Every history of astrobiology has to solve the same narrative problem: the field spans a laboratory flask on a benchtop in 1953 and a spectrometer in orbit at the second Lagrange point seventy years later, and there is no single experiment connecting them. What connects them is a question that has not changed shape even as the instruments have: does chemistry become biology on its own, under conditions we can specify, and if it happened once here, could we detect it happening somewhere else. This article follows four load-bearing episodes in that history — the 1953 Miller-Urey experiment, the 1976 Viking biology experiments, the 1996 ALH84001 meteorite controversy, and the 2009-2025 run from Kepler’s exoplanet census to JWST’s atmospheric spectroscopy — and treats each one the way the historical record actually supports: as a specific, dated claim, made by named people, that was received, contested, and in some cases never resolved.
1953: a spark, a flask, and a very old question made testable
Before 1953 the origin of life was a subject for philosophical speculation more than laboratory work. Alexander Oparin and J. B. S. Haldane had proposed in the 1920s that Earth’s early atmosphere, if chemically reducing rather than oxidizing, could have permitted the spontaneous synthesis of organic compounds from simpler gases — a hypothesis with no experimental test behind it for three decades. Stanley Miller, then a graduate student working with Harold Urey at the University of Chicago, built the test. He sealed a mixture of methane, ammonia, hydrogen, and water vapor into a closed glass apparatus, ran a continuous electrical discharge through the gas to simulate lightning, and circulated the vapor through a condenser and collection trap for about a week [1]. When he analyzed the resulting residue by paper chromatography, he found glycine, alanine, and several other amino acids — the building blocks of proteins — formed from inorganic starting materials under conditions meant to approximate what he and Urey believed early Earth’s atmosphere might have looked like [1].
The result was reported in a single page in Science in May 1953, “A Production of Amino Acids Under Possible Primitive Earth Conditions,” and its effect on the field was immediate and disproportionate to its length: it took an origin-of-life question that had been the province of speculation and made it a bench experiment that other laboratories could repeat and vary [2]. That is the fact. The claim layered on top of the fact — then and for decades after — was broader: that this showed how life’s chemical building blocks did arise on early Earth. That inference does not follow from the experiment alone, and it has been substantially revised since. Geochemist Jeffrey Bada’s 2013 review of Miller’s own retained samples, reanalyzed with modern high-performance liquid chromatography, found a wider range of amino acids than Miller originally reported and confirmed the robustness of the basic chemistry — but it also underlined that the gas mixture Miller chose, dominated by methane and ammonia, no longer matches most geochemists’ best estimate of early Earth’s actual atmosphere, which is now thought to have been closer to neutral, dominated by carbon dioxide and nitrogen [2]. Under a neutral atmosphere, spark-discharge amino acid yields drop sharply. The experiment’s chemistry was real and repeatable; its claim to represent the actual early atmosphere was a simplifying assumption that later evidence weakened. Both things are true at once, and the field’s own literature says so.
By the 2000s a second hypothesis had grown to rival the atmospheric-spark scenario: that life’s chemistry began not in an atmosphere-driven surface pool but at alkaline hydrothermal vents on the early ocean floor, where a natural pH gradient between alkaline vent fluid and mildly acidic seawater could have driven proton-motive chemistry resembling the core energy metabolism still found in modern archaea and bacteria. William Martin and Michael Russell laid out this hypothesis in detail in 2007, arguing that mineral-walled vent pores containing iron-nickel-sulfur minerals could have acted as natural electrochemical reactors, with the vent’s chemical gradient substituting for the membrane-bound proton gradients that all known cells use to make ATP today [3]. This is best read as a scenario rather than a demonstrated pathway: it is grounded in a real chemical parallel between vent geochemistry and cellular metabolism, and it is testable — laboratory flow reactors have since been built to reproduce vent-like gradients — but no experiment has taken such a reactor to a self-replicating system, and the hypothesis remains one of several competing frameworks for what conditions capitalized early biochemistry, not a settled account.
What both scenarios share, and what separates 1950s-2000s prebiotic chemistry from the earlier era of Oparin and Haldane, is falsifiability. A spark-discharge yield can be measured and challenged by a better atmospheric model; a vent-reactor’s chemical output can be compared against known metabolic pathways and found wanting. The field’s later history is largely the history of applying that same discipline — dated, sourced, checked-against-geochemistry — to increasingly distant and increasingly indirect kinds of evidence.
1976: Viking asks the same question of another planet, in real time
Twenty-three years after Miller’s flask, NASA sent the question to Mars directly. The twin Viking landers, which touched down in July and September 1976, each carried three biology experiments designed to test Martian soil for signs of active metabolism, on the reasoning that if microbial life existed in the soil, it should respond measurably to nutrients or light the way terrestrial soil microbes do. The most consequential of the three was the labeled-release experiment, led by Gilbert Levin, which wetted small soil samples with a dilute aqueous nutrient solution labeled with radioactive carbon-14 and then monitored the headspace gas for radioactive carbon dioxide, on the logic that a metabolizing organism consuming the nutrients would release labeled gas as a byproduct [4]. Both landers, separated by roughly four thousand miles on the Martian surface, returned a similar result: an initial burst of radioactive gas release reaching a plateau within hours of the nutrient injection, a signal that in a comparable terrestrial soil test would normally be read as a metabolic response [4].
That is the observed fact. What it means has been argued for fifty years without resolution, which is itself the historically important part of the story. A companion Viking instrument, the gas chromatograph-mass spectrometer, found no organic molecules at all in the same soil to within its detection limits — a result hard to square with an active microbial ecosystem, since even dead microbial biomass should leave organic residue [5]. The mission’s own science team majority concluded, at the time and since, that the labeled-release signal was more likely produced by a reactive, non-biological oxidant in the Martian soil — a conclusion strengthened decades later when the Phoenix lander detected perchlorate salts at another Martian site in 2008, since perchlorates, especially when irradiated by ultraviolet light, can drive exactly the kind of oxidative reactions that would release labeled carbon dioxide from an organic nutrient without any organism present [5]. Levin, the experiment’s principal investigator, has maintained until his death in 2021 that the simplest reading of his own data was a positive biological detection, and has argued that the burden of proof was never met for ruling life out [4]. Neither position has been closed out by a follow-up measurement that would settle it definitively, which is why serious review articles as recently as 2025 describe the Viking labeled-release result as still open [5].
The disagreement is worth stating plainly in its component parts, because it is a template the field would use again. The fact: two landers recorded a repeatable chemical signal consistent with metabolism. The vendor-adjacent claim, made by the instrument’s own investigator: that signal was life. The analysis, from the wider mission team: an abiotic oxidant, later identified as a plausible perchlorate mechanism, explains the same data without invoking biology and is favored by the weight of corroborating evidence, including the absence of any detected organics. Astrobiology after Viking is, in large part, the discipline of building instruments and reasoning that could resolve this kind of ambiguity before it hardens into fifty years of argument — a discipline that, as later sections show, it has not yet fully achieved.
1996: a four-pound rock and a claim that reshaped a field’s funding
In August 1996, a NASA-led team announced in Science that a meteorite designated ALH84001, recovered from the Allan Hills region of Antarctica in 1984 and identified as a fragment of Mars ejected by an ancient impact, contained several lines of evidence they argued were “possible relic biogenic activity”: polycyclic aromatic hydrocarbons on freshly fractured interior surfaces, carbonate globules with textures the team compared to some terrestrial bacterially induced carbonate structures, and fine-grained magnetite and iron sulfide grains within those globules resembling, in size and crystal habit, magnetite chains produced by some terrestrial magnetotactic bacteria [6]. The paper was explicit that no single line of evidence was conclusive on its own and that the team considered inorganic formation possible for each feature individually; its claim was that the combination, taken together, was consistent with biological origin and could not be dismissed [6].
The announcement, made at a press conference with the sitting U.S. President commenting publicly the same day, triggered a wave of both public attention and rapid scientific rebuttal. Over the following years, independent groups proposed non-biological explanations for each of the paper’s three main lines of evidence: the PAHs were shown to be consistent with contamination from Antarctic ice and handling, or with abiotic synthesis during the meteorite’s shock heating; the carbonate globules were shown to have formed at temperatures too high for the terrestrial bacteria comparison to hold under some later thermal analyses; and the magnetite grains, while genuinely resembling magnetotactic bacterial magnetite in some populations, were also reproduced experimentally by purely inorganic decomposition of iron-bearing carbonate under laboratory heating. None of these rebuttals individually falsified the original claim beyond argument, and some of the original authors have continued to defend parts of their 1996 interpretation in later years — the case, unlike Viking’s, did produce a rough scientific consensus, but that consensus took years of published back-and-forth to form, not a single conclusive follow-up measurement.
What makes ALH84001 historically significant independent of how the specific claim resolved is institutional: the controversy is widely credited with catalyzing the formal creation of NASA’s Astrobiology Institute in 1998 and with elevating the search for evidence of past or present life on Mars from a peripheral mission goal to a central one shaping the Mars Exploration Rover program, the Mars Science Laboratory, and the sample-caching strategy now underway with the Perseverance rover. A single contested claim reorganized a field’s funding and mission architecture around the possibility, whether or not the specific meteorite ever settles the question of Martian biology.
2009–2025: from counting planets to reading their air
For most of the twentieth century, astrobiology’s second half — the search beyond the solar system — had no planets to search. That changed with NASA’s Kepler space telescope, launched in 2009, which monitored the brightness of more than 150,000 stars simultaneously, watching for the small, periodic dimming produced when a planet transits in front of its star as seen from Earth. The mission’s first published results, in 2010, described five new exoplanets and confirmed the transit method’s power to detect and characterize planets down to sizes smaller than Neptune, at a rate and precision no ground-based survey had achieved [7]. By the time Kepler’s primary mission ended, it and its successor surveys had pushed the confirmed exoplanet count from a few hundred worlds to several thousand, transforming “does a given star have planets” from an open question into a statistical baseline, and making “how common are potentially habitable rocky planets” a question with actual denominators [7].
Kepler could measure a planet’s size and orbital distance from its star, from which a rough surface temperature and thus a “habitable zone” classification could be inferred — but it could say almost nothing about atmospheric composition, the property that would actually distinguish a plausibly inhabited world from a sterile one of similar size and distance. That gap defined the next phase of the field, and the James Webb Space Telescope, launched in December 2021, was the first instrument sensitive enough to attempt it for a meaningful subset of known exoplanets. Beginning in 2023, a team led by Nikku Madhusudhan reported that JWST transmission spectra of the sub-Neptune-sized exoplanet K2-18 b, a planet orbiting a red dwarf about 124 light-years away, showed methane and carbon dioxide consistent with a hydrogen-rich atmosphere over a possible global liquid-water ocean — a planet type the team termed “hycean” — and a tentative, statistically weak signal they attributed to dimethyl sulfide, a compound that on Earth is produced overwhelmingly by marine phytoplankton and other microbial life. A follow-up analysis using a third JWST instrument in 2024 reported a strengthened, though still contested, detection of dimethyl sulfide along with a related compound, dimethyl disulfide.
That claim did not hold up as cleanly as the initial reporting suggested, and the correction is itself part of the history. A 2025 reanalysis using JWST’s MIRI instrument, applying alternative statistical treatments and a wider comparison set of candidate molecules, found that the data did not require dimethyl sulfide or dimethyl disulfide at all — the same spectral features could be explained by other, unambiguously non-biological molecules, and the statistical significance originally claimed for the biosignature interpretation did not survive more conservative analysis choices [10]. Science journalism covering the reanalysis described the episode plainly as “a claimed hint of alien life” that “whips up spirited debate” rather than a confirmed detection [11]. The disagreement here is the same shape as Viking’s, compressed into two years instead of fifty: an ambiguous spectral signal, an initial biology-favoring interpretation from the team that found it, and a subsequent analysis from other researchers arguing the same data does not require that interpretation.
This is why the field now treats “biosignature” claims through an explicit false-positive framework rather than a simple presence-or-absence test. Oxygen is the clearest illustration. A 2018 review led by Victoria Meadows catalogued the ways a planet could accumulate abundant atmospheric oxygen with no biosphere at all — most importantly, photochemical loss of a planet’s primordial water and subsequent hydrogen escape to space, which can leave an oxygen-rich atmosphere behind on a planet that was never habitable in any biological sense, particularly around small, active red dwarf stars whose intense early ultraviolet output can drive this process efficiently [8]. A companion 2018 review by Edward Schwieterman and coauthors laid out the wider landscape of candidate biosignature gases — oxygen, ozone, methane in disequilibrium with oxidizing species, nitrous oxide, and organic haze signatures among others — and stressed that essentially every candidate gas has at least one plausible abiotic production pathway under some planetary conditions, meaning no single molecule detected in isolation can serve as unambiguous proof of biology [9]. The methodological consensus that emerged from this literature, and that both the K2-18 b claim and its rebuttal were argued within, is that a credible biosignature detection requires a specific molecule or combination of molecules, a planetary and stellar context that disfavors the known abiotic pathways for producing it, and independent lines of evidence, rather than a single spectral feature crossing a statistical threshold.
What the record actually supports
Read end to end, the four episodes do not describe a field converging steadily on an answer. They describe a field that has gotten progressively better at stating its own uncertainty precisely, which is a different and more durable kind of progress. Miller’s flask showed a real chemical possibility and, decades later, prompted its own correction about the atmosphere it was meant to model. Viking produced a signal that the instrument team and its principal investigator have never agreed how to read, and no subsequent Mars mission has definitively closed the question either way. ALH84001 produced a claim later substantially, though not unanimously, disfavored by follow-up work, and it left behind an institutional legacy — dedicated funding, dedicated missions — larger than the claim itself. K2-18 b produced a biosignature candidate that a more careful reanalysis found the data did not require. In each case the discipline that ultimately mattered was not the initial instrument reading but the willingness of other scientists to reanalyze the same evidence and publish a different conclusion.
Where the field is heading is visible in the shape of that improvement rather than in any single upcoming result. Missions and instruments now under construction or in early operation — the European Space Agency’s PLATO mission for transiting rocky planets around sun-like stars, proposed NASA concepts for direct imaging of Earth-sized exoplanets, and continued JWST time on the small number of temperate sub-Neptunes and rocky worlds bright enough to permit atmospheric spectroscopy at all — will multiply the number of ambiguous signals long before they multiply the number of settled ones. A reasonable forecast, on a roughly ten-to-twenty-year horizon, is that additional tentative biosignature candidates will be reported and contested in the same pattern as K2-18 b, each one requiring the observable indicators the false-positive framework specifies: a molecular combination not easily produced abiotically under that star’s specific ultraviolet and X-ray environment, ideally cross-checked against an independent tracer of the planet’s history, such as evidence for or against ongoing atmospheric escape. The condition that would disconfirm this forecast is a single, unambiguous biosignature detection meeting that full standard within the window — an outcome the field’s own literature treats as possible but, on current instrument sensitivity and the small sample of accessible planets, unlikely to arrive quickly. Until then, the honest continuation of the history told here is not a discovery but a widening set of contested spectra, read the way Viking’s data still is: carefully, in public, and without a forced verdict.
