Oxygen Isotopes as Applied to Phosphate:
Development and Use of a Paleoproxy
Introduction:
-Studying the δ18O values of materials precipitated in sea-water (esp. biogenic
CO32-) has provided detailed Cenozoic temperature and climate information.
-Carbonate δ18O is limited with respect to providing absolute δ18O values
of the water in which the material was precipitated.
-To address this problem, researchers studied δ18O values from other mineral
phases that form under similar conditions (SiO2 and Ca5(PO4)3).
- Longinelli and Nuti (1973) developed the following phosphate δ18O temp
equation:
tºC = 111.4 -4.3(δp-δw)
- Unfortunately, this equation falls along the same slope as the carbonate δ18O
temperature equation.  See Fig. 1 (Faure 1986)
Development of a Useful Paleo-proxy:
-The utility of carbonate δ18O values also is limited over longer
periods of geologic time. Carbonates are susceptible to increasing diagenetic
alteration with age, which tends to destroy the original isotopic signal within the
sample.
- There existed a need for a system that was robust through larger sections of
geologic history. Phosphatic materials provide that system, as they exhibit
many of the characteristics that qualify carbonates as paleo-environmental
indicators, while having a decreased loss of signal potential, compared to
carbonate.
- As discussed in Faure (1986), phosphates possess several desirable properties as
a paleo-climate proxy –
1) Enzyme catalyzed oxygen isotope exchange between phosphate and
water is very rapid – operating on time scale of minutes
2) Isotope exchange at low temp. between phosphate and water is
extremely slow, even on a geologic time scale
3) The isotope composition of oxygen in the bones and teeth of fish in
equilibrium with the water and is not affected by phosphate in the diet
4) Paleotemperatures can be determined from a wide range of materials
including fish teeth, conodonts, inarticulate brachs, phosphorites, etc.
-however, study of unaltered carbonates provide a useful corroborative tool for
phosphate δ18O record during the Paleozoic.
Methods:
-Direct analysis of biogenic apatite Ca5(PO4,CO3,F)3(OH, F,Cl,CO3)
has proven difficult due to the occurrence of other oxygen bearing species that
can occur along with (PO4)3-. Developing accurate results is dependant on the
formation of an intermediate species that contains the (PO4)3- ion, but none of
the additional oxygen bearing compounds found in biogenic apatite
- Two methods dominate studies that seek to develop the
phosphate δ18O record.
1)Formation of a BiPO4 compound which is reacted with BrF3 (or a
similar compound) to release the oxygen, which is then
converted to CO2 and analyzed for its δ18O value.
-Requires 10-25mg of sample
-This method was widely used in initial phosphate oxygen work
although it (BiPO4) is hygroscopic as well as being labor-intensive
to produce in the lab.
- Tudge (1960)
- However it was shown that Ag3PO4 can be precipitated from a phosphate
bearing solution. Due to its stability and simpler production, Ag3PO4 has
become the favored intermediate for δ18O determinations
2) O’Neil et al. (1994) develops a method for preparation of Ag3PO4 from
phosphate bearing samples, and is described as follows:
1) Dissolution of ~ 25mg sample in 2M HNO3
2) Addition of 2M KOH (raise pH to ~ 5) followed by the addition
of 2M HF (causes precipitation of CaF2, which is separated via
centrifuge)
3) Phosphate bearing fluid is transferred to clean test tube
4) Addition of a silver amine solution followed by heating to 50ºC
causes precipitation of Ag3PO4 from solution. This occurs as
ammonia is driven from solution, which lowers the pH (~10)
to a value of ~ 8.5 where precipitation begins
5) Samples are then filtered on a Millipore filter and rinsed (x3)
6) Samples are dried and placed into to silica tubes along with
graphite
7) The tube is then placed under vacuum and sealed
8) The sealed tube is reacted at 1200ºC for 3 minutes which allows
the mixture to evolve CO2 ; which is then analyzed via mass
spectrometry.
-Venneman et al. (2002) tested three methods of Ag3PO4 analysis
1) Fluorination – (similar to Tudge 1960)
2) Sealed Silica Tube reaction (see above)
3) High Temperature Reduction
- 0.2-1.0 mg Ag3PO4 is placed in a silver foil capsule
- capsule is then loaded into autosampler of a TC-EA (hightemperature conversion elemental analyzer)
- sample falls into graphite crucible heated to 1440ºC
- CO gas is produced and carried into Finnigan Delta Plus XL
mass spectrometer via ConFlo interface (allows for
introduction of CO reference gas prior to and after each run)
-Fluorination and SST require large (10-25 mg) samples, necessitates use
of samples with high phosphatic abundance or the lumping of smaller
samples. Work tended to focus on phosphorite and inarticulate
brachiopods, while excluding all but bulk sample analyses of conodont
phosphatic material. (Luz et al., 1984)  see Figure 2
-Conodont flouro-apatite, a dense crystalline substance, holds great
potential for accurately recording and storing δ18O signals from the
waters it lived in. Additionally, conodonts exhibit much greater
distribution across varying marine conditions (and corresponding
lithologies) and occur throughout the Paleozoic.
-However, they (conodonts) are rarely found in the abundances required
by fluorination or SST.
-HTR method allows study of very small samples which increases the
data resolution (less lumping of sample) and allows for high resolution
δ18O studies of Paleozoic rocks. (Bates et al., 2003)  see Fig 3-6
Discussion of Wenzel et al., (2000).
-The authors studied Silurian brachiopods (articulate and inarticulate)
and conodonts from Gotland. These samples were deposited at tropical
latitudes in shallow marine shelf conditions. Carbonate and phosphate samples
are compared in an attempt to decouple primary oxygen isotope signals from
those of secondary alteration.
-Many of the same issues are raised concerning the mechanics of using carbonates
and phosphates in the Paleozoic as paleo-climate indicators.
-Samples of articulate brachiopod carbonate were analyzed for δ18O values using
standard methods (powdering sample, reacting with phosphoric acid in a Kiel
automated preparation line coupled to a Finnigan MAT 252 mass spectrometer).
-Bulk samples of inarticulate brachiopods were prepared using the SST process
discussed earlier. δ18O values of these samples were obtained from a VG Prism
mass spectrometer.
-Samples consisting of 5-20 conodont elements (0.1-0.5 mg) were converted to
Ag3PO4 based on the method outlined earlier. These samples were analyzed
utilizing a laser-based microsampling technique, which heats the sample using an
infrared laser which produces CO2. The δ18O value of this gas is then measured
using a Finnigan MAT 252 mass spectrometer.
-The authors report δ18O values as follows:
-inarticulate brachiopods - 13.0-17.5 ‰ V-SMOW
-conodonts
- 17.5-19.5 ‰ V-SMOW
 see fig 7
- The majority of samples show a more consistent (valid?) δ18O signal
from conodont samples, although the authors point out two samples that
show greater variability. This variation is attributed to the contribution
of basal filling apatite, which has a lower preservation potential than that
of the more densely crystalline crown apatite.
-The difference between δ18O values from the inarticulate brachiopods and
conodonts is attributed to diagenetic alteration of the brachs.
-The authors move on to develop a comparison of the δ18O values from phosphate
with those from calcitic brachiopods. This comparison sought to address the
apparent offset between unaltered brachiopods and conodont δ18O values.
-Several possible mechanisms for the offset are discussed:
- The “unaltered” brachiopods may actually have undergone some
alteration that shifts the δ18O values lower, while not changing
appearance or crystalline structure.
- The samples(conodonts and brachs) may have formed in isotopically
different bodies of water.
-Under some circumstances, biogenic carbonates may be formed out of
isotopic equilibrium with the water they form in.
- Varying concentrations of CO32- in water, which alter both the δ13C and
δ18O of the carbonate that forms in that water.
- While only speculation, the authors favor the idea of changing CO32concentrations as a mechanism for shifting the fractionation of O and C
in the brachiopod samples.
References:
Bates, S., Lyons, T., Brown, L., Rexroad, C., and Bright, C. (2003) Conodont
geochemical records of late Paleozoic Paleoenvironmental variability in midcontinent North America. Abstracts with Programs - Geological Society of
America 35; 6, 567.
Faure, G., (1986) Principles of isotope geology. John Wiley and Sons. 429-455.
Longinelli, A., and Nuti S. (1973) Revised phosphate-water isotopic temperature scale.
Earth and Planetary Science Letters,19, 373-376.
Luz, B., Kolodny, Y., and Kovach, J. (1984) Oxygen isotope variation in phosphate of
biogenic apatite, III Conodonts. Earth and Planetary Science Letters, 69, 255-262.
O’Neil, J., Roe, L., Reinhard, E., and Blake, R., (1994) A rapid and precise method of
oxygen isotope analysis of biogenic phosphate. Israel Journal of Earth Science,
43, 203-212.
Tudge, A., (1960) A method of analysis of oxygen isotopes in orthophosphates – its use
In measurements of paleotemperatures. Geochemica et Cosmochimica Acta, 18,
81-93.
Wenzel, B., Lecuyer, C., and Joachimski, M., (2000) Comparing oxygen isotope records
of Silurian calcite and phosphate - δ18O compositions of brachiopods and
conodonts. Geochemica et Cosmochimica Acta, 64, 1859-1872.