Flare-driven habitability: Expanding life’s potential around low-mass stars

Abstract

The traditional definition of the circumstellar habitable zone (HZ) focuses on liquid water, but neglects the crucial role of ultraviolet (UV) radiation in prebiotic chemistry. Low-mass stars typically emit insufficient UV radiation for photochemistry throughout the liquid water HZs (LW-HZs) during quiescent states. However, frequent flares can provide substantial UV fluxes, potentially fostering habitable conditions. We refine the concept of a UV radiation HZ (UV-HZ) by incorporating a temperature-dependent model for RNA precursor synthesis. Furthermore, we explore a parameterized spectral energy distribution model and adopt an empirical flare frequency distribution for flares on different stars to quantify their UV contribution. Applying this framework to different flaring stars, we find that the UV-HZ around low-mass stars can extend to inner regions and overlap with the traditional HZ in wide ranges. Applying the analysis to 9 planets around Kepler flaring stars, three planets are located within both the refined UV-HZ and the LW-HZ without causing ozone depletion. Our findings highlight the significant role of flares in expanding the potential for life around low-mass stars, offering a revised perspective on exoplanet habitability criteria.

Graphical abstract

Public summary

• •An improved flare spectrum model evaluates the effect of stellar flares on planetary habitability.
• •Photochemical reaction rates varying with temperature constrain a physical ultraviolet habitable zone (UV-HZ).
• •Flares extend the HZ around low-mass stars to sustain both liquid water and prebiotic UV environments.
• •One rocky planet around Kepler flaring stars is identified in the UV-HZ and is worth further observation.

Introduction

Detecting habitable planets has been a primary objective of the astronomical community for decades. Recent discoveries of numerous exoplanets have revealed various planetary architectures, with a higher prevalence of terrestrial worlds over giant planets.¹ To assess the habitability of the terrestrial planets, it is essential to gain a deeper understanding of their intrinsic properties and the characteristics of their host stars.² A critical aspect of this assessment is the potential for the rocky planets to sustain liquid water on their surfaces, which defines the liquid water habitable zone (LW-HZ).³

Searching for exoplanets in the HZ is crucial for identifying promising targets for follow-up observations and understanding their planetary properties.^4,5 Previous studies have defined the conservative HZ’s boundaries through simulations of water loss and the maximum greenhouse effect, while the optimistic HZ is based on empirical criteria from “recent Venus” and “early Mars.”^3,6 However, these definitions primarily consider the star’s steady incident flux and overlook the ultraviolet (UV) radiation, which is particularly significant for stars with frequent high-energy flares.

Excessive UV radiation can inhibit photosynthesis and damage biological systems, while moderate UV radiation supports essential biological processes.^7,8 According to the “principle of mediocrity,” Buccino et al.⁸ proposed that planets in the UV radiation HZ (UV-HZ) should receive a UV flux between half and twice that of early Archean Earth. Since pyrimidine RNA precursors are essential building blocks for life, understanding the pathway to form pyrimidines is crucial.^9,10 This pathway involves seven steps with similar photochemical rate constants, starting from a mixture of HCN and ${\text{SO}}_3^{2-}$ or HS⁻. Rimmer et al.¹⁰ estimated the amount of UV light for photochemistry by experimentally measuring the rate constants for UV-driven photochemical reactions (light chemistry, required for prebiotic synthesis) and bimolecular reactions occurring in the absence of UV light (dark chemistry). They selected the region where UV radiation (>45 erg cm⁻² s⁻¹) can provide at least a 50% yield per step, resulting in an overall yield (>0.1%) sufficient to sustain stable prebiotic chemical reactions. According to the Arrhenius equation,¹¹ chemical reaction rates are highly sensitive to temperature. Rimmer et al.¹⁰ also provide the near-UV (NUV, 200–280 nm) flux required for abiogenesis at different temperatures, referred to as the abiogenesis flux (f_Abio).

As stellar temperature (T_eff) decreases with decreasing stellar mass, UV radiation from low-mass stars is weaker. Thus, quiet, low-mass stars may not provide sufficient UV energy for synthesizing biological mechanisms within the LW-HZ.^12,13

Since many low-mass stars are more magnetically active, these flares, caused by magnetic reconnection on the stellar surface, release bolometric radiation.^14,15,16 The flares also emit enhanced UV radiation compared to their quiescent stages,^17,18,19 which may support biogenetic processes.⁹ However, frequent superflares can deplete the ozone layer, allowing much more harmful UV doses to reach the planetary surface and damage proteins and lipids.^20,21 This raises a critical question: is the UV radiation moderate due to stellar flares, to sustain the HZ around low-mass stars? Understanding habitability, especially for planets around low-mass flaring stars, requires more information about the UV radiation and flare frequency of the host star.

The flare frequency distribution (FFD) is a key indicator of stellar magnetic activity and crucial for assessing planetary habitability due to flares.^17,22 However, instrument precision and observational cadence can lead to incomplete detection of low-energy flares.^18,23 Additionally, the lack of simultaneous multi-wavelength observations introduces significant inaccuracies in modeling flare spectral energy distributions (SEDs).^19,24 These limitations hinder accurate FFD determination for different stars. To estimate total flare energy and obtain flare SEDs from optical light curves, both radiative-hydrodynamic simulations and empirical models are used.^25,26

In this study, we focus on flaring stars, which exhibit frequent flare events that boost UV and bolometric radiation on orbiting planets. We reassess the UV-HZ around these stars using an empirical and corrected cumulative FFD (CFFD) for different stellar masses and also construct an SED model for stellar flares in the NUV band. Unlike previous studies,^13,27 we provide a more comprehensive approach by both considering the minimum requirements of UV radiation that vary with planetary surface temperatures and improving the estimation of UV-HZ boundaries based on the UV radiation required for RNA precursor synthesis. Additionally, we apply our methods to known Kepler planets around flaring stars to evaluate their habitability. This paper aims to identify which types of flaring stars can sustain temperate regions where liquid water and moderate UV radiation can enable the synthesis of genetic molecules or organisms.

Materials and methods

Flare SED model for different flaring stars

The flare SEDs are generated via both thermal and non-thermal mechanisms, depending on various environments, e.g., the local magnetic structure, local plasma conditions, and the rate of energy deposition.^28,29,30 These SEDs are expected to explain the observed emissions and provide clues to physical processes in solar and stellar flares. We collect the characteristics of flare SEDs as follows.

• (1)Continuum radiation in the optical and NUV wavebands during flares is thought to be the response of the atmospheric heating, magnetic energy release, and electron acceleration at coronal altitudes.³¹ The hot-blackbody assumption (a thermal photospheric spectrum with T ≈ 9,000 K) is able to replicate some spectral regions of flare spectra (e.g., the white-light band) while ignoring emission lines.³² The NUV continuum is significantly higher than the hot-blackbody assumption, most likely caused by hydrogen recombination.¹⁹

• (2)Hydrogen Balmer lines are enhanced and broadened during flares, particularly in the NUV band. Below the Balmer jump wavelength (at ∼0.365 μm), flare fluxes can increase by several times compared to predictions from hot-blackbody models. However, only a few solar flares exhibit an obvious Balmer jump.³³

• (3) The emission line contribution, especially in the NUV band, is non-negligible.^24,34 Fortunately, this contribution can be included by integrating the hot-blackbody spectrum with an appropriate pseudo-continuum temperature.¹⁹

Several flare SED models based on the 9,000 K hot-blackbody spectrum, combined with the far-UV (FUV) and NUV emission lines taken by the Hubble Space Telescope (HST), better match actual observations than the hot-blackbody model (e.g., the MUSCLES model³⁵). Based on the research of Transiting Exoplanet Survey Satellite (TESS) white light flares for M dwarfs and combined with the GALEX UV photometric data, Jackman et al.¹⁹ extrapolated stellar FFD at the UV band to test different flare SED models. They found that the GALEX NUV (0.177–0.283 μm) energies on M0-M2 stars predicted by the 9,000 K hot-blackbody model are underestimated by up to a factor of 2.3 ± 0.6. A more accurate SED of flares requires multiple waveband observations for each event, including low-energy flares.^35,36

To improve the flare SED model on various star types, we employed a power-law continuum superimposed with a hot-blackbody spectrum from 0.100 to 0.365 μm in the UV band and a hot-blackbody model in the white-light band, considering the Balmer jump. Our flare SED model is described by Equation 1,

$$F_{\lambda }\left(T_{\text{flare}}\right)=\{\begin{array}{l}{F}_{\lambda ,\text{HBB}}\left(T_{\text{flare}}\right),\text{for}\lambda >0.365\mathrm{\mu }\mathrm{m},\\ F_{\lambda ,\text{HBB}}\left(T_{\text{flare}}\right)+F_{0.365\mathrm{\mu }\mathrm{m},\text{HBB}}\left(T_{\text{flare}}\right)\times f_{\mathrm{b}}\times {\left(\frac{\lambda }{0.365\mathrm{\mu }\mathrm{m}}\right)}^{\gamma },\text{for}0.1\mathrm{\mu }\mathrm{m}\le \lambda \le 0.365\mathrm{\mu }\mathrm{m},\end{array}$$ (Equation 1)

where F_{λ, HBB}(T_flare) = B_λ(T_flare) + S_λ, i.e. the combination of the hot-blackbody spectrum B_λ(T_flare) and the stellar spectra at quiet state S(λ). T_flare is the temperature of the stellar flare active region. The factor f_b is the enhancement ratio at 0.365 μm, and γ is the power-law index related to the wavelength dependency.

We set T_flare to 9,000 K, leaving two parameters for the flare SED model: the enhancement of the continuum at wavelengths shorter than the Balmer jump (f_b) and the power-law index (γ) describing the UV continuum trend. This model corrects the underestimation of UV radiation using the 9,000 K hot-blackbody assumption.^19,24

Stellar and solar flares are manifestations of magnetic activity and share a similar physical process.³⁷ However, on stars of different masses, with varying convective layer thicknesses, flares exhibit distinct characteristics.^31,38 We fitted f_b and γ with observations and the spectrum model of M and G stellar flares and used linear interpolation to model the SED for different types of stars (see Table S1).

Figure 1 shows several flare SED models. Previous observations^19,33 indicate that for M stars, f_b is higher than for solar flares, while γ is lower. This suggests that UV-optical radiation contrast during flares is more pronounced on M stars compared to G and K stars. In our model, as T_eff ranges from 3,400 to 6,000 K, f_b(T_eff) varies from 1.013 to 0.011, and γ(T_eff) from −0.500 to 1.032 (see Figure S1). Note that this work investigates changes in the HZ considering flares for specific spectral types. Therefore, f_b and γ represent the typical or average flare SED for a stellar type, not an individual star. The energy range of detected solar and stellar flares spans from 10²⁸ to 10³⁸ erg.^17,23,39,40 Details on our flare SED model validation and the increase in stellar luminosity due to flares are provided in the section average luminosity of flaring stars of the supplemental information. The averaged stellar NUV radiation is a necessary condition for the accumulation of photochemical reaction products, which are essential for forming the macromolecular prebiotic inventory (see the HZs around different flaring stars subsection for details).

Figure 1.

Flare SEDs on different spectral types

The blue, orange, and green solid lines represent flare SED models for stars with different effective temperatures of 3,400, 4,500, and 6,000 K, respectively. Two arrows highlight the flare SEDs for M2- (T_eff = 3,400 K) and G0- (T_eff = 6,000 K) type stars, respectively. For stars with T_eff< 6,000 K, flares can enhance the NUV band compared to the T_flare = 9,000 K hot-blackbody model (the green dotted line). The black solid line represents the “great flare” spectra of AD Leo, with points with error bars marking its wavelength-binned observed continuum fluxes from Kowalski.²⁵ The brown dashed line is the best fit using a two-component radiative-hydrodynamic model from the same study. Colored backgrounds show the relative band passes of different filters, with gray vertical dashed lines marking the mean wavelengths of the GALEX FUV band (0.155 μm), GALEX NUV band (0.230 μm), Evryscope g′ band (0.464 μm), and TESS band (0.745 μm).

Calculation of HZ around flaring stars

LW-HZ

For quiet stars, the LW-HZ boundaries are determined using the methods of Kopparapu et al.^3,6 The inner boundary corresponds to recent Venus at 57°C⁴¹ and the outer boundary to “early Mars” at 0°C. For flaring stars, stellar flares increase the flux reaching the planet. Although magnetically active regions are randomly distributed across the stellar surface, only flares directed toward the planet affect its habitability. We calculated the increase in stellar luminosity caused by flares, as detailed in the supplemental information. Considering existing flare spectra and the SEDs for M to G stars, the integrated wavelength range is set from 0.1 to 3.0 μm.

By adjusting the LW-HZ boundaries for quiet stars, considering the increased stellar bolometric luminosity, we determined the LW-HZ boundaries for flaring stars. This adjustment is uniformly applied to both the inner and outer limits,

$$d_{\text{flare},\text{LW}}=d_{\text{quiet},\text{LW}}\times {\left(1+\frac{L_{\text{flare}}}{L_{\ast }}\right)}^{0.5}\text{,}$$ (Equation 2)

where d_{flare, LW} is the distance from the host star to the boundary of LW-HZ when considering stellar flares and d_{quiet, LW} is the distance from the host star to the boundary of LW-HZ without considering stellar flares. L_flare is the time-averaged bolometric luminosity of the star with stellar flares (as explained with Equation S6, and L_∗ is the bolometric luminosity in the quiescent state.

UV-HZ

The UV-HZ mainly focuses on biological processes. On Archean Earth about 3.8 billion years ago, NUV radiation provided essential energy sources for synthesizing biochemical compounds. Previous studies defined the UV-HZ considering the UV flux required for abiogenesis only at 0°C. However, according to the Arrhenius equation,^10,11 chemical reaction rates are highly sensitive to temperature. On planets like Earth, with an average temperature of 15°C, reaction rates significantly increase. Consequently, more UV radiation is needed to sustain RNA synthesis.

To obtain the average NUV flux required for abiogenesis (f_Abio), we fitted the relationship between f_Abio and the planet surface temperature (T_surf) using the Arrhenius equation, according to Figure 3A in Rimmer et al.¹⁰ Similarly, we chose the 50% curve from the figure to ensure sufficient flux for robust prebiotic chemistry. The fitting results are as follows:

$$f_{\text{Abio}}=f_0\times e^{\left(T_{\text{surf}}-273\mathrm{K}\right)/10K}\text{photons}{\text{cm}}^{-2}{\mathrm{s}}^{-1}\text{,}$$ (Equation 3)

where f₀ = 4.08 × 10¹².

The surface equilibrium temperature T_surf is proportional to d^−0.5 if we ignore the atmospheric greenhouse effect of a planet. However, when considering the planet’s atmosphere, the temperature decrease with distance is less significant. To model the correlation between T_surf and d, we adopted an exponential function:

$$T_{\text{surf}}=T_0{\left(\frac{d}{1\text{AU}}\right)}^{\chi }\text{.}$$ (Equation 4)

To determine T₀ and χ in Equation 4, we use the inner and outer boundaries of the LW-HZ defined by Kopparapu et al.^3,6 The inner boundary is set at the moist greenhouse limit, where the surface temperature is set as 330 K,⁴¹ while the outer boundary is at the freezing point of water (T_surf = 273 K). By comparing these boundary temperatures (i.e., T_surf) with their respective distances (i.e., d), we can derive the value of χ using Equation 4 for different stars. In the case of a solar-type star, T₀ = 310.12 K and χ = −0.22. Consequently, by substituting Equation 4 into Equation 3 and adopting the first-order approximation, we have f_Abio ∝ d⁻⁷ within the LW-HZ. Figure 2 shows the f_Abio and modeled surface temperatures at different locations for typical M, K, and G stars.

Figure 2.

The required NUV flux for prebiotic chemistry in LW-HZ around typical M, K, and G stars

The green solid line represents the f_Abio requirements at different locations, while the blue solid and dashed lines represent the stellar NUV insolation with and without flares, respectively. The inner boundary of the UV-HZ is the intersection of the green and blue solid lines. The NUV flux of Archean Earth (F_Arch) is also shown using a dot-dashed line. The red line shows the variation of modeled surface temperature (T_surf) on the planet, and the green background marks the region with temperatures between 0°C (273 K) and 57°C (330 K).

To obtain the quiescent NUV flux of different stars, stellar photospheric models (e.g., PHOENIX^42,43) are usually adopted, which only include stellar photospheric radiation. However, radiation contributions from the chromosphere and transition region cannot be neglected. Especially for cooler stars (K to M stars), the photospheric flux between 200 and 280 nm is relatively weak in comparison to chromospheric emission.⁴⁴ To account for contributions from other mechanisms, we corrected the NUV flux obtained from the PHOENIX model by dividing by different factors. These factors are the fraction of modeled radiation to observed radiation in the GALEX NUV band (0.177–0.283 μm). We calculated the median factors for different stars according to the statistical sample in Wang et al.⁴⁵ (see Figure S3). After correction, the total NUV radiation of different stars in the quiescent stage can be obtained, including contributions from the photosphere, chromosphere, transition region, and corona.

To estimate the contribution of NUV due to stellar flares, we adopted the flare SED model and the empirical CFFD based on a sample of young stars ranging from G to M types.¹⁸ We calculated the total energy of flaring stars over a long-term period and added the time-averaged increase in NUV flux to the quiescent stellar flux (see enhancement of stellar luminosity caused by flares in the supplemental information). Note the NUV enhancement due to stellar flares is time dependent. The typical timescale of flare events is several hours. According to previous studies, the reaction timescales of certain prebiotic precursor molecules, such as 2-aminooxazole, range from minutes to hours.⁴³ However, for planetary atmosphere or surface, the subsequent chemistry of the products is often coupled with the global dynamics of materials, which typically operate on timescales much longer than flare durations. For instance, the global impact of stellar flares on the atmospheric ozone column depth can last for months to years, according to Segura et al.⁴⁶ Thus, adopting the time-averaged increase due to stellar flares can represent the secular and global influence on photochemistry.

After obtaining the averaged NUV flux of flaring stars, we used the lower limit of f_Abio to estimate the UV-HZ boundary. By comparing the average stellar insolation and f_Abio (see Equation 3) above the planet’s atmosphere at different distances and surface temperatures, we determined the inner boundary of UV-HZ with and without flares (see Figure 2). As the distance from the star decreases, the abiogenesis flux rapidly increases (f_Abio ∝ d⁻⁷ for solar-type stars), while stellar radiation increases more gradually (∝d⁻²). Therefore, there is a physical inner boundary of UV-HZ to guarantee the minimum requirements of f_Abio. Beyond the inner boundary, stellar radiation consistently meets the abiogenesis flux required for life (see Figures 2, S6, and S7). Considering the need for liquid water to support life, the outer boundary of UV-HZ is therefore set to align with the outer boundary of LW-HZ.

The limits of CFFD for ozone depletion

Since the inner boundary of UV-HZ is defined by the lower limit of UV flux, there should also be an upper limit of the UV flux, as excessive UV radiation can dissociate prebiotic molecules. According to Abrevaya et al.,⁴³ even if the received UVC (200–280 nm) flux is ∼15 times higher than the current Earth’s flux outside the atmosphere (i.e., 92 W m⁻²), a small fraction (∼10⁻⁴–10⁻⁵) of microorganisms can still survive. However, this UVC flux refers to radiation outside the atmosphere. Since atmospheric components, such as ozone, have varying abilities to shield UV radiation, the surface UV flux should be lower than the estimated flux outside the atmosphere. Furthermore, UV-photounstable molecules can extend their lifetimes through a self-shielding mechanism (e.g., 2-aminooxazole) or the presence of other UV-absorbing molecules (e.g., purine ribonucleosides).⁴⁷ Additionally, turbid water containing UV-absorbing species (e.g., ${\text{Fe}\left(\text{CN}\right)}_6^{4-}$) can strongly shield UV radiation.⁴⁸ Thus, some molecules dissolved or accumulated in water may be preserved. We therefore did not really use the upper limit on UV flux via prebiotic chemistry.

However, for mature planetary systems, such as terrestrial planets around Kepler flaring stars, ozone in Earth-like atmospheres becomes crucial for shielding UV radiation and protecting potential life. Thus, we constrained the CFFD of superflares based on previous studies to avoid significant ozone depletion. Frequent superflares can severely deplete the ozone layer via energetic particles and UV radiation. An Earth-like ozone layer around M dwarf stars such as GJ1243 would be greatly depleted by flares of 10³⁴ erg at a frequency of 0.4 day⁻¹.^17,21 Therefore, we scaled an upper limit on CFFD to account for ozone depletion, i.e., if the CFFD of a given flaring star exceeds ${10}^{34}\times {\left(\frac{d}{0.16\text{AU}}\right)}^2$ erg flares, the planet is considered inhabitable. d represents the distance between the planet and the host stars. Fortunately, all empirical CFFDs adopted for different stars are below these upper limits, even when considering the uncertainties in the fitted correlations.

Results

HZs around different flaring stars

Life on Earth originated 3.8 Gyr ago,⁴⁹ when the Sun was approximately 0.8 Gyr old. Therefore, a timescale of ∼1.0 Gyr is required for life to originate.^50,51 Chromospheric activity and coronal emission of G, K, and M stars gradually keep a high level before 1 Gyr and decrease due to the reduced dynamo production of magnetic fields as the star spins down.⁵² It is hard to know when the prebiotic molecules can be produced for Archean Earth, but the M stars with longer lifetimes in the main-sequence stage can have adequate time to accumulate the prebiotic molecules when the UV radiation becomes moderate. To estimate the NUV radiation of young stars during life origination, we selected 1.0 Gyr as the typical age for calculating stellar insolation.

For the general case, we consider flaring stars ranging from 0.3 to 1.1 solar masses (M_⊙), spanning spectral types M2 to G0. We employed the stellar evolution model at an age of 1.0 Gyr from Baraffe et al.⁵³ to calculate the effective temperature (T_eff) and radius (R_∗) of these stars. Using flaring samples from Gao et al.,¹⁸ we established an empirical relationship between the CFFD and T_eff and calculated the HZ boundaries considering the averaged luminosity of flaring stars. More calculation details are provided in the materials and methods section.

As shown in Figure 3, the UV-HZ always exists around early K- and G-type stars (M_∗ = 0.76–1.10 M_⊙, T_eff = 4,700–6,000 K), regardless of flares. Considering the quiescent stellar NUV radiation, including contributions from the photosphere, chromosphere, transition region, and corona, the UV-HZ is narrow (e.g., less than 0.007 AU for a quiet star below ∼0.36 M_⊙), and the planet can hardly be located inside. If only considering the NUV flux contributed by the photosphere, the UV-HZ disappears for quiet stars below ∼0.76 M_⊙. However, if the star has flares, the increased UV flux can extend the UV-HZ. For G-type flaring stars, the UV-HZ extends slightly because the enhancement of UV flux due to flares is only a small fraction of quiescent radiation. Due to a higher flare frequency, the UV-HZ extends more noticeably for lower-mass stars. Additionally, K-type stars around 0.68 M_⊙ exhibit a narrow UV-HZ. Notably, the flare SED significantly enhances NUV luminosity rather than the total luminosity; thus, the LW-HZ experiences negligible change (<0.006 AU) across all stellar types.

Figure 3.

The habitable zones around flaring stars at an age of 1.0 Gyr

The magenta solid line indicates the flare-driven inner UV-habitable zone (HZ) boundary, while the magenta dashed line shows the quiet inner UV-HZ boundary. The lime solid line indicates the flare-driven LW-HZ boundary, while the lime dashed line shows the quiet LW-HZ boundary. The boundary of the flux required for abiogenesis at 0°C without (w/o) and with flares is plotted as blue dashed and solid lines, respectively. Considering that prebiotic chemistry also requires a liquid water environment, we adopt UV-HZ with the same outer boundary as LW-HZ. The position of Earth is also marked with a circled cross in the figure.

To assess the impact of flare SED models on UV-HZ changes, we examined two specific flare SEDs: the G0-type flare SED, which includes only the traditional 9,000 K hot-blackbody assumption, and the M2-type flare SED, which accounts for the Balmer jump of flares on M-type stars and is thus more accurate. As shown in Figure S9, although the UV-HZ range using the M2-type flare SED is significantly larger, both SEDs can extend the UV-HZ obviously and lead to a wide overlapping range with LW-HZ.

HZs around Kepler flaring hosts

Yang and Liu²³ presented a catalog of Kepler flaring stars, including five with planetary candidates and four with confirmed exoplanets listed as habitable planetary hosts in the NASA Exoplanet Archive (https://exoplanetarchive.ipac.caltech.edu/). Their habitability is defined by an equilibrium temperature of 180–310 K or received insolation between 0.25 and 2.2 times that of Earth.

Using the stellar and planetary parameters (see Table S4), we calculated the LW-HZ boundaries based on Equation 3 from Kopparapu et al.³ We performed flare detection on the Kepler light curves of these hosts and fitted the CFFD parameters for each star (see Table S3). We measured the NUV radiation flux of five hosts using GALEX NUV data (see Table S4). For four hosts without GALEX NUV or Swift UVOT observations, we adopted their NUV radiation flux as the statistical median values from Wang et al.⁴⁵ (see Figure S3). For the Sun, we adopted 1.4 times the modern solar spectrum from Willmer⁵⁴ to account for the NUV radiation evolution of G-type stars with an age of 1.0 Gyr.⁵⁵ We then derived the LW-HZ and UV-HZ boundaries, considering both the presence and absence of stellar flares. Figure 4 shows the changes in stellar insolation and the changes in the HZ boundary before and after considering flares around Kepler-438. During the quiescent stage, the UV-HZ is as narrow as 0.04 AU. After accounting for the enhancement of flares, the width of the UV-HZ increases to 0.16 AU, resulting in a 4-fold increase in the likelihood of accommodating an additional terrestrial planet in the HZ.

Figure 4.

Stellar insolation around Kepler-438 without and with flares

(A and B) NUV band insolation (200–280 nm) without flares (A) and bolometric insolation under the same quiet conditions (B).

(C and D) These same insolation types from (A) and (B) but with flares. Lime and magenta vertical lines mark the LW-HZ and UV-HZ boundaries without/with flares, respectively.

Figure 5 illustrates the HZ boundaries of the planet hosts. KOI-7706.01 and Kepler-1512 b are no longer in the LW-HZ with updated stellar parameters (see Table S4). Although Kepler-1540 b, Kepler-438 b, and Kepler-155 c are in the LW-HZ, they are not in the UV-HZ. The three planets, i.e., KOI-7703.01, KOI-8047.01, and KOI-8012.01, are all located in the overlapping region of the two HZs (regardless of whether flaring is considered or not). Coincidentally, these three stars are G, K, and M stars, respectively. Note that KOI-5879.01 was originally in the LW-HZ. After considering the flares, the NUV radiation can satisfy the abiogenesis flux, but it is no longer in the LW-HZ.

Figure 5.

LW-HZs and UV-HZs sorted by stellar mass

Lime dashed boxes represent quiet LW-HZs. Lime backgrounds represent flare-driven LW-HZs. Magenta dashed boxes represent quiet UV-HZs. Magenta backgrounds represent flare-driven UV-HZs. Black dots represent planets or candidates located at their semi-major axes with errors (see Table 1). Due to unavailable CFFD power-law indices for KOI-7703 and the Archean Sun, their flare-driven HZs are not considered.

As described in subsection the limits of CFFD for ozone depletion, we also examined the limits of CFFD for ozone depletion, based on the fitted CFFD of these 9 planet hosts via their Kepler light curves. The frequency of superflares with certain energy must not exceed 0.4 day⁻¹. Otherwise, such superflares would lead to ozone depletion on their planets.²¹ Figure 6 shows an example of the observed and fitted CFFD of KOI-8012, considering the constraints from the ozone depletion. For KOI-7706, its flare frequency could potentially cause ozone depletion on its planet, as the CFFD with energy of 1.35 × 10³⁴ erg is very close to the limit of 0.4 day⁻¹ (see the details in Figure S8).

Figure 6.

The cumulative FFD of KOI-8012

The blue dot represents the cumulative FFD (CFFD) directly from observations. The blue line illustrates the fitted CFFD, with two parameters (α_C and ${\beta }_{\mathrm{C}}^{\prime }$, in blue text) marked in the bottom left corner. The orange area denotes the ozone depletion forbidden region, where flare rates are ≥0.4 day⁻¹ with a flare bolometric energy larger than 0.58 × 10³⁴ erg. The cyan dot-dashed line represents the lower limit of flaring required by the abiogenesis flux at 0°C. The red dashed line represents the frequency for twice the NUV flux received by the Archean Earth.⁸ Note that the cyan dot-dashed and red dashed lines are only drawn in this figure for comparison with previous studies and are not used in the definition of the UV-HZ.

In summary, the habitability of the nine planets around Kepler flaring hosts is as follows.

• (1) One rocky planet (KOI-8012.01) and two super-Earths (KOI-8047.01 and KOI-7703.01) are in both the UV-HZ and the LW-HZ. The CFFDs of their host stars do not exceed the limit of ozone depletion. Thus, they provide radiation environments more suitable for abiogenesis.

• (2) Three planets (Kepler-1540 b, Kepler-438 b, and Kepler-155 c) are located in the LW-HZ. However, they are not within the UV-HZ defined by abiogenesis requirements under modeled surface temperatures. For Kepler-438 b and Kepler-155 c, if the temperature in some regions of their surface can be maintained at 0°C, the received UV radiation can satisfy the requirements for abiogenesis (see Figure S8 for further explanation).

• (3) KOI-7706.01 and Kepler-1512 b are not located in the LW-HZ or the UV-HZ. Moreover, the CFFD of KOI-7706 nearly reaches the threshold for ozone depletion, making these two planets less likely to support prebiotic chemistry.

• (4) KOI-5879.01 was originally within the LW-HZ and the UV-HZ if its host star were quiescent. However, flares from the host star cause an outward expansion of the LW-HZ, and it is therefore no longer within the HZ.

Discussion

Although many exoplanets have been studied statistically, assessing the habitability of individual planets in the HZ is still challenging from both astrobiological and observational perspectives. Evaluating HZs around stars in various aspects helps us better understand exoplanet habitability.^5,56 By re-evaluating the HZs and creating a comprehensive catalog of planets within them, we can infer that terrestrial planets in both LW-HZs and UV-HZs are more likely to support life.

In this paper, we highlight that the requirements of UV radiation for abiogenesis vary rapidly with temperature. When analyzing planets around flaring stars, we use a modeled temperature to estimate the inner boundary for UV-HZ. Note that the surface temperature is correlated with different aspects, e.g., the atmosphere effect, surface albedo, and thermal redistribution on the day and the night side.⁵⁷ Thus, Kepler-1540 b, Kepler-438 b, and Kepler-155 c deserve further investigation by large telescopes to better constrain their temperatures. Even KOI-7703.01, KOI-8047.01, and KOI-8012.01, which are among the most promising targets, need to be characterized in terms of their atmospheric properties. We also calculate the transmission spectroscopy metric (TSM) of these planets, as shown in Table 1.

Table 1.

Boundaries of the habitability zones for candidate and confirmed exoplanets orbiting Kepler flaring stars

Planet	aa (AU)	din, FLW (AU)	dout, FLW (AU)	din, FUV (AU)	Located inb	TSMc
Kepler-1540 b	0.4426 ± 0.0251	0.383	0.941	0.788	LW-HZ	2.271
KOI-7703.01d	1.0911 ± 0.0618	0.798	1.869	0.693	LW-HZ, UV-HZ	0.956
KOI-8047.01	0.8332 ± 0.0472	0.438	1.065	0.621	LW-HZ, UV-HZ	0.576
Kepler-155 c	${0.2254}_{-0.0139}^{+0.0092}$	0.211	0.533	0.333	LW-HZ	0.553
KOI-5879.01e	0.3027 ± 0.0172	0.327	0.835	0.179	–	0.332
Kepler-1512 b	${0.1097}_{-0.0079}^{+0.0045}$	0.117	0.304	0.164	–	0.059
Kepler-438 b	${0.1660}_{-0.0420}^{+0.0510}$	0.179	0.459	0.300	LW-HZ?f	0.059
KOI-7706.01g	0.1860 ± 0.011	0.288	0.719	0.185	–	0.028
KOI-8012.01	0.1219 ± 0.0069	0.062	0.162	0.069	UV-HZ, LW-HZ	0.006
The Archean Earthd	1.0	0.650	1.539	0.790	UV-HZ, LW-HZ	–

d_{in, FLW} and d_{out, FLW} denote the inner and outer boundaries of the flare-driven LW-HZ, respectively; d_{in, FUV} represents the inner boundaries of the flare-driven UV-HZ. Note that the outer boundaries of the flare-driven UV-HZ are set the same with d_{out, FLW}. (This table is available in its entirety in machine-readable form.)

^aFor KOI-8012.01 and KOI-7706.01, which lack semi-major axis uncertainty in the source reference, we applied a typical 17% uncertainty for stellar mass from Thompson et al.⁵⁸ to estimate the uncertainty via error propagation.

^bUV-HZ and LW-HZ represent the ultraviolet habitable zone and the liquid water habitable zone driven by flares, respectively.

^cThe transmission spectroscopy metric (TSM) is calculated following Kempton et al.⁵⁹ to assess the viability of follow-up exo-atmospheric observations.

^dKOI-7703 and the Archean Sun lack sufficient flare events to fit the CFFD. Therefore, the boundaries for scenarios without stellar flares are presented.

^eFor KOI-5879, a more appropriate integration range for flare energies was chosen from 10³⁰ to 10³⁸ erg (see footnote “b” of Table S3 for details).

^fMeasurement uncertainties in the semi-major axis prevent a clear conclusion of whether this planet resides within the LW-HZ (see Figure 5).

^gFlares on KOI-7706 may potentially deplete ozone layers of KOI-7706.01 (see Figure S8).

An averaged UV flux is adopted in this paper due to flare events. However, the influence of time-dependent enhancements in UV flux on life or chemistry needs to be compared with the continuum UV enhancements. Through the simulation of phenomena such as solar activity, Ranjan et al.⁶⁰ found that secular time-dependent variations in the UV band have little influence on prebiotic chemistry, but they also suggested caution when extrapolating from high-fluence regimes to natural conditions. By modeling flares via pulsed UV radiation, microorganisms can tolerate high fluences of UV radiation in quantities and at even shorter wavelengths not experienced by microorganisms on present-day Earth.⁶¹ Thus, our assumption of an averaged UV flux is reasonable. Of course, more quantitative studies on the influence of time-dependent UV radiation on prebiotic chemistry would be preferred to refine the average assumption.

To accurately model or determine the UV radiation of stars, we need more NUV observations of various stars, especially young stars. In the quiescent stage, a survey including a large sample of stars is necessary to characterize the NUV radiation correlated with different stellar parameters, e.g., rotation period, age, metallicity, and effective temperature. Secular monitoring, such as Kepler and TESS, would also be beneficial for characterizing the CFFD of flaring stars. Additionally, time-series observations or spectra during stellar flare events are crucial for understanding the physical processes and refining the SED model of flares.

Astrobiology experiments to explore the influence of UV flux on prebiotic chemistry are also crucial to determine the limits of UV radiation. UV radiation has both beneficial and detrimental effects on the accumulation of biological macromolecules such as RNA precursors. Extreme UV radiation can destroy probiotic chemistry. However, the formation of life is a process of selective evolution over long periods. Certain mechanisms protect these biomacromolecules, ensuring their survival at specific rates and probabilities, either in particular locations or within specific environments. Studying the shielding mechanisms^48,62 in different environments can help refine the UV requirements for abiogenesis. Furthermore, testing the amount of cumulative products over a longer timescale under a wide range of UV radiation levels is critical for habitability studies. The quantitative assessment of the effects of UV radiation on the origin of life remains an open issue, requiring more data and interdisciplinary collaborations.

After the origin of life, many other aspects can also influence the sustenance of life. For example, the synthesis of genetic molecules is influenced by various factors. X-rays, coronal mass ejections (CMEs),^63,64,65 and stellar energetic particles (SEPs)^66,67 can all cause water loss and affect planetary habitability via chemistry. Optimistically, when life exists on a planet, it can resist high UV radiation through mechanisms such as clumping and biofilm formation.⁴³ Therefore, further laboratory research is needed to study how microbes, such as Deinococcus radiodurans²¹ and tardigrades, perform in extreme UV environments to determine the upper limit of UV radiation for sustaining life.

Conclusion

Traditional HZs for liquid water are widely accepted for detecting habitable planets, while many works consider UV radiation to show its importance for life origins via prebiotic chemistry. Adopting the NUV (200–280 nm) requirements from previous experimental requirements of abiogenesis and considering the effects on chemical reaction rates at different temperatures, we constrain the inner boundary of UV-HZ naturally. After correcting the quiescent stellar NUV radiation for young stars (∼1 Gyr), based on the observation from GALEX, the UV-HZ can overlap with the LW-HZ in a narrow region for low mass stars. However, planets around these low-mass stars can hardly be located in the overlapped regions. To calculate the UV enhancement due to flares, we construct an SED model for flares and adopt an empirical model of CFFD correlated with different stars. We find that flaring stars can tolerate both LW-HZ and UV-HZ in the same region, much wider than the quiescent low-mass stars. Therefore, the flaring low-mass stars can sustain the habitability of planets in the overlapped HZ. Specifically, we investigate nine Kepler flaring stars and refine their LW-HZ and UV-HZ, according to the observed CFFD. One rocky planet (KOI-8012.01) and two super-Earths (KOI-8047.01 and KOI-7703.01) are located within both HZs, avoiding the erosion of ozone in the atmosphere. The other three planets, i.e., Kepler-1540 b, Kepler-438 b, and Kepler-155 c, deserve further investigation of their surface temperatures to assess their habitability. Our work provides a positive perspective on searching for habitable planets and potential life around flaring low-mass stars.

Resource availability

Materials availability

This study did not generate new unique materials or reagents.

Data and code availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the supplemental information. Additional data and code related to this paper are available from the lead contact upon request.

Acknowledgments

These results are based on observations obtained with Kepler. We are grateful to the Lightkurve⁶⁸ for the valuable assistance in gathering our Kepler data. We thank Song Wang, Xue Li, and Jia-Hui Wang from the National Astronomical Observatories, Chinese Academy of Sciences, for their helpful discussions on stellar NUV radiation and stellar ages. We thank the National Key Research and Development Program of China (2024YFA1611801), the National Natural Science Foundation of China (12150009), the China Manned Space Project (CMS-CSST-2025-A16), the Instrument Education Funds of Shandong University (yr20240205), the Civil Aerospace Technology Research Project (D010102), the Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China (JYB2025XDXM105), and the Shanghai Municipal Education Commission’s Artificial Intelligence Innovation Program for Empowering Discipline Development and Reforming Research Paradigms.

Author contributions

H.-G.L. proposed and designed this study. D.-Y.G., H.-G.L., and M.Y. investigated the observation data and pertinent literature. D.-Y.G. conducted the calculations and analysis. J.-L.Z. provided key suggestions on this work. All authors discussed the results and reviewed the manuscript.

Declaration of interests

The authors declare no competing interests.

Published Online: January 14, 2026

Lead contact website

Dong-Yang Gao: https://orcid.org/0000-0001-6643-2138; Hui-Gen Liu: https://orcid.org/0000-0001-5162-1753; Ming Yang: https://orcid.org/0000-0002-6926-2872; Ji-Lin Zhou: https://orcid.org/0000-0003-1680-2940.