Draft version September 30, 2020 Typeset using LATEX twocolumn style in AASTeX63

On The Biomass Required To Produce Phosphine Detected In The Cloud Decks Of Venus Manasvi Lingam1, 2 and Abraham Loeb2

1Department of Aerospace, Physics and Space Sciences, Florida Institute of Technology, Melbourne, FL 32901, USA 2Institute for Theory and Computation, Harvard University, Cambridge, MA 02138, USA

ABSTRACT The detection of phosphine in the atmosphere of Venus at an abundance of ∼ 20 ppb suggests that this gas is being generated by either indeterminate abiotic pathways or biological processes. We consider the latter possibility, and explore whether the amount of biomass required to produce the observed flux of phosphine may be reasonable. We estimate that the typical biomass densities predicted by our simple model are potentially orders of magnitude lower than the biomass density of Earth’s aerial biosphere in the lower atmosphere. We briefly discuss how small spacecraft could sample the Venusian cloud decks and search for biomarkers.

1. INTRODUCTION premise was extensively investigated by Limaye et al. It has been more than half a century since Morowitz (2018) in light of recent observational data and theoret- & Sagan(1967) speculated about the prospects for life ical modeling. A detailed assessment of the habitabil- in the Venusian atmosphere based on its clement con- ity of the Venusian atmosphere was presented in Seager ditions compared to the surface; this milieu was adum- et al.(2021), along with a potential life cycle for the brated earlier by Sagan(1961). 1 After these pioneering putative Venusian microbes. Izenberg et al.(2020) de- papers, a number of studies have posited that the cloud veloped a broad framework akin to the famous Drake layer of Venus situated at ∼ 50 km above the surface equation (Drake 1965), and concluded that the proba- constitute promising habitats for life because of their bility of current extant life on Venus is . 10%. moderate pressures and temperatures, sufficient protec- Despite these insightful theoretical analyses, no tion against radiation and high-energy particles, and prospective gaseous biomarkers had been identified in ostensibly adequate energy and nutrient sources (Grin- the Venusian atmosphere. In September 2020, how- spoon 1997; Schulze-Makuch et al. 2004; Grinspoon & ever, Greaves et al.(2020) reported the ≤ 15 σ detection Bullock 2007; Dartnell et al. 2015; Schulze-Makuch & of phosphine, regarded as a candidate biosignature gas Irwin 2018). On the other hand, a number of significant (Sousa-Silva et al. 2020), at a concentration of ∼ 20 ppb obstacles would have to be overcome by the hypothe- in the Venusian cloud decks. The reanalysis of Pioneer- sized Venusian lifeforms such as combating acidity and Venus Large Probe Neutral Mass Spectrometer data is maintaining osmoregulation (Cockell 1999). also indicative of the presence of phosphine in the Venu- There have been a number of notable developments sian cloud layer at ∼ 50-60 km (Mogul et al. 2020). It in the realm of Venusian astrobiology over the past few was proposed by Greaves et al.(2020) and Bains et al. years. Three-dimensional climate simulations suggest (2020) that none of the known abiotic pathways are ca- that Venus may have retained surface water and habit- pable of reproducing the observed abundance. Hence, if able conditions until ∼ 0.7 Ga (Way et al. 2016; Way & the detection of phosphine holds up to further scrutiny, Del Genio 2020), thereby providing an interval of nearly there are two major possibilities to consider: either this gas is being generated by an unknown abiotic pathway arXiv:2009.07835v2 [astro-ph.EP] 29 Sep 2020 4 Gyr for abiogenesis to occur. The presence of strong ultraviolet absorbers in the Venusian atmosphere (whose or because of biological functions. origin remains unknown) has been long conjectured to In this Letter, we examine the feasibility of the latter arise from microbial activity (Grinspoon 1997), and this scenario, namely, that the detected phosphine is origi- nating from metabolic processes. We examine the con- straints on the hypothetical metabolic pathways in Sec. Corresponding author: Manasvi Lingam 2, the possible biomass and average biomass density in mlingam@fit.edu Sec.3, and the prospects for in situ detection in Sec.4, where we also discuss some of the model limitations. 1 The notion of the “atmospheric habitable zone” has been ex- tended to gas giants and brown dwarfs (Shapley 1967; Sagan & Salpeter 1976; Yates et al. 2017; Lingam & Loeb 2019, 2021). 2. METABOLIC CONSTRAINTS 2

104 left with a number of unspecified parameters. In or- -2 fbio = 10 der to proceed further, a couple of additional simpli- -3 fying assumptions are required. It is unlikely that mi- fbio = 10 crobes could exist in the “open” for a variety of rea- 1000 -4 fbio = 10 sons, the most prominent of which is desiccation (Cock- -5 fbio = 10 ell 1999; Seager et al. 2021). Hence, we will presume that the phosphine-producing microbes are embedded 100 in droplets. Let us denote the average number den- ALLOWED sity of droplets with sizes greater than > 0.2 µm by ηd, where this limit was chosen to preserve compatibility with theoretical biophysical constraints as well as em- 10 pirical data (Knoll et al. 1999; Luef et al. 2015; Ander- ( kJ / mol ) energy free Gibbs sen et al. 2016). We will further presume that a fraction fbio of the droplets are inhabited by living microbes, and 1 that the mean number of microbes per droplet is ζbio. 10-21 10-20 10-19 10-18 10-17 In this case, the biomass is roughly given by Survival power per cell(W) Mbio ∼ ηd fbio ζbio Mcell AH, (2)

Figure 1. The Gibbs free energy (in kJ/mol) versus the where H represents the height of the atmospheric hab- power required for survival per cell (in W) for different itable zone. Note that V = AH is the volume of this choices of the fraction of biogenic particles (fbio). The re- region. It is not unreasonable to assume that ζbio ∼ 1 gions above the straight lines demarcate the potentially al- because the majority of particles in the temperate cloud lowed parameter space. layer are . 1 µm (Knollenberg & Hunten 1980; Limaye et al. 2018). By equating (1) and (2), we end up with

We will describe a heuristic procedure for determin- |∆G | Φ η f ζ P H. (3) ing the biomass that adopts the methodology delin- r r & d bio bio cell eated in Seager et al.(2013, Section 3) - see also Sholes By utilizing the simple inequality ζbio ≥ 1, the above et al.(2019) and Lingam & Loeb(2020) - for the so- equation can be rewritten to yield called Type I biosignatures, which are produced when organisms extract their energy for sustenance from re-   |∆Gr| 25 −1 H dox chemical gradients. The biomass in the temperate & 3 × 10 mol s Pcell fbio 10 km region (Mbio) can be loosely estimated as follows: η  Φ −1 × d r , (4) Mcell |∆Gr| ΦrA η0 Φ0 Mbio . , (1) Pcell 8 −3 where we have defined η0 ≡ 3 × 10 m . We have where Mcell and Pcell denote the characteristic mass and normalized ηd as per Limaye et al.(2018, Section 4) “survival” power sensu lato of a generic putative mi- and H in accordance with Cockell(1999); Seager et al. crobe, respectively, Φr is the globally integrated pro- (2021); Greaves et al.(2020). An important point worth duction flux of phosphine over the cloud layer (which is mentioning here is that the variables on the right-hand- inferred from observations), and A ≈ 4πR2 is the total side (RHS) of this inequality are seemingly constrained cross-sectional area, where R ≈ 0.95 R⊕ is the radius of to within an order of magnitude of their fiducial values, Venus. In the above formula, |∆Gr| embodies the Gibbs whereas those manifested in the left-hand-side (LHS) free energy available to the organism to extract from are unknown and therefore poorly constrained. the environment in the process of producing phosphine We have plotted this LHS in Fig.1, while holding from environmentally available chemicals. We have in- the parameters on the RHS fixed. The regions lying troduced |∆Gr| in place of ∆Gr since the above for- above the straight lines identify the allowed parameter mula is valid only for exergonic (thermodynamically fa- space. We have chosen the minimum power per cell to −21 vorable) reactions, i.e., when ∆Gr < 0. We discuss be Pmin ≡ 10 W based on the synthesis of state-of- the ramifications of this important assumption in Sec. the-art theoretical models and empirical data (LaRowe 4.1. In contrast, for ∆Gr > 0, we note that Mbio = 0 & Amend 2015; Bradley et al. 2020). The parameter because energy extraction would be rendered nonviable space we investigate corresponds to fbio  1 because and biomass cannot be sustained accordingly. it is anticipated that only a minuscule fraction of all Of these parameters, A is known, and the maximal the particles in the cloud layer are actually biogenic in −13 production flux was estimated to be Φr ∼ Φ0 ≡ 10 nature (Seager et al. 2021). By inspecting this figure, −2 −1 mol m s by Greaves et al.(2020). We are still we see that all choices of Pcell yield constraints on the 3

Gibbs free energy that might be satisfied by the con- optimal conditions specified in the figure, we end up with 7 −4 −3 jectured metabolic pathway for sufficiently small values Mbio . 4 × 10 kg and ρbio . 10 mg m . Both of of fbio. On the other hand, if we examine the extreme these estimates are not unrealistic prima facie, in view limit wherein a significant fraction of all particles are bi- of the harsh physicochemical conditions prevalent in the ological, it is apparent that |∆Gr| must be exceptionally Venusian cloud decks. Furthermore, by comparing (6) high or Pcell must be extremely low. with (2) and employing our fiducial choices, we find that fbio  1 is likely to be fulfilled. 3. POTENTIAL BIOMASS AND BIOMASS By drawing upon the measured density of aerosols in DENSITY the layer, Limaye et al.(2018) suggested that a max- −3 We remark that prior publications have sought to esti- imum biomass density of ρV ∼ 0.1-100 mg m was mate the mass of the Venusian aerial biosphere by devel- plausible for Venus, whereas Earth’s aerial biosphere oping analogs of (2) via mass-loading calculations (Li- is characterized by localized regions with a density of −3 maye et al. 2018). We will, instead, draw upon Seager ρ⊕,max ≈ 44 mg m (see Amato et al. 2007). On the et al.(2013, their Equations 11-15) and construct the other hand, the average biomass density in the lower −3 −3 analog of (1) for the biomass and the average biomass troposphere of Earth is lower at ρ⊕ ∼ 10 mg m (Fr¨ohlich-Nowoisky et al. 2016, their Table 1). It is ap- density (ρbio), under the assumption that the organisms are mixed throughout the region of height H. The po- parent that our results in Fig.2 are much smaller than these values, which indicates at the minimum that our tential biomass Mbio is expressible as predictions are compatible with past studies. It must |∆G | Φ A  E  be recognized, however, that our estimates for ρbio were M r r exp a , (5) bio . C RT solely for hypothetical microbes that give rise to phos- phine. In the same vein, if we were to repeat the anal- where T could be envisioned as the mean temperature in ysis for Earth using (7), it is anticipated that ρbio is the region where the bulk of the biomass is concentrated, negligible relative to the aforementioned densities. This while C and Ea are the survival power per unit biomass “discrepancy” would be along expected lines, because and Ea is the activation energy, respectively; note that R Earth’s aerial biomass is not primarily sustained by is the gas constant. Neither of the latter two parameters metabolic pathways producing phosphine. are known for extraterrestrial lifeforms, owing to which We can adopt a different strategy and compute the it is necessary to suppose that their typical values are upper bound on the biomass (Mmax) by making use of similar to those on Earth. This assumption might be (1) and adopting Pcell ≈ Pmin for this purpose. We can partly valid for carbon- and water-based life, but it is compute Mcell by noting that most of the particles in the probably inaccurate for exotic biochemistries. cloud decks have a diameter of ∼ 0.4 µm (Knollenberg For a generic anaerobic metabolic pathway, taking & Hunten 1980). If roughly 50% of the volume of the our cue from the fact that the Venusian atmosphere is droplet contains the microbe, which is taken to possess mostly devoid of molecular oxygen, we obtain a density close to that of water (Milo & Phillips 2016, pp. 71-74), we obtain a cell mass of ∼ 1.7 × 10−17 kg.     −9 |∆Gr| Φr This value is nearly identical to the average cell mass of Mbio . 2.1 × 10 kg −17 1 kJ/mol Φ0 M0 ≡ 1.4 × 10 kg in sediments beneath the seafloor (Kallmeyer et al. 2012). By substituting these numbers  C −1 8347 E  × exp a (6) in (1), the maximal biomass is estimated to be C0 T E0   |∆Gr| 10 −1 −1 4 M ∼ 6.4 × 108 kg where C0 ≡ 2.2 × 10 kJ kg s and E0 ≡ 6.94 × 10 max −1 1 kJ/mol J mol (Tijhuis et al. 1993). Next, we calculate ρbio =       Mcell Φr Pcell Mbio/V, thereby leading us to × . (8) M0 Φ0 Pmin     |∆Gr| Φr ρ 4.5 × 10−28 kg m−3 As a benchmark, Earth’s total biomass - which in- bio . 1 kJ/mol Φ 0 cludes the aerial, surface and subsurface components -  C −1  H −1 8347 E  is M ∼ 1015 kg (Bar-On et al. 2018), after applying × exp a .(7) tot,⊕ C0 10 km T E0 the conversion from organic carbon content to biomass (Ma et al. 2018). The upper bound for the biomass den- We have plotted (6) and (7) in the left and right panels sity (ρmax) derived from (8) is given by of Fig.2. In both panels, the temperature range was     chosen based on Limaye et al.(2018, Section 4) and the −4 −3 |∆Gr| Mcell 3.5 −1 ρmax ∼ 1.4 × 10 mg m Gibbs free energy range was truncated at 10 kJ mol , 1 kJ/mol M0 because exergonic reactions within cells on Earth rarely Φ   P   H −1 exceed this value (Cooper 2019). Even under the most × r cell . (9) Φ0 Pmin 10 km 4

Figure 2. Left panel: potential biomass present in the aerial biosphere (in kg). Right panel: possible biomass density in the biosphere (in mg m−3). In both panels, the vertical axis depicts the characteristic temperature (in ◦C) of the temperate layer, while the horizontal axis quantifies the Gibbs free energy (in kJ mol−1) for the phosphine pathway. Note the logarithmic scale on the horizontal axis, e.g., the location 2.0 is equivalent to 100 kJ mol−1.

If we specify |∆Gr| ∼ 10 − 100 kJ/mol in (9), we obtain atmosphere at the present, there are a couple of general −3 −2 −3 ρmax ∼ 10 −10 mg m . Upon comparing this value points that should be borne in mind. against ρV and ρ⊕,max delineated earlier, it is apparent The reduction of phosphate to phosphine is exergonic that ρmax is orders of magnitude smaller than either in theory at moderate temperatures provided that the of the latter duo. We reiterate that this is consistent environment is adequately acidic, as shown by Bains with prior publications, because we have only examined et al.(2019, Table 9) and Bains et al.(2020, Figure 10). the constraints on biomass imposed by one specific (and However, this reduction was enabled by the presence of possibly marginal) metabolic pathway. biological substances (e.g., NADH and iron-sulfur pro- In closing, we note that Mmax and ρmax are likely to teins), which do not fall under the category of environ- overestimate the actual biomass and biomass density as- mentally available chemicals. Yet, at the minimum, this sociated with this pathway by at least a few orders of datum suggests that the existence of similar reducing magnitude. The main reason is that, even in energy- agents in the Venusian atmosphere ought not be dis- limited environments on Earth such as sediments be- missed a priori. Our knowledge of the inorganic com- neath the seafloor, only a minuscule fraction (∼ 0.02%) pounds in the Venusian cloud decks is largely limited of microbes are capable of functioning at the minimal (Titov et al. 2018), especially when it comes to phos- power of Pcell ≈ Pmin (Bradley et al. 2020). In addition, phorus compounds because only their gross abundance at the typically warm temperatures of the cloud layer, is known and not the composition of the P-containing the basal power requirements are likely to be enhanced. species (Esposito et al. 1997; Krasnopolsky 2006). As both the sign and magnitude of ∆Gr are unknown 4. DISCUSSION in actuality, we have treated ∆Gr as a “free” positive In conclusion, we will describe some of the primary parameter in our models, i.e., it constitutes the basis of drawbacks of our modeling, and briefly explore the im- our ansatz. This approach of employing an unknown pa- plications for future life-detection missions to Venus. rameter to encapsulate the inherent uncertainty is com- mon to many areas of astronomy and physics ranging 4.1. Limitations of the model from the Shakura-Sunyaev viscosity in accretion discs Here, we explicate some of the salient drawbacks of (Shakura & Sunyaev 1973; Pringle 1981) to dark mat- the models utilized. To begin with, the framework is ter and dark energy (Weinberg 2008; Amendola & Tsu- applicable when phosphine constitutes a metabolite in jikawa 2010; Feng 2010). We reiterate, however, that an exergonic process entailing the exploitation of chem- if the phosphine is generated through alternative bi- ical energy. Hence, these models are valid if and only if ological processes (e.g., phototropy), the methodology chemical compounds in the cloud decks of Venus exist would become invalid. Furthermore, the estimates for such that the production of phosphine is favored on ther- the biomass, namely (1) or (5), implicitly operate un- der the premise of maximal efficiency. Due to the costs modynamic grounds, i.e., amounting to ∆Gr < 0. While no such compounds are confirmed within the Venusian 5 incurred by spillover reactions, ceteris paribus, these for- Let us suppose that the total energy flux of sunlight mulae probably overestimate the biomass to an extent. available in the cloud decks is Ψν of which a fraction In the course of deriving the order-of-magnitude esti- ε is harnessed for metabolic purposes, thereby result- mates for the relevant quantities, we ignored the spatial ing in the production of phosphine. It is important to and temporal variability of the model parameters as well recognize that ε encodes information about the wave- as absence of reliable knowledge (e.g., Gibbs free energy) length range of the solar spectrum accessible to pho- in some instances; to bypass this issue, we drew upon toautotrophs, the conversion efficiency of light energy Earth-based data, but this approach is not guaranteed to chemical energy, and much more (Zhu et al. 2008). A to yield accurate results. Moreover, our treatment was heuristic criterion for phosphine production at the de- exclusively centered on thermodynamic considerations, tected levels is thus given by which makes the absence of kinetics more conspicuous in view of its undoubted importance. Future studies εΨν & ∆GrΦr, (10) are needed to evaluate the rate laws for the pertinent where we have used ∆Gr in place of |∆Gr| because metabolic reactions, while assimilating thermodynamic the reaction is now taken to be endergonic in nature. effects in a self-consistent fashion (Jin & Bethke 2007). As a result, ∆Gr > 0 serves as a measure of the Another crucial point deserves to be underscored amount of energy needed to synthesize the detected here. The calculations pertaining to the biomass (or phosphine. In writing down this expression, we have the biomass density) were exclusively predicated on pu- implicitly presumed that phototrophy is energy-limited tative biochemical pathways leading to phosphine pro- with Φr ∝ Ψν , but there are many other factors in re- duction. In other words, Mbio quantifies the amount ality that are capable of restricting phototrophy such as of biomass that is linked with phosphine-generating mi- the availability of reactants, nutrients and water along crobes. Thus, it is not equivalent to the total biomass, with suitable temperatures (Lingam & Loeb 2021). We the latter of which may comprise organisms that rely have also supposed that the production of phosphine on other metabolic pathways. To take the example of −12 is continuous, and driven solely by phototrophy. After Earth, the biogenic flux of phosphine is ∼ 2×10 mol simplifying the above equation, we end up with m−2 s−1 in specialized environments (Sousa-Silva et al. 2020; Greaves et al. 2020), whereas the net global flux of  ∆G  Φ  Ψ −1 ε 1.7 × 10−13 r r ν , (11) O2 added to the atmosphere by oxygenic photosynthe- & 1 kJ/mol Φ Ψ sis by way of carbon burial is ∼ 6 × 10−10 mol m−2 s−1 0 0 −2 (Holland 2002). It is plausible, by the same token, that where the normalization factor of Ψ0 = 600 W m is the total biomass on Venus might exceed our phosphine compatible with the lower bound for the total solar en- calibrated biomass Mbio by orders of magnitude. ergy flux in the cloud decks of Venus (Seager et al. 2021, Lastly, we emphasize that the preceding analysis ex- Section 2.5). In comparison, we remark that the upper amined a hypothetical scenario in which the phosphine bound for ε is of order 0.01 for C3 and C4 photosynthesis is engendered due to biogenic activity. If future obser- on Earth (Zhu et al. 2008). Consider, for instance, that vations, laboratory experiments or numerical modeling ∼ 1% of the entire solar flux is suitable for phototrophy, demonstrate that the detection of phosphine was spuri- that ∼ 0.01% of this photon energy is used for generating ous or that it could be produced by abiotic processes, phosphine via hypothetical biochemistry, and that the naturally our discussion will be rendered irrelevant. On energy required for this process is high (∼ 103 kJ/mol). account of the preceding caveats, it is necessary to inter- Under these restrictive conditions, the maximal produc- pret our results with the appropriate degree of caution. tion flux of phosphine is nevertheless potentially four or- ders of magnitude higher than Φ0. Hence, even if a tiny fraction of the radiation is harnessed by putative Venu- 4.2. Generation of phosphine via phototrophy sian microbes at a very low quantum yield, the above We have hitherto operated under the assumption that relationship is still likely to be fulfilled. the production of phosphine took place via exergonic processes involving chemical substances available in the 4.3. Sampling the Venusian atmosphere Venusian atmosphere. We will now consider the oppo- A number of prospective biomarkers have been iden- site scenario, whereby phosphine production is ender- tified and extensively analyzed in the context of life- gonic in nature, i.e., ∆Gr > 0. In order to drive this detection missions (Summons et al. 2008; Chan et al. reaction - which is now treated as thermodynamically 2019; Neveu et al. 2020). Notable examples in this cat- unfavorable (uphill) - an external energy source becomes egory include, inter alia, enantiomeric excess in amino necessary. A number of such sources are expected to ex- acids and sugars, polymers composed of nucleotides or ist on Venus - the most dominant among them is prob- amino acids or with repeating charges in the scaffolding, ably solar radiation if the example of Earth is anything and distinct depletion of carbon and nitrogen isotopes. to go by (Deamer 1997), owing to which we will focus on Let us suppose that a fraction χ of the biological mat- the synthesis of phosphine by hypothetical phototrophs. ter (organic detritus or microbes) remains viable after 6 contact with the probe(s) equipped with instrumenta- ilar cross-sectional areas (Ouyang & Cooks 2009). By tion for in situ analysis. The mass of effective biological substituting these numbers in (12), we duly end up with material (M ) that is collected over a time ∆t by a probe b   of cross-sectional area Ap and travelling at relative hor-  χ  ρbio Mb ≈ 3.6 µg −4 −3 izontal speed vp is consequently expressible as 0.01 10 mg m  A   v   ∆t  × p p , (13) Mb ≈ χ ρbio Ap vp ∆t. (12) 0.01 m2 0.1 km s−1 1 hr

−1 This relationship applies to probes that are designed to where we normalized vp by 0.1 km s because dispers- actively scoop material in the Venusian atmosphere by ing this small spacecraft in the horizontal direction (from propelling through the cloud layer. Alternatively, bal- a larger one) may be feasible at this speed. The normal- loons floating in the atmosphere could collect particles ization for ρbio is taken to be the value described below slowly drifting downward, as proposed in Hein et al. (7); note that this choice is vastly smaller than either (2020). We will not examine this architecture herein. ρV or ρ⊕,max. For many biosignature candidates, in situ missions require the retrieval of ∼ 1 g of material Given a rough estimate of ρbio and the desired Mb, it is feasible to tune the mission design parameters in accor- in total (Neveu et al. 2020, Table 3), but of which only a minuscule fraction (to wit, of order ∼ 1 ng) actually dance with (12). The magnitude of vp depends on both the fuel costs incurred by braking as well as the type of has to be biological in nature. It is plausible, therefore, instrumentation and biomarkers being sought. For ex- that small spacecraft could possess the capacity to de- ample, in the case of impact ionization mass spectrom- tect biomarkers in the Venusian cloud decks even if the −1 probes are functional over short timescales. eters, speeds of vp ∼ 4-6 km s are probably optimal (Klenner et al. 2020). One other key physicochemical It should be recognized that our analysis is heuristic. constraint becomes crucial at high speeds. The bond Among other factors, it did not rigorously incorporate dissociation energies of many chemical bonds and or- the constraints imposed by the environment (e.g., tur- ganic molecules in particular are of order 100 kcal/mol bulence), instrument sensitivity and efficiency, and the (Blanksby & Ellison 2003; Luo 2007), which translates optimal sampling of biotic material. What it does indi- to ∼ 4 eV/bond. This value corresponds to the kinetic cate, however, is that life-detection missions might yield energy associated with a nucleon moving at ∼ 28 km meaningful results even if the density of microbes in s−1. At higher speeds, scooping material with a hard the Venusian atmosphere is orders of magnitude smaller surface could break molecules upon impact and heat the than Earth’s aerial biosphere. Hence, the importance of impacting material to thousands of degrees Kelvin, thus carrying out detailed engineering studies that fulfill the destroying any traces of biogenic material. desired scientific objectives is evident. When it comes to Ap, one can either deploy many small spacecraft akin to the Sprites on board the KickSat-2 satellite,2 or a few big spacecraft. Minimiz- ACKNOWLEDGMENTS ing the probe size permits dispersal over a larger area, This work was supported by the Breakthrough Prize but comes at the cost of reduced flexibility vis-`a-visin- Foundation, Harvard University’s Faculty of Arts and strumentation. We will appraise the former strategy Sciences, and the Institute for Theory and Computation 2 and adopt the fiducial value of Ap ∼ 0.01 m , because (ITC) at Harvard University. miniature mass spectrometers are characterized by sim-

REFERENCES Amato, P., Parazols, M., Sancelme, M., et al. 2007, Atmos. Bains, W., Petkowski, J. J., Sousa-Silva, C., & Seager, S. Environ., 41, 8253, doi: 10.1016/j.atmosenv.2007.06.022 2019, Sci. Total Environ., 658, 521, doi: 10.1016/j.scitotenv.2018.12.086 Amendola, L., & Tsujikawa, S. 2010, Dark Energy Bains, W., Petkowski, J. J., Seager, S., et al. 2020, (Cambridge: Cambridge University Press) Astrobiology, arXiv:2009.06499. https://arxiv.org/abs/2009.06499 Andersen, K. H., Berge, T., Gon¸calves, R. J., et al. 2016, Bar-On, Y. M., Phillips, R., & Milo, R. 2018, Proc. Natl. Annu. Rev. Mar. Sci., 8, 217, Acad. Sci. USA, 115, 6506, doi: 10.1073/pnas.1711842115 doi: 10.1146/annurev-marine-122414-034144 Blanksby, S. J., & Ellison, G. B. 2003, Acc. Chem. Res., 36, 255, doi: 10.1021/ar020230d Bradley, J. A., Arndt, S., Amend, J. P., et al. 2020, Sci. 2 https://www.nasa.gov/ames/kicksat Adv., 6, eaba0697, doi: 10.1126/sciadv.aba0697 7

Chan, M. A., Hinman, N. W., Potter-McIntyre, S. L., et al. LaRowe, D. E., & Amend, J. P. 2015, Front. Microbiol., 6, 2019, Astrobiology, 19, 1075, doi: 10.1089/ast.2018.1903 718, doi: 10.3389/fmicb.2015.00718 Cockell, C. S. 1999, Planet. Space Sci., 47, 1487, Limaye, S. S., Mogul, R., Smith, D. J., et al. 2018, doi: 10.1016/S0032-0633(99)00036-7 Astrobiology, 18, 1181, doi: 10.1089/ast.2017.1783 Cooper, G. M. 2019, The Cell: A Molecular Approach, 8th Lingam, M., & Loeb, A. 2019, Astrophys. J., 883, 143, edn. (New York: Sinauer Associates) doi: 10.3847/1538-4357/ab3f35 Dartnell, L. R., Nordheim, T. A., Patel, M. R., et al. 2015, —. 2020, Astrophys. J. Lett., 901, L11, Icarus, 257, 396, doi: 10.1016/j.icarus.2015.05.006 doi: 10.3847/2041-8213/abb608 Deamer, D. W. 1997, Microbiol. Mol. Biol. Rev., 61, 239 —. 2021, Life in the Cosmos: From Biosignatures to Drake, F. D. 1965, The Radio Search for Intelligent Technosignatures (Cambridge: Harvard University Press) Extraterrestrial Life, ed. G. Mamikunian & M. H. Briggs Luef, B., Frischkorn, K. R., Wrighton, K. C., et al. 2015, (Oxford, United Kingdom: Pergamon Press), 323–345 Nat. Commun., 6, 6372, doi: 10.1038/ncomms7372 Esposito, L. W., Bertaux, J. L., Krasnopolsky, V., Moroz, Luo, Y.-R. 2007, Comprehensive Handbook of Chemical V. I., & Zasova, L. V. 1997, in Venus II: Geology, Bond Energies (Boca Raton: CRC Press) Geophysics, Atmosphere, and Solar Wind Environment, Ma, S., He, F., Tian, D., et al. 2018, Biogeosciences, 15, ed. S. W. Bougher, D. M. Hunten, & R. J. Phillips, 415 693, doi: 10.5194/bg-15-693-2018 Feng, J. L. 2010, Annu. Rev. Astron. Astrophys., 48, 495, Milo, R., & Phillips, R. 2016, Cell Biology by the Numbers doi: 10.1146/annurev-astro-082708-101659 (New York: Garland Science) Fr¨ohlich-Nowoisky, J., Kampf, C. J., Weber, B., et al. 2016, Mogul, R., Limaye, S. S., Way, M. J., & Cordova, Jr, J. A. Atmos Res., 182, 346, doi: 10.1016/j.atmosres.2016.07.018 2020, Nat. Astron., arXiv:2009.12758. Greaves, J. S., Richards, A. M. S., Bains, W., et al. 2020, https://arxiv.org/abs/2009.12758 Nat. Astron., doi: 10.1038/s41550-020-1174-4 Morowitz, H., & Sagan, C. 1967, Nature, 215, 1259, Grinspoon, D. H. 1997, A New Look Below The Clouds Of doi: 10.1038/2151259a0 Our Mysterious Twin Planet: A New Look Below The Neveu, M., Anbar, A., Davila, A. F., et al. 2020, Front. Clouds Of Our Mysterious Twin Planet (Cambridge: Astron. Space Sci., 7, 26, doi: 10.3389/fspas.2020.00026 Perseus Publishing) Ouyang, Z., & Cooks, R. G. 2009, Annu. Rev. Anal. Chem., Grinspoon, D. H., & Bullock, M. A. 2007, Geophys. 2, 187, doi: 10.1146/annurev-anchem-060908-155229 Monogr. Ser., 176, 191, doi: 10.1029/176GM12 Pringle, J. E. 1981, Annu. Rev. Astron. Astrophys., 19, Hein, A. M., Lingam, M., Eubanks, T. M., et al. 2020, 137, doi: 10.1146/annurev.aa.19.090181.001033 Astrophys. J. Lett., arXiv:2009.11826. Sagan, C. 1961, Science, 133, 849, https://arxiv.org/abs/2009.11826 doi: 10.1126/science.133.3456.849 Holland, H. D. 2002, Geochim. Cosmochim. Acta, 66, 3811, Sagan, C., & Salpeter, E. E. 1976, Astrophys. J., Suppl. doi: 10.1016/S0016-7037(02)00950-X Ser., 32, 737, doi: 10.1086/190414 Izenberg, N. R., Gentry, D. M., Smith, D. J., et al. 2020, Schulze-Makuch, D., Grinspoon, D. H., Abbas, O., Irwin, Astrobiology, arXiv:2007.00105. L. N., & Bullock, M. A. 2004, Astrobiology, 4, 11, https://arxiv.org/abs/2007.00105 doi: 10.1089/153110704773600203 Jin, Q., & Bethke, C. M. 2007, Am. J. Sci., 307, 643, Schulze-Makuch, D., & Irwin, L. N. 2018, Life in the doi: 10.2475/04.2007.01 Universe: Expectations and Constraints, 3rd edn. (Cham: Kallmeyer, J., Pockalny, R., Ram Adhikari, R., Smith, Springer), doi: 10.1007/978-3-319-97658-7 D. C., & D’Hondt, S. 2012, Proc. Natl. Acad. Sci. USA, Seager, S., Bains, W., & Hu, R. 2013, Astrophys. J., 775, 109, 16213, doi: 10.1073/pnas.1203849109 104, doi: 10.1088/0004-637X/775/2/104 Klenner, F., Postberg, F., Hillier, J., et al. 2020, Seager, S., Petkowski, J. J., Gao, P., et al. 2021, Astrobiology, 20, 179, doi: 10.1089/ast.2019.2065 Astrobiology, 21, arXiv:2009.06474, Knoll, A., Osborn, M. J., Baross, J., et al. 1999, Size Limits doi: 10.1089/ast.2020.2244 of Very Small Microorganisms: Proceedings of a Shakura, N. I., & Sunyaev, R. A. 1973, Astron. Astrophys., Workshop (Washington, DC: National Academy Press) 500, 33 Knollenberg, R. G., & Hunten, D. M. 1980, J. Geophys. Shapley, H. 1967, Beyond the observatory (New York: Res., 85, 8039, doi: 10.1029/JA085iA13p08039 Charles Scribner’s Sons) Krasnopolsky, V. A. 2006, Planet. Space Sci., 54, 1352, Sholes, S. F., Krissansen-Totton, J., & Catling, D. C. 2019, doi: 10.1016/j.pss.2006.04.019 Astrobiology, 19, 655, doi: 10.1089/ast.2018.1835 8

Sousa-Silva, C., Seager, S., Ranjan, S., et al. 2020, Way, M. J., Del Genio, A. D., Kiang, N. Y., et al. 2016, Astrobiology, 20, 235, doi: 10.1089/ast.2018.1954 Geophys. Res. Lett., 43, 8376, Summons, R. E., Albrecht, P., McDonald, G., & Moldowan, doi: 10.1002/2016GL069790 J. M. 2008, Space Sci. Rev., 135, 133, Weinberg, S. 2008, Cosmology (Oxford: Oxford University doi: 10.1007/s11214-007-9256-5 Tijhuis, L., Van Loosdrecht, M. C. M., & Heijnen, J. J. Press) 1993, Biotechnol. Bioeng., 42, 509, Yates, J. S., Palmer, P. I., Biller, B., & Cockell, C. S. 2017, doi: 10.1002/bit.260420415 Astrophys. J., 836, 184, Titov, D. V., Ignatiev, N. I., McGouldrick, K., Wilquet, V., doi: 10.3847/1538-4357/836/2/184 & Wilson, C. F. 2018, Space Sci. Rev., 214, 126, doi: 10.1007/s11214-018-0552-z Zhu, X.-G., Long, S. P., & Ort, D. R. 2008, Curr. Opin. Way, M. J., & Del Genio, A. D. 2020, J. Geophys. Res. Biotechnol., 19, 153, doi: 10.1016/j.copbio.2008.02.004 Planets, 125, e06276, doi: 10.1029/2019JE006276