The Only Constant of Life’s Origin Is Change: The Importance of Chemical Fluxes in Prebiotic Environments

Abstract

High concentrations have long been thought to be important in prebiotic chemistry as they offer a way to circumvent a lack of available enzymatic catalysis to overcome kinetic barriers. Here, we argue that fluxes and timescales are also of critical importance. Fluxes and timescales determine, in part, whether an environment can achieve high concentrations of reactants and, in particular, place a critical constraint on whether high concentrations of product molecules can be maintained. We focus on closed basin lakes, which offer a viable way to concentrate molecules relative to background sources under benign conditions. From the perspective of P, HCN and its derivatives, and S, these systems may yield competitively high concentrations of reactants. Nonetheless, closed basin lakes often have limited fluxes of reactants, which places tight constraints on the concentrations of product molecules that can be maintained at steady state. In conjunction with experimentally measured reaction kinetics, an opportunity exists to discriminate between the plausibility of environments on the basis of their simulated ability to generate desired concentrations of products over relevant timescales. Crucially, to make such an evaluation is extremely difficult to do with confidence without quantitatively dealing with fluxes and timescales. Therefore, future work should routinely and systematically consider these aspects alongside molecule concentrations in environmental systems of interest and in experiments.

Keywords

Phosphorus—Cyanide—Sulfide—Prebiotic—Timescales—Fluxes

1. Introduction

The study of the origin of life on Earth exists at the intersection of chemistry, geology, biology, physics, astronomy, and yet more varied fields. Each contributes something unique to the total sum of information we can bring to bear on the problem. These fields fuse together well when pursuing a concrete goal. One important example is the concept that prebiotic chemistry cannot proceed without attaining high concentrations of reactants (Fahrenbach et al., 2017; Monnard, 2016; Walton et al., 2022). Ergo, if we want to understand life’s origins, we must understand which environments had the right reactants at the right concentrations. Such a question naturally calls upon contributions from all the aforementioned fields:

Chemistry: Which reactants and at what concentrations?

Geology: Which environments?

Biology: Difference between prebiotic and modern inhabited versions of environments? Difference between timescales of biological versus prebiotic reactions?

Physics: Detailed understanding of reaction mechanisms to enable modeling of the system?

Astronomy: Impact on environmental chemistry by the early Sun, asteroid impacts and cosmic dust, and so on?

Indeed, the search for high reactant concentrations has a generally sound basis in the following logic: since the origin of life chemistry necessarily preceded the existence of enzymes, at least to begin with, many of the thermodynamically or kinetically unfavorable reactions accelerated by enzymes today must have been originally driven at least in part by high reactant concentrations alongside other factors such as pH, temperature, and pressure (Dill and Agozzino, 2021). Given that reaction yields are rarely 100%, the concentration problem becomes ever more difficult to solve if many reactions are needed to go from simple precursors to functional biomolecules and, crucially, enzymes.

This latter issue is referred to as the arithmetic demon (White and Rimmer, 2024a), wherein a multistep reaction series with low yields at each step will produce a low steady-state concentration of product molecules. If concentration is important for driving the forward reaction, this issue can quickly lead to the forward reaction being inhibited in long reaction chains.

In part searching for an elusive solution to the arithmetic demon, the fields mentioned above contribute each year to assessing the concentrations of the elements, species, and molecules that could potentially have been sustained in prebiotic environments. This information drives an iterative process of chemical and environmental discovery, filtering out environments where the concentrations are considered to be too low and emphasizing environments where concentrations may be high enough. A good example is the finding that phosphate is highly effective as a buffer and catalyst for promoting clean reactions in cyanosulfidic chemistry (Patel et al., 2015), which prompted a search for phosphate-rich environments and led to a focus on P-rich soda lakes (Toner and Catling, 2020), which in turn are emerging as a strong candidate for cyanide accumulation (Toner and Catling, 2019).

This deductive process provides a solid logical basis for multiple fields to work in a coherent way, but is it enough to meaningfully constrain the origins of life? It is our contention that concentrations alone are only one part of a bigger picture when it comes to evaluating environmental scenarios for prebiotic chemistry.

In this review article, we argue that, as a consequence of the dynamic character of prebiotic chemistry, fluxes are at least as important as concentrations. It is sometimes assumed that an environment identified to have a high concentration of a reactant of interest will be a suitable location to host prebiotic chemistry that involves that reactant. However, this assumption is likely to be flawed, as we will discuss in this article.

Throughout, we use the environment of a closed basin lake as a conceptual framework to demonstrate key concepts. We develop a toy box model of said lakes for use in demonstrating key illustrative points. This model is greatly simplified and used to put into context the importance of fluxes, lifetimes, and volumes—which provides a simple quantitative framework in which to discuss key open questions about the prebiotic plausibility of several emerging land-based scenarios for the origins of life.

2. Fluxes in Biogeochemistry

Flux is the change of some quantity per unit time per unit area, which is related to the change of that quantity per unit time per unit volume, or rate. Here, we consider the flux of molecules, Φ (molcm⁻² s⁻¹), and the (related) rate of molecular production or destruction R (molcm⁻³ s⁻¹). The number of examples is almost innumerable, but a particularly good one is the existence of our oxygen-rich atmosphere. In attempting to explain this observation, one must contend with the fact that atmospheric oxygen is highly reactive. If there were no continuous replenishment, the oxygen in the atmosphere would react with iron and other elements in the crust to form oxides and, in time, be completely lost (Lyons et al., 2014). Yet, at the same time, oxygen is not accumulating over time. It is at a quasi-steady state. As much oxygen enters as leaves the atmosphere on long timescales.

The question, then, is why the current steady state exists with 21% oxygen and not 0.21%, or indeed 0.0000000021%? There is proxy evidence that all three of these concentrations have been relevant for Earth’s atmosphere at steady state during the last 4 billion years (Tostevin and Mills, 2020), with movement from one to the next indicating major changes in production and loss terms.

This means that oxygen was not always held in a steady state, either over timescales long enough to perceive slow secular evolution of molecular concentrations or because of instabilities that lead to deviations from steady state. Often, in the latter case, these deviations from steady state shift into a new steady state, because the loss of molecules tends to be proportional to the concentration of those molecules and so will often compensate to balance formation and destruction, especially when rapidly perturbed. To explain why each steady-state regime existed when it did, and to hope to explain why, for each regime, steady state was lost, we must contend with fluxes. When we do so, we see that the options to modify the steady-state oxygen concentration in the atmosphere include:

a higher flux of oxygen into the atmosphere, or;

a lower flux of oxygen out of the atmosphere (since the loss rate must always be the same as production at steady state).

Inevitably, it turns out that a great many other fluxes indirectly determine the oxygen flux in and out. For example, oxygen production fluxes depend on nutrient fluxes such as P weathering, N fixation, and Fe deposition, and oxygen loss depends on fluxes such as volcanic outgassing, hydrogen escape to space, hydrothermal iron precipitation (Alcott et al., 2019; Lenton et al., 2018; Tostevin and Mills, 2020), and the direct production and consumption of oxygen by life (Lyons et al., 2014). In addition, oxygen fluxes are also regulated by photochemistry, which will be affected by the light of the star and the transparency of the atmosphere to that light. Hence, oxygen fluxes on any planet will depend upon molecules that do not directly interact with oxygen at all but that either affect the climate (and therefore the abundance of oxygen-containing condensable molecules, like water, in the atmosphere) (Wordsworth and Pierrehumbert, 2014) or that affect the transport of ultraviolet light through the atmosphere (Rimmer et al., 2019).

Crucially, it is only by taking into account the temporal element of all of the above mechanisms that they can be assembled into a model that satisfactorily explains how particular steady-state regimes are achieved. This applies to oxygen and, at a fundamental level, to modeling most aspects of the Earth system that are not at thermodynamic equilibrium. This includes all life and, therefore, underpins modeling of the biosphere. It also necessarily applies to modeling the prebiotic chemistry that gave rise to the origin of life on Earth.

Many of those studies that seek to understand the Earth system take a box model approach. These models divide complex abstract systems into different reservoirs of defined volume, which can exchange materials over some timeframe. While clearly a simplification, such box models have proven to be powerful tools to explain key aspects of Earth’s changing biogeochemical cycles over time. While there are dangers of overfitting, the models retain predictive power due to the fact that the system in question is so complex. In other words, we have measurements (proxy records) from rocks of numerous model losses, as well as accurate measurements from nature of numerous fluxes, making it difficult to fit all proxy records at once by simply tuning those fluxes that are poorly known (Alcott et al., 2019; Lenton et al., 2018; Tostevin and Mills, 2020).

Indeed, there is increasing confidence in the robustness of the box model approach due to predictions that have been independently confirmed. For example, recent box models of atmospheric oxygenation produce proxy records for oxygen best when there are two step changes in crustal carbonate fraction, one at the time of the Great Oxygenation Event and the other during the early Cambrian Period (Alcott et al., 2024). Studies of Earth’s rock record independently confirm that carbonate fractions did increase at these points in time and by the requisite amount needed in models to drive the levels of oxygenation indicated in proxy records (Walton and Shorttle, 2024). This may yet prove to be a noncausal correlation, but it is an example of the sort of iterative hypothesis testing that box models allow.

The increasingly apparent success of box models in wrestling with the complexity of the Earth system is rooted in constraints offered by observation. It would be impossible, for example, to construct almost any aspect of the box models without first having measured volcanic outgassing rates, hydrogen escape rates, photosynthetic oxygen production rates, and so on. All of those rate measurements involve measuring the timescales and fluxes of natural processes, both abiotic and biological. Though biological processes are often—though not always—faster than abiotic processes, this distinction is immaterial for the mathematical construction of simple box models, which provide one means to link together the relationships between key processes that drive the system being modeled. Notably, the wealth of information in Earth’s rocks allows us to test model predictions against reality regardless of the degree of complexity of the model.

Unfortunately, constraining the relevant environmental parameters and the timescales of key processes in a prebiotic context is extremely difficult. Because there are no rocks from the relevant time of Earth’s history (Harrison, 2009), we must rely instead on analog studies and theory to place constraints on fluxes in our models.

Finally, the timescales that govern prebiotic chemistry are largely unknown, which makes it even more challenging to build box models, let alone to meaningfully constrain them. This is, in part, because not all relevant chemical reactions are yet known and because, of those that are known and thought to be relevant, measurements of their kinetics do not yet exist (White and Rimmer, 2024a). The lack of data is often due to the time-consuming process of measurement, which requires systematically exploring pressure, temperature, ionic strength, pH, and so on. Lacking this information, it is very difficult to model robustly the fluxes that are so essential to determining how prebiotic chemistry will play out in different environments.

Here we argue that obtaining accurate information about fluxes will be critical to developing robust and sophisticated models of prebiotic chemistry and, hence, also to the iterative process of environmental discovery/discarding that has become central to the field. Identifying environments with high (initial) concentrations of particular species will not be enough, since what we ultimately care about is the transformation of those species into prebiotic building blocks and ultimately life. We believe that constraining concentrations, fluxes, and timescales, and then deploying this information in models, will ultimately provide one of the most powerful ways to discriminate between plausible environmental settings for the origins of life.

3. Concentrations Say Little About Fluxes (and Why That Matters)

An illustrative example of the limitations of incorporating only concentrations in models of prebiotic chemistry is the phosphorus (P) problem (Schwartz, 2006). There is broad consensus that P is critical for terrestrial life (Lang et al., 2018), that P is often scarce in terrestrial surficial environments (Schwartz, 2006), and that the evidence from prebiotic chemistry suggests that high P concentrations are needed to drive a host of key reactions—both those in which P is directly taken up and those in which P is instead a catalyst or pH buffer (Islam and Powner, 2017; Liu et al., 2020; Morasch et al., 2019; Pasek, 2019; Patel et al., 2015; Xu et al., 2020). These insights have led geochemists to survey environments where P can accumulate to unusually high concentrations, such as closed basin soda lakes (Haas et al., 2024; Toner and Catling, 2020).

Closed basin soda lakes have by far the highest known P concentrations of any aqueous environment on Earth (Haas et al., 2024; Toner and Catling, 2020). In this way, geochemists have contributed to the ongoing search for environmental scenarios wherein P concentrations are suitably high to be compatible with prebiotic chemistry. However, issues with this approach arise when we move from considering concentrations to fluxes.

Consider a hypothetical closed basin lake that sits atop a completely impermeable rock body, that is, a lake into which water flows but from which there is no outflow (Fig. 1). Inflow of water to the reservoir is balanced solely by evaporation, which leads to accumulation of nonvolatile species, for example, P (Fig. 1). In our (toy) box model of this lake, it is at steady state with respect to its volume (V, liters). If one considers only the concentration of P in the lake, then it depends solely on the time it takes to export a given atom of P from solution into the long-term sedimentary sink, which depends upon the efficiency of all of the internal P sinks (Fig. 1).

FIG. 1.

Schematic illustration of a simple box model of a closed basin lake. The input flux of water (from multiple sources) is balanced entirely by the outflow flux (evaporation). The inflow of water (with its suspended and dissolved components) results in a production flux of, for example, a type of organic molecule that—in a prebiotic world—is then subject to a variety of forward reactions, hydrolysis, and sedimentation sinks. When the concentration of that organic molecule is at steady state, the sum of all of its sink fluxes must be equal to its production flux.

A given P atom will spend a long time in solution if internal P sinks (which export from solution to sediment) are inefficient and vice versa. We quantify this as follows. The time spent in solution is the residence time (in years) or, for an unstable prebiotic molecule, the lifetime (again in years), both represented by τ. The production or inflow rate of the chemical species (i) is R_i (mol L⁻¹ yr⁻¹). We can calculate the steady-state concentration (mol L⁻¹) from the rate as follows:

C_{i} = R_{i} τ_{i} .

(1)

For a fixed P production flux, for a fixed volume lake, when P (sum of all species) concentration is at steady state, we can see that a high P concentration demands a high residence time and vice versa (Fig. 2). Simply by changing the efficiency of internal P sinks, the system can switch from having a very high to a very low steady-state P concentration. Such a change may occur due to a shift in temperature or pH, promoting some reactions and hindering others. Yet, despite that shift, the P fluxes through the system are identical in both end-member regimes when the system is at steady state.

FIG. 2.

Schematic illustration of the relationship between the residence time of phosphorus (P) (in all inorganic species; likely to be dominated by orthophosphate) and its steady-state concentration. More efficient uptake of P by prebiotic chemistry will result in a shorter residence time and a lower steady-state concentration. The concept of residence time, which is common in the biogeochemistry literature, is mirrored in the prebiotic chemistry literature by the concept of a species lifetime.

In this way, we can immediately see how steady-state concentration is only part of a complex set of factors that govern how prebiotic chemistry will proceed in an environment. A high steady-state concentration may be maintained with minimal P fluxes, as in small seasonal soda lakes such as Last Chance Lake (Haas et al., 2024, 2025; Toner and Catling, 2020), though it might not necessarily be coupled to a high prebiotic productivity, i.e., flux of molecules built by incorporating that P (Fig. 3).

FIG. 3.

Schematic illustration of how production fluxes, reaction/sink efficiency, and molecule residence times/lifetimes interact. (a) Low production fluxes can give rise to high reactant molecule concentrations if internal sinks (such as forward reactions) are inefficient. Yet, since input fluxes must equal losses at steady state, the productivity fluxes of the product molecule will be low. This will result in low product molecule concentrations if the molecule has a short lifetime. (b) The same production flux can give also give rise to low reactant molecule concentrations if the internal sinks are efficient. If there is no kinetic barrier to forward reactions, product molecule fluxes, and concentrations may be the same as in scenario (a). (c) An idealized scenario is one where production fluxes are high, internal sinks are inefficient, and product molecule lifetimes are long. These interactions are explored quantitatively in Figures 5–8.

In contrast, a lake with a somewhat lower P concentration, if that concentration is sufficient to drive prebiotic productivity, may be capable of sustaining much higher concentrations of product molecules due to the high fluxes of P into the system relative to its volume (Figs. 3 and 4) (Walton et al., 2025). In this way, the high throughput of prebiotically relevant material through the system plays an important role in achieving the desired result of a high production flux of product molecules. By analogy to the concept of biological productivity, which is a key factor of many aqueous environments studied today, we refer from here-on-in to the production flux as “prebiotic productivity” (units of mmol/(m³ yr) per unit volume, or mmol/(m² yr) per unit area).

FIG. 4.

(a) Aerial photograph of Mono Lake (CC2.0, credit to Ron Reiring). (b) Aerial photograph of Last Chance Lake (copyright Google Earth).

FIG. 5.

A tale of two lakes—concentrations versus fluxes in Mono Lake and Last Chance Lake. (a) The same as Figure 2, but with real data for the hydrologic and chemical parameters [see Walton et al. (2025) for details]. (b) Phosphorus (P) productivity fluxes in each lake. Corrections are made for a prebiotic scenario where organic uptake is the sole internal P sink [see Walton et al. (2025) for details].

Overall, the steady-state P concentration is controlled by fluxes and timescales. Since prebiotic chemistry may operate over short timescales, and the maximum P fluxes in nature are limited, this places great emphasis on optimizing environmental dynamics to obtain high P concentrations without relying on extremely sluggish timescales.

4. Real-Life Examples and Their Prebiotic Analogs

We now consider some existing lakes of possible relevance as prebiotic analog environments. We consider two alkaline soda lakes that have recently been studied in a prebiotic context (Haas et al., 2024, 2025; Toner and Catling, 2019, 2020; Walton et al., 2025). The interest in these dry-land-hosted scenarios for the origin of life comes principally from the fact that they are restricted or “closed” basins, meaning that nonvolatile elements or species that arrive in the lake will tend to accumulate there. This provides a simple concentration mechanism that might be able to overcome the noted problems of generally low environmental concentrations for key species on the prebiotic Earth (Schwartz, 2006).

Beyond this, these closed basin lakes are indeed currently enriched in P. This means that the possibility of P-rich prebiotic closed basin lakes is not merely theoretical. We have observed modern analogs, which demonstrate that the geochemical parameter space can be realized in nature. The question, however, is whether prebiotic equivalents would host similar concentrations of P, less, or perhaps even more? Furthermore, as we will discuss, it is also crucial to understand whether the fluxes involved would be sufficient to sustain prebiotic chemistry and initiate life.

Mono Lake (ML) is an extremely large, closed basin soda lake in California, USA, that plays host to exceptionally high levels of productivity, that is, fluxes of organic matter (in which P is incorporated). Phosphorus concentrations in ML (all inorganic species, though dominated by orthophosphate) are also relatively high (0.6 mM) (Walton et al., 2025). In contrast, we consider Last Chance Lake (LCL) in British Columbia as an example of a low productivity lake with exceptionally high P concentrations (varying from 1 to 37 mM during a seasonal evaporative cycle) (Haas et al., 2024, 2024; Toner and Catling, 2020).

These are informative examples to focus on because they have received extensive previous study and because they represent two different forms of a highly competitive environment for prebiotic chemistry, with the main difference between the two being productivity, that is, organic molecule fluxes and seasonality (i.e., volume variability throughout the year).

The fluxes, volumes, and observed concentrations of P in each lake indicate that ML has a P residence time of only 10 years, while the equivalent for LCL is on the order of 1000 s of years. This outcome is consistent with recent findings (Haas et al., 2025) (Fig. 5). Naively, we can infer that LCL only has such high P concentrations because it has a slow internal P cycling rate due to limited biological activity. Comparing fluxes, we can see that ML has around a 100-fold higher P productivity than LCL (Fig. 5b).

At face value, this suggests that in a prebiotic scenario in which organic P extraction is of equal efficiency in each lake, ML would have higher fluxes of prebiotic molecules as well as higher P concentrations. In other words, more P is arriving per unit time, such that the production flux of prebiotic molecules must be higher (i.e., assuming the threshold concentration of P is reached to achieve the chemistry that started the prebiotic forward reaction of this element initially). However, there are several important caveats to this comparison that should be considered.

First, we must account for the fact that some P in ML is from avian vectors. This lowers the P we can expect to enter a prebiotic analog of this system by just over half (Walton et al., 2025) (Fig. 5b). Second, since LCL has a biological productivity that is limited by some factor other than P availability, we must normalize productivity in the two lakes such that organic uptake of P is equal in both (Fig. 5b) (Haas et al., 2025). This correction greatly increases the relative productivity of LCL. Finally, to get an idea of the productivity that can be achieved without biology, we must account for the absence of oxidative recycling of organic matter, which decreases the estimated productivity in both lakes (Fig. 5b) (Kipp and Stueken, 2017).

After completing the above corrections (Fig. 5b), we find that these two lakes are not so different after all. The total productivity gap between the two decreases to only one order of magnitude instead of two (Fig. 5b). Given the finding of Haas et al. (2025) that limited biological P uptake plays a major role in explaining the millennial-scale residence time of P in LCL, we can make the reasonable assumption that if biological P uptake is governed in LCL, as it is in ML, then the residence times of P will be much shorter. If P residence times are assumed to be equal in the two lakes under those conditions, a prebiotic LCL should still have elevated average P concentrations compared with a prebiotic ML (Fig. 5a). However, as noted above, the total prebiotic P productivity remains an order of magnitude lower.

The exact prebiotic steady-state P concentration in these lakes will depend on the timescales of prebiotic P uptake, as captured by the residence time (in years; Fig. 5). Again, without enzymes, we can expect this to be much less efficient than biological uptake, which leads to long residence times and high concentrations. It is reasonable, therefore, to consider that the saturation ceiling of P may come into play in these lakes. In that case, an ML scenario will struggle to accumulate much more P than the 0.6 mM currently observed (Walton et al., 2025), whereas we know that LCL scenarios are capable of reaching accumulations at least 50-fold higher than this (37 mM highest recorded) (Haas et al., 2024).

Overall, we can see that P cycling in these lakes represents two sides of the same coin. ML and LCL largely explore different regions of the same shared parameter space that governs P concentrations in closed basin lakes. If we are optimizing for total productivity and can accept only 1 mM P concentrations (Gull and Pasek, 2013) for prebiotic chemistry, then ML is superior, while if we require higher than 10 mM P to drive prebiotic chemistry, then LCL is necessarily the winner.

A final point to consider is whether total productivity is as important as relative productivity. ML is around 250-fold deeper than LCL. When total productivity is normalized for volume, we can expect a prebiotic LCL to be almost one order of magnitude more productive than ML. This observation may swing the balance of probability back in favor of LCL—at least from the perspective of P concentrations and productivity fluxes.

Then again, do we need to choose? It remains possible that in a prebiotic context, some ideal situation exists halfway between the two, where low volumes, high P saturation ceilings, and high P productivity can be combined in a single environment. With these points in mind, it is now worth pondering in more depth just how “high” prebiotic productivity needs to be during the origins of life.

5. How “High” Does Prebiotic Productivity Need to Be?

What defines a “high” prebiotic productivity depends largely on what is capable of defeating the arithmetic demon. The criteria are stringent. We must:

maintain the right concentration of reactants to continually drive the forward reaction;

sustain a high enough prebiotic productivity to maintain a high concentration of product molecules, the necessary value of which will be set by (i) the lifetime in solution of the product molecule, as set by destructive sinks such as hydrolysis, photolysis, and so on; and (ii) the concentration of the product molecule needed to drive the next step in the reaction chain.

Indeed, the prebiotic productivity in the initial stages of the reaction network must be capable of driving forward reactions throughout the entire chain. In this way, the necessary productivity is set by that which achieves the “final” step, whatever that may be. The more reactions that must be linked together, the higher the prebiotic productivity required in step one to obtain the same yield of the final product. These requirements emphasize the fundamental role of timescales (fluxes and lifetimes) alongside concentrations in determining whether an environmental scenario can plausibly host sufficiently productive prebiotic chemistry.

We have laid out above the theoretical case for why high and low concentrations of P on their own offer little insight into how P-dependent prebiotic chemistry will actually occur in the environment of interest. To discover this, we must consider species lifetimes and prebiotic productivity, the latter of which depends critically on fluxes.

We now model reactant and product molecule concentrations maintained in ML and LCL scenarios for molecules built by using not just P but also cyanoacetylene, hydrogen sulfide, and ferrocyanide (Liu et al., 2021; Rimmer and Shorttle, 2019, 2024; Todd et al., 2022). The model considers some production flux into a closed basin lake of some volume (V). At steady state, we can consider:

F_{production} = F_{forward} + F_{hydrolysis},

(2)

where F_production is the production flux, F_forward is the forward reaction flux, and F_hydrolysis is the flux term for the hydrolysis sink (units of moles/yr for each). This is a simplified toy model. In nature, there will be more sinks than these two alone.

This is a toy model, intended to illustrate concepts rather than derive realistic estimates of product molecule concentrations. Therefore, we explore a simplified scenario in which hydrolysis and forward reaction for a given species are always set to be equal. This assumption allows us to calculate:

F_{forward} = \frac{F_{production}}{2},

(3)

where the flux of a specific product molecule (i) for some reaction yield (j, value between 0 and 1) at a reaction step (x) is then:

F_{i}^{^{j, x}} = \frac{F_{production}}{2} j .

(4)

We can further modify the above equation to account for the product flux of a species i at the next reaction step in the series (x + 1), as follows:

F_{i}^{j, x + 1} = \frac{F_{i}^{j, x}}{2} j .

(5)

In each case, the steady-state concentration of a molecule in the system will depend on the production flux (or forward reaction flux) and the molecule lifetime (years), as laid out in Eq. (1).

Figures 6 –8 collate flux and timescale data for these species from prior studies (Table 1) and deploy them into the box model outlined in Figure 1 using Eqs. (1)–(5).

FIG. 6.

Steady-state concentrations of P and product molecules formed by P-uptake in Mono Lake (ML) and Last Chance Lake (LCL) scenarios. Observed P concentrations in each lake today are plotted. LCL is a seasonal lake, therefore not fully conforming to the steady-state assumptions of the toy model. The effects of evaporation in the dry season are shown as a 37× multiplier on the wet season toy model output, based on the maximum 37 mM P observed in the dry season.

Table 1.

Literature Values For Fluxes of Prebiotically Relevant Species, Used In Box Model Calculations

Flux type	Receiving reservoir	Species	Value error	Lower estimate	Upper estimate	Units	Reference
Atmospheric deposition	Oceans	NOx		2.5E+05	6.5E+08	cm-2 s-1	Ranjan et al (2019)
Outgassing into atmosphere	Atmosphere	SO2	2E+10	1E+10	3E+10	cm-2 s-1	Ranjan et al (2018)
Outgassing into atmosphere	Atmosphere	H2S		3.1E+08	7.7E+09	cm-2 s-1	Ranjan et al (2018)
Post-impact atmosphere rain	Earth surface	HCN	1E+09			cm-2 s-1	Wogan et al (2023)
Post-impact atmosphere rain	Earth surface	HCCCN	1E+09			cm-2 s-1	Wogan et al (2023)
Post-impact atmosphere rain	Earth surface	NO	1E+08			cm-2 s-1	Christensen et al. (2024)
Atmospheric rainfallout	Earth surface	HCN	1E+07			cm-2 s-1	Tian and Zahnle (2011)

We consider the following criteria for a “successful” environmental scenario for prebiotic chemistry:

concentration of initial species of 1 mM and

concentration of final product molecules of 1 mM.

Figures 6–9 emphasize that the degree to which closed basins are an effective concentration mechanism for a given species depends strongly on the molecule lifetime (Figs. 2, 3). While our toy model does not attempt to account for sinks other than hydrolysis and forward reaction, the concentrations we naively obtained indicate that closed basins with constant inflow can be efficient at concentrating initially dilute molecules of prebiotic interest, though only if lifetimes are on the order of hundreds of years or much more.

5.1. Phosphorus

Consistent with the results in Figure 5, we find that LCL is around 1 order of magnitude more efficient at concentrating P than ML for a given P lifetime in solution, despite having a lower estimated prebiotic P productivity flux (Fig. 5). This effect is due to its shallow depth compared with ML (250-fold shallower), such that when productivity is normalized by per unit volume instead of by per unit surface area, a prebiotic LCL is roughly 1 order of magnitude more productive than an equivalent ML scenario (Fig. 5).

Running through the results for the simulated series of forward reactions, we can see that for a product formed after four steps in a reaction network, we require a species lifetime on the order of 10,000 years in ML and 1000 years in LCL to achieve 1 mM concentrations. The effect is even stronger in the dry season (Fig. 6), when volumes in the lake decrease rapidly, but it will apply only for a short period of time before the system fully dries out.

We can conclude that, at least for P, LCL is a strong candidate for sustaining high productivity fluxes and steady-state concentrations of P and product molecules. Millennia-scale lifetimes for P also appear quite reasonable for prebiotic chemistry (Haas et al., 2025). However, even for this exceptional environment, long molecule lifetimes are required to achieve 1 mM concentrations of any product molecules. It is not obvious, however, that this is a reasonable expectation given that hydrolysis and photochemical sinks are also efficient for many P-bearing prebiotic molecules (Farr et al., 2023).

5.2. Cyanoacetylene

Based on available information for riverine inflow into ML, we estimate that 10,000-year lifetimes are required to reach 1 mM concentrations for cyanoacetylene (HCCCN). For this same lifetime, HCCCN in LCL is only 0.01 µM (Fig. 7). This is an offset of 5 orders of magnitude. However, the gap can be closed somewhat in the dry season for an LCL analog, where concentrations are enhanced due to the high degree of evaporation (Fig. 7c).

FIG. 7.

Steady-state concentrations of HCCCN and product molecules formed by the HCCCN forward reaction in ML and LCL scenarios. A 37× effect of evaporation in the dry season is shown for LCL output. HCCCN, cyanoacetylene.

FIG. 8.

Steady-state concentrations of HCCCN and product molecules formed by the HCCCN forward reaction in ML scenarios considering a reducing CN-rich hydrothermal production.

The offset in HCCCN concentrations between the two scenarios becomes even more stark if we consider that around 10% of water production to ML each year is hydrothermal (Fig. 8). If we model this as a reducing hydrothermal source, as in Rimmer and Shorttle (2019), where inflows have 1 mM HCCCN, ML will accumulate 1 mM HCCCN with lifetimes of only 100 years, increasing the offset to 7 orders of magnitude versus the accumulation of 1 mM HCCCN in LCL, which is fed by atmospheric deposition (Fig. 8).

A lifetime of 100 years is long for HCCCN. It has been shown that HCCCN will hydrolyze with a half-life of only 10 days at pH 9 and 30°C (Ferris et al., 1968). The half-life is lower at higher pH and temperatures but much longer (roughly 1000 days) at neutral conditions. Lower temperatures will also extend the half-life. ML approaches 0°C in winter but is as warm as 20°C in the summer. Therefore, a direct ML scenario will have a short half-life for HCCCN hydrolysis. It is possible that a prebiotic analog would have closer to 1000-day half-lives if the lake were more acidic, for example, due to being in equilibrium with an atmosphere of higher carbon dioxide content (Toner and Catling, 2020).

Considering our simple forward reaction model, no atmospheric deposition model can achieve 1 mM product molecule concentrations derived from HCCCN at step 3 with lifetimes shorter than in the millions (Fig. 7). Hydrothermal sources will much more easily sustain 1 mM product concentrations (Fig. 8), yet even in this case the yields must be high if inflows are only moderately concentrated (1 mM). If yields are low (Fig. 8) under these circumstances, million-year lifetimes are again required. If hydrothermal sources are highly concentrated (90 mM), comparatively short lifetimes are required even for low-yield reactions at step 3 (1000 s of years) and decrease to 10 s of years for high-yield reactions (Fig. 8).

5.3. Ferrocyanide

If we assume that all HCN that enters an ML-type system reacts over time to form ferrocyanide, we can also roughly estimate the ferrocyanide reservoir that may accumulate. Given a conservative HCN input through direct atmospheric deposition and via rivers (Table 1), ML would accumulate 0.2 mM concentrations of ferrocyanide only if the lifetime of the species was on the order of 100 years. If lifetimes remain at 100 years, 1000 mM ferrocyanide is achieved from an initially 90 mM HCN hydrothermal source that entirely converts to ferrocyanide.

The lifetime of ferrocyanide in aqueous alkaline solution is not well constrained. However, it is known that ferrocyanide will react quickly when exposed to UV light (Todd et al., 2022, 2024a, 2024b). Under these conditions, we can expect a lifetime of mere minutes to hours. Our options for accumulating ferrocyanide, therefore, involve sluggish vertical mixing rates as well as some form of UV self-shielding.

ML is one of the largest and deepest soda lakes in the world, at around 48 m deep. Today, it is a monomictic lake (Jellison et al., 1993), that is, full vertical mixing occurs only once per year. Therefore, in an end-member case, we can roughly estimate that 90% of the time may be spent with the deeper, darker, denser, and colder layers, shielded from UV. This is also the reservoir into which any upwelling hydrothermal fluids would initially emerge.

Assuming full conversion of HCN to ferrocyanide and no other ferrocyanide sinks other than UV, a basal layer of 10% volume of the total lake, given a hydrothermal source equivalent to 10% of the total riverine input with 1 mM HCN concentration, would 0.06 mM ferrocyanide would accumulate each year in the basal layer. This basal layer would then mix with the rest of the water column during the annual mixing event, rapidly undergoing photolysis in the process. If the hydrothermal source instead contained 90 mM HCN, then 5.4 mM ferrocyanide would accumulate each year in the basal layer.

Our results suggest that soda lake reservoirs will not accumulate high concentrations of ferrocyanide unless the UV photodegradation efficiency is kept relatively low. There are some physical ways to extend the lifetime against photolysis in a deeper lake like ML, but not in shallow systems like LCL.

Second, if ferrocyanide production rates are sufficiently high or if other UV-sensitive side product molecules accumulate, self-shielding effects may attenuate UV penetration (Ranjan et al., 2022; Todd et al., 2022, 2024a). In this way, the accumulation of prebiotic organics in soda lakes could lead to a secular evolution in chemical composition as those organics gradually change their local UV environment.

While this self-shielding, driven accumulation could help extend the lifetimes of key intermediates, it may also inhibit essential UV-driven reactions (Green et al., 2021; Todd et al., 2022, 2024a; Xu et al., 2020). This highlights a broader challenge in complex prebiotic systems, namely that side products generated alongside key intermediates can affect the overall chemical environment. As has been noted in discussions of “messy” chemistry (Guttenberg et al., 2017), such byproducts may accumulate to higher concentrations than the intermediates themselves and potentially interfere with productive pathways. From a flux perspective, this makes understanding both the formation and removal of these species critical. This argument applies to ML and LCL but requires further work to properly quantify.

Overall, given the deeper water column of ML, we expect that the bottom half of the lake will remain effectively dark until the annual mixing episode, whereas a fully well-mixed LCL scenario would cycle its organics through the UV-active zone on a timescale of hours to days. On the basis of available evidence, we are led to conclude that cyanosulfidic chemistry in closed basin lakes will proceed most effectively if fueled by a concentrated hydrothermal source and with lifetimes for reactants, intermediates, and products extended by some form of internal UV-shielding that may be enabled more easily in deeper water columns than in shallow ones, that is, in ML versus LCL.

5.4. Sulfur

Closed basin lakes are often extremely rich in oxidized sulfur. Mono Lake (ML) has around 120 mM sulfate concentrations, while, for example (Phillips et al., 2017), LCL seasonally reaches 1 M sulfate concentrations (Haas et al., 2024; Toner and Catling, 2020). Reduced S has long been recognized to be of possible importance for prebiotic chemistry (Cody et al., 2004). However, recent advances highlight that oxidized S species are of great utility, especially in cyanosulfidic pathways (Xu et al., 2018, 2020). In this context, it is useful to note that recent calculations suggest that prebiotic Earth’s surficial S cycle was probably dominated by oxidized S (Hao et al., 2017; Ranjan et al., 2018).

Prebiotically relevant oxidized S species include bisulfite (S-IV) (Xu et al., 2020), rather than the abundant sulfate (S-VI) found in ML. However, we must consider that ML today is highly biologically active. On a more reducing early Earth with sulfur arriving in the form of sulfur dioxide (Ranjan et al., 2018) and lacking the metabolic pathways that tend to accumulate sulfate, calculations indicate that the interim accumulation of S-IV species like bisulfite is likely (Ranjan et al., 2023). However, prebiotically relevant sinks like photolysis would also likely have limited S-IV concentrations to below 100 µM (Ranjan et al., 2023).

Analogous to the above discussion of ferrocyanide, the concentration effect of soda lakes combined with some degree of shielding may have allowed for higher concentrations than elsewhere in ML- and LCL-type environments. This topic deserves detailed future modeling efforts.

6. Are Small or Large Closed Basin Lakes Best for Prebiotic Chemistry?

There are pros and cons for both small and large closed basin lakes in the context of their ability to host prebiotic chemistry. Phosphorus productivity normalized by volume, as well as total P concentrations, is roughly 1 order of magnitude higher in LCL than in ML analog lakes under prebiotic conditions. In contrast, an LCL analog is likely to be relatively inefficient at accumulating molecules vulnerable to destructive photolysis, such as ferrocyanide and bisulfite.

This contrast is enhanced if a reducing hydrothermal source to an ML scenario is considered. Such a source is highly plausible for large-scale soda lakes, given their common origin in crater lakes and rift zones. And this situation will affect productivity within the basal UV-dark layer that often forms in such lakes and allow for the accumulation of higher concentrations of relevant molecules than might occur in fully well-mixed lakes.

However, our toy model calculations make clear that long lifetimes are required for all reactant molecules, in all scenarios, to accumulate concentrations that match those used in most published laboratory experiments. Whether or not a prebiotic ML scenario could have accumulated more P than the modern example, or a prebiotic LCL could have accumulated more cyanide-derived molecules and/or overcome vulnerability to UV, will require extensive future work to resolve. Indeed, it may be that neither scenario is relevant for prebiotic chemistry, but making such an assessment with some confidence will require greater knowledge of the applicable timescales and fluxes for these environments.

Finally, just because large, closed basin lake environments may be effective at sustaining high fluxes of certain molecules does not mean we can only consider the bulk fluxes of such an environment. A large environment will inevitably be comprised of, or give rise to, other smaller environments, such as evaporating shorelines, frozen edges, rock pores, and regularly re-hydrating ponds near the main lake. In this way, a rich tapestry of prebiotic chemistry likely occurred at smaller scales, fueled by the high fluxes of prebiotic molecules to the parent closed basin environment.

Indeed, the general point we make here about the importance of timescales and fluxes should be used as an analytical filter for considering all types of environments. Focusing on the theme of this special issue, land-based models should consider how these criteria hold up for the following:

hydrothermal fields with restricted pools,

crater lakes,

lava fields,

margins of braided river systems, and

estuarine environments.

This is not an exhaustive list. The full scope of the interacting complexity of Earth’s prebiotic environments should be leveraged in the search for viable environmental scenarios for the origin of life. Yet, throughout that search, timescales and fluxes should be kept at the forefront of the analysis.

7. Hydrothermally Active Closed Basin Lakes: A Prebiotic Holy Grail?

It emerges from the above discussion that closed basin evaporative concentration is itself unlikely to be sufficient to yield highly productive prebiotic environments. Inputs from aerial deposition, post-impact deposition, riverine input, and so on are simply too low to lead to the sorts of fluxes and concentrations that would sustain suitable downstream concentrations of intermediate species and ultimately sustain productive prebiotic chemistry.

In contrast, evaporative concentration emerges as suitably powerful when coupled to an already concentrated hydrothermal inflow. Today, many closed basin alkaline lakes form in volcanic craters or rift valleys and are indeed hydrothermally active. However, the C- and N-rich composition of source magmas and reducing conditions required to obtain organic-rich inflow do not appear to occur on Earth today. While early Earth may have provided different conditions (Rimmer and Shorttle, 2019, 2024), we are therefore left without certain prebiotic analogs to study for what would otherwise clearly be a prebiotic holy grail.

In many ways, it would be unsurprising if hydrothermalism is the missing ingredient in dry-land scenarios for the origin of life. There is a reason that hydrothermal vents have captured the attention of the field for so long, which comes down, in large part, to the fact that they obviously satisfy the requirement for a constant source of energy and chemical ingredients. The fact that those ingredients may then be diluted or be the wrong ingredients is something that evaporative concentration in a reducing and hydrothermally active closed basin lake would seem to satisfy.

In other words, it may be that both “sides” of the dry land versus vent debate have had it right this entire time, but we simply need to integrate the strongest aspects of each scenario in one environment: vents on land, rather than vents at the bottom of the sea. Time will tell whether this scenario is supported by geochemical evidence, with the possibility of temporary reducing pulses on early Earth offering perhaps the most plausible scenario in which ultra-reducing vents might occur (Benner et al., 2020).

8. Conclusions and Future Work

High concentrations are important in prebiotic chemistry, as they offer a way to circumvent a lack of available enzymatic catalysis to overcome kinetic barriers. We have argued here that fluxes and timescales are also of critical importance. Fluxes and timescales determine, in part, whether an environment can achieve high concentrations of reactant and, in particular, place a critical constraint on whether high concentrations of product molecules can be maintained. We have shown how high concentrations of reactant molecules may not imply high concentrations of product molecules, and vice versa.

A major challenge is that lab experiments are often not set up with sustained fluxes and, thus, cannot achieve a steady state. The experimental constraints on both prebiogeochemical fluxes and reaction kinetics with which to inform steady-state models also remain broadly unknown (White and Rimmer, 2024a). Without these experimental constraints, it is difficult to know whether a prebiotic chemical scenario would be successful in an environment of interest.

Our limited ability to model fluxes in prebiotic chemistry, as determined by the environment of interest, is one of the most significant roadblocks to understanding which environmental scenarios are presently leading and which face the greatest hurdles. Drawing an analogy to biogeochemistry, we are presently trying to understand why the Earth’s atmosphere has so much oxygen without strong constraints on the relevant production fluxes of oxygen or the efficiency of the processes that destroy it.

On the other hand, kinetics, fluxes, and environmental box models are increasingly deployed in the prebiotic chemistry literature and have a strong presence, in particular, in the modeling of abiotic planetary atmospheres. This provides a basis for future work to expand upon (White and Rimmer, 2024a). Furthermore, the fluxes used for production rates in biogeochemical box models of the Earth system already provide a reasonable foundation for modeling prebiotic chemistry. There is a wealth of information waiting to be used to drive models of prebiotic chemistry and indeed to constrain experimental designs.

However, these rates governing production fluxes must first be collated and/or measured. While some useful work has already been accomplished toward this goal (Rimmer et al., 2018; Todd et al., 2024b; White et al., 2024b), a more focused effort is required. We suggest that future models should incorporate fluxes, in the form of rate constants, to facilitate a better understanding of the complex dynamics of potentially overlapping chemical pathways. In this way, the aggregation of chemical networks may allow us to observe previously unrealized behavior. Further, incorporating flux terms that respond to specific environmental settings should enable a more rigorous assessment of prebiotic plausibility by clarifying how physicochemical properties of a system influence its reaction dynamics, for example, rate dependency on pressure, temperature, pH, ionic strength, and so on, all of which will depend, in part, on the environmental setting.

Existing evidence continues to highlight that defeating the arithmetic demon in any prebiotic environment is still a major challenge. Across all species and scenarios considered here, the molecule lifetimes needed to obtain 1 mM product molecule concentrations stray into the millions after only three reaction steps. The key to overcoming the arithmetic demon is not to obtain extremely high concentrations of reactants but, instead, to obtain high fluxes of reactants as well as long lifetimes for product molecules. In this way, even relatively low steady-state concentrations of reactants can give rise to high concentrations of desired products (provided that the forward reaction does occur, that is, kinetic barriers are overcome).

Closed basin lakes offer a viable way to concentrate molecules relative to background sources under benign conditions. From the perspective of P, HCN and its derivatives, and S, these systems are top contenders. Nonetheless, it should be clear from the optimistic assumptions made in our calculations that realistic ways around the concentration problem and the arithmetic demon remain elusive. Even closed basin lakes with concentrated hydrothermal fluxes of reactant species struggle to furnish the levels of productivity capable of sustaining product molecule concentrations of prebiotic interest (1 mM), given reasonable product molecule lifetimes.

On this basis, we can conclude with confidence that either our knowledge of prebiotic reactions/timescales remains severely incomplete, or that microenvironments are required to alter in a fundamental way the chemical landscape, for example, lipid enclosure, mineral surfaces, and coacervates (Busch et al., 2025; Deamer, 2012; Georgelin et al., 2015; Milshteyn et al., 2018). In any case, viable solutions to these challenges will be most easily identified through the quantification and utilization of prebiotic timescales.

In conjunction with experimentally measured reaction kinetics, an opportunity exists to discriminate between the plausibility of environments on the basis of their simulated ability to generate desired concentrations of products over relevant timescales. Crucially, making such an evaluation is extremely difficult to do with confidence without quantitatively dealing with fluxes and timescales. Therefore, future work should routinely and systematically consider these aspects alongside molecule concentrations in environmental systems of interest and in experiments.

Finally, we note that the threshold concentrations often cited for prebiotic relevance are poorly constrained. While high concentrations of mM are reported in lab experiments, the origin of life could have taken place over geological timescales. There is, hence, a parameter space for trading off concentration with timescales, where long-term timescales may have allowed key reactions to occur despite lower concentrations. Since such sluggish chemical scenarios cannot easily be tested directly in the lab, models rigorously grounded in experimental constraints on the kinetics and thermodynamics of reactions will be needed to explore those nonhuman timescales.

Footnotes

Data Availability

All data needed to evaluate the conclusions in the article are present in the article.

Author Disclosure Statement

The authors declare no competing interests.

Funding Information

C.R.W. acknowledges financial support from the Cambridge Leverhulme Center for Life in the Universe; funding support from Trinity College (Cambridge) in the form of a Junior Research Fellowship; and funding support from ETH Zurich and the NOMIS foundation in the form of a research fellowship.

Associate Editor: Victor Parro

References

Alcott

, Mills

BJW

, Poulton

. Stepwise earth oxygenation is an inherent property of global biogeochemical cycling. Science 2019;366(6471):1333–1337.

Alcott

, Walton

, Planavsky

, et al. Crustal carbonate build-up as a driver for earth’s oxygenation. Nat Geosci 2024;17(5):458–464.

Benner

, Bell

, Biondi

, et al. When did life likely emerge on earth in an RNA-first process? ChemSystemsChem 2020;2(2):e1900035.

Busch

, Garenne

, Zhang

, et al. How droplets can accelerate reactions—Coacervate protocells as catalytic microcompartments. J Am Chem Soc 2025;147(12):5479–5491.

Cody

, Boctor

, Brandes

, et al. Assaying the catalytic potential of transition metal sulfides for abiotic carbon fixation. Geochim Cosmochim Acta 2004;68(10):2185–2196.

Deamer

. Liquid crystalline nanostructures: Organizing matrices for non-enzymatic nucleic acid polymerization. Chem Soc Rev 2012;41(16):5375–5379.

Dill

, Agozzino

. Driving forces in the origins of life. Open Biol 2021;11(2):200324.

Fahrenbach

, Giurgiu

, Tam

, et al. Common and potentially prebiotic origin for precursors of nucleotide synthesis and activation. J Am Chem Soc 2017;139(26):8780–8783.

Farr

, Hao

, Liu

, et al. Archean phosphorus recycling facilitated by ultraviolet radiation. Proc Natl Acad Sci USA 2023;120(30):e2307524120.

10.

Ferris

, Sanchez

, Orgel

. Studies in prebiotic synthesis III. Synthesis of pyrimidines from cyanoacetylene and cyanate. J Mol Biol 1968;33(3):693–704.

11.

Georgelin

, Jaber

, Fournier

, et al. Stabilization of ribofuranose by a mineral surface. Carbohydr Res 2015;402:241–244.

12.

Green

, Xu

, Sutherland

. Illuminating life’s origins: UV photochemistry in abiotic synthesis of biomolecules. J Am Chem Soc 2021;143(19):7219–7236.

13.

Gull

, Pasek

. Is struvite a prebiotic mineral? Life (Basel) 2013;3(2):321–330.

14.

Guttenberg

, Virgo

, Chandru

, et al. Bulk measurements of messy chemistries are needed for a theory of the origins of life. Philos Trans A Math Phys Eng Sci 2017;375(2109):20160347.

15.

Haas

, Sinclair

, Catling

. Biogeochemical explanations for the world’s most phosphate-rich lake, an origin-of-life analog. Commun Earth Environ 2024;5(1):28.

16.

Haas

, Tutolo

, Catling

. Soda lake phosphorus fluxes controlled by biological uptake imply abundant phosphate in plausible origin-of-life environments. Geochim Cosmochim Acta 2025;393:63–74.

17.

Hao

, Sverjensky

, Hazen

. A model for late Archean chemical weathering and world average river water. Earth Planet Sci Lett 2017;457:191–203.

18.

Harrison

. The hadean crust: Evidence from >4 Ga zircons. Annu Rev Earth Planet Sci 2009;37(1):479–505.

19.

Islam

, Powner

. Prebiotic systems chemistry: Complexity overcoming clutter. Chem 2017;2(4):470–501.

20.

Jellison

, Miller

, Melack

, et al. Meromixis in hypersaline Mono Lake, California. 2. Nitrogen fluxes. Limnol Oceanogr 1993;38(5):1020–1039.

21.

Kipp

, Stueken

. Biomass recycling and earth’s early phosphorus cycle. Sci Adv 2017;3(11):eaao4795.

22.

Lang

, Fruh-Green

, Bernasconi

, et al. Deeply-sourced formate fuels sulfate reducers but not methanogens at Lost City hydrothermal field. Sci Rep 2018;8(1):755.

23.

Lenton

, Daines

, Mills

BJW

. COPSE reloaded: An improved model of biogeochemical cycling over Phanerozoic time. Earth Sci Rev 2018;178:1–28.

24.

Liu

, Wu

L-F

, Kufner

, et al. Prebiotic photoredox synthesis from carbon dioxide and sulfite. Nat Chem 2021;13(11):1126–1132.

25.

Liu

, Wu

L-F

, Xu

, et al. Harnessing chemical energy for the activation and joining of prebiotic building blocks. Nat Chem 2020;12(11):1023–1028.

26.

Lyons

, Reinhard

, Planavsky

. The rise of oxygen in earth’s early ocean and atmosphere. Nature 2014;506(7488):307–315.

27.

Milshteyn

, Damer

, Havig

, et al. Amphiphilic compounds assemble into membranous vesicles in hydrothermal hot spring water but not in seawater. Life (Basel) 2018;8(2):11.

28.

Monnard

P-A

. Taming prebiotic chemistry: The role of heterogeneous and interfacial catalysis in the emergence of a prebiotic catalytic/information polymer system. Life (Basel) 2016;6(4):40.

29.

Morasch

, Liu

, Dirscherl

, et al. Heated gas bubbles enrich, crystallize, dry, phosphorylate and encapsulate prebiotic molecules. Nat Chem 2019;11(9):779–788.

30.

Pasek

. Thermodynamics of prebiotic phosphorylation. Chem Rev 2020;120(11):4690–4706.

31.

Patel

, Percivalle

, Ritson

, et al. Common origins of RNA, protein and lipid precursors in a cyanosulfidic protometabolism. Nat Chem 2015;7(4):301–307.

32.

Phillips

, Fairbanks

, Wells

, et al. Dynamics of biogeochemical sulfur cycling in Mono Lake. AGU Fall Meeting Abstracts B13K-01; 2017.

33.

Ranjan

, Abdelazim

, Lozano

, et al. Geochemical and photochemical constraints on S [IV] concentrations in natural waters on prebiotic Earth. AGU Advances 2023;4(6):e2023AV000926.

34.

Ranjan

, Kufner

, Lozano

, et al. UV transmission in natural waters on prebiotic Earth. Astrobiology 2022;22(3):242–262.

35.

Ranjan

, Todd

, Sutherland

, et al. Sulfidic anion concentrations on early earth for surficial origins-of-life chemistry. Astrobiology 2018;18(8):1023–1040.

36.

Rimmer

, Shorttle

. Origin of life’s building blocks in carbon- and nitrogen-rich surface hydrothermal vents. Life (Basel) 2019;9(1):12.

37.

Rimmer

, Shorttle

. A surface hydrothermal source of nitriles and isonitriles. Life (Basel) 2024;14(4):498.

38.

Rimmer

, Shorttle

, Rugheimer

. Oxidised micrometeorites as evidence for low atmospheric pressure on the early Earth. Geochem Perspect Lett 2019;9:38–42.

39.

Rimmer

, Xu

, Thompson

, et al. The origin of RNA precursors on exoplanets. Sci Adv 2018;4(8):eaar3302.

40.

Schwartz

. Phosphorus in prebiotic chemistry. Philos Trans R Soc Lond B Biol Sci 2006;361(1474):1743–1749; discussion 1749.

41.

Todd

, Lozano

, Kufner

, et al. UV transmission in prebiotic environments on early earth. Astrobiology 2024a;24(5):559–569.

42.

Todd

, Lozano

, Kufner

, et al. Ferrocyanide survival under near ultraviolet (300–400 nm) irradiation on early earth. Geochim Cosmochim Acta 2022;335:1–10.

43.

Todd

, Wogan

, Catling

. Favorable environments for the formation of ferrocyanide, a potentially critical reagent for origins of life. ACS Earth Space Chem 2024b;8(2):221–229.

44.

Toner

, Catling

. Alkaline lake settings for concentrated prebiotic cyanide and the origin of life. Geochim Cosmochim Acta 2019;260:124–132.

45.

Toner

, Catling

. A carbonate-rich lake solution to the phosphate problem of the origin of life. Proc Natl Acad Sci USA 2020;117(2):883–888.

46.

Tostevin

, Mills

BJW

. Reconciling proxy records and models of earth’s oxygenation during the Neoproterozoic and Palaeozoic. Interface Focus 2020;10(4):20190137.

47.

Walton

, Hao

, Schonbächler

, et al. Large closed-basin lakes sustainably supplied phosphate during the origins of life. Sci Adv 2025;11(8):eadq0027.

48.

Walton

, Shorttle

. Phanerozoic biological reworking of the continental carbonate rock reservoir. Earth Planet Sci Lett 2024;632:118640.

49.

Walton

, Shorttle

, Hu

, et al. Ancient and recent collisions revealed by phosphate minerals in the Chelyabinsk meteorite. Commun Earth Environ 2022;3(1):40.

50.

White

, Rimmer

. Do-nothing prebiotic chemistry: Chemical kinetics as a window into prebiotic plausibility. Acc Chem Res 2024a;58(1):1–10.

51.

White

, Rimmer

, Liu

. Shedding light on the kinetics of the carboxysulfitic scenario. ACS Earth Space Chem 2024b;8(11):2133–2144.

52.

Wordsworth

, Pierrehumbert

. Abiotic oxygen-dominated atmospheres on terrestrial habitable zone planets. APJ 2014;785(2):L20.

53.

, Chmela

, Green

, et al. Selective prebiotic formation of RNA pyrimidine and DNA purine nucleosides. Nature 2020;582(7810):60–66.

54.

, Ritson

, Ranjan

, et al. Photochemical reductive homologation of hydrogen cyanide using sulfite and ferrocyanide. Chem Commun (Camb) 2018;54(44):5566–5569.