You are on page 1of 15

Available online at www.sciencedirect.

com

Geochimica et Cosmochimica Acta 75 (2011) 4441–4455


www.elsevier.com/locate/gca

Biotic and inorganic control on travertine deposition at


Bullicame 3 spring (Viterbo, Italy): A multidisciplinary approach
F. Di Benedetto a,b, G. Montegrossi c,⇑, A. Minissale c, L.A. Pardi d, M. Romanelli a,
F. Tassi b, A. Delgado Huertas e, E.M. Pampin f, O. Vaselli b,c, D. Borrini b
a
Dip. Chimica, Università di Firenze, Via della Lastruccia, 3, 50019 Sesto F.no, Italy
b
Dip. Scienze della Terra, Università di Firenze, Via G. La Pira, 4, 50121 Firenze, Italy
c
CNR-IGG Istituto di Geoscienze e Georisorse, Via G. La Pira, 4, 50121 Firenze, Italy
d
CNR-IPCF Istituto per i Processi Chimico-Fisici, Via Moruzzi, 1, 56010 Pisa, Italy
e
Estación Experimental del Zaidı́n (CSIC), Prof. Albareda, 118008 Granada, Spain
f
Genetica (Produccion Animal), Facultad de Veterinaria, Universidad Complutense, 28040 Madrid, Spain

Received 3 June 2010; accepted in revised form 6 May 2011; available online 18 May 2011

Abstract

A multidisciplinary characterization of an active thermal spring in central Italy has been undertaken with the aim of (i)
ascertaining whether microbiological activity plays a relevant role in hot-depositing travertines and (ii) establishing an exper-
imental protocol able to identify similar effects in fossil travertines. Water, gas, and travertine samples were investigated by
chemical (ICP/MS, SEM/EDS), physical (DTA–DTG), isotopic (d18O, dD, and d13C), mineralogical (XRPD), and spectro-
scopic (EPR) techniques. Twenty-four samples (three for each phase) were collected every 5 °C temperature drop, along a
100 m long artificial channel near Viterbo (Bullicame 3, Latium, central Italy). A microbiological characterization was car-
ried out in parallel, sampling the channel every 10 °C temperature drop.
The Bullicame 3 system is revealed to be composed of two markedly different subsystems: a water/gas interface, where a
kinetically fast exchange allows equilibrium of components both in water and in gases; a solid/water interface, where traver-
tine precipitation occurs, influenced by microbiological activity. A peculiar lattice shrinking of calcite was identified, as well as
an anomalous value of the zero-field splitting parameter from the EPR measurements. The interpretation of these anomalies is
confirmed by the identification of calcifying cyanobacteria throughout the channel path.
Our results point out that microbiological activity can play a significant role in travertine deposition from hot springs. Fur-
thermore, the proposed approach, representing a tool to identify crystal chemical remnants of past microbiological activity,
could be applicable to fossil travertines.
Ó 2011 Elsevier Ltd. All rights reserved.

1. INTRODUCTION (e.g. Andrews, 2006), neotectonics and paleohydrology (e.g.


Hancock et al., 1999; Minissale et al., 2005); (ii) geochemi-
In the last decades thermogene and/or meteogene (tufa) cal repository of pollutants (e.g. Di Benedetto et al., 2006);
travertine deposits have received an increasing interest, par- and (iii) building stones involving technological and cul-
ticularly after the discovery of the involvement of bacteria tural heritage (Attanasio, 1999).
in their formation (Chafetz and Folk, 1984) and the role The role of biological activity (e.g. bacteria, algae, micro-
they played in the appearance of life on the planet (Farmer, organisms, and so forth) in travertine formations is the focus
2000). Travertine deposits are important geological materi- of a close scientific debate, namely concerning whether the
als since they are used as: (i) proxy in climate change studies presence of biotic activity could hinder the reconstruction
of paleoclimatic environments (Andrews and Riding, 2001;
⇑ Corresponding author. Tel.: +39 055 275 7477. Fouke, 2001). Two opposite approaches have been pro-
E-mail address: montegrossi@igg.cnr.it (G. Montegrossi). posed: (a) Guo et al. (1996), by studying the Rapolano (Italy)

0016-7037/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.gca.2011.05.011
4442 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

active springs (low temperature, T = 29–34 °C), suggested (northern Latium, Italy), hosting several hot water
microbial activity to significantly affect the carbon isotopic discharges in a small homonymous natural park.1 One of
signature (13C/12C ratio increases downstream), probably them, known as Bullicame 3 (Duchi et al., 1985; Pentecost,
due to the preferential metabolization of 12CO2 by living 1995; Minissale et al., 2002), is artificially elevated over the
organisms; (b) Fouke et al. (2000), in a study of the Angel ground few meters and it is canalized to form a simple
Terrace springs (Yellowstone Park, USA; T = 71–73 °C), open-air Roman-style thermal bath. The thermal discharge
suggested a totally inorganic process controlling the fraction- (57 °C) is confined in a 3 m in diameter pool and is accom-
ation of 12CO2 during travertine deposition. panied by a vigorously bubbling gas (CO2-rich) phase. A
In recent studies, two main topics were considered. On relatively narrow (14 cm) channel allows the thermal water
the one hand, theoretical studies showed that a non- to gently flow for about 112 m to the end of the channel,
equilibrium isotopic signature cannot fully discriminate be- where a cooler pool (frigidarium) is equilibrated with the
tween concurring kinetically-driven and biogenic deposi- ambient temperature (Fig. 1). All artificial structures, chan-
tions (Skidmore et al., 2004; Scholz et al., 2009). On the nels and pools, made up by bricks, are covered by a 10–
other hand, Pokroy et al. (2006) and Zolotoyabko et al. 20 cm thick travertine deposit.
(2010) identified structural and chemical evidences in bio- The Bullicame thermal spring discharges have been
genic calcite, namely: (a) the presence of a permanent lattice known since the Etruscan age. The Etruscan and the Ro-
strain (elongation along the c-axis) unrelated to the chemi- man civilizations were particularly attracted by thermal
cal composition of calcite (in particular to the occurrence of baths and, consequently, fostered the development of this
cation/anion replacement) and (b) the presence of biogenic area. The first documented evidence of the use of this loca-
sulfur, i.e. S-bearing biomolecules co-precipitated or tion as a spa resort dates back to the I–II century B.C.,
embedded in the carbonate fabrics. although it was not associated to any important conurba-
Moreover, Lee et al. (2006) discussed the capability of the tion, but likely related to private country houses; urban set-
Sinechococcus cyanobacteria of bio-precipitating calcite, and tlements are indeed described in the VII century B.C.
the effects of external chemical and physical constraints (e.g. (Giannini, 1969). Successively, Viterbo and the Bullicame
carbonate/bicarbonate availability, pH) on the bacterial site were developed by Papal State during the XIII century
activity. Montegrossi et al. (2006), in a study on some Italian AD. This thermal area is also cited by Dante in the Divina
travertine deposits, revealed an unusual variability of the Commedia: “Quale del Bullicame esce ruscello/che parton poi
zero-field splitting parameter by Electron Paramagnetic Res- tra le lor peccatrici/tal per la rena giù sen giua quello./Lo
onance (EPR) spectroscopy, suggesting that chemical order- sfondo suo ed ambo le pendici/fatt’eran pietra, è margin di
ing of cations in travertine can be investigated, as already lato” (Dante Alighieri, Hell XIV, 79–84).
proposed in the pioneering study by Kretz (1982). Calcite is The spring of Bullicame 3, already studied in detail by
known, in fact, to retain many non-equilibrium depositional Duchi et al. (1985) and Pentecost (1995), has a flow section
features (e.g. Wang and Merino, 1992; Klein and Hurlbut, of 150 cm2 with a relatively constant water speed (7.7 cm/s)
1993; Humphrey and Howell, 1999). and flow rate (0.7 L/s). Travertine fabrics are constant
A multidisciplinary study was undertaken with the aim of throughout the whole channel and can be described in
exhaustively monitoring the gas/water and water/solid inter- terms of the “shrub” assemblages, already pointed out by
faces in an active thermal spring (named Bullicame 3; 57 °C, Pentecost (1995) and Guo et al. (1996) (Fig. 2).
according to Duchi et al., 1985; otherwise cited as “Piscina Eight samples (Tr1–Tr8) of travertine, water, and gas
Carletti”), located north of Rome, close to the Etruscan city were collected along the channel, in which temperature var-
of Viterbo (central Italy), in order to gain complete informa- ies between 57 and 22 °C. The sampling was performed at
tion on the physical, chemical and biological parameters. constant temperature decrease of 5 °C (Fig. 1b). An addi-
Bullicame 3 represents an excellent natural laboratory to tional gas sample (Tr0) was taken at the thermal spring dis-
investigate the parameters acting on the travertine precipita- charge (i.e. the free bubbling gas at the center of the spring,
tion process, such as (i) temperature, (ii) CO2, and (iii) bacte- Fig. 1b).
rial activity (Pentecost, 2005). Physical, chemical, and
isotopic features of the gas, liquid, and solid phases were 2.2. Gas sampling and analysis
determined, and the resulting information was combined
with lattice deformation, thermal analysis (DTA–DTG) The free gas phase at the spring was collected by using
and Mn(II) EPR measurements of the travertine samples. an upside-down funnel connected to a pre-evacuated
Moreover a thorough investigation of the microbiological 60 mL glass flask equipped with a Teflon stopcock and
species present in the thermal water was performed. filled with 20 mL of a 4 N NaOH and 0.15 M Cd(OH)2 sus-
pension (Montegrossi et al., 2001).
Water samples for the analysis of the dissolved gases
2. SAMPLING AREA AND EXPERIMENTAL
were collected into one-way, pre-evacuated 125 mL glass
PROCEDURES
flasks equipped with Teflon valves that were filled with
2.1. Location and brief outlines of the Bullicame thermal
discharge system 1
As reported in the may 2010 volume (Italian edition) of
National Geographic, the thermal spring and the park are presently
Bullicame (42.25N/12.03E) is a thermal area located threatened by the project of a large airport that should be realized
3 km west of the Etruscan and after papal city of Viterbo in the area.
Biotic and inorganic control on travertine deposition 4443

Fig. 1. (a) Aerial photograph and map of the investigated area and (b) location of the travertine, water, and gas samples.

by using liquid N2 and a mixture of liquid N2 and trichlo-


roethylene (Evans et al., 1988; Vaselli et al., 2006). Internal
(Carrara and San Vincenzo marbles) and international
(NBS18 and NBS19) standards were used for the estimation
of external precision. Analytical error for d13C values is
±0.05&.

2.3. Water sampling and analysis

Two aliquots of filtered (at 0.45 lm) water samples were


collected in 125 and 50 mL plastic bottles (after adding
0.5 mL of concentrated HNO3), for anion and cation anal-
ysis, respectively. Unfiltered water (20 mL) was collected in
glass bottles for isotopic analysis (O, H, and C-DIC, Dis-
solved Inorganic Carbon) after adding few mg of HgCl2
to prevent any microbial activity. Temperature, pH,
Fig. 2. SEM micrograph of a travertine fragment.
HCO3 (acidimetric titration) and NH4+ (Nessler colori-
metric method) contents were determined in situ. The water
50–60 mL of channel water (Tassi et al., 2008). The deter- samples were analyzed with a Dionex DX100 ion chro-
mination of the gas species in the head-space was per- matograph (F, Cl, Br, NO3, and SO4) and a Perkin-Elmer
formed by gas-chromatography (Vaselli et al., 2006), and AAnalyst 100 atomic absorption spectrophotometer (Na,
the total amount of the dissolved gases was calculated K, Ca, Mg, Fe, and Mn); dissolved CO2 concentrations
according to Henry’s law. The contents of CO2, N2, O2, were determined via acidimetric titration (Tassi et al.,
H2S, CH4, Ar, Ne, and H2 were determined using Shimadzu 2004).
14A and 15A gas-chromatographers equipped with Flame Oxygen isotopes were analyzed by using the CO2–H2O
Ionization Detector (FID) and Thermal Conductivity equilibration method (Epstein and Mayeda, 1953). The
Detector (TCD), respectively. The 13C/12C isotopic ratio hydrogen isotopic measurements were carried out by
of CO2 (expressed as d13C & V-PDB) was determined by reducing the water sample to hydrogen with hot zinc at
a Finnigan Delta S mass spectrometer after a two-step 550 °C, according to the analytical procedure described
extraction and purification procedure of the gas mixtures by Coleman et al. (1982). Oxygen and hydrogen isotopic
4444 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

compositions (expressed as d18O and dD & V-SMOW,


respectively) were determined using a Delta Plus XL and
MAT 251 (Finnigan) mass spectrometers. The analytical er-
ror for d18O and dD values was ±0.1& and ±2&, respec-
tively. The d13C values of Dissolved Inorganic Carbon
(DIC) were determined with the same mass spectrometer
used for the d13C–CO2 analysis, after the reaction of
3 mL of water with 2 mL of anhydrous phosphoric acid
in a under-vacuum line (Salata et al., 2000). The recovered
CO2 was analyzed after the two-step extraction and purifi-
cation procedures. Analytical error for d13C was ±0.05.

2.4. Travertine samples

Travertine fragments were sampled from the bottom of


the channel at the contact point with the collected waters
and dissolved gases. Samples were obtained after removing
the encrustations, and then gently powdered in an agate Fig. 3. EPR derivative plot (arbitrary units) of the sixth hyperfine
line of the Mn(II) spectrum, putting in evidence the relevant points
mortar after drying at 40 °C.
for the D parameter evaluation. In the inset, second derivative plot
Chemical analyses of travertine samples were performed of the same spectrum, putting in evidence the relevant points for
at the Acme Analytical Laboratories Ltd. (Vancouver, the WD parameter evaluation. Magnetic field values are expressed
Canada), after aqua regia digestion, by means of Atomic in Gauss.
Absorption Spectrophotometry (AAS) and Induced Cou-
pled Plasma–Mass Spectrometry (ICP/MS), whereas C difference between the field positions corresponding to the
and S were analyzed by Leco Gas chromatography with relative maximum and minimum of this partial spectrum
an IR detector. gives the D6 parameter (Montegrossi et al., 2006).
X-ray powder diffraction (XRPD) patterns were ob- Subsequently, after a smoothing of the raw spectrum to re-
tained using a Philips PW 3710 diffractometer with Cu anode duce the experimental noise, a numerical derivative curve
and graphite monochromator, equipped with PC-X’Pert Pro was calculated. The difference between the field positions
software for data acquisition and handling. Experimental corresponding to the two relative maxima (Fig. 3) yields
conditions were 20 mA current, 40 kV voltage, 20–80° 2h the W6 parameter. The D6 and W6 parameters can be re-
range, step size 0.02° 2h, 2.5 s per step. The powder struc- lated, via the quantification of the zero-field interaction,
tural refinement was carried out through the Rietveld algo- to the strength of the crystal field and to the structural dis-
rithm using the Rietquan software (Lutterotti et al., 1998). order, respectively. These values were converted to the cor-
Thermal factors for Ca, C, and O are from Markgraf and responding D (most frequent value of the axial ZFS
Reeder (1985). A reference synthetic calcite sample (Rudi parameter) and WD (distribution of the axial ZFS interac-
Pont, Turin, Italy) was used as external standard for the dif- tion) values by means of the following relations2:
fractometer calibration (lattice constants, peak shape, and pffiffiffiffiffiffi
2h offset). A raw travertine encrustation was also analyzed D ¼ 0:24294 D6 ð1Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
without pre-treatment, in order to check whether the grind- W D ¼ 0:48588  W 6  D6 ð2Þ
ing procedure would favor the conversion of aragonite onto
calcite (e.g. Criado and Trillo, 1975). Scanning electron micrographs (up to 104 magnifica-
Conventional EPR spectroscopy measurements were tion) of selected samples were obtained by using a SEM
performed on free powdered samples, in amorphous silica Philips 515 microscope interfaced to a personal computer
tubes. Data were collected at room temperature using a through a digital acquisition software. Point and raster mi-
Bruker ER 200D-SRC spectrometer operating at X-Band cro-analytical data were obtained by means of an energy-
(9.5 GHz) interfaced with DS/EPR software to a PC for dispersive analyzer and an EDAX 9800 analytical system.
data acquisition and handling. The actual operating A constant beam was maintained during the investigations,
frequency value was determined by using DPPH radical the current and voltage being 20 nA (on the sample) and
[2,2-di(4-tert-octyl-phenyl)-1-picrylhydrazyl, g = 2.0037] as 25 kV, respectively.
external standard. Due to the narrow line-width of Mn(II) Thermal analyses were performed on a Netzsch STA 409
spectra in calcite (Vassilikou-Dova, 1993), particular care apparatus (accuracy ±3 °C) on finely ground travertine
was paid to the instrumental field modulation and sweep samples. Pure fine grained (<63 lm) alumina, a-Al2O3,
rate. The chosen parameters were 0.5 G modulation ampli- was used as reference standard. Powders were contained
tude and 100 kHz modulation frequency. Scan speed was in alumina open crucibles, heating rate was 3 °C/min, max-
set to 2 G/s. The full Mn(II) spectrum was preliminarily imum programmed temperature was 280 °C, focusing on
identified in each sample, in the magnetic field range
2800–4000 G. Successively, a detailed spectrum was 2
Relations (1) and (2) are defined for D, WD, D6, W6 values
recorded in the range 3600–3700 G, where the sixth hyper- expressed in Tesla (1T = 10000 G), according to Montegrossi et al.
fine line of the central ½ ) ½ transition appears. The (2006).
Biotic and inorganic control on travertine deposition 4445

the temperature range of sulfur melting. Besides an Al2O3/

CO2(aq)
Al2O3 run, to account for instrumental effects, temperature

9.68
6.91
6.94
4.85
4.63
2.99
3.22
2.96

6.80
and calorimetric calibration were also performed, using


low-melting pure standard substances, namely phenan-

 a
threne (C14H10; 99.3 °C), elemental In (156.61 °C) and ele-

166
153
151
138
142
132
130
129
190
F


mental Se (217 °C).

 a
2.5. Isolation of microorganisms

113
107
110
83
71
77
81
80
HS



The presence of fungi, bacteria and phytoplankton was

NO3 a
investigated in four water samples having different temper-

100.8
12.5
12.9
28.2
84.7
2.4
2.4
1.8

All values are expressed in meq/L, with the exception of those analytes marked with superscript ‘a’ and CO2(aq) that are in leq/L and mmol/L, respectively.
atures (57, 47, 37, and 27 °C, respectively). One hundred



microliters of each water sample was placed in Petri dishes

 a
with two culture media: Sabouraud Dextrose Agar (Bio-

0.94
0.78
0.91
0.94
1.10
1.10
0.94
1.10
Br
merieux, Madrid, Spain) and Columbia Blood Agar (Bio-



merieux, Madrid, Spain) for the isolation of fungi and

2
bacteria, respectively. The incubation was performed at

21.88
21.88
21.88
22.92
22.92
23.96
21.88
20.83
23.00
22.8
SO4
25 and 37 °C. Yeast identification was made by macro-

Physical–chemical features of the Bullicame 3 thermal spring discharge and of the water samples collected along the artificial channel.
scopic observation and microscopic morphology, by apply-

Chemistry (anions)

0.416
0.416
0.409
0.367
0.430
0.451
0.367
0.388
0.400
ing standard taxonomic criteria (de Hoog et al., 2000).

0.40

Cl
Bacteria identification was made by macroscopic and
microscopic observation, and by simple biochemical tests,


applying standard taxonomic criteria (Quinn et al., 2000).

HCO3
16.08
16.17
16.11
13.69
12.66
11.64
11.96
11.01
15.60
14.6
Microalgae were identified in fixed samples (4% PBS-buf-
fered formalin) in settling chambers using an inverted
microscope (Axiovert 35, Zeiss, Oberkochen, Germany).
+ a

18.7
18.7
20.2
18.7
17.3
18.7
18.7
18.7
21.3
The identification of algae was carried out in accordance
Li


with John et al. (2002) procedure. NH4+ a

860.0
25.7
17.9
15.0
7.1
3.6
3.6
3.6
3.6
2.6. Theoretical modeling of isotopic signatures


Measured isotopic signatures were modeled through the-
0.86
0.90
0.88
0.94
0.91
0.91
0.89
0.89
0.88
0.85
+

oretical calculations, performed on the basis of (i) measured


K

d13C-DIC ratios, (ii) temperature, and (iii) bulk chemical


composition of waters, gas, and travertine. Numerical simu-
+

1.39
1.37
1.34
1.35
1.39
1.43
1.39
1.41
1.52
2.80
Na

lations of transport and isotope fractionation provide, in


Chemistry (cations)

fact, a method to quantitatively interpret isotope fraction-


2+

9.38
9.71
9.88
9.71
9.96
9.88
9.63
9.88
9.20
ation profiles based on evaporation, gas diffusion and calcite
11.5
Mg

precipitation. The temperature-dependent equilibrium con-


stants of the considered isotopic species, occurring in the dif-
23.55
23.50
23.60
23.15
22.55
24.05
20.30
20.45
26.40
2+

ferent phases of the system, (taken from the database of


28.0
Ca

PHREEQC, Parkhurst and Appelo, 1999) were included


into the thermodynamic database of the reactive transport
6.54
pH
6.2
6.7
7.3
7.6
7.8
7.9
7.7
7.8
6.6

code TOUGHREACT (Xu and Pruess, 2001; Spycher and


Pruess, 2005; Xu et al., 2006), by means of the EOS2 model.
The different diffusive transport properties of the considered
T (°C)

56.3

species (Palmer, 1994; Tamini et al., 1994; Frank et al., 1996)


57
52
47
42
37
32
27
22
57

and water density dependence on NaCl and CO2 (Garcia,


Physical features

2001) were also considered. These calculations have been


Distance (m)

used to illustrate the evolution of stable isotope profiles


along the channel, i.e. outside the spring pool (Fig. 1b).
26.34
50.14
56.54
60.84
67.44
80.04
94.64

The channel was modeled by an array of 99  30  31


0

0
0

elements, each one having dimension 1.0  0.005 


0.005 m3, respectively. The bottom and lateral boundaries
Duchi et al. (1978)

(walls of the channel) are simulated by two layers of calcite


Pentecost (1995)

elements (porosity: 6%, permeability: 1015 m2). The upper


boundary layer simulates atmosphere, having a fixed buffer
of 400 lmol/mol of CO2. This layer also plays the role of
Table 1

Sample

thermal buffer, being maintained at a constant temperature


Tr1
Tr2
Tr3
Tr4
Tr5
Tr6
Tr7
Tr8

of 15 °C. An in-flow is located at the warmer side of the


4446 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

Table 2
Dissolved gas samples composition (in % by vol.) from the thermal spring discharge to the end of the artificial channel at Bullicame 3.
Sample O2 N2 CO2 Ar Ne
Tr0 (emission) 0.17 1.12 98.66 0.02 0.000012
Dissolved gas samples
Tr1 6.71 12.76 80.21 0.321 0.00019
Tr2 15.19 29.92 54.15 0.746 0.00050
Tr3 15.73 30.93 52.56 0.773 0.00052
Tr4 16.28 31.95 50.98 0.800 0.00053
Tr5 17.77 35.48 45.86 0.892 0.00053
Tr6 17.75 35.48 45.89 0.878 0.00059
Tr7 18.11 36.18 44.82 0.892 0.00060
Tr8 25.99 51.97 20.76 1.278 0.00078

simulation array, according to the experimental values for


flux and velocity. At the cooler side of the array, the
out-flow is simulated as an extraction well. The initial con-
ditions (at the warmer side) are set up according to the
chemical and physical data obtained for sample 1 (Table 1).
The system is then allowed to evolve, simulating the flow
through the channel.

3. RESULTS

3.1. Water and gas chemistry

The Bullicame spring system is characterized by a Ca-


SO4-HCO3 composition typical of most waters circulating
in the Mesozoic Carbonate Sequence belonging to the Tus-
can Nappe (e.g. Minissale, 2004). The temperature and the
chemical composition of the Bullicame 3 spring, determined
in the present study, is fully comparable to that reported by
Duchi et al. (1985). This compositional stability over such a
long time must be ascribed to the relatively long circulation
path of the thermal fluids. These, in turn, originate mainly
from meteoric precipitations gathered inside the main
Apennines Range (e.g. Minissale, 2004).
The composition of the free bubbling gas phase at the
emergence pool of the spring is dominated by CO2
(98.66 vol.%; Tr0 sample) with minor N2 (1.12% by vol.;
Table 2), a common feature of thermal springs in central
Italy (Minissale, 2004). The dissolved gases in this pool
(sampled few centimeters away from the emergence) are
characterized by significantly lower CO2 (80 vol.%; Tr1
sample) and higher N2 (12.7 vol.%) and O2 (6.71 vol.%)
amounts (Table 2). These compositional trends are likely
caused by the dissolution of atmospheric components
(air) in the water according to the Henry’s law. However,
a dissolved gas contribution related to a hypothetical pho-
tosynthetic activity in the pond cannot be ruled out.
Water samples collected along the channel show a pro-
gressive decrease of the dissolved CO2 concentrations from
9.7 to 3.0 mmol/L (Table 1 and Fig. 4a), from the discharg-
ing area to the end of the channel. A similar decreasing pat-
tern is also observed for [HCO3], whose content varies
from 16.08 down to 11.01 meq/L. The pH increases from Fig. 4. (a) pH values and concentration of the carbon (HCO3 in
6.22 to 7.79. Such chemical–physical changes are likely meq/L and dissolved CO2 in mmol/L) species; (b) d13C (& V-PDB)
due to both CO2 degassing and travertine precipitation, values of Dissolved Inorganic Carbon (DIC) and of dissolved CO2,
the latter being responsible for a decrease of the Ca concen- calculated (continuous line) d13C of dissolved CO2; (c) relative
trations, from 23.55 to 20.45 meq/L. A further significant content of CO2 (in % by vol. in dissolved gas).
Biotic and inorganic control on travertine deposition 4447

Table 3
Isotopic ratios from water samples (dD, d18O, and d13C of Dissolved Inorganic Carbon), from gas samples (d13C of dissolved CO2), and from
travertine samples (d13C and d18O).
Sample Water Gas (CO2) Travertine
dD & V-SMOW d18O & V-SMOW d13C DIC & V-PDB d13C-CO2 & V-PDB d13C & V-PDB d18O & V-PDB
Tr1 38.56 6.25 1.16 3.35 8.24 18.20
Tr2 38.91 6.16 2.80 3.44 6.36 20.11
Tr3 37.97 5.95 3.63 3.19 6.06 20.05
Tr4 37.79 5.78 4.04 2.19 6.17 20.13
Tr5 36.79 5.43 4.74 2.14 5.95 19.74
Tr6 36.32 5.23 5.09 0.97 6.00 20.11
Tr7 35.69 5.16 4.88 3.60 6.28 21.43
Tr8 35.36 5.03 5.27 0.72 6.36 21.20

more positive with decreasing temperature (Table 3 and


Fig. 4b). Thus, both observations suggest that a fast ex-
change with atmospheric CO2 is occurring.
Measured values of d13C in travertine (Fig. 5) abruptly
decrease from an initial value >8 & at 57 °C down to a
nearly constant value of about 6.2&. The lack of a direct
correlation between d13C of dissolved CO2 and d13C in trav-
ertine is a clear fingerprint of an out-of-equilibrium frac-
tionation process. In order to have an insight into this
process, we performed sample calculations, as described
above. Calculated d13C data of dissolved CO2 (Fig. 4b)
and of travertine (Fig. 5) agree with the experimental ones
along the whole series but for the first point.
The calculation results show that diffusion of carbonate
species in solution from water–atmosphere to water–traver-
tine interface is the rate determining step. The calculation,
Fig. 5. Measured and calculated d13C (& V-PDB) values versus in fact, generates vertical concentration profiles of dissolved
temperature (°C) in travertine samples. CO2, as shown in Fig. 6a for the field samples. Isotopic sig-
natures of three levels in the vertical profiles of the channel
(i.e. the bottom, 5 cm and top levels) are shown in Fig. 6b.
compositional variation along the channel is represented by Lighter dissolved CO2 is almost quantitatively preserved at
the progressive decrease of NH4+ and HS contents, cou- the bottom of the system, whereas the top layer readily at-
pled with the simultaneous increase of [NO3] and tains a steady isotopic ratio, due to a balance between
[SO42] (Table 1). This can be interpreted as a consequence degassing of lighter CO2, dissolution of atmospheric gases
of the non-stoichiometric oxidation of the reduced species and diffusion of CO2 in the water phase. An enhanced effect
to the most stable oxidized species. No other significant of the preferential degassing of lighter CO2 is evidenced by
variations of water and dissolved gas chemistry are ob- the ratios calculated at 5 cm (from bottom of the channel),
served (Table 1). where the combined effects of the mass transport and isoto-
pic exchange lead to a progressive increase of d13C value
3.2. Isotopic signature of water, gas, and travertine downstream (see e.g. Chafetz et al., 1991). The relative sta-
bility of the bottom layer of the system (Fig. 6a and b) re-
In water samples, the dD and d18O values become pro- sults in a small variation of isotopic composition of the
gressively less negative along the channel (Table 3). In trav- depositing calcite (Fig. 5, calculated values). According to
ertine encrustations d18O values moderately increase with this calculation, therefore, a full inorganic process could
decreasing temperature (Table 3). This behavior is typical be operative in the system.
of travertine-depositing thermal springs with temperature
>35–40 °C (Gonfiantini et al., 1968; Fouke et al., 2000). 3.3. Mineralogical and chemical composition of the travertine
The variation of d18O in water (from 6.25& to 5.03& deposits
versus V-SMOW) and travertine (from 18.2& to 21.2&
versus V-SMOW) corresponds to that found for the traver- Travertine encrustations are characterized by the mas-
tine–water fractionation process (Thorstenson and sive presence of CaCO3 that occurs as two main poly-
Parkhurst, 2002). The d18O values show an inverse correla- morphs: calcite and aragonite, although the former is by
tion with both the total amount of dissolved CO2 (Fig. 4a far the most abundant (>94 wt%). Pentecost (1995) points
and Table 1) and its relative abundance (CO2 vol.%) in out the presence of strontianite (SrCO3), quartz and gyp-
the dissolved gases (Fig. 4c and Table 2). Moreover, the sum at Bullicame 3 but in the present study they were
measured d13C ratios of DIC and dissolved CO2 become always found to be below the detection limit of 1.0 wt%
4448 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

Fig. 6. (a) Calculated vertical profiles of CO2 content (mmol) versus temperature (°C) in the channel; (b) calculated profiles of d13C versus
temperature (°C) for selected depths in the channel.

(Table 5). The low aragonite/calcite ratio (0.06/0.01) in the Subordinate Fe amounts (up to 0.41% by wt., Tr3),
present XRPD patterns is likely due to a partial intercon- attributable to either Fe-bearing calcite and/or Fe oxy-
version between the two calcium carbonate polymorphs hydroxides, were also detected. Low Al concentrations
(Criado and Trillo, 1975). A XRPD characterization of (up to 0.23% by wt.) are likely related to the presence of
an uncrushed fraction of travertine sample was thus per- alumino-silicate minerals (Table 4). Potassium (as K2O),
formed, and no significant variations in terms of relative Na (as Na2O) and P (as P2O5) are 60.5% by wt. In Table 4
amounts of the two minerals were detected. the content of selected trace elements (Ba, Cr, Mo, Mn,
The chemical composition of all travertine samples is al- Cu, Zn, Cd, Te, As, Sb, Pb, and Th) are also reported.
most constant (Table 4) and is dominated by calcium (up to Mn concentrations are in the range of few hundreds
56 wt% as CaO: Tr5) and CO2 (42.2%: Tr4). The contents ppm, showing a significant decrease from the spring to
of Mg (as MgO), Sr (as SrO) and S range from 0.8 to 1.0, the pool. Among the other trace elements, Cr, Mo, Cu,
from 0.39 to 0.47 and from 0.89 to 1.76 wt%, respectively, Zn, Cd, Te, Sb, Pb, and Th have lower concentrations,
and are comparable with those obtained by SEM. The mea- whereas As and Ba range from 43 to 57 and from 85 to
sured Mg/Ca ratios are consistent with the temperature- 213 ppm, respectively.
dependent solubility of Mg in calcite (Mackenzie et al., The chemical speciation of S in travertine was investi-
1983). Sr shows affinity to both aragonite and, to a minor gated by means of DTA. In all the DTA/DTG thermographs
extent, calcite (Pingitore et al., 1992; Humphrey and Ho- a double endothermic effect appears at peak temperatures of
well, 1999); however, the relatively low Sr contents did 120 °C. No thermal effects appear on the cooling curves.
not allow us to attempt a pertinent attribution via Rietveld As reported in Fig. 7, the calibrated peak temperatures
refinement, and therefore Sr was assumed to be distributed are 115.5 and 123.9 °C, respectively, and the second one is
in both CaCO3 polymorphs. rather broader and somewhat less intense. The observed
Table 4
Major (in % by wt.) and trace element (ppm) composition of the travertine deposits collected along the artificial channel at Bullicame 3 after leaching with aqua regia (HNO3:HCl 1:3).
Sample Major elements (wt%) Minor elements (wt%) Total
CaO CaO CO2 CO2 S S MgO MgO SrO SrO FeO Al2O3 K2O Na2O P2O5
(ICP– (SEM) (GC) (SEM)a (GC) (SEM) (ICP–MS) (SEM) (ICP– (SEM) (ICP– (ICP–MS) (ICP– (ICP–MS) (ICP–MS)
MS) MS) MS) MS)

Biotic and inorganic control on travertine deposition


b
Tr1 42.2(1) 53.0(1) 40.3(2) 42.9 1.76(6) 1.7(1) 0.68(5) 1.0(1) 0.398(1) 0.6(2) 0.03(4) 0.06(6) 0.04(4) 0.036(4) 0.02(1) 85.5(5)
Tr2 52.5(1) 53.9(1) 41.9(2) 43.6 1.07(6) 1.2(1) 0.93(5) 1.0(1) 0.470(1) 0.5(2) 0.21(4) 0.13(6) 0.04(4) 0.046(4) 0.03(1) 97.4(5)
Tr3 54.5(1) 53.9(1) 40.2(2) 43.6 1.05(6) 1.0(1) 0.93(5) 1.0(1) 0.432(1) 0.5(2) 0.41(4) 0.23(6) 0.05(4) 0.063(4) 0.01(1) 97.9(5)
Tr4 53.6(1) 54.1(1) 42.2(2) 43.6 1.03(6) 1.0(1) 0.91(5) 0.8(1) 0.457(1) 0.5(2) 0.04(4) 0.06(6) 0.04(4) 0.042(4) 0.02(1) 98.5(5)
Tr5 56.0(1) 53.9(1) 41.8(2) 43.6 1.00(6) 1.1(1) 0.86(5) 0.8(1) 0.447(1) 0.5(2) 0.05(4) 0.04(6) 0.04(4) 0.050(4) 0.02(1) 100.3(5)
Tr6 54.2(1) 53.9(1) 41.5(2) 43.6 1.13(6) 1.0(1) 0.83(5) 0.8(1) 0.459(1) 0.6(2) 0.01(4) 0.02(6) 0.04(4) 0.049(4) 0.02(1) 98.3(5)
Tr7 53.5(1) 54.0(1) 41.4(2) 43.6 0.89(6) 0.9(1) 0.73(5) 0.8(1) 0.425(1) 0.6(2) 0.01(4) 0.02(6) 0.04(4) 0.050(4) 0.01(1) 97.1(5)
Tr8 54.4(1) 53.7(1) 41.4(2) 43.3 1.04(6) 1.0(1) 0.76(5) 0.8(1) 0.435(1) 0.5(2) 0.01(4) 0.04(6) 0.05(4) 0.046(4) 0.01(1) 98.2(5)
Sample Trace elementsc (ppm)
Mo Cu Pb Zn Mn As Th Cd Sb Cr Ba Te
Tr1 0.04(3) 0.35(3) 0.49(3) 1.1(3) 121(3) 85.5(3) 0.3(3) 0.03(3) 0.03(6) 1(2) 43(2) 0.14(6)
Tr2 0.06(3) 1.06(3) 0.71(3) 2.5(3) 168(3) 176.5(3) 0.1(3) 0.08(3) 0.03(6) <1 50(2) 0.19(6)
Tr3 0.13(3) 11.54(3) 2.07(3) 4(3) 184(3) 198.7(3) 0.1(3) 0.18(3) 0.07(6) 5(2) 57(2) 0.15(6)
Tr4 0.04(3) 0.39(3) 0.92(3) 1.5(3) 114(3) 193.8(3) 0.2(3) 0.04(3) 0.03(6) <1 49(2) 0.21(6)
Tr5 0.04(3) 1.66(3) 0.80(3) 0.8(3) 114(3) 213.8(3) <0.1 0.04(3) 0.03(6) 1(2) 52(2) 0.18(6)
Tr6 0.05(3) 0.21(3) 0.44(3) 0.7(3) 107(3) 192.9(3) 0.1(3) 0.11(3) 0.02(6) <1 50(2) 0.18(6)
Tr7 0.01(3) 0.12(3) 0.17(3) 0.4(3) 92(3) 173.4(3) <0.1 0.03(3) <0.02 <1 48(2) 0.17(6)
Tr8 0.03(3) 0.15(3) 0.26(3) 0.6(3) 98(3) 182.9(3) 0.1(3) 0.02(3) <0.02 <1 49(2) 0.14(6)
a
SEM CO2 data were recalculated considering CaO, MgO, and SrO contents as referred to CaCO3, MgCO3, and SrCO3.
b
CaO wt% for sample Tr1 was obtained with AAS.
c
Other trace elements were below the operative detection limit (cf. Acme Analytical Laboratories Ltd.): Hg < 10 ppb; Ag, Au < 40 ppb; Tl < 0.02 ppm; Bi < 0.04 ppm; Co, U, W, Se < 0.2 ppm;
Ni, Ga, Sc < 0.6 ppm; V < 2 ppm; La < 3 ppm; Ti < 10 ppm, B < 50 ppm.

4449
4450 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

25.2(1)
28.7(1)
45.1(1)
45.1(1)
43.9(1)
50.8(1)
43.0(1)
47.9(1)
WD
110.4(1)
114.4(1)
111.3(1)
112.2(1)
111.1(1)
111.0(1)
112.0(1)
111.0(1)
EPR parameters (in Gauss, G)

D
25.74(4)
28.19(4)
30.48(4)
30.87(4)
30.14(4)
31.66(4)
30.34(4)
30.97(4)
W6

Fig. 7. Thermograph of the Tr1 sample: calibrated T (in °C, black)


20.74(4)
22.25(4)
21.07(4)
21.40(4)
21.01(4)
20.96(4)
21.35(4)
20.96(4) and DTA (gray) curves are shown. The vertical lines indicate the
maxima of the two thermal effects.
D6
17.032(1)
17.021(1)
17.021(1)
17.022(1)
17.024(1)
17.026(1)
17.031(1)
17.029(1)
c0 (Å)
Theoreticalb

4.9846(1)
4.9827(2)
4.9826(2)
4.9828(2)
4.9832(2)
4.9835(2)
4.9843(2)
4.9840(2)
Structural and EPR parameters of calcite from the travertine deposits collected along the artificial channel at Bullicame 3.

a0 (Å)
365.41(7)
365.47(7)
365.47(7)
365.76(7)
365.63(7)
365.76(7)
366.07(7)
365.59(7)
367.50(5)
V0 (Å3)

367.61

Fig. 8. Lattice constants (in Å) and cell volume (in Å3) for the
investigated samples, from the beginning (Tr1) to the end (Tr8) of
0.2596(5)
0.2599(5)
0.2615(5)
0.2613(5)
0.2677(5)
0.2603(4)
0.2588(5)
0.2631(5)
0.2580(4)
XRPD calcite structural parameters

the artificial channel. Std: synthetic reference sample (Markgraf


0.2567

and Reeder, 1985).


x

temperatures are in agreement with the known values for


17.094(1)
17.086(1)
17.093(2)
17.097(1)
17.094(1)
17.097(1)
17.098(1)
17.088(1)
17.064(1)

thermal effects of elemental crystalline sulfur, the first peak


17.061
c0 (Å)

ascribable to the rhombic–monoclinic transition (113 °C)


and the second to the melting of monoclinic S (123 °C). On
Experimental

the basis of the performed calorimetric calibration, the mea-


Calculated according to the ICP Mg and Sr contents of Table 4.

surement of the area of the melting peak allowed to quanti-


4.9682(2)
4.9699(2)
4.9688(3)
4.9700(2)
4.9697(2)
4.9703(2)
4.9721(2)
4.9703(1)
4.9868(2)

tatively estimate the S-amount present in the sample, that


4.9880
a0 (Å)

resulted in 2.7 wt% in Tr1 sample. Taking into account the


relative uncertainties of the different methods, the obtained
value agrees with the chemical data reported in Table 4.
According to Pokroy et al. (2006), some living organisms
Ref.: Markgraf and Reeder (1985), see text.
XRPD phase composition (wt%)

can control the precipitation of calcite by using S-bearing


Aragonite

macromolecules, the presence of which determines a prefer-


ential route for the growth of bio-calcite. At the end of the
2.6(2)
2.2(2)
2.3(4)
5.6(2)
2.8(2)
2.6(2)
0.9(2)
1.1(2)

process, the macromolecules would remain “embedded” in


the crystallites. On the basis of the present data, the occur-
rence of this second mechanism could not be ruled out,
although it could play a minor role.
Calcite
97.4(5)
97.8(4)
97.7(7)
94.4(5)
97.2(5)
97.4(4)
99.1(4)
98.9(4)

3.4. Structural characterization

The structural data are reported in Table 5. The value of


lattice constants, a0 and c0, refined on the whole series of
travertine samples, are nearly constant; however, they sub-
Table 5

Sample

Ref.a

stantially differ from the literature’s values reported for this


Tr1
Tr2
Tr3
Tr4
Tr5
Tr6
Tr7
Tr8
Std

a
b

range of temperatures (e.g. Reeder, 1983). A standard


Biotic and inorganic control on travertine deposition 4451

synthetic calcite was analyzed, to verify the correctness of and are in agreement with literature values (e.g. Shepherd
the acquisition of the XRPD powder patterns and/or the and Graham, 1984; Di Benedetto et al., 2005). The experi-
refinement procedure: the determined lattice constant val- mental D6, D, W6, and WD values are reported in Table 5.
ues excellently agree with those obtained by Markgraf The experimental D values are roughly constant, whereas
and Reeder (1985) on a natural crystal (Table 5). W6 and WD significantly change, i.e. they increase as tem-
Bullicame 3 samples show negative Da0 (0.4%) and posi- perature decreases from 57 °C (W6: 25.7 and WD: 25.2) to
tive Dc0 (+0.2%) with respect to the reference values for pure 47 °C (W6: 30.5 and WD: 45.1). At lower temperatures they
calcite (Fig. 8). Remarkably, differences in a0 and c0 have remain almost constant. These data suggest that the mean
opposite signs. As a result of this lattice variation, an overall strength of the crystal field interaction of Mn(II) does not
decrease of the cell volume, V0, is observed (Fig. 8). The change in the whole temperature range, whereas the struc-
refinement of the x fractional coordinate of the oxygen atom, tural disorder, i.e. the presence of lattice strains and defects
which is the only free atomic coordinate in the structure, gave nearby Mn(II) ions, increases by lowering the temperature
no reliable results. Some variations of this parameter are until a steady value is attained. It must be pointed out that
observed along the series, but the poor agreement between the WD increase with decreasing temperature cannot be ex-
literature data and the standard calcite suggests X-ray data plained in terms of inhomogeneous (dipolar) broadening of
not to be reliable for a more detailed analysis (Table 5). the line, because Mn concentration actually decreases as
Theoretical a0 and c0 values, calculated according to the temperature decreases (Table 4).
Mg and Sr contents of each sample (Table 4), are also re- The mean D value, about 111 G, is unexpectedly high
ported in Table 5. The Ca–Mg substitution (Mg is the most when compared with pure calcite’s value (about 80 G, e.g.
abundant major ion substituting for Ca, Table 4) is known, Shepherd and Graham, 1984). This value cannot be ex-
in fact, to induce a decrease of the lattice constants, due to plained even considering the existence of dolomite-like mi-
the difference in the ionic radii between Mg2+ (0.65 Å) and cro-domains, since Mn(II) in Mg-dolomite sites is reported
Ca2+ (0.99 Å). On the other hand the substitution of Ca with to have a D value of the order of 150 G (Shepherd and
Sr, which was conservatively assumed to be contained in cal- Graham, 1984). Moreover, no pre-edge peak characterizing
cite (although in general preferentially partitioned in arago- EPR spectra of Mn(II) in dolomite has been observed at
nite, Sunagawa et al., 2007), causes an increase of the lattice 3100 G (Attanasio, 1999).
constants, Sr2+ (ionic radius: 1.13 Å) being larger than Ca2+ No evidence of other paramagnetic impurities, e.g.
(Pingitore et al., 1992). Thus, both Mg and Sr should be effec- Fe(III), of calcite nor other magnetic associated phases,
tive in changing (decrease/increase) both lattice constants, such as MnO2 or Fe oxy-hydroxides, were observed by
even if this effect would be more pronounced for c0 (Paquette EPR, thus ruling out their possible interference in the
and Reeder, 1990). The experimental values for a0 and c0 of Mn(II) spectral analysis. As a consequence, the relatively
Bullicame 3 travertine are, therefore, apparently due to a low Fe-content (Table 4) can be attributed to the presence
different cause than simple Mg and/or Sr substitution. of the EPR-silent Fe(II) ion in calcite.

3.5. EPR spectroscopy 3.6. Water microorganisms

The experimental EPR spectra show Zeeman and hyper- The main microorganisms isolated along the channel of
fine parameters to remain constant over the whole series Bullicame 3, from high to low water temperatures are

Table 6
Microbial diversity observed in the water samples collected along the artificial channel of Bullicame 3.
Water Fungi Bacteria Microalgae
samples (T)
Tr1 (57 °C) Enterobacteriaceae (Gram bacteria, Cyanobacteria: Synechococcus eximius,
oxidase ) Chroococcus cf. yellowstonensis
Tr3 (47 °C) Corynebacteriaceae (Gram+ bacteria) Cyanobacteria: Synechococcus eximius,
Enterobacteriaceae (Gram bacteria, Lygnbia limnetica, Phormidium sp.
oxidase )
Tr5 (37 °C) Corynebacteriaceae (Gram+ bacteria) Cyanobacteria: Synechococcus eximius, Oscillatoria sp.
Enterobacteriaceae (Gram bacteria, Bacillariophyta: Navicula sp.
oxidase ) Chlorophyta: Dyctiosphaerium sp.,
Pseudomonadaceae (Gram bacteria, Scenedesmus acuminata
oxidase +)
Tr7 (27 °C) Sporidiobolaceae: Cyanobacteria: Nostoc comune, Anabaena circinalis,
Rhodotorula sp. Oscillatoria sp.
Bacillariophyta: Navicula radiosa, Pinnularia sp.,
Gyrosigma alternatum
Chlorophyta: Chlorella sp., Chlamidomonas sp.,
Pediastrum cf. biradiatum, Scenedesmus acuminata
Euglenaceae: Phacus sp.
4452 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

Enterobacteriaceae and Microalgae at 57 °C, Corynebacte- 4.2. Structural and EPR features
riaceae, Enterobacteriaceae, and Microalgae at 47 °C;
Corynebacteriaceae, Enterobacteriaceae, Pseudomonada- The travertine samples are characterized by a double
ceae, and Microalgae at 37 °C and Sporidiobolaceae (Rho- evidence of unexpected structural features: both XRPD
dotorula sp.) and Microalgae at 27 °C (Table 6). A relatively and EPR, in fact, exhibit anomalous values of the lattice
high biodiversity of Microalgae (Cyanobacteria, Bacilla- and the crystal field interaction, respectively.
riophyta, Chlorophyceae, Euglenacea) was observed in The observed modified lattice of Bullicame 3 travertine
the water samples at 27 °C, whereas at higher temperature cannot be simply explained in terms of its elemental chem-
Cyanobacteria and Enterobacteriaceae are dominating. Set- istry. In fact, the substitution of Mg and Sr for Ca in the
ting aside the Rhodotorula sp. pigmented yeast found at calcite lattice is expected to induce opposite volume effects
27 °C, no fungal diversity was identified (Table 6). and cannot give rise to the observed variation of the lattice
constants with a negative value for Da0 and a positive one
4. DISCUSSION for Dc0. On the other hand, no other minor/trace compo-
nent, apart from Mg and Sr, is abundant enough to play
4.1. Isotopic fractionation among gas, liquid, and solid phases a significant role in the production of lattice variations.
Bischoff et al. (1983), Paquette and Reeder (1990), and,
A relevant difference is observed between the experimen- more recently, Pokroy et al. (2006) point out that the c0/a0
tal isotopic d18O and d13C values in the system: the former ratio can be a useful parameter to indicate the presence of
values appear consistent with a balanced equilibrium tem- bio-mineralizing activity: bio-mineralized calcite is indeed
perature-dependent isotopic exchange between the gas always characterized by c0/a0 ratios higher than those re-
and the liquid phase (Scholz et al., 2009), whereas the latter ported as a function of the Mg content. Bischoff et al.
values mark a solid phase clearly far from the equilibrium (1983) suggest a c0/a0 ratio of 3.437, particularly close to
with the contact liquid phase (Skidmore et al., 2004). Thus, the value determined in the present work (3.439).
experimental data would apparently indicate that isotopic In the EPR spectrum of Mn(II), as in general for transi-
fractionation of C and O follows different pathways. tion metal ions, ZFS values are directly related to the chem-
The observed d13C values for travertine samples can be ical and structural environment of the paramagnetic center.
in principle attributed to either a kinetically controlled inor- In Mn(II), hyperfine and Zeeman anisotropies are actually
ganic deposition process or a bio-assisted precipitation of negligible compared to the crystal field. In calcite, the Ca
calcite. According to the first hypothesis, the geometrical site is a trigonally distorted octahedron, with D3d point
dimensions of the spring–channel system would not allow symmetry. The Mn(II) replacing Ca(II) gives rise only to
a complete T-dependent re-equilibration of the chemical minor local distortions (Cheng et al., 2001), inducing an ax-
species in the whole volume of water, at the given boundary ial zero-field splitting. The anisotropy of this interaction,
conditions (flux of the water, time scales of degassing and parameterized by the D term, increases by increasing the de-
precipitation processes, time scale of the diffusion processes gree of distortion of the polyhedron.
involving the considered species) (Skidmore et al., 2004; The observed D values, significantly larger than those re-
Scholz et al., 2009). According to the second hypothesis, ported in literature for pure calcite (e.g. Wildeman, 1970;
during bio-mineralization isotopic fractionation processes Angus et al., 1979) indicate a strong distortion of the geom-
caused by bio-mediated precipitation would produce 12C etry around the Mn(II) ion. The observed shrinking of the
enrichment in carbonates (e.g. Guo et al., 1996; Zhang unit cell evidenced by X-ray data was tentatively associated
et al., 2004). Several authors attributed out-of-equilibrium to the observed ZFS values. However, the structural
isotopic signatures of travertines from many sites to this changes of the Ca octahedra are too small to induce the ob-
process (Folk, 1992, 1993a,b, 1994, 1997; Pentecost, 1995; served relevant increase (+39%). For comparison, one has
Vasconcelos et al., 1995; Chafetz and Guidry, 1999; Kirk- to consider the D value observed for Mg-sites in dolomite
land et al., 1999) including some sites from the Viterbo (about 150 G; Shepherd and Graham, 1984) that is caused
Area (Chafetz and Lawrence, 1994; Allen et al., 2000). by a reduction of the cation–oxygen distance of 12%: here
The simulation model presented in this study calculates the distance decrease is about 1%. Thus, the ZFS values ob-
the evolution of the chemical species distribution in the sys- served in this work cannot be attributed either to Mn in a
tem in absence of biotic effects. The agreement between sim- distorted Ca site or to a Mn in dolomite-like sites. Accord-
ulated and experimental isotopic fractionation values at ing to Kretz (1982), the tendency of Mn(II) to be systemat-
both the gas/liquid and liquid/solid interfaces (Figs. 4a ically related either to Ca- or to Mg-sites in Mg-bearing
and 5) suggests that the simple diffusion-controlled kinetics calcites or dolomites reflects the conditions of rock forma-
of transport and distribution of the chemical species could tion. In the case under study, it is therefore conceivable that
justify a precipitation process of travertine with an anoma- the large D parameter has to be related to the proximity of
lous isotopic signature. However, the simultaneous occur- Mn(II) ions to highly distorted Mg containing sites. The ab-
rence of bio-assisted precipitation processes cannot be sence of any dependence on Mn(II) concentration (Table 4),
ruled out. In our opinion, no further reliable indications temperature and pH, suggests that the co-precipitation of
can be extracted from our analysis concerning isotopic sig- Mn(II) with calcite does not follow a random path; a given
nature, and discrimination between the two hypotheses is association between Mn and Mg is retained in the whole
not allowed, as already pointed out by Skidmore et al. series of travertine encrustations deposited from the source
(2004). down to the pool.
Biotic and inorganic control on travertine deposition 4453

The circumstantial evidence of a non-random clustering (Pokroy et al., 2006). According to inorganic mechanisms,
of Mn and Mg is reinforced by the trend observed in the Mg and Mn ions, less accepted in the Ca-carbonate bacte-
WD parameter, related to the spread of the ZFS interaction rial metabolism, would probably co-precipitate with
and to the lattice chemical/structural disorder. Indeed, in a CaCO3., but it is noteworthy to mention that Pokroy
random process this parameter should decrease with Mn(II) et al. (2006) do not exclude the fixing of Mg through bacte-
concentration, since the probability of finding Mn in non- rial activity. Finally, the intracellular bacterial activity
equivalent structural/chemical environments is a function would yield a constant precipitate masking the external
of the Mn content (e.g. Angus et al., 1979; Kralj et al., physico-chemical input. Indeed, Pokroy et al. (2006) stress
2004). In the Bullicame 3 samples, the WD behavior indi- that sulfur anomalies associated to the shrinking of the cal-
cates an increasing local disorder as temperature decreases, cite unit cell may be attributed to “biologic S”, thus linking
in spite of the sharp decrease of the Mn concentration along the two experimental observations as effects of a unique
the channel. This supports the idea of non-random process bio-mediated precipitation process. As previously noted,
acting in the precipitation, which in turn is a clue of a bio- also in the Bullicame 3 samples a sulfur anomaly is
assisted precipitation. occurring.

5. CONCLUSIONS
4.3. Model of the spring system and of travertine deposition
According to the results presented in this study, we can
On the basis of the complete set of experimental data, describe the Bullicame 3 system (hot spring + artificial
the Bullicame 3 system, constituted by the thermal spring, channel + pool) as constituted by two markedly different
the channel and the final pool, can be considered formally subsystems:
divided into two sections, each having peculiar composi-
tional and isotopic features.
– water/gas interface, where a fast exchange between the
Deep hot water, enriched in gas, is taken up to the sur-
solute and the dissolved gas species allows the system to
face of the spring, and subsequently flows down to the pool
achieve chemical and isotopic equilibrium as the water
at a constant rate and with decreasing temperature. Along
flows along the system, while temperature decreases;
the channel, partial H2O as well as excess CO2 evaporation
– solid/water interface, where fast travertine precipitation
occurs. As a consequence, pH increases whereas HCO3 de-
occurs; carbon isotopic ratios in the solid phase, suggest
creases. This process is tracked by water and gas chemistry,
non-equilibrium conditions, while chemical, structural
and can be described as a fast process. The isotope fraction-
and spectroscopic data point to a bio-assisted process.
ation follows the fast exchange at the water/atmosphere
This conclusion is also strongly supported by the identi-
interface, and the isotopic anomaly of the spring is progres-
fication of calcifying cyanobacteria throughout the
sively shifted towards values typical of meteoric water. The
channel.
HCO3–CO2 isotopic equilibrium rapidly attained at each
temperature can be conveniently modeled through isotopic
The fact that an active role played by bacteria is proba-
fractionation calculations. Simultaneously, precipitation of
bly operating in a high temperature spring system, as Bulli-
travertine occurs at the water/solid interface, where experi-
came 3, opens interesting perspectives for the study of more
mental data appear to be more ambiguous: the isotopic
active systems with similar characteristics, whose composi-
fractionation, the calcite structural features, and the spec-
tional and isotopic profiles could be properly analyzed in
troscopic data clearly indicate an apparent non-equilibrium
the framework of the theoretical modeling herewith pre-
process. However, even if the isotopic signatures of traver-
tine are assumed to depend on precipitation under a kinet- sented. One of the most relevant aspect of this characteriza-
tion, i.e. the anomalous ratio between the lattice constants,
ically-controlled regime of transport, the same cannot be
could be used as a proxy for microbiological activity in cal-
said regarding structural and spectroscopic evidence. This
fact deserves, in our opinion, further consideration. cite-bearing geomaterials.
The approach presented in this work combines isotopic,
ZFS indicates preferential clustering pathways for Mg
compositional, structural and EPR spectroscopic investiga-
and Mn that cannot be attributed to a pure stochastic pro-
tions. The results obtained in this study suggest that this ap-
cess. On the other hand, structural XRPD analysis of the
proach can be profitably applied to the study of other
calcite fraction clearly indicates a bio-mineralizing process
systems, as fossil travertine deposits, where the absence of
as a possible candidate to explain all the observed anoma-
the coexisting parent waters prevents the direct knowledge
lies: the lattice distortion can indeed be considered a strong
of the microbiological environment active during the depo-
indication for such explanation.
sition. On this basis, future applications on fossil travertine
Furthermore, the presence of the Sinechococcus cyano-
will be relevant in order to reconstruct the environmental
bacteria in three out of the four analyzed water samples
supports the idea that the travertine deposit, mantling the (and potentially paleoclimatic) conditions existing during
the deposition.
artificial channel of the Bullicame 3 thermal spring dis-
charge, likely exhibits some bio-assisted precipitation, as ACKNOWLEDGMENTS
these bacteria are known to participate in the formation
of biogenically-precipitated calcite (Lee et al., 2006). The manuscript draft benefited of the critical reading by E.
The structural deformation of the calcite lattice would Valsami-Jones (Imperial College, London), G. Lampronti (Univer-
be produced by protein activity of bio-precipitation sity of Bologna), F. Pineider and G.P. Bernardini (University of
4454 F. Di Benedetto et al. / Geochimica et Cosmochimica Acta 75 (2011) 4441–4455

Florence), to whom the authors wish to express their warmest grat- Duchi V., Minissale A. and Romani L. (1985) Studio geochimico su
itude. The authors acknowledge also Filippo and Claudio Di acque e gas dell’area geotermica lago di Vico-M. Cimini (Viterbo).
Benedetto, who provided the historical information about the site, Atti Soc. Toscana Sci. Nat.: Memorie A92, 237–254 (in Italian).
and P. Costagliola (University of Florence) for the useful discus- Epstein S. and Mayeda T. (1953) Variations of the 18O/16O ratio in
sions about the structural remnants of microbiological activity. natural waters. Geochim. Cosmochim. Acta 4, 213–224.
Miria Borgheresi is acknowledged for her help in the sampling Evans W. C., White L. D. and Rapp J. B. (1988) Geochemistry of
campaign. FDB benefited of a research grant funded by the Regio- some gases in hydrothermal fluids from the southern Juan de
nal Administration of Tuscany. Luca Pardi and Giordano Monteg- Fuca Ridge. J. Geophys. Res. 93(B12), 15305–15313.
rossi thank the Italian National Research Council (CNR) for Farmer J. D. (2000) Hydrothermal systems doorways to early
support. Authors wish to express their gratitude to the anonymous biosphere evolution. Geol. Soc. Am. Today 10(7), 1–9.
referees, for their stimulating criticisms, which largely improved the Folk R. L. (1992) Bacteria and nannobacteria revealed in
text. hardgrounds, calcite cements, native sulfur, sulfide materials,
and travertines. In Geological Society of America Annual
Meeting, Program Abstracts, p. 104 (abstr.).
REFERENCES Folk R. L. (1993a) Dolomite and dwarf bacteria (nannobacteria).
In Geological Society of America Annual Meeting, Program
Allen C. C., Albert F. G., Chafetz H. S., Combie J., Graham C. R., Abstracts, A-397 (abstr.).
Kieft T. L., Kivett S. J., McKay D. S., Steele A., Taunton A. E., Folk R. L. (1993b) SEM imaging of bacteria and nannobacteria in
Taylor M. R., Thomas-Keprta K. L. and Westall F. (2000) carbonate sediments and rocks. J. Sed. Petrol. 63, 990–999.
Microscopic physical biomarkers in carbonate hot springs: Folk R. L. (1994) Interaction between bacteria, nannobacteria, and
implication in the search for life on Mars. Icarus 147, 49–67. mineral precipitation in hot springs of central Italy. Geog. Phys.
Andrews J. E. (2006) Paleoclimatic records from stable isotopes in Quatern. 48, 233–246.
riverine tufas: synthesis and review. Earth Sci. Rev. 75, 85–104. Folk R. L. (1997) Nannobacteria: surely not figments, but what
Andrews J. E. and Riding R. (2001) Depositional facies and under heaven are they? Available from: <http://natural-
aqueous–solid geochemistry of travertine-depositing hot springs science.com/ns/articles/01-03/ns_folk.html>.
(Angel Terrace, Mammoth Hot Springs, Yellowstone National Fouke B. W. (2001) Depositional facies and aqueous–solid
Park, USA)—discussion. J. Sed. Res. 71, 496–497. geochemistry of the travertine-depositing hot springs (Angel
Angus J. A., Raynor B. and Robson M. (1979) Reliability of Terrace, Mammoth Hot Springs, Yellowstone National Park,
experimental partition coefficients in carbonate systems: evi- USA)—reply. J. Sed. Res. 70, 497–500.
dence for inhomogenous distribution of impurity cations. Fouke B. W., Farmer J. D., Des Marais D. J., Pratt L., Sturchio N.
Chem. Geol. 27, 181–205. C., Burns P. C. and Discipulo M. K. (2000) Depositional facies
Attanasio D. (1999) The use of electron spin resonance spectros- and aqueous–solid geochemistry of travertine-depositing hot
copy for determining the provenance of classical marbles. Appl. spring (Angel Terrace, Mammoth Hot Springs, Yellowstone
Magn. Reson. 16, 383–402. National Park, U.S.A.). J. Sediment. Res. 70, 565–585.
Bischoff W. D., Bishop F. C. and Mackenzie F. T. (1983) Frank M. J. W., Kuipers J. A. M. and Van Swaaij W. P. M. (1996)
Biogenically produced magnesian calcite inhomogeneities in Diffusion coefficients and viscosities of CO2 + H2O,
chemical and physical properties: comparison with synthetic CO2 + CH3OH, NH3 + H2O, and NH3 + CH3OH liquid mix-
phases. Am. Mineral. 68, 1183–1188. tures. J. Chem. Eng. Data 41, 297–302.
Chafetz H. S. and Folk R. L. (1984) Travertines depositional Garcia J. E. (2001) Density of Aqueous Solutions of CO2. Open
morphology and the bacterially constructed constituents. J. Report LBNL-49023. Available from: <http://escholar-
Sed. Petrol. 54, 289–316. ship.org/uc/item/6dn022hb>.
Chafetz H. S. and Lawrence J. R. (1994) Stable isotopic variability Giannini P. (1969) Centri etruschi e romani del Viterbese carta
within modern travertines. Geog. Phys. Quatern. 48, 257–273. archeologica della Tuscia. Quatrini ed. (Viterbo), p. 139 (in
Chafetz H. S. and Guidry S. A. (1999) Bacterial shrubs, crystal Italian).
shrubs, and ray-crystal shrubs: bacterial vs. abiotic precipita- Gonfiantini R., Panichi C. and Tongiorgi E. (1968) Isotopic
tion. Sediment. Geol. 126, 57–74. disequilibrium in travertine deposition. Earth Planet. Sci. Lett.
Chafetz H. S., Rush P. F. and Utech N. M. (1991) Microenviron- 5, 55–59.
mental controls on mineralogy and habit of CaCO3 precipi- Guo L., Andrews J., Riding R., Dennis P. and Dresser Q. (1996)
tates: an example from an active travertine system. Possible microbial effects on stable carbon isotopes in hot-
Sedimentology 38, 107–126. spring travertines. J. Sed. Res. 66, 468–473.
Cheng L., Sturchio N. C. and Bedzyk M. J. (2001) Impurity Hancock P. L., Chalmers R. M. L., Altunel E. and Cakir Z. (1999)
structure in a molecular ionic crystal atomic-scale X-ray study Travitonics: using travertines in active fault studies. J. Struct.
of CaCO3 Mn2+. Phys. Rev. B 63, 144104–144109. Geol. 21, 903–916.
Coleman M. L., Shepherd T. J., Durham J. J., Rouse J. E. and de Hoog G. S., Guarro J., Gene J. and Figueras M. J. (2000) Atlas
Moore G. R. (1982) Reduction of water with zinc for hydrogen of Clinical Fungi, second ed. Centralbureau voor Schimmelcul-
isotope analysis. Anal. Chem. 54, 993–995. tures, Utrecht, & Universitat Rovira I Virgili, Reus, p. 1126.
Criado J. M. and Trillo J. M. (1975) Effects of mechanical grinding Humphrey J. D. and Howell R. P. (1999) Effect of differential stress
on the texture and structure of calcium carbonate. J. Chem. on strontium partitioning in calcite. J. Sed. Res. 69, 208–215.
Soc.: Faraday Trans. I 71, 961–966. John D. M., Whitton B. A. and Brook A. J. (2002) The Freshwater
Di Benedetto F., Montegrossi G., Pardi L. A., Minissale A., Paladini Algal Flora of the British Isles. Cambridge University Press,
M. and Romanelli M. (2005) A multifrequency EPR approach to Cambridge, UK, pp. 327–408.
travertine characterisation. J. Magn. Reson. 177, 88–94. Kirkland B. L., Lynch F. L., Rahnis M. A., Folk R. L., Molineux
Di Benedetto F., Costagliola P., Benvenuti M., Lattanti P., I. J. and McLean R. J. C. (1999) Alternate origins for
Romanelli M. and Tanelli G. (2006) Arsenic incorporation in nannobacteria-like objects in calcite. Geology 27, 347–350.
natural calcite lattice: evidence from electron spin echo spec- Klein C. and Hurlbut C. S. (1993) Manual of Mineralogy. John
troscopy. Earth Planet. Sci. Lett. 246, 458–465. Wiley & Sons, New York, p. 681.
Biotic and inorganic control on travertine deposition 4455

Kralj D., Kontrec J., Brečevic L., Falini G. and Nöthig-Laslo V. Scholz D., Mühkinghaus C. and Mangini A. (2009) Modeling d13C
(2004) Effect of inorganic anions on the morphology and and d18O in the solution layer on stalagmite surfaces. Geochim.
structure of magnesium calcite. Chem. Eur. J. 10, 1647–1656. Cosmochim. Acta 73, 2592–2602.
Kretz R. (1982) A model for the distribution of trace elements Shepherd R. A. and Graham W. R. M. (1984) EPR of Mn2+ in
between calcite and dolomite. Geochim. Cosmochim. Acta 46, polycrystalline dolomite. J. Chem. Phys. 81(12), 6080–6084.
1979–1981. Skidmore M., Sharp M. and Tranter M. (2004) Kinetic isotopic
Lee B. D., Apel W. A. and Walton M. R. (2006) Calcium carbonate fractionation during carbonate dissolution in laboratory exper-
formation by Synechococcus sp. strain PCC 8806 and Synecho- iments: implications for detection of microbial CO2 signatures,
coccus sp. strain PCC 8807. Biores. Tech. 97(18), 2427–2434. using d13C DIC. Geochim. Cosmochim. Acta 68, 4309–4317.
Lutterotti L., Ceccato R., Dal Maschio R. and Pagani E. (1998) Spycher N. and Pruess K. (2005) CO2–H2O mixtures in the
Quantitative analysis of silicate glass in ceramic materials by the geological sequestration of CO2 center dot: II. Partitioning in
Rietveld method. Mater. Sci. Forum 278–281, 93–98. chloride brines at 12–100 °C and up to 600 bar. Geochim.
Mackenzie F. T., Bischoff W. D., Bishop F. C., Loijens M., Cosmochim. Acta 69(13), 3309–3320.
Schoonmaker J. and Wollast R. (1983) Magnesian calcites low- Sunagawa I., Takahashi Y. and Imai H. (2007) Strontium and
temperature occurrence, solubility and solid solution behav- aragonite–calcite precipitation. J. Miner. Petrol. Sci. 102, 174–
iour. Rev. Mineral. 11, 97–144. 181.
Markgraf S. A. and Reeder R. J. (1985) High-temperature Tamini A., Rinker B. and Sandall O. C. (1994) Diffusion
structure refinements of calcite and magnesite. Am. Mineral. coefficients for hydrogen sulfide, carbon dioxide, and nitrous
70, 590–600. oxide in water over the temperature range 293–368 K. J. Chem.
Minissale A. (2004) Origin, transport and discharge of CO2 in Eng. Data 39, 330–332.
central Italy. Earth Sci. Rev. 66, 89–141. Tassi F., Montegrossi G. and Vaselli O. (2004) Metodologie di
Minissale A., Kerrich D., Magro G., Murrell M. T., Paladini M., campionamento ed analisi in fase gassosa. Internal Report
Rihs S., Sturchio N., Tassi F. and Vaselli O. (2002) Geochem- CNR-IGG No. 1/2004, Florence, p. 17 (in Italian).
istry of quaternary travertines in the region north of Rome Tassi F., Vaselli O., Luchetti G., Montegrossi G. and Minissale A.
(Italy): structural, hydrologic and paleoclimatic implications. (2008) Metodo per la determinazione dei gas disciolti in acque
Earth Planet. Sci. Lett. 203, 709–728. naturali. Internal Report, CNR-IGG No. 2/2008, Florence,
Minissale A., Paladini M. and Sturchio N. C. (2005) Travertine of Italy, p. 11 (in Italian).
central Italy relation between genesis, tectonics and paleocli- Thorstenson D. C. and Parkhurst D. L. (2002) Calculation of
mate. In Proceedings of the First Congress on Travertine, 21–25 Individual Isotope Equilibrium Constants for Implementation
September, 2005, Pamukkale, Turkey, pp. 106–115. in Geochemical Models. U.S. Geological Survey Water-
Montegrossi G., Tassi F., Vaselli O., Buccianti A. and Garofalo K. Resources Investigations, Open File Report 02-4172, p. 129.
(2001) Sulfur species in volcanic gases. Anal. Chem. 73, 3709– Vasconcelos C., McKenzie J. A., Bernasconi S., Grujic D. and Tien A.
3715. J. (1995) Microbial mediation as a possible mechanism for natural
Montegrossi G., Di Benedetto F., Minissale A., Paladini M., Pardi L. dolomite formation at low temperatures. Nature 377, 220–222.
A., Romanelli M. and Romei F. (2006) Determination and Vaselli O., Tassi F., Montegrossi G., Capaccioni B. and Giannini
significance of the Mn(II) Zero-Field Splitting (ZFS) interaction in L. (2006) Sampling and analysis of volcanic gases. Acta
the geochemistry of travertines. Appl. Geochem. 21(5), 820–825. Vulcanol. 18, 65–76.
Palmer B. J. (1994) Calculation of thermal-diffusion coefficients Vassilikou-Dova A. B. (1993) Characterisation of the crystal
from plane-wave fluctuations in the heat energy density. Phys. quality of calcites by electron paramagnetic resonance. Phys.
Rev. E 49(3), 2049–2057. Status Solidi B178, 465–476.
Paquette J. and Reeder R. J. (1990) Single crystal X-ray structure Wang Y. and Merino E. (1992) Dynamic model of oscillatory trace
refinements of two biogenic magnesian calcite crystals. Am. element zoning in calcite: inhibition, double layer, and self-
Mineral. 75, 1151–1158. organization. Geochim. Cosmochim. Acta 56, 587–596.
Parkhurst D. L. and Appelo C. A. J. (1999) User’s Guide to Wildeman T. R. (1970) The distribution of Mn2+ in some
PHREEQC (Version 2)—A Computer Program for Speciation, carbonates by electron paramagnetic resonance. Chem. Geol.
Batch-Reaction, One-Dimensional Transport, and Inverse 5, 167–177.
Geochemical Calculations. U.S. Geological Survey Water- Xu T. and Pruess K. (2001) Modeling multiphase nonisothermal
Resources Investigations Report 99-4259, 312p. fluid flow and reactive geochemical transport in variably
Pentecost A. (1995) Geochemistry of carbon dioxide in six saturated fractured rocks: 1. Methodology. Am. J. Sci. 301,
travertine-depositing waters of Italy. J. Hydrol. 167, 263–278. 16–33.
Pentecost A. (2005) Travertine. Springer Verlag, Berlin, p. 445. Xu T., Sonnenthal E. L., Spycher N. and Pruess K. (2006)
Pingitore, Jr., N. E., Lytle F. W., Davies B. M., Eastman M. P., TOUGHREACT: a simulation program for non-isothermal
Eller P. G. and Larson E. M. (1992) Mode of incorporation of multiphase reactive geochemical transport in variably saturated
Sr2+ in calcite: determination by X-ray absorption spectros- geologic media. Comput. Geosci. 32, 145–165.
copy. Geochim. Cosmochim. Acta 56(4), 1531–1538. Zhang C. L., Fouke B. W., Bonheyo G. T., Peacock A. D., White
Pokroy B., Fitch A. N., Marin F., Kapon M., Adir N. and D. C., Huang Y. and Romanek C. S. (2004) Lipid biomarkers
Zolotoyabko E. (2006) Anisotropic lattice distortions in and carbon-isotopes of modern travertine deposits (Yellow-
biogenic calcite induced by intra-crystalline organic molecules. stone National Park, USA): implications for biogeochemical
J. Struct. Biol. 155(1), 96–103. dynamics in hot-spring systems. Geochim. Cosmochim. Acta
Quinn P. J., Carter M. E., Markey B. K. and Carter G. R. (2000) 68(15), 3157–3169.
Clinical Veterinary Microbiology. Mosby, Edinburgh, p. 720. Zolotoyabko E., Caspi E. N., Fieramosca J. S., Von Dreele R. B.,
Reeder R. J. (1983) Crystal chemistry of the rhombohedral Marin F., Mor G., Addadi L., Weiner S. and Politi Y. (2010)
carbonates. Rev. Mineral. 11, 1–47. Differences between bond lengths in biogenic and geological
Salata G. G., Roelke L. A. and Cifuentes L. A. (2000) A rapid and calcite. Cryst. Growth Des. 10(3), 1207–1214.
precise method for measuring stable carbon isotope ratios of
dissolved inorganic carbon. Mar. Chem. 69, 153–161.
Associate editor: Roy A. Wogelius

You might also like