by Ikeya (1993), Grün (2007), Rink (1997), and Jonas
(1997). Most recently, Schellmann et al. (2008) have
reviewed the accuracy of ESR dating of different materials
like coral, mollusc shells, and quartz; Blackwell (2006)
has focused on ESR dating in Karst environments; and
Grün (2006) has given a detailed description of ESR dating of human teeth.
As already stated, the ESR dating method uses the phenomenon that common minerals, like the aragonite fabric
of corals, act as natural dosimeters. An ESR age is derived
from the ratio of the accumulated radiation dose D E
(equivalent dose) to the natural radiation rate D’ (dose
rate), to which the sample was exposed in the past
(Figure 1). D E is determined by an additive dose method.
For this method, 20 or more aliquots of the same sample
are stepwise irradiated with increasing artificial g-doses
(e.g., using a
60
Co gamma source). The growing amplitude of the ESR signal at g = 2.0007 of a coral sample is
proportional to the number of trapped charges (D E ), which
is dependent on the strength of natural irradiation (dose
rate, D’) and on the radiation time (= age of the sample).
A living coral does not contain the ESR dating signal at
g = 2.0007, which means that the ESR intensity is zero.
D E is determined by extrapolating the resulting dose
response curve to zero ESR intensity. In the case of dating
corals, the D E –D max plot procedure (D–DP procedure)
should be used for the calculation of D E as described by
Schellmann and Radtke (2001). The former natural dose
rate (D’) of a coral sample is estimated on the basis of
the content of uranium (other radioactive elements like
Th and K are of no relevance) and the intensity of the cosmic dose rate. The correct determination of D’ and D E is
essential for any ESR age estimation.
Improvements in annual dose rate (D’) estimation and
the newly developed approach of equivalent dose (D E )
determination (D E –D max plot procedure) increase the precision of ESR dating of Holocene and Pleistocene corals
(Schellmann et al., 2008). This is strongly supported by
comparing ESR dating results with other numeric dating
methods such as radiocarbon (
14 C) and TIMS U-series
analysis (TIMS
230
Th/
234
U).
Case study 1: ESR dating of Pleistocene coral reef
terraces
Since several early systematic studies have been carried
out (e.g., Ikeya and Omuhra, 1983; Radtke and Grün,
1988), dating of corals is one of the most common applications of ESR dating. It has been widely applied for dating
sequences of raised coral reef terraces with ages of up to
500–600 ka, e.g., on Sumba island (Pirazzoli et al., 1991);
on Huon Peninsula, New Guinea (Grün et al., 1992); on
Curaçao, Netherlands Antilles (Radtke et al., 2003;
Schellmann et al., 2004b); and on Barbados (e.g., Radtke
et al., 1988; Schellmann and Radtke, 2004a, b).
The ESR dating of the fossil coral reef terraces on Barbados provides an impressive example of the application
of this method (e.g., Schellmann and Radtke, 2004a, b).
On this Caribbean island, several uplifted Pleistocene
coral reef terraces are well preserved, rising up to 122 m
above modern sea level in southern Barbados (Figure 2).
By using morpho- and chronostratigraphic methods, the
area between the present coastline and the center of southern Barbados can be subdivided up to 13 clearly distinguishable main coral reef terraces (T1 to T13) and some
less prominent sub-levels (for example T-6a and T-6b,
T-1a 1 , and T-1a 2 ). The youngest formations are located
at the lowest elevations close to the present coast, while
older reef terraces have higher elevations and generally
a greater distance from the coast. Due to the relatively
slow uplift rate in southern Barbados (approx.
0.27 Æ 0.02 m/1,000 a), the youngest Pleistocene coral
reefs (T-1a 1 ) have been formed at approx. 74 ka (ESR)
or 77 ka (TIMS Th/U) and the oldest and highest coral reef
(T-13) during a sea-level highstand approx. 410 ka (MIS
11). This correlation between age and height of the terraces allows us to calculate neotectonic movements and
to estimate paleo sea-level changes (e.g., Schellmann
et al., 2004a; Radtke and Schellmann, 2005).
Generally, the timing of coral reef growth calculated by
both dating methods agrees well, especially when the error
ranges are considered. However, a detailed comparison of
dating results of both methods shows that many ESR ages
are up to 5–10% younger than the corresponding U-series
ages, perhaps the result of a slight overestimation of the
alpha-efficiency (0.05) used for ESR age calculations or
of weak recrystallization processes that influence the
ESR more strongly than the U-series data. The comparison also shows that the apparently minor analytical errors
(precision) of U-series data (less than 1%, ESR between
5–8%) give no evidence of a better time resolution
(Schellmann et al., 2004a). The latter can be demonstrated
well by the relatively high scattering of both ESR and
TIMS U-series data inside an individual Late Pleistocene
coral reef deposit. For example, comparing the results of
ESR and TIMS U-series dating of MIS 5a-1 corals, carried
out on paired samples from Inch Marlowe Point location
on Barbados (Figure 3), shows that the spread of individual ages (not considering errors) for both methods is about
4,000 years, although all ages were produced on individuals from the same branch of coral which are hence most
likely of the same age (grown within not more than
a few hundred years). Therefore, age differentiations
between single Late Pleistocene coral reef deposits that
have developed in time periods of less than 4,000–5,000
years are rarely reliable, independent of the applied dating
technique (ESR or TIMS U-series).
Case study 2: ESR dating of Holocene corals
The accuracy of ESR dating on paired samples of Holocene corals has been tested by Radtke et al. (2003) using
21 radiocarbon-dated Late and Middle Holocene coral
samples from Tsunami deposits found on Bonaire and
Curaçao (Netherlands Antilles) (Figure 4). The calibrated
radiocarbon ages spread between recent and 3,644 years
370
ELECTRON SPIN RESONANCE DATING (ESR)
(1997). Most recently, Schellmann et al. (2008) have
reviewed the accuracy of ESR dating of different materials
like coral, mollusc shells, and quartz; Blackwell (2006)
has focused on ESR dating in Karst environments; and
Grün (2006) has given a detailed description of ESR dating of human teeth.
As already stated, the ESR dating method uses the phenomenon that common minerals, like the aragonite fabric
of corals, act as natural dosimeters. An ESR age is derived
from the ratio of the accumulated radiation dose D E
(equivalent dose) to the natural radiation rate D’ (dose
rate), to which the sample was exposed in the past
(Figure 1). D E is determined by an additive dose method.
For this method, 20 or more aliquots of the same sample
are stepwise irradiated with increasing artificial g-doses
(e.g., using a
60
Co gamma source). The growing amplitude of the ESR signal at g = 2.0007 of a coral sample is
proportional to the number of trapped charges (D E ), which
is dependent on the strength of natural irradiation (dose
rate, D’) and on the radiation time (= age of the sample).
A living coral does not contain the ESR dating signal at
g = 2.0007, which means that the ESR intensity is zero.
D E is determined by extrapolating the resulting dose
response curve to zero ESR intensity. In the case of dating
corals, the D E –D max plot procedure (D–DP procedure)
should be used for the calculation of D E as described by
Schellmann and Radtke (2001). The former natural dose
rate (D’) of a coral sample is estimated on the basis of
the content of uranium (other radioactive elements like
Th and K are of no relevance) and the intensity of the cosmic dose rate. The correct determination of D’ and D E is
essential for any ESR age estimation.
Improvements in annual dose rate (D’) estimation and
the newly developed approach of equivalent dose (D E )
determination (D E –D max plot procedure) increase the precision of ESR dating of Holocene and Pleistocene corals
(Schellmann et al., 2008). This is strongly supported by
comparing ESR dating results with other numeric dating
methods such as radiocarbon (
14 C) and TIMS U-series
analysis (TIMS
230
Th/
234
U).
Case study 1: ESR dating of Pleistocene coral reef
terraces
Since several early systematic studies have been carried
out (e.g., Ikeya and Omuhra, 1983; Radtke and Grün,
1988), dating of corals is one of the most common applications of ESR dating. It has been widely applied for dating
sequences of raised coral reef terraces with ages of up to
500–600 ka, e.g., on Sumba island (Pirazzoli et al., 1991);
on Huon Peninsula, New Guinea (Grün et al., 1992); on
Curaçao, Netherlands Antilles (Radtke et al., 2003;
Schellmann et al., 2004b); and on Barbados (e.g., Radtke
et al., 1988; Schellmann and Radtke, 2004a, b).
The ESR dating of the fossil coral reef terraces on Barbados provides an impressive example of the application
of this method (e.g., Schellmann and Radtke, 2004a, b).
On this Caribbean island, several uplifted Pleistocene
coral reef terraces are well preserved, rising up to 122 m
above modern sea level in southern Barbados (Figure 2).
By using morpho- and chronostratigraphic methods, the
area between the present coastline and the center of southern Barbados can be subdivided up to 13 clearly distinguishable main coral reef terraces (T1 to T13) and some
less prominent sub-levels (for example T-6a and T-6b,
T-1a 1 , and T-1a 2 ). The youngest formations are located
at the lowest elevations close to the present coast, while
older reef terraces have higher elevations and generally
a greater distance from the coast. Due to the relatively
slow uplift rate in southern Barbados (approx.
0.27 Æ 0.02 m/1,000 a), the youngest Pleistocene coral
reefs (T-1a 1 ) have been formed at approx. 74 ka (ESR)
or 77 ka (TIMS Th/U) and the oldest and highest coral reef
(T-13) during a sea-level highstand approx. 410 ka (MIS
11). This correlation between age and height of the terraces allows us to calculate neotectonic movements and
to estimate paleo sea-level changes (e.g., Schellmann
et al., 2004a; Radtke and Schellmann, 2005).
Generally, the timing of coral reef growth calculated by
both dating methods agrees well, especially when the error
ranges are considered. However, a detailed comparison of
dating results of both methods shows that many ESR ages
are up to 5–10% younger than the corresponding U-series
ages, perhaps the result of a slight overestimation of the
alpha-efficiency (0.05) used for ESR age calculations or
of weak recrystallization processes that influence the
ESR more strongly than the U-series data. The comparison also shows that the apparently minor analytical errors
(precision) of U-series data (less than 1%, ESR between
5–8%) give no evidence of a better time resolution
(Schellmann et al., 2004a). The latter can be demonstrated
well by the relatively high scattering of both ESR and
TIMS U-series data inside an individual Late Pleistocene
coral reef deposit. For example, comparing the results of
ESR and TIMS U-series dating of MIS 5a-1 corals, carried
out on paired samples from Inch Marlowe Point location
on Barbados (Figure 3), shows that the spread of individual ages (not considering errors) for both methods is about
4,000 years, although all ages were produced on individuals from the same branch of coral which are hence most
likely of the same age (grown within not more than
a few hundred years). Therefore, age differentiations
between single Late Pleistocene coral reef deposits that
have developed in time periods of less than 4,000–5,000
years are rarely reliable, independent of the applied dating
technique (ESR or TIMS U-series).
Case study 2: ESR dating of Holocene corals
The accuracy of ESR dating on paired samples of Holocene corals has been tested by Radtke et al. (2003) using
21 radiocarbon-dated Late and Middle Holocene coral
samples from Tsunami deposits found on Bonaire and
Curaçao (Netherlands Antilles) (Figure 4). The calibrated
radiocarbon ages spread between recent and 3,644 years
370
ELECTRON SPIN RESONANCE DATING (ESR)