1

Naturally Occurring Microbially-Induced Smectite-to-illite Reaction 1

Jinwook Kim1*, Hailiang Dong2,3,*, Kiho Yang1, Hanbeom Park1, W. Crawford 2

Elliott4, Arthur Spivack5, Tae-hee Koo1, Gilyoung Kim6, Yuki Morono7, Susann 3

Henkel8, Fumio Inagaki9, Qiang Zeng2, Tatsuhiko Hoshino7 and Verena B. Heuer10 4

1Department of Earth System Sciences, Yonsei University, Seoul 03722, Korea. 5

2State Key Laboratory of Biogeology and Environmental Geology, China University of 6

Geosciences, Beijing 100083, China. 7

3Department of Geology and Environmental Earth Science, Miami University, Oxford, 8

OH 45056, USA. 9

4Department of Geosciences, Georgia State University, Atlanta, GA 30302-3965, USA. 10

5Graduate School of Oceanography, University of Rhode Island, RI 02882, USA. 11

6Petroleum & Marine Research Division, KIGAM, Daejeon 34132, Korea. 12

7Kochi Institute for Core Sample Research, JAMSTEC, Kochi 783-8502, Japan. 13

8AWI Helmholtz Centre for Polar and Marine Research, Am Handelshafen 12, 27570 14

Bremerhaven, Germany. 15

9Research and Development Center for Ocean Drilling Science, JAMSTEC, Yokohama 16

236-0001, Japan.17

10MARUM Center for Marine Environmental Sciences, University of Bremen, 28359 18

Bremen, Germany. 19

*To whom correspondence should be addressed to jinwook@yonsei.ac.kr and20

dongh@cugb.edu.cn 21

22

GSA Data Repository 2019200

2

MATERIAL AND METHODS 23

Sampling Location 24

Core samples were drilled by the International Ocean Discovery Program (IODP) 25

through a water depth of 4,776 m into the sediment/basement interface at 1,180 meters 26

below seafloor (mbsf). Expedition 370 of the drilling vessel Chikyu (Temperature Limit 27

of the Deep Biosphere off Muroto, 10. Sept. – 23. Nov. 2016) established Site C0023 at 28

32⁰22.0018’ N and 134⁰57.9844’ E. The Nankai Trough is located at the active 29

subduction zone between the Philippine Plate and the Eurasian plate, east of Japan 30

between the Shikoku Basin and the Southwest Japan arc (Moore and Karig, 1976). The 31

subduction rate of the Shikoku Basin is ~2-4 cm/year. A thick sediment pile 32

accumulated on the young (~16Ma) basaltic basement (Seno et al., 1993). At Site 33

C0023, the sediment column consisted of varying lithologies (Unit II-V) (Heuer et al., 34

2017): II) trench-style deposition (silt and sand (IIA), terrigenous turbidites (IIB), 35

hemipelagic mud (IIC); 0-494 mbsf); III) basin style deposition (upper Shikoku Basin 36

facies including tuff and tuffaceous sedimentary rock; 494-635 mbsf); IV) 37

volcaniclastic-bearing mudstone (lower Shikoku Basin facies; 635-1,112 mbsf); and V) 38

acidic volcaniclastic-bearing mudstone (mudstones and felsic ash; 1,112-1,125mbsf) 39

(Fig.1). Details on site summary and drilling procedure should be referred to IODP Exp. 40

370 proceeding (Heuer et al., 2017). 41

Clay mineral preparation 42

A total of 49 clay samples (253.39-1,176.52 mbsf) were prepared following the 43

Jackson’s procedure (Jackson, 1969). Some volcanic ash layers were identified in the 44

core. For consistency in the analysis of S-I reaction, these layers were not sampled and 45

analysed. Core sediments (13 g) were placed in a 250 ml polypropylene copolymer 46

3

(PPC) bottle (Nalgene, USA) with 10 ml of 30% hydrogen peroxide (H2O2) and sodium 47

acetate (NaOAc) -acetic acid (HOAc) buffer solution (pH 5.0) for 24 hours to disperse 48

clay minerals by removing organic matters. Samples were washed with distilled water 49

and then centrifuged using a Labogene 1236R centrifuge at 4,000 rpm for 30 minutes to 50

separate supernatant. Clay size fraction smaller than 2µm was collected through 51

centrifugation at 1,000 rpm for 2 minutes. 52

X-ray diffractometry 53

From the sediment cores (189 to 1,180 mbsf), clay minerals less than 2 µm in size 54

were separated from subsamples every ~50 m. Clay samples were applied to a slide 55

glass using a filter-peel method with a 0.45 µm membrane filter paper (Moore and 56

Reynolds, 1989). Samples were solvated in ethylene glycol in a desiccator overnight. 57

These oriented clay mounts were scanned over angles from 3º to 15º 2θat a speed of 58

1.5º/min, a step sizes of 0.02º with a receiving slit size of 0.3 mm and a divergence slit 59

size of 1.25º using Rigaku MiniFlex II with CuKα radiation source at 30 kV and 15 60

mA. Crystallographical Search-Match software (version 2.0.3.1) was used to identify 61

the clay mineralogy, and semi-quantitative analysis was followed (Biscaye, 1965; 62

McCarty et al., 2009) to estimate the relative proportions of smectite and illite. X-ray 63

diffraction (XRD) peaks occurring at 5.09o and 8.82o 2θ correspond to smectite and 64

illite respectively. The peaks at 9.84o and 16o 2θ correspond to illite/smectite mixed 65

layers (I/S). The positions of the I/S diffraction peaks were used to quantify the relative 66

abundances of smectite and illite (Fig. 1). Semi-quantification of illite and smectite (Fig. 67

1) was performed with both Biscaye (Biscaye, 1965) and Sybilla methods (McCarty et 68

al., 2009) showing compatible results (Supplementary DR3). 69

Kinetic modelling of smectite to illite transformation 70

4

The mathematical model of the smectite to illite process was done utilizing a kinetic 71

expression for the loss of smectite. The loss of smectite was set equal to a gain in illite. 72

The loss of smectite was calculated as the product of an Arrhenius-type rate constant for 73

illitization, the concentration of smectite, and concentration of K+ in solution. The 74

illitization rate constant was calculated as the product of a pre-exponential rate constant 75

(sec-1), activation energy, Ea (kjoule/mole), and absolute temperature as shown below 76

(Altaner, 1989; Huang et al., 1993; Pytte and Reynolds, 1989; Pytte, 1982; Velde and 77

Vasseur, 1992). These models were found most applicable to certain geologic settings 78

such as burial diagenesis at high geothermal gradients such as the Salton Sea(Huang et 79

al., 1993), Tertiary and Paleozoic foreland basins (Altaner, 1989; Elliott et al., 1991); 80

Paris Basin (Velde and Vasseur, 1992); and Gulf Coast (Elliott and Matisoff, 1996; 81

Huang et al., 1993). The expressions for loss of smectite and illitization rate constants 82

are given below. 83

𝛿𝛿 𝑆𝑆𝑆𝑆𝛿𝛿 𝑡𝑡

= −𝜅𝜅 𝑆𝑆 𝛼𝛼[𝐾𝐾+]𝛽𝛽 1 84

Where S is the concentration of smectite, α and β are dimensionless order parameters, κ 85

is the illitization rate constant defined below. 86

κ = A exp (-Ea/RT) 2 87

Where A is a pre exponential constant (sec-1). R is the gas constant. T is absolute 88

temperature. Ea is the activation energy. The order of the reaction is the sum of α and β. 89

In this study, we employed the kinetic parameters developed for burial at high 90

geothermal gradients (Huang et al., 1993). A is 80800 sec-1, α is equal to 2. β is equal to 91

1. Ea is equal to 117.5 kJ/mole. 92

A published temperature-depth model (Heuer et al., 2017) was used for geothermal 93

gradient. The amount of illite was formed at 0.1 Ma time steps by employing a kinetic 94

5

model that successfully simulated the conversion of smectite to illite using geothermal 95

gradients (Elliott and Matisoff, 1996; Huang et al., 1993). Weight percent of illite 96

formed is calculated below at a given time step, temperature and K+ where the K+ in 97

solution was converted to wt.% K2O, and the amount of smectite S at the i time step and 98

the previous i-1 time steps. The quantity (1-Si-1/9.13) was the amount of smectite 99

removed written in terms of the amount of illite formed, where 9.13 wt.% was the 100

maximum K2O in muscovite or illite (Elliott and Matisoff, 1996). 101

𝑆𝑆𝑆𝑆𝑆𝑆 𝑜𝑜𝑜𝑜 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑜𝑜𝑜𝑜𝑓𝑓𝑆𝑆𝑖𝑖𝑓𝑓 𝑎𝑎𝑖𝑖 𝑖𝑖 𝑖𝑖𝑖𝑖𝑆𝑆𝑖𝑖 = −𝜅𝜅 (1 − 𝑆𝑆𝑆𝑆−19.13

) 𝛼𝛼[𝐾𝐾+]𝛽𝛽 + S i-1 3 102

The calculated amount of illite was sensitive to temperature and K+ in solution. A 103

maximum value for K+ in solution was used (9.344 mmole K+ /L or 0.044 wt.% K2O) 104

given the measured K+ in the pore solutions. 105

TEM, EELS, and SEM 106

Variations in structure and elemental composition in smectite and illite were 107

measured on the lattice fringe images, selected-area electron diffraction (SAED) 108

patterns, and Energy Dispersive X-ray Spectroscopy (EDXS) analyses. TEM samples 109

were impregnated with L.R. White resin (Kim et al., 1995) to prevent structural collapse 110

of smectite layers resulting in a layer spacing identical to that of illite, caused by the 111

dehydration of clay minerals in the high TEM vacuum. The cured samples were sliced 112

to 70 nm in thickness using an ultra microtome (ULTRACUT TCT; Leica) and then 113

placed on a holey carbon TEM Cu-grid. The lattice fringe images were then used to 114

directly differentiate the I/S mixed layers (~23 Å) from the smectite (12 Å) and illite 115

layers (10 Å) with EDXS. Illite polytypes and illite packet-size distribution were 116

measured on SAED patterns and illite lattice images, respectively (Inoue et al., 1988). 117

The quantification of Fe redox-states in smectite and illite structure was determined 118

6

using a chemical shift of Fe L2 and L3 edges, and integral intensity ratio of Fe-L2,3 edges 119

(Fe-L3 at 709 eV and Fe-L2 at 722 eV) for three phases (Van Aken et al., 1998). The 120

operational conditions for EELS acquisition included an entrance aperture of 2.0 mm, 121

energy dispersion of 0.1 eV/channel, and by setting the full width at half maximum 122

(FWHM) to 1.0 eV for the zero-loss peak calibration. Fe L-edge spectra were acquired 123

with an acquisition time of 5.0 seconds in Scanning Transmission Electron Microscope 124

(STEM) mode. The statistical optimum signal-window parameters for the integral ratio 125

of Fe-L2,3 edges were calculated using the Gatan Digital Micrograph software (Kim and 126

Dong, 2011). In particular, the background intensities were removed from EELS spectra 127

through the application of the standard power law and double arctan functions. A 128

window δ (30 eV) was selected in the background to extrapolate the curve over the post-129

edge window and the electron intensity under the extrapolated line was subtracted from 130

the total intensity (Yang et al., 2016). The Fe oxidation states of the samples were 131

determined using a universal curve as a function of the integral ratio of Fe-L3/L2 and the 132

total ferric and ferrous Fe concentration (Van Aken et al., 1998). Alteration of Fe-133

sulfides with elemental composition were examined by SEM. Samples were prepared on 134

the adhesive carbon tape then coated with Au in vacuum coater for 60 seconds prior to 135

SEM observation. 136

Geochemical Analysis 137

Concentrations of K, Si, Fe2+, SO42- and CH4in pore water were measured on board 138

during the Expedition. Details concerning sample treatment and analyses can be found 139

in the IODP Exp. 370 proceeding (Heuer et al., 2017). 140

141

REFERENCES CITED 142

7

Altaner, S. P., 1989, Calculation of K diffusional rates in bentonite beds: Geochimica et 143

Cosmochimica Acta, v. 53, no. 4, p. 923-931. 144

Biscaye, P. E., 1965, Mineralogy and sedimentation of recent deep-sea clay in the 145

Atlantic Ocean and adjacent seas and oceans: Geological Society of America 146

Bulletin, v. 76, no. 7, p. 803-832. 147

Elliott, W. C., Aronson, J. L., Matisoff, G., and Gautier, D. L., 1991, Kinetics of the 148

smectite to illite transformation in the Denver Basin: clay mineral, K-Ar data, 149

and mathematical model results (1): AAPG Bulletin, v. 75, no. 3, p. 436-462. 150

Elliott, W. C., and Matisoff, G., 1996, Evaluation of kinetic models for the smectite to 151

illite transformation: Clays and Clay Minerals, v. 44, no. 1, p. 77-87. 152

Heuer, V., Inagaki, F., Morono, Y., Kubo, Y., Maeda, L., Bowden, S., Cramm, M., 153

Henkel, S., Hirose, T., Homola, K., Hoshino, T., Ijiri, A., Imachi, H., Kamiya, 154

N., Kaneko, M., Lagostina, L., Manners, H., McClelland, H.-L., Metcalfe, K., 155

Okutsu, N., Pan, D., Raudsepp, M. J., Sauvage, J., Schubotz, F., Spivack, A., 156

Tonai, S., Treude, T., Tsang, M.-Y., Viehweger, B., Wang, D. T., Whitaker, E., 157

Yamamoto, Y., and Yang, K., 2017, Temperature Limit of the Deep Biosphere 158

off Muroto: Proceedings of the International Ocean Discovery Program. 159

Huang, W.-L., Longo, J. M., and Pevear, D. R., 1993, An experimentally derived kinetic 160

model for smectite-to-illite conversion and its use as a geothermometer: Clays 161

and Clay Minerals, v. 41, p. 162-162. 162

Inoue, A., Velde, B., Meunier, A., and Touchard, G., 1988, Mechanism of illite 163

formation during smectite-to-illite conversion in a hydrothermal system: 164

American Mineralogist, v. 73, no. 11-12, p. 1325-1334. 165

8

Jackson, M. L., 1969, Soil Chemical Analysis-Advanced Course: Soil Chemical 166

Analysis-Advanced Course., no. Edition 2. 167

Kim, J.-W., Peacor, D. R., Tessier, D., and Elsass, F., 1995, A technique for 168

maintaining texture and permanent expansion of smectite interlayers for TEM 169

observations: Clays and Clay Minerals, v. 43, no. 1, p. 51-57. 170

Kim, J., and Dong, H., 2011, Application of electron energy-loss spectroscopy (EELS) 171

and energy-filtered transmission electron microscopy (EFTEM) to the study of 172

mineral transformation associated with microbial Fe-reduction of magnetite: 173

Clays and Clay Minerals, v. 59, no. 2, p. 176-188. 174

McCarty, D. K., Sakharov, B. A., and Drits, V. A., 2009, New insights into smectite 175

illitization: A zoned K-bentonite revisited: American Mineralogist, v. 94, no. 11-176

12, p. 1653-1671. 177

Moore, D. M., and Reynolds, R. C., 1989, X-ray Diffraction and the Identification and 178

Analysis of Clay Minerals, Oxford, Oxford university press. 179

Moore, J. C., and Karig, D. E., 1976, Sedimentology, structural geology, and tectonics 180

of the Shikoku subduction zone, southwestern Japan: Geological Society of 181

America Bulletin, v. 87, no. 9, p. 1259-1268. 182

Pytte, A., and Reynolds, R., 1989, The thermal transformation of smectite to illite, in 183

N.D., N., and T.H., M., eds., Thermal history of sedimentary basins: New York, 184

Springer, p. 133-140. 185

Pytte, A. M., 1982, The kinetics of the smectite to illite reaction in contact metamorphic 186

shales: Dartmouth College. 187

9

Seno, T., Stein, S., and Gripp, A. E., 1993, A model for the motion of the Philippine Sea 188

plate consistent with NUVEL‐1 and geological data: Journal of Geophysical 189

Research: Solid Earth, v. 98, no. B10, p. 17941-17948. 190

Van Aken, P., Liebscher, B., and Styrsa, V., 1998, Quantitative determination of iron 191

oxidation states in minerals using Fe L2,3-edge electron energy-loss near-edge 192

structure spectroscopy: Physics and Chemistry of Minerals, v. 25, no. 5, p. 323-193

327. 194

Velde, B., and Vasseur, G., 1992, Estimation of the diagenetic smectite to illite 195

transformation in time-temperature space: American Mineralogist, v. 77, no. 9-196

10, p. 967-976. 197

Yang, K., Kim, J.-w., Kogure, T., Dong, H., Baik, H., Hoppie, B., and Harris, R., 2016, 198

Smectite, illite, and early diagenesis in South Pacific Gyre subseafloor sediment: 199

Applied Clay Science, v. 134, p. 34-43. 200

201

SUPPLEMENTARY FIGURE CAPTIONS 202

Supplementary DR1. The percentages of illite which are taken as the percentages of 203

illite in I-S were calculated using equation 3 in text using the kinetic model parameters 204

(Kim et al., 1995), a published temperature-depth (age) model (Heuer et al., 2017), 0.1 205

Ma time step, and K+ (9.344 mmole K+ /L or 0.044 wt.% K2O). 206

207

Supplementary DR2. Distribution of Deltaproteobacteria with depth at Site C0023 208

obtained by 16S gene-tagged sequences (Heuer et al., 2017), showing an increase up to 209

5% in the whole microbial community at the interval II where cell count also shows an 210

abrupt increase. 211

10

212

Supplementary DR3. Relative amounts of illite with depth, calculated from the clay 213

mineral assemblage of smectite + illite + chlorite + kaolinite = 100% with Biscaye 214

modelling (Biscaye et al., 1965) (solid circles). These results were confirmed with 215

Sybilla (Pytte, 1982) modeling (open circles). 216

Fig. DR1

Fig. DR2

Fig. DR3