вход по аккаунту


Thermally activated mineralogical transformations in archaeological hearthsinversion from maghemite ╨Ю╤ЦFe2O4 phase to haematite ╨Ю┬▒Fe2O4 form.

код для вставкиСкачать
Archaeological Prospection
Archaeol. Prospect. 13, 207–227 (2006)
Published online 28 March 2006 in Wiley InterScience ( DOI: 10.1002/arp.277
Thermally Activated Mineralogical
Transformationsin Archaeological
Hearths:Inversion from Maghemite
Fe2O4 Phase to Haematite
Fe2O4 Form
Archaeo-Physics, LLC, 4150 Digit Ave, #110 Minneapolis, MN 55406, USA
Statistical Research Inc., PO Box 31865,Tuscon, AZ 85751, USA
Courson Oil and Gas,1800 South Main Street, PerrytonTX 79070, USA
A series of laboratory experiments were conducted in an effort to understand why magnetic field
gradient survey techniques failed to detect hearths at a prehistoric archaeological site in southern
California. The study used various methods of environmental magnetism to examine the effects of
exposing soil samples to a temperature of 650 C over a period 26 h. Results of the study indicate that
the failure was associated with a reduction in soil magnetic susceptibility to below background levels
within hearth soils. This reduction was due to high-temperature transformation of iron oxides from a
highly magnetic form to a relatively non-magnetic form. The reduction in susceptibility is thought to
have proceeded via the oxidation of primary (lithogenic) magnetite, Fe3O4 to maghemite, Fe2O4
followed by the inversion of maghemite to haematite, Fe2O4. The study suggests that in some
instances high temperature inversion can reduce the proportion of ferrimagnetic minerals within a
hearth to below initial concentrations (resulting in a negative magnetic field gradient anomaly,
the opposite of what is normally expected). A field experiment was also conducted to determine what
soil temperatures might be achieved 2 cm beneath a hearth.The experiment recorded temperatures
ranging from approximately 400 C to 650 C, with an average temperature of about 470 C. Soil
colour changes and magnetic susceptibility enhancement observed at the conclusion of the field
experiment indicate that these temperatures were sufficient to activate some mineralogical
changes, possibly including inversion to haematite. The implications of high temperature inversion
to archaeological prospection are discussed, as is a potential archaeological application.
Copyright 2006 JohnWiley & Sons,Ltd.
Key words: magnetic susceptibility; oxidation; inversion; fractional conversion; archaeological
This paper describes a series of laboratory experiments that were conducted on soil samples
from the West Bluff Project Area, located near the
* Correspondence to: D. Maki, Archaeo-Physics, LLC, 4150
Digit Ave, # 110, Minneapolis, MN 55406, USA.
Copyright # 2006 John Wiley & Sons, Ltd.
city of Los Angeles, in southern California. The
measurements were undertaken in an effort to
understand why magnetic field gradient survey
failed to detect hearth features that were discovered after mechanical stripping had removed
the non-cultural soil overburden from this
prehistoric archaeological site. The hearths had
been heavily bioturbated, but the displacement
of culturally modified soil and stones from the
Received 23 March 2005
Accepted 19 December 2005
hearths appears to have been relatively localized.
Turbation of the features suggests that the
thermoremanent magnetization of the hearths
may have been randomized, however, many
features appeared sufficiently intact to expect a
positive induced magnetization anomaly resulting from culturally enhanced soil within the
The paper will seek to explain why enhancement of magnetic minerals in soil associated with
these hearths did not create significant and readily detectable magnetic anomalies. Controlled
tests examined the initial magnetic susceptibility
enhancement in soil samples upon exposure
to high temperatures. This was followed by an
examination of the subsequent mineralogical
transformations that occurred with prolonged
exposure to a temperature of 650 C. Finally,
a field experiment sought to determine whether
650 C is a realistic temperature that might
be expected of past conditions in soil under
archaeological hearths.
Susceptibility enhancement
and inversion mechanisms
Several different mechanisms can cause an
enhancement (increase) in the concentration of
magnetic minerals in archaeological features (or
archaeological soils), including fire and pedogenic production of ferrimagnetic minerals by
organic and inorganic pathways. Dalan and
Banerjee (1998) and Weston (2002) provide summaries of soil magnetic susceptibility enhancement pathways in archaeological contexts, and
Peters et al. (2001) and McClean and Kean (1993)
discuss the contribution of magnetic minerals
from wood ash to archaeological features and
soils. Controlled experiments conducted by
Linford and Canti (2001) demonstrated that the
combined effects of wood ash and thermal
enhancement of soil susceptibility resulting
from short-term camp fires often produce readily
distinguishable magnetic anomalies.
This paper will focus on just one enhancement
pathway—fire (also referred to as the burning
mechanism). Exposure of soil to high temperatures typically causes an enhancement (increase)
in the magnetic susceptibility. This enhancement
Copyright # 2006 John Wiley & Sons, Ltd.
D. Maki, J. A. Homburg and S. D. Brosowske
is due to conversion of iron oxides and hydroxides in the soil from a weakly magnetic form to
a strongly magnetic form; the conversion often
proceeding via reduction to magnetite, followed
by reoxidation to maghemite. Several factors
may influence the level of enhancement that is
achieved in the soil surrounding an archaeological hearth. For example, anaerobic conditions are
necessary to provide a reducing environment,
and oxygen is necessary if subsequent reoxidation is to occur. Hence the atmospheric conditions within the soil under and around a hearth is
a variable affecting the level of enhancement. Soil
atmospheric conditions are, in turn, associated
with the porosity of the soil, size and tortuosity
of soil pores, shape and ventilation of the hearth,
thickness of ash deposits lying on the soil, and
the amount of water and organic matter initially
present in the soil matrix. Additional variables
that influence the level of enhancement occurring
at high temperature include the availability of
iron minerals in the upper soil horizon, temperatures achieved during heating, and the thermal
conductivity and diffusivity of the soil.
Recent research suggests that the level of enhancement that may be achieved within a soil is
finite, and may be partially or wholly reversible
(Weston, 2002; Crowther, 2003). For example,
heating experiments conducted by Weston
(2002) measured an increase in soil magnetic
susceptibility of up to 171 times the initial susceptibility value after ignition at 650 C, however,
the magnetic susceptibility of these samples
significantly decreased after reheating to 850 C.
This reduction in susceptibility may have been
due to high temperature oxidation and inversion.
High temperature oxidation and inversion
have been the subject of numerous studies in
non-archaeological fields of interest (Bando et al.,
1965; Tite and Linington, 1975; O’Reilly, 1983;
Ozdemir and Banerjee, 1984; Ozdemir, 1987,
1990; Ozdemir and Dunlop, 1993; Dunlop and
Ozdemir, 1997; Adnan and O’Reilly, 1999; Brown
and O’Reilly, 1999). These studies have shown
that at high temperatures maghemite, Fe2O4 is
metastable, and can invert to stable haematite,
Fe2O4. Inversion is a thermally activated process, with inversion temperatures ranging from
250 to 900 C. The inversion rate appears to
vary with temperature, magnetic mineral grain
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
size, the degree of oxidation, pressure and lattice
impurities such as substitution by aluminum or
titanium. Naturally occurring maghemites often
have much higher inversion temperatures than
synthetic maghemites.
The inversion of maghemite is structural
in origin and takes place without a change in
the bulk chemical composition. Maghemite is
arranged in a spinel structure, which is responsible for its ferrimagnetic properties. Haematite is
arranged in a rhombohedral structure and
has canted-antiferromagnetic properties. This
structural transformation is very significant
as the magnetic susceptibility of canted-antiferromagnetic haematite is approximately 200 times
weaker than that of ferrimagnetic minerals (Evans
and Heller, 2003). Inversion to haematite is also
associated with the reddening of soil colour often
observed within archaeological hearths.
West Bluff Project Area soils
and sampling strategy
Archaeological investigations have been conducted since the mid-1980s at the West Bluff
Project Area, which encompasses the Del Rey
Site (CA-LAN-63), the Bluff Site (CA-LAN-64)
and CA-LAN-206 (Van Horn, 1986, 1987;
Altschul, 1997, 1999; Van Horn and White,
1997; Altschul et al., 2000). In September of
2000, a geophysical investigation of CA-LAN-63
and CA-LAN-64 was undertaken. The investigation consisted of magnetic field gradient survey
over an area of 3.14 ha using a Geoscan FM36
fluxgate gradiometer operated at the 0.1 nT sensitivity level. Data were collected using a transect
spacing of 0.50 m, with eight samples per metre
along each transect (16 samples m2). The results
of the magnetic field gradient survey are shown
in Figure 1.
After the geophysical investigation, portions of
the sites were mechanically stripped of noncultural topsoil overburden and several archaeological features were revealed. The majority of
these features consisted of concentrations and
scatters of burnt stone, and associated flaked
and ground stone tools and hammer stones,
which were tentatively identified as hearths or
hearth clean-outs (Douglass and Altschul, 2003).
Copyright # 2006 John Wiley & Sons, Ltd.
Other features include artefact concentrations
and scatters, human burials, refuse deposits, a
faunal bone concentration and one historicalperiod feature. Radiocarbon dates obtained
thus far indicate that CA-LAN-206 dates to the
Early Period (ca. 6500–7000 yr BP) and that the
other two sites date to the Middle Period (ca.
1800–2800 yr BP). Research on these three sites is
ongoing, but previous investigators have suggested that the sites functioned as temporary
base camps (Van Horn and White, 1997) or as
parts of an interdependent settlement system of
sites located both on the bluff and in the wetlands below (Grenda et al., 1994).
All three sites are located on the Westchester
Bluffs about 2 km north of Los Angeles International Airport and about 1 km south of Ballona
Creek. Soils consist of loamy sands and sandy
loams formed in Quaternary aeolian sediments
of the El Segundo Sand Hills. These aeolian
sediments originate from beach deposits on the
Pacific coast 2 km to the west that have been
reworked extensively during the Holocene to
form sheets of sand. The sand grains from the
very fine fraction are dominated by quartz and
magnetite (approximately 18% quartz and 67%
magnetite), with lesser quantities of zircon, tourmaline and calcite (Mbila and Homburg, 2000).
The coarser sand is dominated by quartz. The
sandy soils are well drained, with slow to rapid
runoff and moderate permeability. Soils in most
of the project area consist of thin brown to greyish brown A horizons overlying a sandy subsoil
with thin, discontinuous clay lamellae. A shell
midden deposit of dark brown to dark greyish
brown sandy loam approximately 1 m thick was
found on the shoulder of a swale in the Del Rey
Site, and cultural features were concentrated on
the adjacent summits.
A post-excavation analysis determined that
the correlation between hearth (and hearth
clean-out) features and magnetic anomalies was
quite poor. Several soil samples were collected
from a geomorphology test unit in an effort to
understand why enhancement of magnetic
minerals in soil associated with these hearths
did not result in significant induced magnetic
anomalies. Unfortunately no soil samples from
excavated features were available for testing. The
soil samples came from a profile located in
Archaeol. Prospect. 13, 207–227 (2006)
Figure1. Location ofgeophysical surveyareas and site boundaries at theWest Bluff Project Area.The magnetic field gradient surveyresults are depicted as grey-scale images.
The large rectilinear magnetic lows identify the location of machine excavated trenches from previous archaeological investigations. Signal clutter created by near-surface
iron debris, igneous rock and induced magnetic fields associated with plough furrows dominate the survey results.
Copyright # 2006 John Wiley & Sons, Ltd.
D. Maki, J. A. Homburg and S. D. Brosowske
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure 2. Trench1001unit profile.
Trench 1001, located near the centre of Bluff Site
(Figure 2). The soil grain size, organic content
and pH from a vertical profile within Trench
1001 are provided in Figure 3, and common soil
magnetism parameters are provided in Figure 4
(see next section for a brief description of these
Throughout the West Bluff Project Area the
top of the main occupation horizon ranged from
20 to 38 cm below surface. A soil sample (Sample
A) was selected from Trench 1001 at a depth of
35 cm below surface (A1 soil horizon) for the
high temperature experiments described in the
following section.
Figure 3. Organic content, grain size and pH data from Trench 1001. High temperature experiments were conducted on soil
samples from 35 cm below surface.
Copyright # 2006 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
Figure 4. Soil magnetism parameters fromTrench1001.
High temperature experiments
High temperature laboratory experiments were
conducted simultaneously on Sample A and
additional soil samples from several archaeological sites for comparison. The additional sites
are briefly described as follows: 48CA3030—the
3030 Winchester Site is a Late Prehistoric I period
occupation in northeast Wyoming and soils at
the site consist of sandy silts; 11S1131—the
Grossmann Site is an early Mississippian period
settlement located in the uplands about 18 km
east of Cahokia in Illinois and soils at the site
consist of silty clays; 31CD218—a prehistoric
occupation located in central North Carolina
Copyright # 2006 John Wiley & Sons, Ltd.
and soils at the site consist of sandy loams;
34BV100—the Odessa Yates Site is a Plains Village period settlement located in western
Oklahoma and soils at the site consist of a clay
loams; LA106780—a multicomponent Jornada
Mogollon habitation in south-central New
Mexico and soils at the site consist of silty loams.
An initial high temperature experiment compared the enhancement of magnetic susceptibility in soil samples from several of these sites
by applying a ‘maximum conversion’ procedure
as described by Clark (1996). The procedure was
slightly modified in that a temperature of 750 C
was used rather than 650 C, and a 2-h oxidation
period was used rather than 1 h. A Bartington
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure 5. A comparison of susceptibility values from five archaeological sites before and after experimental heating.
MS2 AC susceptibility meter was used to measure the low frequency (0.46 kHz) magnetic susceptibility after the experimental heating. The
results of the conversion study are presented in
Figure 5. Susceptibility enhancements were
observed in four of the five samples subjected
to the study. Enhancements ranged from 2 to
63 times the initial susceptibility values. The
largest enhancement occurred in the sample
with the lowest initial susceptibility (31CD218),
whereas the smallest enhancement occurred in
the sample with the highest initial susceptibility
(LA106780). Soil Sample A was atypical in that
the final susceptibility was approximately 66%
lower than the initial susceptibility.
Initial susceptibility values suggest that unheated soil from the West Bluff Project Area
has a relatively high concentration of magnetic
minerals (Figure 5). A Day plot (Figure 6) of soil
samples from five archaeological sites suggests
that the ferrimagnetic mineral grain size of
Sample A falls within the multidomain (MD) or
Figure 6. A comparison of the domain states of soil samples from five archaeological sites.
Copyright # 2006 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
superparamagnetic (SPM) range, whereas the
remaining four sites were all pseudosingledomain (PSD) (Day et al., 1977).
The frequency dependence of susceptibility
(fd% ¼ [(0.46 kHz 4.60 kHz)/0.46 kHz] 100) of
Sample A was measured using a Bartington
MS2 susceptibility meter in an effort to learn
whether SPM grain sizes might be present. A
relatively low average frequency dependence of
susceptibility of 1.48% (n ¼ 10) suggests that SPM
grains are probably not present and that MD
grains are predominant.
To investigate the reduction in susceptibility
observed in Sample A further, some bulk magnetic properties of the sample were measured as
a function of time spent at 650 C. Variations in
the magnetic mineral concentration and composition with time spent at this isotherm were
tracked by making room temperature measurements of the mass magnetic susceptibility (at
0.46 kHz) and a backfield parameter termed the
S-ratio at regular intervals between exposures to
high temperature.
A detailed description of these properties and
their relevance to environmental magnetism can
be found elsewhere and is beyond the scope of
this paper (Maher, 1986; Thompson and Oldfield,
1986; Hunt, 1991; Dunlop and Ozdemir, 1997;
Evans and Heller, 2003). A brief description of
these parameters, as well as the parameters presented in Figure 4, is provided below.
Mass magnetic susceptibility () is the ratio
of magnetization induced by a small applied
field to the intensity of the magnetizing
field. Susceptibility is roughly proportional
to the concentration of ferrimagnetic minerals within a sample. Susceptibility is also
affected by the magnetic mineral grain size.
(ii) Frequency dependence of susceptibility
(fd) is a measure of the variation in susceptibility with frequency. Extremely small
magnetic minerals in the superparamagnetic (SPM) range contribute significantly
to fd (1–20%), whereas larger grains in
the single domain (SD) to multidomain
(MD) range contribute very little (< 1%).
The parameter is defined as follows: fd
(%) ¼ [(0.46 kHz 4.60 kHz)/0.46 kHz] 100.
Copyright # 2006 John Wiley & Sons, Ltd.
Anhysteretic remanent magnetization
(ARM) is produced by the combined actions of a large alternating field (AF) and a
smaller constant direct current field (DC).
An ARM was imparted by slowly reducing
a peak AF of 100 mT to zero while at the
same time applying a constant DC field of
0.05 mT. The ARM is sensitive to both
magnetic mineral concentration and grain
size. Small magnetic mineral grains in the
SD size range are especially susceptible to
ARM acquisition.
(iv) Saturation isothermal remanent magnetization (SIRM) is the highest level of magnetic remanence that can be given to a
sample through the application of a large
DC field (1000 mT in the present case). The
SIRM responds to the total concentration of
magnetic minerals in a sample. The SIRM is
also affected by the grain size of magnetic
minerals, although to a lesser extent than
(v) The interparametric ratio ARM/SIRM is a
useful indicator of grain size. This ratio
responds preferentially to ferrimagnetic
minerals in the SD size range.
(vi) The S-ratio is a backfield parameter that
measures the degree of loss of remanence
at selected reverse fields. It is an effective
means of measuring relative changes in the
proportion of ferrimagnetic minerals to
antiferromagnetic minerals. Values can
range from þ1 to 1, larger positive values
imply a higher proportion of low coercivity
ferrimagnetic minerals such as magnetite
or maghemite, whereas smaller positive
values (and occasional negative values)
indicate an increasing concentration of
high coercivity canted antiferromagnetic
minerals such as haematite or goethite.
Determination of the S-ratio proceeded as
follows. A forward field of 1 T was applied
using a Princeton Applied Research Vibrating Sample Magnetometer and the remanence imparted by this field was measured
on a 2-G Superconducting Rock Magnetometer. Reverse fields of 100 mT and
300 mT were then applied and the resulting
remanence measured in like manner. The
S-ratio was calculated from the measured
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
remanence imparted by these forward
and reverse fields. The S-ratio calculated from the two reverse fields
are defined as follows: S-ratio100mT ¼
ratio300mT ¼ (IRM300mT/SIRMþ1000mT).
Initial measurements were made on the samples prior to heating (time ¼ 0). Subsequent measurements were then made at room temperature
after the samples had been exposed to heating
cycles. Each cycle consisted of heating approximately 50 g of soil in an open crucible for a period
of 1.3 h. Eighteen minutes of each cycle were
spent ramping up to 650 C, followed by 1 h at
this isotherm. The samples were then rapidly
cooled to room temperature and a portion was
removed and packed into a plastic P-1 cube
(5.28 cm3) and weighed for conversion to mass
normalized values. The magnetic susceptibility
was then measured on a Bartington Instruments
MS2 susceptibility system.
The variation in magnetic susceptibility with
time spent at 650 C is presented in Figure 7. Soils
from all of the five sites presented in Figure 7
experienced an initial susceptibility enhancement. Soils from three of the sites experienced a
reduction in susceptibility from peak values
upon extended heating. The magnetic susceptibility of Sample A dropped to below its initial
susceptibility after approximately 17 h. Soil from
one site continued to experience an enhancement
in susceptibility for the duration of the experiment, while susceptibility from the remaining
site remained relatively constant after the initial
Variations in S-ratios with time spent at 650 C
are presented in Figure 8. The measured S-ratio
of Sample A decreased with time spent at this
isotherm. After 26 h at this temperature the ratio
resulting from both a backfield of 100 mT and
300 mT were lower than the initial (unheated)
ratio, although the change was significantly
greater in the 100 mT backfield ratios. The S-ratio
from the four remaining archaeological sites
experienced an initial increase, followed by relatively constant values. In all four of the comparative sites the S-ratios remained above their initial
(unheated) values.
Copyright # 2006 John Wiley & Sons, Ltd.
The results of the high temperature experiments suggested that after a small initial
enhancement, the proportion of ferrimagnetic
minerals in Sample A decreased with time spent
at 650 C. It is suspected that this was the
result of high temperature oxidation of magnetite to maghemite, followed by the subsequent
inversion (or partial inversion) of maghemite to
haematite. In an effort to test this hypothesis, the
end product mineralogy after heating for 26 h
(Sample A—heated) was compared with the
mineralogy of Sample A prior to heating (Sample
Sample A mineralogy before
and after experimental heating
Sample A (heated) and Sample A (unheated)
were both physically demagnetized by unpacking, mixing the soils, and repacking in P-1 cubes.
The low temperature remanence behaviour of
Sample A before and after experimental heating
was then examined in an effort to better understand the mineralogical transformations that had
occurred at high temperatures. A saturation isothermal remanence magnetization (SIRM) was
applied to both samples at 10 K, after having
been cooled to this temperature in the absence
of a magnetic field. The variation of remanent
magnetizations of the samples were then measured as they warmed to room temperature
(Figure 9) using a Quantum Designs MPMS2
cryogenic susceptometer. Sample A (unheated)
experienced a large drop in SIRM between 100 K
and 120 K. This drop in remanence—known as
the Verwey transition—is diagnostic of magnetite. The suppression of a Verwey transition in
Sample A (heated) suggests magnetite minerals
no longer dominate the magnetic mineral assemblage, although it should be noted that partially
oxidized non-stoichiometric magnetite can also
result in a suppressed Verwey transition. A
Morin transition at 258 K would have provided
evidence of haematite in the end product, but
this was not observed. The lack of a Morin
transition could indicate that haematite was not
present, although this cannot be stated for certain
as the transition may also be suppressed
by substitution by titanium, or in haematite
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
Figure 7. Variation in mass magnetic susceptibility with time spent at 650 C. All data have been normalized by the initial
susceptibility value.
particles with grain sizes less 0.1 mm (Dunlop and
Ozdemir, 1997).
A method of unmixing magnetic mineral
assemblages was next applied in an effort to
further clarify what mineralogical transformations had occurred with extended heating (again,
after physically demagnetizing the samples). The
unmixing method is based on the quantification
of magnetic coercivity components by the analysis of isothermal remanent magnetization (IRM)
Copyright # 2006 John Wiley & Sons, Ltd.
acquisition curves. In this procedure, the first
derivatives of the measured IRM acquisition
curves are mathematically modelled using a
number of separate log-normal probability density functions that are characterized by their
SIRM, mean coercivity and dispersion (Kruiver
et al., 2001; Heslop et al., 2002). The log-normal
probability density functions that most closely
approximated the observed behaviour of the
IRM acquisition curves were selected after a
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure 8. Variation in S-ratio values with time spent at 650 C. The S-ratios represented by solids lines are defined as follows:
(IRM100mT /SIRMþ1000mT). Dashed lines represent S-Ratios defined as follows: (IRM100mT /SIRMþ1000mT).
visual and statistical examination of the fit
between the real and modeled data.
The IRM acquisition curves were obtained for
the unmixing study by exposing each sample to a
stepwise-increasing uniaxial field imparted by a
Princeton Applied Research Vibrating Sample
Magnetometer. The IRM data were acquired in
1 mT steps from 1 to 10 mT, in 10 mT steps from
10 to 100 mT, and in 100 mT steps from 100 to
1000 mT. The resulting remanence was measured
after each step using a 2-G Superconducting
Rock Magnetometer.
The IRM acquisition data are presented in
Figure 10a. The noise characteristics of the raw
Copyright # 2006 John Wiley & Sons, Ltd.
IRM acquisition data were reduced by applying a
smoothing spline. Spline smoothing was desirable because the unmixing analysis utilizes the
first derivatives (gradients) of the acquisition
curves, which are extremely susceptible to noisy
data. Spline smoothing was accomplished using
the fit spline function in JMP# statistical analysis
software. Values of the spline smoothing parameter () are given in Figure 10a, as are the sum
of squares error (SSE). The SSE provides a measure of the ‘goodness’ of fit by summing the
squared distances (residuals) from each raw
data point to the fitted spline. Examination of
Figure 10a reveals that Sample A (unheated)
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
Figure 9. Thermal demagnetization of zero-field-cooled SIRM. Suppression of the Verwey transition is apparent in the heated
continued to acquire remanence throughout the
IRM acquisition procedure (implying that the
sample is non-saturated), whereas Sample A
(unheated) reached peak IRM values prior to
the maximum applied field (implying the sample
is saturated).
The first step of the unmixing analysis used
an utomated procedure based on the expectation-maximization
Version 2.2). The automated algorithm simply
requires the user to define whether the samples
have reached saturation during IRM acquisition,
and the number of individual mineral phases to
be modelled in the final solution (Heslop et al.,
2002). The statistically optimum number of
model components was estimated by this algorithm to be five. The number of model components was then systemically reduced in an effort
to determine the minimum number of model
components that still produced a realistic and
satisfactory fit. This reduction in the number of
model components was undertaken because
although, ‘the fit of a finite mixture model will
always improve as the number of components is
increased, it is generally best to favour simplicity
over complexity’ (Heslop et al., 2002). This procedure determined that a satisfactory fit was
Copyright # 2006 John Wiley & Sons, Ltd.
obtained using three model components, and
the initial unmixing parameters of these model
components (see below) was estimated.
The second step of the analysis consisted of
making small incremental improvements to the
fit between real and modelled data using the
interactive computer program (IRM CLG). This
procedure provided SSE ‘goodness’ of fit estimates after each incremental change in unmixing
parameters (Kruiver et al., 2001). In this manner
the SSE between real and modelled data was
The results obtained from steps 1 and 2 of
the study are expressed using three unmixing
B1/2: the applied field at which the mineral
acquires half of its saturation IRM (SIRM).
This parameter provides a measure of the
mean coercivity of that population.
(ii) DP: a dispersion parameter expressing the
distribution of each mineral phase as one
standard deviation of the log-normal
(iii) The relative contribution from each model
component expressed as a percentage.
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure10. Results ofthe three component unmixing study. (a) Raw IRMacquisition data with splined smoothed acquisition curves.
(b) IRM gradient plot of the spline smoothed acquisition curve from the experimentally heated sample. (c) IRM gradient plot of the
spline smoothed acquisition curve from the natural undisturbed sample.
Copyright # 2006 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
Table1. Results of the three model component unmixing study
Sample A (unheated)
Sample A (unheated)
Sample A (unheated)
Sample A (heated)
Sample A (heated)
Sample A (heated)
Log (B1/2)
The results of the unmixing study are presented
in Table 1. The fit between the three model
components and the spline smoothed IRM acquisition curve gradients are presented graphically
in Figures 10b and 10c. Examination of Table 1
shows that the contribution due to high coercivity minerals is significantly greater in Sample
A (heated). The B1/2 values of the high coercivity
components range from 308 to 333 mT in the
unheated and heated samples, whereas the lower
coercivity components were estimated to range
from 5 to 32 mT.
The parameter B1/2 is also often referred to as
B0 cr, the remanent acquisition coercive force.
Dankers (1978) determined that the B0 cr values
for natural occurring magnetites, titanomagnetites and maghemites ranged from 8.5 to 67.5 mT.
Naturally occurring haematites ranged from
55 to 450 mT, although values more commonly
ranged from 100 to 400 mT. A comparison with
Dankers (1978) data shows that the high coercivity component of Sample A (heated) falls within
the normal range for haematite, providing evidence that this mineral is an end product of the
high temperature phase transformations. The
high coercivity component increases from
approximately 11% in Sample A (unheated) to
31% in Sample A (heated). Further evidence
supporting inversion to haematite is provided
by a soil colour change from brown (10YR 4.5/3)
in Sample A (unheated) to reddish yellow (7.5YR
6/8) in Sample A (heated).
The low temperature remanence and unmixing study provided evidence that significant
mineralogical transformations had occurred
after exposing Sample A to a temperature of
650 C for extended periods of time. In the next
section we have attempted to determine whether
this temperature might realistically be expected
in the soil beneath an archaeological hearth.
Copyright # 2006 John Wiley & Sons, Ltd.
B1/2 (mT)
Contribution (%)
Soil temperature experiment
A field study was conducted to determine what
soil temperatures might be expected in the soil
beneath a fire hearth. The experiment was held
during the 2003 University of Oklahoma field
programme at the Buried City, a Plains Village
Period archaeological site located in the Texas
Panhandle region.
A type K thermocouple was buried 2 cm
beneath an experimental hearth (Figure 11). A
fire was started and kept active over a period of
11.5 days according to a loosely defined schedule. The fire was stoked by field school students
or volunteers in the morning before the days
excavations began. The fire was stoked again at
midday during lunch break, and was tended
continuously after work until approximately
2200 hours to midnight. The fire had generally
extinguished by morning, but still contained
warm coals and ashes. Ashes were cleaned
from the hearth approximately every 24 h. Mesquite (Prosopis pubescens), locust (Gleditsia triacanthos) and cottonwood (Populus deltoides var.
Figure 11. Measurement of soil temperatures beneath an experimental hearth.
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure12. Results of the soil temperature experiment.Temperature versus time (top) and a histogram of temperature data (bottom).
occidentalis) were the primary fuels. Temperature
data from the thermocouple were recorded every
4 min on an Extech Instruments EA15 digital
data logger for approximately 280 h, at which
point the data logger’s batteries failed and the
experiment was concluded.
Results of the soil temperature experiment are
presented in Figure 12. A total of 4204 temperature measurements were recorded. The average
temperature over the duration of the experiment
was 447 C, the maximum temperature achieved
was 649.5 C. One interesting aspect of the experimental data is that temperatures generally
increased over the course of the first 48 h then
remained relatively constant, although significant local variations were observed. The slow
ramp up to relatively stable temperatures may be
due to the gradual loss of soil moisture during
the first 48 h. The average temperature from 48 h
to the end of the experiment was 479 C (that is,
Copyright # 2006 John Wiley & Sons, Ltd.
temperature data excluding the first 48 h). The
temperature also increased significantly after
each ash cleaning, increases from 50 C to 150 C
were common. The soil temperature exceeded
500 C for a total of 71 h during the experiment
(25% of the total time). At the conclusion of the
experiment the dark brown silty loam soil had
changed colour to a reddish brown, indicating
some inversion to haematite had occurred.
Two years after the soil temperature experiment was conducted the hearth was relocated,
exposed by shovel scrapping, and photographed
(Figure 13). The soil colour changes resulting
from the heating experiment were recorded and
representative samples were taken from each
colour (Figure 13). The magnetic susceptibility
and frequency dependence of susceptibility of
these five samples are presented in Table 2. Soil
colour changes and magnetic susceptibility
enhancements within the hearth were rather
Archaeol. Prospect. 13, 207–227 (2006)
D. Maki, J. A. Homburg and S. D. Brosowske
Summary and discussion
Figure 13. Planview photograph of the experimental hearth
after shovel scraping in August 2005. The locations of the
five samples collected for magnetic susceptibility analysis are
variable. Soil at the edge of the oxidized
zone was minimally enhanced (1.1 times the
natural soil), whereas the maximum enhancement occurred in the olive yellow oxidized soil
(6.4 times the natural soil).
Canti and Linford (2000) report that soil
temperatures seldomly exceed 500 C beneath
experimental fire hearths, with soil reddening
occurring rarely. The somewhat higher temperatures and definite soil reddening observed during this experiment may be due to the regular
removal of ash, a rather low initial soil moisture
content (the experiment took place in semi-arid
northwest Texas), and the use of a protective
metal ring around the fire (Figure 11) that may
have acted as a windbreak and a heat reflector.
The relatively long duration of the experiment
and natural soils with a lower inversion temperature also may be factors that contributed to
the observed soil reddening.
Prolonged exposure of soil samples from the
West Bluff Project Area to a temperature of
650 C resulted in a decrease in the magnetic
susceptibility, falling to below the initial susceptibility value after 17 h. A reduction in S-ratio
values with time spent at this isotherm suggests
that the proportion of ferrimagnetic to antiferromagnetic minerals decreased with prolonged
exposure to high temperatures. The reduction
in these two bulk magnetic properties is thought
to be the result of high temperature mineralogical transformations.
Suppression of the Verwey transition and
results from the unmixing study show that these
transformations caused a decrease in stoichiometric magnetite and an increase in the high
coercivity component of approximately 30%.
This increase is thought to have proceeded via
oxidation of the multidomain primary (lithogenic) magnetite to maghemite, followed by partial inversion to haematite, although further tests
would be required to confirm that the high
coercivity component is not composed of
goethite or a goethite/haematite mixture. These
tests might take the form of IRM acquisition
curves using applied fields approaching 4 T,
which unfortunately is beyond the range of
applied fields that are possible with the equipment used during this study (France and Oldfield, 2000).
Additional evidence of inversion to haematite
was provided by a distinct colour change from
brown to reddish yellow. Inversion may result in
a ‘core’ of magnetite surrounded by maghemite and/or haematite. After 26 h at 650 C the
frequency dependence of susceptibility had
Table 2. Magnetic properties of soil samples from the experimental hearth
exp 2
exp 3
exp 4
exp 5
Natural soil: greyish brown 2.5YR 5/2
Oxidized soil: olive yellow 2.5YR 6/6
Charcoal and wood ash from near centre
of hearth
Oxidized soil: strong brown 7.5YR 5/6
Edge of oxidized soil: very dark greyish
brown 2.5YR 3/2
Copyright # 2006 John Wiley & Sons, Ltd.
Mass normalized
Frequency dependence
magnetic susceptibility (m3 kg1) of susceptibility (%)
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
increased from 1.48% to 4.32%, suggesting a
small portion of the ferrimagnetic grain-size
assemblage may have been reduced to SPM
proportions. The soil within the experimental hearth that experienced the largest level
of enhancement also experienced a relatively
large increase in frequency dependence of susceptibility (see Table 2). Again, this increase
in SPM minerals may be the result of a shrinking
‘core’ of ferrimagnetic minerals surrounded by
a growing ‘shell’ of haematite.
The results of the laboratory testing suggest
that the failure to detect hearth features by magnetic field gradient survey methods at the West
Bluff Project Area may have been primarily
associated with two factors:
a rather small initial enhancement of soil
susceptibility associated with exposure to
high temperatures (increase in < 20%);
(ii) a drop in susceptibility to below initial
values associated with prolonged exposure
to high temperatures.
Other factors affecting the survey results include
randomization of the TRM component by bioturbation, and signal clutter created by bioturbation,
agricultural plow furrows, disturbance from previous excavations, randomly oriented igneous
and metamorphic rock at or near the ground
surface, and modern ferrous iron debris.
A post-excavation comparison of the location
of archaeological features with the magnetic survey results was conducted in an effort to test
whether inversion did result in negative magnetic field gradient anomalies. This comparison
identified several hearths where a correlation
between a negative magnetic anomaly and the
underlying feature was found; two of these correlations are depicted in Figure 14. It should be
noted that the negative field gradient anomalies
identified in Figure 14 are relatively subtle,
with intensities ranging from 1.9 nT m1
to 8.4 nT m1, whereas overall the magnetic
field gradient data possessed a standard deviation ranging from 5 to 7 nT m1. These signalto-noise (and clutter) characteristics would have
Figure 14. Two examples where negative magnetic field gradient anomalies exist at the location of excavated hearths.
Images were created from processed magnetic data. Processing included the application of a zero mean traverse algorithm,
data reduction in the north^south direction and expansion in the east^west direction by linear interpolation (final display data
density is 0.25 m 0.25 m). A low pass filter (one data point radius) was used to reduce the noise characteristics of the data.
Copyright # 2006 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 207–227 (2006)
made recognition of these anomalies rather difficult without prior knowledge of the feature
locations. The inconsistent and weak nature of
the negative magnetic anomalies suggests that
the reduction in soil magnetic susceptibility may
be partially (or wholly) offset by the contribution
of high susceptibility material from wood ash.
Additionally, several hearths were identified that
correlate with a positive magnetic anomaly (possibly due to wood ash contributions), and many
hearths created no discernable magnetic anomaly.
The soil temperature experiment suggests that
650 C is at the high end of temperatures that
might be expected in an archaeological hearth.
As noted earlier, the average soil temperature
2 cm beneath the experimental hearth ranged
from about 450 C to 480 C. Soil colour changes
and susceptibility enhancement indicate that the
temperatures and atmospheric conditions under
the experimental hearth were sufficient to activate some mineralogical transformations, probably including high temperature inversion to
Implications to archaeological
The primary implication of this study is that
prolonged heating of soil within a hearth may,
in some cases, result in a reduction in susceptibility rather than an increase or enhancement.
A localized decrease in susceptibility can result
in a negative induced magnetic field gradient
anomaly, the opposite of what is normally
expected. The probability of a negative magnetic
anomaly is greatly increased when bioturbation
has randomized the thermoremanent component
of the feature.
A potential application in archaeology
As haematite is very stable, it is possible that
fired archaeological features may retain a record
of their inversion history for extended periods of
time. Inversion history may eventually find practical applications in archaeology. For example,
the relative concentration of haematite could be
Copyright # 2006 John Wiley & Sons, Ltd.
D. Maki, J. A. Homburg and S. D. Brosowske
used to estimate the ‘use life’ of archaeological
hearths by the method outlined below.
An expression for the rate of inversion at
elevated temperatures is provided in the form
of a first-order differential equation by Adnan
and O’Reilly (1999). Predictions concerning the
variation in magnetic susceptibility levels with
time can be made using a solution to this
rate expression. This solution takes the form,
(t) ¼ maximum exp (t/), where (t) is the
magnetic susceptibility at time t and maximum is
defined as the maximum magnetic susceptibility
that occurs as a result of high temperature susceptibility enhancement (e.g. the maximum susceptibility values from each site in Figure 7). In
this expression, the time at which the maximum
magnetic susceptibility is reached is defined as
t ¼ 0, and is a rate constant related to the
Arrhenius equation for a thermally activated
During the high temperature experiment, soils
from the West Bluff Project Area, 48CA3030
(Wyoming) and 31CD218 (North Carolina) all
experienced an initial enhancement in magnetic
susceptibility, followed by a decrease in susceptibility, apparently due to inversion. The time
constant (for heating at 650 C) was determined
for soils from these three sites by fitting the
inversion function (above) to the experimental
heating data from Figure 7. The time constant ()
was adjusted until a ‘best fit’ was arrived at, as
determined by sum of squares error (SSE) analysis (Figure 15). Examination of Figure 15 will
show a large variation in the rate constant at
this isotherm. It was not possible to determine
the time constant of soils from 11S1131 (Illinois)
or 34BV100 (Oklahoma) because the maximum
magnetic susceptibility was not achieved by the
conclusion of the high temperature experiment.
Experimental determination of the inversion
time constant () may find a practical application
in archaeology by making it possible to estimate
the time a given soil sample has spent at high
temperatures. By way of example, several heavily oxidized soil samples were collected from the
wall of a small diameter pit hearth/oven at
48CA3030—one of the comparative sites examined during this study (Jones and Munson, 2005;
Maki, 2005). Oxidized samples collected from the
interior wall of the cylinder were red (2.5YR 4/6),
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Figure15. Estimation of the rate constant for a thermally activated process.
whereas the natural soils from which the hearth/
oven was constructed were brown (10YR 5/3).
The magnetic susceptibility values of oxidized
soil samples were twice those of the natural soils,
but only about 2% of the maximum magnetic
susceptibility value recorded during the high
temperature heating experiment described in
this report (see Figure 7). The relatively low
magnetic susceptibility of the oxidized soil samples (compared with the maximum susceptibility
value) and the distinct reddening of soil colour
suggests that high temperature inversion had
If we assume that inversion to haematite was
responsible for reducing soil susceptibility to 2%
of its maximum experimental value, a prediction
can be made using the experimentally determined time constant () and the inversion function introduced above. The inversion function
(using ¼ 103) provides an estimate of about
400 h at 650 C to reduce the magnetic susceptibility to 2% of its maximum experimental value.
In other words, the hearth’s ‘use life’ is estimated
to be 400 h at this temperature.
Of course such mathematical estimates must
be used with extreme caution, as there are
numerous sources of uncertainty, not the least
of which is the estimated temperature of the soil
Copyright # 2006 John Wiley & Sons, Ltd.
surrounding the hearth. The heating experiment
described in this report, and previous experiments described by Canti and Linford (2000),
reveal significant variation in soil temperatures
achieved beneath experimental hearths. Given
the uncertainty associated with soil temperatures, a more rigorous and useful approach to
modelling the inversion process is recommended. This approach should proceed by utilizing the method developed by Adnan and
O’Reilly (1999) for determining the two components of , the activation energy and a frequency
factor. Determination of these two components
allows one to use the inversion function to model
changes in magnetic susceptibility at any isotherm. Unfortunately the equipment necessary
to determine by this method were not available
during this research.
The experiments described in this article have
documented a reduction in the concentration of
ferrimagnetic minerals—and an increase in the
concentration of high coercivity minerals—with
prolonged exposure to high temperatures. These
mineralogical changes appear to be due to high
Archaeol. Prospect. 13, 207–227 (2006)
temperature inversion to haematite. In the case of
soils from the West Bluff Project Area, the final
experimental susceptibility values were less than
initial values. This implies that under some soil
conditions high temperature inversion can result
in a local decrease in susceptibility, which would
create a negative induced magnetic field gradient anomaly (the opposite of what is normally
The temperature and environmental conditions necessary for inversion to proceed appear
to be highly variable. At present the proportion
of archaeological sites at which inversion may be
a factor to consider when interpreting magnetic
survey results is unknown. For now care should
be taken when interpreting magnetic data from
sites with soils similar to those found in the
West Bluff Project Area; that is, sites with relatively high initial concentrations of multidomain
primary lithogenic magnetite (magnetite is widespread in parent rocks of the Los Angeles
Basin, including the granitic rocks that dominate most of the surrounding mountains). In a
more general sense, high temperature inversion may be a factor to consider in soils formed
in sandy deposits associated with dunes,
alluvium and marine sediments that are now
The authors would like to thank Rinita Dalan for
providing a thorough and insightful review of an
early draft of this manuscript. The manuscript
also greatly benefited from the comments of
Chris Gaffney and two anonymous reviewers at
Archaeological Prospection. We would also like
to thank Harold and Kirk Courson (of Courson
Oil and Gas) Perryton, Texas for continuing
support of this research, and the personnel of
the Institute for Rock Magnetism, University of
Minnesota for valuable guidance and advice
throughout the project. The Institute for Rock
Magnetism is supported by the USNSF and the
W.M. Keck Foundation (Los Angeles). The unmixing algorithms used during this research
were obtained from the paleomagnetic laboratory ‘Fort Hoofddijk’ via their home page at
( forth/).
Copyright # 2006 John Wiley & Sons, Ltd.
D. Maki, J. A. Homburg and S. D. Brosowske
Adnan J, O’Reilly W. 1999. The transformation of
Fe2O4 to Fe2O4: thermal activation and the
effect of elevated pressure. Physics of the Earth
and Planetary Interiors 110: 43–50.
Altschul JH. 1997. A Cultural Resources Assessment of
the West Bluff Project, Westchester/Playa del Rey,
California. Technical Report 97-8, Statistical Research, Tucson, Arizona.
Altschul JH. 1999. National Register Evaluation of CALAN-63, CA-LAN-64, and CA-LAN-206, West Bluff
Project. Technical Report 99-45, Statistical Research, Tucson, Arizona.
Altschul JH, Stoll AQ, Grenda DR, Ciolek-Torrello
R. 2000. Historic Properties Treatment Plan for
the Bluff Site, LAN-64, West Bluff Project,
Westchester/Playa del Rey, California. Technical
Report 00-32, Statistical Research, Tucson, Arizona.
Bando Y, Kiyama M, Takada T, Kachi S. 1965. The
effect of particle size of Fe2O4 on the transformation from -phase to -form. Japanese Journal of
Applied Physics 4: 240–241.
Brown AP, O’Reilly W. 1999. The magnetism and
microstructure of pulverized titanomagnitite,
Fe2.4Ti0.6O4: the effect of annealing, maghemitization and inversion. Physics of the Earth and Planetary Interiors 116: 19–30.
Canti MG, Linford N. 2000. The effects of fire on
archaeological soils and sediments: temperature
and colour relationships. Proceedings of the Prehistoric Society 66: 385–395.
Clark AJ. 1996. Seeing Beneath the Soil. Prospecting
Methods in Archaeology. B.T. Batsford Ltd.:
Crowther J. 2003. Potential magnetic susceptibility
and fractional conversion studies of archaeological soils and sediments. Archaeometry 45: 685–
Dalan RA, Banerjee SK. 1998. Solving archaeological problems using techniques of soil magnetism.
Geoarchaeology 13: 3–36.
Dankers PHM. 1978. Magnetic properties of dispersed natural iron-oxides of known grain-size.
Doctoral Thesis, Rijkuniversiteit Te Utrecht,
Day R, Fuller MD, Schmidt VA. 1977. Hysteresis
properties of titanomagnetites: grain size and
composition dependence. Physics of the Earth and
Planetary Interiors 13: 260–266.
Douglass JG, Altschul JH. 2003. Preliminary Report on
Archaeological Monitoring and Data Recovery at Sites
CA-LAN-63, CA-LAN-64, and CA-LAN-206A, West
Bluff Project, Westchester/Playa del Rey, California
(draft report). SRI Technical Report 03–77, Statistical Research, Redlands, California, and Tucson,
Archaeol. Prospect. 13, 207–227 (2006)
Thermally Activated Mineralogical Transformations in Hearths
Dunlop DJ, Ozdemir O. 1997. Rock Magnetism: Fundamentals and Frontiers. Cambridge University
Press: Cambridge.
Evans ME, Heller F. 2003. Environmental Magnetism:
Principles and Applications of Environmagnetics.
Academic Press: San Diego, CA.
France DE, Oldfield F. 2000. Identifying geothite
and hematite from rock magnetic measurements
of soils and sediments. Journal of Geophysical
Research 105: 2781–2795.
Grenda DR, Homburg JA, Altschul JH. 1994. The
Centinela Site (CA-LAN-60): Data Recovery at a
Middle Period Creek-edge Site in the Ballona Wetlands,
Los Angeles County, California. Technical Series
No. 45, Statistical Research, Tucson, Arizona.
Heslop D, Dekkers MJ, Kruiver PP, van Oorschot
IHM. 2002. Analysis of isothermal remanent
magnetization acquisition curves using the expectation-maximization algorithm. Geophysical
Journal International 148: 58–64.
Hunt PC. 1991. Handbook from the Environmental
Magnetism Workshop. University of Minnesota,
Minneapolis, Minnesota, June 5–8.
Jones G, Munson G. 2005. Geophysical survey as an
approach to the ephemeral campsite problem:
case studies from the Northern Plains. Plains
Anthropologist 193: 31–44.
Kruiver PP, Dekkers MJ, Heslop D. 2001. Quantification of magnetic coercivity components by the
analysis of acquisition curves of isothermal remanent magnetisation. Earth and Planetary Science
Letters 189: 269–276.
Linford NT, Canti G. 2001. Geophysical evidence
for fires in antiquity: preliminary results from an
experimental study. Archaeological Prospection 8:
Maher BA. 1986. Characterization of soils by
mineral magnetic measurements. Physics of the
Earth and Planetary Interiors 42: 76–92.
Maki D. 2005. Lightning strikes and prehistoric
ovens: determining the source of magnetic
anomalies using techniques of environmental
magnetism. Geoarchaeology 20: 449–459.
Mbila MO, Homburg JA. 2000. Mineralogical and
micromorphological characterization of selected
archaeological deposits in the Ballona Wetlands
of Southern, California. Paper presented at the
Copyright # 2006 John Wiley & Sons, Ltd.
63rd Annual Meeting of the Soil Science Society of
America, November 5–9, Minneapolis, Minnesota.
McClean RG, Kean WF. 1993. Contributions of
wood ash magnetism to archaeological properties
of fire pits and hearths. Earth and Planetary Science
Letters 119: 387–394.
O’Reilly WO. 1983. The identification of titanomaghemites: model mechanisms for the maghemitization and inversion processes and their
magnetic consequences. Physics of the Earth and
Planetary Interiors 31: 65–76.
Ozdemir O. 1987. Inversion of titanomaghemites.
Physics of the Earth and Planetary Interiors 46: 184–
Ozdemir O. 1990. High-temperature hysteresis and
thermoremanence of single-domain maghemite.
Physics of the Earth and Planetary Interiors 65: 125–
Ozdemir O, Banerjee SK. 1984. High temperature
stability of maghemite (Fe2O4). Geophysical Research Letters 11: 161–164.
Ozdemir O, Dunlop DJ. 1993. Chemical remanent
magnetization during Fe2OOH phase transformations. Journal of Geophysical Research 98(B3):
Peters C, Church MJ, Mitchell C. 2001. Investigation
of fire ash residues using mineral magnetism.
Archaeological Prospection 8: 227–237.
Tite MS, Linington RE. 1975. Effect of climate on the
magnetic susceptibility of soils. Nature 256: 565–566.
Thompson R, Oldfield F. 1986. Environmental Magnetism. Allen and Unwin: London.
Van Horn DM. 1986. Surface Mapping and Auger
Sampling at LAN-63 and LAN-64, City of Los Angeles. Archaeological Associates: Sun City,
Van Horn DM. 1987. Excavations at the Del Rey Site
(LAN-63) and the Bluff Site (LAN-64) in the City of
Los Angeles. Archaeological Associates: Sun City,
Van Horn DM, White LS. 1997. LAN-206 and the
West Bluff Property. Archaeological Associates:
Sun City, California.
Weston DG. 2002. Soil and susceptibility: aspects of
thermally induced magnetism within the dynamic pedological system. Archaeological Prospection 9: 207–215.
Archaeol. Prospect. 13, 207–227 (2006)
Без категории
Размер файла
455 Кб
forma, transformation, цfe2o4, hearthsinversion, maghemites, archaeological, haematite, fe2o4, mineralogical, phase, thermally, activated
Пожаловаться на содержимое документа