EUV Late Phase in Solar Flares
- EUV late phase is a phenomenon defined by a secondary peak in warm coronal EUV emission (Fe XVI 33.5 nm) occurring 10–180 minutes after the main flare peak.
- It stems from larger, higher-lying loop systems with unique magnetic topologies and thermodynamic evolution, distinguishing it from the main-phase flare emission.
- This phase significantly influences solar-terrestrial interactions by contributing sustained coronal irradiance, impacting both energy partitioning and space weather forecasts.
The extreme-ultraviolet (EUV) late phase is a distinct phenomenon in solar flares, manifesting as a secondary peak in warm coronal EUV emission (typically Fe XVI 33.5 nm, formation temperature ≃ 2.5–3 MK), occurring tens of minutes to hours after the main soft X-ray (SXR) peak. First systematically identified using the EUV Variability Experiment (EVE) on SDO, this feature is characterized by unique magnetic topologies, loop system geometries, and thermodynamic evolution, with significant implications for coronal physics, magnetic reconnection models, solar-terrestrial impacts, and flare energy partition. The following sections elucidate the observational signatures and diagnostic criteria, magnetic and geometric context, quantitative energetics and timescales, physical mechanisms, variations and exceptions, and heliophysical and space weather relevance of the EUV late phase.
1. Definition, Diagnostic Criteria, and Observational Signature
EUV late phase flares are defined by a double-peaked irradiance profile in warm coronal lines such as Fe XVI 33.5 nm, with the secondary maximum (the "late phase") delayed by Δt ≃ 10–180 minutes relative to the SXR or hot EUV (e.g., Fe XX/XXIII 13.3 nm) peak (Liu et al., 2015, Ornig et al., 19 Aug 2025, Chen et al., 2020). This late-phase emission:
- Is seen as a distinct peak in warm (T ≃ 2–4 MK) coronal lines, without a corresponding enhancement in GOES SXR, and with no significant maximum in hotter lines (Fe XX, Fe XXI, >10 MK).
- Originates from a physically distinct system of longer, higher-lying loops, spatially resolvable in SDO/AIA images (Sun et al., 2013, Hock et al., 2012).
- Meets quantitative criteria: (i) secondary peak >30% of main-phase maximum; (ii) delays >10 min; (iii) intervening local minimum; (iv) no hot-line increase near the secondary peak; (v) spatial association with the same active region (Ornig et al., 19 Aug 2025).
- The amplitude ratio (late/main) ranges widely: 0.3–5.9, with "extreme" ELP defined as ratio > 1 (late-phase peak exceeds main-phase) (Wang et al., 9 Dec 2025).
A typical event displays:
- Main-phase (impulsive and early gradual) peak in Fe XX/XXIII and GOES SXR.
- Delayed late-phase peak in Fe XVI 33.5 nm, with corresponding cooling sequence traced in AIA 335/211/193/171 Å.
- Absence of SXR or hot-EUV enhancement during the late-phase peak (Liu et al., 2015, Zhang et al., 2022).
2. Magnetic and Geometric Context
EUV late phase events are inextricably tied to multipolar active-region magnetic topologies, typically involving two (or more) distinct loop systems connected by reconnection at null points or quasi-separatrix layers (QSLs):
- Main-phase arcade: Compact, lower, bipolar system (footpoint separation ≃ 2.1–2.8×10⁴ km, field ≃ 1000–1200 G), produces impulsive SXR/hot EUV emission.
- Late-phase arcade: Much larger, higher-lying system (separation ≃ 7.2–8.7×10⁴ km, field ≃ 600–800 G), connected at one footpoint to the main region—a so-called “asymmetric magnetic quadruple” or fan–spine topology (Liu et al., 2015, Sun et al., 2013, Li et al., 24 Oct 2024).
Circular-ribbon and two-ribbon flares, as classified by flare ribbon morphology in AIA 1600 Å, both occur, with ELP loops being systematically longer and higher. In circular-ribbon cases, a “dome-plate” QSL is common, with the dome enclosing compact “fan” loops (main phase) and the plate QSL (sometimes with spine linkage) sourcing the longer late-phase loops (Chen et al., 2020, Masson et al., 2017).
NLFFF and magnetic extrapolations confirm that late-phase loops trace either null-point–associated field lines (fan–spine or plate QSLs) or overlying extended arcades in quadrupolar configurations (Hock et al., 2012, Masson et al., 2017, Sun et al., 2013, Zhong et al., 2021).
3. Quantitative Energetics, Timescales, and Loop Cooling Physics
The core physical distinction arises from the differing physical scales and thermodynamic evolutions of the two loop systems:
- Main-phase loops: Short (L ≃ 1–2.5×10⁴ km), high initial T (>10 MK), densities up to 3×10¹⁰ cm⁻³, fast conductive/radiative cooling (minutes).
- Late-phase loops: Much longer (L ≃ 7–18×10⁴ km), initial T ≃ 2–10 MK, lower densities (1–10×10⁹ cm⁻³), much slower cooling (tens of minutes to hours) (Liu et al., 2015, Zhang et al., 2021, Wang et al., 9 Dec 2025).
The cooling time of late-phase loops is well-approximated by the formula [Cargill et al. 1995, (Liu et al., 2015, Chen et al., 2020)]:
where is the loop half-length [cm], is initial electron temperature [K], and is density [cm⁻³].
For L ≃ 7–14×10⁹ cm, T₀ ≃ 2.5–14 MK, n₀ ≃ 2–10×10⁹ cm⁻³, predicted τ_cool ≃ 60–120 min, matching observed ELP delays (Liu et al., 2015, Li et al., 24 Oct 2024, Sun et al., 2013, Liu et al., 2015). Loop cross-sectional expansion with apex height (Γ = A_apex/A_base ≃ 1.3–3.0) further retards the conductive cooling, as the “funnel-shaped” geometry reduces heat and mass leakage, yielding “longer-than-expected” τ_cool without requiring additional heating (Li et al., 24 Oct 2024).
Energy partition analysis in confined flares shows total heating inputs for late-phase loops erg, with peak thermal energies erg and radiative output in the 70–370 Å band exceeding erg (Zhang et al., 2021). In “extremely energetic” events, >4× more energy can be radiated in the 335 Å passband during the late phase than during the main peaks (Wang et al., 9 Dec 2025).
4. Physical Mechanisms: Cooling versus Secondary Heating
There are two principal mechanisms for the EUV late phase:
4.1. Long-Lasting Cooling (Simultaneous Heating; “Cooling Scenario”)
- The late-phase loops are energized essentially simultaneously with the main phase (as traced by near-coincident footpoint brightenings and remote ribbon timings), but due to their greater length, cool much more slowly, producing the delayed warm-EUV emission (Liu et al., 2015, Hock et al., 2012, Chen et al., 2020).
- In the classic scenario, a single impulsive heating episode is followed by a monotonic cooling (conductive, then radiative), with no requirement for a distinct second energy input at late times (Dai et al., 2018, Zhong et al., 2021).
4.2. Delayed/Episodic Secondary Heating (Additional Energy Input)
- In some flares, especially those with “extreme” late phases (late/main >1), the late-phase peak cannot be solely attributed to cooling; model-data comparisons and DEM inversions reveal excess emission, density plateaus, or temperature “jumps” during the decay (Liu et al., 2015, Wang et al., 9 Dec 2025, Dai et al., 2018).
- Direct signatures: Re-brightening and secondary expansion of remote flare ribbons in AIA 1600 Å, smooth delayed microwave enhancements at footpoints (as analyzed with EOVSA imaging spectroscopy), or sustained brightenings in hot-bandpass and DEM data (Zhang et al., 2022).
- Quantitatively, delayed secondary heating injected into the long loops produces a fast rise and slow decay in the Fe XVI light curve, contrasted with the slow-rise-fast-decay expected in the pure-cooling scenario (Dai et al., 2018).
In select cases (e.g., non-eruptive sigmoids or failed flux-rope eruptions), “continuous reconnection” in the multipolar topology injects persistent energy into the long loops, leading to an extremely energetic late phase with observed emission far exceeding what is predicted by cooling models alone (Wang et al., 9 Dec 2025, Li et al., 24 Oct 2024).
5. Variability: Statistical Properties, Morphological Classes, and Exceptions
Statistical analyses confirm that ELP events are infrequent but non-negligible:
- 10 % of all C3.0+ flares (2010–2014) show a late phase, with frequency decreasing from solar minimum (~15 %) to maximum (~5 %) and anti-correlating with sunspot number (Pearson’s ρ = –0.75) (Ornig et al., 19 Aug 2025, Woods, 2014).
- The majority (67 %) of ELP flares are confined (no CME); the remainder are eruptive. “Extreme” ELPs are more commonly seen in confined M- or X-class flares (Chen et al., 2020, Ornig et al., 19 Aug 2025).
- The relative strength R = late/main peaks: mean 1.54 (±0.97) (Ornig et al., 19 Aug 2025).
- A robust linear relationship exists between the late-phase delay Δt and its duration D (D ≃ 1.42 Δt + 0.11 h; Pearson’s ρ = 0.81) (Ornig et al., 19 Aug 2025).
Morphological diversities:
- The “plateau-like” late phase results from the superposition of multiple long-loop systems of graded length, whose cooling timescales spread the emission into a broad, flat maximum rather than a singular peak (Chen et al., 2023).
- In multi-stage reconnection events (e.g., X2.1 on 2011 Sep 6), the late phase maps to the least energetic, most extended episode of reconnection, forming loops only heated to warm coronal temperatures (Dai et al., 2013).
- Rarely, extremely energetic ELPs (late/main > 4) are seen in C-class flares involving a non-eruptive sigmoid undergoing continuous reconnection (Wang et al., 9 Dec 2025).
6. Diagnostic and Modeling Tools
Rigorous identification and physical interpretation of ELP flares require coordinated analysis:
- Irradiance and Imaging: SDO/EVE provides disk-integrated spectral time series; SDO/AIA offers high-cadence, multi-band imaging for spatial and thermal diagnosis.
- Loop Diagnostics: Differential emission measure (DEM) inversion characterizes T–n–EM structure; 1D/0D hydrodynamic loop modeling (EBTEL) quantifies cooling times and heating profiles (Hock et al., 2012, Li et al., 2014, Dai et al., 2018).
- Magnetic Field Extrapolation: NLFFF models reconstruct coronal connectivity, field topology, and reconnection geometry (Masson et al., 2017, Zhong et al., 2021, Wang et al., 9 Dec 2025).
- Diagnostic Criteria: Slope ratio of Fe XVI late-phase light curve (rise/decay times), footpoint ribbon timing, and magnetic diagnostics (QSL mapping) distinguish cooling-only from secondary-heating scenarios (Dai et al., 2018, Li et al., 2014).
- Space Weather Impact: Modeling requires inclusion of warm coronal irradiance proxies for accurate ionospheric response prediction (Bekker et al., 19 Jul 2024).
7. Broader Heliophysical and Space-Weather Significance
The EUV late phase has substantial relevance beyond flare physics:
- Coronal Heating and Magnetic Reconnection: ELP events demand reconnection models that accommodate multi-system, simultaneous energy release and non-trivial magnetic connectivity (asymmetric quadruple, dome-plate QSL, null-point, fan–spine configurations) (Liu et al., 2015, Masson et al., 2017, Sun et al., 2013).
- Energy Partitioning: In both confined and eruptive cases, late-phase emission is an order-of-magnitude less energetic than total SXR output but can dominate the coronal radiative budget in warm lines, affecting thermal energy distribution (Zhang et al., 2021).
- Space Weather: ELP-induced irradiance lasts hours after the SXR peak, often accounting for up to a third of the total ionospheric TEC enhancement in large flares (Bekker et al., 19 Jul 2024). Conventional models relying solely on GOES X-ray proxies underestimate the sustained geoeffective EUV input, necessitating warm-EUV monitoring for ionospheric and thermospheric modeling (Ornig et al., 19 Aug 2025).
- Solar-Cycle Dependence: ELP occurrence is anti-correlated with sunspot number and more frequent near solar minima, reflecting changing AR complexity and magnetic connectivity (Ornig et al., 19 Aug 2025, Woods, 2014).
The EUV late phase represents a confluence of coronal magnetic complexity, multi-timescale energy release, and large-scale magnetic coupling, providing stringent observational constraints on reconnection physics and coronal-loop thermodynamics, as well as playing a critical role in Earth's near-space environment.