 19 10 3.00 285 4.94 22 7.7 Average % Error = 9.7% Samples 2 hoprg Conc. Cu Y (Ab) is (Ab) t (Ab) % error 2 total Added error m1 0.500 36 1.41 6.3 17 27 2.00 56 2.82 12 21 31 5.00 101 2.82 12 12 22 10.0 108 4.24 19 17 27 25.0 118 0.707 3.2 2.7 12 50.0 19 0.707 3.2 16 26 75.0 55 1.41 6.3 11 21 100 196 4.24 19 10 20 Average % total error = 23% 2.0215 0.500 8.5 0.707 3.1 36 46 2.00 37 2.82 12 32 42 5.00 51 0.00 0.0 0.0 9.7 10.0 45 4.24 19 42 52 25.0 43 0.707 3.1 7.2 17 50.0 99 3.66 16 16 26 75.0 22 1.41 6.3 28 38 100 74 2.12 9.5 12 12 Average % total error = 30% As in the earlier experiments the error calculated is extremely high due to the small sample size. The same results are obtained for the other repetitions of this experiment and are therefore not shown. 125 TABLE A9 Continued These errors show the confidence level one can place in the determination by atomic absorption but, they do not account for the inaccuracy inherent in the experiment. The temperatures varied by t 1.500, and the sediment mixtures, even though homogenized. were mixtures of many inhomogeneous compounds, colloids, etc. These sediments and their chemistries change by season and location. Thus, the machine error shown above is minimal compared to the variation in a natural system. The numbers and data given will show only trends. trends with confidence levels much exceeding the calculated machine error. The inherent error in these data will be given in Tables A10 and A13. ....._‘ 126 TABLE A10. Data and figures used to obtain the rate constants K (day'l) in Table 2 of the Co++ and Cu++ reactions with Burke Lake sediment. The raw data are given in Table A9. The correla- tion coefficient (r) shows the fit of data to the line and the slope m is equal to the rate constant. K. The value «2 or %<72 is defined as percent variability and represents the goodness; of the fit of the data to the theoretical linear cu e. The percent variability (%a ) is given by (100(l-r2)) (Wang Laboratories. 1970). Data for Cu++ plotted in Figure Al. where 2.3 log N0 = Kt is the first order plot. No’x Cone. of Cu++ added = 0.5 mg = N X = conc. of Cu adsorbed o Time N :K 2.3 LOG (N N 2&1 0.2 0.053 2.25 2.0 0.048 2.35 3.0 0.024 3.04 5.0 0.023 3.11 6.0 0.012 3.75 8.0 0.012 3.75 9.0 0.011 4.78 Correlation coefficient (r) = 0.97 Slope (m) = 0.262 Intercept (a) = 2.05 ++ Conc. Cu added 2 5.0 mg = N X = Cone. of Cu adsorbed 0 Time N :5. 2.3 LOG (En N -x Way) ...0 14.1 0.2 0.150 3,50 2.0 0.130 3.66 3.0 0.096 3.96 4'0 0.098 3093 5.0 0.059 u,uu 6.0 0.049 4.62 8.0 0.031 5.08 9.0 0.014 5.89 r = 0.96 m = 0.259 a = 3.18 02= 7.8% 127 TABLE A10-Continued Conc. Cu++ added = 25 mg = N o X = Conc. of Cu adsorbed Time 30:1, 2 Lo N [Ho:§l. da 8 0.2 0.18 4.92 2.0 0.16 5.06 3.0 0.11 5.38 4.0 0.10 5.47 5.0 0.065 5.93 6.0 0.066 5.91 8.0 0.032 6.65 9.0 0.020 7.13 r = 0097 m = 0.252 a = 4.64 r2= 6.0% Conc. Cu added = 100 mg = NO X = Cone. of adsorbed Cu c = N -X This 13 a second order plot of l/c vs t Time :E. 112 da 0.2 15.5 0.064 2.0 4.88 0.20 3.0 2.40 0.42 4.0 1.80 0.56 5.0 1.18 0.85 6.0 1.02 0.98 8.0 0.77 1.30 9.0 0.52 1.92 11.0 0.51 1.96 r = 0.98 m = 0.195 a3: -0012? V = 4.0% 128 TABLE A10--Continued Rate data for 92, The first order rate equation LOG (NO/No-X) where N = original Co concentration and X = adsorbed Cu is used. N0 = 100 mg da 0.2 18.1 1.71 2.5 14.8 1.91 6.0 11.0 2.21 11.0 5.40 2.92 r = 0.99 m = 0.112 a = 1.64 ¢2= 109% N0 = 50 mg 1% 10:1 2-3 L0G (4.1101). 0.2 5.13 2.27 2-5 3.29 2.71 6.0 1.83 3.31 11.0 1.30 3.66 r = 0097 m = 0.128 a2: 2.36 0“ s: 5.9% No 25 mg Time N03); 2.: LOG (“0110229. a 0.2 0.941 3.27 2.5 0.416 4.09 6.0 0.281 4.48 11.0 0.258 4.58 r = 0.89 In = 0.104 a: = 3.50 cr2= 20.8% 129 TABLE A10 Continued N0 = 5 mg Time N :X 2. LOG N N -X T—da ) .0 -2-—1.01.04 0.2 0.105 3.86 2.5 0.091 4.23 6.0 0.039 4.85 11.0 0.012 6.03 r = 0.99 m = 0.201 a2= 3.75 0' = 109% The fit of data to each of the lines is given 2 for by percent variability 02. The average percent a the Co runs are 7.62% and for the Cu runs 4.96%. This shows that the Cu data fit the theoretical curve better than the Co data. The results of the listed data are plotted for Cu++ (Figure Al) and for Co++ (Figure A2) reaction rates with Burke Lake Sediment. ADSORBED CH:(2L3Log(No/No—X»(lstordefl 130 Cu(mg) ‘7 25 ‘6 5 ‘5 f: 0.5 Q «4 3 .U C CV ’3 #38 U '3 U \ 2- v -2 “i /// 100 % / ,05/