Evidence of the gamma-ray counterpart from nova FM Cir with Fermi-LAT

We report the analysis results of X-ray and gamma-ray data of the nova FM Cir taken by Swift and Fermi-LAT. The gamma-ray emission from FM Cir can be identified with a significance level of 3sigma within 40 days after the nova eruption (2018 January 19) while we bin the light curve per day. The significance can further exceed 4 sigma confidence level if we accumulate longer time (i.e., 20 days) to bin the light curve. The gamma-ray counterpart could be identified with a Test Statistic (TS) above 4 until 180 days after the eruption. The duration of the gamma-ray detection was longer than those reported in the previous studies of the other novae detected in the GeV range. The significant X-ray emission was observed after the gamma-ray flux level fell below the sensitivity of Fermi-LAT. The hardness ratio of the X-ray emission decreased rapidly with time, and the spectra were dominated by blackbody radiation from the hot white dwarf. Except for the longer duration of the gamma-ray emission, the multi-wavelength properties of FM Cir closely resemble those of other novae detected in the GeV range.


INTRODUCTION
Classical novae are thermonuclear eruptions that occur in binary systems, where a white dwarf accretes matter from its binary companion.The energy release resulting from the thermonuclear eruption leads to a dramatic expansion and ejection of the accretion envelope.It has been observed that the ejected matter expands into the surrounding environment at a speed of hundreds to thousands of km s −1 (Gallagher & Starrfield 1978).Multi-wavelength observations in the past decade have revealed that novae are transient sources in the broadband energy bands from radio to gamma-rays (Chomiuk et al. 2021;Acciari et al. 2022;H. E. S. S. Collaboration et al. 2022).It is argued that the shock is formed due to the collisions of the multiple ejecta (internal shock) or the interaction between the ejecta and preexisting medium surrounding the binary (Della Valle & Izzo 2020; Aydi et al. 2020b;Chomiuk et al. 2021).
Fermi Large Area Telescope (Fermi-LAT) had discovered GeV emissions from 19 novae and potential emissions from 6 sources 1 .Abdo et al. (2010) reported the first nova detected in the GeV bands, V407 Cyg, which is the binary system composed of a white dwarf and a red giant companion.It is interpreted that the shock was formed as a result of the collision between the nova ejecta and the dense wind from the red giant companion.Another nova, RS Oph, with a red giant companion, was detected in TeV energy bands, and its TeV light curve shows a temporal evolution similar to the GeV emission, indicating GeV and TeV emissions from RS Oph have a common origin each other (H.E. S. S. Collaboration et al. 2022).Most of the novae detected in the GeV range are classified as classical nova and have a main-sequence star as the companion.For such a system, it has been argued that the γ-ray emission originates from the internal shock, because the stellar wind from the main-sequence star is too weak to create a strong shock (Chomiuk et al. 2021).It has been suggested that a shock of the ejected matter by nova eruption accelerates particles to relativistic energy, leading to the production of gamma-rays through leptonic and/or hadronic processes (e.g.Vurm & Metzger 2018;Chomiuk et al. 2021).In the leptonic scenario, for example, the accelerated electrons produce the γ-rays via bremsstrahlung and/or inverse Compton scattering processes.In the hadronic scenario, the pion decay produces the majority of the gamma-rays, and the secondary electron/positron pairs also contribute to the emission through the bremsstrahlung or inverse-Compton scattering processes.
Observed X-ray emission of novae is usually characterized by thermal emission of the hot white dwarf and/or the shocked matter (Orio et al. 2001;Mukai et al. 2008;Chomiuk et al. 2014).The soft X-ray emission with an effective temperature of < 0.1 keV can reach a luminosity of L X > 10 36 erg s −1 , and it is considered that the so-called super-soft emission originates from a hot white dwarf sustained by residual nuclear burning.The X-ray emission from the majority of the novae detected in the GeV range appears after the γ-ray flux falls below the detector sensitivity of the Femi-LAT.This is interpreted that it in the earlier epoch is absorbed by the ejecta and is reprocessed into the emission in lower energy bands (Metzger et al. 2014;Li et al. 2017;Aydi et al. 2020a).As the ejected mate-rial spreads out, the environment becomes optically thin and makes the soft X-ray emission from the hot white dwarf visible (Page et al. 2020).
A naked-eye nova, FM Cir (also known as Nova Cir 2018 or PNV J13532700-6725110), was discovered by John Seach on 2018 January 19 2 , in the constellation of Circinus.FM Cir has been classified as a classical Fe II nova through optical spectroscopy (Strader et al. 2018), and the binary system has a 3.4898-day of orbital period (Schaefer 2021).FM Cir was extensively monitored optically by the AAVSO (American Association of Variable Star Observers) group, and its light curve exhibited multiple peaks and an absorption of the dust (Molaro et al. 2020).
In this paper, we report analysis results of the GeV and X-ray observations for FM Cir, and we find evidene that the duration of the γ-ray detection is longer to those in other novae detected in the GeV range.In Section 2, we describe the data analysis.The results from the γ-ray and X-ray data analysis are presented in Section 3, along with a comparison of nova FM Cir with other classical novae.We make a conclusion in Section 4.

LAT data analysis
We performed a binned analysis using the standard Fermi-LAT Sci-enceTool package, which is available from the Fermi Science Support Center 3 .The data for the fourth Fermi-LAT catalog (4FGL DR4, gll_psc_v32.fit)were taken during the period August 2008 to August 2022 covering 14 years (Ballet et al. 2023;Abdollahi et al. 2022), which contains ∼ 7200 gamma-ray sources.To avoid Earth's limb contamination, we only included the events with zenith angles below 90 degrees.We limited our analysis to the events from the point source or Galactic diffuse class (event class = 128) and used data from both the front and back sections of the tracker (evttype = 3).We constructed a background emission model that incorporates both the Galactic diffuse emission (gll_iem_v07) and the isotropic diffuse emission (iso_P8R3_SOURCE_V3_v1).A γ-ray emission model for the whole ROI was built using all sources in the fourth Fermi-LAT catalog (Abdollahi et al. 2020) located within 20 • of the nova, and the nova FM Cir is included in the model at the nova position of (R.A., decl.)=(13o 53 ′ 27.61 ′′ , −67 o 25 ′ 00.9 ′′ ).
We utilized the Test Statistic (TS) map to scrutinize the morphology of the emission originating from the FM Cir source region (Figure 1).To examine the temporal evolution of the gamma-ray emission, we selected the Fermi-LAT data (> 100 MeV) taken from MJD 58037, which is ∼100 days before the eruption (at 2018 January 19 = MJD 58137), to MJD 58430, after which no GeV detection is confirmed.We used gtlike to search for the emission with a 20day time window in a two-day time step and plot the data points with the TS value above 1 (see Figure 2).We also created the daily light curve, which allows a more precise measurement for the epoch of the gamma-ray emission as shown in Figure 3.To generate the spectrum, we performed the likelihood analysis using the data obtained from MJD 58137 to MJD 58178, around the epoch of the optical peak and of the gamma-ray detection with a TS value of 9 (i.e., > 3 σ).

Swift data analysis
Swift had continuously monitored FM Cir from the discovery of the nova eruption until to ∼MJD 58425.We created the light curve and hardness ratio of the X-ray emission using the XRT web tool4 (Evans et al. 2007(Evans et al. , 2009)).To investigate the spectral properties, we downloaded the archival data from HEASARC Browse5 and performed the analysis with the HEASOFT version 6.31.1 and the SWIFTDAS package with the updated calibration files.The clean event lists were obtained using the task xrtpipeline of the HEASOFT and extract the spectrum using Xselect.We grouped the source spectra to ensure at least 1 count per spectral bin and fit the spectra using Xspec.
We employed the blackbody radiation and/or optically thin plasma emission to fit the observed spectra.Figures 4 and A1 present the X-ray spectra taken at the different epoch.

GeV emission properties
Figure 1 illustrates the TS map of the nova FM Cir region in 0.1-300 GeV energy bands with the data taken at MJD 58137-58178.
The maximum TS value is around 22, which corresponds to a detection significance of larger than 4σ (i.e., √ TS is about detection significance in σ). Figure 2 shows the light curve created with 20day time wind.We can find in the figure that TS value increased after nova eruption, reaching TS∼ 30 at MJD 58175 and MJD 58300.In the daily light curve of Figure 3, we can also confirm the emission with T S > 9 at MJD 58175 and MJD 58300.The delay of the γ-ray peak from the epoch of the nova eruption is one of the common properties among the novae detected in GeV range, and it may be due to the timescale for particle acceleration in the ejecta or γ-ray absorption by a dense ejecta (Ackermann et al. 2014;Metzger et al. 2015).After the first peak occurred at around MJD 58175, both TS values and flux decreased with time.As shown in Figures 2 and 3, however, there is evidence of the second peak at around MJD 58300, which is about 125 days after the first peak.In the light curve of daily bins, the emission with TS >4 (σ > 2) were confined until ∼ 180 days after the nova eruption.
We fitted the spectrum obtained in MJD 58137-58178 (around the first peak) using a power-law function with an exponential cut-off of where we fixed to γ 2 = 2/3.We obtained a power-law index of γ 1 = 1.88(2) and a cut-off energy of E c = 1.00(5)GeV, which are similar to those of the novae detected in GeV bands (Franckowiak et al. 2018;Chomiuk et al. 2021).We obtained an averaged energy flux of F γ = 7.1(3) × 10 −12 erg cm −2 s −1 in 0.1-300 GeV bands.For the second peak in MJD 58280-58312, we fitted the spectrum using a pure power law function and obtained a photon index of 2.12(2).The time averaged energy flux is F γ = 2.52(1) × 10 −12 erg cm −2 s −1 in 0.1-300 GeV bands.
According to GAIA archive6 , the distance of nova FM Cir is estimated as 3.23 +0.84  −0.54 kpc or 7.48 +1.07 −3.05 kpc in geometric or photogeometric measurements (Bailer-Jones et al. 2021).To estimate the total emitted energy in the γ-ray bands, we integrated the daily flux detected with T S > 4. Figure 5 compares the total emitted energy and duration of the gamma-ray emission of the novae detected in the GeV bands.We can see in the figure that the duration of FM Cir can be the longest among those of Fermi-LAT novae, while its total emission is consistent with the others.a The unabsorbed flux is measured in 0.3-10.0keV and recorded in units of 10 −11 erg cm −2 s −1 .Li et al. (2017) reported a strong correlation between the optical and gamma-ray light curves in nova ASASSN-16ma, in which the γ-ray peak aligned with the optical peak, suggesting the optical emission of ASASSN-16ma was a reprocessing of the shock emission rather than that of the emission from the hot white dwarf.For FM Cir, the correlation analysis by utilizing the discrete correlation function indicated a lag of about 10 days between the gamma-ray peak and the optical peak.There is the second peak of the γ-ray emission at 163 day after the nova eruption.For FM Cir, hence, we could not find the correlation between the optical and gamma-ray emission with the current data.

X-ray emission properties
As Figure 3 shows, the significant detection of the X-ray emission of FM Cir started around the epoch when the detection of the γ-ray emission with T S > 4 was ended.This behavior of emergence of the X-ray emission after the γ-ray emission is similar to those of other novae detected in the GeV range (Gordon et al. 2021).We can also see that the epoch of the emergence of the X-ray emission coincided with the end of the dust absorption feature in the optical light around MJD 58300.These multi-wavelength observations suggest that absorption causes the absence of X-rays after the nova eruption until the environment becomes transparent.We fit the observed spectrum with the blackbody radiation (bbodyrad in Xspec) and/or the optically thin plasma emission (mekal) models, as shown in Table 1.The X-ray spectrum just after its emergence (top panel of Figure 4) can be fitted by a mekal model with a temperature of k B T mekal ∼ 3 keV (Table 1), where k B is the Boltzmann constant.It may be reasonable to assume that the component of the optically thin thermal emission originates from the shocked heated plasma, whose temperature may be described as (Steinberg & Metzger 2018): k B T ≈ 1.2 keV v 10 3 km s −1 2 , where v is the shock velocity.After the emergence of the X-ray emission, the hardness ratio rapidly decreased (bottom panel in Figure 6) and the second emission component appears as Figure 4 indicates.We can see in Table 1 that the fitting column density N H tends the decreasing.After the peak in the observed count rate (Figure 6), on the other hand, N H is not well constrained.We, therefore, fix the column density to N H = 2.94 × 10 21 cm −2 estimated from the sky position 7 .It is found that the temperature of the blackbody radiation of the soft component (k B T BB in Table 1) is k B T BB ∼ 20-40 eV, suggesting the emission from the hot white dwarf surface; the effective radius at the initial stage of the observed X-ray emission is of the order of 10 8 cm.
As Figure 2 indicates, the emission was not detected in the single Swift observation before MJD 58300.If the all data taken before MJD 58300 are stacked, on the other hand, the emission can be detected with a significance level of 4σ.The time averaged count rate is estimated to be (2.34 ± 0.53) × 10 −3 count/s, which is about one order of magnitude lower than (2.75 ± 0.43) × 10 −2 count/s of the first detection after MJD 58300.This may indicate that the X-ray emission is significantly absorbed before MJD 58300; we could not fit the observed spectrum because of the low count rates.
For the navae from which the X-ray emission was detected during the gamma-ray detection, it has been observed that the ratio of the GeV luminosity over the X-ray luminosity is L γ /L X > 10 − 10 2 , and such a high ratio could be explained by several emission senarios, for example, multiple shocks, very high efficiency of the acceleration process, etc., (Gordon et al. 2021).As we described in the previous paragraph, the time averaged count rate before MJD 58300 is about one order of magnitude smaller than that of the first significant detection with a single Swift observation.If we just scale the unabsorbed flux using the count rate, we obtain the ratio as L γ /L X ∼ 10.We may, however, expect that the low count rate is due to a stronger absorption, and the time averaged unabsorbed flux is of the order of that of the first detection after MJD 58300.In such a case, we estimate as L γ /L X ∼ 1.The current data, therefore, could not provide a tight constraint on the L γ /L X .

CONCLUSION
We presented an analysis of Fermi-LAT observations of nova FM Cir and identified evidence of γ-ray emission.The detection significance could exceed > 4σ for 20-day bins and reached ∼ 3σ for daily bins in the data taken at ∼ 40 days after the eruption.We found that the γ-ray emission with a TS value larger than 4 lasted about 180 days, which is the longest among the known novae detected in the GeV range.On the other hand, the integrated energy flux ∼ 10 41−42 erg was consistent with those of the novae detected in the GeV range.The X-ray emission emerged at MJD 58318 after the GeV emission was ended.The X-ray spectrum just after the emergence was fitted by the emission from the optically thin thermal plasma, which likely originates from the shock.It was observed that the hardness ratio rapidly decreased with time, and the spectra taken later time were dominated by the blackbody radiation from the hot white dwarf.The multi-wavelength properties in the optical to GeV bands are similar to those of other novae detected in the GeV range, suggesting that nova FM Cir is likely a nova in the GeV range.

Figure 3 .
Figure3.The photon flux with 1-day bin using Fermi-LAT data.The red triangles and the light blue dots in upper panel are the flux with the TS value large than 4 (corresponding to a detection significance of ∼ 2 σ) and 1 (∼1 σ), respectively.Second panel: the evolution of TS value.Third panel: the light curve from optical data in site of http://aavso.org/lcg.Fourth panel: the count rate of X-ray data which come from the S wi f t telescope in 0.3-10.0keV.The vertical dash line shows the epoch of nova eruption in optical.

Figure 4 .
Figure 4. X-ray spectrum of nova FM Cir taken at MJD 58318 (the first significant detection by the Swift observation) and MJD 58340.All spectra taken at different epoch are presented in Figure A1.

Figure 5 .
Figure 5.Total emitted GeV energy vs. duration of GeV emission of the novae detected in the GeV range.The duration is defined by the epoch during which the emission with TS > 4 lasted.The data illustrated with triangles are taken from Albert et al. (2022).The symbols with the square and filled circles correspond to the emission energy of nova FM Cir estimated with the geometric distance and photogeometric distance, respectively.

Table 1 .
The fitting parameters of X-ray spectra using the model of blackbody radiation and/or optically thin thermal plasma emission (mekal).