IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.319.102.2010

Geochemistry

Mud gas monitoring

Expedition 319 was the first IODP cruise that conducted real-time drilling mud gas extraction and analysis in addition to sampling for postexpedition studies (e.g., noble gas and stable isotopes) using a system developed at the Deutsches GeoForschungsZentrum (GFZ) (Erzinger et al., 2006). This section provides a short description of the online drilling mud gas extraction system and analyses performed during Expedition 319. More detailed explanations of technical aspects of scientific drilling mud gas monitoring can be found in Erzinger et al. (2006).

Gas extraction

The gas dissolved in the returning drilling mud was extracted under a slight vacuum using a custom built water-gas separator (Fig. F13). The separator consists of a steel cylinder with an explosion-proof 220 V electrical motor above the cylinder, which drives a stirring impeller mounted inside the cylinder. In contrast to commercially built gas separators, the separator used during Expedition 319 is not permanently flushed with air because this could lead to contamination by atmospheric gases prior to sampling and analysis. This is important for gas species that might be only slightly enriched in the formation compared to their atmospheric concentration (e.g., helium). Nevertheless, when making a new pipe connection, the mud flow was stopped and the separator was exposed to air for a short time.

A membrane pump was used to create a vacuum and pump the extracted gas through a polyethylene tube (3 mm inner diameter, 9 mm outer diameter) into a laboratory trailer, which was installed on deck (Fig. F13). The flow rate was adjusted to between 1 and 2 L/min, which resulted in a traveltime of <2 min. Wiersberg and Erzinger (2007) showed that because of the short gas traveltime, diffusion loss along the gas line is negligible. The gas space in the separator head depends on the mud level and the applied vacuum, which was ~20 mbar below ambient air pressure. When the amount of extracted gas was small, the mud would rise in the separator tank because of the vacuum and enter the gas line. In this case, a condensation trap placed directly behind the separator acted as a security valve and opened the separator to the atmosphere.

During drilling of the 12¼ inch hole from 703 to 1510 m MSF (drilling Phase 2 operations; see C0009_T1.XLS in GEOCHEM in "Supplementary material"), the gas separator was placed in the bypass between the gumbo shakers and close to the outlet of the mud flow line in order to minimize air contamination and degassing of the drilling mud before reaching the separator (Fig. F14). With an average ROP of 20–30 m/h, it took ~1.5–2 h to make a new pipe connection, during which time the degasser was exposed to air. With a separator headspace volume of ~30 L and gas pumping rate of 1–2 L/min, it took up to 1 h to fully exchange the gas volume after air flushing, particularly when the gas composition of drilling mud significantly differed from air. The low gas exchange rate in the separator headspace volume in combination with a high gas load of the drilling mud resulted in smoothing of peaks in gas composition. The uncertainty in depth of origin is estimated at ±20 m during this drilling phase.

To obtain data at higher spatial resolution, a smaller separator was chosen (headspace volume ~10 L) during coring (drilling Phase 3) and hole opening procedures (drilling Phase 8 operations; see C0009_T2.XLS in GEOCHEM in "Supplementary material"). Furthermore, the separator was moved from the bypass between the gumbo shakers to the mudline in front of the shale shakers (Fig. F49 in the "Site C0009" chapter). Between the gumbo shakers, the separator could only be placed directly in front of the mudline outlet pipe, where a strong mud flow caused an effect similar to a water-jet pump. When drilling strata with low gas concentration, the separator did not produce gas but instead drew it from the line. The reduced headspace volume and different positioning of the separator improved the resolution of gas sample locations to ±10 m during Phase 8, to ~1260 m MSF. Below ~1260 m MSF, the ROP was significantly increased and the gas-loaded drilling mud was not circulated out, resulting in consistently high background methane. In addition, because of the lower mud weight during Phases 2 and 3 relative to Phase 8, the gas concentrations are higher during these phases. Commercial safety gas monitoring during the same period shows the same trend of lower gas concentrations during Phase 8. Although gas concentrations are different during the two phases, trends with depth are comparable.

Gas analysis

In the laboratory trailer (Fig. F13, right), residual water gas was condensed in a trap installed behind the membrane pump before the gas entered the analytical devices. Prior to gas analysis, the dried gas flowed through a sampling device. This sampler collected up to four gas samples from the gas line in glass cylinders and copper tubes for postexpedition laboratory and isotope analysis. The sampler is remotely controlled by the mass spectrometer; a remote signal is given when a preset threshold concentration of a given type of gas is exceeded. The sampler is equipped with a bypass line to make sure that gas flow to the analytical devices continues even after all samples are taken.

After passing through the sampler, the gas was analyzed by a gas mass spectrometer and a gas chromatograph (GC). Concentrations of N2, O2, Ar, CO2, CH4, He, and H2 were determined using an OmniStar (Pfeiffer Vacuum, Germany) quadrupole mass spectrometer (QMS). The QMS requires a gas flux of ~30 mL/h of gas. A complete QMS analysis with detection limits of 1 parts per million by volume (ppmv) for He, H2, CH4, and Ar, as well as 10 ppmv for O2, N2, and CO2 was achieved with this setup after an integration time of 16 s. However, a 1 min sampling interval was chosen to reduce the amount of data produced. Hydrocarbons (CH4, C2H6, C3H8, i-C4H10, and n-C4H10) were analyzed only during Phase 8 at 10 min intervals with an automated standard field GC (SRI 8610) equipped with a flame ionization detector (FID). Detection limits for the hydrocarbons are ~1 ppmv. In addition, electrical conductivity, pH, and temperature of the returning drilling mud were determined continuously at 5 min intervals during all drilling phases.

After acquisition, data were corrected for artifacts (e.g., connection gas, trip gas) and correlated to the lag depth that takes into account the traveltime of the drilling mud from the drill bit to the surface. Lag time and resulting lag depth were calculated by the mud logging company from the borehole volume and mud pump rate (strokes) and cross-checked by carbide tests twice during drilling operations.

Several analytical instruments (GC, radon detector, one of two mass spectrometers) as well as one hydrogen generator and one computer failed before and/or during drilling. For some of those incidents, a link between failure and apparent fluctuation in the power supply is likely, suggested by simultaneous failure. Hence, a clean power supply independent from rig power is highly recommended for possible future use of this experimental setup.

Organic geochemistry

Sample preparation

Inorganic carbon, TC, and TN were determined for 111 cuttings samples (1–4 mm fraction) from 1037.7 to 1592.7 m MSF and for 34 core samples from 1509.8 to 1591.5 m CSF. Prior to sample analysis, cuttings material (~10 cm3) was washed with seawater and deionized water, sieved, freeze-dried, and ground to powder (see "Cuttings handling"). Core material (~10 cm3) was only dried and ground.

Inorganic carbon

Inorganic carbon was measured using a Coulometrics 5012 CO2 coulometer. Approximately 10–12 mg of freeze-dried powder was weighed and treated with 2M HCl. The released CO2 was titrated, and the change in light transmittance was measured with a photodetection cell. The weight percent of calcium carbonate was calculated from inorganic carbon content, assuming that all extracted CO2 was derived from dissolution of calcium carbonate according to the reaction

CaCO3 + 2H+ → CO2↑ + H2O + Ca2+, (8)

by the following equation based on molecular weight ratio:

CaCO3 = 8.33 × inorganic carbon. (9)

It is assumed that all carbonate minerals were composed of CaCO3. The standard deviation based on 63 reference measurements was less than ±0.71%.

Total carbon and nitrogen

TC and TN concentrations were determined using a Thermo Finnigan Flash EA 1112 CHNS analyzer. Calibration was performed using the synthetic standard sulfanilamide, which contains C (41.81 wt%) and N (16.27 wt%). About 10–20 mg of freeze-dried sample powder was weighed and placed in a tin container for carbon and nitrogen analyses. The samples were combusted at 1000°C in an oxygen stream. NOx were reduced to N2, and the mixture of CO2 and N2 was analyzed by a GC equipped with a thermal conductivity detector. Total organic carbon was calculated by subtracting inorganic carbon from TC. Standard deviation of carbon and nitrogen is less than ±0.1%. The analytical accuracy for carbon was determined with 0.188 ± 0.014 wt% for N and 0.1755 ± 0.014 wt% for C, using Soil NCS reference material with 0.195 wt% N and 0.176 wt% C.

Interstitial water geochemistry

Interstitial water extraction

Interstitial water samples were obtained from 35–46 cm long whole-round core sections. Whole-round sediment samples were cut from each core at the core cutting area and then immediately transferred to the CT laboratory to be scanned for important features such as lithologic boundaries and structures and to confirm that the section was relatively homogeneous and unfractured. The sample was then taken to the quality assurance/quality control laboratory and was extruded from the core liner in a nitrogen-flushed glove bag. In the glove bag, each sediment sample was cleaned to prevent drilling contamination (e.g., seawater, drilling fluid). The cleaned residual sediment was placed in a Manheim-type titanium squeezer (Manheim, 1966). If all pore water squeezers were in operation, the samples were stored at 4°C prior to extrusion; in all cases, squeezing started within 24 h of receiving core on deck.

Because of the small amount of water extracted from the samples, two squeezers were filled with the maximum amount of sample material and then manually presqueezed to obtain filter cakes. Filter cakes were then placed together in one squeezer for automatic squeezing. The squeezing procedure comprised 10 min of squeezing with 15,000 lb (pound–force), 100 min of squeezing with 20,000 lb, and final squeezing with 24,000 lb until 4 cm3 of pore water was extracted. Squeezing times are reported in Table T12 in the "Site C0009" chapter.

Collected interstitial water was passed through two rinsed filter papers fitted with 2–4 300 mesh stainless steel screens at the base of the squeezer. Fluids from the squeezing process were then passed through a 0.45 µm disposable filter into an acid-washed (10% HCl) 50 mL plastic syringe. Because of the small volumes obtained, interstitial water was sampled for shipboard analyses only. High-density polyethylene (HDPE) sample vials intended for minor and trace element analysis were cleaned by immersion in 55°C 10% trace metal grade 12N HCl for a minimum of 24 h, rinsed with Millipore 18.2 MΩ·cm Type 1 ultrapure (Milli-Q) water, and then dried in a class 100 laminar flow clean hood. All samples designated for shipboard minor and trace element analysis were acidified with 6N HCl with a ratio of 4 mL of 6N HCl per liter of sample for at least 24 h. This was followed by either inductively coupled plasma–atomic emission spectroscopy (ICP-AES) or inductively coupled plasma–mass spectrometry (ICP-MS) analysis to dissolve any metallic oxide precipitates that may have formed after squeezing.

Interstitial water analysis

Because of the limited amount of water extracted from core (~4 cm3 per sample), not all standard shipboard analyses could be performed. Determination of alkalinity and pH was not carried out, as these measurements alone would have required 3 cm3 sample volume. Interstitial water samples were analyzed for salinity by an Atago RX-5000α refractometer. The refractive index was converted to salinity based on repeated analyses of International Association of Physical Sciences of the Oceans (IAPSO) standard seawater. Precision for the refractive index measurement was ±0.0004. Chlorinity was determined from the interstitial water samples using a Metrohm autotitrator and silver nitrate solution as a titrant. The relative standard deviation for chlorinity is ±0.15%. The average precision of the chlorinity titrations, expressed as 1σ standard deviation of means of multiple determinations of IAPSO standard seawater, is ±0.15%.

Sulfate and bromide concentrations were measured with a Dionex ICS-1500 ion chromatograph (IC) using subsamples that were diluted 1:100 (10 µL in 990 µL) with Milli-Q water. This dilution provided quality peak detection for chloride, bromide, and sulfate. Because of the high chloride concentration, these data were used only to check the quality of the dilution step. 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. The average precision of the sulfate and bromide analyses were ±2.3% for sulfate and ±5.7% for bromide, expressed as 1σ standard deviation.

The concentrations of certain major cations (Mg2+, Ca2+, Na+, and K+) were determined with an ICS-1500 IC equipped with a cation exchange column. Samples were acidified with HCl (Tamapure-AA-110 grade) in 0.4% of the sample volume and then diluted by a factor of 200 using a Hamilton diluter. The samples were then 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 every eight samples. In order to determine the concentration of each ion, measurements of a standard material were performed to obtain calibration curves from the measured peak area and the known concentration. A standard solution containing Li+ (0.5 mg/L), Na+ (2 mg/L), NH4+ (2 mg/L), K+ (5 mg/L), Mg2+ (5 mg/L), and Ca2+ (5 mg/L) was diluted to five standard solutions (10%, 25%, 50%, 75%, and 100%). The resulting precision for the measurements was ±0.4 for Na+, ±1.7% for K+, ±2.8% for Mg2+, and ±1.3% for Ca2+.

Minor element B, Ba, Fe, Li, Mn, Si, and Sr concentrations were determined by ICP-AES (Horiba Jobin Yvon Ultima2). To analyze these elements, 0.5 mL of sample solution was added to 9.5 mL of ultrapure bidistilled nitric acid solution and spiked with 10 ppm Y solution as an internal standard. Because of the high concentration of matrix salts in the interstitial water samples at a 1:20 dilution, matrix matching of the calibration standards is necessary to achieve accurate results by ICP-AES. A matrix solution that approximated IAPSO standard seawater major element concentrations was prepared from the following salts in 1 L of Milli-Q water acidified with 4 mL of optima-grade 6N HCl: 26.90 g NaCl, 3.81 g MgCl2, 1.00 g CaCO3, and 0.75 g KCl. Blanks were measured using a dilution of 1% nitric acid solution in the Y solution where only the slope of the calibration curve was used for quantification.

A stock standard solution was prepared from ultrapure primary standards (SPC Science PlasmaCAL) in the 1% nitric acid solution, and then diluted in the same 1% ultrapure nitric acid solution used for pore water samples to concentrations of 50%, 25%, 10%, 5%, and 1%. A 1.25 mL aliquot of each stock solution was added to 8.75 mL of matrix solution to generate a series of standards that could be diluted using the same method as the samples for consistency.

The final matrix-matched 100% standard solution contained the following concentrations of elements: B = 3000 µM, Li = 400 µM, Si = 1000 µM, Mn = 50 µM, Fe = 50 µM, Sr = 400 µM, and Ba = 200 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. Relative standard deviations were ±0.8% for B, ±0.5% for Ba, ±2.3% for Fe, ±3.1% for Li, ±1.5% for Mn, ±2.0% for Si, and ±1.4% for Sr.

V, Cu, Zn, Mo, Rb, Cs, Pb, and U were analyzed by ICP-MS (Agilent 7500ce ICP-MS) equipped with an octopole analyzer to reduce isobaric interferences from polyatomic and double-charged ions. To calibrate for interferences between some of the transition metals (V, Cu, and Zn) and some major element oxides, solutions containing these elements were prepared with concentrations similar to IAPSO standard seawater values. These solutions were analyzed at the beginning of each measurement, and an interference correlation was applied based on the average ion counts per second (cps) measured on the standard solutions divided by the abundance of the interfering elements. This ratio was multiplied by the known concentration of the major ions in the samples based on previous analysis, and the result was then subtracted from the measured counts (cps) of the sample.

A 100 µL aliquot of 500 parts per billion (ppb) indium standard was added before further dilution. Aliquots of 150 µL of this sample solution were then diluted with 4.85 mL of 1% HNO3 based on previous determination of the detection limits and the low concentrations of the elements of interest. A primary standard solution was made that matched the maximum range of predicted concentrations based on published results of deep-sea pore fluid compositions in a variety of settings. The composition of the standard is as follows: V = 20 ppb; Cu, Mo, Pb, and U = 40 ppb; Zn = 140 ppb; Rb = 500 ppb; and Cs = 5 ppb.

This primary standard was then diluted with 1% nitric acid solution to relative concentrations of 50%, 25%, 10%, 5%, and 1%. A 0.15 mL aliquot of these standards was then further diluted by addition of 0.15 mL of a 560 mM NaCl solution and 4.7 mL of 1% HNO3 solution to account for matrix suppression of the plasma ionization efficiency. The 25% standard was diluted accordingly and analyzed together every eight samples throughout each analysis series for precision and in order to correlate the results from different analysis dates. Blanks were also analyzed 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 optima-grade 6N HCl per liter. The average precision of multiple determinations of the 10% ICP-MS standard was ±1.7% for 51V, ±3.6% for 66Zn, ±1.2% for 85Rb, ±10.5% for 95Mo, ±0.5% for 133Cs, and ±3.6% for 238U.