PP Handbook Peter Blum November 1997

7/22/2019 PP Handbook Peter Blum November 1997

1/117

11PP Handbook , Peter Blum , November, 1997

1. INTRODUCTION

1.1. Objectives of Physical Properties Measurements

Physical properties of rocks and sediments are indicators of composition,

formation, and environmental conditions of the deposits. Some physical properties

can be measured rapidly and easily at high spatial resolution (core logging) and

serve as proxies for processes such as paleoclimatic changes. Physical properties

data are usually well defined and quantitative, which helps constrain the complex

mineralogical and fluid systems in rocks and sediments. They are used

increasingly by a wide scientific community for various scientific objectives. For

these reasons, physical properties data form the bulk of all core data collected on

board theJOIDES Resolution on each leg.

In soft and semiconsolidated sediment sections, physical properties data servemostly as proxies for sediment composition, which is controlled by provenance,

depositional and erosional processes, oceanographic and climatic changes, and

postdepositional processes such as consolidation, and early diagenesis. In

consolidated sediments and igneous rocks, diagenetic processes, including

cementation, major lithological changes, and major faults, tend to dominate many

physical properties. Hydrothermal circulation can be detected in sediment and rock

environments by using physical property measurements.

A major application of data collected at small sampling intervals (a few

centimeters), such as magnetic susceptibility, color reflectance, gamma-ray

density, and natural gamma radiation, is for core-to-core and hole-to-hole

correlation and for correlating core data to wireline log data. These correlation

procedures are essential for stratigraphic studies, and some of the most important

ocean drilling projects are unthinkable without the high-performance acquisition

of physical properties data.

1.2. Shipboard Laboratory Stations and Sampling

OVERVIEW

After cores arrive on deck they are cut into 1.5-m-long sections and stored in racks

for temperature equilibration. The first measurement station is the multisensor

track (MST), where the whole-core sections are loaded on a motorized core

conveyor boat for the automatic measurement of gamma-ray density,

compressional (P-)wave velocity, magnetic susceptibility, and natural gamma

radiation. The MST is used most effectively with cores completely filled with soft


2/117

12 PP Handbook , Peter Blum , November, 1997

to semiconsolidated sediments that were retrieved with the advanced hydraulic

piston corer (APC). Intact sedimentary or igneous rock cores cut with the extended

core barrel (XCB) or rotary core barrel (RCB) also give good MST measurements.

Coring disturbance such as severe biscuiting (typical for XCB cores) and

fracturing (typical for RCB cores) associated with torquing significantly reduces

the accuracy and usefulness of MST measurements, sometimes to a degree that

MST measurements should not be performed.

For soft sediment cores, the second station is the thermal conductivity station,

where needle probes are inserted into the whole cores. Next, the cores are split

either with a wire (soft sediment) or with a saw. The half-cores are designated as

archive-half cores and working-half cores. Figure 11 shows the relative core

orientation conventions established to place core measurements, particularly

paleomagnetic data, in a geographic reference frame using absolute core

orientation measurements when the core is cut. The same conventions are used for

other physical properties measurements that can be performed in multiple

directions and that may reveal anisotropy (e.g., acoustic measurements) or for

structural measurements. The archive-half cores are preserved in a pristine

condition whereas the working-half cores are available for measurements that

physically disturb parts of the cores and for theremoval of specimens for shipboard

as well as shore-based studies.

The archive-half core is used for the visual core description, paleomagneticmeasurements using the cryogenic magnetometer, noncontact color reflectance

measurements (to be implemented), and photography. A track system is in

development that will measure the two physical properties of magnetic

susceptibility and color reflectance along with the acquisition of color images of

the core surface. After core photographs have been taken, the archive-half cores

are stored in plastic tubes and refrigerated.

UP

Workinghalf

y(90)

z

Split-coreface

x(0)

(Double line)

UP

Archivehalf

z

Split-coreface

-x(180)

(Single line)

Figure 11 Core orientation conventions.


3/117


The working-half core is used for the measurement of color reflectance (the

present mode of manual operation requires contact with sediment), P-wave

velocity by using probes that are inserted into the soft sediment, vane shear

strength by inserting a miniaturized vane into the sediment, and similar strength

measurements with the hand-held Torvane and penetrometer devices. Half-core

pieces of rocks are used for the measurement of thermal conductivity by using the

half-space needle probe. In the future, a gamma-ray densiometer will be added

to the working-half station. Along with the use of a caliper (associated with the P-wave system on this track) gamma-ray densities may be more accurate and precise

than those obtained currently from the MST.

For the final physical properties measurement, specimens are extracted from the

working-half core to determine moisture content and average mineral density

(MAD station). P-wave velocity can also be determined on specimens of

sedimentary or igneous rock extracted using a parallel-blade or cylindrical saw.

The working-half core then proceeds to the sampling table where one to three

individuals extract specimens for analysis on shore. The sampling voids are filled

with Styrofoam, and the working-half core is stored in plastic tubes and

refrigerated along with the archive-half core.

MULTISENSOR TRACK (WHOLE-CORE MST) STATION

Measurement Systems The MST is an automated core conveying and positioning system for logging core

physical properties at small sampling intervals. At present, the MST system

includes the following measurements:

gamma-ray attenuation densiometry (GRA)

P-wave velocity logging (PWL)

magnetic susceptibility logging (MSL)

natural gamma ray (NGR) measurements

The MST is one of the most routinely used devices onboard theJOIDES

Resolution. No other shipboard instrument produces a comparable amount of core

data, and the MST data set is among the most widely used ODP data and

represents a worldwide standard of core analysis. The MST is designed to handle

the sampling of whole cores automatically, and all measurements except the PWL

can also be used on split cores and for measurements on individual core

specimens. A new flexible, intuitive control interface was implemented in 1996.

Sampling One of the most useful new features is the improved sampling parameter interface.

The user can set sampling intervals and periods for all sensors and the program

returns the calculated total measuring time for a core section based on anoptimized measuring sequence. A graphical display shows the sampling points

with depth. Typically, the time permissible for a whole core (typically seven core

sections) is about 1 hr on legs with high core recovery (about 4 km of core or

more). Therefore, if full-time attention is given to the MST, about 10 min can be

allowed for measuring one core section. An overview of useful sampling

parameter settings is given in this section. More data and information are presented

in the individual sensor sections as appropriate.


4/117


When selecting sampling intervals, consideration should be given to the depth

interval each sensor can resolve (see Table 11). For the GRA and PWL sensors,

the depth intervals are less than 1 cm, for the MSL loop it is about 4 cm, and for

the NGR it is about 15 cm. Because the sensitivity of the MSL and NGR sensors

decreases away from the center of the sensor, better resolution can still be achieved

by taking measurements at intervals smaller than the intrinsic interval of influence.

Generally, ideal sampling intervals for the GRA, MSL, and PWL are 1 cm and

should not exceed 5 cm. For the NGR, the best depth resolution possible is at about5 cm. Intervals should not exceed 30 cm, which is about the depth resolution of

downhole logging tools.

Sampling periods are directly related to the data quality (precision) particularly for

the nuclear sensors. Because of the high flux provided by the 137Ce gamma-ray

source, 2-s sampling with the GRA is sufficient. The MSL has an internal

integration time of 0.9 s (1.0 range) or 9 s (0.1 range); it should be set at 1 s. The

MST program is best set to 2-s sampling time to allow for minor electronic and

communications delay. The NGR is most sensitive to the sampling period because

of the low intensity and random nature of natural gamma ray emissions. The more

counts are accumulated, the more reliable the signal (the error is is proportional to

N-0.5, where N is the number of counts; see Natural Gamma Radiation chapter

for more discussion). If spectral analysis is attempted to estimate abundance of K,

U, and Th (which is not implemented for routine application yet), at least 1 min

should be counted. (One hour would probably be more appropriate to reduce the

statistical error to a level that would yield a good estimate of K, U, and Th). If only

a total counts signal is desired, as little as 15 s is sufficient in terrigenous

sediments, whereas 30 s should be measured in carbonates. The PWL system takes

five measurements (data acquisitions or DAQs) at each point that are averaged for

the sample and provide a sufficiently precise value.

For optimized sampling parameter settings it is important that intervals and

periods are multiples of each other. This ensures that the idle time of sensors isminimized and data quantity and quality are maximized for a given total core

section scan period. For example, if GRA is set to 2 cm and MSL to 3 cm, one of

the two sensors is partly idle while the other is taking a measurement. It is more

efficient to set both at 2 cm so they measure simultaneously. Similarly, if the core

stops every 1 cm for GRA and MSL measurements and 4 s are required for the

MSL, the GRA sampling period should also be 4 s rather that 2 s because the

additional time improves data quality but it does not require any additional time.

Table 11 MST sampling parameters.

Sensor Sensitivity Interval (cm) Period (s)

interval (cm) Best Typical Maximum Best Typical Minimum

GRA


5/117


A further optimization can be considered for NGR measurements. Rather than

taking a 20-s reading every 20 cm and leaving the other sensors mostly idle during

that time, a 5-s reading can be taken every 5 cm, simultaneously with the other

readings. This shortens total scanning time considerably. To get data quality

(statistical error range) equivalent to a 20-s counting time, the user simply runs a

moving average with a four-point window on the data.

THERMAL CONDUCTIVITY (TC) STATION

Measurement Systems Thermal conductivity is the only property measured at this station. Two systems

are available currently:

Thermcon-85 system customized for ODP use and

new TK04 system not customized for ODP.

A project plan exists to replace these with a fully integrated system that would

incorporate the best features of both existing systems. However, no resources have

been allocated yet.

Soft-sediment cores are measured before they are split because the larger volume

of material surrounding the needle probe reduces geometrical problems (edgeeffects). If the core material is too hard to be penetrated by the needles without

excessive force, thermal conductivity is measured on working-half core pieces

using the half-space needle probes.

Sampling Given the minimum time available until a soft sediment core must be split (about 1

hr), at least 5-10 measurements can be performed (1- to 2-m sampling interval).

This is usually sufficient because thermal conductivity variations are strongly

proportional to, but less sensitive and less precise than, bulk density

measurements. Density can be used as a proxy and calibrated against a limited

number of thermal conductivity measurements if higher spatial resolution is

required.

ARCHIVE-HALF CORE LOGGER (A-LOGGER, TO BE IMPLEMENTED)

Measurement Systems

(to be implemented)

The archive-half core logger is under development and scheduled for deployment

later this year (1997). It will include the following measurement systems:

color line-scan images,

color reflectance spectrophotometry and colorimetry, and

magnetic susceptibility.

The main goal for this development is to acquire color images of the cores (not

discussed in this note) and to automate the routine acquisition of visible light colorreflectance measurements. In addition, the spacial resolution and sensitivity of

magnetic susceptibility logging will be improved with a point-sensor that

requires contact with the core surface. Although line scans are truly noncontact

and nondestructive (i.e., ideally suited for archive-half logging),

photospectrometry and magnetic susceptibility require contact with the core

surface and these implications still must be evaluated. These measurements may

haveto be obtained from working-half cores.


6/117


Present Measurement

System

The present proto-A-logger consists of a manually operated track for color

reflectance measurements. Measurements are usually performed on working-half

cores because imprints are left on the core surface from the manual operation. A

simple computer program writes the data directly to disk and assists the operator

further by incrementing sampling intervals automatically.

Sampling Color reflectance should be measured at the smallest intervals possible because it

is very sensitive to compositional changes. Variations in color reflectance serve asan excellent proxy for detailed correlation and compositional interpretation. A

measurement with the Minolta spectrophotometer covers an 0.8-cm-diameter area.

The manual mode sampling intervals used by shipboard scientific parties are 2 to

20 cm. With the future automated system, intervals should be set at 1 cm or less.

WORKING-HALF CORE STATION (W-LOGGER)

Measurement Systems The working-half core station is semiautomated currently. It includes the following

measurements (Figure 12):

P-wave velocity with the PWS1, PWS2, and PWS3 systems,

Shear strength using the automated vane shear (AVS),

Shear strength using the manual Torvane (TOR), and

Compressional strength using a pen-size penetrometer (PEN)

A component analyzer is available for resistivity measurements, but these

measurements are not supported by ODP at present. Users are required to provide

their own probes, perform their own calibrations, and develop their own

procedures.

P-wave velocity Shear strengt

AVSPWS1 PWS2 PWS3

y xz

n/a any directionn/a

Split-core:

Measurement direction

Specimens:

z-y plane

n/a

Figure 12 Schematic view of the semiautomated instrumentation on theworking-half core track. n/a = not applicable.


7/117


Sampling Sampling intervals for these measurements are mainly a function of available time

at a given core recovery rate and how much core destruction (particularly using the

AVS system) is permissible. The minimum sampling frequency on soft sediment

cores is one per core section; a more typical sampling rate is two per section (75-

cm sampling interval). If numerous measurements are desired on specimens that

must be extracted from the working-half core or that disturb the core, the ODP

staff representative must be consulted.

Whenever possible, the same sampling location should be coordinated for P-wave

velocity and strength measurements, as well as for subsequent extraction of

specimens for moisture and density measurements, carbonate, X-ray diffraction

(XRD), and/or magnetic rock properties measurements.

For velocity measurements on split cores in liners, no sample preparation is

necessary. An undisturbed interval is chosen for the measurement. For

measurements on specimens that require two parallel faces to obtain optimum

values, there are several ways to obtain such samples. In semiconsolidated

sediment, use a spatula or knife to cut a cube of approximately 20 cm3. For

indurated sediment, use a hammer and chisel or the Felker saw. The Torrance

double-bladed saw cuts good parallel faces. The easiest way to obtain a velocitysample in hard rock is to drill cylindrical minicores. These samples are

particularly useful for sharing with the paleomagnetics laboratory (note the

orientation when taking the sample).

MOISTURE AND DENSITY (MAD) STATION

Measurement Systems At the MAD station, the following are measured:

wet-bulk mass and dry mass of the same specimen (for moisture content

and density) and

volume of dry (and optionally wet-bulk) specimen using gas pycnometry.

From these measurements, basic phase relationships such as porosity, bulk density,

grain density, dry density, and void ratio can be calculated. At present, a

convection oven is used to dry the specimens. Ideally, a freeze-dryer should be

used to avoid excessive extraction of interlayer water from clay minerals,

particularly smectite.

Sampling Sampling is typically 1-2 specimens per section, 10-mL volume per specimen. If

possible, the same sample interval should be used as for strength and/or P-wave

velocity measurements. Where numerous lithologic changes occur, denser

sampling may ensure measurements from all significant lithologies throughout the

core. Where cyclic changes in gamma-ray density are observed, a denser samplingprogram over a characteristic interval may be desirable. In XCB and RCB cores,

which commonly show the biscuiting type of disturbance, particular care should be

taken to sample undisturbed parts of the core sections and to avoid the drilling

slurry.


8/117


1.3. New Shipboard Data Management Environment

BACKGROUND

In the early 1990s, the JOIDES advisory structure, through input from shipboard

participants identified the need to design and implement a new database system on

the ship as well as on shore. The complexity and level of productivity of the

shipboard data acquisition environment made this a multiyear, multimillion dollar

project. The physical properties laboratory was the first shipboard laboratory to be

integrated into the new data management environment once the basic operational,

curatorial, and depth calculation functions were redefined and implemented.

The process of redefining the entire ODP data structure offered the opportunity to

implement more rigorous data acquisition, calibration, and control measurement

protocols for physical properties measurements and to give the user access to these

quality control data. A uniform data structure, compatible with the rules of

relational data management, was created wherever possible. Leg 173 (April to

June 1997) was the official acceptance leg for the new data management system,

as described in this first edition of the note.

From the users perspective, the data management system includes the following

components:

data acquisition interfaces and controls,

data upload utilities,

database and data models, and

data access and standard queries.

The following section briefly introduces these components.

COMPONENTS OF SHIPBOARD DATA MANAGEMENT

Data Acquisition

Interfaces and

Controls

DAQ programs are written in various programs depending on the most suitable

software tools and available expertise and hardware at the time and place they

were written. During the past two years, two dominating standards have evolved:

Neuron Data for operational and curatorial functions and descriptive data types

(excellent for PCs, but performs poorly on Macintosh computers); and Labview

for instrumental data (Macintosh or PC). The Neuron Data applications are

integrated into a common user interface, called the Janus Application. Most

physical properties DAQ programs are written in Labview now, including the

MST control, MAD program, P-wave velocity and vane shear strength on half

cores (PWS, AVS), and control of the Minolta photospectrometer (COL). Thermal

conductivity remains in a state of development, and both available systems

controls are written in QuickBasic.

Data Upload Utilities Once data are acquired and located on a local drive, they must be uploaded to the

Oracle database. Although procedure this could be fully automated and become

part of the DAQ program, it was decided that an interactive user quality control

should separate the two functions. Invalid or erroneous data are frequently


9/117


acquired, particularly on highly automated systems that are operated in a

conveyer-belt mode. The user has the option to delete such data from the local

directory before triggering upload to the database, which avoids excessive editing

within the database, a process that involves significantly more risk and effort.

Data upload utility programs are written in Neuron Data and are closely integrated

with DAQ programs written in Neuron Data. For DAQ programs written in

Labview or another language, a separate data upload utility must be operated. This

is the responsibility of the ODP technical support representative, but scientists may

learn the procedure and operate it themselves.

Database and Data

Models

The new ODP Oracle database is designed specifically for ODPs unique

shipboard environment and user needs. The system includes more than 250 data

tables in a complex relational scheme, capturing data from the initiation of a leg,

through core recovery and curation, physical and chemical analyses, core

description, and sampling. Physical properties alone use 65 tables at present, not

counting related tables for sample identification and depth data shared with other

laboratories, and will involve more than 100 tables once the remaining

measurement systems are integrated. The tables pertaining to a particular

measurement system are presented in the Data Specification sections to help the

user understand how the data are structured and how they can be accessed.

Data Access and

Standard Queries

At this early stage of using the new database, there are three different technical

approaches to data access, and the next few legs will show which is the most

efficient and user-friendly one. The three approaches are referred to as

Janus Application,

Report Access Program, and

World Wide Web Data Access.

The first solution integrates an off-the-shelf reporting utility, Business Objects,

into the Janus Application. Many reports are available through this maininterfacefrom which the user selects a particular report from a submenu.

The Report Access Program (RAP) was written as an alternative manager for the

Business Objects reports. The advantage is that the user does not have to log on to

the Janus Application, which may be somewhat time-consuming, and that access

to and expansion of Business Objects reports and queries could be more efficiently

managed by ODP. This environment allows the user to create special reports using

existing Business Object macros relatively easily.

The third approach is for ODP personnel to write standard queries in C-language

and make them available through a World Wide Web (WWW) browser. This

approach has the advantages that routines are directly suitable for global dataaccess and that accessing data on the ship on the local web may be the fastest

method. It will not provide the freestyle access to the database that Business

Objects in the RAP environment offers to the user. However, recent information

indicates that Business Objects will not continue to be supported on the Macintosh,

which rules out its future use. The third approach will therefore most likely be

fully implemented.


10/117


SAMPLE IDENTIFIERS AND DEPTH CALCULATION

Links to Curatorial

Identifiers

In the relational ODP database, redundancy of information is minimized for

efficient data management. For example, site, hole, core, and section information

is entered in specific tables linked in a logical way, and all measurement locations

in a particular section are linked to the table. Similarly, if a core

specimen is extracted for shipboard or shore-based analysis, the basic curatorial

information is accessed through the table, which is linked to the table, etc. In the physical properties database models presented in the

following chapters, the field alone or with the fields

and are the links to the more specific information in the

appropriate tables. The and tables are listed in Table 12.

Depth Types Depth below seafloor of a core specimen or measurement location can be

calculated in different ways. The standard way is to measure the distance in the

recovered and physically expanded core and add it to the measured drill string

depth datum for the top of the core. This depth scale is known as meters below

seafloor (mbsf). Of course, this is only an approximation to the true depth below

seafloor. Problems inherent in this scale are that the recovered core length may begreater than the interval advanced by the drill string, and some of the material from

this interval was lost between successive cores. With APC material, this results in

apparently overlapping sections between successive cores when in fact there is a

coring gap.

If a complete stratigraphic section is to be constructed, multiple holes are drilled at

the same site and a composite section is developed at the meters composite

depth (mcd) scale. This scale is at the physically expanded state of the recovered

cores and does not match the drilled interval. However, it is a much more

continuous scale that can be fit approximately to the drilled interval using the core-

top data (mbsf) or fit more precisely if good-quality downhole logging data are

available.

There are additional corrections that can be applied to derive a more accurate

approximation to depth below seafloor. These and other depth issues are explained

in detail in a workshop report (Blum et al., 1995), and a technical note dedicated to

these issues will be produced. The redefined concepts are integrated in the new

database, which features a depth map that allows the rapid calculation of any depth

type provided that pertinent data have been acquired and entered (see Table 12 ).


11/117


Standard data queries prompt the user to specify the desired depth type. The

default map type (mbsf) is referred to (map_type_name) as standard.

1.4. Physical Properties Standards

Standard materials used to calibrate instruments are an essential part of the

analyses and should be integrated into the measurement systems accordingly. The

goal is to enter all standards used into database tables so that calibration data and

results can be tracked to the particular standard used at a given time. Our plan is to

populate a table shown in Table 13. The table is

generic enough to accommodate any type of standard, and the value of any

property can be linked to any calibration utility and file in the physical propertiesenvironment. This table may principally include standards from other laboratories

as well.

A table of existing standards is in preparation.

Unfortunately, ODP has not made significant efforts to share standards and

calibration procedures with other core laboratories (with rare exceptions). Such

efforts would benefit ODP as well as other laboratories, and therefore the scientific

drilling community, because reliable and widely endorsed calibration standards for

systems that measure complex natural systems are difficult to find.

Table 12 Database model for some essential s.

Map Type Depth Map Section Sample

map_type [PK1] section_id [PK1] [FK] section_id [PK1] sample_id [PK1]

description map_type [PK2] [FK] section_number location [PK2]

map_type_name sect_interval_top [PK3] section_type sam_section_id . section_id

map_type_date sect_interval_bottom [PK4} curated_length sam_archive_working

map_interval_top liner_length top_interval

map_interval_bottom core_catcher_stored_in bottom_interval

section_comments piece

leg sub_piece

site beaker_id . mad_beaker_id

hole volume

core entered_by

core_type sample_depth

sample_comment

sam_repository . repository

s_c_leg . leg

s_c_sam_code . sam_code

sam_sample_code_lab . s_c_l

Table 13 Physical properties standards database model.

Physical Properties Standard Physical Properties Std Data

standard_id [PK1] standard_id

standard_name property_name

standard_set_name property_description

date_time_commissioned proper ty_value

date_time_decommissioned property_units

lot_serial_number

comments


12/117


2. MOISTUREAND DENSITY

(BY MASSAND VOLUME)

2.1. Principles

PHYSICAL BACKGROUND

Moisture content and mineral density are basic sediment and rock properties that

are determined most accurately through mass and volume determinations. Core

specimens of approximately 8 cm3 are extruded from the working-half core for

this purpose. Moisture content is determined by measuring the specimens mass

before and after removal of interstitial pore fluid through drying. The drying

method is the most critical part of the entire procedure. At present, a convection

oven is used for this purpose for 24 hours at temperatures varying from 100 to

110C. This method is suspected to remove a substantial portion of the interlayer

(hydrated) water from clays such as smectite in addition to interstitial water, which

may result in porosity errors of up to 20%. Alternative methods such as microwave

or freeze-drying have other potential problems and have not replaced the

convection oven.

Moisture content, porosity, and void ratio are defined by the mass or volume of

extracted water (assumed to be interstitial pore fluid), corrected for the mass and

volume of salt evaporated during the drying process (see also ASTM, 1990). The

mass and volume of the evaporated pore-water salts are calculated for a standard

seawater salinity (35), seawater density at laboratory conditions (1.024 g/cm3),

and an average seawater salt density (2.20 g/cm3). Any gases that may be present

are allowed to escape during core retrieval, core splitting, and specimen extrusion.

The volume of a specimen can be measured in three ways:

method A: wet-bulk volume measured with special volume sampler,

method B: wet-bulk volume measured by gas pycnometry, and

method C: dry volume measured by gas pycnometry

Method A is the least standardized method. The device to be used depends on user

preference and can be a simple steel ring (fixed volume, available on the ship) or

some sort of syringe (volume is measured after the sample has been taken). The

advantage of method A, according to some users, is that a larger number of

specimens can be measured than with gas pycnometry in a given time. The

disadvantages are (1) the method works only in soft, non-sticky sediment (mainly

homogenous carbonate oozes to a depth of about 200 mbsf), (2) the volume

measured includes potential cracks or other spaces filled with air, (3) there is no

precision estimate for this method, and (4) there is no standard for this method.


13/117


This method should therefore be used only if there is ample justification, and

measurements must be calibrated with an appropriate number of pycnometry

results.

Methods B and C use the same gas pycnometer. The measurement principle of this

device is briefly described in the following. Gas pycnometry works with pressure

ratios of an ideal gas (helium), which are sensitive to contamination with partial

pressures of other fluids. The material to be measured should therefore be dry. The

ODP database contains thousands of examples from specimens measured with

both method B and method C. A systematic error is clearly discernible in

comparing calculated results, with bulk densities 1%5% too high and grain

densities about 5%10% too high for method B. It is therefore recommended that

only method C be used.

The following relationships can be computed from two mass measurements and

one or two volume measurements. First, if methods B or C are used, the beaker

mass and volume, which are determined periodically and stored in the programs

lookup table, are subtracted from the measured total mass and volume

measurements. If method A is used, only the beaker mass is subtracted (the user

must specify the use of method A in the program). This results in the followingdirectly measured values:

Mb: bulk mass,

Md: dry mass (mass of solids,Ms, plus mass of evaporated salt),

Vb(A or B): bulk volume, method A or method B, and

Vd(C): dry volume = volume of solids, Vs(C), plus volume of evaporated

salt, Vsalt.

Variations in pore-water salinity, s (s = S/1000), and density, pw, that typically

occur in marine sediments do not affect the calculations significantly, and standard

seawater values at laboratory conditions are used:

s = 0.035 (1)

pw = 1.024. (2)

Pore-water mass,Mpw, mass of solids,Ms, and pore-water volume, Vpw, can then

be calculated:

Mpw = (Mb Md) / (1 s) (3)

Ms=Mb Mpw = (Md sMb) / (1 s) (4)

Vpw =Mpw/pw = (Mb Md) / [(1 s) pw]. (5)

Additional parameters required are the mass and volume of salt (Msaltand Vsalt,

respectively) to account for the phase change of pore-water salt during drying. It

should be kept in mind that for practical purposes the mass of salt is the same insolution or as precipitate, whereas the volume of the salt in solution is negligible.

Msalt=Mpw (Mb Md) = (Mb Md) s / (1 s) (6)

Vsalt=Msalt/salt= [(Mb Md) s / (1 s)] /salt, (7)

where the salt density value salt= 2.20 g/cm3 is a value calculated for an average

composition of seawater salt (Lyman and Fleming, 1940; Weast et al., 1985).


14/117


Moisture content is the pore water mass expressed either as percentage of wet bulk

mass or as percentage of the mass of salt-corrected solids:

Wb =Mpw/Mb= (Mb Md) /Mb (1 s) (8)

Ws =Mpw /Ms = (Mb Md) / (Md sMb). (9)

Calculation of the bulk volume for method C and volume of solids depend on the

volume measurement method used:

Vs(A or B) = Vb(A or B) Vpw (10)

Vs(C) = Vd(C) Vsalt (11)

Vb(C) = Vs(C) + Vpw. (12)

Bulk density, b, density of solids or grain density, s, dry density, d, porosity, P,

and void ratio, e, are then calculated accordingly for each method:

b(A,B,C) =Mb /Vb(A,B,C) (13)

s(A,B,C) =Ms /Vs(A,B,C) (14)

d(A,B,C) =Ms /Vb(A,B,C) (15)

P(A,B,C) = Vpw /Vb(A,B,C) (16)

e(A,B,C) = Vpw /Vs(A,B,C). (17)

ENVIRONMENTAL EFFECTS

Core Expansion Cores, particularly sediment cores from a few hundred meters below the seafloor,

expand upon recovery for a number of reasons, which include

elastic recovery,

gas expansion, and

mechanical stretching.

Expansions of solids can be neglected. Pore water expands by about 4% for every

1000 bar (100 MPa) pressure release. This is what the pore water of a seafloor

sample from about 10,000-m water depth would experience, or in ocean drillingterms, what a sample buried by about 2000 m of water and about 3000 m of

sediment would experience. For the bulk of ODP cores, pore-water expansion is in

the order of 1% and therefore negligible compared with the analytical error.

Free gas expands by orders of magnitude, according to the simple relationship

P1V1 = P2V2. A few percent of free gas in the sediment can produce an explosive

sediment-gas mixture that has torn apart plastic core liners on the ship on several

occasions. Most gas escapes before the cores are analyzed and can produce

microfractures, which appear as porosity with methods based on core unit-volume

measurements, such as the gamma-ray attenuation bulk density method.

Mechanical stretching may also cause microfracturing. The MAD methodmeasures the mass and volume of the solid and liquid phases only and is therefore

not affected by this type of artificial porosity. The original contribution of the gas

to in situ porosity cannot be measured with our routine core analysis program.

Composition of

Seawater

Different water masses of the world oceans have different chemical compositions

and physical properties. For the purpose of correcting oven-dried sediment

specimens for the evaporated salt from the pore water, the standard compositions


15/117


after Lyman and Fleming (1940) and salt densities after Weast et al. (1985) are

used (Table 21).

aLyman and Fleming (1940).

bWeast et al. (1985).

Given the uncertainty in regard to the crystalline structure of some evaporated

components, the average density of the standard seawater salt is between 2.10 and

2.24 g/cm3. A value of 2.20 g/cm3 is used routinely for the MAD calculations.

Density of Pore water Density of pore water is a function of temperature (T), salinity (S), and pressure

(P). Equations of state for seawater (Millero et al., 1980; Millero and Poisson,

1981) can be used to illustrate the variability of pore-water density as a function of

these three parameters (Figure on page 5).

Typical salinity values for pore waters are 30 to 40, although more extreme values

exist. At laboratory pressure and temperature, this range of salinity change affects

pore-water density change of less than 1%, which is negligible compared with the

analytical uncertainty. We therefore use a standard value of 35 for all MAD

calculations and leave it up to the user to apply corrections if warranted.

The typical temperature change experienced by nonlithified sediment upon

recovery is from about 100C at depth to few degrees at the seafloor to about 22C

in the laboratory. At standard salinity and laboratory pressure, a 100C change

results in about a 2% change in seawater density. These effect is not figured in our

MAD calcualations because it is close to the uncertainty.

The effect of pressure change on density is of a similar magnitude. For a high-

porosity mud sample (for example, from 100 mbsf) at a water depth of 3000 m, the

pressure release is about 320 bar (32 MPa). According to Figure 21, if the in situ

Table 21 Composition of sea water.

Salt Mass fractiona

(x 103)

Densityb

(g/cm-3)

NaCl 23.476 2.165

MgCl2 4.981 2.316-2.33

Na2SO4 3.917 1.46 (monocl.)

2.68 (orthorh.)

CaCl2 1.102 2.15

KCl 0.664 1.984

NaHCO3 0.192 2.159

KBr 0.096 2.75

H3BO3 0.026

SrCl2SrCl2.2H2O

0.024 2.671 (leaf)

3.052 (cub.)

NaF 0.003 2.08Total 34.481

Weighted average 2.10-2.24


16/117


temperature is about the same as laboratory temperature and salinity is 35, pore-

water density decreases by about 1% upon recovery.

Figure 21 Density of seawater as a function of pressure, salinity, and tempera-ture, using equations from Millero and Poisson (1981). The pressure range from 0to 1000 bar covers most ODP situations. Standard salinity of 35 and two extremesalinities (0 and 70) are plotted as a function of a temperature between 0 to 40C(experimental temperature range of Poisson and Millero, 1981). The arrow indi-cates standard laboratory conditions.

USE OF MAD DATA

MAD data are the only data that provide a direct estimate of porosity and void

ratio and the average density of the minerals. Porosity variations are controlled by

consolidation and lithification, composition, alteration, and deformation of the

sediments or rocks.

MAD data can be used to calibrate the high-resolution gamma-ray attenuation bulk

density data sampled automatically at much smaller intervals than would be

possible for MAD data. If mineral density can be defined with sufficient precision,

GRA bulk density can be expressed as porosity.

990

1010

1030

1050

1070

1090

0 200 400 600 800 1000

Density

(kg/m

3)

Pressure (bar)

0C

S = 7

S = 3

S = 0

20C

40C

0C

20C

40C

0C

20C

40C


17/117


2.2. Moisture and Density System

EQUIPMENT

Balance Mass is determined with two Scientech 202 electronic balances to compensate for

the ships motion. A set of mass standards ranging from 1 to 20 g is used for

calibration and on the reference balance during measurements.

Gas Pycnometer The helium displacement pycnometer with five cells (penta-pycnometer),

manufactured by Quantachrome Corp., employs Archimedes principle of fluid

displacement to determine the volume of solid objects. The five measurement cells

contain custom-fabricated inserts that reduce the chamber to a cylindrical space

that holds exactly one 10-mL Pyrex beaker. The measurement chamber must

contain as little air space as possible to maximize measurement precision. (The

user should also ensure that the Pyrex beakers are filled as completely as possible

with core material.)

Each sample cell of volume VChas an input valve (from the gas tank) and an

output valve (to the pressure transducer). An additional reference cell of volume

VA is located downvent of the sample cells, with an input valve (which separates

VA from the pressure transducer) and a vent valve (Figure on page 7). All cell

volumes must be calibrated periodically (see calibration section). The following is

the operation sequence of the pycnometer during the measurement of a specimen.

The specimen to be measured is placed in a cell of known volume, VC. It is

pressurized, using helium, to an exactly measured pressure of about 18 psi (~120

kPa). The solenoid valve between sample cell and the reference cell of known

volume VA is opened and the helium from the pressurized chamber is ported to the

reference cell. The subsequent pressure in the system is measured. Using the ideal

gas law, the sample volume can be calculated from the pressure ratio. The

following is the sequence of operation (Figure 22).

1. Gas input valves to all five cells are closed (corresponding light-emitting

displays [LEDs] on pycnometer are unlit). The five sample cell output

valves, the reference cell input valve, and the vent valve are open, ensuring

that all cells are at the ambient pressure, Pa.

2. For all cells in use all valves are opened and cells are purged in parallel (for

a 1-min minimum). Cells not being used (not identified by the user) are

isolated by closing the input and output valves.

3. At the end of the purge period, processing begins on the first cell to be used.

(Cells are run in ascending numerical order regardless of the order in whichthey were specified). When a stable ambient pressure is reached, the vent

valve of the reference cell closes and the pycnometer acquires and stores a

zero pressure value.

4. The reference cell input valve is closed to isolate VA from the cell.

Approximately 6 s later, the current sample cell input valve opens and the

cell is pressurized to approximately 17 psi or until 3 min elapse.

5. When the cell pressurization is complete, the current cell input valve closes.


18/117


19/117


under system pressure P1 (about 17 psi). Green lines and cells are under a

reduced system pressure P2.

The sample volume can be calculated using the ideal gas law. By opening the

solenoid valves on one sample cell with volume VC, the system is brought to

ambient pressure Pa after being purged with helium. The state of the system is then

defined as

PaVC= nRTa , (18)

where n is the moles of gas occupying volume VCat pressure Pa,R is the gas

constant, and Ta is the ambient temperature in degrees Kelvin.

When the solid sample of volume VSis placed in the sample cell, the equation can

be written as

Pa (VC VS) = n0RTa . (19)

After pressurizing to about 17 psi above ambient pressure, the state of the system

is given by

P1 (VC VS) = n1RTa . (20)

Here, P1 indicates a pressure above ambient and n1 represents the total moles of

gas contained in the sample cell. When the solenoid valve is opened to connect the

added volume VA to that of the cell VC, the pressure falls to the lower value P2

given by

P2 (VC VS+ VA) = n1RTa + nARTa , (21)

where nA is the moles of gas contained in the added volume when at ambient

pressure.

The term PaVA can be used in place ofnARTa in Equation on page 8yielding

P2 (VC VS+ VA) = n1RTa + PaVA . (22)

Substituting P1 (VC VS) from Equation on page 8 for n1RTa:

P2(VC VS+ VA) = P1(VC VS) + PaVA (23)

(P2 P1) (VC VS) = (Pa P2) VA (24)

VC VS= (Pa P2) / (P2 P1) VA . (25)

Equation on page 8 is further reduced by adding and subtracting Pa from P2 and P1

in the denominator, giving

VS= VC {[(Pa P2) VA / [(P2 Pa) (P1 Pa)]} (26)

= VC+ VA / {1 [(P1 Pa) / (P2 Pa)]}. (27)

Because Pa is zeroed prior to pressurizing:

VS= VC+ VA / [1 (P1/P2)]. (28)

This is the working equation employed by the penta-pycnometer.

Convection Oven The convection oven can maintain 105 5C.


20/117


This drying process has two main problems: (1) clay mineral interlayer water is

largely lost in addition to interstitial water and (2) specimens dried in a convection

oven become brick hard and are rarely useful for any other analyses that require

substantial sample volumes. Use of freeze-drying would partly eliminate these

problems. In particular, stable isotope analyses on foraminifers would be possible

from freeze-dried samples, but not from oven-dried samples. The convection oven

is used based on advice from the relevant JOIDES advisory panel, because drying

at 105 5C for 24 hr is a well-established soil science standard.

CALIBRATION

Beaker Mass and

Volume

Beaker mass must be measured and entered into the MAD program for all beakers

to be used on a leg. Beaker volume is not convenient to measure because the low

material volume to void ratio in the pycnometer cell gives inaccurate values. We

have therefore determined the density of the Pyrex beaker material accurately by

filling a beaker with chips of other beakers, measuring its mass and volume, and

calculating its density. The density of 2.2 g/cm3 is stored in the MAD program,

which returns the volume corresponding to each beaker mass.

Custom-made aluminum beakers were used until Leg 168. These beakers were

difficult to clean, corroded with time, and were expensive to manufacture. For

historical data migration purposes, those beaker materials had a density of 2.78 g/

cm3 (determined by P. Blum, 1996).

Balance Calibration The ship is an environment of cyclically changing gravity, and the measured

weight Wof a massMis significantly affected by the ship's motion. IfWis

measured over a period of time several times the periodicity of the ships

acceleration a, the average can be related toM. By using two balances, mass

determination can be significantly accelerated. The following two equations can be

written for two balances:Fs =Msa(t) =As +BsVs(t) (29)

Fr=Mra(t) =Ar+BrVr(t), (30)

where Fs and Frare average measured weights andMs andMrare known mass

standards on the sample and the reference balance, respectively, a is the ship's

average acceleration, Vis the average voltage measured, andA andB are constants

characteristic for the balances. The calibration principle is to measure multiple

standards (typically 1, 5, 10, 20, and 30 g) to determineA andB for each balance.

For the calibration, measuring time should be at least 30 s to cover several of the

78 heave cycles ofJOIDES Resolution.

Equations on page 9 and on page 9 can be solved for a(t), which is assumed to be

equal for both balances:

Ms = [As +BsVs(t)] Mr/ [Ar+BrVr(t)], (31)

which is identical to

Ms (unknown) /Ms (calculated) =Mr(known) /Mr(calculated). (32)


21/117


The first right-hand term in Equation on page 9 is the first approximation to the

calculated sample mass. This value is uncorrected for motion and is returned

instead of 0 if the user setsMr(known) = 0. The second right-hand term in

Equation on page 9 uses the ratio between a known massMron the reference

balance and its corresponding calculated value to correct the first term for the

ships motion.

The MAD program performs the linear regression for multiple standards and

stores the coefficients until a new calibration is performed. A balance calibration

takes up to 15 min. It is recommended that a few control measurements be taken

after a calibration to verify the correct mean value and a percent standard deviation

of less that 1% for 100 or more measurements taken over approximately 30 s.

Pycnometer

Calibration

The pycnometer has an internal calibration procedure. The user is guided through

the procedure step by step by the program. First, cell 4 must be used to calibrate

the reference volume (pressure) VA. Then, the calibration sphere is cycled through

all five cells to determine the empty cell volume (pressure). The calibration values

are stored in the pycnometer and used until a new calibration is performed. A

pycnometer calibration takes up to 30 min.

The instrument calibrates VA by performing two pressurizations, once with the

sample cell empty (VS= 0) and once with the calibration standard of volume Vstd

in the same sample cell.

Equation on page 8 derived previously for a sample measurement for these two

conditions can be written as

VS= 0 = VC- VA / [(P'1/P'2) - 1] (33)

and

VS= Vstd= VC- VA / [(P1/P2) - 1]. (34)

Combining these two equations yieldsVA = Vstd/ {[1/(P'1/P'2) - 1)] - [1/(P1/P2) - 1]}. (35)

The instrument calibrates the volume VCof each cell with one pressurization of

each cell holding the appropriate sample holder and the calibration standard.

Equation on page 10 is then used and can be written as:

VS= Vstd= VC+ VA / [1 - (P1/P2)] (36)

VC= VA {1 / [(P1/P2) - 1]}. (37)

PERFORMANCE

Precision Standards of 1and 20 g are measured to confirm balance calibration, and thereadings should be within 1 mg, or better than 0.1%. Repeatability of specimen

mass at sea should also be within 0.1%.

For the pycnometer, a standard sphere is measured (e.g., 7.0699 cm3) and

precision should be within 0.1% (0.005 cm3 for the sphere mentioned). Repeat

measurements on sediment samples yield a precision of about 1%, probably


22/117


resulting from changes in ambient pressure and temperature and the material

during handling.

Accuracy Mass error: 0.1%.

Volume error: 1%.

MEASUREMENT

The user is guided through data entry by the MAD program, which controls the

balance as well as the pycnometer. The sample ID needs to be entered only once

for the entire process. The pycnometer key pad is not used during measurement.

The following is the general measurement protocol:

1. Typical sampling frequency for MAD measurements is two per section. One

per section is considered a minimum; more than two per section on

medium- to high-recovery legs is rather demanding with the present staff

assignments.

2. Fill a numbered 10-mL Pyrex beaker with sediment to about 3 mm below

the rim so that material is not lost during handling of the beaker. The largest

errors in MAD measurements probably stem from lost material during theprocess and from volume measurements with incompletely filled beakers. It

is the operators responsibility to find the optimum. Place a special PP

Styrofoam plug into the hole left from where the sample was taken from the

working-half core.

3. Enter the sample and beaker number into the Sample program at the

sampling table. This information will then be in the database; only the

beaker number is used at the MAD station to select samples.

4. Measure the mass. Do not let the sample stand without covering it with

plastic film, being careful not to lose material.

5. Optionally, measure the wet volume in the pycnometer. However, this is not

necessary and years of experience have shown that wet volumemeasurements (method B) appear to have a large error.

6. Place the sample in the oven at 105 5C for 24 hr. Place the sample in a

desiccator after it is removed it from the oven.

7. Measure the mass and volume of the dry sample and beaker.

8. Place the residue in a sample bag, attach a completed label, seal, and box.

9. Clean the beaker.


23/117


DATA SPECIFICATIONS

Database model

Notes: The Sample table is used for all ODP core samples. MAD samples are identified by sampling code; the ODP standard designation islinked through the beaker_id. If method A is used the fixed_volume must be set to be TRUE. The MAD calibration history table is a logof calibrations but does not hold the calibration data.

Standard Queries

Table 22 MAD database model.

Sample MAD sample data MAD control data MAD beaker history

sample_id [PK1] sample_id [PK1] [FK] mad_control_id [PK1] mad_beaker_id [PK1]location [PK2] location [PK2] [FK] run_date_time beaker_date_time [PK2]

sam_section_id . section_id mad_beaker_id ctrl_standard_id beaker_number

sam_archive_working beaker_date_time control_type beaker_type

top_interval fixed_volume expected_value beaker_mass

bottom_interval mass_wet_and_beaker pyc_cell_no beaker_volume

piece mass_dry_and_beaker measured_value

sub_piece vol_wet_and_beaker measured_stdev MAD beaker

beaker_id . mad_beaker_id vol_wet_and_beaker_stdev mad_beaker_id [PK1]

volume vol_wet_and_beaker_n

entered_by vol_wet_and_beaker_cell

sample_depth vol_dry_and_beaker MAD calibration history

sample_comment vol_dry_and_beaker _stdev mad_calibratin_id [PK1]

sam_repository . repository vol_dry_and_beaker_n calibration_date_time

s_c_leg . leg vol_dry_and_beaker_cell calibration_type

s_c_sam_code . sam_code comments

sam_sample_code_lab . s_c_l sample_date_time

Table 23 MAD query A (results).

Short description Description Database

Sample ID ODP standard sample designation Link through [Sample]sample_id

Depth User-selected depth type Link through [Sample]sample_id

Wb Water content, relative to bulk mass see MAD Query B

Ws Water content, relative to solid mass see MAD Query B

Calculations depend on the volume measurement method used: A, B, or C

Bulk density Bulk density, method A, B, or C see MAD Query B

Dry density Dry density, method A or B see MAD Query B

Grain density) Grain density, method A or B see MAD Query B

Porosity Porosity, method A or B see MAD Query B

Void ratio Void ratio, method A or B see MAD Query B

Table 24 MAD query B (results, measurements, and parameters) (to be implemented).


Sample ID ODP standard sample designation Link through [Sample]sample_idDepth User-selected depth type Link through [Sample]sample_id

Method A Indicates if method A was used [MAD Sample Data] fixed_volume

Mb+beak Bulk mass of sample + beaker [MAD Sample Data] mass_wet_and_beaker

Md+beak Dry mass of sample + beaker [MAD Sample Data] mass_dry_and_beaker

Vb+beak Bulk volume of sample (+ beaker for B) [MAD Sample Data] vol_wet_and_beaker

sd(Vb+beak) Std. dev. of n vol. measurements (for B) [MAD Sample Data] vol_wet_and_bkr_sd

n(Vb+beak) No. of vol. measurements (for B) [MAD Sample Data] vol_wet_and_bkr_n

c(Vb+beak) Cell no. used for vol. measurement (for B) [MAD Sample Data] vol_wet_and_bkr_cell

Vd+beak Dry volume of sample + beaker [MAD Sample Data] vol_wet_and_beaker


24/117


sd(Vd+beak) Std. dev. of n vol. measurements [MAD Sample Data] vol_wet_and_bkr_sd

n(Vd+beak) No. of vol. measurements [MAD Sample Data] vol_wet_and_bkr_n

c(Vd+beak) Cell no. used for vol. measurement. [MAD Sample Data] vol_wet_and_bkr_cell

Comments Comments comments

Date/Time Date and time of measurement sample_date_time

Beaker Beaker number [MAD Beaker History] beaker_number

Mbeak Mass of beaker [MAD Beaker History] beaker_mass

Vbeak Volume of beaker [MAD Beaker History] beaker_volume

Mb Bulk mass = (Mb+beak) - Mbeak

Md Dry mass (includes evaporated salt) = (Md+beak) - Mbeak

Mpw Mass of porewater = (Mb - Md) / 0.965

Ms Mass of solids (salt-corrected) = (Md - 0.035*Mb) / 0.965

Vpw Volume of porewater = Mpw / 1.024

Msalt Mass of evaporated salt = Mpw - (Mb - Md)

Vsalt Volume of evaporated salt = Msalt / 2.20

Wb Water content relative to bulk mass = Mpw / Mb

Ws Water content relative to solid mass = Mpw / Ms

For volume method A

Vb(A) Bulk volume (method A) = (Vb+beak)

Vs(A) Volume of solids (methods A) = Vb(A) - Vpw

For volume method B

Vb(B) Bulk volume (method B) = (Vb+beak) - Vbeak

Vs(B) Volume of solids (methods B) = Vb(B) - Vpw

For volume method C

Vd(C) Dry volume (method C) = (Vd+beak) - Vbeak

Vs(C) Volume of solids (method C) = Vd(C) - Vsalt

Vb(C) Bulk volume (method C) = Vs(C) + Vpw

For volume method A or B

Bulk density Bulk density, method A or B = Mb / Vb(A,B)

Dry density Dry density, method A or B = Ms / Vb(A,B)

Grain density) Grain density, method A or B = Ms / Vs(A,B)

Porosity Porosity, method A or B = Vpw / Vb(A,B)

Void ratio Void ratio, method A or B = Vpw / Vs(A,B)

For volume method C

Bulk density Bulk density, method C = Mb / Vb(C)

Dry density Dry density, method C = Ms / Vb(C)

Grain density) Grain density, method C = Ms / Vs(C)

Porosity Porosity, method C = Vpw / Vb(C)

Void ratio Void ratio, method C = Vpw / Vs(C)

Table 24 MAD query B (results, measurements, and parameters) (to be implemented).

Table 25 MAD query C (control measurements) (to be implemented).


Date/Time Date/time of control measurement. [MAD Control Data] run_date_time

Standard Standard identification [MAD Control Data] ctrl_standard_id

Type Type of control meas. (mass or vol.) [MAD Control Data] control_type

Expected Expected value [MAD Control Data] expected_value

Cell If pycnometer, cell number used [MAD Control Data] pyc_cell_no

Measured Measured value [MAD Control Data] measured_value

Stdev. Std. dev. of multiple vol. meas. [MAD Control Data] measured_stdev

Table 26 MAD query D (beaker data) (to be implemented).


Date/Time Data/time of beaker meas. [MAD Beaker History] beaker_date_time


25/117


Beaker Beaker number [MAD Beaker History] beaker_number

Type Type of beaker (e.g., Pyrex 10 mL) [MAD Beaker History] beaker_type

Mbeak Measured mass of beaker [MAD Beaker History] beaker_mass

Vbeak Calculated volume of beaker [MAD Beaker History] beaker_volume

Table 26 MAD query D (beaker data) (to be implemented).

Table 27 MAD query E (calibration log) (to be implemented).

Short description Description Database Date/Time Date/time of calibration calibration_date_time

Type Type of calibration (mass or vol.) calibration_type


26/117



27/117


3. GAMMA-RAY DENSIOMETRY

3.1. Principles

PHYSICAL BACKGROUND

Bulk density of sediments and rocks is estimated from the measurement of

gamma-ray attenuation (GRA) (Tittman and Wahl, 1965; Evans, 1965). The

familiar acronym GRAPE (Evans, 1965) stands for GRA porosity evaluator,

referring to the computer that Evans attached to the density measurement device to

compute porosity using an assumed grain density. The measurement device does

not estimate porosity, and is therefore referred to as GRA densiometer.

The principle is based on the facts that medium-energy gamma rays (0.11 MeV)

interact with the formation material mainly by Compton scattering, that the

elements of most rock-forming minerals have similar Compton mass attenuation

coefficients, and that the electron density measured can easily be related to the

material bulk density. The 137Ce source used transmits gamma rays at 660 KeV. A

scintillation detector measures the gamma-ray beam transmitted through the core

material. If the predominant interaction is Compton scattering, transmission of

gamma rays through matter can be related to the electron density by:

Yt= Yiensd, (1)

where Yi is the flux incident on the scatterer of thickness d, Ytis the flux

transmitted through the scatterer, n is the number of scatterers per unit volume or

the electron density, and s is the Compton cross section for scattering per scatterer

in square centimeters per electron. Bulk density of the material is related to the

electron density by

n = NAv (Z/A), (2)

whereZis the atomic number or the number of electrons,A is the atomic mass of

the material, andNAv is the Avogadro number. Bulk density estimates are therefore

accurate for a wide range of lithologies if theZ/A of the constituent elements is

approximately constant. Variations ofZ/A are indeed negligible for the most

common rock-forming elements. The GRA coefficient is defined as

= (Z/A)NAv s (cm2/g). (3)

For the medium energy range of gamma rays and for materials withZ/A of about 1/

2, such as the most common minerals, the Compton is approximately 0.10

cm2/g, increasing with decreasing energy. For water, is about 11% higher than

for common minerals at a particular energy (e.g., Harms and Choquette, 1965).

Sediments can therefore be regarded as two-phase systems in regard to GRA

(mineral-water mixtures).


28/117


Equation on page 1 can now be written in the more frequently referenced form

Yt= Yied (4)

and the expression for the bulk density becomes

= ln (Yt/Yi) /d. (5)

If the coefficient could be determined with sufficient accuracy, it could be used

directly to compute bulk density. However, is a function of detected gamma-ray

energy and is therefore dependent on the particular device, including source,detector, spectral component used, and the material itself (degree of scattering). A

more practical and accurate method is to calibrate the gamma radiation with bulk

density standards as described later in this chapter.


Attenuation

Coefficient of

Minerals

An important assumption of this densiometry method is that for a given

measurement system the average attenuation coefficient is constant for the

measured materials. For a more accurate density estimate, variations in the average

composition of the material must be taken into consideration. If mineralogical

analysis determines that the average 1 deviates significantly from the standard ,

the following correction can be applied:

1 = /1 , (6)

where the ratio of average coefficients can be calculated from reference tables.

Core Thickness The GRA routine calculations assume a constant core diameter of 66 mm. If voids

or otherwise incompletely filled core liner segments occur because of gas pressure,

gas escape, or other coring disturbances, the density estimate will be too low. (The

highest values are therefore the most reliable ones in disturbed cores.) Using a

thickness log obtained from core photographs or by other means, density can

easily be corrected for varying core thickness using

1 = d/d1 . (7)

USE OF GRA DATA

GRA data provide a precise and densely sampled record of bulk density, an

indicator of lithology and porosity changes. The records are frequently used for

core-to-core correlation. Another important application is the calculation of

acoustic impedance and construction of synthetic seismograms.

3.2. MST (Whole-Core) GRA System

EQUIPMENT

Gamma-ray Source The 137Ce source used transmits gamma rays at 660 KeV.


29/117


Scintillation Counter A standard NaI scintillation detector is used in conjunction with a universal

counter.

CALIBRATION

New Procedure GRA calibration assumes a two-phase system model for sediments and rocks,

where the two phases are the minerals and the interstitial water. Aluminum has an

attenuation coefficient similar to common minerals and is used as the mineralphase standard. Pure water is used as the interstitial-water phase standard. The

actual standard consists of a telescoping aluminum rod (five elements of varying

thickness) mounted in a piece of core liner and filled with distilled water (Figure

31). The standard element i has an average bulk density i of

i = di /D Al + (D di)/D water (8)

whereD is the maximum aluminum rod thickness (inner diameter of core liner, 6.6

cm), di is the diameter of the aluminum rod in element i, and Al and waterare the

densities of aluminum and water, respectively. The first element (porosity of 0%)

has a bulk density of aluminum (2.70 g/cm3) and the last element (porosity of

100%) has a bulk density of water at laboratory temperature (1.00 g/cm3).

Intermediate elements are used to verify the linearity of the ln(Y) to density

relationship, as well as the precise alignement of core and sensor. A linear least

squares fit through three to five calibration points (ln(counts/tcal), ) yields the

calibration coefficients m0 (intercept) and m1 (slope, negative). Total measured

counts are automatically divided by the counting time, tcal, to normalize the

coefficients to counts per second. Sample density is then determined:

core = m0 + ln (counts/tsample)m1 , (9)

where the measured counts are again normalized to counts per second using the

sampling period, tsample , before the calibration coefficients are applied.

Old Procedure The present calibration procedure has been implemented only since Leg 169

(August 1996). Before that time, calibration was performed with two aluminum

cylinders of different thickness, but without water. The thinner aluminum rod was

cut to a diameter of 25 mm to give an aluminum density of 1.00. The counts

returned from measuring the thin aluminum rod were not compatible with the

Compton attenuation coefficient for water, however, and when measuring water

the density was about 11% too high. A fluid-correction had to be applied to the

initial density estimate. This procedure is obsolete now, and no fluid correction is

required because water is used in the calibration procedure.


30/117


MEASUREMENT

The GRA is logged downcore automatically..

Figure 31 Schematic of GRA calibration. A. Physical standard used. B. Mea-surement geometery. C. Calibration principle. D. Application of calibration tocore measurement

PERFORMANCE

Precision Precision is proportional to the square root of the counts measured because

gamma-ray emission is subject to Poisson statistics (see Natural Gamma

Radiation chapter for additional explanation). The statistical uncertainty ist Nz (t N)1/2, (10)

where Nis the count rate (counts per second, cps), tis the sampling period (s), and

z is the number of standard deviations for the normal distribution (0.68 probability,

or confidence, forz = 1; 0.95 for z = 1.96, etc.). Measurements with the present

system have typically count rates of 10,000 (dense rock) to 20,000 cps (soft mud).

If measured for 4 s, the statistical error is therefore less than 40,000 200, or

ln(counts/tcal)

Density (g/cm3)

m0

(g/cm3) m1(g/cm3)

counts= total measured counts

tcal= calibration counting period (s)

core= 'coredcore/dstandard

S1 S3S2 S4 S5

Distilled water

Aluminum

49 mm2.28 g/cm3

32 mm1.83 g/cm3

66 mm2.72 g/cm3

16 mm1.42 g/cm3

0 mm1.00 g/cm3

Rod thickness:Average density:

Core liner

Center/support disk

Thin rod

providesalignment

control

GAMMA-RAY ATTENUATION DENSIOMETRY

Scintillation

detector

A

B C

D

137Ce

source

dcore

Two-phase model: minerals = aluminum; pore water = distilled water

tsam= sampling period (s) dcore values are determined separately, standard report assumes

full core liner, so that dcore= dstandard(= 66 mm for ODP)

core= m0+ m1 ln(counts/tsam)


31/117


0.5%. This shows that the high flux of the 137Ce source does not require excessive

counting times.

Accuracy Accuracy is limited by the assumption that the measured material has the same

attenuation coefficent as the calibration standards used. For general sediment-

water mixtures, this should be the case and errors should be less than 5%.

Spatial Resolution The GRA system allows high spatial resolution of about 0.5 cm.

DATA SPECIFICATIONS

Database Model

Notes: GRA control 1 are control measurements run the same way as a core section. GRA control 2 are measurement taken before run. GRAcontrol 3 are control measurements from a standard mounted on the core boat.

Standard Queries

Table 31 GRA database model.

GRA section GRA control 1 GRA control 3 GRA calibration

gra_id [PK1] gra_ctrl_1_id [PK1] [FK] gra_ctrl_3_id [PK1] density_calibration_id [PK1]

section_id run_number run_number calibration_date_time

run_number run_date_time run_date_time run_numberrun_date_time core_status requested_daq_period system_id

core_status liner_status actual_daq_period liner_status

liner_status requested_daq_interval density_calibration_id requested_daq_period

requested_daq_interval requested_daq_period standard_id density_m0

requested_daq_period density_calibration_id meas_counts density_m1

density_calibration_id standard_id density_mse

mst_gra_ctrl_2_id comments

mst_gra_ctrl_3_id GRA control 2

gra_ctrl_2_id [PK1] GRA calibration data

GRA section data GRA control 1 Data run_number density_calibration_id [PK1] [FK]

gra_id [PK1] [FK] gra_ctrl_1_id [PK1] [FK] run_date_time mst_top_interval [PK2]

mst_top_interval [PK2] mst_top_interval [PK2] requested_daq_period standard_id [PK3][FK]

mst_bottom_interval mst_bottom_interval actual_daq_period mst_bottom_interval

actual_daq_period actual_daq_period density_calibration_id standard_density

meas_counts meas_counts meas_counts actual_daq_period

core_diameter core_diameter meas_counts

Table 32 GRA report.


A: Results

Sample ID ODP standard sample designation Link through [GRA Section]section_idDepth User-selected depth type Link through [GRA Section]section_id

Bulk density = [GRA Calibration] density_m0 +

ln ([GRA Section data] meas_counts)

/ [GRA Section data] actual_daq_period)

* [GRA Calibration] density_m1

B (optional): Parameters and measurements

Run Run number [GRA Section] run_number

Date/Time Run date/time [GRA Section] run_date_time

Core Status HALF or FULL [GRA Section] core_status


32/117


Liner Status NONE, HALF or FULL [GRA Section] liner_status

Req. Interval User-defined sampling interval (cm) [GRA Section] requested_daq_interval

Req. Period User-defined sampling period (s) [GRA Section] requested_daq_period

Period Measured sampling period (s) [GRA Section Data] actual_daq_period

Counts Measured counts (not normalized) [GRA Section Data] meas_counts

Core Dia. Core diameter, default = 6.6 cm [GRA Section Data] core_diameter

Cal. Date/Time Calibration date/time [GRA Calibration] Calibration_date_time

Cal. m0 Calibration intercept (g/cm3) [GRA Calibration] density_m0

Cal. m1 Calibration slope ([g/cm3)]/cps) [GRA Calibration] density_m1

Table 32 GRA report.

Table 33 GRA control 1 measurements (to be implemented).


Bulk density =[GRA Calibration] density_m0 +

ln ([GRA Ctrl 1 Data] meas_counts

/ [GRA Ctrl 1 Data] actual_daq_period)


Run Run number [GRA Ctrl 1] run_number

Date/Time Run date/time [GRA Ctrl 1] run_date_time

Core Status HALF or FULL [GRA Ctrl 1] core_status

Liner Status NONE, HALF or FULL [GRA Ctrl 1] liner_status

Standard Standard name [Phys. Properties Std.] standard_name

Std. Set Standard set name [Phys. Properties Std.] standard_set_name

Std. Expected Expected value (range) (g/cm3) [Phys. Prop. Std. Data] property_value

Interval Interval top [GRA Ctrl 1 Data] mst_top_interval

Req. Interval User-defined sampling interval (cm) [GRA Ctrl 1] requested_daq_interval

Req. Period User-defined sampling period (s) [GRA Ctrl 1] requested_daq_period

Period Measured sampling period (s) [GRA Ctrl 1 Data] actual_daq_period

Counts Measured counts (not normalized) [GRA Ctrl 1 Data] meas_counts

Core Dia. Core diameter, default = 6.6 cm [GRA Ctrl 1 Data] core_diameter
















Cal. m0 Calibration intercept (g/cm3) [GRA Calibration] density_m0Cal. m1 Calibration slope ([g/cm3)]/cps) [GRA Calibration] density_m1







33/117


3.3. Split-core GRA System

ODP has purchased a split-core GRA system that will be implemented as soon as

resources become available. This system must be implemented together with the

latest model GEOTEK P-wave logger which provides the caliper measurement

required to correct split-core GRA measurements for uneven split-core thickness.














Table 36 GRA calibration data (to be implemented).


Date/Time Calibration date/time [GRA Calibration] calibration_date_time



Cal. mse Calibration mean squared error [GRA Calibration] mse

Run Run number [GRA Calibration] run_numberLiner Status NONE, HALF or FULL [GRA Calibration] liner_status

Req. Period User-defined sampling period (s) [GRA Calibration] requested_daq_period

Comments Comments [GRA Calibration] comments




Density Density value from MST control [GRA Calibration Data] standard_density

Interval Interval top [GRA Calibrat ion Data] mst_top_interval

Period Measured sampling period (s) [GRA Calibration Data] actual_daq_period

Counts Measured counts (not normalized) [GRA Calibration Data] meas_counts


34/117


4. MAGNETIC SUSCEPTIBILITY

4.1. Principles

PHYSICAL BACKGROUND

Magnetic susceptibility is the degree to which a material can be magnetized in an

external magnetic field. If the ratio of the magnetization is expressed per unit

volume, volume susceptibility is defined as

=M/H, (1)

whereMis the volume magnetization induced in a material of susceptibility by

the applied external fieldH. Volume susceptibility is a dimensionless quantity. The

value depends on the measurement system used:

(SI) = 4(cgs) = 4 G Oe1, (2)

where G and Oe are abbreviations for Gauss and Orstedt, respectively. The SI

system should be used.

Mass, or specific, susceptibility is defined as

= / , (3)

where is the density of the material. The dimensions of mass susceptibility are

therefore m3/kg.

Magnetic susceptibility measured by the common methods is an apparent value

because of the self-demagnetizing effect associated with anisotropy connected

with the shape of magnetic bodies, such as magnetite grains (Thompson and

Oldfield, 1986). When a substance is magnetized its internal magnetic field is less

than the externally applied field. i, the intrinsic susceptibility, relates the induced

magnetization to the internal magnetic field, whereas e, the extrinsic

susceptibility which we actually observe, relates the induced magnetization to the

externally applied field. The relationship between the two susceptibilities can be

shown to be

e = i / (1 +Ni), (4)

whereNis the demagnetization factor. For a strongly magnetic mineral, such as

magnetite,Ni > 1, and e ~ 1/N. IfNis known, there is a simple relationship

between the concentration of ferrimagnetic grains and the magnetic susceptibility.

This is the case for natural samples where the concentration of ferrimagnetic

minerals is a few percent or less. The measured susceptibility can be

approximated:

= e ~ /N, (5)

where is the volume fraction of ferrimagnetic grains. It is found that for natural

samplesNis reasonably constant with a value close to 1/3. Thus, if the grain


35/117


shapes are roughly spherical and the dominant mineral is magnetite, the volume

fraction (


36/117



Cores should be equilibrated to room temperature before measurement.

USE OF MAGNETIC SUSCEPTIBILITY

Magnetic susceptibility is used mostly as a relative proxy indicator for changes in

composition that can be linked to paleoclimate-controlled depositional processes.The high precision and sensitivity of susceptibility loggers makes this

measurement extremely useful for core-to-core and core-downhole log correlation.

The physical link of magnetic susceptibility to particular sediment components,

ocean or wind current strength and direction, or provenance, usually requires more

detailed magnetic properties studies in a specialized shorebased laboratory.

Iron-titanium oxides

Hematitea 500 to 40,000 10 to 60 to 760

Maghemitea 2,000,000 to 2,500,000 40,000 to 50,000

Ilmenitea 2,200 to 3,800,000 46 to 200 to 80,000

Magnetitea 1,000,000 to 5,700,000 20,000 to 50,000 to110,000

Titanomagnetite 130,000 to 620,000 2,500 to 12,000

Titanomaghemite 2,200,000 57,000

Ulvospinel 4,800 100

Average rock values

Sandstones, shales, limestones 0 to 25,000 0 to 1,200

Dolomite -10 to -940 -1 to -41

Clay 170 to 250 10 to 15

Coal 25 1.9

Basalt, diabase 250 to 180,000 8.4 - 6,100Gabbro 1,000 to 90,000 26 to 3,000

Peridotite 96,000 to 200,000 3,000 to 6,200

Granite 0 to 50,000 0 to 1,900

Rhyolite 250 to 38,000 10 to 1,500

Documents

PP Handbook Peter Blum November 1997