Tectonophysics

Temperature-dependent frictional properties of heterogeneous Hikurangi Subduction Zone input sediments, ODP Site 1124

Carolyn Boulton1, André R. Niemeijer2, Christopher J. Hollis3, John Townend1, Mark D. Raven4, Denise K. Kulhanek5, Claire L. Shepherd3

1School of Geography, Environment, and Earth Sciences, Victoria University of Wellington, Wellington, New Zealand

2 Faculty of Geosciences, HPT Laboratory, Utrecht University, Utrecht, the Netherlands3GNS Science, Lower Hutt, New Zealand

4CSIRO Land and Water/Mineral Resources, Urrbrae, South Australia 5IODP, Texas A&M University, College Station, Texas, USA

Appendix A

This appendix contains supporting information in the form of eight figures, with captions, and six tables, with descriptions. Figure A.1 comprises five representative X-ray diffraction patterns of the types of sediments used in the hydrothermal friction experiments. Figure A.2 summarizes the results of particle size analyses performed on all samples used in the hydrothermal friction experiments. Figure A.3 contains images of the sample assembly and ring shear apparatus used to perform the hydrothermal friction experiments. Figures A.4, A.5, A.6, and A.7 are plots of the friction rate parameters and slip weakening distance(s) as a function of sliding velocity and temperature.

Table A.1 shows the presence of biostratigraphically useful nannofossil taxa, assemblage abundance and preservation, and nannofossil zone and absolute age for sediments investigated in the hydrothermal friction experiments as well as additional samples within the Paleocene to Eocene interval. Tables A.2 to A.5 list the friction rate parameters (a–b) measured from velocity steps and plotted in Figure 7 in the text. Detailed information, along with relevant methods and equipment used to perform the analyses, occurs in the figure or table description.

1

Figures

Text A.1. Qualitative and Quantitative X-ray diffraction (XRD) analyses were performed on sediments recovered from Hole 1124C. Approximately 1.5 g of the as-received samples were ground for 10 minutes in a McCrone micronizing mill under ethanol. The resulting slurries were oven dried at 60 °C and then thoroughly mixed in an agate mortar and pestle before being lightly back pressed into stainless steel sample holders for XRD analysis. The XRD patterns of the micronized samples showed evidence of dehydration of the interlayer of the smectite phase, so they were calcium saturated to restore hydration and maintain a more stable two-water interlayer expansion. This involved dispersing the samples in a 1M CaCl2 solution, centrifuging at 5150 g for 10 minutes, calcium saturating a second time, washing with water and then ethanol (centrifuging at 5150 g for 10 minutes after each step), and finally oven drying at 60 °C. The dried materials were thoroughly mixed in an agate mortar and pestle before being lightly back pressed into stainless steel sample holders for X-ray diffraction analysis. XRD patterns were recorded with a PANalytical X'Pert Pro Multi-Purpose Diffractometer using Fe-filtered Co K radiation, automatic divergence slit, 2° anti-scatter slit and fast X'Celerator Si strip detector. The diffraction patterns were recorded from 3 to 80° in steps of 0.017° 2 with a 0.5 second counting time per step for an overall counting time of approximately 35 minutes. Qualitative analysis was performed on the XRD data using in-house XPLOT and HighScore Plus (from PANalytical) search and match software. Quantitative analysis was performed on the XRD data using the commercial package SIROQUANT from Sietronics Pty Ltd. The results are normalized to 100%, and hence do not include estimates of unidentified or amorphous materials. All analyses were performed at the Commonwealth Scientific and Industrial Research Organization (CSIRO).

2

3

Figure A.1. Representative X-ray diffraction patterns for the Hole 1124C sediments studied, along with the identified minerals and their corresponding peaks. All diffraction patterns were obtained from Ca-saturated samples; Ca-saturation results in the more stable two water layer expansion of the smectite (montmorillonite) interlayers. Diffraction patterns pictured were obtained from: (a) 181-1124C-43X-CC (419.1 mbsf), clay-bearing nannofossil chalk; (b) 181-1124C-44X-5, 50-53 (425.8 mbsf), clay; (c) 181-1124C-44X-7, 4-6 (428.34 mbsf), clay; (d) 181-1124C-44X-7, 4-6 (428.34 mbsf), clay, Ca-saturated (black) and heat treated for 1 hour at 350 °C (red). Note ~50% reduction in heulandite-clinoptilolite peak height following heat treatment, indicating a mixture of both zeolites in the sample (Mumpton, 1960). (e) 181-1124C-45X-5, 53-56 (435.53 mbsf), clay-bearing nannofossil chalk.

4

Text A.2. Particle size analyses were performed on the sediment samples after they were disaggregated and powdered for the friction experiments. The particle size analyses were conducted using a Malvern Laser Particle Sizer equipped with a 300RD lens suitable for grain sizes in the range of 0.05-900 m. The active beam length was 2.4 mm, and the samples were dispersed in DI water using a MS17 automated sample dispersion unit with built-in stirrer and ultrasound; stirring and ultrasonication intensity varied from sample to sample depending on its obscuration ratio and tendency to aggregate. Each sample was measured 2 to 3 times per run to ensure reproducibility. All analyses were performed at the High Pressure and Temperature Laboratory, Utrecht University.

Figure A.2. Plot of volume frequency percent against particle size for milled and powdered Hole 1124C samples. Samples are listed by top of sampled interval, in meters below sea floor (mbsf). The results show that the clayey chalk and clays have a slightly larger mean grain size than the clay-bearing nannofossil chalks, likely due to elongate clay particle morphology and/or clay flocculation. Where D90 means that 90% of all measured particles are finer than a given grain size, the average D90 for all samples is 5.9 m (n=9). The average D90 of the clayey chalk and clays is 6.3 m (n=5), and the average D90 of the clay-bearing nannofossil chalks is 5.4 m (n=4).

5

Text A.3. Hydrothermal friction experiments were performed in a hydrothermal ring shear apparatus, High Pressure and Temperature (HPT) Laboratory, Utrecht University. The apparatus is described in detail in Niemeijer et al. (2008) and den Hartog et al. (2012a), and original labelled photographs appear below in Figure A.3.

Figure A.3. Images of the hydrothermal ring shear apparatus used to measure the frictional properties of Hole 1124C sediments, where (a) is a labelled photograph of the apparatus, and (b) depicts the two pistons which, fitted together with an annulus of gouge between them, are inserted into the pressure vessel and sheared under controlled conditions of temperature, effective normal stress, and velocity (see publication text for details of sample preparation and experimental conditions).

6

Text A.4-A.7. The friction rate parameter (a–b) was calculated from coefficient of friction measurements performed following signal gathering and processing methods outlined in http://doi.org/10.5880/icdp.5052.002, an International Continental Scientific Drilling Program (ICDP) data archive. Acquired raw torque and normal force data were processed to obtain shear stress and normal stress measurements respectively.

Externally measured torque was corrected for fluid pressure and shear displacement-dependent friction of the Teflon-coated O-ring seals using calibration values obtained in runs with a dummy sample of carbon-coated PolyEtherEtherKeton with a known sliding friction; seal friction is typically around 0.03 kN (equivalent to ~1.5 MPa shear stress). The contribution of the Molykote-coated confining rings to the measured friction is negligible (see also den Hartog et al., 2012a). The applied normal stress was corrected for the stress supported by the internal seals, the level of which is clearly visible during initial loading and was generally around 0.5 kN (equivalent to ~2 MPa normal stress acting on the sample).

For each velocity step plotted in Figure 6 in the main text, the friction rate parameter (a–b) was determined, where a is the direct effect, and b is the evolution effect, taken to be the sum of b1 and b2 where two state variables were required to fit the velocity step (see details of the fitting procedure in the main text). All rate and state friction parameters were determined using the XLook program and the inversion technique described by Saffer and Marone (2003).

We modelled velocity steps that resulted in unstable stick slips. Although stick-slip instabilities only occur in rate-weakening materials, the magnitude of the (a–b) values obtained from steps containing stick-slip events should be treated with caution. For velocity steps that result in stick slips, we do not report the individual values of the rate parameters a, b1 and b2, or the critical slip distance(s) Dc1 and Dc2.

In the following four figures, we plot the individual rate parameters a, b1 and b2, the critical slip distance(s) Dc1 and Dc2, and (a–b) for clays (52-57% montmorillonite; experiments u644, u713, u639) (Fig. A.4 and A.5) and chalks (66-85% calcite; experiments u643, u645, u657, and u656) (Fig. A.6 and A.7) as a function of velocity and temperature. Any observable trends are described in the figure captions.

7

Figure A.4. Friction rate parameters (a) a, (b) b1, (c) b2, and (d) (a–b) as a function of sliding velocity for velocity steps performed on Hole 1124C clays (52-57% montmorillonite; experiments u644, u713, u639). As sliding velocity increases from 0.3 to 30 m/s, a increases (a), and b1 increases as sliding velocity increases from 1 to 30 m/s (b). We observe no correlation between sliding velocity and b2 or (a–b) (c, d). The critical slip distance(s) (e) Dc1 and (f) Dc2 also show no correlation with sliding velocity. Symbols in (a-c) and (e-f) are colour coded by temperature, where experiments performed at 25 °C are denoted by purple circles, 75 °C by blue circles, 150 °C by orange circles, and 225 °C by red circles (see Fig. A.5). Outliers affect plotted mean values; where outliers are present, trends are shown by dashed lines drawn between median values rather than solid lines drawn between mean values.

8

Figure A.5. Friction rate parameters (a) a, (b) b1, (c) b2, and (d) (a–b) as a function of temperature for velocity steps performed on Hole 1124C clays (52-57% montmorillonite; experiments u644, u713, u639). As temperature increases from 25 °C to 225 °C, (a) a exhibits small increases in median and mean values, a trend also observed with increasing velocity. With increasing temperature, (b) b1 and (c) b2 decrease, with mean and median values of b2 becoming negative at 225 °C. Increasing values of a and decreasing values of b1 and b2 result in the observed transition from rate-weakening (negative a–b) to rate-strengthening (positive a–b) behaviour with increasing temperature (c). The critical slip distance(s) (e) Dc1 and (f) Dc2 show no clear correlation with temperature. Symbols are colour coded by temperature. Outliers (circles outlined in red with labelled values) affect plotted mean values; where outliers are present, trends are shown by dashed lines drawn between median values rather than solid lines drawn between mean values.

9

Figure A.6. Friction rate parameters (a) a, (b) b1, (c) b2, and (d) (a–b) as a function of sliding velocity for velocity steps performed on Hole 1124C chalks (66-85% calcite; experiments u643, u645, u657, and u656). We observe no correlation between sliding velocity and (a) a or (b) b1. At 30 m/s, mean and median values of (c) b2 decrease. There appears to be no correlation between sliding velocity and (d) (a–b) or the critical slip distance(s) (e) Dc1 and (f) Dc2. Symbols in (a-c) and (e-f) are colour coded by temperature, where experiments performed at 25 °C are denoted by purple circles, 75 °C by blue circles, and 150 °C by orange circles (see Fig. A.7). All velocity steps at 225 °C resulted in stick-slips, so individual parameters from those steps are not plotted; (a–b) values from velocity steps performed at 225 °C are plotted in (d) and in Fig. 7d in the main text. Outliers (circles outlined in red with labelled values) affect plotted mean values; where outliers are present, trends are shown by dashed lines drawn between median values rather than solid lines drawn between mean values.

10

Figure A.7. Friction rate parameters (a) a, (b) b1, (c) b2, and (d) (a–b) as a function of temperature for velocity steps performed on Hole 1124C chalks (66-85% calcite; experiments u643, u645, u657, and u656). As temperature increases from 25°C to 225°C, (a) a, (b) b1, and (c) b2 exhibit no systematic trend. With increasing temperature, median and mean values of (d) (a–b) decrease. Deducing the cause of the decrease in (a–b) with increasing temperature requires values of the individual parameters at 225 °C, but they could not be measured reliably because of stick slip instabilities. The critical slip distance(s) (e) Dc1 and (f) Dc2 show no clear correlation with temperature. Symbols are colour coded by temperature. Outliers (circles outlined in red with labelled values) affect plotted mean values; where outliers are present, trends are shown by dashed lines drawn between median values rather than solid lines drawn between mean values.

11

Tables

Table A.1. Samples for nannofossil analysis were prepared using standard smear slide techniques (Bown and Young, 1998). A small amount of sediment was mixed with water on a coverglass using a toothpick, and then smeared evenly across the coverglass. The coverglass was dried at low temperature (~60 ºC) on a hot plate, affixed to a glass microscope slide using Norland Optical Adhesive No. 61, and cured under ultraviolet light for a minimum of 5 to 10 minutes. Assemblage data were collected using two different microscopes at GNS Science, Lower Hutt, New Zealand. D.K. Kulhanek used a Leitz Ortholux II POL-BK microscope to count 400 nannofossil specimens per sample in random fields of view at 1000 magnification. A minimum of 2 additional traverses of the 40-mm coverglass were scanned for rare taxa. C.L. Shepherd used an Olympus BX53 microscope at 1000 magnification to document the overall abundance of nannofossils and presence/absence of individual taxa by scanning at least 2 coverglass traverses. Cross-polarized and plane-transmitted light were used by both investigators. Taxonomic references for species can be found at the online resource Nannotax (http://www.mikrotax.org/Nannotax3/?dir=Coccolithophores). Results were correlated to the nannofossil zonation of Martini (1971), with absolute ages for biostratigraphic events tied to the Geological Time Scale 2012 (GTS2012; Gradstein et al., 2012). Data are reported as presence/absence of biostratigraphically important species or species groups in Tables S1.

12

The lowermost samples examined (181-1124C-46X-CC and 46X-5, 50-53 cm [445.54-445.20 m below seafloor (mbsf)]) are assigned to nannofossil Zone NP4 based on the presence of Chiasmolithus bidens (first appearance datum [FAD] at 62.07 Ma). The presence of Sphenolithus primus (FAD at 61.98 Ma) in sample 46X-4, 57-60 cm (443.66 mbsf) indicates upper Zone NP4. The FAD of Fasciculithus tympaniformis (FAD at 61.51 Ma) marks the base of Zone NP5 and is found in sample 46X-1, 50-53 cm (439.20 mbsf). Sediment between this sample and 45X-1, 17-21 cm (429.17 mbsf) is assigned to Zone NP5 based on the absence of Heliolithus cantabriae, which has a FAD in latest Zone NP5 at 59.6 Ma. Within this interval, the last appearance datum (LAD; 60.73 Ma) of Lithoptychius pileatus occurs in sample 45X-3, 50-53 cm (432.50 mbsf). These datums indicate a relatively consistent sedimentation rate of ~10 m/Myr for the Paleocene interval of Hole 1124C studied here.

Three samples from the core catcher (bottom) of core 44X (44X-CC) were examined; two from the dark lithology and one from the light lithology (Figure 3). The nannofossil assemblage in the light-coloured lithology is Paleocene in age, whereas the assemblages in the dark-coloured lithology are a mix of mid-Paleocene and middle Eocene taxa. This suggests that either the Paleocene sediment is reworked into the middle Eocene sediment, or conversely, that the middle Eocene sediment has been mixed into the Paleocene sediment via bioturbation. Given that the sediment in the core catcher is dominantly dark-coloured (Figure 3), it is likely that it represents the background sedimentation and light-coloured Paleocene clasts have been reworked into the section. Above the core catcher, sample 44X-7, 4-6 cm is barren of calcareous nannofossils and cannot be assigned to a zone. Samples 44X-6, 50-53 cm to 44X-4, 50-53 cm (427.30-424.30 mbsf) are assigned to middle Eocene Zone NP16 based on the presence of Reticulofenestra umbilicus (FAD 41.94 Ma). Thus, the hiatus separating mid-Paleocene and middle Eocene sediments at Site 1124 is at least 17 Myr in duration.

Table A.1. Ranges of selected calcareous microfossils used for biostratigraphy. All analyses were performed at GNS Science.

13

Table A.2. Full quantitative X-ray diffraction results. Methods are listed in Text A.2

14

15

Tables A.3-A.6. Friction rate parameters obtained from inversions and plotted in Figure 7 in the text. Conditions imposed and samples used in the experiments are listed in Table 1. We quantified the friction rate parameter (a–b) using an iterative least squares method incorporating the Dieterich (1979) rate-and-state friction equation:

❑ss=❑0+a ln( VV 0 )+bi ln(V 0 θi

Dci)

Here, V0 and V are the initial and final load point velocities, respectively, 0 and ss are the initial and final, steady-state, friction coefficients, a, bi, and Dci are empirical constants, and i is a state variable that evolves with time according to

d θ i

dt=1−

V θi

Dci

For each step, the friction rate parameter a–b was determined using the XLook program and the inversion technique described by Saffer and Marone (2003). To fit most velocity steps, two state variables were required, and the subscript i is equal to 1 for steps with one state variable and 1 and 2 for steps with two state variables. For steps with two state variables, b is the sum of b1 and b2. Some modelled velocity steps resulted in unstable stick slips. Stick-slip steps occurred at temperatures of 150 °C and 225 °C in experiments u643 and u657 performed on clay-bearing nannofossil oozes (83 to 85% calcite); stick-slip steps occurred only at 225 °C in experiments u645 and u656 performed on clay-bearing nannofossil oozes (66-67% calcite). Stick-slip instabilities can lead to rate-and-state models that have no solution, particularly if the velocity step occurs directly after or right before a stick-slip event. In those cases, it was necessary to move the modelled point of the velocity step forward or backward to obtain a solution. Although stick-slip instabilities only occur in rate-weakening materials, the magnitude of the (a–b) values obtained from steps containing stick-slip events should be treated with caution. In addition, in the tables below, some values for the initial 1 m/s to 0.3 m/s velocity step are not listed where the sample was not at steady-state when the step occurred.

16

Table A.3. Friction rate parameter (a–b) data corresponding to Figure 7a in the text. Conditions imposed and samples used in the experiments are listed in Table 1. Methods are described in the Tables A.3-A.6 description. Repeat steps were performed at 75 °C in experiment u644 to check intra-experiment reproducibility. Experiments u644 and u713 were performed on the same sampled interval (181-1124C-44X-7, 4-6 cm) to check inter-experiment reproducibility.

17

Table A.4. Friction rate parameter (a–b) data corresponding to Figure 7b in the text. Conditions imposed and samples used in the experiments are listed in Table 1. Methods are described in the Tables A.3-A.6 description.

18

19

Table A.5. Friction rate parameter (a–b) data corresponding to Figure 7c in the text. Conditions imposed and samples used in the experiments are listed in Table 1. Methods are described in the Tables A.3-A.6 description.

20

Table A.6. Friction rate parameter (a–b) data corresponding to Figure 7d in the text. Conditions imposed and samples used in the experiments are listed in Table 1. Methods are described in the Tables A.3-A.6 description.

Dataset A.1. A text file accompanies this appendix with the complete experimental dataset. Data for each experiment listed in Table 1 appear as columns of displacement (mm) and coefficient of friction (). Data have been processed using methods described in the main text and appendix text.

Reference

Mumpton, F.A., 1960. Clinoptilolite redefined. Am. Mineral. 45, 351–369.

21