THE ORIGIN, DISTRIBUTION, AND DYNAMIC CHARACTER OF SOME CHLORITIZED VERMICULITE SOIL CLAYS Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY DAVID A. LIETZKE 1972 .rji‘t! This is to certify that the thesis entitled THE ORIGIN, DISTRIBUTION, AND DYNAMIC CHARACTER OF SOME CHLORITIZED VERMICULITE SOIL CLAYS presented by ,1 ' DAVID A. LIETZKE has been accepted towards fulfillment of the requirements for Ph.D. deg-66in Soil Science , v Major pro/fem Date_Ma¥ 18; 1972 0-7639 LIBRARY _ Michigan Saw University ? BINDING av ? H HDAG & SUNS' : Iuannx amnm me. I I qua :‘v FINGERS -- “win-h MICHIGAN. ABSTRACT THE ORIGIN, DISTRIBUTION, AND DYNAMIC CHARACTER OF SOME CHLORITIZED VERMICULITE SOIL CLAYS BY David A. Lietzke Large quantities of potassium fertilizer were required to obtain Optimum yields and quality of tomatoes grown on a low terrace of the St. Joseph river at the Sodus Experimental Farm in Berrien County, Michigan, in 1968. Dr. Mortland found the soil clay contained a con- siderable quantity of vermiculite. The major concerns of this study were to determine the origin, distribution and dynamic character of the vermiculitic soil clays along the St. Joseph river. Seven landscape units were differentiated on the basis of the surface age of the deposits. The oldest landscape unit, level U1, consisted of Valparaiso moraine and outwash deposits. Level UZ consisted of river terrace and lakeplain deposits of Lake Glenwood stage. Level Tl David A. Lietzke consisted of St. Joseph River deposits of Lake Nipissing stage. Level T2 consisted of well and moderately well drained alluvium of Late Lake Nipissing and early Lake Algoma. Level T3 consisted of somewhat poorly drained alluvium of late Lake Algoma with a veneer of recent flood deposits. Level T 4 consisted of poorly drained alluvium that floods every year. Level "B" was designated as un- differentiated poorly drained alluvium. Sandy, well drained soils from Levels U1, U2, and T1 contained considerable quantities of chloritized ver— miculite in surface layers, while surface soil layers from Levels T2, T3, T4 and B contained vermiculite in varying quantities. Level T2 generally contained the highest con- centration of vermiculite. Apparently the chloritized vermiculite soil clays were eroded from the well drained acid upland soils and transported to the neutral to mildly alkaline alluvial floodplain environment of the river. The change in pH plus the additional quantity of organic matter and the poorer natural drainage triggers a dechlori- tization process involving an aluminate reaction sequence, or another type of complexing reaction. Interlayer material is removed by leaching rainwater and vermiculite is formed. As the pH drops due to continued leaching, rechloritization occurs. An interlayer removal study showed the dynamic character of the chloritized vermiculite in the upland levels U1, U2 and terrace level T1, the probable source David A. Lietzke areas of the vermiculite found on Levels T2, T3, T4 and B. One sample of a stable dioctahedral chloritized vermicu— lite soil clay from Level U2 was cleaned up in a conven- tional manner, while another sample from the same site was dispersed in distilled water with a sonification treatment. This particular sonified sample contained most of the organic matter and oxide coatings of the surface soil layer plus the natural cations on the clay surfaces. These two samples exhibited different behavior when sub- jected to a dilute NaOH interlayer removal treatment sequence. A standard pure prochlorite was used as a check. CEC increased markedly within a few days, but as pH dropped when OH ions were used up in the aluminate or similar complexing reaction, the CEC began to decrease and rechloritization was occurring in the sonified sample. The conventionally cleaned soil clay sample did not show a decrease in pH when treated with NaOH, but the CEC in- creased markedly. The prochlorite sample when treated with NaOH showed an increase in CEC but no increase in vermiculite content as the chloritized vermiculite samples did. Deuteration experiments using a natural dioctahe- dral chloritized vermiculite soil clay, a natural dioctahedral dechloritized vermiculite, and a prochlorite as check readily showed the easy accessibility of inter- layer OH groups in the chloritized vermiculite. David A. Lietzke Interlayer deuteration started at 68 C° in the chloritized vermiculite, and 150 C° in the dechloritized vermiculite. At 200 C° in the prochlorite, dechloritized vermiculite and chloritized vermiculite lattice OH deuteration evidently had occurred. This study shows the need for multi-disciplinary studies in the earth sciences. Both field and laboratory studies are necessary. All too often in studies of soil clay mineralogy the manipulated soil clay in the laboratory bears little resemblance to the natural system. Soil clays that are dispersed in distilled water with a sonifi- cation procedure may be one way to keep the system in a more natural state. THE ORIGIN, DISTRIBUTION, AND DYNAMIC CHARACTER OF SOME CHLORITIZED VERMICULITE SOIL CLAYS BY < I‘X ' 0 AI) ‘ o DaVld Nu Lietzke A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of CrOp and Soil Sciences 1972 ACKNOWLEDGMENTS The author wishes to thank Dr. Maynard M. Miller who, in 1966 on the Juneau Icefield, first started the author to think seriously about attaining a Ph.D. degree. Secondly, he expresses his appreciation to Dr. Max M. Mortland for suggesting the research problem and his guidance. Thirdly, he expresses his appreciation to Dr. Eugene Whiteside and Dr. Ray Cook for their help and encouragement. He also expresses his appreciation to the Crop and Soil Sciences Department and to the Agricultural Experiment Station of Michigan State University for financial and technical assistance on aspects of this research problem. ii TABLE OF CONTENTS INTRODUCT ION I O O O O O O O O O B. C. PART I. GEOMORPHIC SETTING IN BERRIEN COUNTY Physical Location . . . . . LITERATURE REVIEW . . . . . l. Glacial Geology of Berrien County . . . a. Kalamazoo Interlobate Moraine Complex . . . . . . b. Valparaiso Moraine Complex . . . . . c. Lake Border Moraine The Early Great Lakes in Basin . . . . . . . . . a. Lake Chicago . . . . b. Lake Algonquin . . . c. Lake Nipissing . . . d. Lake Algoma . . . . e. Lake Michigan . . . the Michigan Sequence of Events that Formed the St. Joseph River System, and Land Forms . . . . . . . Its Surrounding a. The Landscape along the St. Joseph River near Sodus Farm b. St. Joseph River Landscape from Berrien Springs to the Indiana State Line . . . . . MATERIALS AND METHODS . . . RESULTS AND DISCUSSION . . . 1. Relationship of Clay Mineralogy to Landscapes . . . . . . . iii Page . l O (bib-N 0 00000on 0‘! . 18 Page a. Clay Mineralogy of Watersheds Other Than the St. Joseph River in Berrien County . . . . . . . . . . . . . 18 b. Clay Mineralogy of the St. Joseph River Watershed in Berrien County . . . 26 c. Summary of Clay Mineralogy Data along the St. Joseph River . . . . . . . 40 Proposed Model . . . . . . . . . . . . . . . 44 Evidence to Support Model . . . . . . . . . 45 a. Evidence from Two Transects across the St. Joseph River Valley . . . . . 46 b. Clay Mineralogy Within Soil Profiles . . 54 CONCLUSIONS OF PART I . . . . . . . . . . . . . 59 PART II. CLAY MINERALOGY LITEMTURE REVIEW 0 O O O O O O O O O O O O O O 62 1. 2. 3. 4. Interlayered Soil Clays . . . . . . . . . . 62 Laboratory Studies . . . . . . . . . . . . . 68 Effect of Clay Cleanup Procedures in Changing Soil Clay Mineralogy . . . . . . . 72 Deuteration Studies . . . . . . . . . . . . 76 METHODS OF ANALYSIS 0 O O O O O O O O O O O O O 7 9 Cleanup and Dispersion . . . . . . . . . . . 79 X-ray Diffraction . . . . . . . . . . . . . 80 CEC Determination . . . . . . . . . . . . . 80 Total Potassium . . . . . . . . . . . . . . 82 Interlayer Removal Experiment, Materials and MethOd S O O O O O O O O O O O I I O O O 8 2 Deuteration Experiment, Materials and MethOdS O I O O O O I O O O O O O O I O O O 8 4 RESULTS AND DISCUSSION . . . . . . . . . . . . . 87 1. Clay Mineralogy . . . . . . . . . . . . . . 87 iv 3. SUMMARY Interlayer Removal Experiments a. Initial Trials . . . . . . b. Experiment #1 . . . . . . c. Experiment #2 . . . . . . d. Experiment #3 . . . . . . e. Experiment #4 . . . . . . f. Experiment #5 . . . . . . 9. Discussion . . . . . . . . Deuteration Experiments . . . CONCLUSIONS 0 O O O O O O O O O O O 0 RESEARCH NEEDS O O O O O O O O O O O O BIBLIOGMPHY O O O O O O O O O O O O 0 APPENDIX Page 98 98 108 109 114 116 120 122 125 132 136 138 140 146 LIST OF TABLES Table l. 10. 11. 12. 13. Summary of late Wisconsonian history of the Michigan lobe and history of the Great Lakes in the Michigan Basin . . . . . . . . . . . . Vermiculite content of the clay fraction in the alluvium of the major watersheds in Berrien County other than the St. Joseph River watershed . . . . . . . . . . . . . . . St. Joseph River transect . . . . . . . . . . Data from landscape Transect I (Figure 10) . Data from landscape Transect II (Figure 11) . Soil profile data from three landscape levels I O O O O O O O O O O O I O O O O O 0 Data from Sample 2-lT2 and its fractionated components (Figure l3) . . . . . . . . . . . Summary of data for different chemical treat- ments on Sample 24-lUl (Figure 15) . . . . . Summary of data for increasing number of iron removal treatments on Sample 24-lUl (Figure 15) O O O O O O O O O C O O O O O O 0 CEO of two soil clays effected by the increasing number of iron removal treat- ments (SCDB) O O O O I O O O I O O I C O O 0 Effect of standard pretreatment procedures vs. ultrasonic treatment on the CEO of soil clays and vermiculite content . . . . . Experiment #1. Data from Figure 16 . . . . . Experiment #2. Data from Figure 17 . . . . . vi Page 21 30 48 51 55 94 100 103 104 106 109 111 Table Page 14. 15. A1. CEC data from Experiment #4 . . . . . . . . . . 120 Experiment #5. Acid dissolution of an intergrade soil clay (Sample 35-lU2) with OOIN HCl I O O O O O O O O O C I I O O O O O O 121 Probable classification of soils that were sampled in this study . . . . . . . . . . . . . 153 vii LIST OF FIGURES Figure 1. Major river systems and glacial geology map of Berrien County . . . . . . . . . . . . . 2. Elevations along the St. Joseph River . . . 3. Air photo mosiac of a portion of the St. Joseph River I I I I I I I I I I I I I I I 4. Site location map of Berrien County . . . . 5. X-ray, CEC, and total K data for Samples 13-lB, ll-lB, and ll-laB . . . . . . . . . 6. X-ray, CEC, and total K data for Samples lOI ll. 12. 13. 14. lz-lB' 9"pr and lS-lB o o o o o o o o o 0 X-ray, CBC, and total K data for Samples 19-lB and 20-1B . . . . . . . . . . . . . . Cross-section of terrace levels T1, T2 and T4 at Sodus Farm at the tomato plot site . Vermiculite and total K content of some samples from levels T1, T2, T3 and T4 along the St. Joseph River in Berrien County . . Transect I: Landscape relationships at SOduS Farm I I I I I I I I I I I I I I I I Transect II: Landscape relationships 13 miles upstream from Sodus Farm . . . . . . X-ray, CEC, and total K data for Samples 35-1U2 and HVTZ I I I I I I I I I I I I I I X-ray, CBC and total K data for Sample 2-lT2 and its fractionated components . . . Infrared analysis of Sample 2-1T2 and its fractionated components . . . . . . . . . . viii Page 10 15 19 20 25 27 38 41 47 49 88 92 96 Figure 15. l6. 17. 18. 19. 20. 21. 22. A1. A2. A3. A4. A5. A6. A7. Effect of different treatments on the stability of the 14A° component of sample 24-1111 0 o o o o o o o o o o o o o 0 Experiment #1. Sample 35-laU2; pH trace, change in CEC, and extractable A1 with time I I I I I I I I I I I I I I I I I I I Experiment #2. Sample 35-laU2; pH trace, change in CEC, and extractable A1 with time I I I I I I I I I I I I I I I I I I I Experiment #3. Prochlorite: pH trace, and change in CEC with time . . . . . . . . X-ray and CEC data for Sample 35-1U2 after the completion of Experiment #4, and prochlorite after the completion of Experiment #3 . . . . . . . . . . . . . . . Experiment #4. Sample 35-1U2: pH trace and CEC change with time . . . . . . . . . Deuteration effects on Sample HVT2 and prochlorite with increasing temperature . . Deuteration effect on Sample 35-lU2 with increasing temperature . . . . . . . . . . X-ray, CBC, and total K data for Samples 4-lT3, 5-1T4, l4-lT4, l6-lT2 and 18-lT2 . . X-ray, CEC, and total K data for Profile 24Ul: Samples 24-laU1, 24—2 and 24-3 . . . X-ray, CBC, and total K data for Samples 23-1Ul, 22-1U2 and 21-1U2 o o o o o o o o 0 X-ray, CEC, and total K data for Profile 3T2: Samples 3-lT2, 3-2 and 3-3; and Profile 1702: Samples l7-lU2, l7-2 and 17-3 . . . X-ray and CEC data from Samples 31-lB and 32-1U2 I I I I I I I I I I I I I I I I I I X-ray, CEC, and total K data for Samples 1-1T]. and 33-1T1 o o o o o o o o o o o o 0 X-ray and CEC data for Samples 28-lT2, 29-1.1.2 and 30-1T2 o o o o o o o a o o o o 0 ix Page 99 110 112 115 117 118 127 128 146 147 148 149 150 151 152 NOMENCLATURE AND ABBREVIATIONS -8 centimeter. A° - Angstrom unit. 10 CEC - Cation Exchange Capacity expressed as milequivalent weight per 100 grams of oven-dry material, either clay or whole soil. Chloritization - A natural process where gibbsite, com- posed of hydroxy aluminum becomes fixed in the interlayer part of vermiculite or smectites. Chloritized Vermiculite - See vermiculite. The interlayer part has been partially filled with gibbsite. The mineral has properties midway between vermiculite and chlorite. Dechloritization - A natural process whereby interlayer gibbsite is removed. Deuteration - Process whereby the H of OH groups in the interlayer or lattice is replaced by D ion from D20. Feldspar - A framework silicate. Feldspars were noted in the clay fraction of several soil clay samples in this study. It has diffraction peaks at 4.10- 4.llA°, 3.18, and 3.lOA°. It is probably a low K content oligoclase. Hydroxy Interlayered Vermiculite - See chloritized vermicu- lite. Illite - See mica. A clay thought to be more hydrated and to contain less potassium than muscovite mica and typically of clay particle size, <2u equivalent diameters. Interlayered Vermiculite - See chloritized vermiculite. Interlobate Moraine Complex - Debris pushed up by two separate advancing ice sheets that mingle together so that the debris is highly mixed. Land form ex- pressed by strong relief, and kames are a common feature. Interstratified Clay - A clay composed of layers of mica, vermiculite and smectite in a random sequence. KAD - A clay specimen that is potassium saturated and air dry. K110 - A K saturated clay specimen heated to 110°C for one hour. K300 - A K saturated clay specimen heated to 300°C for two hours. K550 - A K saturated clay specimen heated to 550°C for two hours. Kaolinite - A class of 1:1 lattice type of clay that has a 7.07 A° spacing in x-ray diffraction patterns. MgAD-gly - A clay specimen that is magnesium saturated, glycerated and air dried. Mica - A class of 2:1 lattice non-expanding clay. Soil micas have a lOA° spacing in x-ray diffraction patterns. Moraine - Debris pushed up by an advancing glacier. Land form expressed by rolling hills usually one to five miles wide. Moraine Complex - Debris pushed up by initial glacial advance and then further pulsations modify the land form. A typical complex is usually five to ten or more miles across. Outwash Plain - Wide expanse of stratified, usually coarse-textured, debris deposited by glacier melt water. Land form expressed by nearly level or gently sloping plains. Pitted outwash plains are hilly but have concordant hills. Quartz - A framework silicate non-clay but in the clay particle size class. It has a 4.25 A° spacing in x-ray diffraction patterns. Smectite - Class of 2:1 lattice type of expanding clays. The interlayer part accepts two layers of glycerol molecules or more when magnesium saturated. These clays, when treated thus, show a typical spacing of 17 to 18 Angstroms in x-ray diffraction patterns. xi Sonification - Dispersion technique using sound waves at 20,000 cycles per second or higher. Till Plain - Debris deposited, usually by a downwasting ice sheet. Land form expressed by nearly level to gently sloping relief. Sometimes referred to as "ground moraine.‘ Torrent Channel - See valley train. Ultrasonic Treatment — See sonification. Valley Train - Narrow band of outwash deposits. Vermiculite - A class of 2:1 lattice type of expanding clays. The interlayer part accepts one layer of glycerol molecules when magnesium saturated. The typical spacing is 14A° in x-ray diffraction patterns after this treatment. xii INTRODUCT ION This study began in 1968 as the result of fertility trials on tomatoes at the Sodus Experimental Farm in Berrien County, Michigan in 1967. Large quantities of potassium (about 1500 pounds of K per acre) fertilizer were needed to produce maximum crop yield and quality. Analysis of a soil sample indicated that about 58 percent of the clay was vermiculite. This clay mineral can fix large quantities of K in its structure. The purpose of this study was to determine the origin of this vermiculite, and its distribution in the landscape. Did the clay arrive in the terrace position directly from the uplands? Was it changed enroute? Or was it formed after deposition? As the investigation progressed, the dynamics of the system, chloritized vermiculite i=3 vermiculite were also studied in an attempt to learn more of the natural dechloritization . processes which may lead to vermiculite formation. PART I GEOMORPHIC SETTING IN BERRIEN COUNTY Physical Location Berrien County is located in extreme southwest Michigan. Three glacial moraine complexes occur in the county (Figure 1). In addition, there are extensive, acid, sandy, outwash and lake plains. The early great lakes also influenced parts of the county. The Sodus Experimental Farm is located mainly on the Lake Glenwood stage terrace at an elevation of about 640 feet (USGS, 1927). The soils on this terrace are mostly of the Oshtemo and Kalamazoo series, both acid and leached free of lime to depths exceeding four feet. Part of the Farm is located on lower terraces and alluvium of the St. Joseph River, 35 to 40 feet below the Lake Glen- wood terrace, at an elevation of 595 to 605 feet. The tomato fertility trials were located on a low, well drained terrace, 10 to 15 feet above the present St. Joseph River. A . LITERATURE REVIEW 1. Glacial Geology of Berrien County a. Kalamazoo Interlobate Moraine Complex The Kalamazoo interlobate moraine complex in the southeast part of Berrien county marks the maximum extent of the Michigan Lobe during the Cary glacial sub-stage of VAN BUREN CO. 4 I |:] = I scum: IILE CLEAR AREAS ARE: l. oumsn sums TEASE." q" _ 2. cucut SPILuuvs ' 3. PONOED IATERS, sum 53“, V 'm L _ 4. LAKE sens, sun 4 IS" I]... P m S *2 X Q INDIANA IOTE'. FIOI "07 STATE OF IICNIGAI, DEPT. OF COISEIVATIOI GEOLOCIC SURVEY DIVISION IOIAIIIC SVSTEIS 0F IICNIIAI Figure 1. Major river systems and glacial geology map of Berrien County. VAN BUREN CO. CO. CASS Wisconsinan glaciation (Figure l). The moraine has been approximately dated at 20,000 to 23,000 years before present (B.P.) (Zumberge and Potzer, 1956). (See Table l.) The Kalamazoo moraine complex ranges in elevation from 800 to 900 feet (U.S.G.S., 1927) and consists of dominantly coarse textured, low-lime materials. When the ice retreated from the moraine a vast torrent of meltwater poured southward between the moraine and the retreating ice front into the Kankakee drainage system (Wayne and Zumberge, 1965). The elevation of this sandy torrent channel ranges from 780 to 700 feet and grades to the south. b. The Valparaiso Moraine CompIex After developing the Kalamazoo moraine complex, the Michigan Lobe downwasted and retreated for some dis- tance and then re-advanced, pushing up and mixing debris to form the broad Valparaiso moraine, till plain and out- wash complex, 6 to 8 miles to the west of the Kalamazoo moraine (Figure l). The age of the Valparaiso moraine complex has been set at about l7-20,000 years B.P. (Zumberge and Potzer, 1956). The separation between the Kalamazoo and Valparaiso moraine complexes has generally been considered to be the torrent channel, which was probably never closed off by the ice advance which pro- duced the Valparaiso moraine complex. The outwash that Table 1. Summary of late Wisconsonian history of the Michigan Lobe and history of the Great Lakes in the Michigan Basin.* Years B.P. Wisconsonian Chronology Lake Stage (Strand- line elev. in feet above M.S.L.) ("little ice age" of Present the Cordilleran Lake Michigan (581) Regions) 2,500 Algoma (595) Interglacial Thermal Maximum Nipissing Great 3,500 (XEROTHERMIC INTERVAL) Lakes (605) 5,000 §°rth Bay ice be°°m°3 Lake Chippewa (230) ree 6 000 General downwasting Payette and other ' period "Lower Algonquin" Lakes of Stanley Retreat Lake Algonquin (605) Toleston (605) VALDERS MAXIMUM Calumet II (620) Intraglacial TWO CREEKS Bowmanville (below 11,000 INTRACLACIAL 480) Calumet I 13,000 C Port Huron Glenwood (640) 15,000 A Lake Border 17,000 R Tinley Y Valparaiso Minooka Substage Tazewell and older Wisconsonian substages 20-23,000 23-39,000 Kalamazoo Moraine Farm-dale Substage *Zumberge and Potzer, 1956; Wayne and Zumberge, 1965. was formed in front of the eastern border of the Valparaiso moraine complex slopes to the east and south (USGS, 1927) and merges into the old torrent channel (D. A. Lietzke, 1971, PawPaw Twp. Soil Survey, Van Buren County, unpub- lished notes). The Valparaiso moraine complex is a broad feature in Berrien County, where three distinct parts occur (Figure 1). The inner or west border of the moraine com- plex lies at an elevation of 650 to 690 feet while the crestal portions lie at an elevation of 750 to 800 feet (USGS, 1927). c. Lake Border Moraine The Lake Border moraine is a lesser feature in Berrien County (Figure 1). It was formed by a readvance or pulsation about 15,000 years B.P. (Zumberge and Potzer, 1956). Through most of Berrien County the Lake Border moraine consists of reworked Valparaiso morainal material that was pushed up onto the Valparaiso moraine. After the ice front downwasted from the Lake Border moraine the pre- decessors of present day Lake Michigan came into being. 2. The Early Great Lakes in the Michigan Basin a. Lake Chicago. Lake Chicago was formed when the Michigan ice lobe began to retreat into its basin (Wayne and Zumberge, 1965). Lake Chicago consisted of three stages: Glenwood stage, Calumet stage and Toleston stage (Leverett, 1898). The Glenwood stage came into being about 14,000 years B.P. and lasted for some 2,000 years (Hough, 1958) until the outlet at Chicago was eventually lowered to around 620 feet and Calumet I at 620 feet came into being (Hough, 1958; Leverett and Taylor, 1915). Calumet I existed through the Port Huron retreat (Wayne and Zumberge, 1965). A low water stage, Bowmanville, occurred during Two Creeks time and during Valders advance lower lying outlets were cut off and Calumet II also at 620 feet was created. According to Wayne and Zumberge (1965) this stage was not static but outlet downcutting to bedrock at Chicago gradually lowered the outlet until a still stand occurred. This stillstand is known as the Toleston stage, 605 feet elevation, of Lake Chicago. b. Lake Algonquin For a short period before Lake Algonquin there was a lower-lake stage known as the Kirkfield, (Wayne and Zumberge, 1965). Lake Algonquin was created at an eleva- tion of 605 feet when the lakes in the Michigan and Huron basins coalesced. This occurred as Valders ice retreated about 10,000 years B.P., and lasted until 9500 years B.P. when the North Bay outlet was opened by ice retreat (Hough, 1958). From about 9500 years B.P. until about 4000 years B.P. there was a period of low (Lake Chippewa, 230 feet) lake levels. Then lake levels gradually rose as crustal rebound raised the North Bay outlet, until the 605 foot outlets of Chicago and St. Clair were again used. c. Lake Nipissing Lake Nipissing came into being at about 4000 years B.P. (Hough, 1958). This Lake stage occupied the Michigan Superior and Huron basins. This lake stage lasted for about 1000 years. During this period the bedrock floored Chicago outlet was abandoned due to more rapid downcutting of the St. Clair outlet. d. Lake Algoma Lake Algoma was a transitory lake that lasted until the St. Clair outlet was downcut to 580 feet which is about the present day elevation of Lakes Michigan and Huron. The downcutting lasted for some 2,000 years (Hough, 1958). e. Lake Michigan Lake Michigan came into being some 2,000 years B.P. Lake levels have fluctuated some i 10 feet since then due to variations in the climate. 3. Sequence of Glacial Events that Formed the St. Joseph River, and its Surrounding Landforms.. The upper reaches of the present day St. Joseph River are in the outwash channel that sloped southward into the Kankakee River (Leverett, 1897), but are not of great concern in this study. The Valparaiso moraine and outwash complex is important since the present St. Joseph River cuts through this moraine complex (Figure 1). Val- paraiso outwash south of Berrien Springs slopes to the south, and north of Berrien Springs, to the west and north (Figure 2). The Valparaiso moraine, till plain and out- wash complex was designated as Landscape Level Ul (U = Upland). After the ice front downwasted from the Lake Border moraine pulse, the Glenwood stage of Lake Chicago came into being. At this time the early St. Joseph River was formed. A new lower outlet forced the river to change its course from the Kankakee outlet and swing around through northern Indiana, across low places in the Val- paraiso and Lake Border moraine complexes and thence into Lake Chicago in the vicinity of Benton Harbor (Leverett and Taylor (1915). The Glenwood stage of Lake Chicago deposits, 640 feet elevation appears as the broad, nearly level sand plain from Benton Harbor upstream to the Sodus Farm (Figure 10 .Ho>Hm £m6m0h .um on» mcoam mcowum>mam 24220.: 92.. soc... 33.5.5.5 mud: .N musmflm a... ”2.4 aam._ \ I I .I. H IIIIIIIIIIIII m wilburm a 3:... 2.322... 8:3 I“ 108 a I I Fa: .c $3.133 $6“ .54“ u _ 8.3 .623 Etta & a: J ox \ z ._ voum SE :3 . r SPFI «II I AM m to 8353 3.5 _ x... .0 _ \' _ \Wu‘a‘ ‘0 \0 ‘oww .\ \ \I ‘ \\ I I I I qu4¢ u \A’OA‘ I I \o. \o)? room .I \ \es .2: l a . rONN I l x a I .?Y\\\ Amy. 1 165 1" \e/d. . 102. ‘J D .v. 46/6 a $53.0 . v, . 1 &/$ 1005 so; . x .7 19:1 I 5 $53 $8 . .5! :3 (“IS") 133:! NOIIVA3'13 11 2). A delta, now situated just downstream from Sodus Farm probably migrated there from its beginning point near Berrien Springs (Leverett, 1898). From Sodus Farm the river terrace grades uniformly upstream (Figure 2). Up- stream from Berrien Springs the St. Joseph River cuts into Valparaiso moraine and outwash deposits. The Lake Glenwood stage lake plain and river terrace materials were designated as Landscape Level U2. The Two Creeks intraglacial period ended the Glen- wood stage of Lake Chicago (see Table 1). With the Valders ice advance the Great Lakes outlet changed again and the Calumet stage of Lake Chicago was formed and stood at an elevation of 620 feet. Leverett and Taylor (1915) reported that most of the evidence of this lake stage had been destroyed by subsequent events. There is, however, a remnant of this lake stage at Sodus Farm. After the Calumet stage ended there was a gradual lowering of the Lake level to about 605 feet. During the existence of Lake Algonquin the St. Joseph River en- trenched itself in the Glenwood stage deposits. It is not known for certain whether the terraces in the St. Joseph River valley lying at an elevation of 605 feet are of Lake Algonquin age or younger. After Lake Algonquin ended there was a period of very low lake levels. The lowest, Lake Chippewa, stood at about 230 feet elevation. During this low lake level 12 period the St. Joseph cut and widened the valley it now flows in. Probably the valley floor was eroded consider— ably lower than its present elevation. The valley width and depth was controlled by the materials down through which the river cut as well as various river and Great Lakes base levels. Lake Nipissing brought a rise to 600-605 feet lake elevation. During this stage there was a rapid period of deposition in the St. Joseph River valley. The deposits consist of very coarse sand and fine gravel with few fines. The elevations of Lake Algonquin and Nipissing are about the same so there is some question as to whether the age of the deposit situated at an elevation of 600-605 feet is 8500 or 3500 years B.P. (Table 1). This deposit was designated as Landscape level Tl (T = Terrace). If there was a period between Lakes Algonquin and Nipissing when there was extensive erosion of the St. Joseph River valley bottom then the Level Tl deposits are probably the product of Lake Nipissing. In any case, the deposits of Lake Algonquin would have been reworked during Lake Nipissing time. Lake Nippissing time was a period of rapid deposi- tion. During Lake Algoma time elevation 590-595 feet, there was a period of some erosion and resorting of Lake Nippissing aged sediments. Present Lake Michigan stands at an elevation of about 580 feet. St. Joseph River has 13 become entrenched into the Lake Algoma terraces. The well and moderately well drained terraces of late Lake Nipissing and early Lake Algoma have been designated as Landscape Level T2. The somewhat poorly drained terraces along the St. Joseph River have been designated as Land- scape Level T3, and are mid to late Algoma in age (Table 1). Landscape Level T4 consists of reworked older deposits. Level T4 deposits are poorly drained and of recent age (Table l). 2;; The Landscape along the St. Joseph River near Sodus Farm In going up the St. Joseph River from Lake Michi- gan, the first noticeable narrowing of the river valley occurs just downstream from Sodus Farm (Figure 3). This narrow part of the valley is 1/8 to 1/3 mile wide, and denotes the place where the St. Joseph River emerges from and turns westward along a section of the Valparaiso moraine. Pipestone Creek flows between ridges of the moraine and enters the St. Joseph River here. Downstream from the constricted part of the valley the river flood plain widens out and the river has developed a mature meander pattern. This part of the river to its outlet has as its base level the level of Lake Michigan. Of in- terest, however, is that the river has straightened itself out thus indicating that it no longer is in equilibrium. 14 Upstream from the Sodus Farm to the dam at Berrien Springs, the river valley has been widened from 3/4 to 1 mile. The river has reached grade, and, as revealed by its sine generated curves (Figure 3), is dynamically self- regulating (Leopold, Wolman and Miller, 1964). Consequent- ly, at the constricted part of the valley, where the St. Joseph River leaves the Valparaiso moraine, there is a boulder pavement or bedrock threshhold which prevented the river stretch upstream from downcutting as rapidly as that downstream during the period of low Great Lakes levels. There probably was a rapids in this constricted part of the valley. The major part of Level T1 occurs between Sodus Farm and Berrien Springs. Most areas of Level T2 are at Sodus Farm. The present river gradient is steeper than the Lake Algoma terraces in the river valley. Thus, the better drained soils are closer to the constricted area at Sodus Farm. They become more poorly drained upstream towards Berrien Springs (Figure 3), with Level T3 terraces, which are somewhat poorly drained, more common. b. St. Joseph River Landscape from Berrien Springs to the Indiana State Line The dam at Berrien Springs raises the water level of the river about 22 feet. Subsequently, all T1, T2, T3 and T4 terrace levels are under water, until just \ I ‘ ‘ .y- An;ExE)T‘.rF/°II3M7IT’“’, (I? It; 1 I! u 5:4.” 3:53 f ‘1 "“1 ““ “i“ :7’. I w f_. Rid $1; 3!: '9 Fig.3 AIR PHO o MOSAIC OF A PORTION OF THE ST. JOSEPH RIVER 16 downstream from Buchanan where Batchelors Island, which is level T1 or T2, occurs. Lake Glenwood stage terraces, U2, are on either side of the river from Berrien Springs up- stream to the state line. This terrace grades upstream uniformly and has been cut into Valparaiso aged outwash which sloPes in the opposite direction (Figure 2). Another dam at Buchanan covers terrace deposits up to Bertrand where Level T4 is just above water level. B. MATERIALS AND METHODS Aerial photographs and topographic quadrangles were used to obtain an overview of the length of the St. Joseph River in Berrien County. Field observations began in 1968. The initial step was to try to delineate the problem area. The soil profile at the tomato plot site was sampled. The second step was to sample all terrace levels within the St. Joseph River valley at the Sodus Farm. Only the surface, 0 to 6 inch depth, was sampled. The third step was to find some benchmark amounts of vermiculite in the alluvial soils of all the major water- sheds in Berrien County. One sample was taken in each from approximately the same position of the recent allu- vium. The fourth step was to sample the major upland soils within 1 or 2 miles of the St. Joseph River. Samples covering all four steps were collected in 1968, 1969 and 1970. The object of this study was to characterize soil clays, not soils. For methods of analyzing soil clays, see the Materials and Methods sec- tion of Part II. 17 C. RESULTS AND DISCUSSION 1. Relationship of Clay Mineralogy to Landscapes This section relates clay mineralogy to landscapes. The assumption is made that the sediments sampled do re- flect their respective landscapes. Surface samples were gathered in most cases. Vertical sections were sampled on Levels U1, U2 and T2 in order to determine changes, if any, in clay mineralogy with depth. a. Clay Mineralogy of Watersheds Other than the St. Joseph River in BerrIen County In order to establish a base level of vermiculite content in the alluvial sediments, samples were taken in all of the major watersheds in the county. Figure 4 shows the locations of the sample sites. The PawPaw River drains a large watershed of mostly acid outwash soils that are in the Kalamazoo and Valparaiso moraine complexes. The clay mineralogy of Sample ll-lB (Figure 5), shows the presence of consider— able smectites, vermiculite, intergrade clay minerals, mica and kaolinite. Quantitative clay data in Table 2 shows that this sample contains 22% vermiculite and 2.7% 18 A N | I: = I SQUARE NILE 19 SI. JOSEPH VAN BUREN CO. EARL L. § V .. 7‘ I, c EA” CLAIRE .N L—(/ 0,510,, t” / " (I? I IIICNAIAI f f / Figure 4. INDIANA Site location map of Berrien County. VAN BUREN CO. CO. CASS *3 “23% 2:9. :8 3.13 a 2:: g: oi and «5'6 v v' o In N” _c_$ :5: I: Ev I | A 7 I. i I - l l " Inn—"T—Tfii—‘i ; ' ( , GALIEN R. I ~ . I .' K ' v , 55° l3-l B I _ .I _‘ #1 4.. 3... “Mg K/NH4 %v «.033 25 I0 3.7m “9'90 AD K550 PAW PAW R. lI-IB “no “lug K/NH4 xv 60 25 22 “9:30 2.7 %K 1‘ A,— Figure 5. PAW PAW R. 'f Caxug KIIIII. %v 42 25 II WM 2.3 H “9'qu A0 l968 flood Deposn X-ray, CEC, and total K data for Samples 13—1B, ll-lB, and ll—laB. 21 Ho>fim smomOh xmouo Enmm .Dm on mumpsoflne h.m Ha H.n a: mo cozmuoum3Iodm m HImH um>wm gammOh . . I .um on humusoflne m H o v n a: Momuu Eumm m H om moaflz um Hm>fim £m6m0h . . I .um on mnmusnflne m H om a A no omflmmzoo m H ma Hm>fim summon . . . own mco need I .Dm ou humusnflue m m ma m m ND a: x U u .m m H m Hm>fim summon . . mom who 0a .um ou humusowue m m om o 5 N: x U x .m m HINH cmmfleoflz oxen . . oucH mmflumEm n m oa m n as sowamo m HIma Ho>am 3mm 3mm mo nonumm coucmm :A m.~ Ha o.k m: .Hs . m RHIHH man coca mmnumem h.~ mm m.m mo .Hs 3mm 3mm m HIHH M mafia Aommc wwwwmma Hm>0q mmeEom Hobos ISUREH0> mm ommomccmq cmnmnoumz omMMm onEmm w w UGMHQD lmv .H .Umnmumum3 uo>wm smomOH .um on» :mcu Hmnuo hucdoo cofluuom cw moonmumpm3 Momma on» mo.EsH>5HHm may cw coauoonm hoao on» no ucmucoo onwasoflfiuo> .N manna 22 total K (percent K x 12 = percent Illite). Sample ll-la B is a sample 2 inches thick that represents the June, 1968 flood deposition and reflects the mineralogy of reworked sediments from somewhere upstream (Figure 5). Of interest is the kaolinite content, which was the highest in any sample gathered in this study. The reason for the high kaolinite content is not known. Perhaps the water turbu— lence at the sample site caused an enhancement of the 2 to 2 micron particle size fraction in which most of the kaolinite fraction is concentrated. In contrast to sample ll-l B this sample contained quantitatively lower amounts of vermiculite and total K (Table 2). In addition, the intergrade or interlayered clay fraction is more heat stable as shown by the unsymmetrical 10 A° peak with the K550 treatment. The Galien River is contained in the Valparaiso moraine complex and most of the morainal materials have a loamy texture. The sample site (Figure 4) is close to the mouth and sample l3-l B is from a similar position to sample 11-1 B from the Paw Paw River. The qualitative clay mineralogy (Figure 5) reflects a higher clay content, as expressed in the mica content by total K analysis (Table 2). Sample 13-1 B, in addition, contains some smectites, vermiculite, interlayered and interstratified mica- vermiculite, and kaolinite. 23 The clays from the Paw Paw and Galien river alluvial sediments show the effect of upland soils on clay mineralogy in total K and vermiculite content. The Galien River up- land soils originally containing more clay and lime are less weathered than the sandy Paw Paw River upland soils. This is reflected in the total K content of the soil clay. Although the age of the upland soils in both watersheds is similar, variations in initial materials and subsequent weathering have caused significant differences in clay mineralogy. Hickory Creek (Figure 4) is almost wholly contained in Glenwood stage, Level U2, deposits. The upland soils are well drained, leached and acid to 3 feet or more. The clay mineralogy of sample 12-1 B (Figure 6), shows the presence of smectites, considerable vermiculite (Table 2), mica and kaolinite. Of interest is that the 17.65 A° smectite peak in the Mg-gly AD treatment collapses to 14A° with the KAD treatment. This reflects the inter- stratified, partially interlayered chlorite smectite nature of the alluvial sediment clays derived from acid upland soils. Pipestone Creek drains loamy Valparaiso moraine soils in the upper watershed and acid leached Glenwood stage sandy soils in the lower watershed. The physical location of Pipestone Creek and the sample site is shown in Figure 3. The clay mineralogy of sample 9-1 B (Figure 24 6), is significantly different from those in any of the other watersheds. It contains a very distinct l4 A° com- ponent, but a very low smectite content. Other than these differences, the quantitative analysis (Table 2) shows this sample to be similar to the Hickory Creek sample 12-1 B. The Dowagiac River drains a large area of acid outwash soils on the Kalamazoo Interlobate moraine complex and old torrent channel. This watershed is the oldest in relation to all of those studied. Sample 15-1 B is located close to where this river empties into the St. Joseph River (Figure 4) and reflects the age of the upland soils. The clay mineralogy (Figure 6) shows the presence of con- siderable smectite and vermiculite. The original illite component has been mostly lost due to weathering processes. The smectite component collapses from 17.65 A° to 14 A° with the K AD component, however, readily collapses to a 10 A° spacing treatment and very little to 10 A°. The 14 A° with the K550 treatment. The CEC data (Table 2) shows about 30% vermiculite. Total K content at 1.6% is very low and reflects the old and leached upland soils. Farm or Farmers Creek drains a watershed composed of Valparaiso moraine complex aged deposits. A soil map (not shown) of the watershed showed the major soils to be well or moderately well drained, with loam or clay loam surface textures. The remainder of the soils are well 25 NM 0- NF 8 8 1‘ I6 I I “I000 -l3.80 'IQA -I765 HICKORY CR. III-I E3 K550 Kjwq K/IIH. xv 60 30 20 W, 2.7 x K A0 K550 PIPESTONE CR. Km 9-I 8 “Mg K/IIII, xv 5| 20 IS _%3"26xx .. .. x550 ‘° " DOWAGIAC R. .. . - l5-I B .. “nag “mu. xv I00 43 38 K” stx no "°|’ A0 Figure 6. X-ray, CEC, and total K data for Samples 12-lB, 9-1B, and 15-lB. 26 drained, acid, outwash soils that are situated on the banks of the creek. Sample 20-1 B (Figure 4) is from the main Farm Creek stream. Sample 19-1 B is from a small subwatershed composed of clay loam soils. The x-ray data of these two samples (Figure 7) reveals considerable differ- ences in the clay mineralogy. Finer textured upland soils contribute more mica to the alluvium. The interlayered l4 A° clay component is also less stable in sample 19-1 B than in 20-1 B. Quantitative data are in Table 2. From Table 2 there is a range of 6 to 30 percent vermiculite, with an average value of 16 percent. Total K content ranges from 1.6 to 3.7 percent with an average value of 2.6 percent. In general those watersheds that contain acid coarser textured outwash derived soils con- tain more vermiculite in the alluvium. These watersheds are the Paw Paw, Hickory Creek and the Dowagiac River. However, study of these watersheds shows that other factors must be considered, such as age of the upland soils, length of residency of alluvium in the floodplain, ori- ginal clay mineralogy and chemical composition, pH of the upland and alluvial soils. by_ Claprineralogy of the St. Joseph River Watershed in Berrien County The discussion and results of this section are organized from the uplands and upper watershed downstream .._-—.... 27 II “In I I I m w _ I- NKSSO .. ,, FARMERS CREEK . l9-IB ° ° K300 '° '° CW9 K/NH4 xv '° -° 39 22 II - .. 3.7 x K KAO _, “rely A0 FARMERS CREEK K550 20-IB K300 Kfioa/Mg K/NH4 75V 40 22 I2 . I.9%K '19-qu a .0 AD 9'"! II I I I II I - I .I Figure 7. X-ray, CEC, and total K data for Samples l9—lB and 20—1B. 28 to the lower terraces and lastly the most recent deposits. The vermiculite and total K contents are shown in Table 3 and the x-ray diffraction results are shown in the Appen- dix or Part II. The samples from Level T4 are most com- parable to those in the other watersheds in Table 2. The upper reaches of the St. Joseph River water- shed is represented only by sample l4-1 T4 (Figure 4). This sample is poorly drained and sandy. It represents the sediment load derived from the Kalamazoo moraine and older moraines and outwash deposits. The clay mineralogy of 14-1 T4 (Figure Al), is very similar to the Paw Paw, Dowagiac and Hickory Creek watersheds. There is a highly interstratified soil clay composed of smectites, inter- layered or chloritized vermiculite and mica. Kaolinite is also abundant. Along Transect II (Figure 4) samples 24-la U1 and 23-1 U1, represent outwash of the Valparaiso moraine. This outwash is the oldest material encountered along the St. Joseph River in Berrien County. The samples are located within 2 miles of the river valley and represent much of the surface erosional sediments that have been transported into the St. Joseph River valley. Sample 24-la Ul is the surface layer of soil pro- file 24 U1. The clay mineralogy (Figure A2) shows an abundance of a moderately heat stable chloritized vermicu- lite but only 5 percent vermiculite is indicated by 29 CEC data. Examination of the soil profile suggested that it had developed at least in part under a grass vegetation. The interlayer material may be an Al-organic complex which may explain the instability of this 14 A° component to heat and its low percent vermiculite. Total K is low-—l.9 per- cent-~suggesting prolonged weathering. Sample 24-2, from the B horizon of this soil profile shows a strong 21t smectite component in addition to the 14 A° component. Total K is low--2 percent--and CEC analysis shows about 11 percent vermiculite. Sample 24-3 is from the B3 horizon of this soil profile. The clay mineral assemblage is re- lated to the horizons above. There is more mica--3.2 per- cent K--but a lower kaolinite component than in 24-la and 24-2. Since the quartz content by x-ray diffraction peak height is similar for all three samples, this leads one to suspect that there has been kaolinite formation as a diagenetic product in the strongly weathered upper hori- zons. Sample 23-1 U1 is similar in clay mineral assemb- lage to sample 24-la U1 (Figure A3). Highly interstrati- fied materials are dominant. The improvement in resolution with the KAD treatment is due to better organization of the clay structures. Table 3 summarizes the data for Level U1. Landscape Unit, Level U2 is some 8 to 10,000 years younger than Level U1 and consists of Glenwood stage 30 Emmuumms moHHE m m.N m Hams maoumnmvoz m.n NB Hum Emmuumms moafle m ammuumm: mmawe RH v.m m Hams mamumnmcoz k.s Ne Hum 1H9 Hoe pumamH mHoHosoumm w.N o cocflmup Ham: v.m Ne HIGH Emmuumms moHflE m N.N m Umcflmnw Ham: m.m H9 Hlmm Enmm motom um m.H o cocflmuc Ham: H.m HE HIH Emouum In: means Nxa A I- m emaflmue Ham: I. ma HINm Eumm mscom pd o.H m pocflmup Ham: m.v ND HImm Atmeflav Enmm mucom u< m.a m Uocflmuc Hana n.m ND Anna ammnnmms mmnfle ma m.m m emeAmAS Ham: m.o m: HIHN Emonumms moafle ma m.N m Docflmnp Ham: H.m ND HINN ammuumas mmafls ma m.~ e mmcflmuu Ham: H.m A: HImm Apoeflav smmAumms means ma m.H m emcfimuc Ham: m.m A: RHISN Eumm mDUom M .> ommcflmuo Aommv Ho>oq .02 on cofluwmom Hmuoe w amusumz mm ommomccmq mamamm .uoomcmuu uo>wm Daemon .um .m manna 31 .oHumEonmN cmexoom .oDOHDUOHo mmMHm QDH3 max Eumm mscom Dd Emonpmms mmHHE m Emouumms mOHHE m Emmuumms moHHE mN Eumm mDDom pd Emmuumc3oc moHHE m Eumm mspom pd Enmm mspom Dd sums manom n< Show mspom DN Eoouummn DHHE H Econummd mHHE H Emwnumms mHHE H Emmuummd mmHHE m mH NH 5H vm HN Nv 5H mm mm vm vN mH vH pochnp mHuoom Umchup aHuoom DochuU >Huoom UmsHmuU mHuoom DochnU hHHoom uma3mEom HHOB mHmumuoUoz HHo3 mHmumuoUoz HHoB HHmz HHo3 HHoz mHmumnoDoz HHmz hHmumumvoz HHDS hHmumuopoz HHUB aHmumuocoz v9 v9 v9 v8 me Na Na Na Na Na Na Na Na NB mHIm mHIb HIvH HIV HImH Him HImN HIN Hlom HImN HImN Hlvm 32 deposits. The soils are mostly well drained, sandy and acid. Sample 22-1 U2 and 21-1 U2 characterize the upper Glenwood stage terrace on the St. Joseph River (Figure 4). The clay mineral assemblage (Figure A3) illustrates the difference in two very similar soils except for depth of leaching. The Oshtemo soil is leached deeper than the Boyer soil. The major differences between these two sam- ples is the total K content and the degree of heat stability of the chloritized vermiculite. The total K content is higher in the Boyer sample, 21-1 U2, 3.6 percent vs. 2.5 percent. The chloritized vermiculite in the Oshtemo, 22-1 U2 sample exhibits a greater heat stability than the Boyer sample. The kaolinite and quartz peaks are similar as are the percent vermiculite by CEC analyses. Assuming then that the total K soil clay content reflects mainly the weathering rate then the Oshtemo sample is more weathered than the Boyer sample, although the deposit ages are the same. The initial lime content, through preferen- tial sorting during deposition was probably higher in the Boyer than in the Oshtemo soil (and the initial clay com- ponents and amounts may have been different). Samples 17-1 U2 and 35-1 U2 represent Glenwood stage deposits farther downstream at Sodus Farm (Figure 3). The clay mineral assemblage of profile 17 U2, Kalamazoo soil (Figure A4) shows much similarity to profile 24 U1 (Figure A2) except for a greater heat stability of the 33 chloritized vermiculite component. The kaolinite content of 17-1 U2 and 24 U1 is considerably lower, as expressed by peak height than that in samples 23-1, 22-1, 21-1 U2 (Figure A3), but quartz contents are similar. This could be the result of preferential sedimentation on the river terrace or the delta on which 17-1 U2 is located. Sample 17-2, from the B21 horizon, contains highly interstrati- fied clays that with the KAD treatment produces a strong 14 A° peak. It also contains more smectite than the over- lying or underlying sample. The higher vermiculite content in sample 17-3, of the B horizon, is probably due to both 3 weathering processes and that inherited from initial materials. Sample 35-1 U2 is from an Oshtemo soil located less than 1/4 mile north of 17-1 U2 (Figure 3). It is from a sandier soil. This is reflected in the stability of the chloritized vermiculite (Figure 12, Part II) and in the lower total K content. The weathering rate of the Oshtemo soil has been greater than in the Kalamazoo soil. (This area has apparently not been limed as has the soil at 17-1 U2.) Sample 32-1 U2 (Figure 3) represents the loamy till soils of the Valparaiso moraine along the St. Joseph River. This sample, from a clean tilled orchard, is from a low area where surface erosion sediments have accumulated. Sample 31-1 B is a sample from a delta that has built out 34 on Level T4 alluvium. The deposited material is primarily derived from the loamy till soils above. The clay mineral assemblage of these two samples is similar (Figure A5). The chloritized vermiculite component of 32-1 U2 is as stable as that from the Level U2 sandy soils. In sample 31-1 B, however, there are larger quantities of mica and kaolinite. No total K data are available. Preferential sorting undoubtedly has enhanced the mica and kaolinite components (but the decreased chloritized vermiculite stability is probably due to some chemical dechloritiza- tion process). The CEC of 31-1 B is higher but there is less vermiculite than in 32—1 U2. The higher CEC of 31-1 B is due, at least in part, to a higher smectite component. The higher smectite component raises the question of its origin. Is the smectite component enhanced by preferential sedimentation, is it derived from atmospheric dust or is it due to a diagenesis reaction (the result of the de- creased chloritized vermiculite stability)? If the chloritized vermiculite in sample 32-1 U2 is undergoing some dechloritization reaction, are smectites one of the products of the reaction? Table 3 summarizes some of the data from Level U2. Landscape Level T1, of Lake Nipissing age, is represented by two samples. Samples l-l T1 and 33-1 Tl (Figure 3) are remnants of a once extensive higher terrace level. The soils are sandy, acid, and well drained. 35 Sample 33-1 T1 is from an Oshtemo soil and 1-1 T1 is from a sandier Coloma soil. The clay mineral assemblage (Figure A6) is similar for both. Both contain a stable chloritized vermiculite component. The quartz component of 1-1 T1 and its lower mica content reflects the sandier sediments. Table 3 summarizes the data from Level Tl. Landscape level T2, of Lake Algoma age, is repre- sented by samples from Batchelors Island, sample 16-1 T2, 17 miles upstream below Buchanan Dam to 18-1 T2, 3 miles downstream from Sodus Farm. Batchelors Island is a ter- race remnant of late Lake Nipissing age (Figure 2). Most areas of the Lake Nipissing terraces between the Berrien Springs and Buchanan dams are under water. Batchelors Island is now only 6 to 8 feet above the present river level, and is blanketed with loamy alluvial deposits not related to the coarse textured Lake Nipissing terrace deposits. For this reason, as stated earlier, the age of this T2 terrace is of late Nipissing and early Lake Algoma and is, therefore, older than the T2 terraces down- stream which are mid to late Lake Algoma in age. The clay mineral assemblage (Figure Al) shows the presence of a fairly stable chloritized vermiculite, no vermiculite by CEC analysis, and a quantity of quartz that is higher than usual but similar to sample l-l Tl (Figure A6), for exam- ple. 36 The sample pH (Table 3) is the lowest on the Level T2 terrace and about the same as the Level Tl terrace samples. Thus the age of Batchelors Island is in doubt. The soil pH and clay mineralogy place the soil on this island in Level Tl but the alluvium blanket is similar to that blanketing the well drained Level T2 terraces at Sodus Farm. Due to the Berrien Springs dam, the next Level T2 exposure is just downstream below the dam. The clay mineral assemblage of sample 6-1 T2, pattern not shown, includes smectites, chloritized vermiculite and mica, that are highly interstratified. Sample 8-1 T2 (pattern not shown) four miles downstream from sample 6-1 T2 site shows an increase in smectite content, more chloritized vermiculite and vermiculite and less mica. Sample 34-1 T2 (pattern not shown) two miles farther downstream shows a stronger smectite peak, more vermiculite, less chloritized vermiculite and less mica than sample 8-1 T2. Table 3 summarizes some of the data from these three samples. All have high pH, but it drOps as one progresses downstream. Mica content decreases, downstream, smectite content in- creases, vermiculite content increases and chloritized vermiculite content or the heat stability decreases down- stream. Samples 28-1 T2, 29-1 T2 and 30-1 T2 are from one 40 acre site (Figure 3) some 8 miles downstream from the 37 Berrien Springs Dam or 1 mile upstream from the Sodus Farm. The clay mineral assemblage of these three sites is interesting (Figure A7). All three samples are com- posed of similar materials as judged by comparing kaoli- nite and quartz peaks. They have similar natural drainage (Table 3) but relatively speaking, 28-1 T2 is more poorly drained than 30-1 T2 and 29-1 T2 lies inbetween. The significant aspect of these three samples lies in their increasing ages as named. The pH decreases with increas- ing age, as judged by elevation and relative degree of drain- age in the moderately well drained drainage class. Vermiculite content increases markedly from 14 to 34 per- cent from youngest to oldest, respectively. The chlori- tized vermiculite heat stability is about the same in Figure A7 for all three samples. Samples 3-1 T2 (Figure A4), 2-1 (Figure 13, Part II), HV T2 (Figure 12), and 25-1 T2 (data not shown) all have similar mineralogy. The age of all is approximately the same but the degree of natural drainage varies from well to moderately well. The degree of natural drainage is related to the pH (Table 3). Figure 8 shows schematically the location of sam- ple sites in relation to the tomato plot site but not all of the samples shown are discussed here. Table 3 summa- rizes the data from these samples at Sodus Farm. Vermicu- lite content reaches a maximum at Sodus Farm, and appears Tl 39 to be related to degree of natural drainage and pH. Total K content remains nearly constant for the samples at the Sodus Farm on level T2. Sample 18-1 T2, located three miles downstream from Sodus Farm (Figure 3) is in a different cross-section of the St. Joseph River that lies below the constriction in the valley width. The clay mineral assemblage (Figure Al) is similar to other samples from level T2 at Sodus Farm except for higher smectite and kaolinite contents. This sample compared with the T2 samples from the Sodus Farm (Table 3) shows a higher pH, and a lower vermiculite content. Level T3 is represented by sample 4-1 T3 (Figure 3). The clay mineral assemblage (Figure Al), shows a highly interstratified system composed of smectite, ver- miculite, and mica. Like the T2 samples at the Sodus Farm, it is high in vermiculite (Table 3). Level T4 is represented by samples 14-1 T4, 7-1a T4, 8-1a T4, and 5-1 T4. Sample 14-1 T4 was discussed earlier, x-ray diffraction data for samples 7-la T4 and 8-la T4 are not shown, and data for 5-1 T4 is shown in Figure Al. Chloritized vermiculite stability in 5-1 T4 is greater, for example, than sample 4-1 T3 and very similar to 14-1 T4 (Figure Al). The data from Level T4 is summa— rized in Table 3. The pH is higher, vermiculite content 40 lower and total K content slightly lower than that of level T2. c. Summary of Clanyineralogy Data along the St. Joseph River Figure 9 shows in a diagram the vermiculite and total K data for terrace levels T1, T2, T3 and T4 from Table 3. Total K content of the clay changes very little along the St. Joseph River. However, the T1 and T4 levels are generally lower in K than the T2 levels. The varia- tions are possibly due to preferential sorting and deposi— tion as the clay sized minerals or silt sized, if flocculated, are transported downstream. The vermiculite distribution in sediments can probably be accounted for by considering the general dyna- mics of the St. Joseph River. Analysis of sample 14-1 T4 showed that essentially no vermiculite was transported down from the upper reaches of the river. The dams at Buchanan and Berrien Springs prevent any analysis of ter- race levels in these two stretches of the river. But, downstream from the Berrien Springs dam, the vermiculite content increases to a maximum at Sodus Farm and then de- creases below the constricted part of the valley. What evidently has occurred is that chloritized vermiculites from levels U1, 02 and T1, during Lake Algoma time, were eroded into the river, probably during severe storms, tran5ported a short distance and then deposited. If 41 .mucaoo coHuuom DH Ho>Hm Daemon .um on» mcoHo v9 can .ma .NB .HB mHm>mH Eoum monEmm meow mo ucmucoo M Hmuou can muHHDOHEHm> .m mHDmHm .& 9m. 2 VN .3 «3:: EB... EE§ Am A. .39. «setussnxéfi 33>. c9022! 33 so: 53...»: 3:2 260m 0 p bill I My b b E _ p O 3..-: 33:23:; $ 3.-.. / «truO/A/ Iv 'h..l~.° NP-l. °._ 11°— 3730 I -.I»~o 278M :36 o.~..o~ d IIIIIIIIIIIII QII I I I 372—0. I IxIIIIIIII III II \IIN 3:8 .38» Q 2 -o o 3.-.. 2...: 3:. I 273 oérvoe ron Npszo . r )I IDIOI °/o ”magma/(Va g 42 conditions were suitable (see Part II) some kind of de- chloritization reaction began. As the river meander system developed, probably during late Lake Algoma time, the flood deposits were resorted, transported another short distance and redeposited. Under stable conditions, as on levels T2 and T3, the dechloritization continued and the vermicu- lite content increased. At the same time fresher materials from upland surfaces and escarpments that meanders impinged on were added to alluvium. The level T4 samples show the mineralogy of the clay minerals that are now (1968) being transported downstream. Vermiculite content increases and chloritized vermiculite content decreases as one pro- gresses downstream. That sample 5-1 T4, a 1968 flood de- posit sample contains 19 percent vermiculite is not surprising considering that the major source area of this deposit is coming from meander erosion 1/4 to 1/2 mile up- stream from sample 34-1 T2 site (Figure 3), and the source materials are derived from primarily level T3 sediments. In order to determine the magnitude in range of vermiculite content, a portion of level T2 from Sodus Farm upstream as far as sample 34-1 T2 site (Figure 3), was selected. All of the samples discussed so far, plus seven additional samples at Sodus Farm for a total of 14, were statistically evaluated. Figure 8 shows the landscape at Sodus Farm. The statistical analysis consisted of es- tablishing the probability of values falling outside a 95 43 percent certainty (Steel and Torrie, 1960). The vermicu- lite content ranged from 14 to 58 percent with a mean value of 30 percent. Ca/Mg CEC ranged from 61 to 115 meg/100 g clay with a mean value of 84. The K/NH4 CEC ranged from 26 to 46 with a mean value of 37. Total K values ranged from 1.6 to 3.0 percent, with a mean value of 2.2 percent. The statistical analysis showed no signi- ficant differences in total K content. Two K/NH4 CEC values fell outside of the 95 percent probability, and significant differences among Ca/Mg CEC values were both above and below the established probability limits. The statistical analysis indicates the relative uniformity of the sediments, expressed by clay total K content, and that the Ca/Mg CEC appears to depend on the degree of dechloritization or rechloritization that has occurred. Further, general analysis of the relation of pH to percent vermiculite for level T2 samples shows that at higher pH's vermiculite content is low (6-1 T2, 8-1 T2, 34-1 T2) and at lower pH vermiculite also drops, 16-1 T2. The Optimum pH for maximum vermiculite content occurs in the range of 6.2 to 7.0. Table A1 lists the probable soil series from which each of the samples were taken. Part II pursues the dynamics of the dechloritization process. 44 2. Proposed Model a. In the acid upland soils there has been a weathering sequence of illite + vermiculite + chloritized vermiculite, with clays from the different soils in all stages of the sequence. Potassium is weathered from the illite to form a dioctahedral vermiculite. As weathering proceeds and the upland soil becomes more acid, aluminum is made more soluble as A1(H20) and [A1(HZO)X°organic] and enters the expandable part of the vermiculite structure. This Al interlayer material occurs as atolls, islands and pillars which effectively block inner CEC sites (Frink, 1969). Low organic matter content appears to accentuate the chloritization process. The most highly chloritized soil clays in this study have developed under oak forest. b. The acid chloritized vermiculite is eroded into a neutral to calcareous lake environment (Frink, 1969), or in the case of the St. Joseph River in Berrien County, to a neutral or calcareous floodplain environment. 0. At positions high enough for leaching to occur, a dechloritization process causes the removal of the Al interlayer material reforming a high charge vermiculite again. Abundant organic matter present in the terrace soil surfaces appears to accentuate the process. d. Rechloritization can occur again in the flood- plain position if the soil is high enough for leaching to I'D-IRIS: AMEN. IEI. It EH. 4.45 _ 45 occur to a further extent, eventually lowering the pH to below 6.5. Sample 16-1 T2 is a good example of this pheno- mena, or if K or NH4 ions are present in sufficient quanti- ties, the vermiculite can revert to an illite type mica. 3. Evidence to Support Model a. Acid upland soils from Levels Ul and U2 contain abundant chloritized vermiculite in the surface and the amount decreases with depth. b. These acid soils have low illite content at the surface, but this increases with increasing depth along with a decrease in chloritized vermiculite (Profile 17 U2, Figure A4). c. High (>20%) vermiculite contents are found in the soil surface of many low terrace environments. High organic matter content plus a neutral, slightly leaching environment appears to enhance the dechloritiza- tion process. d. In the terrace soils the vermiculite content is greatest at the surface and decreases rapidly with depth (See Profile 3 T2, Figure A4). This suggests that leaching removes the Al interlayer material from the sur- faCe or rechloritization is occurring at some depth below the surface or a reversion to illite, although there is no significant increase in total K with depth in Profile 3T2. e. Rechloritization can readily occur in the floodplain environment once leaching has removed the 46 slight amount of lime deposited with the sediments. Sam- ple 16-1 T2 is the prime example. On the highest, best drained places on the T2 level terrace at Sodus Farm rechloritization is also occurring. This can be seen by comparing samples HV T2 with 2-1 T2 and 25-1 T2 (Table 3). a. Evidence from Two Transects across the St. Joseph RIver Valley Transect I across Sodus Farm illustrates the rela- tionship between clay mineralogy and landscape levels. Figure 10 is a diagramatic expression of the topography, while Table 4 lists the relevant data. Landscape Level U2, Glenwood stage sediments, is represented by two sam- ples on this transect. Sample 35-1 U2 is the surface layer of an Oshtemo soil. This soil has a loamy sand or sandy loam surface and a coarse-loamy control section (Soil Taxonomy of the National Cooperative Soil Survey, 1970). Due to the ease with which water can percolate downward in this soil, it has become highly leached and is quite acid. Consequently, the surface layer contains abundant heat stable chloritized vermiculite and a low amount of total potassium (Figure 12). Sample 35-1 U2 represents the coarser-textured sediments of this level. Sample 17-1 U2 represents the finer-textured sediments. This sample is the surface layer of a Kalamazoo soil. The texture of the surface 47 .Eumm mspom um mmHanOHDMHou mmmomtcmq "H Doomcmua .OH oHDmHm .8... I IIIIIIIIIII : .mow IIIIIII I / / m.— uu . . 3.3 \ 53v“ 3 \\ M. as... \ 5:... 7.. w / ~ $80 n.— \\ / \ ”0—. 48 .mcoHDMUOH ouHm mom m oanHm Dome me on m.N mH v.5 NIo #9 Hum we mm m.N em m.n mIo m9 HIH mN mHH m.N mm v.o mIo NB >D mm moH o.m Nv v.5 OHIo NB HIm nm em m.H o H.m OHIo HE HIH AcmeHHv vm mm m.H m n.o mIo ND HIhH HN «N o.H m m.v =mIo ND RHImm H m m2\x 2\mo M > Acmmc AmmnocHo Hm>0q .oz Hmuoe w mm Damon ommomvsmq mHmEmm w onEmm .AoH ousmHmv H Doomcmnu ommompcmq Eoum sumo .v OHQMB ij ? I.“ 49 NOIIVASIB 00 $40 (.040 Gem 0 to N ow» Eoum Emouumms moHHE NH mmHnmcoHumHmH ommompqmq .Enmm mspom "HH pommcmue .HH mHDDHm / (I .I‘ I I /’I //////I M” / / flaflfl In..I...I.4.........:... z. A. .... atri . In. w . _ _ 50 layer is loam and the control section is fine-loamy. Since this Kalamazoo soil contains more clay, water has more difficulty in percolating downward. Consequently, this soil is relatively less leached and is, relatively speaking, less weathered than the adjacent Oshtemo soil, although the surface age of both is the same. Sample 17-1U2 also contains an abundant quantity of chloritized vermiculite, but which is less stable than that in Sample 35-1U2 (See Figures 12 and A4). Landscape transect II is located upstream from Sodus Farm Landscape transect I, 13 miles (Figure 4). Trans- sect II is shown in diagram form in Figure 11 and support- ing data are in Table 5. Samples 21-1U2 and 22-1U2 from transect II (Figure 11) are also from Landscape Level U2, but 13 miles upstream from the Sodus Farm Landscape tran- sect I. These two samples are similar to sample 35-1U2 in profile characteristics, but are relatively less leached. The mineralogy of samples 21-1U2 and 22-1U2 (Figure A3) is slightly different in that the sediments probably initially contained a higher carbonate content than sample 35-lU2. Consequently, these soils, although probably as old as sample 35-1, still have a slightly higher pH in the sur- face layer and relatively speaking are considerably less weathered than sample 35-1, but the weathering sequence is less advanced. Samples 21-1U2 and 22-1U2 contain more illite than samples 17-1U2 and 35-lU2 and a larger quantity 51 .coHuosponm mono now DmEHH oommnsmx MN mm m.m m m.m mIo ND HIHN em we m.N m H.@ mIo ND HINN mN mm m.N v H.m on HD HImN mm we m.H m «m.m =HHIo HD mHIvN sm2\x m2\mo m .> AONE gamma Hm>0q .oz omu w w mm mHmEmm ommomcqmq mHmEmm .AHH oHDDHmv HH nommcmuu manomccmq Eonw mama .m mHQme 52 of vermiculite that is in a less advanced stage of chlori- tization, whereas in samples 17-1U2 and 35-1U2 much of the illite has been converted to vermiculite and most of this vermiculite has been chloritized. Samples 23-1Ul (Figure A3) and 24-laUl (Figure A2) represent Landscape Level U1, the oldest and highest level in this study. The soils that these samples repre- sent have a different mineralogy in that they contain considerable coarse fragments of shale. They also developed under a grass vegetation for some time and this has re- sulted in the darker, thicker, and more organic matter rich surface layers. These two samples are several thousand years older than the samples of Landscape Level UZ and, consequently, even though they are finer textured they are highly leached and quite acid. Due to the highly leached condition of these samples chloritized vermiculite should be abundant in the clay. These two samples do contain a chloritized vermicu- lite component, but not as heat stable, as one might expect. The higher organic matter content may interfere with interlayer formation. In fact, the poor heat stabil- ity of the chloritized vermiculite from these two samples may be due to an organo-alumino interlayer complex. The clay in these two samples also contains a larger smectite component, especially in the lower subsoil layers. The higher organic matter content may change the sequence of 53 mineral alteration. Instead of the vermiculite + chlori- tized vermiculite stage the sequence in these two soil profiles may be vermiculite + aluminous smectite or expand- ing fine grain low charge vermiculite which is translocated to the B horizon and deposited as argillans (Brewer, 1964), to chloritized vermiculite and smectite. Due to the age of Level U1 sediments textural variations should result in coarser-textured soils being relatively much more weathered than finer-textured soils. Samples 24-1Ul (Figure 16a) and 24-1aUl (Figure A2) illus- trate this extremely well. Sample 24-1Ul was the original sample. The sediments in the profile included a fairly high percentage (20-25%) of coarse fragments. Consequently, this soil has a higher percolation rate than a soil of the same texture but without coarse fragments (the coarse fragments dilute the clay content by increasing the volume of total soil material). The clay mineralogy of sample 24-1U1 (Figure 15b) showed the largest quantity of highly heat stable chloritized vermiculite before sample 35-1U2 was taken. This sample, 24-lUl, taken in September, 1968, was to be used for the interlayer dynamics experi- ments which were to follow. The following spring, May, 1969, sample 24-laUl was taken within a 6-10 foot radius of sample 24-1. Samples 24-2 and 24-3 were taken from the same profile as sample 24-laUl. The clay mineralogy was much different from 24—1Ul. A check of sample 24-laU1 54 revealed no coarse fragments such as were abundant in sample 24-lUl. The lesser degree of leaching in sample 24-laUl had produced a considerable quantity of chlori- tized vermiculite but it was not the highly stable material desired for the dynamics study. In order to clear up this situation, five more samples were gathered within a 100 foot radius of the initial sample site of 24-1Ul, and one more sample was located near the sample 23-1Ul site, but in a wooded area. All six samples showed x-ray diffraction patterns similar to 24-laUl (Figure A2). They contained considerable smectites and unstable chloritized vermicu- lite. Also, none of the additional samples contained any appreciable coarse fragments. Thus, the original clay and coarse fragment content affect the weathering rate of soils. b. Clay Mineralogy within Soil Profiles Four soil profiles were sampled in depth in order to ascertain the clay mineralogy and weathering sequences. A maximum of three samples were taken from each profile; the surface organic rich layer, the upper subsoil layer and a lower subsoil layer. The data are contained in Table 6. The profile, 24U1, represented by samples 24-la, 24-2 and 24-3 (Figure A2) is in the Lydick series (Table A1). The surface layer had been limed previously and con- siderable lime had been applied from road dust. The soil 55 ommum DooscoHO oxmqv .Dmpoon mHmHmu mH Dom mums» oom.N cusp mmmH mH omm NB Hm>me«« .m.m ooo.mHImH mHmmeonummm mmm HD Ho>mH¥ .Aoomunou .m.m whom» ooo.NH mHoumEonummm mom ND Ho>oqxx Nm he o.m m m.m omION NO mIm mm mm o.m hN m.> ONIOH Ho NIm Hmm mm mcH o.m me H.» =OHIo ma ma HIm mm mm v.N OH v.h ONIvH m NIN Hocssm mm mm m.N Nm n.m =mIo DN Ne «xxHIN mHHom oomnuma uo>Hm ON mm m.m oH m.m NhImo mm NINH me om n.H m v.m mHIm um NINH oonEmHmm em mm m.H m v.0 =mIo mm ND «xHINH om he N.m HH m.m mmIom mm mIvN He mm m.N HH N.v mmImH um NIvN onesq mm as k.m m m.m .HHIo m< H: .mHIHN . N AHm mHnmev vDz} mz\mo M Hmuoa w «memo enema .Aom mmHuom Hm>oH .oz HHom umo mHHom ADDMHQDV nmmsuso mmcomocmg mHmEmm .mHo>oH mmmomDDMH owns» Eoum mumw oHHmonm HHom .m mHQme 56 representing one portion of Level U1 was the most acid profile sampled. This is understandable since it has been developing for the longest period of time. There is one major difference between samples 24-la and 24-2. This is the smectite clay component in sample 24-2, but the chloritized vermiculite component stability appears to be essentially the same in both samples. Sample 24-3 shows both discrete 14A° and expanding smectite peaks. The pro- file illustrates the weathering sequence of Jackson, gt_al. (1952). The illite content increases with depth. The vermiculite content also increases with depth as is ex- pected. Some of the vermiculite in sample 24-3 is probably inherited from the initial sediment composition. The remainder has been derived by the loss of K from the illite component. The profile, l7U2, represented by samples 17-1, 17-2 and 17-3 is in the Kalamazoo series (Table A1). This profile occurs on Level U2, and is less acid than profile 24U1. This can be due to a lower surface age and perhaps an initially higher carbonate content, affecting the weathering rate. This profile shows the same trend as the 24Ul profile in illite, vermi- culite and chloritized vermiculite contents (Figure A4). Geomorphic Level T2 is represented by profiles 3T2 and 2T2 (Table 6). Profile 3T2 is an example of the Eel series (Table A1). This profile exhibits different characteris- tics than profiles 2401 and l7U2. The vermiculite content 57 is maximum in the surface layer and decreases rapidly with depth while illite content remains constant. The x-ray diffraction data for profile 3T2, samples 3-1, 3-2 and 3-3 (Figure A4) shows a large vermiculite component in 3-1 that collapses fairly readily upon K saturation while samples 3-2 and 3-3 show a more stable chloritized vermi- culite component. Another explanation for the decrease in vermiculite with depth in this profile is fixation of NH4. However, infrared absorption data (not shown) showed no definite increase in NH levels with depth. 4 The constant total K throughout the profile indicates no significant K fixation is occurring. The higher organic matter content of the soils on Level T2 must be involved in the interlayer removal process. The tOp layer is sub- ject to more leaching. The interlayer Al is removed from the interlayer position by leaching and becomes attached to exchange sites on the organic matter where it is trans- located by percolating rain water. The organo-alumino complex is leached to a lower part of the profile where it is precipitated in a higher pH environment, since the high charge vermiculite there is susceptible to rechlori- tization. If the pH becomes acid in the Level T2 soils through excess leaching the vermiculite in the surface layer also becomes rechloritized. Sample 16-1T2 illus- trates this very well (Figure Al). In the interlayer dynamics section (Part II) the stability of the interlayer 58 material is studied in more detail in an attempt to learn of the removal and fixation of interlayer material in natural high charge soil clay vermiculites. D. CONCLUSIONS OF PART I l. The occurrence of vermiculite in the river terrace soils and sediments is due to a diagenesis process, dechloritization. 2. The distribution of vermiculite in river ter- race soils depends on: (a) the nature of the soil materials being eroded into the river, (b) the length of time neces- sary for dechloritization to occur which varies with leaching rate and pH. 3. The occurrence of chloritized vermiculite in upland soils is widespread but the quantity and stability is due to the degree of leaching (original lime content of parent material and clay content), and vegetation. Origi- nally low lime, low clay content materials under poor oak vegetation (which produces little organic matter) have the largest quantity and most stable chloritized vermicu- lite soil clay component. High organic matter content appears to work against chloritization as the chloritized product is a heat unstable organo-alumino interlayer com- plex. 4. The decloritization-rechloritization process appears to be dynamic in that large CEC changes can occur within a short time Span. The vermiculite formed by 59 6O dechloritization has a high charge and is very unstable. A small pH shift to the acid side causes a rechloritiza- tion to occur. Higher organic matter contents of river terrace soils appear to increase the dechloritization rate. 5. The concept of only six landscape surfaces in a complex area of glacial sediments such as there is in the St. Joseph River watershed may over-simplify the situ- ation since one landscape surface can include both till, outwash and lacustrine materials. The concept of landscape surfaces in this paper is one where age is the primary reason for separating one surface from another. 6. The possibility arises that the conclusions reached are based on biased samples. Does a sample of alluvium from a tributary draining a morainic watershed reflect the unbiased clay mineralogy of the upland surface soil material that is being drained into the channel? 7. What effect does river sorting have in causing bias in samples? Much of the fine clay carried by the St. Joseph River lies on the bottom of Lake Michigan. 8. What has been the effect of long periods of dust accumulation on differences in clay mineralogy be- tween acid upland soils and the alluvial terrace soils? Analysis of the results of this study thus far confirm the illite weathering sequence of Jackson, §E_al. (1952). Frink (1969) prOposed a dechloritization reaction Ifluat was occurring in lake bottom sediments. It has been 61 proposed that a similar dechloritization reaction is occurring on the recent well, and moderately well drained terraces of the St. Joseph River as well as in the sedi- ments of other river systems that meet the necessary requirements. Part II pursues the dynamics of the de- chloritization process. PART II CLAY MINERALOGY A . LITERATURE REVIEW 1. Interlayered Soil Clays The occurrence of interlayered soil clays is wide- spread. The presence of 14°A clay minerals that were fairly stable to heat treatments was first noted in the 1950's. Rich (1968) reviewed the early literature on the subject and notes in a table the occurrence of interlayers in soils and sediments. Rich also reports that little evidence has been found for Fe hydroxy interlayers since Fe oxides are more stable. Al hydroxy interlayers are the most common, although Al-organic complexes have been found to occur in interlayer positions (Gjems, 1963; Perez- Rodriquez and Wilson, 1969). Rich (1968) reports the following conditions he feels are necessary for the development of interlayers: (1) moderately active weathering to remove Mg, Ca and K cations and furnish Al ions, (2) moderately acid pH about 5.0, (3) organic matter content should be low, and (4) there should be frequent wetting and drying cycles. pHs reported by others in Rich's paper vary from 4.5 to 5.2. Rich speculates that a lower pH may be required for vermi- culite interlayering because of its higher charge. Rich 62 63 concludes, on the evidence presented, that interlayer formation is generally the greatest in the A horizon be- cause of the greater number of wetting and drying cycles. Jackson, gt_al. (1952) first proposed the weather- ing sequence mica F=é illite e=é vermiculite e=é montmoril- lonite which has become widely accepted. The reversibility of the sequence was noted by the ability of the swelling 2:1 layer silicates to preferentially fix potassium. Residual non-exchangeable K is found in mont- morillonite (Mehra and Jackson, 1959) and may be the remnants of unweathered illite cores or as a few unex- panded interstratified layers. Jackson (1963) reviews interlayering in expansible silicates. In those interlayered expanding clays the initial sharpness of the 18°A peak suggests swelling of the interlayer. The 12, l4, l8 and 24-28°A peaks after heating indicated some partial hydroxy interlayering. The broadened 10°A peak after 550° heating is another indicator as well as 14°A enhancement upon K saturation. Jackson concludes that less extensive interlayer building in vermiculite is necessary than in montmoril- lonite for CEC decrease and stability. The smaller size and greater weathering rate of montmorillonites in soils negates the interlayering effects. Vermiculitic inter- grades, on the other hand, can survive in warm and humid 64 weathering regimes, primarily because of their larger size and a lower stage in the weathering sequence. Jack— son concludes that the l8°A intergrade--primarily mont- morillonite--occurs in more alkaline environments where montmorillonite is more stable and the 14°A intergrade clays are found in acid soils where the following reac- tions occur: Mica e===é Vermiculite §===é 14°A intergrade. Sawhney (1960) in studying a soil catena in Connecticut formed from micaceous schist found a stable 14°A clay mineral with a behavior between chlorite and vermiculite. He found that the stability of interlayered clay to sodium citrate and heat treatments increased toward the surface. CEC was increased 30 to 40 percent after an interlayer removal treatment with sodium citrate. Both vermiculite and montmorillonite were chloritized. The hydroxy interlayers were more stable in well-drained than in poorly-drained soils. Sawhney considers chloriti- zation to be a continuous series with various stabilities and quantities of interlayer material. With his particular interlayer clay Sawhney found that K saturation and 100° heating had no effect on the intensity of the 14°A spacing and only after heating to 550°C did the 14°A decrease to 10°A in-a broad band. He also found that the aluminum in the interlayer material was not exchangeable by the usual CEC measurements. 65 Quigley and Martin (1963) found chloritization effects to 55 inches in a New England till. They also felt that in their situation more Fe than Al was involved in the interlayer material. Weed and Nelson (1962) found chlorite-intergrade minerals to be a common constituent of clays in various geomorphic regions of North Carolina. The authors sug- gested that the intergrade clays were derived from micaceous parent materials. The quantity and stability of the intergrade clays were greatest in the surface. Rich and Cook (1963) also concluded that the weathering sequence of illite to vermiculite was a common process in acid soils. They considered the mechanism one of exchange of potassium by hydrated cations accompanied by lattice expansion. Their study used 5-2 micron K micas where Na+ had originally substituted for some of the K. The authors also took natural soil clays having considerable chloritized vermiculite having a low CEC (30 meq/lOOg) and by treating with sodium citrate raised the CEC to over 100 meg/100 g. They also felt that the pro- cess of interlayer filling could be reversed by organic matter when abundant. Douglas (1965) studied the clay mineralogy of a Sassafras soil in New Jersey. He concluded that the weathering sequence illite -——e»expanding vermiculite ————+ aluminum interlayered vermiculite accounted for the primary 66 clay mineral alterations found in that particular soil. The alteration series was considered to be continuous. Frink (1965) considered that the Al interlayers existed as atolls, thus effectively blocking CEC sites. Other Al interlayer material existed as pillars holding the 14°A spacing until quite severe heat treatments (inter- layer dehydroxylation) caused partial collapse. A l N sodium citrate treatment for two hours broke down part of the atoll interlayer and allowed cations, expecially K, to enter and be fixed. Frink found that the average relative 14°A intensity of K saturated air dry samples to be related to the length of time of citrate extraction, also that citrate extraction tended to produce a broad band between 10°A and 14°A. Removal of interlayer material enhanced the 1st order diffraction while weakening the 2nd order. Frink assumed that the conversion of illite to chloritized vermiculite in acid soils was a measure of the degree of weathering and that the most highly weathered soils contained the most chloritized vermiculite regardless of surface age. Frink (1969) studied the chemical and mineralogical characteristics of eutrOphic lake sediments in an acid upland soil watershed. He found among the chemical pro- cesses acting in the neutral lake bottom sediments that dechloritization of chloritized vermiculite occurred form- ing a high charge vermiculite that was rapidly converted 67 to illite. This was because of the inherent instability of the vermiculite and abundance of potassium. Frink also found that even though the pH of the acid upland soils had been increased from 4.8 to 6.3 by liming the Al interlayer- ing persisted. On the other hand, the dechloritized ver- miculite in the lake bottom sediments contained very little interlayer material. Frink's study has been the only field evidence of dechloritization occurring under relatively high pH conditions. Malcolm, Nettleton and McCracken (1969) have re- ported on natural dechloritization, but under acid pH conditions in the southeast coastal plain soils of the United States. They found montmorillonite to be the dominant clay mineral in surface horizons of some soils while subsurface horizons contained 2:1-2:2 intergrade minerals. They found a set of conditions where this situation occurred: low pH (below 4-5) and high organic matter content. Normally intergrade clay minerals are dominant in the surface horizons of soils in the southeast United States, organic matter contents are low and pH is higher (5-6). The authors theorized that the combined effects of low pH and high organic matter in the surface horizons were responsible for the dechloritization. Montmorillonite was found in the surface horizons of soils regardless of 68 drainage if the horizon was low enough in pH and high enough in organic matter. 2. Laboratory Studies Since natural soil clay systems are so complex and interlayered clays comprise only a part of the system it is very difficult to study naturally interlayered clays. Formation of artificial interlayered clays in the laboratory is one approach in an attempt to understand at least the process of formation of interlayered clays in natural soil systems. Research is also going on in removing natural as well as artificial interlayers. Interest first focused on interlayers in clays when cleanup procedures were used to prepare clays for x-ray identification in the late 1940's and early 1950's (MacEwan, 1950: Pearson and Ensminger, 1949). Rich (1960) conducted some of the first studies using specimen clay minerals and artificially introduced aluminum interlayers. The vermiculite used was triocta- hedral and he states that its properties were probably somewhat different than dioctahedral vermiculites weathered from micas that are commonly found in soils. Rich found that a stable pH of 5.7 for the Al ver— rniculite suspension indicated little hydrolysis. The interesting results he obtained by adding NaOH to this 69 stable suspension and the resulting decrease in pH from 10 down to 7 shed some light on the method by which inter- layer Al was removed from the clay. The CEC also increased as the pH decreased from 10 down to 7. Rich's concern was that KOH or NaOH might fix Al but the data showed that interlayer Al was removed by these treatments. The mechanism for interlayer removal proposed was the forma- tion of aluminate (Al(OH4)) anions at sufficiently high pH. Limited amounts of NAOH had little effect since the OH ions were repulsed by the clay. A certain threshold quantity of OH ions were required before the aluminate formation reaction would proceed. Rich (1968) discussed methods of introducing arti- ficial interlayers into 2:1 silicates. He cites the factors most responsible effecting interlayer formation and stability were OH/cation ratio and pH. Carstea, Harward and Knox (1970) were able to interlayer vermiculite so that a 13.4°A spacing was main- tained at 300°C, a reaction that took one year. During that time the suspension pH decreased from 4.6 to 3.9. Carstea, gt_al., found that interlayers formed at neutral or alkaline pH were not as well developed or stable. Montmorillonite interlayered with aluminum was stable and was essentially chlorite after ten days, while .Al interlayered vermiculite showed partial collapse when K saturated and heated after the same time period. 70 Acid conditions were found to be best for aluminum interlayer formation in vermiculite and montmorillonite, while iron interlayer formation in vermiculite was favored by alkaline conditions. Iron interlayer formation in montmorillonite, on the other hand, was favored under an acid environment. The approach used purified known clays and known environmental conditions under which interlayers were formed. It also pointed out some significant differences between the two specimen clay minerals. Carstea, Harward and Knox.(l970a)also experimented with dissolution of interlayers artificially introduced into montmorillonite and vermiculite. The treatments were of two types; a sequential series of increasing harshness: 2% boiling Na CO 2 3 Buffered sodium citrate-dithionite Boiling in 0.5 N NaOH and HCl treatment Cold shaker for 20 minutes in 0.1N HCl and a repeat of above. Results on the original standard clay minerals showed a slight increase of CEC with all treatments in the sequential dissolution series. Both minerals had lower CEC after the HCl treatments. 71 The conclusions drawn were that hydroxy Al inter- layers formed under acid conditions were more stable for both minerals. Aluminum interlayer in vermiculite could be re- moved by either the sequential or HCl treatments, while in montmorillonite the interlayer removal was much less. Dewan and Rich (1970) titrated acid soils contain- ing varied cation exchange minerals. They obtained generally consistent results when the aluminate reaction was considered when titrating with strong bases. The re- action is as follows: 3Na + Al-X $=$ Na3 - X + Al+3(exchange) (1) +3 Al + 3(OH) e=é A1(OH)3 (neutralization) (2) A1(OH)3 + NaOH EFé NaA1(OH)4 (3) (aluminate formation) DeWan and Rich concluded that in their strongly acid soils that there was no exchangeable hydrogen except for the "cat" clay. Weak acid sites on organic matter in the soils were occupied by hydroxy Al groups or by diffi- cultly exchanged A1. Relatively consistent results were obtained if a base such as Ba(OH)2 was used and correc- tions made for the aluminate reaction. 72 3. Effect of Clay Cleanup Procedures in, Changing Soil.Clay Mineralogy Alkaline solutions for the solubilization and re- moval of Al and silica have been used extensively in clay cleanup procedures. Hashimoto and Jackson (1960) used boiling O.5N NaOH for this purpose. Mehra and Jackson (1960) also advocated the use of the buffered sodium dithionite-bicarbonate-citrate system for clay cleanup. However, Sawhney (1960) and Harward, §t_al, (1962) showed that these treatments often modified the properties espe- cially of soil clay minerals, vermiculite, montmorillonite and intergrades. They specifically suggested the develop- ment of techniques that would permit the identification of clay minerals and that would reflect the properties of soil clay minerals as they exist in the soil. Dudas and Harward (1971) evaluated alkaline dissolu- tion methods for removal of amorphous components and iron removal treatments as to their effect on amorphous- crystalline component clay systems. They found that NaOH treatments resulted in a larger clay weight loss than KOH treatment, but CEC of kaolinite, halloysite, interstrati- fied mica-chlorite and chlorite did not significantly change while the CEC of biotite increased and nontronite decreased with either treatment but there was less change with KOH than NaOH. They also tested sodium citrate- dithionite-bicarbonate extraction (SCDB) in conjunction 73 with NaOH and KOH. The combination NaOH-SCDB extracted more aluminum and more iron than the KOH-SCDB combination. An experiment was conducted using acid ammonium oxalate as a dissolution agent. The experiment revealed that the acid ammonium oxalate was as efficient as SCDB in extract- ing amorphous materials. The author recommended a KOH- acid ammonium oxalate dissolution method when alteration of crystalline components was to be minimized. Perez-Rodriquez and Wilson (1969) also experimented with pretreatment techniques on an interlayered soil clay. Their particular soil clay was interlayered with an organo- metallo complex. They found that any pretreatment tried altered the clay in different ways. They concluded that in their case no pretreatments could be done and still have the same material that existed in the natural setting. Since vermiculite content is important in this study, the subject of vermiculite determination requires an in-depth look. Alexiades and Jackson (1965) first evolved a laboratory procedure for the determination of vermiculite in soils and sediments. The authors conclude that the procedure works since they obtained nearly the same amount of vermiculite from the summation of other clay components as from the CEC determination. Page, gt_§1. (1967) studied the fixation of potassium and ammo- nium by vermiculite soils. They found that the soils they worked with had a definite capacity to fix potassium or 74 ammonium in nearly equal amounts. They also found that fixation could occur in excess of the specimen clay CEC if excess K or NH4 salts were present during oven drying, thus indicating that a fixation process was occurring that was not associated with CBC. However, the soil vermicu- lite clays showed no excess fixation. The authors con- cluded that for certain soils CEC reductions were a reliable means of determining the amounts of NH4 or K fixed by an ion exchange process. Rhoades (1967) studied cation exchange reactions of soil and specimen vermiculites. He found that the soil vermiculites differed considerably from specimen vermiculites in ion exchange and K-sorption properties. On the other hand, there was no essential difference between soil and specimen montmorillonites. The two soil vermiculites had CEC values calculated to be 114 and 94. These are considerably lower than the value of 159 given by Alexiades and Jackson (1965) or of 175 given by McNeal and Sansoterra (1964) to be representative of the soil vermiculites that they used. Rhoades (1967) concluded that it would appear necessary to use a CEC value closer to those observed for soil vermiculites. Rhoades concludes that his data should apply to most ver- miculitic soils in California, but that soil vermiculites in other areas might not behave in the same manner. Post and White (1967) studied a soil clay which contained two components--mica and vermiculite--that were inherited or 75 weathered from mica. The soil vermiculite showed variable charge density. The upper horizons of the soil contained considerable hydroxy interlayered vermiculite. The weathering sequence postulated is mica + high charge vermiculite + low charge vermiculite + kaolinite. The last step was felt to be problematical by the authors. McNeal (1968) discusses the limitations of quanti- tative clay mineralogy. Multi-component soil clays are especially difficult to deal with quantitatively. McNeal feels that the Alexiades and Jackson procedure (1965) for vermiculite determination appeared to be on sounder foot- ing than a single CEC determination. McNeal concludes that variations in vermiculite CEC values of from 155 to 175 meg/100g would produce no significant difference in vermiculite content. Results calculated to the nearest 1 percent using average parameters should not be relied upon too strongly by researchers. McNeal feels that a variation of i 5 percent in the estimation of components is a valid figure to strive for. Harward and Thiesen (1962) discuss the pitfalls of clay mineral identification by x-ray diffraction techni- ques. They conclude that pretreatments can definitely alter the soil clay mineralogy and that an intensive effort is needed to put soil clay mineral identification on a better foundation. 76 Grim (1968) in the second edition of Clay Mineralogy stresses the importance of a multi-faceted attack on the quantitative identification of clay minerals by using different techniques to verify the validity of the conclusions reached. 4. Deuteration Studies Farmer and Russell (1964) reported some deutera- tion studies done in an effort to assign bands to specific structural hydroxyl groups. They consider that in the perphyllite structure each pair of aluminum ions shares two hydroxyl groups that are related by a center of symmetry. This can give rise to two vibration frequencies: 180° out of phase and symmetric. The antisymmetric vibration, infrared active, absorbs at 3675 cm-1. Muscovite and beidellite, where there is increasing Al for Si substitution causes a shift in frequency to lower values with two bands at 3660 and 3627 cm-1. The lower frequency band increased in intensity with increasing Al substitution and the higher frequency component decreased. Farmer and Russell also point out that the A1 ions in the gibbsite-like sheets in kaolinite are related by a three-fold axis that passes through the A1 ions to give three patterns of vibration; a 3697 cm.1 band due to symmetrical vibration of the gibbsite-like sheet; 3669 and 77 3652 cm-1 to symmetrical vibrations of the inplane vibra- 1 band is due tions with lifted degeneracy and a 3620 cm— to hydroxyl groups on tetrahedral sheets. In the musco- vite structure the 3697 cm"1 band does not occur because the gibbsite sheet is sandwiched between two tetrahedral sheets and there is incomplete symmetry due to substitu- tion. The 3669 and 3652 cm.1 vibrations are either infra- red inactive or remain degenerate, while the 3620 cm-1 band is due to OH vibrations in the octahedral sheet. The fact that the 3620 cm.1 band is of lower intensity in many dioctahedral clays is due to the fact that the SiO network deviates from ideal hexagonal symmetry. The Si-O stretching vibration region (1150-960 cm-l) showed no change upon deuteration. Farmer and Russell also ascribe the 910-915 cm-1 band to OH bending vibrations where the vibrations arise from gibbsite layers where the hydroxyl groups move in phase in a direction perpendicular or nearly so to the layers. Russell, gt_al. (1970) deuterated clays for three one hour treatments at 350°C and produced nearly complete deuteration of saponite, hectorite, vermiculite and mont- morillonite. They were most interested in the infrared region between 200-1200 cm-l. The following shift was noted in montmorillonites upon deuteration from 910-915 cm'1 + 687-693 cm"1 which is the result of Al-AlOH bending 78 vibrations. Beidellites, also with AlAl-OH bands, showed a shift from 935 + 705 cm-1. This illustrates a differing symmetry of the octahedral layers of these two minerals. Vedder and Wilkins (1969) deuterated muscovite and biotite. They found decreases in the 3628, 925 and 405 cm-1 bands and subsequent appearance of 2678 and 705 cm-1 bands in the muscovite. Wada (1967) studied hydroxyl groups in expanded 1 band to inner kaolinites and he ascribed the 3620 cm— surface OH groups of the octahedral layer. His studies on deuteration of montmorillonite showed that with his method of deuteration very few, and probably only exposed OHs on broken edges, were deuterated. He also found a D20 band due to entrapment on the inner layer areas. Vedder (1964) correlated infra-red spectra with compositional variations in mica. He reported that the OH stretching occurring at 3628 cm.1 was the result of OH ions adjacent to unoccupied octahedral sites. Subse- quent work in a review by White (1971) showed that this 3628 band occurs where two Al cations occur in adjacent positions in the octahedral layer (AlAl-OH) type of arrangement. B. METHODS OF ANALYSIS 1. Cleannp and Dispersion Soil samples were at first allowed to air dry, until work showed that in some irreversible collapse of the vermiculite occurred. Later they were treated while still moist or kept in plastic bags to prevent drying. Carbonates were removed, if necessary, with pH 5.0 Na acetate. The soil was treated with 30 percent H202 until frothing ceased. The sample color lightened considerably and the reddish iron color was more apparent after H202 treatment. One iron extraction procedure followed (Mehra and Jackson, 1960) the H202 treatment. The cleaned-up sodium saturated soil was dispersed in distilled water and the clay was siphoned off. In order to reduce the water content excess NaCl was added to flocculate the clay. After the excess water was removed, the clay was washed once or twice with distilled water and then stored in a distilled water-methanol solution until needed. Later in the course of the work pH 9.5 Na CO3 was used for disper- 2 sion, but was discontinued since it was observed to cause flocculation of the less stable intergrade clays. (Near the close of the research, ultra-sonic dispersion was tried. The procedure and results are noted elsewhere.) 79 80 2. X-Ray Diffraction The sodium saturated dispersed clay was deposited on a porous ceramic plate by a vacuum technique. After deposition the film was washed with 0.1N MgCl in 10 per- 2 cent glycerol and finally washed free of excess Mg with 10 percent glycerol water solution. The film was allowed to air dry, then x-rayed (Mg-gly AD). After the initial x-ray the film was washed with INKCl, washed free of excess salt, allowed to air dry then x-rayed (KAD). Heat treat- ments of 110°C for one hour (K 300°C for two hours 110)' and 550°C for two hours (K550) followed with (K300). x-raying after each treatment. Where the intermediate heat treatments produced little change in the diffraction pattern they are omitted from the diffraction tracings shown. In essence then, one clay film is used for all the diffraction patterns. This cuts down on the number of variables when evaluating diffraction tracings. Prob- lems were encountered with lifting and cracking of the films of the intergrade clays when K saturated and heated. 3. CEC Determination The method of Alexiades and Jackson (1965) was used with some modifications. The procedure used here was as follows: from 50 to 100 mg of clay in suspension was put in a 15 m1 plastic centrifuge tube. The clay was 81 washed five times with lNCaCl2 and centrifuged at 5000 rpm in a high-speed Sorval centrifuge. After five CaCl2 washings the clay was washed once with distilled water, then five times with 99 percent methanol. Then the clay was washed five times with 5 ml of 1N MgCl2 and 5 ml of H20 and the washings saved in a 100 ml volumetric flask. After these five washings the clay was washed once with distilled water then washed five times in 1N KCl, one time with H20 and five times with 99 percent methanol. The first two methanol washings were done at 5,000 rpm, the third at 7,000 rpm, the fourth at 10,000 rpm and the fifth at 12-15,000 rpm. This was necessary to keep the washings clear. The clay was then dried for a minimum of twelve hours at 110°C in a forced air oven. After drying the clay was washed five times with IN NH4C1 and the washings saved in 100 ml volumetric flasks. The volumetric flasks were filled to the mark with H20. Ex- cess NH4C1 was washed out once with H20 and four times with methanol. Higher speeds were required again at this point. The clay was transferred to dried weighed glass weighing bottles, dried at 110°C for a minimum of twelve hours, and then the oven dry clay weight determined. Calcium and potassium were determined by a Beckman flame photometer at no higher concentrations than 10 ppm. At these low concentrations of standard solutions, dilutions were needed for all samples. The dilutions reduced the 82 salt concentrations to acceptable levels and prevented atomizer plug-up. The Mg level in the calcium determina- tion was 0.05N and the NH Cl concentration in the potassium 4 determination was 0.1N. 4. Total Potassium. Approximately 0.1 g of air dry Na saturated clay was placed in a weighed oven dry platinum crucible. The clay was dried overnight and weighed to obtain an oven dry clay weight. 1 ml of 1:5 HZSO was added, then about 5 ml 4 of hydrofluoric acid. The crucible was heated in a sand bed until the HF was driven off, then after cooling 5 m1 more HF was added and the crucible heated until the H2804 fumed away leaving a dry residue. After cooling the residue was taken up in IN HNO3 in the crucible. The mix- ture was brought to a boil to loosen spatter on the lid and sides. Then the solution was transferred to a 50 ml beaker and taken to complete dryness. The residue was taken up in 0.1N HCl and transferred to a volumetric flask. Potassium was determined in a 0-10 ppm range. 5. Interlayer Removal.Experiment,. Materials and.Methods The Oshtemo soil clay, sample 35-lU2 from the Al horizon was used in the interlayer stability experiments. One soil sample finer than 50 mesh was stirred in a Waring 83 blender for 5 minutes, the suspended silt and clay poured off into a beaker and subjected to ultrasonic vibration for five minutes. After this treatment the sample was suspended in distilled water and the clay sized soil particles drawn off. This sample, 35-1aU2, was treated only enough to put the clay into suspension. Another sam— ple, 35-lUZ, was put through the standard pretreatment procedures: H202, sodium citrate-dithionite-bicarbonate (SCDB) iron removal and suspended in a dilute NaCl solu- tion, the clay drawn off, flocculated by additional NaCl and then washed twice with distilled water which was enough to deflocculate the clay. In preliminary experiments sample 24-1Ul was used, also a very stable intergrade, but no more could be found in the immediate area that the original sample came from, an indication of variability in the field within a ten meter square areal Sample 24-1Ul was put through the H202, SCDB, iron removal, but was suspended in pH9, 2 percent Na2CO3. While in suspension the clay began to flocculate by itself. A pH check revealed the pH to have dropped back to about 5. After this Na2C03 used as a suspending medium for clay separation. was no longer A prochlorite from Wards Scientific, prepared as described later, was used as a basis of comparison in the interlayer removal experiments. 84 The clay suspensions were put in a 1000 ml beaker on a stirring magnet and pH electrodes inserted, and the top sealed with parafilm to slow down evaporation. Dis- tilled water was added periodically to replace water loss due to evaporation. The pH meter, a Beckman Zeromatic, was hooked to a Sargent SR recorder and the pH monitored continuously for most of the period of the experiment. Heat was applied by a hotplate for those experiments which called for it. The clay suspension was stirred at a constant rate. Dilute NaOH was added to raise the pH to 10. At suitable intervals after the pH had dropped, 2 subsamples were withdrawn, and thepH raised back to 10 with dilute NaOH. One subsample was extracted by BaCl2 for aluminum con- tent. Aluminum was determined by an aluminon colormetric procedure. The other subsample was used for a CEC deter- mination. Each subsample was subdivided so that replica- tion could be made. 6. Deuteration Experiment, Materials and Methods? A pure pale green sample of Wards Scientific pro- chlorite was dry ground in a ball mill to a size range of silt to fine clay and then stored in distilled H20. The best natural intergrade clay came from the Al horizon of an Oshtemo loamy sand in the study area. 85 Sample 35-1U2 was the <20 fraction after H202 and (SCDB) Fe removal (Mehra and Jackson, 1960), cleanup procedures. Sample 35-laU2 was the <2u fraction of uncleaned clay brought into suspension by ultrasonics and then stored in distilled H20. Sample 35-1U2 was Na+ saturated while Sample 35-laU2 was saturated by H+ and A1+3 ions (the "natural" state). A natural soil clay after H202 but not iron removal treatments that was found to contain 58 percent vermiculite by CEC determination (Alexiades and Jackson, 1965) (Sample HVT2). Since these clays, except prochlorite, are highly prone to shrinking and cracking, free standing films could not be used. Thin 1-2 mg films of clay were deposited on Irtran disks. After deposition and drying the film was scanned initially with a Beckman I-R 7 SpectrOphotometer, then the disk was placed in a calorimeter bomb, 1 ml of D 0 added, the bomb sealed and placed in an oven overnight 2 (16 hours). Temperatures of 68°, 100°, 150°, 200°C were employed. One film was used for each subsequent heat treatment. Excessive dripping ruined some films but no difference was noted if the same film was used for each heating or a different film was used each time in the relative amount of deuteration obtained since the same quantity of clay was deposited each time. 86 In another experiment prochlorite and 35-laU2 clay were deposited on Irtran disks. The disks were placed over D20 at 20°C for one week. No deuteration was noted with this experiment. C. RESULTS AND DISCUSSION 1. Clay Mineralogy Differences in clay mineralogy between upland levels U1 and U2 and the terrace levels T1, T2, T3, T4 and B are quite evident both qualitatively and quantita- tively (Part I). There are also striking differences between terrace levels T1 and T2. Figure 12 shows the two extreme cases of differing mineralogy. Sample 35-1U2 is some 12,000 years older than sample HVT2. Sample 35-1U2 has reached an equillibrium condition, with stable chloritized vermiculite, kaolinite and quartz the three major components. Mica probably exists only in inner protected areas of flakes. The CEC data shows very few expanding minerals to be present. Sample HVT2 shows, on the other hand, a large quantity of the expanding minerals that partially expand, a mica component plus quartz, and a similar quantity of kaolinite. The assumption is that the terrace soil clay HVT2 has been derived from upland soil clays similar to 35-lU2. If the dechloritization process does occur then the ex- panding components in HVT2 can be derived from the upland chloritized vermiculite. A further assumption made is 87 88 n no ID V o m e —v In can 3 3 33 3 5 3 r: E §§§ 2.3.? a. .. .. K550 K300 35-I U2 KIIo C"/IIIg K/NH4 xv 24 2| 2 I‘NIK KAI) Mo'qlv AD :550 ’°° HV T2 KIIo CO/Mg K/NH4 xv K HS 28 58 ‘° 2.5xK "9'90 AD I II n. I I I I I II I I I I Figure 12. x-ray, CEC, and total K data for Samples 35-1U2 and HVT2. 89 that in the dechloritization process the dechloritizing vermiculite loses small edge particles which gives rise to a high charge aluminous smectite-beidellite or alumi- nous montmorillonite. Evidence will be given shortly for this. Of course, a smectite and vermiculite component is derived from clayey upland soils and may not have gone through any chloritization-dechloritization process. Evidence presented in Part I shows that very little high charge soil clays are derived as such from the upper water- shed of the St. Joseph River so this inherited component is of small amount. The addition of smectites to land surfaces by dust must not be ignored. The smectites in the dust would be rapidly chloritized in the acid upland soils and when eroded into a river floodplain would be dechloritized. This could be another explanation for the concentration of smectites in Levels T2, T3 and T4. In a further comparison of the two diffractograms (Figure 12) the kaolinite and quartz peak heights are similar for both samples, a further indication of similar soil clay ancestry, 3S-lU2 being a common type of precur- sor of the high charge terrace soil clays. One is also given to notice what happens to the kaolinite peak heights for the two samples as the treatments progress. The kaolinite peak height continues to decrease in HVT2 but remains nearly the same in 35-1U2. Also, the double kaolinite peak in 35-1U2 compared to HVT2 is interesting. 90 The chloritized vermiculite has a 14.24 Angstrom spacing while the vermiculite component in HVT2 has a second order spacing that more closely coincides with the first order kaolinite spacing. The gibbsite type interlayer in the chloritized vermiculite causes a slight expansion. Thus, the double 7.13, 7.07A° peak. In order to check the stability of the vermiculi- tic clay component of sample HVT2 the CEC was determined again after a three year period of storage. The original sample in review, was treated with hydrogen peroxide and dispersed ultrasonically and then stored in distilled water. In the intervening three years there was no dis- cernable bacterial or algal growths. The CEC results are as follows: CazMg KZNH: $2 1968 115 27 58 1971 100 28 46 After the 1971 CEC determination a small sample of HVT2 was cleaned up further by one SCDB iron removal treatment. This treatment allowed for better alignment of clay par- ticles which subsequently enhanced the x-ray peak heights. The x-ray diffraction pattern in Figure 12 is of the clay after a complete cleanup. In order to investigate the clay mineralogy of the terrace soils further, sample 2-1T2 was selected. 91 The vermiculite component of this sample, while not as high as in sample HVT2, was representative of geomorphic Level T2. The clay mineralogy of 2-1T2 (Figure 13) is similar to HVT2 (Figure 12) except for higher kaolinite and quartz components. The probable reason for this is that sample 2-1 consisted of somewhat coarser sediments than sample HVT2; the other difference between these two samples is the pH. Sample 2-lT2 is slightly higher and has been leached more than sample HVT2. Consequently, the vermiculite in 2-1T2 has been partially rechloritized. This is observed by comparing the 14°A component in the Mg-gly and KAD treatments. For the two clays the 14°A component shows a slight degree more stability in 2—1T2 than in HVT2. How much rechloritization is necessary to drop the Ca/Mg CEC from 115 to 85? To give a better idea of the effect of pH on clays sample 2-laT2 (x-ray not shown) still higher but on the same terrace and ten feet away from 2—1T2 was compared with sample 2-1T2. Both samples are the same age, texture, and have similar soil profiles. Sample Ca/Mg K/NH4 %V pH(H20) 2-1 T2 85 35 32 6.2 2-la T2 63 38 16 6.0 Data such as this illustrates the dynamic character of the dechloritization-rechloritization process in these 92 - 5.00 07 I 3 40.04 -I 4.20 -l 755 n "s: n =5 2 3.4 v v 9 II ' ' FINE CLAY 00mg “mu. xv I06 72 22 uxx UEDIUU CLAY (Zn/“9 K/NH4 xv ' 1 as 5: 22 2.21m COARSE CLAY 00mg KINH. xv 44 IS I8 3.2 xx 2-l T2 WHOLE CLAY “mo “ma. xv as 35 32 2.8%K X— ray, CEC and total K data for Sample 2-1T2 and its fractionated components Figure 13. 93 particular soils in relation to slight change in pH. In other soils and conditions the process may behave much differently. Sample 2-lT2 was fractionated into three compo- nents: coarse, 2 to 0.2u; medium, 0.2 to 0.08u; and fine, <0.08U clay by equivalent Stokes diameters. The procedure (Tanner and Jackson, 1948; Jackson, 1956) con- sisted of an initial separation of the 2 to 0.2u fraction in a high speed Sorval centrifuge with a tube technique and the fine clay separated using a Sharples continuous feed super centrefuge. The fine clay contained in a large volume of pH 9.5 NaZCO was flocculated by the 3 dropwise addition of HCl. The fine clay was washed im- mediately after pouring off the excess solution. All fractions were stored in distilled water until needed. 060 spacings were determined by orienting a clay film at right angle to the beam and sweeping through 58 to 60° 26. Table 7 lists the data on the fractionated samples. Figure 13 shows the x-ray diffraction data. The vermiculite content summation of the fractions does not agree with the total clay. Suspending all three fractions in Na2C03 probably has caused the change. The mineralogy of the fractions is decidedly dif- ferent. The coarse fractions contain most of the quartz, kaolinite and considerable mica. The medium fraction 94 gzoum OH H.H om.a mm ma moa mqoz Izoaamm mafia om N.m om.H mm Hm mm momma mmuw endows xuma . - o m cm m.m mm.H .wm.w ma ma as Haggucmam ugmmw mmumoo u: m.m I- mm mm mm s. ammo macs: a x m2\m mz\mo Hmsflmfluo Hmuoa msflommm omo >w usmusoo Hoaoo mNHm Mo w Nuumso w omo .Ama musmwmv mpcmcomEoo UmUMGOHuomum mufl cam NBHIN mamfimm Eoum mumo .5 magma 95 contains a quantity of vermiculite and smectites and some mica. The fine fraction consists of what appears to be smectites that are quite highly interlayered. (Does the Alexiades and Jackson procedure (1965) measure vermiculite content? Is this clay a fine vermiculite that expands beyond 14A°? (Or is it an aluminious smectite-beidellite or montmorillonite that fixes K? Or is this fine frac- tion an interlayered system of 14 and 18A° minerals? The K/NH4 CEC shows that there are considerable smectites in the fraction. The question is raised of whether there is a continuous series between vermiculites and smectites in soil clays. Harward, Carstea and Sayegh (1969) con- cluded in their study of specimen vermiculites and smectites that the minerals existed as two distinct p0pu1ations. Infrared analyses on the whole clay and its three fractions (Figure 14) readily showed quartz content dif- ferences between the three fractions. The band at 660 cm"1 is due to absorption by the Irtran substrate disk. The bands at 780 and 800 cm-1 is a quartz doublet. The 910-915 cm_1 band is largely due to dioctahedral layers in clay minerals; Al-AlOH (White, 1971). The broad band between 1000 and 1100 cm.1 is due to Si-O bending vibrations. The broad band centered at 1412 cm.1 is due, in part, to NH4 vibrations. The 1634- l 1640 cm- band is due to the OH bending frequency of 9(5 .muGOGOQEoo @mumcofiuomum mufl was NBHIN mHQEMm mo mammamsm UmumuwcH .va muswflm Tau 23%;: 20 . . . . _ «I _ :30 5013 N» TN =m.0-~ Ed wmmqou :on-~d 2.5 22sz CW 327 g NOISSIWSNVHl % >30 ”.2; cu - III - 3|”- mn- om- on:- o»:- fig: I-“ 00’ ’ 09’ '- 0. l - 00. '- 97 water molecules. The 2375 cm.1 band is due to atmospheric CO adsorption on the fine clay, or machine imbalance. 1 2 The 3360 to 3440 cm- band is due to OH stretching fre- quencies for water molecules. This frequency varies according to the number of water molecules and the thick- ness of the film on the clay surface. The 3620-3625 cm"1 band is due to OH groups in the clay mineral lattice. The exact frequency depends on the type of lattice and spatial arrangement of the OH groups and their relation to other OH groups and other ions present. The 3650 and 3700 cm-1 bands are due to the presence of kaolinite in the sample. The infrared analysis verify the x-ray data on the quartz content of the various fractions. The large water bands of the fine clay fraction show the presence of smectites which can absorb many layers of water. All samples were Na+ saturated. It is possible that the fine clay fraction had excess Na+ which could influence the amount of water present. The presence of beidellite in the fine clay fraction is doubtful since the A1-AlOH banding frequency in specimen beidellite is 935 cm-1 (Russell, §E_§l., 1970). Al-AlOH bending frequencies in montmorillonite, however, do occur at 910 to 920 cm—1 (Russell, gt_§1,, 1970). The infrared data then shows an aluminious montmorillonite to be a major component of the fine clay fraction. White (1971) gives no data 98 for dioctahedral vermiculites. Infrared analysis along with other data is a valuable tool in clay mineralogy. 2. Interlayer Removal Experiments a. Initial Trials In order to test pretreatment effects on clay CEC some samples normally dispersed and flocculated with NaCl were dispersed in pH 9.5 NaZCO3. A highly aluminous in- tergrade soil clay (Sample 24-lU1) within one day started to flocculate by itself. A pH check showed the suspension pH to be down to 5.0-5.5. A large pH change had occurred in a short time period. Due to this effect no more clays were dispersed in pH 9.5 NaZCO3. Sample 24-1Ul (Figure 15a), stable chloritic intergrade clay, showed the strongest effect. Samples of 24-1 were placed in solutions of 10 percent, pH 11, NaOH; 1N, pH 11, NaZCOB; Na citrate for two weeks and stirred occasionally (Table and 1N, pH 7, 8 and Figures 15 d, e f). For this sample the SCDB treat- ments had little effect in removing interlayer material. The only CEC effect noted is the change in the Ca/Mg CEC, which doubled. The K/NH4 CEC being higher than the Ca/Mg CEC is interesting. The hydrated Ca was evidently readily removed by the washings of water and methanol prior to saturation by Mg in the CEC sequence. The high K/NH4 CEC values were noted only in the highly stable " l'Sf v 165.6 § 2.2" 3 3? am: In pIITI M2003 “55° 2-Is.i:IN NOHCITRATE “lug “INN. xv “NO “INN. xv 44 so 9 “30° 35 2| 9 Mm “MI! AD I II I ll Q50 2m.iulox NoOII “300 NO IRON EXTRACTION c“IIIII “INN. °/.v c"I . “INN. xv 53 38 IO I2 8 O “ Mm N“, " ‘ ' - III-9|! Io Ib-é-é-I-é- I'o-I-é-I- . 15d 15a Figure 15. “w III-III A0 52“ ‘Mw “91'! IO 2. 3 ggc ‘7’ 7.: “550 2 IRON EXTRACTIONS “300 0“III. “INN. xv “no ( “w III-III A0 .h “ INITIAL SAMPLE I IRON EXTRACTION c“IIIIII‘INII. xv 27 O Mm .2 A “m l IO r! .'.. I- ' '.- I- 6‘ 15b Effect of different treatments on the statility of the 14A° component of Sample 24-1U1. 100 o OH m m .> m mm mm om Hm wmzxx omo Ha 4m 44 mm 62\mo Aao.vav .ma.oa A«N.vav Ama.oav Aao.oav NS.HH RB.HH Hmsvmlom.oav Hmswmxmo.oav «nm.oa Aom.aav Amm.HHv 2m ooomm .umm M Amm.mav Aao.vav Amm.m v Hmswmxmm.mac Hmsamxma.aav Hmswmlam.oac Amm.mav vo.oa Aom.mav mH.mH «vo.oa Hm.ma Am>.mac 2m oooom .66m x vo.oa .mm.oa vo.HH .nn.HH mm.mH mm.ma «mm.mH mn.ma .Ho.va 2H oooaa .umm m w~.aa no.4H no.4H Ho.va mum “Hm .umm M o m> om oumom m sm.ea am.va 4m.va am.va 6 u mmumwsumm w: Qma musmflm Uma mnsmwm wma ousmwm mma musmwm swam. : a M. Ma 2 a mmH+ o m momz oo.mz mumnuwu usosumone Hmcflmfiuo woa cH Bantam aH .Ama musmfimv HDHIvm mamsmm so muswEumouu HMUHEozo usmHOMMflo How dump mo mumfiasm .m magma 101 chloritized vermiculite intergrade soil clays. In the case of the low Ca/Mg CEC values the hydrated Ca ion is evidently unable to attach to the hydroxy Al saturated surfaces and interlayer areas. The K ion is able to dis- place some of the hydroxy A1 and then because it is not fixed, is displaced in turn by the NH4 ion. Salt entrap- ment was noted only in partially dechloritized clays high in aluminous dioctahedral smectites (Sample 18-lT2 and lS-lB, Figures A1 and 6). CO treated 2 3 samples have the same 4.25 A° quartz peak heights and there- The initial samples and the NaOH and Na fore are comparable in other peak height intensities. The citrate treated sample has more quartz in it than the other three samples. The NaOH treatment was most effective in partially removing interlayer material as shown by compar- ing the 10°A peaks at 550°C. It is interesting to note that the percent vermiculite determined by CEC reduction is the same for all three treatments. In this regard the NaOH and Na2C03 treatments altered the clay the most as seen by the increase in K/NH4 CEC. The citrate treatment showed little increase in the K/NH4 CEC but a threefold increase in Ca/Mg CEC. One wonders what would have happened if the solutions has been changed every day. The NaOH treatment was the most severe of the three and the Na citrate the least severe. In a second experiment the effect of SCDB treatments on the clay lattice alteration of a chloritized vermiculite 102 was evaluated. Sample 24—lUl was used. The experiment con- sisted of determining the CEC and obtaining x-ray diffraction patterns after 0, l, and 2 SCDB treatments. The results are shown in Table 9. Increasing SCDB treatments reduced the heat stability, increased both the Ca/Mg and K/NH4 CEC but did not cause a change in vermiculite content although there was an earlier collapse of the chloritized vermiculite struc- ture. Contrary to reports of others SCDB treatments did not significantly remove interlayer material or cause struc- tural lattice alterations of this particular clay. In contrast to sample 24-1Ul which was highly chloritized another sample, 20-1B (Figure 7) which was only partially chloritized showed a decrease in percent vermicu- lite with increasing numbers of iron removal treatments. Table 10 shows the CEC comparison between these two soil clays. It can thus be concluded that on some soil clays iron removal treatments definitely affect CEC. However, for qualitative clay mineral identification by x-ray diffraction, it is necessary to run soil clays through at least one iron removal treatment in order to obtain good alignment of the clay platelets. After this preliminary work was done a more rigorous experiment with better control of variables was set up. The early experiments exhausted the supply of sample 24-lUl. Additional samples gathered from within a few feet of the original 24-1 site did not contain the stable chloritized vermiculite component. (See sample 24-la, Firuge A2, for comparison.) 103 zoom usmcfieoo« o o o .> x mm on ma ¢MZ\M mm HM ma mz\mo omo mn.ma oa.ma n¢.HH mm.aa om.oa .mus m ao.oa .ma.oa xe~.oa ooomm vo.oa mm.ma .mus N om.~a .ma.ma «Ho.ea oooom vo.oa v~.¢a Ho.va .HM H «mm.ma «mm.mH .mm.ma cooaa .umm M Ho.va Rob.ma vm.va Ho.va hHU Ham .UMm M Ho.va ¢~.4M v~.sa ¢~.4M MH6 new Mamumz mma ,ousmflm .moommz uma musmflm nma musmum an ammumm mam>oamm mm N Hm>08mm mm a Imfia Hm>oamm ,mm oz “m ucofiummne .Ama munmflmv Moauwm mamamm co mucofiummng Hm>oEmH coma mo umbfidg msflmmmnosfl How mumo mo mumfifism .m magma 104 v MN mN N N MN mN H NH ON mm o mHION v wN oN N o @N Ha H 0 ma NH o HDHIvN .>w qmZ\M m2\mu muMMEwMMHB mamfimm .Amaumv musmsummup Hm>oEmH coufl mo HOQESG mcwmmmuocw mg» an wouommmm mm mamao Hwom 03¢ mo omu .OH magma 105 Therefore, in order to obtain a good supply of a stable chloritized vermiculite, an Oshtemo soil was chosen. Sample 35-1U2 (Figure 12), and Sample 24-1Ul (Figure 15b), shows both clays as to their degree of thermal stability. The chloritized vermiculite com- ponent of 35-1U2, although several thousand years younger, is as, or more stable than that in 24-1U1. This leads one to conclude that surface age of a landscape surface does not necessarily mean as much as degree of weathering expressed by the soil profile. Many factors enter into the degree of weathering of one soil on one landscape surface and another soil on an older or younger surface (discussion in Part I). Since pretreatment techniques had altered the clay mineralogy of 24—1Ul a new approach was tried, that of an ultrasonic dispersion procedure (see materials and methods). Samples 35-1U2 and 35-laU2 are the standard pretreatment and ultrasonically treated samples respectively compared with two terrace Level T2 samples, 2—1T2 and 3-1T2, which were treated in the same manner, in order to determine the immediate effects of the new pretreatment procedure. Table 11 shows no significant differences in per- cent V between soil clays dispersed by ultrasonics or by the standard pretreatment procedure of H202, SCDB and dis- persion in dilute NaCl. The higher CEC for the ultrasoni- cally dispersed samples is due to exchange sites on organic 106 Table 11. Effect of standard pretreatment procedures vs. ultrasonic treatment on the CEC of soil clays and vermiculite content. CEC ' K/NH4 ACEC %V. Clay No.. Ca/Mg Ultrasonic* 35-laU2 31 27 3 l Cleaned up* 35-1U2 24 20 4 2 Ultrasonic 2-1aT2 62 45 17 ll Cleaned up 2-1T2 63 38 15 10 Ultrasonic 3-laT2 58 43 15 10 Cleaned up 3-1T2 49 31 18 12 *See materials and methods for procedures and CEC analysis procedures. Sample-35-1U2 highly chloritized acid soil clay. Samples 2-1T2 and 3-1T2 from a terrace on the St. Joseph River. These two samples have dechloritized and have partially rechloritized (see Part I). The small "a" denotes the ultrasonic dispersion procedure. matter. The only problem is the poor x-ray diffraction patterns for the ultrasonically dispersed clays (none are shown). As far as infrared analysis is concerned, there is no difference in patterns between the two treatments (data not shown). The ultrasonic vibration kills most of the orga- nisms. But after setting a month algae colonies were beginning to grow in the sample 3-1aT2, which was higher in organic matter than the other two samples. 107 The ultrasonic dispersion technique is fast and appears to leave the soil clays in the most "natural" con- dition. The dispersion time for soil clays from other areas would need to be worked out in order to keep from changing clay particle sizes or otherwise affecting the clay mineralogy. An alternative procedure would be to ultrasoni- cally disperse the sample and then treat it with peroxide to remove most of the organic matter. This might be enough to produce good x-ray diffraction patterns. After these initial experiments on the dynamics and pretreatment effects of the soil clay systems in this study more rigorous experiments were designed. Since sample 35-1U2 was obtained in a large quantity, it was used throughout the rest of the dynamics experiments. Experiment #1 was designed to show the clay dynamic pro- perties of 35-laU2 at room temperature using dilute NaOH to raise the pH to 10. Experiment #2 was the same except the temperature was held constant at 50°C i 1°C, and it covered a longer period of time. Experiment #3 was a control experiment using a specimen prochlorite but under the same conditions as Experiment #2. Experiment #4 dif- fered in that instead of the ultrasonically treated 35-laU2 clay the standard pretreatment treated 35-1U2 clay was used. Experiment #5 used an acid pretreatment technique to illustrate its effects on sample 35-laU2 clay. 108 b . Experiment # 1 A subsample of sample 35-1 was dispersed in dis- tilled water via ultrasonics, the clay separated and stored in distilled water. No peroxide or SCDB treatment was used in an attempt to leave the clay in as natural a state as possible. This ultrasonically dispersed sample is 35-laU2. The clay suspension in a 1000 ml beaker was stirred constantly at a temperature of 20°C. Dilute NaOH was added initially to raise the pH to 10 and additional was added as the pH dropped. Before additional NaOH was added subsamples for aluminum extractable by BaCl2 and CEC analysis were withdrawn. Table 12 lists the results of the experiment and Figure 16 shows the pH, extractable aluminum, and CEC with time. BaCl2 extractable aluminum was not related to change in CEC. This aluminum may have come from hydroxy Al on organic matter exchange sites, rather than from interlayer Al. If aluminate was being formed, it should react with the Ba ion and precipitate to form Ba(AlOH4)2. The drop in CEC with drop in pH is evident in both this experiment and experiment #2 following. This suggests rechloritization of aluminum into interlayers again. Due to the poor diffraction patterns no x-ray tracings are available to show whether the clay was more or less resis- tant to collapse upon being K saturated and subjected to heat treatments after the NaOH treatment. This experiment 109 produced about the same change in Ca/Mg CEC as the soil clay 24-1Ul two weeks in 10 percent NaOH. Table 12. Experiment #1. Data from Figure 16. Sample Time in Meq Al/lOOg CEC % V No. Hours by BaCl2 Ca/Mg K/NH4 ° 0 0 1.0 32 29 l 24 1.0 42 33 2 69 1.0 53 32 14 3 116 0.4 45 30 9 4 142 2.0 49 32 ll Conditions: Initial soil pH, H20, 5.50 Illite content 12% Total K 0.97% Temperature constant at 20°C Sample 35-laU2 c. Experiment #2 The same soil clay (35-laU2) as in Experiment #1 was used except that the temperature was raised to 50°C and held constant. Dilute NaOH was used to raise the pH to about 10.0 and more added at various time periods as the pH dropped. pH was monitored constantly from time period 0 to 204 hours, and sporadically from then until the experiment was terminated. Table 13 shows the results of this experiment while Figure 17 shows the pH, CEC, and extractable A1 data. No attempt was made to measure the 110 f 6' ”PER.” (1175 @ 20°C I B. IO“) .4 .J 94 2 PH 4 d I 82 2 7 3 7.4 .I 1 “ O III-I "‘ 0 6- 601 ,2 3 ‘ I: _ E .. .1 O > ‘ ‘ 33 5.4 40-4 ['20 ‘ 7 — o h— .. o _ 2 —I > I- 20- “ -I0 .4 3‘ a _ o 4 ” Bocl [III-AI. CI U—L-U\D/D 12 _. 0 2'4 4‘8 9‘6 I44 TIME-HOURS Figure 16. Experiment #1. Sample 35-laU2; pH trace, change in CEC, and extractable A1 with time. 111 Table 13. Experiment #2, data from Figure 17. Sample Time Meq A1/1009 CEC % V No. Hours by BaCl2 Ca/Mg K/NH4 ' 0 0 1.0 33 29 3 5 24 2.1 42 32 6 6 72 7.9 44 34 6 7 96 0 48 26 14 8 144 10.0 54 35 12 9 178 9.5 108 28 58 10 204 3.3 151 25 82 11 312 6.4 169 26 93 12 816 - 106 30 48 amount of NaOH added since suspension samples were removed at various intervals. The NaOH to clay suspension ratio changed as clay was withdrawn. No x-ray tracings were obtained due to poor quality resolution of the ultrasoni- cally dispersed sample. The clay suspension appeared to be better buffered as NaOH additions continued. The first washing of BaCl2 to sample 10 brought about immediate flocculation in large floccules. Prior samples showed no evident flocculation upon the first addition of BaClz. This particular soil clay (35-laU2) that is ultrasonically dispersed has a fine 112 «:3 3 2 2 o. o 8 ¢ N 259. o; 2.» 8w 8.~ Nm. 3... 8 3 o I _ I OIIIIIIII‘II‘IIO II \ \ I 03:12:. o.2.\ in \~.\o .\o n on I 9 I N I \ M II/ I \ M I >1. I 8T I l l n I Pu. \m H Eu. \ Ora—l \q \0 l a 03:0 \ Q 82.4 so \ w x \ .mfiflu cuHB H¢ manmuomuuxm paw .omo cw mmcmnu .momuu mm “NDMHImm mamamm .N# ucwEHummxm u .8 Q 2:: 55:42 .ha mnsmflm I In N .I I I I O I o I I o_Io~Im I I II M I V I Mu I I W I I I3; IooIo I I Ice Im II T 2:. %" A n W I. .W n 113 particle size and is relatively non-sticky. Sample 12 had different properties than the earlier samples in this ex- periment. It was the first time that the initial addition of CaCl2 brought about immediate flocculation in large floccules and for the first time the clay was sticky. An immediate conclusion that can be drawn is that there is no relationship between BaCl2 extractable aluminum and CEC or percent vermiculite. Rich (1960) pointed out in artificial interlayer studies that the amount of Al in solution did not represent the total Al exchanged because of hydrolysis and subsequent fixation or precipitation. The pH fluctuation is similar to that reported by Rich (1960) in his artificial hydroxy aluminum interlayer .studies. He added NaOH to pH 10 and after eighteen days the pH had decreased to 7.8 and the clay showed an in- crease in CEC from the initial artificial interlayered sample. Experiment #2 also shows a time lag after NaOH was added before there was a significant increase in Ca/Mg CEC. Rich (1960) concluded that the initial small amounts of NaOH added to Al vermiculites had little ef- fect due to low concentrations of the Na ions and OH ion repulsion by the clay. But, when the pH was raised suf- ficiently high and kept there so that conditions were optimum for aluminate anion formation, or some other type 114 of aluminum complexing reaction, interlayer material was rapidly removed. The high point Ca/Mg CEC of 169 meq/lOO grams is remarkable for a soil clay since the charge is diluted by quartz, kaolinite and other components in the sample. The actual charge on the vermiculite must be considerably higher. The fact that the K/NH4 CEC remained essentially constant throughout the duration of the experiment shows that there was little alteration of the clay structures. The decrease in Ca/Mg CEC in the interval between 13 and 25 days is due to rechloritization of interlayer spaces, a result of the drop in pH. d. Experiment #3 A finely ground prochlorite suspension was treated with NaOH for a one week period in the same manner as Experiment #2. Figure 18 shows the pH variation and CEC at the beginning and end of the experiment. There was no conversion in this sample of prochlorite to vermiculite since the Ca/Mg and K/NH4 CEC's rose by equal amounts. In this experiment the NaOH affected the chlorite struc- ture making more cation exchange sites available. The de- crease in pH is due to Mg ions moving out of the brucite layer and complexing with OH ions. The x-ray diffraction patterns showed no variation in stability or intensity 115 IOIO Ecol/53m 333 .meu spas 0mm CH mocmso paw .momnu mm "mufluoaosooum .mw psmfiflnomxm .mH muswam mmaozlmzz v: mm ac o O p _ p _ _ I n. n L \ i a I M \ \ W \q l. I 4 ¢ \ I h 0&0 ¢\_ \\ a I \ \\ a I I t I __ \ \\ \onoo _ :1 \\\ \ I n_I .\\\ .I0 4 ~.Il I. Au 2L I H 1m f: o .8 © $35.23» 116 before or after the NaOH treatment. Figure 19 shows the diffraction pattern for this mineral. e. Experiment #4 Sample 35-1U2 was used again in this experiment but it had gone through the conventional cleanup procedures (H202 and SCDB). This experiment was designed to show whether cleanup procedures affected the clay properties. The results were surprising (Figure 20). After the pH was raised to 10 with NaOH it remained nearly constant for over a one week period, in great contrast to experiments 1 and 2. Only a slight amount of NaOH was added at the 48 hour mark. The ACEC increased markedly from 3 to 84 meg/100g and in a shorter time than in Experiment #2. The change in CEC indicates that Al complexes were removed from interlayer positions or at least shifted enough to allow for entry and exit of Ca and Mg ions and entry and fixation of K ions. But why no pH change? Is organic matter on the clay surfaces necessary for the aluminate reaction to occur at moderately high pH, or do the highly reduced clay surfaces prevent additional complexing of OH ions by the released Al complexes? Is there some other type of reaction involved in dechloritization? According to Dewan and Rich (1970) the following sequence occurs under pH conditions between 8 and 10: 117 .m* ucmfifluomxm mo soflumamfioo may Hmumm ouwuoasooum paw .¢* uqofifluomxm mo coauoamfioo on» Hmumm NDHImm mamamm mom mump umu ps8 mmnIx .ma musmflm 4W2$ um. 9..U mw ”n K“ II WImzw mam .0 TI 4.Hw mm mum7.4 MW 3H92I cxc OMmrI V. 0E3 I Non/mm M. R r e h, m m CamIAIw m V» ”P .m M .9» K K a” KI. I u“ .M W. mm.» I. ©~.nl railI 5.0.0. I ”fin-.I rr.?I ad.) P W I o 118 '0‘ \l\ 9 -4 d’ TEMPERATURE CONSTANT @ 50°C Bfi :: - c; ._ 7‘ PIZO 4 13 _ 1” l T I O O ~ t. °: fly : CEC "'99 IlOOg clay -¢ / _ 50 .. Kl _ 6/ uuqcec ____..0 ~30 5 l l l I 20 0 48 96 I44 TIME - HOURS Figure 20. Experiment #4. Sample 35-1U2; pH trace and CEC change with time. 119 3 3 Na++Al-x e===§ Na -X + Al+ 3 Al+3 + 3 (OH) e===e A1(0H)3 A1(on)3+ Na(OH) e===é Na Al(OH)4 Since the Na is readily dissociated, the aluminate anion is repelled from the clay surface. In this manner OH ions are taken out of the aqueous suspension. This re- action sequence appears to be what is happening in experiments 1 and 2, but not in experiment 4. The before and after x-ray diffraction tracings for experiment #4 (Figures 12 and 19 respectively) reveal that the soil clay, even after treatment in NaOH one week, was more heat stable than at the beginning even though by CEC analysis the percent vermiculite went from 3 to 53 per- cent. An explanation for this is that the NaOH treat- ment resulted in the partial breakage of the gibbsite outer ring of the interlayer to allow for ion exchange. However, the treatment allowed for a shifting and better orientation of the remaining material, thus allowing for greater stability to heat. I do not believe salt entrap- ment is the cause for the large increase in the CEC since the duplicate samples agree with each other so well (see Table 14). The possibility exists that the Ca/Mg CEC might have risen further if the experiment had lasted longer. 120 Table 14. CEC data from Experiment #4. —_ L Dup. Ca/Mg K/NH4 % V. CEC after one week in NaOH at pH 10 (Figure 19) 1 115 31 54 2 111 30.5 52 CEC before NaOH treatment lFigure 12) l 24 21 3 2 24 20 3 f. Experiment #5 Researchers have reported the use of 0.1N HCl as an interlayer extractant in studies of artificial clay interlayer systems (Carstea, Harward and Knox, 1970) and found that dilute HCl worked in removing artificial inter- layers. In this experiment sample 35-1aU was used without any cleanup procedures and a removal procedure used by Carstea, Harward and Knox, 1970, utilizing 0.1N HCl with a 20 minute shaking period. The results are shown in Table 15. After the second sequential treatment the Ca/Mg CEC decreased by 10 meq while the K/NH4 CEC decreased by 3 121 Table 15. Experiment #5. Acid dissolution of an inter- grade soil clay (Sample 35-1aU2) with 0.1N HCl. - 4‘— L 4— Dup. Ca/Mg K/NH4 CEC at start of experiment I l 30'3 26 Initial 2 30.6 25.6 sample CEC after one 20 minute shaking in 0.05 N HCl l 22 24 II 2 22 24 CEC after second 20 minutejshaking in 0.1N HCl III 2 20.5 22 meq from the initial sample CEC. Evidently CEC sites have been destroyed. The dilute HCl hydrolyzed some of the organic matter because hydrophobic organic substances stuck to the test tube walls and also coated the neck of the volumetric flasks during CEC analysis. In any case, no interlayer material was removed. Some of the clay probably was dissolved, a partial explanation for the de- crease in CEC. Unfortunately, the use of this particular sample prevented x-ray diffraction patterns with enough 122 resolution to determine whether changes had occurred after treatment with HCl. g3 Discussion Every intergrade chloritized vermiculite soil clay that has been reported in the literature has different pro- perties. One feature that all appear to have in common is that they are dioctahedral and aluminum is the main ion in the interlayer position. Other than that they have dif- ferent heat stabilities, CEC in the "natural state," and ease with which different dissolution agents can remove the interlayer material. It must also be emphasized that artificially introduced interlayers in highly crystalline mono-component clays only approximate the dirty organic matter laden multi-component soil clay systems. In the case of the natural system along the St. Joseph River where natural dechloritization was occurring, to attempt in the laboratory to remove the interlayer material from the source clays seemed a desirable thing to do. The natural process could, of course, be only approxi- mated since it may take hundreds or up to a couple of thousand years for the process to act under the conditions given in the geomorphology section (Part I). Another prob- lem arises in that the upland intergrade soil clay has three main components that show by x-ray diffraction; chloritized vermiculite, kaolinite, and quartz. The 123 natural high vermiculite soil has smectites, vermiculite, mica, some feldspar, and quartz components as well as considerably more organic matter and different ions on the exchange surface. Whether under natural dechloriti- zation conditions smectites are produced or if the river carried them in from some areas nearby that have smectites in the clay fraction is not known, but it is suspected that both are happening, or smectites are carried in by dust. Of course dechloritization has not only happened along the St. Joseph River but can happen along any river that has acid upland soils with intergrade clays and a neutral to calcareous floodplain where dechloritization can occur. The use of NaOH to remove interlayers was tried after sodium citrate-bicarbonate—dithionite (SCBD) did not materially remove interlayers. Both the Ca and K CEC rose by the same amount indicating clay alterations produced more exchangeable sites for both ions to be ex? changed in the SCBD system. Rich and Cook (1963) and Sawhney (1960) both used sodium citrate which removed interlayer from the soil clays they worked with. Others (Carstea, Harward and Knox,1970a) have advocated the use of NaOH or KOH dis- solution treatments when their particular clays were not appreciably altered by the treatments. One can only con- clude that with soil clays as with other soil components 124 one must experiment until he comes up with a cleanup system that works well with the clays he is working with and with as little alteration as possible. In this study pH 9.5 Na2C03 could not be used as a dispersing agent since it altered the intergrade clays. The best treatment probably was where all the organic matter was left on the clay and the clay dispersed in distilled water with ultrasonic treatment. But more experimenta- tion is needed for obtaining a suitable ultrasonic treatment procedure. In this case, regardless of the pretreatment, there was no appreciable CEC difference between ultra- sonic and standard pretreatment system (at least for the Oshtemo soil clays, 35-lUl). The only drawback is that x-ray diffraction patterns were very poor due to poor alignment of the clay platelets. During the NaOH treat- ment the organic matter on the clay surfaces appears to have enhanced the rate of formation of the aluminate anion and subsequent reduction of OH ions in solution. The cleaned-up sample, when treated with NaOH changed its Ca/Mg CEC greatly but appeared to be more heat stable. In this case, instead of the aluminate anion forming, the atoll edges were broken enough to allow for the entrance and exit of Ca and Mg and the entrance and fixation of K, but the buildup of hydroxy Al enhanced the ability of the clay to withstand heat treatment and maintain the 14°A 125 spacing. Or, if there was no buildup, the hydroxy Al became better oriented in the interlayer space. In any case, it is possible to produce a high charge multi-component soil clay by NaOH treatment (169 meq/100g clay). As this clay aged the pH continued to drOp, the CEC dropped too, as hydroxy Al was refixed in interlayer positions (Figure 17). It is remarkable that the K/NH4 CEC remained nearly constant throughout the NaOH treatments, an indication of little clay alteration other than interlayer removal. This substantiates that the interlayer material is labile and also very dynamic. BaCl extractable aluminum had no correlation with change 2 in Ca/Mg CEC, but as shown in Figure 17, a definite thres- hold amount of OH ions were needed before the aluminate or other complexing reaction proceeded with considerable rapidity. The cleaned-up sample, 35-1U2, Experiment #4, did not show any extensive aluminate reaction since OH ion concentration (activity) remained nearly constant. 3. Deuteration Experiments In order to further show the dynamics of inter- grade clays, a different approach was tried. If the inter- grade material is labile, then the replacement of the OH groups by OD groups in a deuteration experiment should show whether or not intergrade material OH groups are more accessible than lattice OH groups to replacement. 126 Therefore, three samples were selected for the experiment, a highly intergrade soil clay (35-1U2) a naturally dechlori- tized soil clay (HVT2) and as a standard, a pure prochlorite. (See methods for the manner of preparation and treatment sequence.) If the interlayer material is largely composed of aluminum hydroxide and it can be deuterated, then infrared analysis should detect changes in the 910-915 cm-1 and 3620 cm.1 regions which are due to Al-AlOH bend- ing and stretching frequencies respectively. Figures 21 and 22 show the infrared tracings for the deuteration experiments. The prochlorite (Figure 21) shows no evident deuteration until 200°C when a broad band centered at 2650-2680 cm_1 appears. Sample HVT2, the dechloritized soil clay (Figure 21) shows slight deuteration at 150°C and considerably more at 200°C. Since this clay (HVT2) has lost most of its interlayer material, it should show very little deuteration. It is very similar in degree of deuteration to the prochlorite sample. Sample 35-1U2, the highly intergrade soil clay (Figure 22) shows deutera- tion beginning at 68°C with pronounced deuteration at 100°C and almost complete disappearance of the 3620 cm-1 band and 915 cm‘1 band at 200°C. The shift of the 915 cm—1 upon deuteration falls to the 660-670 cm"1 region which unfortunately is masked by the strong 660 cm-1 irtran absorption band. 127 — 000 - 700 700 - 000 3000 3030 3000 3000 " 3200 - 3000 - 2000 ‘ 2600 r 2600 - 2400 - 2200 ' 2000 - IIOO - I000 - 9|5 - 900 PROCHLORFTE Cm'I Figure 21. Deuteration effects on Sample HVT2 and prochlorite with increasing temperature. 128 .muoumuomfiou mcwmmwuocfl nufiz NDHImm onEmm so pommmm coaumumuswo .mm muomflm zoom 35:. o.mw (N: in 312% o.oo_ coum BEE 129 The deuteration that has occurred in the pro- chlorite at 200°C is probably on broken lattice edges which are most accessible to the exchange, or Brucite OH groups. The dechloritized soil clay HVT2 shows some deuteration effects at 150°C and considerably more by 200°C. It can be concluded that all of the remaining intergrade material is deuterated by this 150°C tem- perature. Then at 200°C many of the lattice hydroxyls have been deuterated. Or, if it can be assumed that most of the interlayer material is gone, then the deu- teration that appears at 150°C can be ascribed to lattice hydroxyls of the octahedral layer. The intergrade clay, 35-lU2, shows deuteration beginning at 68°C, and considerably more at 100°C. This definitely shows the easy accessibility of the interlayer hydroxyls to deuteration. Of interest, too, is the almost complete disappearance of the 915 cm.-1 and 3620 cm.1 bands at 200°C. The 915 cm-1 band is due to Al-AlOH bending vibrations and is shifted to about 600 cm-1, where it is obscured by the irtran absorption band at 660 cm-1. This shows that lattice hydroxyls in this particular clay are more susceptible to deuteration than sample HVT2 lattice hydroxyls. Deuterated samples did not lose any deuteration ions over a two week period that the sample stood in air. As to why 35-lU2 had more deuteration than HVT2 at 150-200°C may be due to a finer 130 particle size or to differences in clay lattice struc- tures to discrete gibbsite in the clay fraction or the presence of amorphous Al(OH)3 on clay surfaces. In order to narrow down the possibilities, a sub-sample of 35-1U2 was cleaned up further with another iron re- moval treatment (Mehra and Jackson, 1960) and then put through the same series of treatments. The infrared tracings showed no difference. The iron removal treat- ment should have removed any gibbsite or amorphous Al(OH)3. Since both 35-laU2 and HVT2 were dispersed untrasonically, particle size should not be a factor. Consequently, deuteration rates of lattice hydroxyls at 150 to 200°C is a matter of differing accessibility of these hydroxyls to diffusion in of D+ ions and out- ward diffusion of H+ ions. Russell and Fraser (1971) discuss the deuteration of lattice OH groups in mont- morillonite. They concluded that there was conclusive evidence for migration and interaction of D+ ions with lattice OH groups. One experiment that should have been done is to use a sample of 35-laU2 from Experiment 2 or 4 that showed a high CEC and deuterate it. How- ever, the CEC analysis exhausted the supply. The deuteration experiment shows that the OH groups of interlayer material are more easily deuterated in the intergrade clay 35-1U2 than either the prochlorite or the high vermiculitic soil clay HVT2. This ease of 131 deuteration is due to the greater accessibility of OH groups in the interlayer material and that less energy is required to replace an H+ ion by a D+ ion in the interlayer clay. The possibility still exists of Al(OH)3 material on outer clay surfaces of the inter— layer clay even after cleanup. 9.}.1. OD correspond to the shift produced in the interlayer soil The theoretical shift of = 1.364 does not clay. Since there is a coupling of adjacent Al ions in the octahedral layer with OH ions, Al-AlOH the result- ing shift is of the order 1.350 to 1.356. The prochlo- rite 3750 + 2650 cm-1 shift is related by a factor of 1.411. The intergrade soil clay shift of 3620 + 2678 is related by a factor of 1.350 to 1.353. Vedder and Wilkins (1969) found that the factor of 1.356 fit their specimen muscovite in deuteration studies. SUMMARY Clay mineralogy is a valuable analytical tool in soil geomorphic studies. Soil clays are highly subject to diagenesis, and therefore readily reveal the natural dynamics and environment in which they occur. Of course other techniques can be used in similar geomorphic studies. Sand grain analysis can, for example, reveal many different kinds of things such as sediment source areas and what has occurred during the transportation of the sand grains. The tool or tools of analysis to be chosen depend on the type of output that is desired. In this study analytical techniques involving clay mineralogy and soil chemistry are well suited to the situation where the dynamics of a system are in- volved over a fairly short time period. That there is a high K-fixing soil clay is shown by both fertility trails and laboratory analysis. Since the presence of vermiculitic soil clays cannot be observed directly in the field, laboratory analysis involving x-ray diffraction and CEC analysis are necessary to show their amounts and extent in the field. Once the conditions that are conducive to the formation of chloritized vermiculite in the upland soils of a 132 133 watershed are known it is possible to predict that the alluvial soils should contain abundant vermiculite in some river basins other than the St. Joseph. The fact that dechloritization is occurring under rather mild pH (7.0-7.5) conditions in Alluvial soils leads one to won- der if the process will not begin eventually in the long limed upland soils that contain abundant chloritized vermiculite and eventually cause the tie—up of all the K fertilizer that is applied. The soil clays containing abundant chloritized vermiculite are highly dynamic is readily shown by the interlayer removal and deuteration studies. The inter- layer removal study showed the importance of organic matter in enabling the dechloritization process to occur at lower pH conditions than if none or very little organ- ic matter is present. Rechloritization can also occur very rapidly, at least in the laboratory in a reaction vessel when the pH was allowed to drop to a low enough level. Rechloritization is shown to occur in the field when leaching of the alluvial soils lowers the pH to a low enough level even in the presence of abundant organ- ic matter. More research is needed in interlayer removal studies in order to reveal more of the mechanisms and conditions responsible for dechloritization to occur. The deuteration study showed how readily accessible the 134 OH ions in the interlayer position are, and also how labile they are. The use of a multi—faceted approach in research lends credence and credibility to the results. A thorough knowledge of soil genesis and morphology, glacial geology and geomorphology, plus laboratory analy— sis techniques, involving clay mineralogy and soil chem- istry enabled the author to pursue this study to the fullest extent within the objectives set forth. Again, as in all research, new and interesting problems arise. Some of these problems have been par- tially explored in this study over and above the original objectives. The interlayer removal and deuteration studies, which came up at about the midpoint of the re- search, would add a great deal to the geomorphology and clay mineralogy section. The author learned at least two things from this research problem, (1) Soil clays are "dirty" to work with, (2) No two soil clays are alike in their component proportions or properties, and soil clays with their associated organic matter and amorphous materials need not, and usually do not, behave as the clean organic matter free single component specimen clay minerals do. I do not deride the efforts of those who experi- ment with simple clean systems. Their work is necessary, for we must first understand simple systems before the 135 science can progress into impure natural soil clay sys- tems. The time is approaching when techniques must be devised to work with soil clays in as natural a state as possible. CONCLUSIONS 1. Many acid soils contain a quantity of chlori- tized vermiculite. The chloritized vermiculite varies in stability from one soil to another. 2. Natural dechloritization is shown to occur in a suitable environment, resulting in a high charge vermi- culitic soil clay that can fix large quantities of K+ and NH: ions from fertilizers. 3. It has been demonstrated in the laboratory that intergrade soil clays can be readily dechloritized and converted to high charge vermiculitic soil clays and that rechloritization can occur rapidly. 4. The role of organic matter is important in natural reactions of chloritization, dechloritization and subsequent rechloritization in both the field and labora- tory. 5. The deuteration experiment demonstrated the dynamic character of a soil clay composed of a large quantity of chloritized vermiculite compared to a standard sample and a natural dechloritized vermiculitic soil clay. 6. The combination of field observations and laboratory analysis is important in showing that no "arti- ficial" soil clays were produced due to laboratory 136 137 manipulations, and that the laboratory data is useful in making proper interpretations of conditions in the field. Both disciplines are necessary and a requisite for mean- ingful applied research. RESEARCH NEEDS Chloritized vermiculite is widespread in well- drained leached soils. However, there appears to be a continuous variation in the quantity and stability of the interlayer material from one soil to another. Chemi- cal extractants do different things to different soil clays. The researcher has an obligation to determine the effects of cleanup procedures on his particular soil clays and then use the one that causes the least amount of change from the 'natural' soil system or that best characterize the 'natural' soil system. Although it has been reported that there are discrete boundaries that separate specimen vermiculites from smectites, there is no good evidence to show the same in soil vermiculites and smectites. This brings up the point that although the use of specimen clays in the laboratory is important to gain a basic understand- ing of processes, one must not extrapolate too far to the natural soil system. The role of organic matter, especially as metallo-organo complexes, in natural chloritized vermiculites requires further study. More research on dechloritization processes is needed. Will the continued application of lime to acid 138 139 outwash soils eventually result in the tieup of K ferti- lizers? Soil columns could be leached with Ca saturated water and undergo wetting and drying cycles at optimum field temperatures. A variety of soil clays need to be evaluated in this manner. BIBLIOGRAPHY BIBLIOGRAPHY Ahlrichs, J. L. 1968. Hydroxyl stretching frequencies of synthetic Ni, A1, and Mg hydroxy interlayers in expanding clays. Clays and Clay Minerals, 19:63-71. Alexiades, C. A., and M. L. Jackson. 1965. Quantita- tive determination of vermiculite in soils. Soil Sci. Soc. Amer., 29:522-527. American Society for Testing and Materials. 1945. Alphabetical index of x-ray diffraction pat- terns. Am. Soc. Testing Mater. Philadelphia. Brewer, Roy. 1964. Fabric and Mineral Analysis of Soils. John Wiley and Sons. New York. Carstea, D. D.; M. E. Harward; and E. G. Knox. 1970. Comparison of iron and aluminum interlayers in montmorillonite and vermiculite: I. Formation. Socil Sci. Soc. Amer. 34:517-521. . 1970a. Comparison of iron and aluminum inter- layers in montmorillonite and vermiculite; II. Dissolution. Soil Sci. Soc. Amer. Proc. 34: 522-526. Dewan, D. D., and C. I. Rich. 1970. Titration of acid soils. Soil Sci. Soc. Amer. Proc. 34:38-44. Dixon, J. B., and M. L. Jackson. 1962. Properties of intergradient chlorite-expansible layer silicates of soils. Soil Sci. Soc. Amer. Proc. 26:358-362. Douglas, Lowell A. 1965. Clay mineralogy of a Sassafras soil in New Jersey. Soil Sci. Soc. Amer. Proc. 29:163-167. Dudas, M. J., and M. E. Harward. 1971. Effect of Dis- solution treatment on standard and soil clays. Soil Sci. Soc. Amer. Proc. 35:134-140. 140 141 Farmer, V. C., and J. D. Russell. 1964. The infrared spectra of layer silicates. Spectrochim. Acta. 20:1149-1173. and J. L. Alrichs. 1968. Characterization of clay minerals by infrared spectrosc0py. Trans. 9th. Int. Congr. Soil Sci., Adelaide 3:101-110. Flint, R. F. 1957. Glacial and Pleistocene Geology. John Wiley and Sons, New York. Frink, C. R. 1965. Characterization of Aluminum inter- layers in soil clays. Soil Sci. Soc. Amer. Proc. 29:379-382. . 1969. Chemical and mineralogical characteris- tics of autrOphic lake sediments. Soil Sci. Soc. Amer. Proc. 33:369-372. Gjems, O. 1963. A swelling dioctahedral clay mineral of a vermiculite-smectite type in the weathering horizons of Podzols. Clay Minerals Bulletin, Vol. 5, No. 29, pp. 183-194. Grim, R. E. 1968. Clay Mineralogy. 2nd ed. McGraw- Hill publ. Harward, M. E., and A. A. Theisen. 1962. Problems in clay mineral identification. Soil Sci. Soc. Amer. Proc. 26:335-341. and D. D. Evans. 1962. Effect of iron removal and dispersion methods on clay mineral indenti- fication by x-ray diffraction. Soil Sci. Soc. Amer. Proc. 26:535-541. Harward, M. E.; D. D. Carstea; and A. H. Sayegh. 1969. Properties of vermiculites and smectites: ex- pansion and collapse. Clays and Clay Minerals 16:437-447. Hashimoto, I., and M. L. Jackson. 1960. Rapid dissolu- tion of allophane and kaolinite-halloysite after dehydration. Clays and Clay Minerals 7:102-113. Hough, J. L. 1958. Geology of the Great Lakes. Illi- nois Press. Urbana. 142 Jackson, M. L.; Y. Hseung; R. B. Corey; E. J. Evans, and R. C. Vanden Heuvel. 1952. Weathering se- quence of clay size minerals in soils and sedi- ments. II. Chemical weathering of layer silicates. Soil Sci. Soc. Amer. Proc. 16:3-6. Jackson, M. L. 1956. Soil Chemical Analysis-Advanced Course. Published by the author, Department of Soils, University of Wisconsin. Madison, Wis. . 1963. Interlayering of expansible layer silicates in soils by chemical weathering. Clays and Clay Minerals. 11:29-46. Leopold, Luna B.; M. Gordon Wolman; and John P. Miller. 1964. Fluvial Processes in Geomorphology. W. H. Freeman and Co., San Francisco. 522 pages. Leverett, Frank. 1989. The Illinois Ice Lobe. U.S.G.S. Monograph No. 38. , and F. B. Taylor. 1915. Pleistocene of Indiana and Michigan and the History of the Great Lakes. U.S.G.S. Monograph No. 53. Mac Ewen, D. M. C. 1950. Some notes on the recording and interpretation of x-ray diagrams of soil clays. J. Soil Sci. 1:90-103. Malcolm, R. L.; w. D. Nettleton; and R. J. McCracken. 1969. Pedogenic formation of montmorillonite from a 2:1-2:2 intergrade clay mineral. Clays and Clay Minerals, 16:405-414. Mehra, O. P., and M. L. Jackson. 1959. Constancy of the sum of mica unit cell potassium surface and interlayer sorption surfaces in vermiculite- illite clays. Soil Sci. Soc. Amer. Proc. 23: 101-105. . 1960. Iron oxide removal from soils and clays by a citrate-dithionite system buffered with sodium bicarbonate. 7th. Natl. Conf. on Clays and Clay Minerals, pp. 317-327. 143 McNeal, B. L., and T. Sansoterra. 1964. Mineralogical examination of arid land soils. Soil Sci. 97: 367-375. McNeal, B. L. 1968. Limitations of quantitative soil clay mineralogy. Soil Sci. Soc. Amer. Proc. 32:119-121. Page, A. L.; W. D. Bunge; T. J. Ganje; and M. J. Garber. 1967. Potassium and ammonium fixation by vermi- culitic soils. Soil Sci. Soc. Amer. Proc. 31: 337-341. Pearson, R. W., and L. E. Ensminger. 1949. Types of clay minerals in Alabama soils. Soil Sci. Soc. Amer. Proc. 13:153-156. Perez-Rodriguez, J. L., and M. J. Wilson. 1969. Effects of pretreatment on a 14°A swelling mineral from Gartly, Aberdeenshire. Clay Minerals Bulletin. 8:39-46. Post, Donald R., and Joe L. White. 1967. Clay mineralogy and mica-vermiculite layer charge density in the Switzerland soils of Indiana. Soil Sci. Soc. Amer. Proc. 31:419-424. Quigley, F. M., and R. T. Martin. 1963. Chloritized weathering products of a New England glacial till. Clays and Clay Minerals. 10:96-106. Rhoades, J. D. 1967. Cation exchange reactions of soil and speciman vermiculites. Soil Sci. Soc. Amer. Proc. 31:361-365. Rich, C. I. 1960. Aluminum in interlayers of vermicu- lite. Soil Sci. Soc. Amer. Proc. 24:26-32. , and M. G. Cook. 1963. Formation of dioctahed- ral vermiculite in Virginia soils. Clays and Clay Minerals. 10:96-106. Rich, C. I. 1968. Hydroxy interlayers in expansible layer silicates. Clays and Clay Minerals. 16:15-30. Russell, J. D.; V. C. Farmer; and B. Vede. 1970. Re- placement of OH by OD in layer silicates and identification of the vibrations of these groups in infrared spectra. Mineralogical Magazine. 37:869-879. 143 McNeal, B. L., and T. Sansoterra. 1964. Mineralogical examination of arid land soils. Soil Sci. 97: 367-375. McNeal, B. L. 1968. Limitations of quantitative soil clay mineralogy. Soil Sci. Soc. Amer. Proc. 32:119-121. Page, A. L.; W. D. Bunge; T. J. Ganje; and M. J. Garber. 1967. Potassium and ammonium fixation by vermi- culitic soils. Soil Sci. Soc. Amer. Proc. 31: 337-341. Pearson, R. W., and L. E. Ensminger. 1949. Types of clay minerals in Alabama soils. Soil Sci. Soc. Amer. Proc. 13:153-156. Perez-Rodriguez, J. L., and M. J. Wilson. 1969. Effects of pretreatment on a 14°A swelling mineral from Gartly, Aberdeenshire. Clay Minerals Bulletin. 8:39-46. Post, Donald R., and Joe L. White. 1967. Clay mineralogy and mica-vermiculite layer charge density in the Switzerland soils of Indiana. Soil Sci. Soc. Amer. Proc. 31:419-424. Quigley, F. M., and R. T. Martin. 1963. Chloritized weathering products of a New England glacial till. Clays and Clay Minerals. 10:96-106. Rhoades, J. D. 1967. Cation exchange reactions of soil and speciman vermiculites. Soil Sci. Soc. Amer. Proc. 31:361-365. Rich, C. I. 1960. Aluminum in interlayers of vermicu- lite. Soil Sci. Soc. Amer. Proc. 24:26-32. , and M. G. Cook. 1963. Formation of dioctahed- ral vermiculite in Virginia soils. Clays and Clay Minerals. 10:96-106. Rich, C. I. 1968. Hydroxy interlayers in expansible layer silicates. Clays and Clay Minerals. 16:15-30. Russell, J. D.; V. C. Farmer; and B. Vede. 1970. Re- placement of OH by OD in layer silicates and identification of the vibrations of these groups in infrared spectra. Mineralogical Magazine. 37:869-879. 144 Russell, J. D., and A. R. Fraser. 1971. I. R. spectro- scopic evidence for interaction between hydronium ions and lattice OH groups in montmorillonite. Clays and Clay Minerals. 19:55-66. Sawhney, B. L. 1960. Weathering and aluminum interlayers in a soil catena: Hollis-Charlton-Sutton- Leicester. Soil Sci. Soc. Amer. Proc. 24:221-226. Soil Taxonomy of the National COOperative Soil Survey. 1970. Soil Conservation Service, United States Department of Agriculture, Washington, D. C. Steel, Robert G. D., and James H. Torrie. 1960. Princi- ples and Procedures of Statistics. McGraw-Hill Co., New York. 481 pages. Tanner, C. B., and M. L. Jackson. 1948. Nomographs of sedimentation times for soil particles under gravity or centrifugal acceleration. Soil Sci. Soc. Amer. Proc. 12:60-65. United States Department of Interior. U.S. Geologic Survey, Washington, D.C. Topo maps of the Niles and Benton Harbor Quadrangles. 1927. Vedder, W. 1964. Correlations between infrared spectrum and chemical composition of mica. Amer. Mineral. 49:736-768. , and R. W. T. Wilkins. 1969. Dehydroxylation and rehydroxylation, oxidation and reduction of micas. Amer. Mineral. 54:482-507. Wada, Koji. 1967. A study of hydroxyl groups in kaolin minerals utilizing selactive deuteration and infrared spectrosc0py. Clay Minerals. 7:51-63. Wayne, William J., and James H. Zumberge. 1965. Pleisto- cene geology of Indiana and Michigan. In, The Quarternary of the United States. Edited by H. E. Wright, Jr., and David G. Frey. Princeton Uni- versity Press, Princeton. pp. 63-84. Weed, S. B., and L. A. Nelson. 1962. Occurrance of chlorite-like intergrade clay minerals in coastal plain, piedmont, and mountain soils of North Carolina. Soil Sci. Soc. Amer. Proc. 26:393- 398. 145 White, Joe L. 1971. Interpretation of infrared spectra of soil minerals. Soil Science. 112:22-31. Zumberge, J. H., and J. E. Potzer. 1956. Late Wiscon- sin Chronology of the Lake Michigan Basin correlated with pollen studies. Bull. Geol. Soc. Amer. 57:271-288. APPENDIX 146 .mnvN paw man—”Iva .vBHIvH .vBHIm .mBHIv mmHmEdm How mummy M .2509 65m. .Umo .hmulx 1nd munmflm 2 h.9 .a 5.3 3. a 9 fi 3. .55. .28 «2:2 :9... :3 3 3 >4. £3. .23.... «tum. 923m. wmodzohkm 02.:- “ c J . . . nu.“ _ 8 3 5 £3. .28 v... T! :3 2 a. z 3. £3. .28 :7» a... 1‘: sin.“ 3 o. no 3. £3.35 n» Tv 0303' out on- m m m on- out n:- ”II; “'0" "U- 0”. - no In . h: 0 n 8 :2 : :§ 3s 5 g;; II I I _ - 7 t v _ Km _ - «m PROFILE 24 UI Mm 24-M M ‘ ho “mg “mu. xv 44 36 5 “no - 53-4 24-2 32:: 00/049 K/NH4 xv fig. 58 4| ll zosx '“1?” 24-3 33 “lug “mu. xv 47 30 u 121x Figure A2. X-ray, CBC, and total K data for Profile 24U1: Samples 24-1aU1, 24-2 and 24-3. 148 .NDHTHN pom NoalmN .HDHIMN moamfimm How mumu M acne“ can .Umo .hmnlx . mm wusmfim :3 :2 :3 o 8 on 3 t. e a R 3, £5 2:5 >4. £3. 2:8 3 £22 2:8 a< ¢u>om la cautzmo s< oo~4. :2; 2:5 >153. 2:8 n m 51...? «um mu: 9:. mun a: on: on!“ 3 6...... a-.. . :n :2 a mm as - 3 I; .53. 58 . 0 NV 2.. _ q o «m 2 >4. IZ\x 230 x l p; ~-m m . 03x2 , I I. s: . x83 rogue: runs. n 3 mm 1.: 3 £22 £5 2. an 8. 2 3 3, £53. 8:8 3 I. 3 3. N: t 378% 2; onnx .. NS 3:2: M." 4 51 D 19‘ 59' 0f . . _ 7w W .m U c. a n m I w 150 $15? é a? I! .. .. 1‘550 , 32-Iu2 " K300 " . ‘ Caxug K/NH4 %v 24 20 3 1 “MI! A0 - k “550 .. Km Kno ‘ 3l-IB C0/Mg K/NH4 xv " 34 32 l " "9‘qu A0 In... 3 II I... I u: 1 Figure A5. X-ray and CEC data from Samples 31-1B and 32-1U2. O (D Q _ m '0 I I ‘20 .3. In Figure A6. 151 -W '0 NM v v m 't 0! P“ <2 <2“: <2 °4 C” l l I I u T “:I- I , K550 l-lTl C“/Mg K/NH4 xv .311 137' () KM, l.6%K Mo-qw 40 - N K550 . ,. SF-4 ‘° ‘° K300 33-I Tl SF-B ng K/NH4 xv 20 IS 3 ”Km 22%K ” Km ° Mq-qw A0 X-ray, CEC, and total K data for Samples 1-1Tl and 33-lT1. 152 .NBHIom wcm maalmm .NBHImN mmamfimm How mumc ONO flaw amulx .bé munmflm gas“ 3 a 8 a 9. t 2 a a a. £2... .38 a. £2... 2...... a. £2... .28 3 To... 2 I... w. 78 0 QC .3. 1 . . .. 3. .4 , 0 a 2... . 2.... .. _ 2... .. - - _. . - . 8mg . I 6 u w .8 a s... .. .. _ .. . - .. a... .. .. Wmmm mm. m “mm. mm WW... W ”U- ml 001! 900 {II 1.01 001; - N" - 031 - m {fl 9" . J .6 O 153 owmoE .pmxfia .mEmoHlmmumoo .maaosvmammm oflucmovm>sam nmuoonoo mauom owmoE .poxfis .aEmoaumsHm .maaosqmammm oaucoavm>sHm cmon mHImH owmms .pmxfla .mEmolech “announoouusm caucmsvcw>sam me meanma Gamma .Umxfle .Hmumflwxm mpcmm Ho quMm Ho>o hemoaumcaw .mwamnsammm oamha oonmEmme «Danna Gamma .poxfls .hpcmm umHHosvmammm Daucm>5Hm NBHImH Gamma .pmxwfi .mucmsmeEmmm UAHHOZ muopcmaw malma canoe .onwe .mucmsvaEMmm owaaoz muoocoaw «Balva OHmmE .pmxHE .mamoalocflm .maaosvammm oaucmsvm>sam cmoam mauma oammE .pmxHE .wfimoHlmmHmoo .maaOdqsammm oaucmdqm>5Hm oomwumo mHImH oamms .UmxHE .hEmoHImmHmoo .mHHosqummm oeucmsqm>sam cmuoonou mmalaa owmoE .pmxHE .wsmoHlomnmoo .maaosvammm oaucmdwm>sam cmuoonou mHIHH OHmmE .Uflomcoc .UmxflE .mEmoHlmcHw .mucosqm>sHm owumd mamonm manm canoe .pmxHE “musmsvmaammm owaaoz muoccmao vamaum oame .tmxfle .mEmoalmCHm .mummngoouusm oflucm>5Hm mommmsmo madam canoe .UmxHE .mucmsmeEMmm oaaaoz myopcmaw wBMaun oamoE .Umxfle .hEmoalmch “monounoouusm caucmswm>sam How «Balm oawma .waflfi .hEmoalmch .maaosvmammm oeucmsvm>sam cmon vaalm oamofi .pflomcos .pmxHE .mamoHlmawm “musmdvm>dam Gamma mamonm mealv oawmfi .UmxHE .hEmoaumcwm “mummunoouusm oaucmdmm>sam How meanm canoe .pmxfia .mpcmm .maaopdammm caucm>zam Hosfism NBHIN oamofi .pmxfle .mwcmm .mHHOUSHmmm owuco>oam umcadm NB>m Gamma .Umxfla .%WGMm .maa0pdammm caucmEEmmm mxcflmm HBHIH coflumofimammmHu moauwm.__ mamamm. .hpsum wasp ca pmHQEMm mums umsn maflom mo soflumoflmwmmmao wabmnoum .H¢ magma 154 .Aoan. mm>udm HHom o>wumumm0Ioo Hmcowumz may no afioaoxwa .n..nom..F owmofi .pmst .mamoHlmmumoo .mmattdammm Danae UHmmE .pmxwfi .mfimoalpsflm .mummunoouusm owucmsmm>sHm owmoE .pmxwfi .mEmoalmmHmoo .mmamosammm owmma pwomco: .owmmE .pwxfifi .afimoHlmch .musmanHmmm owmaa Gamma .poxwa .mEmoHIosHm .maaosvmammm oaucmowm>dam Demos .vmxfla .mamoalmmumoo .maaopSHmmm causm>dam owmofi .pmxwa .wsmoHlmsflm “mummunoouusm oeucmsvm>sam Gamma .pmxfla .hEMOflnmsfim «mummunoonasm owucmsvm>sam canoe .pmxwfi .MpGMm .mHH0©5Hmmm oeuso>sam canoe .pmxHE .mamoHloswm .mmampsHmmm oaaaoz oamofi .pmxHE .mEmoalmch .mmampsammm UHHHOS Gamma .UmxHE .Hmumaoxm MUGMm Ho hpcmm um>o mfimoalmcfim .mmamvsHmmm Gamma canoe .Umxflfi .mamoHnmmHmoo .mMHMUSHmmm aflmha owmms .pmxwa .wamoHlomumoo .mmampsHmmm owmaa 0&9...me Hmm oEOuSmo zmcmDSmmz cmoam mmpcmn Hmm Hum Hmcfinm xoflwmq -x0flomn Ho moaxoo oonEmme oewunmo “whom NDHImm NBHIwm HBHImm NDHINM mHIHm NBHIom «BHImN NBHImN NBHImN HDMHIvN HDHIvN HDHIMN NDHINN mDHIHN