Chemical Geology 175 Ž2001. 319–341
www.elsevier.comrlocaterchemgeo
Causes of colour and fluorescence in speleothems
Philip van Beynen a,) , Rick Bourbonniere b, Derek Ford a , Henry Schwarcz a
a
b
School of Geography and Geology, McMaster UniÕersity, Hamilton, Ontario, Canada L8S 4K1
National Water Research Institute, EnÕironment Canada, Burlington, Ontario, Canada L7R 4A6
Received 8 June 1999; accepted 26 May 2000
Abstract
Speleothems fluoresce, when illuminated with UV light, between 410 and 460 nm. In this study, we attempted to
determine the nature of the fluorophores, thought to be either trace elements or organic matter trapped in the calcite.
Fluorescence of solid speleothems and organic species extracted from the calcite were measured to quantify their
contribution to the observed fluorescence of the speleothems. All speleothems and extracts gave similar spectra with broad
emission maxima centred around 410–430 nm, and two excitation maxima at approximately 255 and 330 nm. The organic
compounds were partly characterized using fulvic acid ŽFA. –humic acid ŽHA. separation and molecular size fractionation.
Trace elements, determined by neutron activation analysis, do not appear to be responsible for the observed spectra. Organic
matter, particularly FAs, were found to be the dominant fluorophore in the calcite. Of the FA, the dominant fractions were
the hydrophilics. Darker speleothems, although having higher concentrations of FA and HA than light speleothems, had
lower emission intensities, due to self-absorption. Average particulate organic matter ŽPOM., FA, HA, and total organic
matter ŽTOM. concentrations for the dark speleothems were twice that of their light counterparts. q 2001 Elsevier Science
B.V. All rights reserved.
Keywords: Caves; Fluorescence; Fulvic acids; Humic acids; Speleothems
1. Introduction
Speleothem is a generic term given to stalagmites,
stalactites, and flowstones consisting of calcite or
other mineral cave deposits. They display two important types of optical properties. To begin, when
viewed either in reflected or transmitted visible light,
)
Corresponding author. Fax: q1-905-546-0463.
E-mail addresses: [email protected] ŽP. van Beynen.,
[email protected] ŽR. Bourbonniere., [email protected]
ŽD. Ford., [email protected] ŽH. Schwarcz..
speleothems display varying degrees of colour hue
and intensity. Secondly, when illuminated by ultraviolet light, they emit fluorescent light in the visible
range. Both of these effects are presumably connected to the presence of minor constituents in these
calcitic deposits since chemically pure calcite is neither coloured nor fluorescent. Some of these properties have been attributed to the presence of trace
elements in the calcite, but since the early work of
Gilson and MacCarthney Ž1954., it has been suggested that this fluorescence is caused by organic
substances, a supposition supported by Gascoyne
Ž1977., Lauritzen et al. Ž1986., and White and Brennan Ž1989.. Lauritzen et al. Ž1986. in particular
0009-2541r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 5 4 1 Ž 0 0 . 0 0 3 4 3 - 0
320
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
demonstrated that much of the colour must be due to
the presence of organic substances Žhumic compounds. dispersed in the calcite. Similarly, Shopov et
al. Ž1994. argued that the fluorescent properties of
speleothems were the result of the presence of organic compounds trapped in speleothems. Ramseyer
et al. Ž1997. found that low molecular weight fulvic
acids ŽFAs. were the dominant constituent of organic
matter derived from speleothems, a result confirmed
by White Ž1997.. Finally, Baker et al. Ž1998. showed
a strong relationship between the luminescence properties of speleothems and the extent of soil humification and rainfall above a cave.
Calcitic cave deposits contain paleotemperature
records as derived from oxygen isotopes ŽSchwarcz,
1986; Gascoyne, 1992. and paleovegetation records
as seen in the carbon isotope records ŽDorale et al.,
1992.. These records can be dated with U-series
analysis to a precision of ; 1% up to 350 ka ŽLi et
al., 1989.. Broecker et al. Ž1960. using 14 C dates
found laminae in parts of speleothems that were
annual in resolution. Such laminae are differentiated
through changing colour of the calcite and are composed of a couplet of light and dark calcite. Their
presence has also been revealed using fluorescence
microscopy ŽBaker et al., 1993; Genty and Quinif,
1996; Shopov et al., 1989, 1994.. Production of
annual laminae within calcite is to be expected:
organic matter in litter and soil decomposes due to
activity of microorganisms and percolation waters
then transport the products through the soil and rock
to the cave. Throughout the passage, organic matter
may be altered by further decomposition or removed
by the filtration processes occurring in the mineral
medium and rock. However, some organic matter
remains in the solution. A complicating factor is that
some ancient organic substances may also be obtained from the dissolution of the limestone bedrock.
Upon reaching the cave, degassing of CO 2 from
aqueous solution may lead to calcium carbonate
saturation and calcite or aragonite precipitation. Calcite has been shown to be an efficient absorber of
organic matter, especially of lipid material and amino
acid substances ŽSuess, 1970; Carter, 1978.. Mitterer
Ž1968. found that organic matter may even induce
CaCO 3 precipitation by concentrating calcium ions.
The principal subjects of this paper are the precise
sources of the fluorescence in speleothems and the
relationships between fluorescence and colour hue
and density. These should explain the major chemical differences between light and dark calcite. To
explain the source of fluorescence in speleothems, it
must be recognized that organic substances are not
the only natural entities, which might fluoresce in
speleothems. Trace and rare earth elements such as
samarium, dysprosium, europium, terbium and manganese have been shown to produce fluorescence
when coprecipitated with calcite; in particular, europium and dysprosium fluoresce at the same wavelengths as the organic acids ŽMason and Mariano,
1990.. The calcium ion can be replaced in the lattice
by these smaller divalent and trivalent fluorescing
cations through substitution ŽTerakado and Masuda,
1988..
Organic substances in soil pore waters may be
divided into three components that are operationally
defined by their solubility in acid–base ŽClapp and
Hayes, 1999.. FAs are soluble in all conditions of
acid–base; humic acids ŽHAs. precipitate when the
solution becomes acidic Žas pH approaches 2.; humin
is insoluble in any solution, although it may be
mechanically flushed from soils. The residence times
of these compounds in soils may range from decades
to hundreds of years for FAs and thousands of years
for HAs ŽSchlesinger, 1977.. The molecular weights
of FAs tend to be smaller than HAs ŽSuffett and
MacCarthy, 1987.. Further differentiation of the FAs
is possible by partitioning them into hydrophilic
acids ŽHPIA. and hydrophobic acids ŽHPOA., bases,
and neutrals ŽLeenheer, 1985.. The structural complexity of FAs has hindered their further characterization ŽErtel, 1988.; however, major functional
groups identified thus far include carboxyls, phenolic
hydroxyls, and alcohols ŽVisser, 1983.. NMR studies
have revealed ketoneraldehyde, carboxyl, aliphatic
and aromatic hydroxyl groups. Spectroscopic studies
by Wilson et al. Ž1988. confirmed these findings.
The typical molecular weight of FAs is 800–1000
Da. HAs are found to range greatly, from 2000 to
300,000 Da ŽSuffet and MacCarthy, 1987.. Their
differentiation can be undertaken by gel permeation
chromatography or ultrafiltration ŽWershaw and
Aiken, 1985.. Much work has been done on the
fluorescence of these substances ŽMiano et al., 1988;
Senesi, 1990; Senesi et al., 1991; Simpson et al.,
1997a.. FA wavelengths of fluorescence emission
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
321
are usually microscopic in scale. To investigate their
differences, larger samples that exhibit more uniform
colour are required. We selected a set of 12
speleothems from a collection at McMaster University, ranging in colour from translucent white to
opaque to nearly black. Their fluorescence was measured in the solid calcite and in solution. Solutions
were fractionated into the FAs, HAs, and particulate
organic matter ŽPOM.. The hydrophilic and hydrophobic components of FAs were then separated
for measurement of their individual fluorescence and
dissolved organic carbon ŽDOC.. Sub-samples of the
solid calcite were subjected to neutron activation
analysis to determine whether there could be any
significant trace element or rare earth element contributions to the fluorescence spectra. These procedures
are summarised in a flowchart, Fig. 1.
Fig. 1. Flowchart of analysis undertaken on speleothems: ŽA.
dissolved speleothems, and ŽB. solid speleothems.
3. Provenance of the sample speleothems
maxima and excitation peaks vary between 410–520
and 320–470 nm, respectively, depending on their
environment ŽSenesi, 1990; Senesi et al., 1991.. HAs
have longer excitation and emission wavelengths,
with respective peak intensities at 480 and 540 nm
ŽHayase and Tsubota, 1985..
2. Research design and procedures
Where alternating light and dark calcite bands,
denoting annual laminae, occur in speleothems they
In addition to colour hue and density, an important criterion for selecting the speleothems was the
current variety of climate and natural vegetation
found at the site of collection. A considerable natural
range was obtained. The samples were selected from
the following locations.
Crowsnest Pass Ž49830X N, 114830X W. and Bow
Valley ŽRatsnest Cave, 51845X N, 115830X W. in the
Rocky Mountains of Alberta, British Columbia. The
settings range from boreal forest to alpine shrub
tundra. Winters are cold and there is a strong spring
Table 1
Present climate conditions at sampling locations
Site
Mean
annual
temperature
Ž8C.
Mean
daily Jan.
temperature
Ž8C.
Mean
daily July
temperature
Ž8C.
Mean
total
precipitation
Žmm.
Mean
cave
temperature
Ž8C.
Crowsnest Pass
Cold Water
Ratsnest
Jewel
McFail’s
Kimberley
Cayman Brac
OFD
3
8.5
8.4
6.6
7.5
24.1
25.5
8.3
y10
y7.5
y6.5
y4.8
y4.7
30.4
22.5
2.5
15
23
22.3
18.1
19.1
17.8
28.5
14
1000
820
263
525
960.4
284
1025
2200
1.8
8.5
8.4
9.4
7.5
24.1
25.5
8.3
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
322
thaw. Samples — CNP, CNPLB, and CNPDB Žlight
and dark layers from the same speleothem., RNCP
and RNCF2.
Jewel CaÕe, Black Hills, South Dakota Ž43844X N,
103855X W.. The setting is sub-humid today, at the
lower limit of ponderosa forest. There is a wellmarked spring thaw. Sample — JC11.
Cold Water CaÕe, Iowa Ž4283X N, 90839X W. is
beneath a deciduous forest but close to the tall
grass–prairie transition. There is a spring thaw. Sample — CW4.
McFail’s CaÕe, New York Ž38822X N, 86820X W. is
located in the Carolinian forest zone with milder
winters than the previous sites but with spring and
earlier snap thaw events. Samples — MF1 and MF2.
Ogof Ffynon Ddu, Wales Ž528N, 3830X E. is in a
temperate maritime setting beneath a deciduous Žash,
oak. forest. Sample — OFD.
Rebecca’s CaÕe, Cayman Brac Ž19844X N,
79848X W. has a tropical maritime climate with a
summer rainfall bias. There is scrub vegetation on
phytokarst above the cave. Sample — RCB.
Kimberley Ranges, Western Australia Ž268S,
1208E. represents the warm arid limits for speleothem
deposition today, with desert–savannah vegetation
and sparse summer rains. Sample — WAH.
Further details of the locations and colours of the
sample speleothems are given in Tables 1 and 2. The
speleothems mentioned in Table 2 are arranged according to colour gradation, lightest to darkest. Six
Table 2
Colour of the solid speleothems
Sample
Light WAH
MF1
CNPLB
RNCF2
CW4
RNCP
Dark MF2
OFD
JC11
RCB
CNPDB
CNP
Colour
Description
ŽMunsell Soil Chart.
N 8r0
5Y 7r1
10YR 7r1
5Y 8r1
5Y 8r2.5
5Y 8r2.5
10YR 6r1
5YR 4r8
5YR 5r3
5R 4r4
7.5R 3r2
7.5R 1.7r1
Grayish white
Light gray
Light gray
Light gray
Light gray_pale Yellow
Light gray_pale yellow
Brownish gray
Reddish gray
Dull reddish brown
Dull reddish brown
Dark reddish brown
Reddish black
light and six dark coloured speleothems were selected.
4. Analytical methods
4.1. Isolation of humic substances
The traditional definitions of soil humic substances are based simply on their acid and base
solubility. Thus, soil organic matter, which is insoluble in both acid and base, is called humin. HAs are
soluble in base but insoluble in acid, while FAs are
soluble in both acid and base ŽStevenson, 1985;
Swift, 1985.. Decalcification Žwith HCl. is commonly carried out on soil samples prior to base
extraction of humic substances using a variety of
alkaline extractants ŽStevenson, 1982.. The definition for separating HAs from the base extract is
usually pH s 1, but pH 1.5 and 2 have been used
ŽSwift, 1985; Richmond and Bourbonniere, 1987;
Bourbonniere, 1989.. In this study, speleothems are
treated as soils with unusually high calcite contents,
adhering to the traditional operational definitions of
soil humic substances.
Twenty grams of powdered speleothem were
added to 500 ml of Milli-Q w water. Then, sufficient
4 N HCl was added to dissolve the calcite, releasing
the humic substances. The pH of the resulting solution was adjusted to approximately 8 with 4 N
NaOH so that the HAs would not precipitate. The
solution was then split in two equal portions. Both
portions were filtered using preweighed Whatman
GFrF glass fibre filters Ž0.7 mm. to remove the
POM. To examine both the fluorescence and the
concentration of the POM, one filter was flushed
with Milli-Q w water to remove most of its POM and
the other filter was dried in an oven and weighed.
After POM removal, both filtrates were recombined
for FA–HA separation.
POM should not be confused with humin, the
other fraction commonly produced when fractionating soil organic matter. For the purposes of this
paper, POM is operationally defined as any matter
above 0.7 mm and could contain a variety of material
such large organic matter Žcolloids. and clays which
are in suspension but not necessarily in solution. The
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
dissolution of the speleothem calcite requires highly
acidic conditions, a situation quite different from the
neutral–base conditions of the cave water in which
the POM originally reached the speleothem. The
acidic solution could promote coagulation of certain
organic matter to create compounds that did not exist
in the speleothem. Such a situation is unavoidable as
the calcite must be dissolved to release organic
substances.
HAs were precipitated from the recombined filtrate by adjusting the pH to 2. The solution was left
to stand for 24 h and was then filtered with Whatman GFrF filters. The filtrate contained organics
now operationally defined as FAs. Filters were then
rinsed with 0.1 N NaOH to resolubilize the HAs. To
collect any extra FAs that may have been trapped
with HAs, the HA solution was acidified to pH of 2
and filtered for a second time. This filtrate was then
added to the earlier FA solution. The HAs were
washed from the second filter with 0.1 N NaOH. All
fractions were refrigerated to slow any microbial
activity. Ten milliliters of the FA–HA solutions was
removed for fluorescence spectroscopy and DOC
analysis. The remaining FArHA solutions were then
divided; the FAs portion from the three light and
three dark speleothem solutions was used for further
FA fractionation; and the FA–HA solutions of the
remaining six speleothems was recombined and used
for size fractionation. The final volume of the solutions for FAs was approximately 800 and 150 ml for
the HAs.
4.2. FA fractionation
The acid soluble fraction of the organic matter
isolated from the speleothems is designated here as
the FA fraction ŽpH s 2.. Note that similar to the
traditional soil definitions, FAs have undergone no
purification other than removal of HAs. Further characterization of the FA fraction according to its sorptive and acid–base character was done using macroporous resin techniques. The primary guide to the
procedure used here is the paper by Aiken et al.
Ž1992. which used XAD-8 resin to isolate hydrophobic components of DOM and to define the hydrophobic–hydrophilic break, while XAD-4 resin was used
to isolate the hydrophilic components. Prior to the
application of the resin procedure, FA fractions re-
323
sulting from decalcification of speleothems were diluted with E-Pure w water to reduce their salt content
to - 0.5 M, thus reducing their density and eliminating problems observed with floatation of resin beads.
Distribution coefficients for dissolved organic acids
on XAD resins were shown to be unaffected by salt
content up to 0.5 M ŽPietrzyk and Chu, 1977. and
these resins have been successfully used to isolate
dissolved organics from seawater ŽAiken, 1995..
The FA fraction at pH s 2 is applied first to a
column of XAD-8 resin using 36 ml of resinrl of
sample. This amount of resin is about 20% more
than that recommended by Malcolm Ž1991. for
preparative isolations of dissolved aquatic hydrophobic organic matter. Thus, breakthrough of hydrophobic organic components with a kX ) 50 should not
have occurred under these conditions according to
stringent theoretical considerations ŽMalcolm, 1991..
Adsorption to XAD-8 under these conditions defines
organic solutes as hydrophobic in character, while
passing through XAD-8 defines hydrophilic organic
solutes ŽLeenheer, 1981; Malcolm, 1991; Aiken et
al., 1992.. HPOA are eluted from the XAD-8 column with 0.1 N NaOH ŽLeenheer, 1981; Aiken et
al., 1992. and those hydrophobic components from
the sample that are not eluted with base are termed
hydrophobic neutrals ŽHPON. by Leenheer Ž1981..
The effluent from the XAD-8 column contains
hydrophilic components and is applied to a column
of the same size packed with XAD-4 resin. The
hydrophilic components, which adsorb to this resin
and are subsequently eluted with 0.1 N NaOH are
termed HPIA by Aiken et al. Ž1992.. Not all of the
hydrophilic components, which adsorb to XAD-4 are
eluted by the base. Those components, which are
more strongly held to this resin are thought to bond
by p – p interactions with the aromatic matrix of
XAD-4 ŽMalcolm and MacCarthy, 1992.. Here, the
bound Žnot eluted with 0.1 N NaOH. hydrophilic
components are called AXAD-4 AcidsB ŽX4AC., a
term used by Malcolm and MacCarthy Ž1992. to
mean all of the acids which adsorbed to the XAD-4
resin. Those components, which pass through the
XAD-4 resin without binding, are primarily hydrophilic neutrals ŽHPIN. ŽMalcolm and MacCarthy,
1992. but could contain acidic or basic components
which were not adsorbed by either column under the
stated conditions. Nevertheless, the term HPIN are
324
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
used here to designate this residual, non-adsorbable,
fraction.
In summary, all of the organic matter contained in
the original speleothems is accounted for in the
seven operationally defined fractions. The acid or
base insoluble organic matter is POM, the base
soluble and acid insoluble dissolved organic matter
is HA and the acid and base soluble dissolved organic matter is FA. FA can also be fractionated into
five sub-fractions: HPOA, HPON, HPIA, X4AC and
HPIN. The fluorescence results for the POM, HA,
and FA for all the speleothems were compared and
the FA sub-fractions were used to indicate the character of the organic components, which percolated
through the overlying soil with water and incorporated into the speleothems.
4.3. DOC
DOC was determined using a Dohrmann DC-190
Carbon Analyzer with platinum on alumina catalyst
at 9008C. Standards and blanks were run daily to
determine the system blank which was then used to
correct all sample results. Inorganic carbon was first
removed from the samples by the carbon analyzer
through acidification with 20% H 3 PO4 and by purging with carrier gas Žair.. All dissolved fractions,
namely, the bulk FA, bulk HA, HPIA and HPOA,
basic and neutral FA, were determined this way.
4.4. Calculation of concentration of organic matter
fractions
DOC analysis only measured the organic carbon
content of the solution, but carbon accounted for
; 51% of the total in soil FAs ŽSFAs. and ; 57%
in soil HAs ŽSHAs. ŽSuffett and MacCarthy, 1987;
Simpson et al., 1997b.. The origins of the humic
substances in the speleothems must be from soils,
hence, only soil examples were included in the correction. To provide a more accurate estimate of total
organic matter ŽTOM. in the FAs and HAs, these
proportions were then used to calculate TOM from
the measured DOC values. The TOM now represents
the total amount of organic carbon in 20 g of calcite.
Finally, all these values Žin mgr20 g calcite. were
divided by 20 to convert them to total weight per-
centage of organic matter in the calcite for each
fraction. Steelink Ž1985. reported slightly different
ranges for %C of soil humic substances ŽFAs: 40–
51%; HAs: 54–59%.. While our HA corrections are
in an appropriate range, our FA corrections could
slightly overestimate the TOM of the FAs, but only
by a few percent which is within the measurement
error of the DOC analysis. Therefore, they are within
"5% of the actual value.
POM was measured on the dried preweighed filter
paper and the amount doubled because only half the
total amount was dried on the filter. The POM
concentrations for the fluorescence spectroscopy are
reported in parts per million units because they
represent the concentration of POM in the immersion
oil solution. Half of the total POM was resuspended
and homogenized in 10 ml of the immersion oil
Ždiscussed below. from which parts per million values could then be calculated.
4.5. Preparation of solid speleothems for absorbance
and fluorescence
The solid speleothems were cut along their growth
axes and finely polished to a thickness of 2 mm "
10% with silicon carbide Žgrain size ; 25–30 mm.
to produce smooth surfaces which prevented scattering of the fluorescent excitation beam. Each polished
section was taken from the same location in a sample
as the 20 g of calcite used for the fractionation work.
The strength of the calcite was such that no backing
or reinforcement was needed.
4.6. Absorbance spectroscopy
Absorbance spectra of the solid speleothems were
measured with a Perkin-Elmer UVrVisible Lambda
6 spectrophotometer over the range 200–600 nm to
characterize the concentration and nature of organic
substances in the speleothems. Pure calcite is known
to have negligible absorption in this region ŽMachel
et al., 1991..
4.7. Fluorescence spectroscopy
A Perkin-Elmer LS-5 fluorescence spectrophotometer recorded the fluorescence spectrum for
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
each solid calcite sample and solution. Slit-widths of
5 nm were used for the monochromators, with a
default scan speed of 60 nmrmin. Milli-Q w water
ŽDOC ; 0.5 ppm. and the empty cuvette were
analyzed each day to test for stability and cleanliness
of the apparatus; fluorescence intensity of all tests
was within the machine background of 0.1–0.3 units.
To allow semi-quantitative comparison of fluorescence intensities between solute samples, sub-samples of the various fractionated, and unfractionated
dissolved speleothems, Rhodamine WT dye Ž1 ml of
1 ppm standard. was added to each sample after
adjusting the pH to 10. All fluorescence spectra were
measured at pH 10 as this provides the best discrimination between organic species, particularly for the
FAs from different sources ŽMobed et al., 1996.. The
maximum intensity of the Rhodamine WT standard
also occurs at pH 10. Rhodamine WT was selected
because it is a fairly stable dye and its fluorescence
did not interfere with the fluorescence of the organic
substances ŽSmart and Laidlaw, 1977.. A blank solution with the dye was used as a standard for fluorescence intensity and all the spiked sample solution
intensities normalized to this standard ŽGoldberg and
Weiner, 1991.. This procedure compensates for machine drift and intrasample irregularities. Both excitation and emission spectra were measured for all
fractions, with numerous scans being made to find
the wavelengths of peak excitation and emission.
4.8. Size fractionation
The remaining FA–HA speleothem solutions not
used as the FA fractionation were recombined and
separated according to their molecular sizes by ultrafiltration. The initial size fractionation that had already been made with the POM being removed,
hence all the remaining fractions were - 0.7 mm.
The addition of the concentrations of the FA and HA
fractions provides the Total DOC ŽTDOC. concentrations for each speleothem. An aliquot of 10 ml
was removed from the recombined solution for each
100 and 10 kDa centrifuge tube: precleaned MSI
centrifuge tubes were used. The tubes were centrifuged until all the solution had passed through the
filters. The resulting solutions were then measured
for DOC to determine each fractions contribution to
325
the TDOC. The - 1 kDa fraction could not be
measured using the centrifuge method and therefore
required dialysis using Spectrum Spectra Por 6 dialysis bags. An aliquot of 30 ml of each recombined
solution was sealed within the bag and placed in
Milli-Q w filtered water. Stirring of the solution facilitated transfer of the - 1 kDa fraction over a 24-h
period. The DOC of the Milli-Q w water was measured to ascertain the levels of the - 1 kDa fraction
and was corrected for dilution. The subtraction of
each progressively smaller fraction from its larger
counterpart provided the DOC concentration for the
following size divisions: 0.7 mm–100 kDa, 100–10,
10–1, and - 1 kDa. The six speleothems used in
this particular analysis covered the colour range from
light to dark: Ratsnest Cave ŽRNCP, RNCF2.,
Crowsnest Pass ŽCNP., McFail’s Cave ŽMF2., Kimberley Ranges ŽWAH., and Ogof Ffynon Ddu ŽOFD..
4.9. POM fluorescence analysis
Because the POM began to settle out in water,
Cargille Immersion Oil ŽType FF. was used because
its viscosity Ž170 centistokes wcStx. kept the POM in
solution and the manufacturer claimed it had virtually no fluorescence. The lowest intensity of the oil
was at an excitation wavelength of 350 nm and the
emission peak wavelength was well below 400 nm
and did not fluoresce in the region where the solid
speleothems fluoresced. The analytical procedure was
to evaporate the water from the POM, 50% of which
was resuspended in the immersion oil and spiked
with the dye in the same manner used for the solutions. Fluorescence intensities of the spiked samples
were normalized to the fluorescence standard as
described above. The solutions were stirred between
each fluorescence run to prevent settling of the larger
POM particles.
4.10. Neutron actiÕation analysis
Trace element concentrations in the speleothems
were determined by instrumental neutron activation
analysis ŽINAA. in the McMaster University Nuclear
Reactor, using a Ge ŽLi. detector Žde Soete et al.,
1972; McKlveen, 1981.. The following trace elements were measured: As, B, Ba, Br, Ca, Cl, Co, Cr,
326
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Dy, Fe, Ga, I, K, La, Mg, Mn, Mo, Na, Nd, S, Sb,
Sc, Sm, Ti, Th, U, V, W and Zn. The speleothems
were sampled at the same location where the fluorescence was measured. These slices were powdered to
homogenize the calcite for the INAA.
5. Results and discussion
5.1. Causes of fluorescence in speleothems
The first objective is to ascertain whether humic
substances ŽFAs, HAs, and POM. or trace elements
are the dominant fluorophore in speleothems. To
achieve this aim, fluorescence of the speleothems
Žcontaining both humic substances and trace elements. and humic substances extracted from them
are compared to humic substance standards and trace
element fluorescence spectra.
5.1.1. Standards for organic compounds
To provide a comparison for the operationally
defined humic and FAs obtained from the
speleothems, humic and FAs were extracted from a
red oak–white ash forest soil close to the shores of
Cootes Paradise, Hamilton, Ontario. Other international standards are available but most are derived
from aqueous or peat environments which are not the
source of the speleothem organic matter. Due to the
varied locations of the speleothems from around the
world, no particular soil standard would be more
applicable than any other. Therefore, we decided to
extract our own humic and FA standards from a
forest soil as forests are the most common environment above the samples selected for this study. The
extraction procedure followed here is that outlined
by Stevenson Ž1982..
Fig. 2a,b shows contour plots of the fluorescence
of two humic substance standards, Cootes Paradise
SFAs and SHAs, in solution at pH 6.5 and at a
concentration of 10 ppm Ždetermined using a
Dohrmann DC-190 Carbon Analyzer.. Both standards lack any sharp peaks in either excitation or
emission wavelengths. SFAs display two peak excitation wavelengths, at approximately 250–260 nm
and 330–340 nm, and one peak emission wavelength
at 430 nm. SHAs show a similar but less clear
pattern with a peak excitation centred on 250–260
nm and a 430 nm emission peak. The smaller peak
centred on 280 nm excitation and 260 nm emission
is the Raman peak of water. This is only apparent in
the HAs because of their lower fluorescence intensities, which are approximately half that of the FAs;
although at the same concentration as SFAs, their
HA counterparts fluoresce less strongly. Senesi
Ž1990. showed that with increased molecular weight,
broadening occurs in the emission peak and there is
a decrease in fluorescence efficiency, both of which
are attributed to greater proximity of aromatic
fluorophores to each other and deactivation of excited states by internal quenching.
5.1.2. Fluorescence of speleothems
With this information about the fluorescence of
standard humic substances, the fluorescence of the
speleothems can now be better understood. Fig. 2c,d
shows two examples of the fluorescence fingerprints
of solid calcite, CW4, a light-coloured sample from
Cold Water Cave and JC11, a dark sample from
Jewel Cave. CW4 fluoresces more strongly than
JC11 and the two spectra differ quite markedly. JC11
has both longer excitation and emission peak wavelengths. Comparison of the fluorescence of the
speleothems with the fulvic and HA standards suggests that there are some similarities between SFAs,
SHAs, and CW4. All have centres of peak emission
at 430 nm, shown by Miano et al. Ž1988. and Mobed
et al. Ž1996. to be indicative of FAs fluorescence.
However, the spectrum also resembles that of SHAs.
SHAs do not display the distinctive excitation peak
of the speleothem samples, but it does share the
same emission peak at 430 nm as CW4. The slightly
longer wavelength excitation peak of CW4 is probably due to the higher concentration of organic substances in the calcite and its own absorption effect.
Mobed et al. Ž1996. describe such an effect and
showed that it is caused by inner filtering produced
by self-absorbance. At higher concentrations, the
incident light is absorbed by the molecules themselves, thereby removing the shorter excitation and
emission wavelengths, an effect known as inner filtering or self-absorbance. This effect is most pronounced when the fluorescence spectrum of JC11 is
compared. It is a particularly dark calcite presumed
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
327
Fig. 2. Fluorescence intensities at excitation and emission wavelengths for Ža. Cootes Paradise SFAs, Žb. Cootes Paradise SHAs, Žc. Cold
Water Cave ŽCW4., and Žd. Jewel Cave ŽJC11.. JC11 has different scales on axes. Note the similarity between the SFA and the
speleothems’ fluorescent AfingerprintsB.
to have high concentrations of organic substances; it
displays much longer wavelength excitation and
emission peak intensities, with the excitation peak
centred on 400 nm and the emission peak at 490 nm.
Fig. 3 shows the fluorescence emission spectra of
the solid speleothems. They depict very similar spectra, however, there is a clear distinction between the
light and dark speleothems, with the lighter
speleothems fluorescing more intensely and at shorter
wavelengths than the dark samples. Such a result can
be explained by the self-absorbance effect described
in the previous paragraph. The emission maxima
wavelengths are between 430 and 450 nm and longer.
A reversal of the solid speleothem results occurs
when the speleothems are dissolved. Solutions of the
darker speleothems now fluoresce more intensely. It
appears that the self-absorbance effect has been diminished because the organic molecules are at lower
concentrations in solution than in the solid calcite.
Another possible explanation is that POM has been
removed and this suggests that it may have had a
quenching effect on the speleothems fluorescence.
Maximum intensity emission wavelengths are all
centred between 410 and 430 nm. In comparison
with Ramseyer et al. Ž1997., these wavelengths are
shorter by 10–30 nm, yet share the same broad
328
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Fig. 3. Emission spectra of the various fractions of the speleothems showing the reversal in intensities from solid to liquid phase. The
numbered lines correspond to WAH Ž1., RNCP Ž2., MF1 Ž3., CNPLB Ž4., RNCF2 Ž5., CW4 Ž6., MF2 Ž7., OFD Ž8., JC11 Ž9., CNP Ž10.,
CNPDB Ž11. and RCB Ž12.. Excitation wavelength is 255 nm, except for POM, which is 350 nm.
featureless spectral shape. Spectra of the FA fraction
are similar to that of the total dissolved organic
matter ŽTDOM. solution with common intensity
peaks at 425–430 nm, although as with the TDOM,
intensities do vary. The darker speleothem FA fraction intensities vary from sample to sample but on
average are not stronger than the light coloured
speleothems. The TDOM solution fluorescence in-
tensities are much greater for the dark speleothems
than their light counterparts. The HA fraction possesses a very similar emission to the FA fraction,
although its intensities are lower. The POM spectra
reveal great variability in both intensity and maximum emission fluorescence wavelengths.
To allow easier comparison between the three
fluorescing organic components, FAs, HAs, and
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
POM, the spectra of the light and dark speleothems
were averaged after they had been normalized by
dividing the entire spectra’s intensities of each sample by its maximum intensity ŽFig. 4.. The normalization process retains individual sample spectra
variability. Fig. 4 shows a comparison of the FA and
HA standards with the speleothems; the AlightB and
AdarkB lines are the averaged fluorescence spectra
for the 12 light and dark speleothems used in this
study.
Speleothem FAs yield a fluorescence spectrum
slightly different from that of Cootes Paradise SFAs,
as both the excitation and emission peak wavelengths are shorter. Ramseyer et al. Ž1997. also
found that the fluorescence maxima wavelengths
from their dissolved speleothems were shorter in
comparison with the International Humic Substances
Society SFA standard. In the excitation spectra, the
dark speleothems show a shoulder between 300 and
350 nm which encompass the peak intensity wavelength Ž330 nm. of SFAs. The light speleothems
resemble the dark samples’ excitation spectra. Emission peak intensities are centred between 410 and
420 nm, 20 nm shorter than in the SFAs, which
suggests that the cave deposits probably contain
smaller molecular weight FAs than the standards.
The basis for this suggestion is the work of Senesi
Ž1990. who showed that larger organic compounds
fluoresced at longer emission wavelengths than
smaller organic molecules.
The HA fraction of the speleothems ŽFig. 4.
behaves similarly in both light and dark calcite to the
Cootes Paradise SHAs. They all share the same
excitation peak at 255 nm, although the dark samples
do not have the same plateau between 300 to 320 nm
as do the light samples and SHAs. Emission fluorescence spectra of speleothems and standards are very
similar at an excitation of 255 nm, peaking in intensities of approximately 430 nm. However, with an
excitation wavelength of 340 nm, the light and dark
coloured speleothems differ in maximum emission
fluorescence wavelength by 20 nm. Both higher
concentrations ŽMobed et al., 1996. or larger organic
molecules ŽSenesi, 1990. could account for this difference.
The POM fluorescence depicted in Fig. 4 yielded
a peak excitation intensity at 350 nm when the
emission wavelength was fixed at 430 nm, which is
329
the value of the peak emission obtained from the
dissolved speleothems ŽFig. 3.. The shorter excitation wavelengths were not shown because of the
very strong excitation peaks observed in the immersion oil used for suspension of the POM. Their
strength would totally outweigh that of the POM’s
shorter excitation wavelengths. Emission measurements revealed spectra similar to the solid or dissolved samples. POM of both the light and dark
calcite shows a peak emission wavelength of approximately 420 nm, which compares well with the 430
nm of the FAs and HAs in the dissolved speleothems.
However, the most important feature is the very low
fluorescence intensity of the POM and also the broad,
featureless form of the emission spectra, producing a
white fluorescence. Fluorescence of POM has not
been reported in the literature hitherto. If POM is
trapped in solid speleothems, in low abundance, it
may add to their white fluorescence, but in high
concentrations Žas in MF2., fluorescence may also be
quenched, due to absorption of UV by the POM.
The POM discussed here is simply operationally
defined as any material greater than 0.7 mm. It is
possible for such large matter to reach a cave in the
percolating water. When the authors were working at
Marengo Cave in Southern Indiana ŽVan Beynen et
al., 1997., the cave waters were filtered in the same
manner as the dissolved speleothem calcite and particulate matter was observed on the filters. Velocities
of percolating water moving through epikarst can be
sufficient to transport matter of this size.
5.1.3. Trace elements
The results outlined above strongly suggest that
the fluorescence observed in speleothems is dominated by that which originates from organic compounds. Trace and rare earth element concentrations
were measured to ascertain if they might contribute
to the observed fluorescence. Table 3 reports all
those elements, which were detectable in the
speleothems. The trace elements are discounted from
significant contribution by comparing their published
spectra ŽMunoz and Rubio, 1988; W.B. White, 1998,
personal communication; O’Donnell et al., 1989;
Fujimori et al., 1989; Macler et al., 1989. with
fluorescence of speleothems in the present work at
concentrations greater than the detection levels. Eu
could not be measured by the INAA technique. Eu2q
330
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Fig. 4. Normalized and averaged excitation and emission spectra for FAs, HAs, and POM extracted from speleothems compared to Cootes
Paradise soil FAs rHAs. With the exception of POM, the spectra of the normalized speleothems’ solutions are shorter than the soil
standards’ spectra.
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
yields an emission peak at 416 nm ŽMachel et al.,
1991., close to that of the dissolved speleothems, but
its fluorescent spectrum is quite different and hence,
it can be discounted as a significant fluorescent
contributor.
The high concentration of iodine in our samples is
of interest in its own right. Ullman and Aller Ž1985.
showed that environmental iodine is incorporated by
marine organisms and released to accumulating sediments through organic decomposition. It is not surprising that speleothems may incorporate some iodine as their calcite is composed of CaCO 3 derived
largely from the overlying rocks. Natural abundances
of iodine ŽBecker et al., 1972. have been measured
in limestone Ž- 29 ppm., sandstone Ž- 37.6 ppm.
and calcareous shale Ž- 38 ppm., all of which are
above caves used in this study.
5.1.4. Concentrations of the fractions
The above discussion has shown that FAs and
HAs are the predominant fluorophores in
speleothems, whereas POM probably quenches fluorescence. The next step is to determine which of
these two compounds has the higher concentrations
and hence, is the dominant fluorophore. Fig. 5a
demonstrates each speleothem’s concentrations in
mgrg calcite. The most striking feature is the dominance of POM over the FAs and HAs in all
speleothems except JC11, RCB, and RNCP. The
Table 3
Trace elements present in speleothems Žconcentrations in ppm..
Only values above the detection limit are reported
Element
Ba
Light WAH
MF1
CNPLB
RNCF2
CW4
60
RNCP
Dark MF2
70
OFD
JC11
210
RCB
CNPDB
CNP
D.L.
40
Br
Cr
I
Mn Na
0.6
90
310
0.7
30
20
40
220
11
7
0.4
180
110
2
12.8
0.3
10
Abbreviation: D.L. — detection limits.
2
12
1
20
150
70
40
60
40
230
30
240
310
30
40
10
U
0.3
1.5
2.6
0.3
0.7
0.6
0.7
2.0
5.5
0.1
2.1
4.4
0.1
Zn
60
270
30
15
10
840
460
10
331
McFail’s Cave samples and Crows Nest Pass
speleothems have particularly high abundances.
These high proportional concentrations confirm our
suggestion that the POM has a quenching effect on
the fluorescence of the speleothems, due to self-absorption, contributing to the lower intensities noted
in the solid dark speleothems ŽFig. 3.. However, the
lighter speleothems, which fluoresced more strongly
in the solid, also display high proportions of POM.
Conversely, certain samples tend to have higher
relative abundances of FAs, especially RNCP, RCB,
OFD and JC11. HAs do not appear to be very
prominent, except in CNPDB, CNP, and JC11. The
relative abundances of fulvic and humic acids are
more clearly displayed in Fig. 5b. Generally, HAs
are at least five times less abundant than the FAs.
With such a result, this must have important implications for the contribution of HAs to the fluorescence
of the speleothems. Although not shown in Fig. 2,
SFAs have twice the fluorescence intensity than
SHAs measured at the same excitation and emission
wavelengths, with both having concentrations of 10
ppm. As FAs have twice the fluorescence at the
same concentration as HAs, and the speleothems
possess at least five times the quantity of FAs to
HAs, then FAs must be the dominant fluorophore
due to its concentration. There is also a slightly
higher average concentration of organic compounds
for the darker speleothems compared to the lighter
samples.
5.1.5. Fractionated FA
With FAs being the dominant fluorophore, some
characterization of what compounds make up this
fraction is useful. Of the samples subjected to the FA
fractionations, only JC11 and RCB had high enough
DOC concentrations to provide meaningful results.
Fig. 6 depicts the percentage of TDOC accounted for
by the hydrophilic and hydrophobic fractions. Overall, the hydrophilics are much more dominant than
the hydrophobics. Such dominance is expected because most of the hydrophobics can be removed
from the percolating waters by clay adsorption as
they travel through the soil and microfissures in the
bedrock ŽThurman, 1985.. For both JC11 and RCB,
HPIN is the largest component of the FAs. This
fraction could contain acidic, basic, and neutral fractions, but they are all nevertheless highly hy-
332
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Fig. 5. Ža. POM, FA, and HA concentrations in the speleothems. Žb. Concentrations of FAs and HAs in speleothems. Note the order of
magnitude difference in the concentrations of the various fractions.
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
333
Fig. 6. Relation between fluorescence intensity Žarbitrary units. and concentration of FAs, HAs, and POM in respective solutions. These are
the DOC values of the solutions for the total organic fraction of FAs, HAs, and POC. Shaded triangles are dark speleothems and clear are
light samples.
drophilic. Such components would be expected to
pass through the overlying soil unaffected by adsorption. It is also possible that such components may be
secondary products of the microbial degradation of
the original organic matter leached from the overlying soil, which could have been hydrophobic or
hydrophilic.
5.1.6. Concentration effect on fluorescence
Fig. 7 shows the peak fluorescence intensities
generated at measured concentrations of FAs, HAs,
and POM. Both the FAs and HAs have weak positive relations between intensity and concentration. At
low concentrations of the FAs and HAs, a positive
correlation between concentration and fluorescence
intensity would be expected ŽSenesi, 1990.. More
importantly, the FAs and HAs have similar concentrations, however, the FAs generally fluoresce at
twice the intensity of the HAs, confirming the Cootes
Paradise soil standards result. POM is at much higher
concentrations, and the self-absorption effect ŽMobed
et al., 1996. is therefore very noticeable with an
exponential decay in fluorescence.
5.1.7. Size fractionation
Ultrafiltration of the solutions was undertaken to
confirm that FAs, the dominant fluorophore as determined by the fluorescence spectroscopy, were also
334
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Fig. 7. FA fractionation results for JC11 and RCB showing the dominance of the hydrophilic fraction ŽHPIA and HPIN. compared to the
hydrophobic fraction ŽHPOA and HPON..
the dominant molecular weight fraction. As noted
earlier, FAs typically have molecular weights of
0.8–1 kDa, whereas HAs range from 2 to 300 kDa
ŽSuffett and MacCarthy, 1987.. If FAs or smaller
organic compounds are the dominant fluorophores,
dominance of less than 1 kDa fraction is expected.
TDOC values of the six bulk speleothem solutions
and their fractions were measured and the percentage
composition is shown in Fig. 8. In all the six
speleothems, at least 50% of the soluble organic
substances have molecular weights less than 10 kDa
ŽMWCO.. For RNCP, RNCF2, and OFD the dominant fractions are the FAs Ž- 1 kDa., ranging from
47% to 73% of the TDOC in the speleothems.
However, it is not the dominant fraction in the other
speleothems. HAs, the 1 kDa to - 0.7 mm fractions,
are the dominant components in OFD, WAH, CNP
and RNCF2. However, smaller colloids Ž- 0.7 mm.
could also be contributing to the 0.7 mm–100 kDa
fraction because some have been found to be as
small as 0.01 mm and therefore not trapped as POM
ŽSposito, 1989.. Their presence may explain why the
larger molecular weight fractions are significant in
certain speleothems but the HAs are not ŽFig. 5b..
Colloids, once entrained, remain highly mobile and
consequently, could be flushed with the organic substances into the cave.
5.2. Causes of colour differences between lighter and
darker speleothems
5.2.1. Absorbance spectroscopy
Annual laminae in speleothems are delineated by
differences in the colour of calcite, whereby a couplet of light and dark calcite indicates a year’s
growth. The chemical differences between light and
dark calcite is of interest, but their microscopic size
makes them difficult to investigate. Bulk samples of
calcite will be investigated instead. Although the
distinction between light and dark calcite is subjec-
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
335
Fig. 8. Molecular weight size fractionations using ultrafiltration with the - 10 kDa fractions accounting for most of the DOM extracted
from these speleothems.
tive, every attempt was made to distinguish these
two groups clearly. Only the McFail’s Cave samples,
MF1 and MF2, are from the same cave and yet the
former is light and the latter is dark. The samples
CNPLB and CNPDB are the only samples to come
from one speleothem, which contained thick enough
bands of light and dark calcite to provide 20 gm of
calcite for analysis.
One possible quantitative measure of the difference between light and dark calcite is absorbance.
336
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Absorption spectra ŽFig. 9. show the expected higher
absorbance of darker speleothems in the visible region and correspondingly, the high absorbance in the
near UV. The chromophores, responsible for visible
darkening of the speleothems, also absorb much of
the incident UV, reducing fluorescence through selfabsorption effects.
Absorbance maxima for the speleothems range
from 240 to 300 nm. Such wavelengths, particularly
between 270 and 280 nm, are typical of electron
transitions for phenolic substances, aniline derivatives, benzoic acids, polyenes and polycyclic aro-
matic hydrocarbons ŽBraun et al., 1988.. Some of
these are expected components of humic substances
and have been shown to cause fluorescence in
speleothems, it was anticipated that the peak intensities would occur at these wavelengths.
Fig. 10a shows the correlation between absorbance and TOM ŽTOM s FAs q HAs q POM..
This parameter was chosen because it is not clear
which of the three organic components of the
speleothem would be most responsible for the absorption. There is a weak positive correlation, therefore, TOM could be considered a reasonable proxy
Fig. 9. Absorbance spectra for the speleothems demonstrate the longer absorbance maxima wavelengths of the dark speleothems which are
centred around 300 nm compared with the 240–260 nm of the light speleothems. The same numbers as Fig. 3 are used for the speleothems.
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
337
Fig. 10. Relations between maximum absorbance value of the solid speleothems Žtaken from Fig. 9. and their corresponding Ža. TOM, Žb.
POM, Žc. FA and Žd. HA concentrations. Shaded circles are dark speleothems and clear are light samples. There is a potential that the higher
absorbances Ždarker colours. could be attributed to augmented organic concentrations.
for absorbance and colour. However, dark
speleothems are not consistently higher in TOM
ŽFig. 5a.. Fig. 10b–d demonstrates the relation between absorbance and FAs, HAs, and POM concentrations. It is not immediately apparent which fraction is most responsible for absorption Žcolour.
because they all resemble the TOM distribution
shown in Fig. 10a. However, the POM fraction
resembles the TOM most closely, not surprisingly
since, in most speleothems, POM is the dominant
species and therefore is probably the principal chromophore. Having viewed the colour of the POM and
the colour of the speleothems, it is our view that
there was a great similarity between them. While it is
true that the FA and HA fractions also possessed
similar colours to their parent speleothems, their
concentrations were orders of magnitude lower than
the POM.
While there is little similarity between the fluorescence spectra of the POM ŽFig. 4. and the solid
speleothems, it must be remembered that the wavelengths of excitation were different, but more importantly, that the POM was in a much higher concentration when its fluorescence properties were measured than what it would have been in the speleothem.
Therefore, the inner filtering effect ŽMobed et al.,
1996. created by the high levels of POM would alter
the fluorescence spectra.
Some other characterization of POM could be
useful to determine what this component represents.
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
338
To this end, loss on ignition ŽLOI. analysis was
undertaken on the POM to determine the proportions
of organic to inorganic matter. The LOI procedure
followed was that outlined by Hesse Ž1971.. Whatman w QM-A Quartz filters were used to remove the
POM from solution. The filters were then dried at
1058C for 1 h to remove moisture and weighed to
determine the total POM. The filters were then ignited in a preheated furnace at 9508C for 1 h to
remove the organic matter from the samples and
weighed once more. The quartz filters were selected
because they could withstand the high furnace temperatures required to ignite the organics. Three blanks
were put through the same process and retained their
integrity though the mean weight loss was added to
the final filter weight of the samples. The results of
this experiment are shown in Table 4. There is great
variability in the contribution of organic matter to
the POM fraction, ranging from 0% to 66%. Consequently, the only conclusions that can be drawn are
that the ashrclay content was significant, making up
at least 34% of the total POM. In appearance, the
inorganic component seemed to be fine clay, though
the difficulty of extracting this material from the
filters discouraged any further identification. Clay
does bind organic matter and therefore, it must be
assumed that some of the fluorescence measured in
the POM fraction could have been contributed by
organic molecules bound to the clay.
5.2.2. Size separations
The size separation results also give an indication
of the different size concentrations of the various
light and dark coloured speleothems. Distinctions are
made difficult by the small sample size and the lack
of total unanimity between the light and dark samples as seen in Fig. 8. Not all 12 samples could be
Table 4
Percentage organic matter in POM as derived from LOI analysis.
Only these six samples were analyzed
Sample
% Organic matter
CNP
MF2
OFD
RNCF2
RNCP
WAH
37.2
61.7
44.8
0.0
0.0
66.0
analyzed as the other six speleothems were used for
the FA fractionation. For the light samples, RNCP
and RNCF2 both show the dominance of the smallest fraction, - 1 kDa, while OFD and CNP both
show higher concentrations of the larger molecular
weight fractions. These larger molecular weight fractions are probably the HAs, and it was shown in Fig.
5 that the darker speleothems have on average a
higher concentration of HAs. Regarding the two
smallest size categories Ž- 1–10 kDa., it appears
that all the speleothems have similar proportions
ranging from 60% to 75% of the total soluble organic matter, with the exception of MF2.
5.3. ReactiÕities of organic matter with calcite
Certain organic compounds have high affinities
for sorption on calcite ŽSuess, 1970; Carter, 1978;
Reynolds, 1978.. These variable affinities can lead to
either promotion ŽSuess, 1970. or inhibition of calcite precipitation ŽReynolds, 1978.. If certain substances, such as polyphenols, will preferentially bind
with the calcite and others not, then, this will make
the fluorescence and colour of speleothems differ in
comparison with that of the organic assemblages that
exist in the percolation waters and consequently,
those of the environment above the cave. In an
earlier paper ŽVan Beynen et al., 2000., we found
however, that percolation waters’ fluorescent properties are identical to those of the speleothems measured in this paper. The evidence suggests that any
selective binding of specific organic species on the
calcite is of minor influence in determining the
organic content of the speleothem. The prior observations of the drip waters together with work presented here, suggests all of the organics in the cave
waters are incorporated in the speleothems because
of the macromolecular nature of the organics. While
certain organic components of the macromolecule
may not possess an affinity to bind with calcite,
other components of the molecule do, thereby binding the entire macromolecule to the calcite crystal.
5.4. Comparison of CNPLB and CNPDB
Only one speleothem offered thick enough bands
to compare light and dark calcite, CNPLB and CNPDB. Both came from the same sample, but not from
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
CNP, which was an entirely different speleothem but
which acts as a control. The reason for the difference
in the appearance of the calcite appears to be the
four times higher concentration of HAs in the dark
band compared to the light ŽFig. 5b., as both have
similar POM and FA concentrations ŽFig. 5a.. This
speleothem has two thick pairs of light and dark
bands, suggesting that twice during its growth, there
was a switch from high to low HA concentrations.
Such a transition must reflect a changing environment above the cave but these phases have not been
dated. CNPDB exhibits much stronger fluorescence
than CNPLB in the HA solution fraction. There are
no other differences between these two portions of
the speleothem. This has important implications for
the difference between light and dark calcite, which
comprises annual bands in calcite. CNP has slightly
higher POM levels than CNPDB and CNPLB, but
more importantly like CNPDB, CNP is a dark
speleothem and also has high levels of HAs.
6. Conclusions
The purpose of this study was to investigate the
relation between organic constituents in speleothems
and their fluorescence and colour. It appears that
fluorescence is caused by humic substances and not
trace elements. FAs have higher fluorescent yields
Žper ppm. than HAs, but more significantly because
of their higher relative concentration, appear to be
the dominant contributor to fluorescence. The HPIN
components in FAs are dominant, followed by the
HPIA. The POM in the speleothems may have a
quenching effect on the fluorescence, a finding suggested by the reversal of fluorescent intensities of
light and dark calcite once the POM was removed by
dissolution of the speleothem. However, the enhanced fluorescence can also be caused by the dilution of the FAs–HAs that in the solid calcite absorbed a great deal of the fluorescence.
Average organic concentrations Žin 1 gm of calcite. in the speleothems are as follows: POM — 204
ppm, FAs — 0.893 ppm, and HAs — 0.064 ppm.
These results show that some darker calcites possess
higher concentrations of humic substances, but fluoresce less strongly than lighter calcite due to the
339
higher abundances in the dark calcite of the POM
and its associated fluorescent quenching. Some dark
speleothems also contain larger, more complex organic substances ŽHAs. as well as the simpler compounds ŽFAs. of the lighter calcite. Extrapolating
these findings to the interpretation of annual laminae
in speleothems, periods of high organic concentrations in cave feedwater will produce low fluorescence in the solid speleothem, such as late winter
and the spring when humic substances are flushed
from the soil ŽVan Beynen et al., 1997. due to
self-absorbance or quenching. Conversely, periods of
low concentrations will yield high fluorescence in
the speleothem, such as the summer and fall. Therefore, the presence of annual bands can be determined
using fluorescence microscopic photography.
Multiple speleothems from two caves ŽRatsnest
and McFail’s. exhibited some differences in behaviour, a result that can best be explained by the
nature of the epikarst above the cave which can
produce different flow paths and soil thicknesses
above the speleothems. However, the Crows Nest
Pass speleothems were quite similar. Comparison of
light and dark calcite from this cave showed that
dark calcite ŽCNPDB. contained more HAs than the
light calcite ŽCNPLB., a result which strengthens
arguments about the chemical differences between
light and dark calcite.
Acknowledgements
We greatly appreciate the technical help of George
Timmins, Scott Smith, Karen Edmondson and Alice
Pidruczny. This research was funded by an operating
grant to Ford from the National Sciences and Engineering Research Council of Canada.
References
Aiken, G.R., McKnight, D.M., Thorn, K.A., Thurman, E.M.,
1992. Isolation of hydrophilic organic acids from water using
nonionic macroporous resins. Org. Geochem. 18, 567–573.
Aiken, G.R., 1995. Isolation and concentration techniques for
aquatic humic substances. In: Aiken, G.R., McKnight, D.M.,
Wershaw, R.L., MacCarthy, P. ŽEds.., Humic Substances in
Soil, Sediment, and Water. Wiley, New York, pp. 363–385.
340
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
Baker, A., Smart, P.L., Richards, A., 1993. Annual growth bandings in a cave stalagmite. Nature 272, 24–28.
Baker, A., Genty, D., Smart, P.L., 1998. High-resolution records
of soil humification and paleoclimate change from variations
in speleothem luminescence excitation and emission wavelengths. Geology 26, 903–906.
Becker, V.J., Bennett, J.H., Manuel, O.K., 1972. Iodine and
uranium in sedimentary rocks. Chem. Geol. 9, 133–135.
Bourbonniere, R.A., 1989. Distribution patterns of dissolved organic matter fractions in natural waters from eastern Canada.
Org. Geochem. 14, 97–107.
Braun, D.W., Floyd, A.J., Sainsbury, M., 1988. Organic Spectroscopy. Wiley, New York, pp. 3–23.
Broecker, W.S., Olsen, E.A., Orr, P.C., 1960. Radiocarbon measurements and annual rings in cave formations. Nature 185,
93–94.
Carter, P.W., 1978. Adsorption of amino acid-containing organic
matter by calcite and quartz. Geochim. Cosmochim. Acta 42,
1239–1242.
Clapp, C.E., Hayes, M.H.B., 1999. Sizes and shapes of humic
substances. Soil Sci. 164, 777–789.
de Soete, D., Gijbels, R., Hoste, J., 1972. Neutron Activation
Analysis. Wiley, New York.
Dorale, J.A., Gonzalez, L.A., Reagan, M.K., Pickett, D.A., Murell,
M.T., Baker, R.G., 1992. A high resolution record of Holocene
climate change in speleothem calcite from Cold Water Cave,
northeastern Iowa. Science 258, 1626–1630.
Ertel, J.R., 1988. Genesis: group report. In: Frimmel, F.H., Christman, R.F. ŽEds.., Humic Substances and Their Role in the
Environment. Wiley, New York, pp. 105–112.
Fujimori, H., Matsui, T., Suzuki, K., 1989. Simultaneous determination of U and HNO 3 concentrations in solution by laser
induced fluorescence spectroscopy. J. Nucl. Sci. Technol. 25,
798–804.
Gascoyne, M., 1977. Trace element geochemistry of speleothems.
Proc. 7th Int. Speleo. Cong., Sheffield, England. pp. 205–208.
Gascoyne, M., 1992. Paleoclimatic determination from cave calcite deposits. Quat. Sci. Rev. 11, 609–632.
Genty, D., Quinif, Y., 1996. Annually laminated sequences in the
internal structure of some Belgian stalagmites — importance
for paleoclimatology. J. Sediment. Res. 66, 275–288.
Gilson, J.R., MacCarthney, E., 1954. Luminescence in speleothems
from Devon, UK: the presence of organic activators. Ashford
Speleol. Soc. J. 6, 8–11.
Goldberg, M.C., Weiner, E.R., 1991. Fluorescence spectroscopy
in environmental and hydrological sciences. In: Wolfbeis, O.S.
ŽEd.., Fluorescence Spectroscopy: New Methods and Application. Springer-Verlag, New York, pp. 213–241.
Hayase, K., Tsubota, H., 1985. Sedimentary humic acid and fulvic
acid as fluorescent organic materials. Geochim. Cosmochim.
Acta 49, 159–163.
Hesse, P.R., 1971. A Textbook of Soil Chemical Analysis. Chemical Publishing, New York, p. 210.
Lauritzen, S.E., Ford, D.C., Schwarcz, H.P., 1986. Humic substances in speleothem matrix, paleoclimate significance. Proc.
9th Int. Cong. Speleology, Barcelona, Spain. pp. 77–79.
Leenheer, J.A., 1981. Comprehensive approach to preparative
isolation and fractionation of dissolved organic carbon from
natural waters and waste waters. Environ. Sci. Technol. 15,
578–587.
Leenheer, J.A., 1985. Fractionation techniques for aquatic humic
substances. In: Aiken, G.R., McKnight, D.M., Wershaw, R.L.,
MacCarthy, P. ŽEds.., Humic Substances in Soil, Sediment,
and Water. Wiley, New York, pp. 409–429.
Li, W., Lundberg, J., Dickin, A.P., Ford, D.C., Schwarcz, H.P.,
Williams, D., 1989. High precision mass-spectrometric Useries dating of cave deposits and implications for paleoclimate studies. Nature 339, 534–536.
Machel, H.G., Mason, R.A., Mariano, A.N., Mucci, A., 1991.
Causes and emission of luminescence in calcite and dolomite.
In: Barker, C.E., Kopp, O.C. ŽEds.., Luminescence Microscopy and Spectroscopy. SEPM, pp. 9–25.
Macler, M., Nicolai, J.P., Heaven, M.C., 1989. Electronic spectroscopy and energy transfer pathways for matrix isolated
iodine. J. Chem. Phys. 91, 365–373.
Malcolm, R.L., 1991. Factors to be considered in the isolation and
characterization of aquatic humic substances. In: Boren, H.,
Allard, B. ŽEds.., Humic Substances in the Aquatic and Terrestrial Environment. Wiley, London, pp. 369–391.
Malcolm, R.L., MacCarthy, P., 1992. Quantitative evaluation of
XAD-8 and XAD-4 resins used in tandem for removing
organic solutes from water. Environ. Int. 18, 597–607.
Mason, R.A., Mariano, A.N., 1990. Cathodoluminescence activation in manganese-bearing and rare earth-bearing synthetic
calcite. Chem. Geol. 88, 191–206.
McKlveen, J.W., 1981. Fast Neutron Activation Analysis. Ann
Arbour Sci. Publ.
Miano, T.M., Sposito, Martin, J.P., 1988. Fluorescence spectroscopy of humic substances. Soil Sci. Soc. Am. J. 52,
1016–1019.
Mitterer, R.M., 1968. Amino acid composition of organic matrix
in calcareous oolites. Science 162, 1498–1499.
Mobed, J.J., Hemminsen, S.L., Autry, J.L., McGown, L.B., 1996.
Fluorescence characterization of IHSS humic substances: total
luminescence spectra with absorbance correction. Environ.
Sci. Technol. 30, 3061–3065.
Munoz, F.F., Rubio, O.J., 1988. Fluorescence of tin sensitized Mn
in single-crystalline NaCl. Phys. Rev. B 38, 9980–9986.
O’Donnell, K.P., Marshall, A., Yamaga, M., Henderson, B., 1989.
Vibronic structure in the photoluminescence spectrum of Cr 3q
ions in garnets. J. Lumin. 42, 365–373.
Pietrzyk, D.J., Chu, C., 1977. Amberlite XAD copolymers in
reversed phase gravity flow and high pressure liquid chromatography. Anal. Chem. 49, 757–763.
Ramseyer, K., Miano, T.M., D’Orazio, V., Wildberger, A., Wagner, T., Geister, J., 1997. Nature and origin of organic matter
in carbonates from speleothems, marine cements and coral
skeletons. Org. Geochem. 26, 361–378.
Reynolds, R.C., 1978. Polyphenol inhibition of calcite precipitation in Lake Powell. Limnol. Oceanogr. 23, 585–597.
Richmond, M., Bourbonniere, R.A., 1987. Manual for the fractionation of dissolved organic matter in natural waters. NWRI
Report 87-145, Environment Canada, Ottawa, p. 29.
Schlesinger, W.H., 1977. Carbon balance in terrestrial detritus.
Ann. Ecol. Syst. 8, 51–81.
Schwarcz, H.P., 1986. Geochronology and isotopic geochemistry
P. Õan Beynen et al.r Chemical Geology 175 (2001) 319–341
of speleothems. In: Fontes, J.C., Fritz, P. ŽEds.., Handbook of
Environmental Isotope Geochemistry. The Terrestrial Environment, B. Elsevier, Amsterdam, pp. 271–303.
Senesi, N., 1990. Molecular and quantitative aspects of the chemistry of fulvic acid and its interactions with metal ions and
organic chemicals. Anal. Chim. Acta 232, 77–106.
Senesi, N., Miano, T.M., Provenzano, M.R., Brunetti, G., 1991.
Fluorescence spectroscopy of humic substances. Soil Sci. Soc.
Am. J. 52, 1016–1019.
Shopov, Y.Y., Dermendjiev, V., Buyukliev, G., 1989. Investigation on the variations of climate and solar activity by a new
method — LLMZA of cave flowstone from Bulgaria. Proc.
10th Int. Cong. Speleology, Budapest, Hungary, pp. 95–97.
Shopov, Y.Y., Ford, D.C., Schwarcz, H.P., 1994. Luminescence
microbanding in speleothems: high resolution chronology and
paleoclimate. Geology 22, 407–410.
Simpson, A.J., Burdon, J., Graham, C.L., Hayes, M.H.B., 1997a.
Humic substances from podzols under oak forest and cleared
forest site: II. Spectroscopic studies. In: Hayes, M.H.B., Wilson, W.S. ŽEds.., Humic Substances, Peats, and Sludges:
Health and Environmental Aspects. Royal Society of Chemistry, Cambridge, pp. 83–92.
Simpson, A.J., Watt, B.E., Graham, C.L., Hayes, M.H.B., 1997b.
Humic substances from podzols under oak forest and cleared
forest site: I. Isolation and characterization. In: Hayes, M.H.B.,
Wilson, W.S. ŽEds.., Humic Substances, Peats, and Sludges:
Health and Environmental Aspects. Royal Society of Chemistry, Cambridge, pp. 73–82.
Smart, P.L., Laidlaw, I.M.S., 1977. An evaluation of some fluorescent dyes for water tracing. Water Resour. Res. 13, 15–33.
Sposito, G., 1989. The Chemistry of Soils. Oxford, New York, p.
188.
Steelink, C., 1985. Implications of elemental characteristics of
humic substances. In: Aiken, G.R., McKnight, D.M., Wershaw, R.L., MacCarthy, P. ŽEds.., Humic Substances in Soil,
Sediment, and Water. Wiley, New York, pp. 457–476.
Stevenson, F.J., 1982. Humus Chemistry. Wiley, New York, 443.
Stevenson, F.J., 1985. Geochemistry of soil humic substances. In:
Aiken, G.R., McKnight, D.M., Wershaw, R.L., MacCarthy, P.
ŽEds.., Humic Substances in Soil, Sediment, and Water. Wiley, New York, pp. 14–51.
Suess, E., 1970. Interaction of organic compounds with calcium
341
carbonate-I. Association phenomena and geochemical implications. Geochim. Cosmochim. Acta 34, 157–168.
Suffett, I.H., MacCarthy, P., 1987. Aquatic Humic Substances.
ACS, Washington, xxiii.
Swift, R.S., 1985. Fractionation of soil humic substances. In:
Aiken, G.R., McKnight, D.M., Wershaw, R.L., MacCarthy, P.
ŽEds.., Humic Substances in Soil, Sediment, and Water. Wiley, New York, pp. 387–408.
Terakado, Y., Masuda, A., 1988. The coprecipitation of rare-earth
elements with calcite and aragonite. Chem. Geol. 69, 103–110.
Thurman, E.M., 1985. Humic substances in groundwater. In:
Aiken, G.R., McKnight, D.M., Wershaw, R.L., MacCarthy, P.
ŽEds.., Humic Substances in Soil, Sediment, and Water. Wiley, New York, pp. 87–103.
Ullman, W.J., Aller, R.C., 1985. The geochemistry of iodine in
near-shore carbonate sediments. Geochim. Cosmochim. Acta
49, 967–978.
Van Beynen, P.E., Ford, D.C., Schwarcz, H.P., 1997. Characterization and seasonality of humic substances in Cave Water,
Marengo Cave, Indiana. Proc. 12th Int. Cong. Speleology, La
Chaux de Fonds, Switzerland, p. 262.
Van Beynen, P.E., Ford, D.C., Schwarcz, H.P., 2000. Seasonal
variations in organic substances in surface and cave waters at
Marengo Cave, Indiana. Hydrol. Proc. Žin press..
Visser, S.A., 1983. Fluorescence phenomena of humic matter of
aquatic origin and microbial cultures. In: Christman, R.F.,
Gjessing, E.T. ŽEds.., Aquatic and Terrestrial Humic Materials. Ann Arbor Sci. Publ., Ann Arbor, MI, pp. 183–202.
Wershaw, R.L., Aiken, G.R., 1985. Molecular weights of humic
substances. In: Aiken, G.R., McKnight, D.M., Wershaw, R.L.,
MacCarthy, P. ŽEds.., Humic Substances in Soil, Sediment,
and Water. Wiley, New York, pp. 409–429.
White, W.B., Brennan, E.S., 1989. Luminescence of speleothems
due to fulvic acid and other activators. Proc. 10th Int. Cong.
Speleology, Budapest, Hungary, pp. 212–214.
White, W.B., 1997. Precise measurement of luminescence banding profiles in speleothems for paleoclimatic interpretation.
Proc. 12th Int. Cong. Speleology vol. 1. La Chaux de Fondes,
Switzerland, pp. 89–92.
Wilson, M.A., Collin, P.J., Malcolm, R.L., Perdue, E.M., Cresswell, P., 1988. Low molecular weight species in humic and
fulvic fractions. Org. Geochem. 12, 7–12.