Observation of GRB 221009A Early Afterglow in X-Ray/Gamma-Ray Energy Bands

The early afterglow of a gamma-ray burst (GRB) can provide critical information on the jet and progenitor of the GRB. The extreme brightness of GRB 221009A allows us to probe its early afterglow in unprecedented detail. In this Letter, we report comprehensive observation results of the early afterglow of GRB 221009A (from T 0+660 s to T 0+1860 s, where T 0 is the Insight-HXMT/HE trigger time) in X-ray/gamma-ray energy band (from 20 keV to 20 MeV) by Insight-HXMT High Energy X-ray Telescope, GECAM-C, and Fermi/Gamma-ray Burst Monitor. We find that the spectrum of the early afterglow in 20 keV–20 MeV can be well described by a cutoff power law with an extra power law that dominates the low- and high-energy bands, respectively. The cutoff power law E peak is ∼30 keV, and the power-law photon index is ∼1.8 throughout the early afterglow phase. By fitting the light curves in different energy bands, we find that a significant achromatic break (from keV to TeV) is required at T 0 + 1246−26+27 s (i.e., 1021 s since the afterglow starting time T AG = T 0+225 s), providing compelling evidence of a jet break. Interestingly, both the pre-break and post-break decay slopes vary with energy, and these two slopes become closer in the lower energy band, making the break less identifiable. Intriguingly, the spectrum of the early afterglow experienced a slight hardening before the break and a softening after the break. These results provide new insights into the physics of this remarkable GRB.


Introduction
Gamma-ray bursts (GRBs), the cosmological electromagnetic events produced by stellar deaths, are the most violent explosions in the Universe (e.g., Zhang 2018).They can be simply divided into short GRBs and long GRBs by their duration of prompt emission, which is generally produced by ultra-relativistic and collimated outflow (jet) launched from binary neutron star mergers or massive star core collapses, respectively (Eichler et al. 1989;Narayan et al. 1992;Woosley 1993;Woosley & Bloom 2006;Abbott et al. 2017).While the jet propagates away from the central engine, the prompt emission is produced within the jet by internal shock or magnetic dissipation (Rees & Meszaros 1994;Zhang & Yan 2011), and multiwavelength afterglow is radiated when the jet sweeps the external medium, producing reverse shock and forward shock (see Gao et al. 2013 for a review).
In contrast to the irregular behavior in the prompt emission, the afterglow emission usually can be well described with several power-law decays.Sometimes the break between power-law decays is achromatic (i.e., showing up in different energy bands at the same time), called a jet break, which is directly related to the jet geometry and observer's viewing angle (Panaitescu & Kumar 2003;Liang et al. 2008;Wang et al. 2018).Generally, when the jet opening angle is small, the jet break occurs earlier in the light curve, making the multiwavelength observation of the break very challenging.
As the brightest GRB event since the discovery of GRBs in the 1960s, GRB 221009A was first reported by Fermi/Gammaray Burst Monitor (GBM) while many other gamma-ray telescopes detected it simultaneously but without real-time alert.With the early accurate location provided by Swift/Burst Alert Telescope (Burst Alert Telescope (BAT); Dichiara et al. 2022), follow-up observations have been made by numerous multiwavelength and multi-messenger telescopes (e.g., An et al. 2023;Burns et al. 2023;Kann et al. 2023;Laskar et al. 2023;Levan et al. 2023;LHAASO-Collaboration et al. 2023;Malesani et al. 2023a;Negro et al. 2023;Williams et al. 2023).The redshift of this burst is found to be z = 0.151 (de Ugarte Postigo et al. 2022;Malesani et al. 2023b).
The Insight-Hard X-ray Modulation Telescope (Insight-HXMT) was triggered by GRB 221009A at 13:17:00.050UT on 2022 October 9, (denoted as T 0 )11 during a routine ground search of bursts.With the novel design dedicated to extremely bright bursts, GECAM-C made a uniquely accurate measurement of GRB 221009A prompt emission with its unsaturated data and found that this GRB has the highest total isotropicequivalent energy (E γ , iso = 1.5 × 10 55 erg; An et al. 2023), which is confirmed by other observations (e.g., Burns et al. 2023;Frederiks et al. 2023;Lesage et al. 2023;Rodi & Ubertini 2023).Interestingly, a jet break at T AG + -+ 670 110 230 s is reported in the TeV range (LHAASO-Collaboration et al. 2023), where T AG is the afterglow starting time. 12Joint observations of early afterglow by Insight-HXMT and GECAM-C also show there is a jet break around T AG + -+ 950 50 60 s in the keV-MeV range; however, a detailed analysis of the jet break is difficult due to observation gaps of these two instruments (An et al. 2023).
In this paper, we performed a comprehensive analysis of the GRB 221009A early afterglow (from T 0 +660 s to T 0 +1860 s ) in X-ray and gamma-ray energy band (20 keV-20 MeV) with joint observations of the Insight-HXMT High Energy X-ray Telescope (HE), GECAM-C Telescope, and the Gamma-ray Burst Monitor of Fermi Gamma-ray Space Telescope.We present the detailed observations of the three instruments and the data analysis in Section 2. Then temporal and spectral analysis results are reported in Section 3. Discussion and conclusions are given in Section 4.
Please note that all parameter errors in this work are for 68% confidence level if not otherwise stated.

Insight-HXMT/HE Observation
As China's first X-ray astronomy satellite, Insight-HXMT has made many discoveries on high-energy objects since it was launched to the orbit of an altitude of 550 km and an inclination of 43°on 2017 June 15 (Zhang et al. 2018;Li et al. 2020;Zhang et al. 2020;Li et al. 2023).The main detector of the HE, consisting of 18 NaI(Tl)/CsI(Na) phoswich scintillation detectors, has played an important role in the observation of GRB 221009A in the MeV range (An et al. 2023).
GRB 221009A was visible to HE before the satellite entered the Earth's shadow.However, the effective observation of GRB 221009A afterglow is up to about T 0 +900 s, when Insight-HXMT began to enter the high-latitude region where the HE detectors suffer high background.The high fluence of charged particles in the high-latitude orbit region had a significant impact on observation data (An et al. 2023).The incident angle of GRB 221009A to HE detector is shown in Figure 1(a) during the early afterglow phase, when Insight-HXMT was executing a Galactic (Milky Way) plane scan observation.
Excluding the time region of background contamination, HE data from T 0 +660 s to T 0 +900 s are selected for this analysis, as shown in Figure 2(a).The gray solid lines represent the afterglow light curves in the 600-3000 keV energy band, and the corresponding background (black lines) is estimated by a parametric model, which makes use of some measured parameters to reconstruct the background (You et al. 2021;Liao et al. 2023).For the spectral analysis, the HE data in the good time interval (GTI) is split into two time-resolved spectra as shown in Table 1, which are used for joint analysis with Fermi/GBM data.

Fermi/GBM Observation
As one of the two instruments on board the Fermi Gammaray Space Telescope, the GBM is composed of 14 detectors with different orientations: 12 sodium iodide (NaI) detectors (labeled from N0 to N11) covering the energy range of about 8-1000 keV and 2 bismuth germanate (BGO) detectors (labeled as B0 and B1) covering energies about 0.2-40 MeV (Bissaldi et al. 2009;Meegan et al. 2009).Fermi/GBM was also triggered by GRB 221009A, with the observation results initially reported on the General Coordinates Network (GCN) (Veres et al. 2022) and detailed analysis presented in Lesage et al. (2023).
Since GRB 221009A is extremely bright, GBM suffered pulse pileup and data saturation during the prompt-emission and flare-emission phase (Lesage et al. 2023).Fortunately, the GBM observation of the early afterglow was unaffected by these effects and thus could provide valuable data until it entered the Earth's shadow at about T 0 +1467 s.
The GBM time-tagged events data are performed for temporal and spectral analysis, with two NaI detectors (NaI 4, NaI 8) and two BGO detectors (BGO 0, BGO 1).The incident angles of these four detectors are shown with the red solid lines in Figures 1(b) to (e).The green and blue dashed lines indicate the nearest revisit orbit geographically (with a time shift of ±85610 s with respect to the current orbit of the detection of GRB 221009A), showing that the detector pointing directions were nearly identical.Since the environmental irradiation background was also stable, the count rate in the revisit orbit can be used to estimate the background of the observation of GRB 221009A.This method of background estimation has been proven to be successful for GECAM-C (An et al. 2023).Indeed it is also validated by the GBM light curve itself before T 0 and after the Earth occultation (see Figures 2(b)-(e)).
The light curves from T 0 −300 s to T 0 +2000 s for N4, N8, B0, and B1 are plotted in Figures 2(b)-(e).The gray solid lines represent the total light curve, and the black solid lines represent the background light curve averaged from the revisit orbits.As shown in red lines, we also tried a background fit with the same 4th-order polynomial with the same background time intervals as Figure 3 of Lesage et al. (2023).Because the 4th-order polynomial function is unable to accurately describe the background after the flare, our averaged revisit orbit background seems to be more appropriate to describe the afterglow background.Considering the influence of the Sco X-1 and charged-particle events, we omit data below 200 keV for NaI 4. The NaI 8 data below 20 keV are also omitted (Lesage et al. 2023).BGO data below 400 keV are also ignored since the detector response at low energies is affected by the photomultiplier tubes and their housings at the angle between source and BGO detector, approaching 90 degrees (see Figure 2(d),(e); Lesage et al. 2023).For the spectral analysis, three time-resolved spectra, as shown in Table 1, are extracted from T 0 +660 s to T 0 +1400 s with GBMDataTools 13 (Goldstein et al. 2022).Additionally, finer time-resolved spectra with a time interval of 40 s are achieved with GBMDataTools from T 0 +660 s to T 0 +1420 s, which are used to characterize the spectra evolution in the 20-300 keV energy band.
Thanks to the working mode and design dedicated to an extremely bright burst observation, the GECAM-C made an accurate measurement of GRB 221009A and provided unique data in the main burst, flare, and early afterglow until it entered the Earth's shadow at about T 0 +1860 s (Liu et al. 2022).From T 0 +640 s to T 0 +1350 s, the GECAM-C was over the highlatitude region (An et al. 2023), where the flux of chargedparticle events was extremely high and contaminated some of the observation data (An et al. 2023).Consequently, we select clean data of GRD05 from T 0 +1350 s to T 0 +1860 s for temporal and spectral analysis.The incident angle of GRD05 was stable during this time range as shown in Figure 1(d).The light curves of GRD05 are displayed in Figure 2(f), where gray lines are the total light curves and black solid lines are the background light curves using revisit orbit data (An et al. 2023; C.-W. Wang & Xiong 2023, in preparation).The total light curves agree well with the background light curves after GECAM-C entered the Earth's shadow, supporting that the revisit orbit background is appropriate.

Spectral Analysis
The observation data are divided into four parts, and the time-resolved spectra are obtained (see Table 1).For the spectral analysis, the single power law (PL; Equation (1)) model and cutoffpl (CPL; Equation (2)) model and Band model (Equation ( 3)) is performed to describe the spectral features: where K is the normalization amplitude constant in units of photons per square centimeters per second per kilo-electronvolt (photons cm −2 s −1 keV −1 ) at 1 keV and α is the dimensionless photon index of the PL.
where K is the normalization amplitude constant in units of photons per square centimeters per second per kilo-electronvolt (photons cm −2 s −1 keV −1 ) at 1 keV, α is the PL photon index, and E cut is the e-folding energy of exponential roll-off in keV.
where K is the normalization amplitude constant in units of photons per square centimeters per second per kilo-electronvolt (photons cm −2 s −1 keV −1 ), α is the first PL photon index, β is the second PL photon index, and E c is characteristic energy in keV.
The package of Xspec (V12.12.1;Arnaud 1996) is utilized to perform the analysis of those spectra.The X-ray/gamma-ray spectra in 20-20,000 keV energy band are fitted with the PL, CPL, and Band models respectively.As can be seen from Figure 3, from T 0 +700 s to T 0 +900 s, the spectra fitting results show a significant structure in residual distribution and excess in the low-energy band (see Figures 3(d), (e), and (f)).Consequently, we adopt a combined model (CPL+PL) to fit spectra.The combined model fitting results show much better  ).From T 0 +660 s to T 0 +1800 s, the spectra fitting results are summarized in Table 1.The CPL E peak (∼30 keV) seems to be relatively stable during this time range.The PL model photon indices are also consistent with Insight-HXMT/HE results (∼1.7 from T 0 +660 s to T 0 +900 s) within error (An et al. 2023).
Since the source intensity decays with time and the effective area of the detector is minor, in order to guarantee sufficient statistics to obtain meaningful spectra, the detailed analysis of time-resolved spectra is only carried out in the energy range from 20 keV to 300 keV.We generate 19 time-resolved spectra using Fermi/GBM N4 and N8 detector data from T 0 +660 s to T 0 +1420 s with the same time bin width of 40 s and 8 timeresolved spectra using GECAM-C GRD05 detector data from T 0 +1350 s to T 0 +1750 s with the same time bin width of 50 s.The spectra data from 20 to 300 keV can be described well with a single PL model.The photon indices are shown in Figure 4.The blue diamonds and magenta points represent the Fermi/ GBM (N4 and N8) spectra results and GECAM-C (GRD05) spectra results, respectively.
In addition, the source spectrum can also be measured by Earth occultation analysis (Xue et al. 2023).For Fermi/GBM, the Earth occultation started at ∼T 0 +1430 s and ended at ∼T 0 +1467 s, while it started at ∼T 0 +1810 s and ended at ∼T 0 +1860 s for GECAM-C.According to Earth occultation analysis, the photon index of PL from Fermi/GBM is plotted with a cyan diamond in Figure 4 and that of GECAM-C with a red dot.We note that the photon index measured by these two methods and two instruments is well consistent with each other, demonstrating that the background model used in this work is reasonably reliable.
From the evolution of the photon index in the energy range of 20-300 keV in Figure 4, it seems that the early afterglow first experienced a hardening and then a softening in the spectrum.

Flux Light-curve Analysis
According to the results of the spectral analysis conducted with time resolution, the flux light curves in units of erg per square centimeters per second (erg cm −2 s −1 ) are calculated in five energy bands: 20-50 keV, 50-100 keV, 100-150 keV, 150-200 keV, and 200-300 keV.With the starting time of the afterglow at T AG = T 0 + 225 s, as shown in Figure 5, the blue points represent Fermi/GBM flux, and the green points represent GECAM-C flux in each panel.It is worth noting that the measured flux from both instruments is highly consistent with each other during the overlapping time period, demonstrating that the calibration and background model used in this work are reasonably reliable for Fermi/GBM and GECAM-C.
The flux light curves from T AG +525 to T AG +1635 s are fitted by the single PL model (Equation 4) and broken PL (BPL) model (Equation 5; Zhang et al. 2006), with the emcee software package (Foreman-Mackey et al. 2013), one by one.The free parameters of the broken PL model include amplitude A BPL , break time t j , and two slope indices α 1 and α 2 .The posterior distributions of the broken PL model and break time are plotted in different energy bands (see Figure 5).Detailed  fitting results for each energy band are listed in Table 2 The significance of the break was estimated by the likelihood ratio (LR) test method (e.g., LHAASO-Collaboration et al. 2023).All LR significances for > 50 keV energy bands are higher than 3σ (50-100 keV: 3.8σ; 100-150 keV: 6.4σ; 150-200 keV: 4.4σ; and 200-300 keV: 4.5σ), demonstrating that a break is required significantly.Although the LR significance in the 20−50 keV energy band is only 1.9σ, the χ 2 /d.o.f value of the broken PL (BPL) model (1.15) is clearly preferred over the PL model (1.37).We note that this lower significance in low-energy bands than higher energy bands seems to be understandable and expected from the trend of the energy dependence of the pre-break and post-break slopes (Figure 7).Therefore, we conclude that the BPL model is statistically more suitable to describe the flux decay.We also plot the summed flux light curves in the 50-300 keV and fit with the BPL model from T AG +525 to T AG +1635 s.The Markov Chain Monte Carlo fit results are shown in Figure 6.
The pre-break, post-break, and break time parameters in the keV-MeV and TeV energy bands are shown in Figure 7    We note that the break time seems to have a marginal dependence on the energy band.We testify this possible dependence by fitting the break time as a function of energy with the PL model and constant model.The ΔBIC (BIC PL −BIC const ) of these two models is only −1.1, meaning that there is no statistical evidence for the energy dependence of the break time.Consequently, we conclude that the break time remains constant in the very wide energy band from keV to TeV, providing compelling evidence for the origin of jet break.Interestingly, we find that the light-curve decay slope varies with energy from keV to TeV in an opposite way: the pre-jet break slope α 1 increases with energy, while the post-jet break slope α 2 decreases with energy, as shown in Figure 7.
Following the calculation in An et al. (2023), the jet opening angle can be inferred with (Zhang 2018) , where h g ˜is the ratio between isotropic γ-ray and isotropic kinetic energies and n is the ambient number density in units of per cubic centimeter (cm −3 ), which are left as free parameters.

Discussion and Conclusion
GRB 221009A is the brightest-of-all-time (BOAT) GRB ever detected.In this Letter, we report a comprehensive temporal and spectral analysis on its early afterglow in the keV-MeV energy band (20 keV-20 MeV) from T 0 +660 s to T 0 +1860 s with combined observation data from Insight-HXMT, GECAM-C, and Fermi/GBM.We find that the spectrum of early afterglow in 20 keV-20 MeV can be well described with the CPL+PL model, where the CPL and PL dominate the low-energy and high-energy bands, respectively.Both the CPL E peak (∼30 keV) and the PL photon index (∼1.8)seem generally stable from T 0 +660 s to T 0 +1860 s.We also note that the CPL E peak and low-energy photon indices are broadly consistent with the later afterglow results (E break = 33.6 - + 28.6 16.1 keV, Γ = 1.7 @ T 0 +4.2 ks) within the error (Williams et al. 2023; Tables 1 and 2).Due to the relatively large error on E peak , it is unclear whether and how the E peak evolves with time.Furthermore, the high-energy PL component may also be detected by Swift/BAT in later afterglow because the spectrum residual ratio shows clear evidence for the existence of an excess in the Swift/BAT energy band (Williams et al. 2023; the right panel of Figure 5).This CPL+PL spectrum shape is very unusual in GRB afterglow since the synchrotron radiation mechanism usually gives rise to a band-shape or PL-shape spectrum.But the GRB 221009A early afterglow shows a "V" shape spectrum above 20 keV.Indeed, this kind of spectrum shape was also reported in GRB 990123 afterglow (Corsi et al. 2005;Maiorano et al. 2005).However, the high-energy spectrum from GRB 990123 afterglow was only detected up to 60 keV and at a relatively later time (from ∼20 minutes to 60 minutes after the GRB trigger), while the early afterglow of GRB 221009A was measured in an extremely wide band (including keV, MeV, GeV, and TeV energy bands) and at a very early stage (immediately after the prompt emission and flare episode; e.g., An et al. 2023;Lesage et al. 2023;LHAASO-Collaboration et al. 2023;Zhang et al. 2023b).Further promoted by the exceptional brightness of this GRB, the detailed and accurate measurement of the early afterglow spectrum of GRB 221009A is truly unprecedented in GRB observation history.The origin of this early afterglow spectrum is intriguing and subject to further studies.
For the refined spectral analysis with a shorter time bin, a single PL model can describe the afterglow spectra well in 20-300 keV energy range, based on which we estimate the flux light curves in five energy bands.The PL photon index (∼−2.1) is in good agreement with Swift/BAT results (Γ = −2.08 + -0.03 0.03 @ T 0 +3302 ∼T 0 +4538 s; Williams et al. 2023).
Although the wide band (from 20 keV to 20 MeV) spectral shape (CPL+PL) of the early afterglow did not change significantly, the spectral hardness experienced a pre-break hardening and post-break softening (Figure 4), which is somewhat consistent with the spectral evolution in the TeV energy band (LHAASO-Collaboration et al. 2023, Figure 3).
A significant and achromatic break is found at t j =T AG + -+ 1021 26 27 s in the 50-300 keV light curve of early afterglow, supporting that this is a jet break.Remarkably, this jet break time in 50-300 keV is well consistent with that in TeV energy band (LHAASO-Collaboration et al. 2023).More interestingly, we find that the pre-break and post-break slopes of the flux light curve vary with energy (from keV to TeV) clearly.These two slopes become closer as energy decreases, making the break less identifiable in the low-energy band.Remarkably, our results show that the pre-break and post-break slopes tend to converge to a value of about −1.5 in the low-energy band (20-50 keV), which is well consistent with the pre-break slope of −1.498 - + 0.004 0.004 reported in the soft X-ray energy band (0.3-10 keV; Williams et al. 2023).We therefore suggest that this jet break feature may disappear in the lower energy band (i.e., soft X-ray and even lower energy band), thus forming an apparent peculiar break feature.Although the observations of a break in the X-ray but not in the optical energy band have been reported (Panaitescu et al. 2006;Racusin et al. 2008), here for the first time, we reveal the detailed transition from the break feature in higher energy to the non-break in lower energy.On the other hand, the energy-dependent slopes before and after the jet break are not expected from the ideal top-hat jet model (Zhang 2018); thus, this observation strongly disfavors the simple top-hat jet model and calls for a more sophisticated jet structure.
We note that, compared to our finding of the jet break at an early time (T AG + ∼1021 s), there are several breaks in later time reported in the multiwavelength observation of GRB 221009A afterglow.For example, a break at -+ 7.9 10 s 1.9 1.1 4 is reported in the soft X-ray light curve (Williams et al. 2023), while an achromatic break at ∼0.6 day is claimed in the optical bands (Shrestha et al. 2023).While these breaks are measured in a relatively limited energy band, our jet break reported here is seen in a very wide energy band from keV to TeV band.We also note that, for our jet break, the post-break slope in highenergy band (200-300 keV) is close to −2, which is expected by the afterglow theory (Zhang et al. 2006), while the postbreak slope in soft X-ray and optical is much shallower than theory expectation (Shrestha et al. 2023;Williams et al. 2023).Moreover, the jet break at a much earlier time indicates a very narrow jet core, which is well concordant with the extremely luminous the GRB.

Summary
In this work, we have performed an unprecedented and detailed temporal and spectral analysis of the early afterglow (from T 0 +660 s to T 0 +1860 s) of the BOAT GRB 221009A in the X-ray/gamma-ray energy band with joint observation of GECAM-C, Fermi/GBM, and Insight-HXMT.We discovered many features in the early afterglow of this exceptional burst, including unusual spectrum (CPL+PL, basically a V shape from 20 keV to 20 MeV), achromatic jet break time (from keV to TeV), and energy-dependent flux decay slopes before and after the jet break, etc.We point out that these observational features and properties may set a new benchmark for GRBs, thanks to the exceptional brightness and observation coverage of this GRB 221009A.We argue that these results strongly disfavor the simple top-hat jet model but provide support to the structured jet scenario with a bright narrow core.This is also consistent with the statistical analysis results by comparing GRB 221009A with the entire GRB sample (Lan et al. 2023).In this structured jet model, the narrow jet core with ultrarelativistic Lorentz factor should be surrounded by a wider jet with a lower Lorentz factor, leading to complicated temporal and spectral features (e.g., Panaitescu et al. 2006;Racusin et al. 2008).A joint analysis of the early and late afterglow of GRB 221009A could further deepen our understanding of GRB and jet physics (e.g., Gill & Granot 2023;O'Connor et al. 2023;Ren et al. 2023;Sato et al. 2023;Zhang et al. 2024).

Figure 2 .
Figure 2. Left: the gray lines in all panels are total light curves of Insight-HXMT/HE (0.6-3 MeV), Fermi/GBM N4 (20-900 keV), N8 (20-900 keV), B0 (0.4-45 MeV), B1 (0.4-45 MeV), and GECAM-C GRD05 (20-6000 keV) during the observation of GRB 221009A.The black light curves are the background of the parametric model (a) and the revisit orbit ((b)-(f)) after the trigger time when the satellites were almost in the same orbital locations and orientations.The red lines show the background fitted with the 4th-order polynomial.The orange shaded area in each panel marks the time windows that are confirmed as the early afterglow and are used for spectral analysis.Right: the net light curves of GRB 221009A, for which the background is estimated from the parametric model (a) and the averaged revisit orbital background ((b)-(f)).

Figure 3 .
Figure 3. (a) The count rate spectra in 20-20,000 keV energy band from T 0 +700 s to T 0 +900 s.(b) The spectral energy distribution (vF v ) of CPL+PL model.(c)-(f) The residual distribution of CPL+PL, PL, CPL, and Band model.

Figure 4 .
Figure 4.The PL photon indices of the early afterglow from T 0 +660 s to T 0 +1870 s are derived from Fermi/GBM N8 and GECAM-C GRD05 in 20-300 keV energy range.It shows that the photon index first experienced hardening before the break and then softening after the break.

Figure 5 .
Figure 5.The early afterglow flux evolution from T AG +435 s to T AG +1635 s are given by Fermi/GBM and GECAM-C in five energy bands, namely 20-50 keV, 50-100 keV, 100-150 keV, 150-200 keV, and 200-300 keV from panels (a) to (e).The gray shaded area represents the fitting range.The orange lines show the posterior distribution of BPL, and the gray vertical lines show the posterior distribution of break time.The red dotted lines show the flux decay with the pre-break slope.
with error bars of 1σ uncertainty.Interestingly, the break time is well consistent among all energy bands.Fitting the light curve in 50-300 keV yields the break time t j = T AG + note that this break time derived in the present work is consistent with the results given by the joint analysis of Insight-HXMT/HE and GECAM-C ( keV;An et al. 2023).Our result is also supported by a recent independent study with Fermi/GBM data only(Zhang et al. 2023b).Furthermore, we also fit the flux light curve in the TeV energy band from T AG + 100 s to T AG + 3000 s (ignored time range of T AG + [255, 525] s to avoid potential contamination from the flare) using the BPL model.The flux decay slopes and break time are consistent withLHAASO- Collaboration et al. (2023)  within 3σ uncertainty, as shown in Figure7.

Figure 6 .
Figure 6.Left: the early afterglow flux light curve (50-300 keV) jointly observed by GECAM-C and Fermi/GBM.The gray shaded area represents the fitting range.The orange lines show the posterior distribution of BPL model, and the gray vertical lines show the posterior distribution of break time.The red dashed line shows the flux decay with the pre-break slope.Right: the corner plot represents the BPL fitting result of the flux light curve in the left panel.

Figure 7 .
Figure 7. Top: light-curve decay slopes before and after the jet break in different energy bands, with error bars indicating 1σ uncertainty.Bottom: the break time of early afterglow in different energy bands, with error bars indicating 1σ uncertainty.The results in the keV energy band are from joint analysis of GECAM-C and Fermi/GBM.Our fitting results of the jet break in the TeV energy band is consistent with LHAASO team (LHAASO-Collaboration et al. 2023) within 3σ uncertainty.Notably, the break time seems to be constant from keV to TeV within the error bar.

Table 1
Fitting Time-resolved Spectra during the Time Periods of GRB 221009A Early Afterglow aThe time is relative to T AG .b ΔBIC =BIC CPL+PL −BIC PL

Table 2
Fitting Results of the Flux Light Curves during the Periods of Early Afterglow b slope (α 1 ) break time (t j ) a slope (α 2 ) χ 2 /d.o.f BIC BPL slope(α) χ 2 /d.o.f aThe break time (t j ) is relative to afterglow starting time