Organic Hazes as a Source of Life's Building Blocks to Warm Little Ponds on the Hadean Earth

Atmospheric simulations were performed with the Planetary Haze Research (PHAZER) experimental setup (He et al. 2017). PHAZER is a vacuum flow system and stainless steel chamber used to simulate chemistry and haze production in planetary atmospheres. Within the chamber is the option to attach different energy sources. For this work we attach two electrodes that allow us to produce a cold plasma discharge: a laboratory analog to short-wave upper atmospheric ultraviolet light (≲110 nm; Cable et al. 2012; Pearce et al. 2022a). For a schematic of the PHAZER setup, see Pearce et al. (2022a).

Our experimental protocol is similar to previous PHAZER studies (He et al. 2018a; Hörst et al. 2018a; He et al. 2018b, 2019, 2020b, 2020a; Moran et al. 2020; Pearce et al. 2022a). We start by cleaning the chamber and flanges using Alconox^® detergent powder and a scouring pad. We then use a fine-grained (1500-grit) sandpaper to clean any remaining stuck-on solids from the chamber and flange walls and the electrodes. The tight space between the flange and the electrodes is cleaned by scraping and twisting a wet paper towel into the area with a spatula repeatedly until the paper towel comes out clean. After a visual analysis shows that no solids are present on the chamber walls, flange walls, and electrodes, these pieces are put in an ultrasonic cleaner for 3 × 30 minutes. We then rinse the chamber, electrodes, and flanges with tap water, then HPLC-grade water, and then methanol. Next, we put all the chamber pieces except the flange with the electrodes in the oven overnight at 80°C. We visually inspect the chamber again to make sure no discoloration is present on the chamber pieces before assembling the chamber while wearing nitrile gloves. New copper gaskets are used to seal all flanges. Lastly, valves near the entry and exit points to the chamber are cleaned with methanol and cotton swabs.

We prepare the gas mixtures using high-purity gases (H₂-99.9999%, N₂-99.9997%, CO₂-99.999%, and CH₄-99.999%; Airgas). We vacuum and purge the lines with N₂ multiple times before running our gas mixtures. We then turn on gas flow at 5 standard cubic centimeters per minute and initiate the 170 W m⁻² cold plasma source by applying a voltage differential of 6000 V between the electrodes. The cold plasma discharge produces electrons and ions roughly in the 5–15 eV range and is energetic enough to directly dissociate H₂, CO₂, CH₄, and N₂. The plasma does not significantly affect the neutral gas temperature, which remains approximately at room temperature. The exposure time of the gas to the plasma is approximately 2 s (He et al. 2022).

Experiments are run for approximately 6 days until the mixing tanks are empty. The chamber pressure gauge varies from 2 to 3.5 torr during that time. The readout variation is expected for H₂-dominant gas mixtures, as these pressure gauges are calibrated for direct readout of nitrogen or air. Gas flow and plasma are then turned off, the chamber valves are sealed off, and the chamber is transferred into a dry (<0.1 ppm H₂O), oxygen-free (<0.1 ppm O₂) N₂ glove box (I-lab 2GB; Inert Technology). Here, samples of solid organic haze particles are collected from the walls and base of the chamber with cleaned and heat-sterilized spatulas and stored in small glass vials that were not previously used. Vial caps are screwed on and sealed with parafilm. No controls were produced at this stage.

Organic hazes are produced for two different Hadean atmospheric compositions: high methane content and low methane content. We collected 522 mg of solid haze particles from the high-methane experiment and 72 mg from the low-methane experiment. The compositions of the gas mixtures are H₂:N₂:CO₂:CH₄ = 89:5:1:5 and 93.5:5:1:0.5 for the high- and low-methane experiments, respectively.

Given the potency of methane as a greenhouse gas, the high (5%) methane atmosphere may have had surface temperatures above the boiling point of water. Therefore, as an initial study of the effects of high temperatures on our samples, we take two portions of the organic haze particles collected from the high methane experiment and heat them at 200°C in an oven for 1 and 7 days, respectively. To do this without exposing the sample to oxygen, we place the glass vial (with cap off) containing the organic haze particles in a Fisherbrand heavy glass desiccator while in the N₂ glove box and lubricate the contact between the dessicator base and lid with high-vacuum silicon grease (rated to 204^◦C) to make an airtight seal. We then transfer the desiccator to the oven for heating. Once heating is complete, we transfer the dessicator back to the glove box to remove the sample and store it for later biomolecule analysis.

Biomolecule analyses are performed on an Agilent 7000c triple quadrupole gas chromatograph/mass spectrometer (GC/MS) using a highly selective and sensitive GC/MS/MS protocol developed for detecting nucleobases and amino acids in organic haze particles (Sebree et al. 2018). GC/MS/MS differs from regular GC/MS in that the first quadrupole traps only the selected parent masses within each chosen range of retention times. Then, the ions are transferred to the second quadrupole, where they are fragmented by collision with neutral gas. Finally, the fragment ions are passed along to the third quadrupole, where only two masses associated with the fragmentation (quantifier and qualifier) are allowed to pass through the detector. The benefit of this technique is the increased signal-to-noise ratio (S/N) that arises from (1) the preconcentration of the species of interest at the first quadrupole, (2) the suppression of other species with the same GC retention time as the species of interest (as they are ignored at the first quadrupole), and (3) the further suppression of noise by only allowing two daughter masses from the intended parent mass to pass through the detector. This method differs from an extracted ion chromatogram, for which quantitation ions are extracted from a total ion chromatogram (TIC) containing all masses. The GC/MS/MS protocol instead filters out all masses that we are not looking for (in each gate) prior to detection using the three quadrupoles. In the case where there are two species in the same gate with the same retention times, e.g., uracil and proline (gate 10), we calculate the relative peak percentages using the relative intensities of the quantifier peaks in the mass spectra (see Figure A1). This procedure is best suited to analyze complex samples like ours that require higher sensitivity. The retention times for the GC/MS/MS gates were found and developed in Sebree et al. (2018) by running amino acid and nucleobase standards through regular GC/MS and analyzing at the mass spectra for each GC peak. We refer the reader to Sebree et al. (2018) for more information about that procedure.

We first carry out GC/MS/MS validation, limit of detection, and calibration curve development on two sets of standards: (1) a physiological amino acid standard (Sigma Aldrich) with 27 detectable components at 2.5 μmol mL⁻¹ biomolecule⁻¹, and (2) a nucleobase standard (Sigma Aldrich, all 99.0%) containing seven nucleobases. The seven nucleobases are dissolved in a 0.1M solution of NaOH at a concentration of 0.5 mg mL⁻¹ nucleobase⁻¹. To find limits of detection and quantification and produce calibration curves, dilutions of $\tfrac{5}{100000}$, $\tfrac{20}{100000}$, and $\tfrac{200}{100000}$ are made of each standard to provide a range of base concentrations to sample different volumes from for GC/MS/MS analysis.

The standard sample preparation procedure is as follows: Multiple volumes in the range of 5–50 μL of each base standard solution are added to separate GC vials and dried in the oven at 40°C under N₂ flow. After all visible liquid evaporated, vials are washed with 100 μL of CH₂Cl₂ and dried again at 40°C under N₂ flow. Next, 30 μL of dimethylformamide (DMF) and 30 μL of N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide (MTBSTFA) are added to the dried GC vials, and the solutions are left to derivatize for 30 minutes under N₂ flow at 80^◦C. Finally, 100 μL of CH₂Cl₂ is added to the solutions, and 0.5, 0.8, 1, or 2 μL of the solution are injected into the GC triple quad running in MS/MS mode for analysis. We do multiple injection volumes in order to get several data points for our calibration curves for a single sample concentration. Each sample takes approximately 25 minutes to run.

Separation of compounds is performed using a Restek capillary column (RTX-5MS) with the helium flow held constant at 1.3 mL minute⁻¹. The operating temperature is 100^◦C, ramping up at a rate of 10°C minute⁻¹ to a final temperature of 270°C, held for 11.5 minutes. Collisionally induced dissociation energies of 10–50 eV are used during MS/MS operations in the MRM scan mode.

In Figure A2 we display the GC/MS/MS chromatograms for nine nucleobase standard concentrations, and in Figure A3 we display the GC/MS/MS chromatograms for nine physiological amino acid standard concentrations. We ran the nucleobase standards first, and it became apparent once running some subsequent tests with methanol and an MTBSTFA derivatizer that underivatized nucleobases from the high-concentration runs stick around in the GC inlet. This is to be expected, given that (a) no derivatization protocol yields 100% derivatized biomolecules and (b) MTBSTFA derivatization is a reversible reaction. This hypothesis is validated by two sets of tests. First, we find that nucleobase peaks do not show up in the methanol runs but do show up in the derivatizer runs (see Figure A4). Second, we test a washing protocol for the GC inlet, whereby we track the decrease in the nucleobase peaks after subsequent GC/MS/MS runs with an MTBSTFA derivatizer. If there are underivatized nucleobases leftover in the GC inlet, the MTBSTFA would derivatize them, allow them to vaporize in the inlet, and enter the GC column. This would result in a reduction of the nucleobase peaks with each wash run. In Figure A5, we display the results for the hypoxanthine peak across 17 wash runs. We see that the hypoxanthine peak decreases with every subsequent wash.

Following this discovery, we initiate a new protocol to wash the GC inlet twice with a procedural blank containing MTBSTFA before any individual GC/MS/MS run on organic haze particles. We subtract the second blank chromatogram from the haze particle chromatogram for analysis. We find that there is no retention time difference between haze particle and blank runs. The linearity of our calibration curves suggests that our protocol can be used for quantification, as inconsistent levels of MTBSTFA derivitization or degradation would show up as nonlinearities in the calibration curves.

Peak areas are integrated using the average noise floor for each GC gate (see Table A1 for the list of GC gates). Uncertainties are calculated by integrating the peak areas with noise floor adjusted by ±1σ. We wrote a Python code to automatically perform these calculations for preselected GC gates and noise regions (see Pearce 2023a). The code outputs an interactive visualization of the integrated peaks, allowing the user to validate the retention times and S/Ns. Peaks falling below an S/N of 3 are screened out in this process. See Figure A6 for a screenshot of the Python code and an example peak area calculation.

In Figures A7 and A8, we display the calculated calibration curves for the nucleobases and amino acids/other biomolecules, respectively. Calibration curves allow us to quantify the concentrations of each of these species in our organic haze particles. Calibration curves are calculated by plotting the integrated GC peak areas from our standards for a range of injected masses for each species and performing a least-squares linear fit to the data. Injected masses are calculated using the concentration of each species in the standard, the volume of the standard put into the GC vial, the volume by which it was diluted during derivatization/preparation, and the volume injected from the GC vial into the GC inlet. We select three to five data points for the linear fits that are within the range of peak areas found in our haze particles and whose fit has the lowest χ² value. Urea is the only species for which we did not use the lowest detected injection mass from our standards (see Table A2 for details on the S/N of lowest detected injection masses). Uncertainties are included on the calibration curves based on their variance from a model linearly connecting each data point.

Organic haze particles are prepared for analysis similarly to Sebree et al. (2018). First, the organic haze particles are weighed and dissolved in a 50:50 solution of methanol and acetone at 5.0 mg mL⁻¹. The vials are then shaken and centrifuged at 10,000 rpm for 10 minutes, after which they are rotated 180° and centrifuged for another 10 minutes. The supernatants are separated from the solids with a Pasteur pipette and transferred to storage vials. We dry and derivatize 50 μL of the storage solutions using the same methods described above. 1 μL of the solutions are injected into the GC triple quad running in MS/MS mode for analysis. Two blank solutions of 50:50 methanol/acetone are also carried through the same procedure as a contamination check and to provide the derivatizer to wash the GC inlet as described in the washing protocol above.

To calculate the nucleobase concentrations in warm little ponds from organic hazes, we use a comprehensive sources-and-sinks numerical pond model. The model was originally developed in Pearce et al. (2017) to compare meteorites and interplanetary dust as sources of nucleobases to ponds, and it was modified in Pearce et al. (2022b) to include in situ production from atmospherically sourced HCN. The model is based on ordinary differential equations that are solved numerically and correspond to the sources and sinks for both water and nucleobases. Water in this model comes exclusively from precipitation and leaves via evaporation and seepage (through pores in the base of the pond). The precipitation rates we use match the "Intermediate Environment" from Pearce et al. (2017), which produce naturally occurring seasonal wet–dry cycles. Nucleobases in this model are sourced from organic hazes and are removed via photodissociation, hydrolysis, and seepage. We model a 1 m radius and depth cylindrical pond at 65°C to represent a typical pond on the hot Hadean Earth. See Pearce et al. (2017) and Pearce et al. (2022b) for details on rate equations, sources of common rate data, and model schema. The numerical pond model is written in Python, and a release is made available for download (Pearce 2023b).

In this work, we add nucleobase deposition from organic hazes to the sources-and-sinks pond model. The rate for this deposition is calculated using the following equation:

$$\begin{eqnarray}&&\dot{{m}_{N}}=[N]{\dot{m}}_{H}\end{eqnarray} \tag{ 1 }$$

where [N] is the concentration of nucleobase N within the organic haze particles (grams of nucleobase/grams of haze particles) and ${\dot{m}}_{H}$ is the mass deposition of haze particles on the Hadean Earth (g yr^–1).

We use the mass flux proportionality equation empirically derived by Trainer et al. (2006) to estimate the minimum and maximum haze particle deposition rates for the Hadean Earth at 4.4 bya. The equation is

$$\begin{eqnarray}&&{\dot{m}}_{H}=\beta {F}_{{\rm{T}}}\gamma {A}_{\bigoplus }{\left(\frac{{I}_{0,4.4}}{{I}_{0,{\rm{T}}}}\right)}^{n}\left(\frac{{\chi }_{c,4.4}}{{\chi }_{c,{\rm{T}}}}\right),\end{eqnarray} \tag{ 2 }$$

where β is the empirically derived reduction factor of 0.45 from having H₂-dominant conditions and both CH₄ and CO₂ in similar abundances in the atmosphere (DeWitt et al. 2009), F_T is the haze particle flux deposited on Titan's surface (g cm⁻² s⁻¹), γ is the number of seconds in a year, A_⊕ is the area of Earth's surface (cm²), I_0,4.4 is the solar flux on Earth at 4.4 bya, I_0,T is the solar flux on Titan, n = 1 or 2 for the two photochemical pathways of haze production (A or B) derived in Trainer et al. (2006), χ_c,4.4 is the atmospheric mixing ratio of CH₄ on Earth at 4.4 bya, and χ_c,T is the atmospheric mixing ratio of CH₄ on Titan.

For our calculation, we employ the same values as used in Trainer et al. (2006), except two parameters: (1) for I_0,4.4, we multiply the present-day solar flux on Earth by a factor of 20 to account for the (at minimum) 20 times stronger emission in the extreme-ultraviolet range (1–120 nm) at 4.4 bya (Ribas et al. 2005); and (2) we use 0.5% CH₄ for χ_c,4.4, to match our low-methane, potentially habitable Hadean experiment.

Considering exponential n values of 1 or 2, we obtain a minimum haze particle mass deposition rate of 4.9 × 10¹⁴ g yr⁻¹ and a maximum haze particle mass deposition rate of 8.9 × 10¹⁷ g yr⁻¹. These values are approximately two orders of magnitude higher than the organic haze production rates estimated by Trainer et al. (2006) for the Archean Earth.

In Figure 1, we display the GC/MS/MS chromatograms for the Hadean haze particle samples from the high- and low-methane experiments, as well as the high-methane Hadean haze particle samples heated to 200°C for 1 and 7 days. Seven nucleobases, nine proteinogenic amino acids, and five other organics (including some nonproteinogenic amino acids) were detected in total. We summarize the detected biomolecules and their concentrations in Table 1.

Zoom In Zoom Out Reset image size

Table 1. Measured Mass Fractions of Nucleobases, Amino Acids, and Other Organics in Four Hadean Earth Experiments

Biomolecule	High Methane	Heated High Methane (1 day)	Heated High Methane (7 days)	Low Methane
Nucleobases
Xanthine	${62}_{-1}^{+3}$	${17.8}_{-0.1}^{+5.3}$	⋯	⋯
Hypoxanthine	${47}_{-1}^{+2}$	${18.5}_{-0.3}^{+0.2}$	⋯	${20.8}_{-0.0}^{+0.6}$
Cytosine	${19.2}_{-0.3}^{+0.5}$	${2.9}_{-0.9}^{+0.1}$	⋯	${3.3}_{-0.6}^{+0.1}$
Guanine	${14.8}_{-1.6}^{+2.7}$	${10.8}_{-4.5}^{+0.3}$	⋯	⋯
Uracil	${5.8}_{-0.2}^{+0.7}$	${3.0}_{-1.5}^{+0.0}$	${2.8}_{-1.5}^{+0.0}$	${3.8}_{-1.7}^{+0.1}$
Adenine	${2.1}_{-0.2}^{+0.5}$	${1.58}_{-0.14}^{+0.04}$	⋯	${4.7}_{-0.1}^{+1.6}$
Thymine	${1.8}_{-0.5}^{+0.1}$	${1.6}_{-0.1}^{+0.0}$	⋯	${1.7}_{-0.3}^{+0.0}$
Proteinogenic Amino Acids
Tryptophan	${\lt 50}_{-0}^{+4}$	⋯	⋯	⋯
Threonine	${6.7}_{-0.4}^{+0.1}$	⋯	⋯	${8.8}_{-1.9}^{+1.9}$
Serine	${5.5}_{-0.3}^{+0.4}$	⋯	${6.8}_{-1.3}^{+0.1}$	${6.0}_{-0.2}^{+0.2}$
Alanine	${3.1}_{-1.4}^{+0.6}$	${1.1}_{-0.4}^{+0.0}$	${2.0}_{-0.7}^{+0.0}$	${5.0}_{-1.8}^{+0.1}$
Valine/norvaline	<2.6 ${}_{-0.0}^{+1.2}$	${0.4}_{-0.1}^{+0.9}$	${1.5}_{-0.1}^{+1.0}$	<4.1 ${}_{-3.3}^{+8.0}$
Proline	${2.5}_{-0.4}^{+0.1}$	${1.09}_{-0.04}^{+0.06}$	${1.72}_{-0.02}^{+0.03}$	${1.34}_{-0.20}^{+0.04}$
Glycine	${2.1}_{-0.3}^{+0.1}$	${0.7}_{-0.0}^{+0.2}$	${0.63}_{-0.01}^{+0.08}$	${33}_{-1}^{+6}$
Isoleucine/norleucine	<0.54 ${}_{-0.03}^{+0.03}$	<0.32 ${}_{-0.02}^{+0.03}$	${0.23}_{-0.02}^{+0.02}$	${2.1}_{-0.2}^{+0.4}$
Leucine	${1.5}_{-0.2}^{+0.5}$	${0.24}_{-0.01}^{+0.01}$	${0.15}_{-0.01}^{+0.02}$	${2.8}_{-0.2}^{+0.2}$
Other Biomolecules
Urea	${5204}_{-2}^{+2}$ a	⋯	${65}_{-32}^{+0}$	${170}_{-10}^{+13}$
β-alanine	${12.0}_{-2.4}^{+0.1}$	<0.4 ${}_{-0.1}^{+4.8}$	⋯	${7.4}_{-0.1}^{+0.1}$
Sarcosine	<10.8 ${}_{-0.1}^{+0.2}$	<11.1 ${}_{-0.2}^{+1.9}$	⋯	${23}_{-0}^{+3}$
Ethanolamine	${2.2}_{-0.1}^{+0.0}$	<0.27 ${}_{-0.01}^{+0.18}$	⋯	⋯
β-aminoisobutyric acid	<0.33 ${}_{-0.04}^{+0.40}$	<0.33 ${}_{-0.03}^{+0.42}$	⋯	<2.6 ${}_{-0.3}^{+1.2}$

Notes. Uncertainties are the combination of GC peak area uncertainties and the calibration curve uncertainties (see the Appendix for an explanation). In some cases, only an upper limit is provided for measurements that fall above the limit of detection but below the limit of quantification (see Table A3 for details). Units are μg biomolecule/g haze particles, or ppm.

^a Based on upper-bound extrapolation of calibration curve.

Download table as:  ASCIITypeset image

In Figure 2, we display the nucleobase concentrations in our four Hadean organic haze particle samples. All seven nucleobases are detected in the high-methane organic haze particles, as well as the haze particles heated for 1 day. Guanine and xanthine are not detected in the low-methane organic haze particles, and only uracil is detected in the haze particles heated for 7 days. Concentrations are highest in the high-methane experiment, with the exception of adenine, for which the low-methane haze particles have the highest concentration. The nucleobases with the highest concentrations in the high-methane experiment haze particles are xanthine, hypoxanthine, cytosine, and guanine, ranging from ∼15 to 62 ppm. Concentrations are lowest for thymine, adenine, and uracil, ranging from ∼2 to 6 ppm for this experiment. The low-methane experimental haze particles range from being 6 times lower to 2 times higher in nucelobase concentrations than the high-methane experimental haze particles. In total, the nucleobase content of the high-methane haze particles is 4.5 times higher than the nucleobase content of the low-methane haze particles. The high-methane haze particles heated for 1 day have 1–7 times lower nucleobase concentrations than the unheated high-methane haze particles. Overall, heating the high-methane haze particles for a day leads to a factor of three reduction in total nucleobase content. All nucleobases are destroyed after 7 days of heating, except uracil, which has the same concentration in the samples heated for 1 and 7 days. This may suggest that uracil reaches a steady state to thermal decay at 200°C after only 1 day.

Zoom In Zoom Out Reset image size

In Figure 3, we display the proteinogenic amino acid concentrations in our four organic haze particle samples. Differences in individual proteinogenic amino acid concentrations between the high-methane experimental haze particles and the low-methane experimental haze particles vary by a factor of ∼1–16. However, the total amino acid content of the high-methane haze particles is only a factor of 1.2 greater than the amino acid content of the low-methane haze particles. Tryptophan is only detected in the high-methane haze samples. Proline is higher in concentration in the high-methane experimental haze particles, whereas glycine, leucine, and isoleucine/norleucine are higher in concentration in the low-methane experimental haze particles. Alanine, serine, valine/norvaline, and threonine have similar concentrations in both haze particle samples.

Zoom In Zoom Out Reset image size

Heating the high-methane haze particles for 1 day destroyed a few amino acids, including serine, threonine, and tryptophan, and reduced the concentrations of glycine, alanine, proline, valine/norvaline, and leucine by factors of 1–13. Interestingly, heating the high-methane haze particles for 7 days only reduced the concentrations of leucine and isoleucine/norleucine from their 1-day heating concentrations. Alanine, proline, and valine/norvaline increased in concentration after the additional heating, and serine was detected after not being present in the 1-day heated haze particles. This may suggest that alanine, proline, valine/norvaline, and serine are thermal decay products of the organic haze particles.

Finally, in Figure 4, we display the nonproteinogenic amino acids and other biomolecule concentrations in our photochemically produced organic haze particles. Urea, ethanolamine, and β-alanine concentrations are highest in the high-methane organic haze particles, and sarcosine and β-aminoisobutyric acid concentrations are highest in the low-methane organic haze particles. After 1 day of heating the high-methane haze particles, urea is no longer detected, and the other four biomolecules either remain the same concentration as the unheated haze particles or decrease by a factor of 8–30. Interestingly, after 7 days of heating, the only detectable biomolecule remaining is urea, which was undetected after 1 day of heating. This may suggest that urea is a longer-term decomposition product of other biomolecules in the organic haze particles.

Zoom In Zoom Out Reset image size

In Figure 5(a), we display the concentration of adenine modeled in a cylindrical warm little pond with a 1 m radius and depth over an 8 yr period, from five different sources. The adenine concentrations from organic hazes were modeled in this work and are based on the potentially habitable low-methane (0.5%) experimental haze particles. The adenine concentrations from meteoritic and interplanetary dust particle (IDP) delivery were modeled in Pearce et al. (2017), and the concentrations from in situ production from atmospherically sourced HCN were modeled in Pearce et al. (2022b). The concentrations from steady sources cycle with a 1 yr period owing to the seasonal wet–dry cycles produced from evaporation, seepage, and precipitation. The flat-top pattern in the concentrations is produced in the dry phase when the adenine influx and photodestruction reach a steady state. Meteorites are single deposition events, and it is improbable that a single pond would have had more than one meteorite deposition (Pearce et al. 2017); therefore, adenine concentrations from these sources are fleeting.

Zoom In Zoom Out Reset image size

The range of adenine concentrations from organic hazes is based on the minimum and maximum haze particle deposition rates of 4.9 × 10¹⁴ g yr^–1 and 8.9 × 10¹⁷ g yr⁻¹ (see Section 2), respectively, as well as the uncertainty in the adenine concentration within organic haze particles, as measured in this study (see Table 1).

Interestingly, organic hazes potentially delivered the highest stable adenine concentrations to ponds at 4.4 bya of up to ∼0.7 μM (∼100 ppb). However, we note that the adenine concentrations from organic hazes during the early Hadean (4.4 bya) agree within uncertainty with the adenine concentrations produced within the pond from atmospherically sourced HCN at the same period. This suggests that both of these sources may have played a part in the origin of adenine in warm little ponds. The pattern of adenine concentration over the year is due to pond level variation (i.e., evaporation and precipitation), as well as the adenine sinks within the pond, i.e., photodissociation, hydrolysis, and seepage through pores at the base. Peak adenine pond concentrations from meteorites are one order of magnitude higher (∼10 μM) than the maximum concentrations from organic hazes; however, meteoritically delivered adenine depletes within weeks, whereas organic hazes are a constant feeding source. As solar activity calms throughout the Hadean, haze production will decrease slightly. Ultraviolet intensity in the 92–120 nm range decreases from 20 times present day to 10 times present day from 4.4 to 4.1 bya. Considering this, we expect at least a factor of 4 decrease in the maximum adenine concentrations in ponds from organic hazes over a 300-million-year period during the Hadean.

In Figure 5(b), we display the maximum modeled pond concentrations of the other six nucleobases in this study sourced from organic haze at 4.4 bya. A total of 8 yr of integration are shown; however, a steady solution is achieved in just 2 yr owing to the balance of sources and sinks. The nucleobases that have the highest pond concentrations are xanthine, hypoxanthine, and guanine at ∼1.8–6.5 μM. We note, however, that the xanthine and guanine concentrations for this plot are based on the high-methane (5%) experimental haze particles, as they were undetected in the low-methane (0.5%) experimental haze particles. Maximum pond concentrations of uracil, cytosine, and thymine are ∼0.21–0.55 μM. Differences in pond concentrations are due mainly to the differences in nucleobase abundances measured in our experimental haze particle samples (see Table 1).