IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.338.102.2014

Geochemistry

Interstitial water geochemistry for core samples

Squeezing method

When core recovery and quality allowed, 15–50 cm long whole-round core (WRC) samples were taken from the cored sections. Squeezed interstitial water (IW) was sampled from Holes C0002J, C0002K, C0002L, C0021B, and C0022B. IW was not obtained from Hole C0002H because of low core recovery. Because the chemical composition of IW changes rapidly, WRC samples must be cut from a homogeneous part of the core section soon after recovery on the core cutting deck. Samples are capped in the core cutting area and immediately scanned by X-ray CT to check for the presence of structurally and/or lithologically important features as well as for homogeneity. In the case of important features being identified on X-ray CT images, the samples were preserved for further observation and another portion of the core was cut for IW analysis. Samples approved for IW analysis were placed into nitrogen-filled glove bags, removed from the core liner, and cleaned by scraping off sediment along the outer surface of the WRC that came in contact with seawater or drilling fluid or had experienced smearing or oxidation. When samples were hard, they were crushed into small fragments inside the glove bag. This procedure made IW extraction easier when compared with putting sample blocks directly into the squeezer. The portion of the cleaned WRC to be squeezed was placed in a Manheim-type titanium squeezer (Manheim, 1966).

Following the procedures of the Expedition 319 Scientists (2010b), the samples were presqueezed by a manually operated squeezer until a few drops of water came out. This allowed a maximum amount of IW to be collected. Then, a 25 mL acid-washed (12 N HCl) syringe was installed into the IW sample port of the squeezer, and the samples were subject to automatic squeezing, which comprised six different steps, each lasting for 5 or 10 min, where loads of (1) 15,000, (2) 17,000, (3) 20,000, (4) 21,500, (5) 23,000, and (6) 25,000 lb (pound-force) were applied until enough volume (commonly >80 mL) of IW was collected for analyses. Loads of 15,000 or 17,000 lb are commonly enough to obtain the amount of IW from sediment shallower than ~300 mbsf; however, samples from deeper than ~400 mbsf were loaded to 20,000–25,000 lb, not only because of low IW content but also because of sediment consolidation. High pressure was needed to collapse the network of grains and release the IW.

To avoid contamination by sediment, IW was passed through a paper filter fitted with 2–4 titanium 90 mesh screens at the base of the squeezer and a 0.45 µm disposable filter. The squeezed water was filtered again with a 0.45 µm disposable filter and stored in high-density polyethylene sample vials, previously prepared by immersion in 55°C 10% trace metal grade 12 N HCl for at least 24 h, rinsed with Millipore 18.2 MΩ·cm Type 1 ultrapure (i.e., Milli-Q) water, and dried in a class 100 laminar flow clean hood. An aliquot of the sample water was stored in a plastic bottle without further treatment for the analyses of pH, alkalinity, major anions, and nutrients (phosphate and ammonium ions). Another aliquot of the sample was stored in a plastic bottle acidified with HCl to be 0.4 vol% of 6 N HCl to stably dissolve cations and minor and trace metal ions.

GRIND method

In cases of small sample volumes of recovered sediment (Section 338-C0002J-2R-2) and of low concentration of labile water from deep and lithified sediment (Section 338-C0002H-2R-2), the ground rock interstitial normative determination (GRIND) method was applied to extract IW. Additionally, in order to investigate the accuracy of the GRIND method, the GRIND method was applied to several samples together with the standard squeezing method (see Table T31 in the “Site C0002” chapter [Strasser et al., 2014b]) and improved to acquire reliable data (see “Appendix A” for details).

A 5–10 cm long sample was taken from every IW WRC sample and put inside a glove bag, in which the external surfaces were scraped clean. Afterward, the samples were fragmented down to <1 cm in size (Wheat et al., 1994; Expedition 315 Scientists, 2009a). About 40 or 80 g of the sample (depending on the availability of sample weight) was carefully weighed, placed inside an agate ball-mill cylinder together with five agate balls, and 5 or 10 g (for 40 or 80 g samples, respectively) of Milli-Q water or dilute HNO3 solution with pH adjusted to 3 was added after accurate weighing. Diluted HNO3 solution of pH 3 was used because certain compositions gave more consistent values with those of squeezed water rather than using Milli-Q water. In order to remove dissolved oxygen, the Milli-Q water was bubbled with nitrogen gas for >48 h before being added to the sample. In the original procedure, Milli-Q water was spiked with indium (In) as a standard (concentration of 500 ppb), although it was not added here because the recovery of In was very low and it did not work as a spike for calculating the dilution rate. The water content was determined based on the weight difference before and after drying the sample sediment at 105°C overnight. The concentration of dissolved components of IW was calculated by the dilution rate of (IW content + added solution)/(IW content). This calculation was performed using the weight of IW in the used sample and added solution; thus, the weight of that with added solution was measured before processing.

Grinding the mixture of fragmented sample and HNO3 solution in the ball mill took place for 5 min at 400 rpm, long enough to adequately crush and homogenize the contents and at the same time minimize the risk of reactions between the sample and the solution. Afterward, the ground slurry was squeezed and the extracted water was collected in a syringe, in a similar manner as the squeezing method. Analytical results of the IW obtained by the GRIND method were evaluated based on the comparable methods (see “Appendix A”).

Interstitial water analyses

The standard IODP procedure for IW analysis was modified according to the availability and functionality of onboard instruments (Expedition 319 Scientists, 2010b). However, because of the limited amount of extracted pore water, not all of the standard IODP measurements were conducted.

Concentrations of numerous major and minor components in the IW were analyzed. Chlorinity was measured on a 100 µL aliquot by potentiometric titration using a Metrohm autotitrator and silver nitrate (AgNO3) as a titrant in 30 mL of 0.2 M sodium nitrate (NaNO3) solution. Relative standard deviation (RSD) for chlorinity was better than ±0.2%, based on repeated analyses of International Association for the Physical Sciences of the Oceans (IAPSO) standard seawater, which were conducted between the measurement of every 5 samples.

Bromide and sulfate concentrations were measured with a Dionex ICS-1500 ion chromatograph with an anion column. An aliquot was diluted to 1:100 (10 µL in 990 µL) with Milli-Q water. IAPSO standard seawater aliquots (2.5, 5, 7.5, and 10 µL in a total of 1000 µL) were analyzed at the beginning and end of each run for quality control and to monitor potential drift in sensitivity throughout a particular run. RSDs are ±3% for bromide and ±1% for sulfate.

An ion chromatograph was used to determine the concentrations of major cations such as Na+, K+, Mg2+, and Ca2+. Aliquots of IW samples acidified with 6 M HCl (Tamapure-AA-100 grade) were used for this measurement. These samples were diluted to 1:200. The diluted samples were placed in an autosampler together with five calibration solutions and two blank solutions (Milli-Q water). For quality checks, a 1:200 solution of diluted IAPSO standard seawater was measured between every eight samples. To determine the concentration of each element, standard solutions were analyzed to construct calibration curves from the measured peak area and the known concentration. For the calibration, an IAPSO standard seawater solution containing Na+ (480 mM), K+ (10.4 mM), Mg2+ (54 mM), and Ca2+ (10.6 mM) was diluted to four standard solutions (25%, 50%, 75%, and 100%). The resulting RSD for the measurements was less than ±0.5% for Na+, ±1% for K+, ±0.7% for Mg2+, and ±0.5% for Ca2+.

For nutrients (PO43– and NH4+), colorimetric methods were applied. Both compositions must be analyzed within 24 h because they are quickly degraded by biological activity. Ammonium adsorption of indophenol blue at 640 nm wavelength was measured with a spectrophotometer (Shimadzu UV-2550PC), with an aliquot of 100 µL of sample IW used as the minimum volume. Standard, blank, 2, 4, 6, 8, and 10 mM NH4+ solutions were prepared in the same manner as the sample solutions and analyzed within 5 h. Phosphate (an aliquot of 100 µm IW) was analyzed using adsorption of molybdate blue at 885 nm wavelength with the spectrophotometer used for NH4+. Standard, blank, 0.5, 1, 2.5, 5, and 7.5 mg/L solutions were prepared in the same manner as that of sample solutions. RSDs of repeated analyses of both components are <1%.

Minor element (B, Ba, Fe, Li, Mn, Si, and Sr) concentrations were determined on 500 µL aliquots using an inductively coupled plasma–atomic emission spectroscopy (ICP-AES) (Horiba Jobin Yvon Ultima2). Aliquots of IW acidified with 6 M HCl (Tamapure-AA-100 grade) were diluted to 1:20 with 0.15 M HNO3. Ultrapure primary standards (SPC Science PlasmaCAL) were prepared with a matrix solution of sulfate-free artificial seawater to fit the sample matrix, and 10 ppm Y solution was added as an internal standard. A matrix solution that approximated IAPSO standard seawater major element concentrations was prepared by mixing the following salts in 1 L of Milli-Q water acidified with 4 µL of Tamapure-AA-100 grade 6 M HCl: 26.9 g NaCl, 3.81 g MgCl2, 1.0 g CaCO3, and 0.75 g KCl. A stock standard solution was prepared from ultrapure primary standards (SPC Science PlasmaCAL) in the 1% HNO3 solution and then diluted in the same 1% ultrapure HNO3 solution used for IW samples to concentrations of 100%, 50%, 25%, 10%, 5%, and 1%. A 10 ppm Y solution diluted as 1% HNO3 solution was prepared as a blank. A series of standards were made by adding 1.25 mL of each stock solution to 8.75 mL of matrix solution.

The matrix-matched 100% standard solution contained the following concentrations of elements: B = 145 μM, Ba = 11.4 μM, Fe = 2.80 μM, Li = 22.5 μM, Mn = 2.84 μM, Si = 55.7 μM, and Sr = 17.8 μM. Because values of many of these elements in IAPSO standard seawater are either below detection limits (e.g., Fe and Mn) or variable, a standard prepared in the 10% matrix-matching solution was repeatedly analyzed to calculate the precision of the method. RSDs determined by repeated analyses of the 10% matrix-matching solution were ±2.5% for B, ±1.5% for Ba, ±3.5% for Fe, ±4.1% for Li, ±2.5% for Mn, ±2.5% for Si, and ±2.0% for Sr.

V, Cu, Zn, Rb, Mo, Cs, Pb, and U were quantified on 500 µL IW samples using an inductively coupled plasma–mass spectrometry (ICP-MS) (Agilent 7500ce) equipped with an octupole reaction system to reduce isobaric interferences from polyatomic and double-charged ions. We used the same aliquot after determining major and minor elements using an ion chromatograph and ICP-AES. To correct for interferences between some of the transition metals (V, Cu, and Zn) and some major element oxides, solutions containing the metals with concentrations similar to IAPSO standard seawater values were prepared. These solutions were then analyzed at the beginning of each measurement, and an interference correction was applied based on the average ion counts per second measured on the standard solutions divided by the abundance of the interfering elements.

A 500 µL aliquot of sample IW was diluted with 500 µL with 500 ppb In internal standard solution and 4 mL of 1% HNO3 based on the previous determination of detection limits and low concentrations of the elements of interest. A primary standard solution was made to draw the calibration lines matching the maximum range of predicted concentrations based on published results of deep-sea pore fluid compositions in a variety of settings. The concentrations of the standard are as follows: V = 40 ppb; Cu, Mo, Pb, and U = 40 ppb; Zn = 140 ppb; Rb = 540 ppb; and Cs = 40 ppb. This primary standard was diluted with 1% HNO3 solution to relative concentrations of 100%, 50%, 25%, 10%, 5%, 1%, and blank. A 500 µL split of these standards was then further diluted by addition of the In solution, 3.5 mL of 1% HNO3 solution, and 500 µL of a 560 mM NaCl solution to account for matrix suppression of the plasma ionization efficiency. The 200% and 400% standard solutions were also prepared using 100% solution changing dilution rate (i.e, instead of combination of 500 µL 100% standard, 500 µL In, and 4 mL HNO3 solutions, 1000 µL and 2000 µL of 100% standard solution were diluted with 500 µL In solution and 3.5 and 3.0 mL HNO3 solutions, respectively). The 25% standard was diluted accordingly and analyzed together with eight samples throughout every analysis series for precision and to check the drift during measurements. Blanks were also analyzed between every eight samples, and detection limits were determined to be three times the standard deviation of a procedural blank of Milli-Q water acidified with 4 mL of Tamapure-AA-100 grade 0.15 M HNO3. The average precision of multiple determinations of the 25% ICP-MS standard was ±3.0% for 51V, ±3.0% for 65Cu, ±7% for 65Zn, ±1% for 85Rb, ±2% for 95Mo, ±0.5% for 133Cs, ±4% for 208Pb, and ±1% for 238U.

Organic geochemistry

Total carbon (TC), inorganic carbon (IC), and total nitrogen (TN) were analyzed using samples from cuttings and core samples. Total sulfur (TS) was measured using core samples. Cuttings (~10 cm3) were washed with seawater, sieved, freeze-dried, and ground to powder before analysis. Core samples (~10 cm3) were freeze-dried and ground to powder before analysis.

TC, TN, and TS concentrations were determined using a Thermo Finnigan Flash elemental analysis (EA) 1112 carbon-hydrogen-nitrogen-sulfur analyzer. Calibration was based on the synthetic standard sulfanilamide, which contains 41.81 wt% C, 16.27 wt% N, and 18.62 wt% S. About 15–25 mg of sediment powder was weighed and placed in a tin container for carbon and nitrogen analyses. For sulfur analysis, the same amount of sediment powder was weighed and put into a Ti container with an equivalent mass of V2O5 catalyst. Sediment samples were combusted in an oxygen stream at 900°C for carbon and nitrogen and at 1000°C for sulfur. Nitrogen oxides were reduced to N2, and the mixture of CO2, N2, and SO2 was separated using a gas chromatograph (GC) equipped with a thermal conductivity detector (TCD). The accuracy of the analysis was confirmed using soil NCS reference material (Thermo Scientific, Milan, Italy), sulfanilamide standard (Thermo Scientific), and JMS-1 reference material.

With the same set of samples used for elemental analysis, we determined IC using a Coulometrics 5012 CO2 coulometer. Approximately 15–25 mg of sediment powder was weighed and reacted with 2 M HCl. The released CO2 was titrated, and the change in light transmittance was measured with a photodetection cell. The weight percentage of calcium carbonate was calculated from the IC content, assuming that all the evolved CO2 was derived from dissolution of calcium carbonate, by the following equation:

CaCO3 (wt%) = IC (wt%) × 100/12. (4)

No correction was made for the presence of other carbonate minerals. Standard deviation for the samples was less than ±0.05 wt%. NIST-SRM 88b and JSD-2 (standard reference materials) were used to check accuracy. TOC contents were calculated by subtracting IC from TC contents as determined by EA.

Assessing drilling mud contamination

Very high TOC concentrations in cuttings raised the potential of samples being contaminated by drilling mud. To assess any potential contamination, eight mud water (LMW) and mud pit (LMT) samples from the mud circulation tanks were analyzed for carbonate, organic carbon, and nitrogen concentrations during Expedition 338. Approximately 5 mL of mud water sample was placed in a petri dish and dried at 60°C. Dried samples were ground by an agate mortar and analyzed for IC, TC, and TN concentrations following the same analytical protocol used for cuttings. Background concentration data are listed in Table T13.

Because of high viscosity, mud water and mud pit water samples were diluted 10 times with Milli-Q water and placed in an ultrasonic bath for 1 h so that the solutions were well mixed. Supernatant drilling mud water was centrifuged at 9500 rpm at 4°C for 1 h. Samples were filtered and analyzed for major, minor, and trace elemental concentrations following the standard procedure adopted for cuttings. The results are listed in Table T32 in the “Site C0002” chapter (Strasser et al., 2014b).

Gas analysis

Sampling and analyzing mud gas

Onboard mud-gas monitoring system

Continuous mud-gas monitoring (CMGM) is a standard procedure in the oil and gas industry and is usually carried out to obtain real-time, qualitative information regarding a gas and/or oil reservoir and to assess the need (or advantage) of deeper drilling. In the framework of IODP, CMGM was first carried out during Expedition 319 (Expedition 319 Scientists, 2010a) using third-party tools and was successfully applied during IODP Expedition 337 with onboard instruments (Expedition 337 Scientists, 2013). The onboard set-up follows previous experience with scientific real-time CMGM and sampling in the context of continental drilling (e.g., Erzinger et al., 2006; Wiersberg and Erzinger, 2007, 2011).

Gases released by the drill bit crushing the source rocks were transported upward within the drilling mud. After passing the flow splitter (Fig. F16), the gases were extracted using a degasser, in which a vacuum is applied and which hosts an impeller ensuring fluid circulation. Some of the mud flowed directly into the bypass line for the degasser. The gas was sucked into PVC tubing and transported to the mud-gas monitoring laboratory, passing through a safety valve placed between the degasser and the monitoring laboratory to prevent overflow of drilling mud into the system. The safety valve consists of a 1 m long cylinder with a central tube that opens at both ends. The lower 40 cm of the cylinder and the tubing are covered with water. If the gas pressure is too low or too high, air can be sucked in or blown out, respectively, which causes water either to flow from the central tube to the cylinder or vice versa to compensate the pressure difference. During Expedition 337, problems arose from mud sucked into the tubing. As an early countermeasure, a mud catcher was inserted between the degasser and the safety valve.

Gas extracted from the drilling mud traveled through a 50 m long tube with an inner diameter of 3 mm, taking ~6 min for gas to arrive at the mud-gas monitoring laboratory (Fig. F16E) (lag time was determined based on the time difference between the start of mud flow and the arrival of mud gas in the monitoring laboratory when drilling resumed after periods in which mud flow had been stopped). For such a short gas traveltime, particularly through PVC tubing, diffusion loss during transportation is negligible (Wiersberg and Erzinger, 2007). In the mud-gas monitoring laboratory, particles and water vapor were removed from the incoming gas by a dehydration module, after which the dry and clean gas was distributed to different instruments. Sampling for postcruise analysis on shore was taken from two sampling ports either before or after passing the mist and moisture remover.

On-line analysis of (non)hydrocarbon gases by gas chromatography

A fraction of the drilling mud gas flowed directly into a GC-natural gas analyzer (NGA) (Agilent Wasson ECE 6890N) that theoretically allows the analysis of hydrocarbon gases (methane, ethane, propane, iso-/n-butane, and pentane [i.e., C1–C5]), Ar, He, O2, N2, Xe, CO, and CO2. The main component of the GC-NGA system is a GC equipped with a gas sampling port with a multiposition valve. Contrary to the procedures used during Expedition 337, hydrogen rather than helium was supposed to be used as the carrier gas. However, hydrogen caused baseline and concentration problems for several elements, including Ar and O2; therefore, nitrogen was used as the carrier gas.

Analysis of hydrocarbon gases was conducted by passing the gas flow into a 50 cm capillary column that is able to retain hexane and heavier hydrocarbon components. Lighter hydrocarbon gases are then separated by another 49 cm capillary column that connects to a flame ionization detector (FID), with which measurements were conducted every 20 min. CH4 was separated from the rest of the components by an 8 inch micropack column (Wasson ECE Instrumentation, column Code 2378), whereas CO2 was separated by a 1.27 cm capillary column (Wasson ECE Instrumentation, column Code S036). Both columns are connected to a TCD with a detection limit of 200 ppm for all permanent gases besides CO, which had a detection limit of 400 ppm. For the remaining hydrocarbons, the detection limit was <1 ppm. Unfortunately, because of nitrogen being used as the carrier gas and the absence of proper calibration gases, only H2 could be determined with the TCD for the nonhydrocarbons.

Although the GC-NGA has good sensitivity, the temporal and spatial resolution of the mud-gas analysis with a GC-NGA is limited because of the rather long run time of 20 min.

The GC-NGA was calibrated on a daily basis in order to detect any sensitivity changes. Two standards were used. The standard mixture for calibration of permanent gases contained 1% of Ar, CO, Xe, O2, H2, CO2, and He in a balance of N2. The hydrocarbon standard mixture contained 1% C1–C5 in a balance of N2.

Hydrocarbon concentration is obtained by FID analysis. However, care has to be taken in data interpretation because measured concentrations are influenced by drilling conditions such as ROP and mud flow. Consequently, drilling operations were monitored to allow comparison and interpretation of quantitative data.

Fortunately, in situ conditions can still be investigated by the ratio of hydrocarbon gases, which is only slightly affected by drilling parameters. In particular, the Bernard parameter (C1/[C2 + C3]) is a valuable tool to distinguish between hydrocarbon gases from biogenic and thermogenic sources and allows a first estimation regarding the thermal maturity (see Fig. F17) (Pimmel and Claypool, 2001; Ocean Drilling Program, 1992).

On-line analysis of the stable carbon isotopic composition of methane

Another part of the incoming mud gas was directed to a methane carbon isotope analyzer (MCIA) (Los Gatos Research, Model: 909-0008-0000). The concentration and stable carbon isotopic composition of methane is determined on the basis of cavity ring-down spectroscopy technology. The instrument is composed of three parts: the main body of the MCIA, a gas dilution system (DCS-200), and an external pump. The stable carbon isotopic composition of methane is reported in the δ13C notation relative to the Vienna Peedee belemnite (VPDB) standard and expressed in parts per thousand (per mil):

δ13C = (Rsample – RVPDB)/RVPDB, (5)

where

Rsample = 13C/12C, (6)

and

RVPDB = 0.0112372 ± 2.9 × 10–6. (7)

For CH4, the precision and accuracy are within 1‰ for concentrations >400 ppm but worsen to <4‰ for concentrations of 200–400 ppm. The MCIA comprises a gas dilution system that works with hydrocarbon or zero-free air and allows measurement of methane concentrations in the range of 500–106 ppm (i.e., 100%). However, the dilution system did not function because of technical problems, and the concentration data from the MCIA were not utilized when >1 × 104 ppm and <2 × 105 ppm.

Besides the determination of δ13C, the MCIA also allows the determination of whole methane concentration. A disadvantage of the MCIA compared to the GC-FID is its lower sensitivity. However, the sampling frequency is much higher (a frequency of 1 measurement/s was chosen here), enabling 100–200 measurements per meter of drilled sediment, depending on the ROP and mud flow. The sensitivity of this instrument was checked daily with a standard gas with 2500 ppm CH4 and a δ13C value of –38.8‰.

The determination of δ13C combined with the Bernard parameter is a powerful tool to distinguish between biogenic and thermogenic sources of hydrocarbon gases during mud-gas monitoring (for details, see Whiticar, 1999) (Fig. F18). It allows further detection of mixed and oxidized gases. However, for further investigations, it will be necessary to do onshore δ13C analysis of the higher homologues and δD of all hydrocarbons.

On-line gas analysis by process gas mass spectrometer

Incoming mud gas was also transferred to a process gas mass spectrometer (PGMS) (Ametek ProLine process mass spectrometer), which allowed continuous monitoring of H2, He, O2, Ar, Xe, N2, CO, CO2, methane, ethane, propane, and butane (differentiation between n- and iso-butane not possible). One advantage of the PGMS is the presence of vacuum not requiring a carrier gas, which might alter the individual gas concentration. The PGMS uses a quadrupole mass filter and identifies gases based on the individual molecular masses of the desired compounds. A Faraday cup detector provides an optimal scanning range of mass-to-charge ratio (m/z) 1–100 and an optional scanning range of m/z 1–200 with the mass resolution of 0.5 at 10% peak height. Input gas flow rate is set to 50 mL/min. For quantification of individual gas species, the PGMS was calibrated on a daily basis using the same standards as for the GC-NGA. Although the PGMS is less sensitive (within 1 ppm) than the GC-NGA, full-range measurements (i.e., m/z 1–200) are possible every 20 s, resulting in better depth resolution. However, during Expedition 338, because of the limitations of the dedicated laptop PC, full-range measurement was not possible. Instead, the system was changed to the “trend mode,” where only selected masses were determined. The dwell time was changed to 120 ms, which decreased the sampling period to 5 s.

For regular quality assurance, three different calibrations had to be carried out:

  1. Binary calibration. This calibration was necessary to establish peak ratios of ion fragments and was usually carried out using a mixture of two gases (a noninterfering balance gas like Ar mixed with a gas that was expected in the sample stream with an appropriate concentration). By contrast, here, only CH4 with a concentration of 100% was used.
  2. Blend calibration. This calibration was used to compensate for ionization variations. Two different standard gases were used: one containing 1% of Ar, CO, Xe, O2, H2, CO2, and He in a balance of N2, and the other containing 1% of C1–C5 in a balance of N2. Pure N2 and Ar were used for daily background checks. Unfortunately, the concentrations were far too high for the expected concentrations, which led to overestimation of some noble gases.
  3. Background calibration. This calibration was used to determine the atmospheric values of gases in the vacuum chamber. For this purpose, Ar and N2 calibration gases were used, each having a concentration of 100%.

Ar, N2, and O2 concentrations can serve as proxies for air contamination during drilling operations or from the mud-gas monitoring system. Air can be introduced into the borehole when the pipe is broken to recover core, when mud flow is stopped while new pipe connections are made (every 38 m for one stand of four joints of drill pipe), when pressure drops in the gas separator, when mud gas is flowing from the bypass line into the flow splitter, or when leaks in the degasser or in the mist and moisture remover occur. Very high N2 values might further indicate humic source rocks (Whiticar, 1994). Note that CO2 can be analyzed by the GC-NGA, but the resulting concentrations are not meaningful because drilling mud is highly alkaline.

On-line radon analysis

Radon analysis was carried out using a stand-alone radon monitor (Alpha GUARD PQ2000 PRO) provided by the JAMSTEC Institute for Research on Earth Evolution (IFREE). The apparatus was attached to the auxiliary port of the scientific gas monitoring line parallel to other instruments. Rn itself is radioactive with a half-life time of 3.8 days; thus, the Rn decay can be counted within an ion-counting chamber with a volume of 650 mL (effective volume is ~500 mL). Measurements took place every 10 min with 5 counts per minute (cpm) and a sensitivity of 100 Bq/m3 in the concentration range of 2 to 2 × 106 Bq/m3. Synchronous with Rn measurements, internal temperature, pressure, and relative humidity are monitored and, together with Rn data, can be exported in CSV format.

Sampling for shore-based analysis

Unlike during Expedition 337 (Expedition 337 Scientists, 2013), discrete samples were not only collected in IsoTube samplers (Isotech Laboratories, Inc.) but also in copper tubes for noble gas studies and in glass flasks for stable isotope studies. A sample interval of 500, 200, and 50–100 m was chosen for samples stored in the IsoTubes, copper tubes, and glass flasks, respectively.

The configuration was from a sampling port at the main gas flow line, where PVC tubing was suspended and connected to the glass flasks. The glass flasks had valves at both ends to control the gas flow and allow passing of gas through the glass flask again into PVC tubing, which connected the glass flask and the copper tubes. The copper tubes were placed in a guide rail and allowed the passage of gas through the next section of PVC tubing, which ended in a sampling port connected to the main gas flow line (Fig. F16).

After a certain time, depending on the flow rate, sampling was done by closing the valves at both ends of the glass flasks and placing clamps at both ends of the copper tubes. Afterward, the glass flasks and copper tubes were replaced with empty ones. Additional samples were taken after pipe tripping (trip gas) or when mud-gas monitoring indicated enhanced inflow of formation fluid.

Recording on-line gas analysis and monitoring drilling operations, time, and depth

As mentioned above, the recovery and concentration of gases can be affected by drilling operations. Drilling parameters were monitored and recorded in the SSX database together with gas data and the lag depth determined by technicians from Geoservices (Schlumberger). The results of the on-line measurements were made available in the mud-gas monitoring laboratory and on the onboard server. To correlate the results of the gas analysis with variations in the drilling procedure as well as lag depth determination, ship time (UTC + 9 h) was used. As an exception, the stand-alone Rn monitor used an internal clock set to the time zone UTC + 1 h. Adjusting the internal clock of the MCIA to ship time failed; consequently, it was 90 s ahead during Expedition 338.

Regarding the lag depth (i.e., the difference between the depth of arriving mud gas and actual depth of the borehole), Lag Depth L (as recorded in real time in the SSX database and provided by Geoservices) was used to assign data and samples from mud-gas monitoring to the correct subseafloor depth. Lag Depth L was calculated based on the ROP and borehole volume and was recorded in meters below rig floor (rotary table) (DRF). It further considered the transfer time between the degasser and the different laboratory instruments, which was 6 min to the MCIA and the IsoTube sampling unit and almost 9.5 min for the other instruments. Unfortunately, only one Lag Depth L could be recorded. Consequently, only the transfer time of 6 min was taken into account. Data obtained from the Rn instrument, GC-NGA, and PGMS were corrected after the data were exported and evaluated. Conversion to mbsf was done by subtracting water depth (1939 m) and distance between sea level and rotary table (28.5 m).

All data gathered during mud-gas monitoring by the GC-FID, MCIA, and PGMS were transferred together with the drilling parameters to the SSX database, where all data were synchronized. Unfortunately, the resolution of the PGMS data included in the SSX database was far too low for on-line monitoring purposes, and therefore, raw PGMS data were used. Data were recorded at all times, including periods where no drilling was conducted and/or mud-gas flow was absent. During data processing, time periods where Lag Depth L did not change were removed. In addition, for hydrocarbon gases, concentrations ≤0.0001% were not considered during data evaluation.

Background control, quality checks, and comparison of different sampling techniques

Following the Expedition 337 Scientists (Expedition 337 Scientists, 2013), several potential problems can arise in mud-gas monitoring, which made it necessary to conduct different tests during geochemical mud-gas monitoring.

Background control. In the drilling mud, formation gases usually mix with atmospheric gases already present in drilling mud as well as with remaining gases from previous gas extraction. Consequently, drilling mud was sampled from the tank regularly, and the hydrocarbon gas component was measured on board the ship. The mud samples were taken with 50 mL plastic vials, which were completely filled and sealed with a plastic cap. After the sample was transferred to the laboratory, a subsample was taken and placed into a 20 mL glass vial and immediately sealed with a silicon septum and metal crimp cap.

The headspace analysis was carried out using an Agilent Technologies G1888 Network Headspace Sampler, where the sample was heated at 70°C for 30 min before an aliquot of the headspace gas was automatically injected into the GC-FID. Unfortunately, background concentrations of nonhydrocarbon gases could not be determined with the available instrumentation. The results of the background checks are displayed in Figure F19. Although the variations in the data follow the ones found in the sampled gas (see “Geochemistry” in the “Site C0002” chapter [Strasser et al., 2014b]), the overall background concentrations of hydrocarbon gases were, with up to 1152.92 and 2.72 ppm for methane and ethane, respectively, too low to have a significant effect on the sample gas measurements in the upper part of the borehole. For the lower part of the borehole, however, an effect cannot be excluded because of the overall low gas concentration (see “Geochemistry” in the “Site C0002” chapter [Strasser et al., 2014b]). Although propane was absent during the background checks, concentrations of up to 17.00 ppm of iso-butane were found, which is relatively high and might have influenced the iso-butane concentrations in the sampled gas.

Dehydrator. The dust remover and dehydrator module (CFP-8000, Shimadzu Corp., Japan) is a possible source for air contamination and could cause fractionation, both with respect to gas contents and their isotopic composition (Expedition 337 Scientists, 2013). During Expedition 337, gas standards were measured with and without passing through the gas dehydrator, and consequently, no further analysis was applied here. According to the Expedition 337 Scientists (2013), for methane the effect of the dehydrator was within the analytical uncertainty. The methane content of the dried gas was 3% lower and the δ13C values were 0.4% more positive when compared to the unfiltered gas. By contrast, for the PGMS, mist and moisture affected H2, O2, He, and CO2 concentration calculations. H2 and O2 concentrations in the wet gas were both 2% higher than those in the dried gas, whereas He and CO2 concentrations in the dried gas were 67% and 84% higher than in the wet gas.

Verifying results obtained from the MCIA. The MCIA for on-line mud-gas monitoring was used for the first time in the history of scientific ocean drilling during the Expedition 337. In order to confirm its accuracy, the Expedition 337 Scientists took samples of gas and will analyze them on shore by isotope ratio monitoring gas chromatography–mass spectrometry (Expedition 337 Scientists, 2013). Consequently, no further subsampling was done during Expedition 338.

Check for air contamination. A check for possible leaks and consequent air contamination was carried out (see “Appendix B” for details). For this purpose, the mud trap between the degasser and the safety valve was removed from the degasser. The standard hydrocarbon gas used for this experiment was the same as the one used for calibration of the PGMS and GC-NGA and was introduced in the mud trap and transferred to the field laboratory. The test showed that no leakage was present between the mud trap (first feature after the degasser) and the PGMS (last feature after the degasser). Consequently, the source of air contamination might be found in the configuration of the degasser itself, and thus, all samples that were taken during this expedition might be affected. When looking at the test results more closely, it further seems that the concentrations were also influenced by pump rate and pressure (the best results were obtained at 1.0–1.5 mL/min and 0.1 MPa). This was probably also related to the minimum flow rates necessary for the different instruments. Future expeditions need to carefully address this issue.

Sampling and analyzing gas samples from core

During riserless drilling, gas samples were obtained from cores. When a core came into the core cutting area, the temperature of the core liner was measured using a handheld infrared camera (FLIR Systems ThermaCAM SC640) in order to check the presence of gas hydrates in sediment inside the core liner. If sediments contain gas hydrates, their endothermic dissociation leads to low-temperature anomalies in the core liner. When such anomalies were detected, sediment in the core liner was immediately separated into a section and stored in a deep-freezer at –80°C for shore-based analyses.

When a void space was observed in a core liner, a void gas sample was collected from the void space by inserting a gas-tight syringe into the core liner. A 5 mL aliquot of the void gas in the gas-tight syringe was transferred to a 20 mL glass vial.

For headspace gas sampling, after a section of sediment for IW analyses was cut from undisturbed sediment in the core cutting area, sediment for headspace gas analysis was taken from a freshly exposed end of the section. About 5 cm3 or about a half the volume of a 20 mL glass vial of sediment was taken using a cut-off plastic syringe or a cork borer in the case of consolidated sediment. The sediment sample was put into two 20 mL vials that had been weighed prior to sampling. During this expedition, in addition to the conventional headspace gas sampling, a method using alkaline solution was also carried out using the following procedures (Expedition 316 Scientists, 2009a). For conventional headspace analysis, the vial was capped with a silicon septum and metal crimp cap as soon as possible after its recovery. For the additional analysis using the sample mixed with alkaline solution, 5 mL of 1 M NaOH was added to the vial, and it was immediately sealed with a septum and crimp cap. The vial containing NaOH solution was shaken for 2 min using a tube mixer and was left to stand for at least 24 h at room temperature prior to measurement of isotopic composition of methane. Analysis of isotopic composition of methane using MCIA was subjected to the interference of carbon dioxide so that carbon dioxide in a sample bottle was absorbed into an alkaline solution for a precise determination of isotopic composition of methane.

Offline hydrocarbon gas analysis using a GC-FID

Gas analysis was carried out using an Agilent 6890N GC equipped with an FID. Calibration of the GC was conducted using a standard gas containing low-molecular hydrocarbons (methane, ethane, propane, iso-butane, and n-butane). The measurement scheme was preprogramed as follows: a sample bottle was set on an autosampler attached to the GC and was heated at 70°C for 30 min in the autosampler oven. Subsequently, helium carrier gas was introduced into the bottle and the sample was transferred to a sample loop in the autosampler. The gas in the sample loop was injected into the GC. Hydrocarbon concentrations in IW were calculated using the following equation:

CH4 = [χM × Patm × (VEWS Bulk)]
/(R × T × Wc × WS),
(8)

where

  • χM = molar fraction of methane in the headspace gas (obtained from GC analysis),
  • Patm = pressure in the vial headspace (assumed to be the measured atmospheric pressure when the vials were sealed),
  • VE = volume of the empty vial,
  • WS = weight of the whole sediment sample (after sampling, weight of the vial containing sediment was measured, and the weight of the empty vial measured prior to sampling was subtracted to calculate the weight of the sediment sample),
  • ρBulk = bulk density of the sediment sample (determined from MAD measurements on nearby samples),
  • R = universal gas constant,
  • T = temperature of the vial headspace in Kelvin, and
  • Wc = water content of sediment (determined from MAD measurements on nearby samples).
Offline analysis of the stable carbon isotopic composition of methane

After the measurement of hydrocarbon content in the headspace gas, the carbon isotope ratio of methane in headspace gas was determined using the MCIA. For MCIA analysis, methane content must be <1 ppm (see “On-line analysis of the stable carbon isotopic composition of methane”). Therefore, a portion of the headspace gas was diluted with zero-air (free from hydrocarbons) prior to introducing the sample gas into the MCIA. The needle attached to the syringe was penetrated and 15–20 mL zero-air was transferred into the vial. Then, the headspace gas was sucked into the gas-tight syringe up to 25 mL volume and the diluted gas was injected into the MCIA to measure the isotopic composition of methane. The resulting isotope ratio is given in the same manner as described in “On-line analysis of the stable carbon isotopic composition of methane.”