Nuclear Instruments and Methods in Physics Research B 172 (2000) 592±596

www.elsevier.nl/locate/nimb

Radiocarbon measurements of stromatolite heads and crusts at the

Salgada Lagoon, Rio de Janeiro State, Brazil
a,*
Melayne M. Coimbra , Cleverson G. Silva b, C
atia F. Barbosa b,c
, Ken A. Mueller a

a
PRIME Lab, 1396 Physics Building, West Lafayette, IN 47906-1396, USA
b
Departamento de Geologia, Universidade Federal Fluminense, Niteroi, RJ, Brazil
c
FAPERJ ± Fundacßa~o de Amparo a Pesquisa do Estado do Rio de Janeiro, Matricula:1997.1643.5, Brazil

Abstract

In this work, we prepared and measured some stromatolite carbonate samples, from Salgada Lagoon, Rio de
Janeiro, Brazil. Stromatolites are bio-sedimentary, laminated, carbonate structures produced by sedimentary, chemical
and biological processes related to the development and growth of microbial benthic communities, mainly dominated
by blue algae and cyanobacteria. These structures are present in the geological record in rocks older than 3.0 billion
years and have been used to study the origin of primitive life and variations in past environmental conditions. Detailed
AMS measurements were performed at PRIME Lab (Purdue Rare Isotope Measurement Laboratory, Purdue
University, IN, USA). Ó 2000 Elsevier Science B.V. All rights reserved.

1. Introduction counting, particularly for small samples. There are

simply too few decays in any reasonable measuring
Radioisotopes have long been used as an im- time.
portant source of information in many areas, such In accelerator mass spectrometry (AMS), these
as earth sciences, anthropology, archaeology, bi- radionuclides, extracted from the sample material
ology, etc. Short-lived radioisotopes can usually be in the ion source, are measured by direct atom
measured with high sensitivity by conventional counting with nuclear detection techniques, after
techniques, in which decay products are counted acceleration to energies in the MeV range. This
eciently. AMS technique has several advantages over decay
Long-lived radioisotopes, such as 10 B, 14 C, 26 Al, counting methods: smaller sample size; shorter
36
Cl, 41 Ca and 129 I, in most cases, cannot be mea- measurement time; higher sensitivity. AMS has
sured at natural levels through radioactive decay found success in many areas of application, such
as astrophysics, hydrology, anthropology, ar-
chaeology, biomedicine, materials science, clima-
*
tology, environmental physics and geology among
Corresponding author. Present address: Universidade
Estadual de Londrina, Campus Universitario, Departamento
others.
de Fisica, Caixa Postal 6001, 86051 Londrina, Brazil. In this work, we prepared and measured some
E-mail address: melayne@onda.com.br (M.M. Coimbra). stromatolite carbonate samples [1], from Salgada

0168-583X/00/$ - see front matter Ó 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 5 8 3 X ( 0 0 ) 0 0 3 9 1 - 8
 M.M. Coimbra et al. / Nucl. Instr. and Meth. in Phys. Res. B 172 (2000) 592±596 593

Lagoon, Rio de Janeiro, Brazil, as part of our is time consuming. In general, sample preparation
AMS program in Brazil [2]. Stromatolites are bio- proceeds in three steps: chemical treatment; com-
sedimentary, laminated, carbonate structures pro- bustion to CO2 ; graphitization to elementary car-
duced by sedimentary, chemical and biological bon.
processes related to the development and growth Contaminants are removed before AMS mea-
of microbial benthic communities, mainly domi- surements by physical and chemical means. This
nated by blue algae and cyanobacteria. These procedure is known as pre-treatment.
structures are present in the geological record in The samples' owner, Cleverson G. Silva, did the
rocks older than 3.0 billion years and have been physical pre-treatment, and the chemical pre-
used to study the origin of primitive life and treatment was performed at PRIME Lab chemis-
variations in past environmental conditions. try facilities [5].
It is presumed that the original inorganic mat-
ter is the source of carbon to be dated. Other
2. Experimental setup carbon, which may have entered the sample later,
is considered to be a contaminant. The outer layer
Detailed AMS measurements were performed of carbonate materials may contain secondary
at PRIME Lab (Purdue Rare Isotope Measure- carbonates.
ment Laboratory, Purdue University, IN, USA). After the physical pre-treatment, the secondary
The AMS system is based on an upgraded FN carbonates were removed by etching the sample
(8 MV) tandem accelerator. with dilute hydrochloric acid, which led to a loss
The ion source is a spherical-ionizer cesium of sample material of up to 50%. After that, these
sputter source, with an eight-sample wheel. Neg- pre-cleaned carbonate samples were dissolved in
ative ions are extracted from the ion source, phosphoric acid. The evolved CO2 was collected
pre-accelerated to 95 keV and the radionuclide is and then transferred to the graphitization line.
selected according to mass by the injector magnet. The carbon dioxide is frozen in two steps to a
Negative ions are then accelerated to the terminal known volume area of the line. During this
of the accelerator, at a potential in the range of transfer, water is removed via dry ice-alcohol
3±7 MV. A foil or Ar-gas stripper removes several traps. After removing an aliquot for d13 C mea-
electrons and the positive ions are accelerated surements, the CO2 sample is frozen into a
again to ground potential. graphitization tube, containing Fe and Zn previ-
Isotope ratios are obtained by cycling the in- ously prepared.
jector and analyzer magnets [3]. The stable iso- This graphitization tube is a 6 mm quartz tube.
topes are measured in the analyzer image Faraday With a glassblowing torch, a dimple is made in the
cup. Radioisotopes are measured in the multi-plate side of the tube, 7 cm from the bottom [6]. This
gas ionization detector [4]. tube is loaded with Zn pellets at the bottom. A
3 mm quartz tube, loaded with Fe powder is slid-
ing down inside the 6 mm tube to the dimple. This
3. Chemical sample preparation tube is connected to the graphitization line. Upper
part of the tube and Fe (not including the Zn) are
The main advantage of AMS is the small baked out in a 700°C tube furnace. The lower part
amount of carbon needed for a 14 C age determi- of the tube is baked out at 500°C and the Zn is
nation. Correspondingly, much less original sam- redistilled to the upper part of the tube to form a
ple material is required for AMS. However, high-surface-area Zn mirror, which puri®es the
handling small amounts of sample material in- Zn.
creases the danger of contamination. Since the The graphitization tube, containing the CO2
addition of small amounts of extraneous carbon sample is sealed and placed in a furnace at 700°C
leads to wrong results, the entire sample prepara- overnight. CO2 is reduced by Zn to CO which, in
tion has to be performed with utmost care, which the presence of Fe at 700°C, disproportionates to
594 M.M. Coimbra et al. / Nucl. Instr. and Meth. in Phys. Res. B 172 (2000) 592±596

C and CO2 . The C is catalytically deposited as as coal, lignite, limestone, ancient carbonate,
graphite on the iron. The resulting graphite is marble or swamp wood. By measuring the activ-
found mixed with the iron and is ready to be ity of a background sample, the normal radioac-
loaded into a cathode and run on the accelerator. tivity present while a sample of unknown age is
being measured can be accounted for and de-
ducted.
Since PRIME Lab currently measures samples
4. Measurements
within the range of 1±2.5 mg C, the mass depen-
dence of the blank is negligible.
The data were processed and analyzed using the
This background may arise from contamina-
software package developed at PRIME Lab [7].
tion during sample preparation, from contamina-
Five samples were measured.
tion in the ion source, and/or from the tails of
Since the mission of PRIME Lab is to provide
other ionic species that the detector cannot dis-
isotope ratio measurements for a number of ele-
tinguish from the radioisotope. To monitor the
ments, the software used is not tailored speci®cally
background a blank is measured before and after
to 14 C measurements. This makes some of the
each group of unknowns and the correction is
equations more complicated than would be nec-
made [8].
essary if only 14 C were measured. As a result there
The value R must be corrected for isotope
are many constants in the equations that cancel
fractionation. Isotope fractionation has the po-
out when the ratio of a sample to a standard is
tential to occur in any physical or kinetic process,
measured.
where the isotope species are selected with less
The quantity of interest in AMS is the ratio of
than 100% eciency. This fractionation is mass
the radioisotope concentration to that of the stable
dependent, but is independent of isotope abun-
isotope(s) which is measured using the count rate
dance, and manifests itself as a change in isotope
for 14 C and the current for 13 C. This ratio, R, is
ratios. In tandem AMS, mass fractionation can
calculated using the relation
originate in four places [9]: during sample prepa-
qeCA ration; during sputtering and negative-ion forma-
Rˆ ; 1† tion; in the terminal stripper; in the beam
I
transport system. Fractionation during sample
where R is the uncorrected measured 14 C=13 C preparation is minimized by utilizing processes
isotope ratio for a sample, q the charge state of the that have nearly 100% yield.
ion selected, e the elementary charge constant Because there is some loss of particles between
(1:602  10ÿ19 C), C the time averaged radioiso- the image cup and the detector and because non-
tope count rate (counts/s), A the attenuation of the linear fractionation e€ects may vary with time, the
stable-isotope beam which includes a factor of measured ratio for the standard is usually di€erent
0.011 to correct for the natural abundance of 13 C, from the ``true'' ratio for the standard. NIST
and I is the averaged stable-isotope current. Oxalic Acid SRM4990C is used for this standard.
It is vital for a radiocarbon laboratory to To correct for this, normalization of the un-
know the contribution to routine sample activity known samples is necessary and the normalization
of non-sample radioactivity. Obviously, this ac- factor is calculated to be the ratio for the standard
tivity is additional and must be removed from measured by AMS to the true (assumed) ratio.
calculations. In order to make allowances for The normalized ratio RNORM uses the mean
background counts and to evaluate the limits of normalization factor measured before and after
detection, materials which radiocarbon specialists the unknown to correct the ratio R for machine
can be fairly sure to contain no activity are fractionation.
measured under identical counting conditions as
normal samples. Background samples usually Ru
RNORM ˆ ÿ Rb ; 2†
consist of geological samples of in®nite age, such Rs ÿ Rtrue
 M.M. Coimbra et al. / Nucl. Instr. and Meth. in Phys. Res. B 172 (2000) 592±596 595

where Ru is the ratio R for an unknown sample, errors. The total statistical error shown in Table 1
Rs the ratio R for the oxalic acid standard is given by the combination of statistical errors
(SRM4990C), Rb the ratio R for background a€ecting the measurements of unknown, standard
subtraction and Rtrue is the accepted value for the and blank samples. The ratios R are measured
standard. with 1% precision, resulting in an error in the age
The natural mass fractionation is measured of 80 years.
through the 13 C value, which is de®ned as
" #
13
C=12 C†sample ÿ 13 C=12 C†PDBstand
13
d Cˆ 5. Results
13 C=12 C†
PDBstand

 1000 3† The stratigraphy of the Salgada Lagoon was

integrated with pollen analysis of the lagoon's
and is expressed in per mil (½). PDB is a 13 C=12 C sediments permitting the recognition of distinct
standard prepared from belemnites collected from climatic ¯uctuations during the lagoon's geological
the Peedee Formation of South Carolina. evolution. Thin sections of the stromatolite heads
Since the radiocarbon age is always expressed show diverse internal arrangement of pore space
with the d13 C corrected to a value of ÿ25 [10] it is and laminations from base to top on each indi-
necessary to do a d13 C correction: vidual stromatolite head. This internal structural
variation has been attributed to past climatic
1 ÿ 25=1000† ¯uctuations, which induced changes in salinity and
Dˆ : 4†
1 ‡ d13 C=1000† on the water depth of the lagoon during stromat-
olite development [11]. The observed variation was
The corrected value for R is used to calculate the
further constrained using the AMS radiocarbon
conventional radiocarbon age BP, using the
ages of one stromatolite head, revealing that
equation
stromatolite development started at 2260  80
t ˆ ÿ8033 ln F ; 5† years BP. This age is coincident with drier climatic
conditions developed around 2540  60 years BP
where F is the fraction modern. (conventional radiocarbon date) as revealed by
RNORM D pollen analysis [12]. Stromatolite development
F ˆ ; 6† ceased around 290  80 years BP which is coinci-
RMODERN
dent with humid conditions developed in the last
where RMODERN is the ratio for 1950 C adjusted to 400 years, also revealed by pollen analysis. During
a d13 C of ÿ25. this last humid phase, abundant microgastropods
The AMS results are shown in Table 1. The are present within the upper organic mud layer.
total error a€ecting AMS measurements is given These microgastropods are known to feed on the
by a combination of statistical and systematic algal mats, inhibiting stromatolite growth.

Table 1
14
Results of C AMS measurementsa

Lab ID # R R error d13 C (½) Age BP Age error

PL9802060A 1200 12 +12.4 290 80

PL9802061A 1077 11 +13.0 1160 80
PL9802062A 1030 10 +18.5 1560 80
PL9802063A 998 10 +16.5 1800 80
PL9802064A 942.9 9.4 +16.1 2260 80
a
All ages are reported in years BP.
596 M.M. Coimbra et al. / Nucl. Instr. and Meth. in Phys. Res. B 172 (2000) 592±596

Average growth rate of the stromatolite head The authors would like to thank David Elmore for
was de®ned as 0.05 mm/year from base to top. the use of PRIME Lab.
However, growth rate variations were observed
through time, re¯ecting changes in internal mor-
phological arrangements of the stromatolite struc- References
tures. Higher rates of stromatolite growth are
observed on the lower 6 cm of the stromatolite [1] M.M. Coimbra, C.F. Barbosa, A. Soares-Gomes, C.G.
head, varying from 0.082 to 0.104 mm/year. This Silva, R. Rios-Netto, K.A. Mueller, Nucl. Instr. and Meth.
portion is highly porous and friable, with incipient 172 (2000) 589.
[2] J.C. Acquadro, G.M. Santos, R.M. Anjos, R. Gomes,
laminations and increasing lamina organization
C. Tenreiro, M.M. Coimbra, C.R. Appoloni, Nucl. Instr.
towards the top. Prominent, well-organized lami- and Meth. B 123 (1997) 34.
nations and columnar structures are developed on [3] M. Perry et al., Nucl. Instr. and Meth. B 123 (1997) 1178.
the intermediate section, which shows lower growth [4] D.L. Knies et al., Nucl. Instr. and Meth. B 92 (1994) 134.
rates of only 0.005 mm/year. The upper section, [5] P. Sharma et al., Nucl. Instr. and Meth. B 123 (1997) 199.
[6] A.T. Wilson, Radiocarbon 34 (1992) 318.
1.5 cm thick, is highly discontinuous and disorga-
[7] E.S. Michlovich et al., PRIME Lab report PL-9303, 1993.
nized, being more massive and non-laminated. This [8] D. Elmore et al., Nucl. Instr. and Meth. B 5 (1984) 233.
section presents again a relative increase in stro- [9] D. Elmore, in: Proceedings of the Symposium on Accel-
matolite growth, estimated in 0.015 mm/year. erator Mass Spectrometry, Argonne National Laboratory,
1981, p. 346.
[10] M. Stuiver, H. Polach, Radiocarbon 19 (1977) 355.
[11] R.T. Lemos, C.G. Silva, in: 14th International Sedimen-
tological Congress, Recife-PE, Brazil, 1994, p. B8 (Ab-
Acknowledgements
stracts).
[12] M.V. Toledo, O.M. Barth, C.G. Silva, Late Holocene
Facility support has been provided by the Na- climatic change in southeastern coastal Brazil: a palyno-
tional Science Foundation, grant EAR-9809983. logical and isotopic approach, submitted.