First detection of the atomic ${}^{18}\mathrm{O}$ isotope in the mesosphere and lower thermosphere of Earth

In the lower atmosphere of Earth, oxygen contains a higher fraction of the heavy ${}^{18}\mathrm{O}$ isotope than ocean water does (Dole effect). This isotopic enrichment is a signature of biological activity, set by the equilibrium between oxygenic photosynthesis and respiratory metabolisms in terrestrial and oceanic ecosystems. While the mixing between stratospheric and tropospheric oxygen leads to a slow isotopic homogenization, little is known about the isotopic oxygen enrichment in the mesosphere and thermosphere of Earth. In situ measurements from rocket-borne air samplers are limited to altitudes below the mesopause, while higher layers have only been accessible through the analysis of the oxidation of ancient cosmic spherules. Here we report the detection of the far-infrared fine-structure lines (${}^3{\mathrm{P}}_1\leftarrow {}^3{\mathrm{P}}_2$ and ${}^3{\mathrm{P}}_0\leftarrow {}^3{\mathrm{P}}_1$) of ${}^{18}\mathrm{O}$ in absorption against the Moon, and determine the ${}^{16}\mathrm{O}/{}^{18}\mathrm{O}$ ratio in atomic oxygen from the mesosphere and lower thermosphere in absorption. After correcting for isotopic exchange between atomic and molecular oxygen, our values for the bulk ${}^{16}\mathrm{O}/{}^{18}\mathrm{O}$ ratio of 468 and 382 in February and November 2021, respectively, fall significantly below that found in solar wind samples ($530\pm 2$), and encompass, within uncertainties, the corresponding ratios pertaining to the Dole effect in the troposphere (487), and those found in stratospheric ozone (429 to 466). We show that with existing technology, future, more sensitive measurements will allow us to monitor deviations from isotopic homogeneity in the mesosphere and lower thermosphere of Earth by remote sensing. We demonstrate that the collisional excitation of the fine-structure levels of the ${}^3\mathrm{P}$ ground-state triplet of ${}^{18}\mathrm{O}$ may compete with isotopic exchange reactions, implying a deviation from the Boltzmann distribution that would be established under local thermodynamic equilibrium.

Figure 1

Optical depth at zenith as observed in the (a) ${}^3{\mathrm{P}}_1{\leftarrow }^3{\mathrm{P}}_2$ and (b) ${}^3{\mathrm{P}}_0{\leftarrow }^3{\mathrm{P}}_1$ absorption line on 2021 February 24 (grey-filled histograms). Note the partially logarithmic ordinates (linear scales below optical depth thresholds, indicated by horizontal-dotted lines, of 0.004 and 0.002 in the ${}^3{\mathrm{P}}_1{\leftarrow }^3{\mathrm{P}}_2$ and ${}^3{\mathrm{P}}_0{\leftarrow }^3{\mathrm{P}}_1\text{line}$, respectively). The line profile fits with a parametrized profile function are overlaid in blue (orange and red dots show the components in ${}^{16}\mathrm{O}$ and ${}^{18}\mathrm{O}$, respectively). The residuals of the fit, as described in Sec. 3 A and after transformation by Eq. (A10), are shown as green histograms. Lime colored dots display the chemical model for the MSIS 2 model of the MLT after equilibrium is established, including collisional excitation and radiative decay. Green crosses mark the static LTE model for $\delta {}^{18}\mathrm{O}=24.15‰$. (c) Optical depth of the ${}^3{\mathrm{P}}_1{\leftarrow }^3{\mathrm{P}}_2$ line on 2021 November 19. Symbols and colors as in (a) and (b).

Figure 2

Zoom into the observed lunar area (top and right are selenographic north and east, respectively), displaying Mare Tranquillitatis to the East, Mare Serenitatis to the North, and crater Copernicus to the West. The overlaid graticule indicates selenographic coordinates. The celestial north pole is at $15\stackrel{\circ }{.}3$ selenographic E from N. The position targeted by our absorption experiment is at ${30}^{\circ }\text{E}$ and $8\stackrel{\circ }{.}5$ N. Its distance from the observer at start was 377 677 km (JPL Horizons On-Line Ephemeris System version 4.50). Image from NASA's Lunar Reconnaissance Orbiter Camera (via quickmap.lroc.asu.edu, credit: NASA/GSFC/ASU). The filled circles in the framed subset show the pixel positions and half- power contours of the low-frequency (orange) and high-frequency arrays (blue)

Figure 3

False-color representation of spectral profiles of the optical depths of the (a) ${}^3{\mathrm{P}}_1{\leftarrow }^3{\mathrm{P}}_2$ and (b) ${}^3{\mathrm{P}}_0{\leftarrow }^3{\mathrm{P}}_1$ lines, as a function of the altitude of a hypothetical observer. The frequencies for the ${}^{16}\mathrm{O}$ and ${}^{18}\mathrm{O}$ transitions are indicated by vertical solid and dashed lines respectively. The horizontal-grey lines indicate the altitude where 50% of the final optical depth (as integrated across the spectral profile) is reached, roughly coinciding with the mesopause (around 110 km altitude). The dotted contours are at 50% and 2% of the line-center optical depth at a given altitude; the lower contour indicates a ten times larger optical depth than that expected for ${}^{18}\mathrm{O}$. It crosses the frequency of the ${}^{18}\mathrm{O}$ transition only at 145 km altitude, where the concentration of atomic oxygen drops by an order of magnitude with respect to the mesopause. Below the mesopause, the pressure broadening emerges, yet without destroying the asymmetry of the line profile due to ${}^{18}\mathrm{O}$. The inserts to the right show the corresponding altitude profiles of the underlying opacities (i.e., derivative of optical depth with respect to altitude).

Figure 4

Synoptic presentation of ${}^{16}\mathrm{O}/{}^{18}\mathrm{O}$ ratios in the solar system, separated into isotope ratio mass spectroscopy (left) and (far-)infrared and submillimeter spectroscopy (right), with appropriate labels enhanced by a selection of colors. The study at hand is shown in red and refers to the ${}^3{\mathrm{P}}_1\leftarrow {}^3{\mathrm{P}}_2$ transition on 2021 February 24 (left) and 2021 November 19 (right), after corrections for non-LTE and isotopic exchange with ${\mathrm{O}}_2$. The orange bar indicates the isotope ratio due to the Dole effect. The gray-shaded area indicates the range of $\delta {}^{18}\mathrm{O}$ values in the troposphere and stratosphere of Earth, lunar, and meteoritic samples; the blue-hatched portion refers to $\delta {}^{18}\mathrm{O}$ values from stratospheric ozone [67, 68]. Error bars refer to $\pm 1\sigma$ and are shown if they exceed the symbol size. Label superscripts indicate the used reference.

Figure 5

MSIS 2 model [43] for SOFIA flights #702 and #794 on 2021 February 24 (solid lines) and November 19 (dashed lines), respectively. The displayed profiles can be distinguished either by color or line style and labels. Temperature (lower abscissa) and concentrations of atomic and molecular oxygen (upper abscissa) are shown against altitude above ground (left ordinate) or corresponding to a given position on the sightline (right ordinate). The base point of the sightlines and their elevations correspond to leg medians (geographic longitude ${21}^{\circ }$ W, latitude ${43}^{\circ }$ N, elevation ${49}^{\circ }$ for 2021 February 24, and longitude ${110}^{\circ }$ W, latitude ${46}^{\circ }$ N, elevation ${55}^{\circ }$ for 2021 November 19). For flight #794, data obtained close in space and time by broadband emission radiometry (SABER instrument aboard the TIMED satellite) are marked as crosses (temperature in magenta, and the concentrations of ${}^{16}\mathrm{O}$ and ozone in dark blue and cyan, respectively). Their analysis is described in [66].

Figure 6

Time evolution of correction factors for (a) non-LTE in the fine-structure levels of ${}^{18}\mathrm{O}$, (b) partitioning of the ${}^{18}\mathrm{O}$ enrichment into atomic and molecular oxygen.

Figure 7

Color-scale representation of chemical evolution of ${}^{18}\mathrm{O}$ enrichment in the ${}^3{\mathrm{P}}_0, {}^3{\mathrm{P}}_1$ , and ${}^3{\mathrm{P}}_2$ levels (top and bottom left), and in ${}^{16}\mathrm{O}{}^{18}\mathrm{O}$ (bottom right), with respect to their respective parent species. Shown are $\delta {}^{18}\mathrm{O}$ values relative to the VSMOW standard [4], against time (abscissae) and altitude (ordinates), for the reservoir formed by ${}^{16}\mathrm{O}$ and fine-structure resolved ${}^{18}\mathrm{O}$, and for that containing ${}^{16}\mathrm{O}{}^{18}\mathrm{O}$ and ${}^{32}\mathrm{O}_2$ (bottom right, referred to number of nuclei). The grey dot-dashed, solid, and dashed contours indicate $\delta {}^{18}\mathrm{O}=-25.0,0.0$ and $+25.0‰$. Underlying concentrations of O(${}^3\mathrm{P}$), ${\mathrm{O}}_2$ ($X^3\mathrm{\Sigma }\stackrel{{}^{-}}{{}_{\mathrm{g}}}$) and ${\mathrm{N}}_2$ ($X^1\mathrm{\Sigma }\stackrel{{}^{+}}{{}_{\mathrm{g}}})$ are from the MSIS 2 model [43] for 2021 February 24. The isotopic enrichment of the bulk (i.e., atomic and molecular) oxygen composition is set to $\delta {}^{18}\mathrm{O}=24.15‰$ [30].

Figure 8

Top: Comparison of the probability distributions of the optical depth ratios $\tau \left({}^{18}\mathrm{O}\right)/\tau \left({}^{16}\mathrm{O}\right)$ resulting from the Monte Carlo simulation of the direct fit to the 2021 February 24 data, in the ${}^3{\mathrm{P}}_1$ $\leftarrow {}^3{\mathrm{P}}_2$ and ${}^3{\mathrm{P}}_0$ $\leftarrow {}^3{\mathrm{P}}_1$ lines (orange and blue, respectively, and labeled to facilitate monochrome visualization). Bottom: Approximation of the distribution of the 256 variates from the Monte Carlo test by a normal (Gaussian) distribution (${}^3{\mathrm{P}}_1$ $\leftarrow {}^3{\mathrm{P}}_2$ line, same SOFIA flight).