Gubbins D., Herrero-Bervera E. Encyclopedia of Geomagnetism and Paleomagnetism

Подождите немного. Документ загружается.

Khan, M.A., 1962. The anisotropy of magnetic susceptibility of some

igneous and metamorphic rocks. Journal of Geophysical Research,

67: 2873–2885.

Nye, J.F., 1960. Physical Properties of Crystals. Oxford: Oxford

University Press, 323 pp.

Rees, A.I., 1961. The effect of water currents on the magnetic rema-

nence and anisotropy of susceptibility of some sediments. Geophy-

sical Journal, 5: 235–251.

Rees, A.I., 1966. The effect of depositional slopes on the anisotropy of

magnetic susceptibility of laboratory deposited sands. Journal of

Geology, 74: 856–867.

Rees, A.I., 1968. The production of preferred orientation in a concen-

trated dispersion of elongated and flattened grains. Journal of

Geology, 76: 457–465.

Smith, J.V., 2002. Structural analysis of flow-related textures in lavas.

Earth Science Reviews, 57: 279–297.

Uyeda, S., Fuller, M.D., Belshé, J.C., and Girdler, R.W., 1963. Aniso-

tropy of magnetic susceptibility of rocks and minerals. Journal of

Geophysical Research, 68: 279–291.

Cross-references

Magnetic Anisotropy, Sedimentary Rocks and Strain Alteration

Magnetic Susceptibility, Anisotropy

Magnetic Susceptibility, Anisotropy, Effects of Heating

Magnetic Susceptibility, Anisotropy, Rock Fabric

Magnetic Susceptibility, Low-Field

Susceptibility

Susceptibility, Measurements of Solids

Susceptibility, Parameters, Anisotropy

MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD

All mineral grains are “susceptible” to become magnetized in the pre-

sence of a magnetic field, and MS is an indicator of the strength of this

transient magnetism within a material sample. MS is very different

from remanent magnetism (RM), the magnetization that accounts

for the magnetic polarity of materials. MS in sediments is generally

considered to be an indicator of iron, ferromagnesian or clay mineral

concentration, and can be quickly and easily measured on small sam-

ples. In the very-low-inducing magnetic fields that are generally

applied, MS is largely a function of the composition, concentration,

grain size, and morphology of the magnetizable material in a sample.

It is also somewhat variable in direction (MS has anisotropy) accord-

ing to magnetocrystalline anisotropy, mineral distributions, and grain

morphology. MS has the advantage of being quickly and easily mea-

sured on small, friable, unoriented samples using commercially avail-

able devices such as balanced coil induction systems (susceptibility

bridges). Most of these instruments generate alternating magnetic

fields at from 500 Hz to 5 kHz. Sensitivity is better at higher fre-

quencies so most of the commercially available bridges operate at

around 5 kHz.

Magnetizable materials in sediments include not only the ferrimag-

netic and antiferromagnetic minerals, such as the iron oxide minerals

magnetite and maghemite, and iron sulfide and sulfate minerals,

including pyrrhotite and greigite, that may acquire an RM (required

for reversal magnetostratigraphy), but also any other less magnetic

compounds, paramagnetic compounds. The important paramagnetic

minerals in sediments include the clays, particularly chlorite, smectite,

and illite, ferromagnesian silicates such as biotite, pyroxene, and

amphiboles, iron sulfides including pyrite and marcasite, iron carbo-

nates such as siderite and ankerite, and other iron- and magnesium-

bearing minerals.

In addition to the ferrimagnetic and paramagnetic grains in sedi-

ments, calcite and/or quartz may also be abundant, as are organic

compounds. These compounds typically acquire a very weak negative

MS when placed in inducing magnetic fields, that is, their acquired

MS is opposed to the low magnetic field that is applied. The presence

of these diamagnetic minerals reduces the MS in a sample. Therefore

factors such as changes in biological productivity or organic carbon

accumulation rates may cause some variability in MS values.

Low-field MS, as used in most reported studies, is defined as the

ratio of the induced moment (M

or J

) to the strength of an applied,

very low-intensity magnetic field (H

), where

¼ w

ðdensity-specificÞ (Eq. 1)

¼ k

ðvolume-specificÞ (Eq. 2)

In these expressions, MS in SI units is parameterized as k, indicating

that the measurement is relative to a 1 m

volume and therefore is

dimensionless; MS parameterized as w indicates measurement relative

to a mass of 1 kg, and is given in units of m

–1

. Both k and w have

anisotropy.

Bulk (initial) measurements of MS are often performed without con-

sideration of the anisotropy of samples, and for most samples the bulk

MS error that is due to this anisotropy is usually small. The anisotropy

effect is further reduced when samples are crushed before the MS is

measured. Unlike the remanence of a sample, MS is much less suscep-

tible to remagnetization than is the remanent magnetization in rocks.

MS can be measured on small, irregular lithic fragments, and on the

highly friable material that is difficult to sample for RM measurement.

Applications and interpretations of bulk (initial)

low-field magnetic susceptibility

MS variabilityfrom sample to sample results primarily fromchanges in the

amount of its iron, clay, and ferromagnesian constituents. This variability

in sedimentary sequences is primary, produced during deposition, and sec-

ondary, resulting from alteration of the deposited sequence. The primary

factors include weathering, erosion and deposition rates, and biological

productivity. These factors are tied to climate, tectonic and volcanic pro-

cesses, availability of nutrients, and local (relative) or global sea-level

changes. Secondary processes include pedogenesis, diagenesis, and

within-sediment redox effects driven by sulfate-reducing bacteria.

Studies of marine sedimentary sequences

Marine sediments are composed of terrigenous, detrital and eolian

components, and a biogenic component of the carbonate and/or silic-

eous tests of marine organisms. There is also a component of sediment

derived from marine weathering processes, but this is generally minor.

The MS signature mainly represents contributions from the terrigenous

and biogenic components. Because the biogenic component is only

weakly diamagnetic, in general, the terrigenous component dominates

the MS. Therefore, the processes controlling the influx into the marine

environment of terrigenous components will usually account for the

MS variations observed. Thus, climate cycles that cause erosion and

transport of terrigenous material will result in a cyclic signal in the

MS recorded in the deposited sediments. There are many factors that

can modify this signature, but while the expected variations may be

reduced in magnitude, usually the character of the cycles will remain

and can be extracted. Superimposed on these cyclic variations are con-

tributions from unique events, such as volcanic eruptions or meteorite

impacts, which may be of local, regional, or global significance.

Continuous and discrete measurements on unlithified

sedimentary cores; marine and lacustrine

Routine MS measurements were not possible on sediments until the

1960s when published instructions for building instruments became

available and when the first commercially produced instruments

566 MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD

entered the market. Early work on sediments was difficult because

of the low sensitivity of these instruments. Besides instrumental

problems, there were ambiguities in units reported, making interpreta-

tions difficult. In the 1970s, publications of MS of discrete samples

from marine piston cores began to appear in the literature. In the

1980s, better instrumental sensitivity made it possible to build instru-

ments for continuous MS measurement on cores and these were

applied to marine and lacustrine sequences. Much of this work was

directed toward correlating cores. In unlithified sediments an inverse

correlation between percent carbonate and MS variability, observed as

early as the middle 1970s, was interpreted to result from variations in

marine organic carbonate productivity (Ellwood and Ledbetter, 1977).

More recent work looking at the spectral character of these MS

variations has shown that the cycles observed are correlated to climate

cyclicities.

Ultimately, the MS in most marine and lacustrine sediments can be

used as a proxy for physical processes responsible for delivering

the detrital and eolian components to the system. These processes

are both regional and global and include climate, volcanism and

tectonic processes. In marine sediments they also include relative

sea-level changes and eustacy. (See Evans and Heller (2003) for a

general discussion of unlithified marine and lacustrine sediments.)

Surface sediment measurements from pilot and trigger

core-top samples

MS has proved useful in some cases. For example, the MS has been mea-

sured on a large number of cores from throughout the Argentine Basin in

the South Atlantic Ocean and the variability is characterized (Sachs and

Ellwood, 1988). They showed that the detrital components responsible

for the MS originated from two primary sources. These components

are brought into and along the western margin of the basin from

Antarctica by Antarctic Bottom Water (AABW), and into the basin from

rivers along the eastern margin of South America. Both sources exhibit

relatively high MS values at their sources. These values exhibit a

rapid down-current MS reduction away from the source. AABW and

other bottom currents then disperse detrital components throughout the

basin, so that the majority of MS values from the basin show similar

values.

Results from lithified marine sedimentary rock

sequences

During diagenesis the MS of marine sediment decreases. This occurs

because ferrimagnetic grains that dominate in unlithified sediments

are generally converted to paramagnetic phases during burial and

diagenesis, mainly by the action of sulfate-reducing bacterial organ-

isms. This process transforms much of the iron to paramagnetic phases

such as iron sulfide and carbonate minerals. In general, however, this

process does not remove the iron from the system, so that total iron

content is conserved. These new phases then contribute, along with the

other paramagnetic detrital components present, such as clays, and the

remaining authigenic ferrima gnetic constituen ts, to the MS in lithified

sediments. Measured MS values for most marine limestones, marls, shales,

siltstones, and sandstones range from 1  10

–9

to 1  10

–7

–1

Shales and some marls have the greatest anisotropy, but for most of

these samples it is less than 2%. Limestones, marls, siltstones, and

sandstones are usually less than 1%. Shale and marl samples are friable

and measurement is done generally on small broken fragments, further

reducing the anisotropy effect. Therefore, the error introduced by the

anisotropy of MS is minimal. A test of the MS variability in single

limestone beds has demonstrated the nonvarying nature of the MS

over distances up to 25 km (Ellwood et al., 1999). Diagenesis also modi-

fies some of the paramagnetic constituents that make up the terrigenous

component of marine sediments. In many cases clays dominate much of

the detrital/eolian component, and these are generally not destroyed but

may be altered. The iron is still conserved in the secondary clays, as is

the paramagnetic behavior of these materials. Other common detrital

components observed in marine rocks are the ferromagnesian minerals

including biotite, tourmaline, and other compounds (see Ellwood

et al., 2000 for discussion).

Ultimately, following diagenesis, the MS observed for most marine

sedimentary rocks can be used as a proxy for physical processes

responsible for delivering the detrital/eolian component of these sedi-

ments into sedimentary basins. As observed for unlithified sediments,

these processes in lithified sediments are both regional and global and

include climate, volcanism, tectonic processes, relative sea-level

changes, and eustacy.

Presentation of MS data from lithified sequences

For presentation purposes and inter data set comparisons, the bar log

format, similar to that previously established for magnetic polarity data

presentations, is recommended. These bar logs should be accompanied

by both raw and smoothed MS data sets. Here, raw MS data (dashed

line in Figure M91) are smoothed using splines (solid line with circles

indicating data point locations shown in Figure M91). The following

bar log plotting convention is used; if the MS cyclic trends increase

or decrease by a factor of two or more, and if the change is represented

by two or more data points, then this change is assumed to be significant

and the highs and lows associated with these cycles are differentiated

by filled (high MS values) or open (low MS values) bar logs (Figure

M91). This method is best employed when high-resolution data sets

are being analyzed (large numbers of closely spaced samples) and

helps resolve variations associated with anomalous samples. Such var-

iations may be due to weathering effects, secondary alteration and

metamorphism, longer term trends due to factors such as plate-driven

eustacy, as opposed to shorter term climate cycles, or event sequences

such as impacts (Ellwood et al., 2003), and to other factors. In addi-

tion, variations in detrital input between localities, or a change in det-

rital sediment source is resolved by developing and comparing bar

logs between different localities. It is also recommended that MS data

be represented in log plots due to the range of MS values often

encountered. However, MS data variations in log plots for some ranges

(1–3  10

–x

vs 4–9  10

–x

) may appear to exhibit large variations

that are artifacts of the presentation style. Therefore, bar logs help to

reduce these ambiguities.

Climate cycles and other MS variations

MS results for almost all studies of marine sedimentary rocks show

many levels of cycles. Some of these cycles are very long-term and

are interpreted as resulting from transgressive-regressive variations

(T-R cycles) due to sea-level rise and fall associated with eustacy.

Other cycles are much shorter and are interpreted to result from clima-

tically driven processes (Ellwood et al., 2001a,b). Figure M91 is an

example of cyclic MS data from limestone—marl couplets from

immediately below the Danian-Selandian (Paleocene) stage boundary

at Zumaia, Spain. These couplets have been interpreted to result from

climate variability, shown to represent climate precession (19–23 ka;

Dinares-Turell et al., 2003).

In addition to these cycles, there are distinct MS peaks that are cor-

related to maximum flooding surfaces (MFS; see discussion in section

“Climate Cycles and Other MS Variations”), and to sudden influxes of

detrital material into sedimentary basins by turbidity currents or other

sediment suspensions. There are also large, relatively sudden shifts or

offsets in MS magnitudes within sequences. The Danian-Selandian

boundary illustrated in Figure M91 may be such a case. If “knife edge”

in character, these shifts have been interpreted as possible unconformi-

ties; several tests have demonstrated such an effect. However, not all

such abrupt changes will be “unconformities,” but may instead be

“pseudounconformities” representing missing core sections, unrecog-

nized faults, or other problems. Work by da Silva and Boulvain

(2002) for Upper Devonian rocks in Belgium has shown a direct, high

MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD 567

correlation between both third and fourth order T-R cycles and MS

variations determined for those rocks.

Large magnitude and rapid shifts in MS (represented by multiple

data points) are interpreted to result from geological processes that

reflect fundamental changes in the sediments being delivered to marine

basins. For example, the sharp change from low MS in the Permian to

high MS in the Triassic (Hanson et al., 1999) is argued to result from

a sudden dramatic influx into ocean basins of detrital material that

helped to cause the mass extinctions in the marine fauna at this

time (Thoa et al., 2004). A similar pulse in MS is associated with

the K-T boundary and is attributed to erosion following the meteorite

impact at that time (Ellwood et al., 2003).

Correlations

MS studies of Phanerozoic marine rock sequences have concentrated

on the geological boundaries known as Global Boundary Stratotype

Sections and Points (GSSPs) that are currently being defined by the

International Geological Congress’s (IGC) Sub-commission on Strati-

graphy. A GSSP represents a “point” or level in a stratigraphic sequence

that defines the beginning of a new geological stage in chronostrati-

graphic time. GSSPs are becoming the basis for the modern Geologi-

cal Timescale. Today, about half of the more than 90 GSSPs for the

Phanerozoic have been formally defined. IGC intends to have all

GSSPs formally defined by the year 2008. All of these boundaries are

well studied and much of the biostratigraphy has been published. Most

stage boundaries are defined based on the first occurrence of a new

fossil species, while a few use other criteria. For example, the Paleo-

cene-Eocene GSSP in Upper Egypt near Luxor is defined by a negative

carbon isotope shift in the sequence exposed there and in most other

Paleocene-Eocene sequences. The biostratigraphy for this section is also

well known and the MS signature is distinctive (Figure M92).

The MS of several of these GSSPs has been studied and a number

of results have been published (for examples see Hanson et al., 1998

and Thoa et al., 2004 for MS results from the Permian-Triassic

boundary, and Crick et al., 2001 for the MS of the Silurian-Devonian

boundary. The Silurian-Devonian boundary represents the first GSSP,

established in 1972). The purpose of this work has been to use MS

as a viable tool for correlating geological sequences, especially for

the Paleozoic where polarities are not well defined and where remag-

netization is a significant problem. Published and unpublished work

on GSSPs and well-studied associated sections has demonstrated that

MS variations can be correlated between sections (Crick et al., 1997).

It has been argued that MS tracks processes that control the influx of

detrital and eolian grains into the marine environment. For example,

sequence stratigraphic studies have demonstrated that at times when

sea level is at a maximum, producing an MFS, at the end of a

transgressive cycle there is a large influx of detrital material into the

marine environment. There is a corresponding MS high associated with

these MFSs. However, as a general rule, during transgressive cycles,

especially in distal marine sequences, MS is observed to decrease. This

happens because detrital sediments are usually trapped near shore during

transgressions. During regressions, when base level is lowered due to

falling sea level, erosion flushes detrital sediment into ocean basins

and MS increases in basinal sediments during these times. Thus T-R

cycles play an important role in creating some cyclic MS variations.

Results of the MS work in these rocks have shown that the MS is

mineral-dependent but may be independent of the macrolithology.

Reworking of sediments by ocean currents, and the variable deposition

of clays may produce large MS changes in both limestones and shales,

even though in general the lithology may appear to be relatively non-

varying. In addition, while local subsidence may cause a lithologic

change from shale to limestone or limestone to shale, the MS may

remain relatively constant ( Figure M93).

Figure M91 (a) Uppermost Danian limestone (resistant)—marl couplets at Zumaia, Spain. (b) MS for the interval shown. High MS

values are associated with the marls. Note the character change in the basal Selandian shales (Ellwood et al., in preparation).

568 MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD

Studies of sedimentary sequences exposed during

archeological excavations within limestone caves and

deep rock shelters

Archaeological artifacts found during excavation are usually encased

in sediments. In most cases, the microenvironment within caves and

deep rock shelters protects deposited sediments from the large-scale

alteration effects produced by pedogenesis. Artifacts, lithic fragments,

and other materials are generally interpreted to have been deposited as

part of a stratigraphic sequence and are isochronous with the encasing

sediments. Careful microstratigraphic excavations expose stratigraphic

sequences that can be sampled for many purposes, including the study

of faunal and floral materials, geochemical studies, dating, and charac-

terization of MS variability. For this purpose, sediment is usually

collected as a set of continuous samples that allows evaluation of

a reasonably complete archaeological sequence. Such a sequence is

illustrated in Figure M94a.

The source of these sediments can be fairly complex. In most cases,

pedogenesis of the limestone-surrounding caves is the source of

eroded soil that can be blown, washed, or tracked in from outside

the cave. Sediments may be deposited from streams flowing through

or within caves or derived as residues from dissolution of the contain-

ing limestones. Because the sedimentary sequences in caves have been

shown to accumulate relatively rapidly, limestone residues are consid-

ered to be only a minor source of the observed sediments, leaving flu-

vial and soil materials as the dominant sedimentary sources. Some of

the material transported into caves is loess and the MS usually reflects

the glacial-interglacial variations observed in these loess sequences

(loess MS studies are discussed below.)

The MS of accumulating sediments derived from limestone pedo-

genesis in archaeological stratigraphic sequences varies with changing

climate. In general, periods of warm climate exhibit increases in ped-

ogenesis that produce extremely fine-grained ferrimagnetic iron oxide

minerals, primarily magnetite and maghemite. During cold periods,

pedogenic activity is reduced, thus reducing the production of ferri-

magnetic components. As surface soils are blown, tracked or washed

into caves and rock shelters, the MS reflects the climate at the time

Figure M92 (a) The Paleocene-Eocene GSSP section exposed near Luxor, Egypt. (b) MS variation for the section composed mainly of

shales (Ellwood et al., 2007). Bar log convention as discussed in “Presentation of MS Data from Lithified Sequences” section.

Figure M93 MS variation in a limestone (light) to shale (dark)

sequence measured with a portable susceptibility bridge. Note the

MS during the transition from limestone to shale is essentially

nonvarying (top), whereas MS variation within the limestone

(bottom) shows a large range.

MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD 569

the material is accumulating. This has been tested in a number of cases

against faunal, floral, and lithic data sets, by relating MS data to dated

chronologies, and by studies of the concentration and character of the

dominant magnetic components in these sediments. All are consistent

in most cases with the climate interpretation given here.

Fluvial sediments within caves are usually deposited as quickly accu-

mulating, thinly laminated sequences of clays, silts, and sands, and thus

easily identified as fluvial. In MS studies these sequences usually do not

show much variation. Areas of sedimentary mixing or disturbance

are also distinctive because the mixing process radically reduces

the MS variability. In addition, in some archaeological contexts there

are alteration effects that produce distinctive, much reduced MS magni-

tudes that are essentially nonvarying through sequences that are

expected to show climate variations. In many cases the sediments from

the top-down to midlevel in a sequence show good cyclicity, whereas

below that is observed a distinctive offset to much lower MS values.

Rock magnetic analyses of these materials show that the primary

magnetite or maghemite has been altered and either replaced by

paramagnetic iron oxyhydroxides like goethite, or removed by dissolu-

tion effects.

Results from archaeological excavations within

limestone caves and deep rock shelter s

Correlations

Excellent correlations have been shown among archaeological sites in

Europe using MS (Ellwood et al., 2001b). These sites range in age

from 3000

B.P.to50000 B.P. and are based on a modified version

of the graphic correlation method (Shaw, 1964) and available

C dates

from each site for which data are reported. MS high (warm) and low

Figure M94 (a) Interior excavations in Scladina Cave, Belgium. (b) MS of Scladina Cave sediments. The data points correspond to

samples from several overlapping sections within the cave that were correlated with the archaeological levels shown on the left.

MS measurements have then been used to construct the MS bar log which in turn is compared with the marine oxygen isotope

results for the last 130 ka (Imbrie et al., 1984; Martinson et al., 1987). For comparison with the MS results, the d

O, as a function of

time, equivalent to a bar log that represents the times of relatively warm (hatched) and relatively cool (white) climates. The d

O isotope

stage number assignments are based on the dates obtained by Otte et al. (1993, 1998). The ages for the samples from the cave, in

thousands of years, and their locations within the sequence are indicated, as are the lower Mousterian level (lithic artifacts produced by

Neanderthals) and the location of Neanderthal skeletal remains. The assigned ages for the top and the bottom of the MS warm zones

have been based on the assumption that the top and bottom of each level are equivalent in age to the d

O ages of the isotope substages

(from Ellwood et al., 2004).

570 MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD

(cool) values from individual sequences are differentiated diagramma-

tically by developing bar logs (an example is given in Figure M94b).

To these bar log diagrams are added archaeological industries, levels,

and available dates. The archaeological lithic industries and dates

constrain the MS bar logs to a chronology that allows comparison to

a previously developed (and continuously evolving) standard sequence

(composite standard, CS). The boundaries between warm and cool MS

climate proxies are assigned ages by comparison to the CS making it

possible to compare these boundaries to other cave sequences.

Climate cycles

Results from MS data show cycles, which correlate with climate

cycles. Both short- and long-term cycles have been resolved from cave

sequences. Neoglacial cycles, with a 2500 year cyclicity, well known

from Quaternary climate studies and also reported from work on

the Greenland GISP2 Ice Core, have been observed in a number

of caves in Europe (Ellwood et al., 1996). Longer term cyclicities

(Figure M94b, from Ellwood et al., 2004) have been correlated to

20 ka fluctuations observed in the marine oxygen isotopic record

(Imbrie et al., 1984).

Relative dating

Relative dating can be accomplished by comparing known dates with

MS bar log age assignments, either by using dates assigned by compar-

ison with the MS standard sequence (CS) or with the marine isotope

curve. Results from this method have been used to date Neanderthal

remains in Scladina Cave in Belgium (Ellwood et al., 2004). This result

was further tested using high-precision radiometric dating and found to

be consistent within the errors assigned by these methods.

Studies of loess

MS work on Chinese loess sequences in the early 1980s showed a

strong correlation with the marine oxygen isotopic record and MS

variability. Much of this work has been summarized by Evans and

Heller (2003). It was observed that during interglacial periods MS

values were high, while during glacial times MS values were low. This

was attributed to differences in pedogenic rates, with the production of

more ferrimagnetic minerals during interglacial times. In the main, at

least for the Chinese sequences, the MS variations in loess appear to

provide a proxy for climate fluctuations. However, there are a number

of problems with such a simple interpretation. The original protolith from

which the loess was derived has an important effect on final MS variabil-

ity, and in some cases has been shown to produce an inverse MS relative

to the Chinese loess sequ ence s, with inter g lacials showing lower MS

values than during glacial times. This is due to erosion of coarse magne-

tite from an original, high-magnetite-content protolith, and this magnetite

is accumulated in loess sequences durin g glacial times. However , pedo-

genesis during interglacial times oxidizes the original magnetite, produ-

cing a suite of magnetic minerals with a significantly weaker MS. In

addition, during interglacials, deep pedogenesis can affect underlying

loess sequences, causing ambiguities in the observed MS results.

Grain size estimates based on frequency

dependence of MS

There is a grain-size dependence to MS that results from measurement

at varying frequencies. This effect has been loosely employ ed in some

rock magnetic studies using balanced coil instruments that can mea-

sure MS at varying frequencies. There is only one such instrument cur-

rently commercially available and it allows measurements at two

frequencies, 500 Hz and 5 kHz. Measurements of the same sample

at both frequencies have been compared and results interpreted to

represent grain size differences between samples.

There are at least three significant problems with this method.

1. An instrument that only operates at two frequencies is not capable

of measurement at intermediate frequencies where responses in

varying size ranges may be important.

2. In general, it is assumed that while coarse, multidomain ferrimagnetic

grains respond quickly to inducing magnetic fields at both low and

high frequencies, the more stable, single-domain or pseudosingle-

domain grains do not respond as quickly to high-frequency-inducing

fields. This results in a lag in higher frequency fields that is reflected

in lower MSvalues for samples containing thesegrains. However, there

is a third size range of ferrimagnetic particles, the ultrafine-grained

superparamagnetic grains produced during pedogenesis. While these

are generally unimportant in remanence studies, due to Brownian

motion effects, they respond rapidly to both low- and high-frequency-

inducing magnetic fields, in much the same way as do multidomain

grains, and thus are not distinguishable from larger grains.

3. MS originates mainly from a large range of paramagnetic and ferri-

magnetic grain compositions and grain sizes. Because only two fre-

quencies are used in frequency-dependence studies, the true

behavior of grains at higher, intermediate, and lower frequencies

is poorly known or unknown in most cases. Clearly, much more

work needs to be done before reliable interpretations of MS data

from variable frequency measurements are useful.

Brooks B. Ellwood

Bibliography

Crick, R.E., Ellwood, B.B., El Hassani, A., Feist, R., and Hladil, J.,

1997. Magnetosusceptibility event and cyclostratigraphy (MSEC)

of the Eifelian-Givetian GSSP and associated boundary sequences

in North Africa and Europe. Episodes, 20: 167–175.

Crick, R.E., Ellwood, B.B., El Hassani, A., Hladil, J., Hrouda, F., and

Chlupac, I., 2001. Magnetostratigraphy susceptibility of the Pridoli-

Lochkovian (Silurian-Devonian) GSSP (Klonk, Czech Republic)

and a Coeval sequence in Anti-Atlas Morocco. Palaeogeography

Palaeoclimatology Palaeoecology, 167:73–100.

da Silva, A.-C., and Boulvain, F., 2002. Sedimentology, magnetic sus-

ceptibility and isotopes of a Middle Frasnian carbonate platform:

Tailfer section, Belgium. Facies, 46:89–101.

Dinares-Turell, J., Baceta, J.I., Pujalte, V., Orue-Etxebarria, X.,

Bernaola, G., and Lorito, S., 2003. Untangling the Palaeocene cli-

matic rhythm: as astronomically calibrated Early Palaeocene magne-

tostratigraphy and biostratigraphy at Zumaia (Basque basin, northern

Spain). Earth and Planetary Science Letters, 216:483–500.

Ellwood, B.B., and Ledbetter, M.T., 1977. Antarctic bottom water

fluctuation in theVema Channel: effects of velocitychanges on particle

alignmentandsize. Earth and Planetary Science Letters, 35:189–198.

Ellwood, B.B., Petruso, K.M., and Harrold, F.B., 1996. The utility of

magnetic susceptibility for detecting paleoclimatic trends and as a

stratigraphic correlation tool: an example from Konispol cave sedi-

ments, SW Albania. Journal of Field Archaeology, 23: 263–271.

Ellwood, B.B., Crick, R.E., and El Hassani, A., 1999. The magneto-

susceptibility event and cyclostratigraphy (MSEC) method used

in geological correlation of Devonian rocks from Anti-Atlas

Morocco. AAPG Bulletin, 83: 1119–1134.

Ellwood, B.B., Crick, R.E., El Hassani, A., Benoist, S., and Young, R.,

2000. Magnetosusceptibility event and cyclostratigraphy (MSEC)

in marine rocks and the question of detrital input versus carbonate

productivity. Geology, 28: 1135–1138.

Ellwood, B.B., Crick, R.E., Garcia-Alcalde Fernandez, J.L., Soto, F.M.,

Truyols-Massoni, M., El Hassani, A., and Kovas, E.J., 2001a.

Global correlation using magnetic susceptibility data from Lower

Devonian rocks. Geology, 29:583–586.

MAGNETIC SUSCEPTIBILITY (MS), LOW-FIELD 571

Ellwood, B.B., Harrold, F.B., Benoist, S.L., Straus, L.G., Gonzalez-

Morales, M., Petruso, K., Bicho, N.F., Zilhão, Z., and Soler, N.,

2001b. Paleoclimate and intersite correlations from Late Pleisto-

cene/Holocene cave sites: results from Southern Europe. Geoarch-

aeology, 16: 433–463.

Ellwood, B.B., MacDonald, W.D., Wheeler, C., and Benoist, S.L.,

2003. The K-T Boundary in Oman: Identified Using Magnetic Sus-

ceptibility Field Measurements with Geochemical Confirmation.

Earth Planetary Science Letters, 206: 529–540.

Ellwood, B.B., Harrold, F.B., Benoist, S.L., Thacker, P., Otte, M.,

Bonjean, D., Long, G.L., Shahin, A.M., Hermann, R.P., and

Grandjean, F., 2004. Magnetic susceptibility applied as an age-

depth-climate relative dating technique using sediments from Scla-

dina Cave, a Late Pleistocene cave site in Belgium. Journal of

Archaeological Science, 31: 283–293.

Ellwood, B.B., Kafafy, A., Kassab, A., Tomkin, J.H., Abdeldayem, A.,

Obaidalla, N., Willson, K., and Thompson, D.E., 2007. Magnetos-

tratigraphy susceptibility used for high resolution correlation among

Paleocene/Eocene boundary sequences in Egypt, Spain and the

USA. SEPM Special Publication, in press.

Evans, M.E., and Heller, F., 2003. Environmental Magnetism: Princi-

ples and Applications of Enviromagnetics. San Diego, CA:

Academic Press, p. 299.

Hansen, H.J., Lojen, S., Toft, P., Dolenec, T., Yong, J., Michaelsen, P.,

and Sarkar, A., 1999. Magnetic susceptibility of sediments across

some marine and terrestrial Permo-Triassic boundaries. In Proceed-

ings of the International Conference on Pangea and the Paleozoic-

Mesozoic Transition, March 9–11, 1999. China University of

Geosciences, Hubei, China, pp. 114–115.

Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C.,

Morley, J.J., Pisias, N.G., Prell, W.L., and Shackleton, N.J.,

1984. The orbital theory of Pleistocene climate: support from a

revised chronology of the Marine Delta

O record. In Berger, A.L.,

Imbrie,J.,Hays,J.,Kukla,G.,andSaltzman,B.(eds.),Milankovitch

and Climate, Part I. Boston, MA: Reidel, pp. 269–305.

Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, T.C., and Shackleton,

N.J., 1987. Age dating and the orbital theory of the ice ages: develop-

ment of a high-resolution 0 to 300 000-year chronostratigraphy.

Quaternary Research, 27:1–29.

Otte, M., Toussaint, M., and D., Bonjean 1993. Découverte de restes

humains immatures dans les niveaux moustériens de la grotte Scla-

dina à Andenne (Belgique). Bulletin Mémoires de la Société d’An-

thropologie de Paris, 5: 327–332.

Otte, M., Patou-Mathis, M., and Bonjean D. (eds.), 1998. Recherches

aux grottes de Sclayn, Volume 2, L’Archéologie. Liège, ERAUL 79,

Service de Préhistoire, Université de Siége.

Sachs, S.D., and Ellwood, B.B., 1988. Controls on magnetic grain-size

variations and concentrations in the Argentine Basin, South Atlan-

tic Ocean. Deep Sea Research, 35: 929–942.

Shaw, A.B., 1964. Time in Stratigraphy. New York: McGraw-Hill,

365 pp.

Thoa, N.T.K., Huyen, D.T., Ellwood, B.B., Lan, L.T.P., and Truong,

D.N., 2004. Determination of Permian-Triassic boundary in lime-

stone formations from Northeast of Vietnam by paleontological

and MSEC methods. Journal of Sciences of the Earth, 26:

222–232.

Cross-references

Anisotropy of Magnetic Susceptibility (AMS)

Archeomagnetism

Iron Sulfides

Magnetic Anisotropy, Sedimentary Rock and Strain Alteration

Magnetic Susceptibility

Magnetization, Natural Remanent (NRM)

Magnetostratigraphy

Polarity

MAGNETIZATION, ANHYSTERETIC REMANENT

Anhysteretic remanent magnetization (ARM) is a magnetization an

assemblage of magnetic particles acquires when it is subjected to an

alternating field (AF) of gradually decreasing amplitude (H

) with a

constant decrement (DH

/cycle) simultaneously with a steady, unidir-

ectional DC field (H

). The ARM is measured when both AF and

DC fields are zero. Typically, the DC field is maintained while the AF

is slowly ramped down to zero, and then reduced to zero. For isotropic

samples, the direction of the ARM is parallel to H

but the intensity

will depend on the amplitudes and relative orientations of the AF and

DC fields. The maximum ARM intensity occurs when H

and H

are parallel. For a standard laboratory experiment, H

 H

with

¼ 100–200 mT and H

¼ 0.05–0.1 mT. H

is also referred to

as a bias field because it produces a statistical preference or biases the

direction of ARM along H

. A special case of ARM is a partial

ARM (pARM) produced by a DC field applied only for a limited time

between two AF field values (0 H

, H

H

max

). A related parameter is

the ARM susceptibility, w

arm

¼ dM

arm

/dH

! 0) where M

arm

is the ARM intensity. Typically, w

arm

is determined from a single

measurement using one DC field value, w

arm

¼ M

arm

. For mass-

normalized ARM (in A m

–1

) per unit DC field (in A m

–1

), the SI

unit for ARM susceptibility is kg m

–3

. For volume-normalized ARM

(in A m

–1

), the SI unit for ARM susceptibility is dimensionless.

With the possible exception of lightning-induced remanence (see

Magnetization, isothermal remanent (IRM)), ARM is not a naturally

occurring remanence like TRM, DRM, or CRM. However, labora-

tory-produced ARM is used in a variety of applications in paleomag-

netism. For example, ARM-based methods have been developed to

(1) estimate absolute paleointensity from igneous rocks and relative

paleointensity from lake and marine sediments; (2) characterize mag-

netic carriers and determine domain state and grain size; (3) detect

magnetic fabrics in rocks and sediments; and (4) study the fundamen-

tal aspects of magnetism. Anhysteretic magnetization also forms the

basis for AC-bias magnetic recording.

The combined action of an AF and DC field (H

 H

) is much

more efficient in producing a remanence that is more resistant to

demagnetization than an isothermal remanent magnetization (IRM),

which results from the application of a DC field alone. Two fundamen-

tal characteristics of ARM distinguish it from an IRM given in the

same DC field: (1) M

arm

 M

irm

, and (2) in order to demagnetize the

ARM, AF fields approximately equal to the coercivity (H

H

) must

be applied, whereas AF or DC fields equal to H

are needed to

demagnetize the IRM. In a classic study, Rimbert (1959) was the

first to demonstrate analogous properties between ARM and TRM

(see Magnetization, thermoremanent (TRM)). Therefore, ARM could

be used as a surrogate for TRM in laboratory experiments (Rimbert,

1959; Gillingham and Stacey, 1971; Levi and Merrill, 1976; Dunlop

and Argyle, 1997), thus avoiding the need to heat the sample to high

temperatures and the possibility of mineralogical alteration. The AF

field for ARM acquisition acts as the randomizing agent in analogous

fashion as temperature for TRM acquisition. Consider a single-domain

(SD) grain with uniaxial anisotropy and coercivity, h

. The following

illustration is applicable whether we are dealing with moment rotation

between easy axes in SD particles or wall displacement in multi-

domain particles. In this case, the magnetic moment is pinned in the

positive direction along the anisotropy axis and will undergo an irrever-

sible rotation into the opposite orientation when a field greater than

is applied along the negative direction. Starting from a large AF

field such that H

is greater than h

, the AF field overcomes the aniso-

tropy barrier and the magnetization will follow the oscillating field, i.e.,

the large AF field initially saturates the sample during each cycle.

However, once H

< h

, the particle moment is blocked into a stable

position along the anisotropy axis closest to the DC field direction. As

is reduced to zero, little or no change in magnetization occurs.

For an ensemble of SD particles with randomly oriented anisotropy

572 MAGNETIZATION, ANHYSTERETIC REMANENT

axes, a statistical bias is produced resulting in a net remanence along

.IfH

¼ 0 during the application of the AF field, the net rema-

nence will be zero and the process is simply AF demagnetization (see

Demagnetization).

ARM can be an unwanted spurious magnetization produced during

AF demagnetization of natural remanent magnetization (NRM). This

type of ARM can be generated by (1) insufficient magnetic shielding

of external DC fields or (2) distorted or asymmetric AF waveforms

that produce higher-order even harmonics from the fundamental AF

drive frequency (Collinson, 1983). In either case, a small net DC bias

is present during the AF demagnetization process that can produce an

ARM of similar or even greater magnitude, than the initial or partially

demagnetized NRM signal. Therefore, it is important to provide ade-

quate shielding around the AF coils to attenuate external DC fields,

such as the Earth’s magnetic field or laboratory fields. This is usually

achieved with multilayer, mu-metal shielding. Asymmetries in the AF

waveform can be attenuated by tuning the AF coil to the fundamental

frequency. Spurious ARM can also be significantly reduced by doing a

double-demagnetization at each demagnetization step. In this techni-

que, the sample is (1) demagnetized along three orthogonal axes

(þx, þy, þz), (2) NRM measured, (3) demagnetized along –x, –y,

and –z, (4) NRM remeasured, and (5) the vector sum of steps (2)

and (4) is determined. Collinson (1983) provides detailed reviews of

AF and ARM instrumentation.

Experimental properties of ARM

The grain size dependences for weak-field ARM (H

< 0.2 mT) and

its median destructive field (MDF) in magnetite from various sources

are shown in Figure M95. The intensity of ARM peaks within the

single-domain size range (d < 0.1 mm) and then drops off approxi-

mately as d

–n

(n ¼ 0.5 –1) for pseudosingle-domain (PSD) and multi-

domain grain sizes. Below 0.05 mm, ARM intensity also falls with

decreasing grain size as particles become superparamagnetic and their

Figure M95 (a) ARM susceptibility as a function of magnetite grain size. (b) Median destructive fields of SIRM (symbols in left

legend) and ARM (symbols in right legend). The experimental data are synthetic and biogenic magnetites from various sources.

(Reprinted from Egli and Lowrie, 2002, with kind permission of the authors).

MAGNETIZATION, ANHYSTERETIC REMANENT 573

ability to acquire remanence is lost. This grain size trend is similar in

form to other grain size-dependent trends in magnetic properties (e.g.,

TRM, SIRM, H

) and is a hallmark of the change in micromagnetic

states from superparamagnetic to stable single- to multidomain configura-

tions with increasing grain size. For single-domain particles, ARM

and TRM have many similar experimental characteristics (Dunlop and

Özdemir, 1997) including: (1) comparable resistance to AF demagnetiza-

tion (Levi and Merrill, 1976), (2) linear dependence on H

for weak

fields, H

< 0.2 mT (Dunlop and Argyle, 1997), and (3) pARMs that

obey laws analogous to the three Thellier laws for partial TRM: additivity,

reciprocity, and independence (Yu et al., 2002a,b, 2003; see Magnetiza-

tion, thermoremanent (TRM)). These experimental observations form

the basis for ARM methods of paleointensity determinations (e.g., Shaw,

1974; Kono, 1978; T au xe et al., 1995), which avoid multiple heating to

high temperatures that are necessary for the conventional Thellier-Thellier

method for paleointensity (Dunlop and Özdemir, 1997).

Whereas qualitative similarities exist between ARM and TRM as

exemplified by properties (1)–(3) in the previous paragraph, TRM is

still a more efficient remanence acquisition process than ARM for a

given DC field. This is demonstrated by the grain size dependence

of the ratio of intensities for weak-field TRM and ARM in magnetite

(Figure M96), which show TRM/ARM ratios greater than 1 for all

grain sizes (Levi and Merrill, 1976; Dunlop and Argyle, 1997). While

the TRM/ARM ratio is between 1 and 2 and practically independent of

grain size above 1 mm, ratios exceeding 10 occur within a narrow

size range between 0.1 and 0.5 mm. This contrast in behavior for grain

sizes above and below d ¼ 1 mm is possibly related to different

micromagnetic remanent states produced by field-cooling from above

the Curie temperature for TRM (remanence acquired at elevated tem-

perature) or field-cycling at room temperature for ARM (remanence

acquired at room temperature). For instance, micromagnetic models

suggest that for the 0.1–0.5 mm size range, high-moment metastable

single- or two-domain states are blocked in during TRM acquisition,

but only low-moment vortex ground states result from the AF cycling

during ARM acquisition (Dunlop and Argyle, 1997). In contrast, for

d > 1 mm magnetites, TRM/ARM ratios between 1 and 2 suggest that

similar micromagnetic states, probably with many domains, are produced

during either field-cooling or field-cycling.

ARM theories in single- and multidomain particles

For an assembly of noninteracting SD particles undergoing only field-

assisted switching (i.e., no thermal fluctuations or time-dependent

switching), the intensity of ARM is theoretically independent of H

(and will equal the saturation remanence) and the initial ARM susceptibility

is infinite (Wohlfarth, 1964). However, these theoretical predictions are

not observed in synthetic or natural materials. For instance, Figure

M97 clearly demonstrates that the shape of the ARM acquisition curve

is strongly affected by particle concentration, where high concentrations

of particles result in stronger dipolar (magnetostatic) interactions

between magnetic particles. The effect of increasing interactions is to

lower the efficiency of ARM acquisition. Dunlop and West (1969) used

the Néel-Preisach model for interactions to explain weak-field ARM

and TRM behavior in several synthetic and natural samples containing

Figure M96 The ratio of weak-field TRM and ARM, and susceptibility ratio (w

trm

arm

) as a function of magnetite grain size. The

experimental data are synthetic magnetites from various sources. (Reprinted from Dunlop and Argyle, 1997, with kind permission of

the authors).

574 MAGNETIZATION, ANHYSTERETIC REMANENT

SD particles of iron oxides. In the Néel-Preisach model, interactions are

accounted for in particle assemblies by assuming each particle has an

asymmetric rectangular hysteresis loop, with the amount of asymmetry

being proportional to the strength of the interaction field. A distribution

of interaction fields, determined by separate experiments, is then used

to model the ARM acquisition curves. A shortcoming to this type of

model is that moment switching occurs instantaneously, while it is

known that time-dependent switching via thermal fluctuations is impor-

tant in SD and small PSD particles according to the Néel theory of ther-

mally activated magnetization (Dunlop and Özdemir, 1997). ARM

theories for SD particles incorporating interparticle, dipolar interactions,

and thermal fluctuations have been proposed by Jeap (1971), Walton

(1990), and Egli and Lowrie (2002). The following equation was derived

by Jeap (1971) for ARM for interacting SD grains:

arm

¼ tanhðm

=kTðBH

 lp

arm

B ðM

ÞðT

1=2

where p

arm

¼ M

arm

, M

is the saturation magnetization, M

is the

saturation remanence, k is Boltzmann’s constant, V is grain volume, T

is absolute temperature, l is the mean interaction field, and the

subscripts 0 and b refer to room temperature (T

) and the blocking

temperature (T

). A field-blocking condition, derived from Néel’s

relaxation theory, is assumed whereby the alternating field necessary

to reverse the magnetization in a given time interval corresponds to a

relaxation time that makes ln( f

t) ¼ 25, where f

is the frequency fac-

tor usually taken as 10

Hz. Note, that when B ¼ 1, the equation

reduces to the SD equation for TRM with interactions.

In the most recent theoretical work, Egli and Lowrie (2002) show

that even in noninteracting systems, ARM can be described solely

by thermal fluctuation effects. Therefore, in dilute systems, such as

in many rocks and sediments containing low concentrations of mag-

netic carriers, ARM properties are controlled by intrinsic parameters

(coercivity and grain volume) rather than by interactions. The SD

ARM equation derived by Egli and Lowrie (2002) is

arm

¼ tanhðw

arm

¼ 1:797m

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ



3=2

1=3

0:35F

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ



3=2

where H

is the anisotropy field, ƒ

is the AF frequency, and DH

the decay rate of AF. This model predicts an increase in ARM intensity

proportional to d

within the SD size range (d < 60 nm). ARM inten-

sity is also predicted to be weakly dependent on the characteristics of

the AF. Experimental results on natural and synthetic samples gave

results compatible with theoretical predictions.

For non-single-domain grains, Gillingham and Stacey (1971)

proposed a multidomain theory of ARM incorporating domain wall

translation under the influence of the self-demagnetizing field. The

self-demagnetizing field itself can be thought of as an internal interac-

tion field. Theory predicts that for DC fields in which domain wall

translation occurs (H

< H

), M

arm

/ H

. Reasonable agreement

has been obtained between theory and experiment for magnetite grains

with d > 40 mm, which most likely contain true multidomain states.

However, experimental ARM intensities were significantly larger than

predicted for d < 40 mm or for particle sizes within the pseudosingle-

domain size range (Gillingham and Stacey, 1971). Dunlop and Argyle

(1997) developed a PSD theory for ARM by combining theories incor-

porating a wall displacement term linearly dependent on H

(an MD

process) and a wall moment term proportional to the hyperbolic tan-

gent function (an SD process). Reasonable fits to experimental ARM

acquisition data for magnetite particles in the 0.2–0.5 m m size range

were obtained with this theory.

Applications

Absolute paleointensity

Shaw (1974) proposed an alternative paleointensity method (now

called the Shaw method) to the conventional Thellier-Thellier method,

employing ARM and a single-step laboratory TRM heating. First, the

NRM of the sample is stepwise AF demagnetized. Then an ARM

(designated ARM1) is imparted and subsequently AF is demagnetized.

Next, the sample is given a total TRM by heating the sample once

above its Curie temperature and cooling in the same DC field as used

in the ARM experiment. The TRM is then AF demagnetized using the

same step as the one used for ARM. Finally, a second ARM (desig-

nated ARM2) is given and AF is demagnetized. If the sample does

not experience any chemical alteration during the TRM step then a

plot of ARM1 vs ARM2 using the peak AF as the common parameter

should be linear with a slope of 1.0. In practice, a portion of the coer-

civity spectrum is selected that gives the best fit to a line of slope 1.0.

Within this coercivity window it is assumed that the TRM is likewise

unchanged and a plot of NRM vs TRM using the peak AF as the com-

mon parameter yields an estimate of the paleofield. Idealized Shaw

diagrams are shown in Figure M98. This method is considerably

quicker than the Thellier-Thellier method because less heating is

required. Furthermore, samples are heated only once above the Curie

temperature, which reduces the potential for chemical alteration that

can occur during the multiple heatings used in the Thellier-Thellier

method. Kono (1978) showed experimentally that the Shaw method

could be as reliable as the Thellier-Thellier method for samples that

undergo little or no chemical change during heating. However, a disad-

vantage of the Shaw method is that the entire experiment is unusable

if alteration takes place during the single-step heating to the Curie

temperature. Thermochemical alteration of the sample can lead to

changes in its TRM capacity and produce erroneous paleointensity

results. Modifications of the original Shaw method have been pro-

posed to correct for limited amounts of thermochemical alteration

Figure M97 ARM acquisition curves for samples containing

dispersed grains of magnetite in different concentrations. The

curves are normalized by saturation remanence and the magnetic

field is the DC bias field applied during AF demagnetization. The

magnetite sample contains a mixture of SD and PSD grains.

(Redrawn from Sugiura, 1979).

MAGNETIZATION, ANHYSTERETIC REMANENT 575