During the past several years CO self-shielding has been investigated as the source
of the 16O-poor endmember of the CAI mixing line [1-4]. A key assumption in
most of the models is that the δ17O/δ18O ratio (i.e., ‘slope’) associated with CO
photodissociation under self-shielding conditions (when C16O is optically thick) is unity or
very close to unity. Laboratory CO photodissociation experiments by Chakraborty et al.
[5] found slopes ranging from ~ 1.8 at 107 nm to ~ 0.6 at 94 nm. Such a strong
dependence of slope on wavelength appears to be inconsistent with the CO self-shielding
hypothesis.
Here I present simulations of these CO experiments, and also of the solar nebula, using
complete line-by-line spectra for the oxygen isotopologues of CO. The questions I address
are 1) What is the origin of the wavelength dependence in δ17O/δ18O ratio of photoproduct O
(as CO2)?; 2) Can a line-by-line spectrum capture the wavelength dependence or is a more
complete spectral model necessary?; 3) What are the implications for the self-shielding
theory for the solar nebula?
I computed line-by-line spectra for 37 bands of 12C16O, 12C17O and 12C18O at 300 K
from published collections on CO molecular constant data [6,7]. Bands 31-33 have a
resolution of 1x10-5 nm, and all other bands have a resolution of 3x10-5 nm. Simulations of
photolysis experiments yield slopes for atomic O of 0.7 to 0.8 at 92.6, 94.1 and 97.0 nm.
Measured values [5] were between 0.6 and 1.0, and the calculations agree with the
measurements that slopes are < 1 at 92 and 94 nm. Computed and measured slopes at 107 nm
are > 1, as is also found in previous simulations [8], which results from an optically thick
column of C18O.
Solar nebula modeling was performed with the computed spectra added to the disk model
of [3]. Calculations were done at 30 AU and 50 K, and also at 10 K and 1500 K to explore the
temperature dependence of the slope. A slope of ~0.9 is found for 50 K (and 1500 K), and ~
1.2 for 10 K. Thus, temperature dependence is predicted for the slope, which may be useful in
eliminating some versions (i.e., locations) of the self-shielding theory. The large range of
slopes seen by Chakraborty et al. [5] is not evident in disk models, nor in cloud models
[7].
[1] Clayton, R. N. (2002) Nature 415, 860-861.
[2]Â Yurimoto & Kurimoto (2004) Science 305, 1763–1766.
[3]Â Lyons J. R. and E. D. Young (2005) Nature 435, 317-320.
[4]Â Lee et al. (2008) Meteorit Planet. Sci. 43, 1351–1362.
[5] Chakraborty et al. (2008) Science 321, 1328-1331.
[6] Eidelsberg et al. (1991) Astron Astrophys S.S. 90, 231-282,
[7] Visser et al (2009) Astron. &Astrophys. 503, 323-343.
[8] Lyons, J. R. (2009) 40th LPSC, abstract 2377. |