AstroSat view of MAXI J1535-571: broadband spectro-temporal features

TL;DR

This paper reports AstroSat observations of the black hole binary MAXI J1535-571, analyzing its broadband spectral and timing features, including QPOs and accretion rates, revealing high flux and near Eddington accretion.

Contribution

First broadband spectral and timing analysis of MAXI J1535-571 with AstroSat, including QPO characterization and physical modeling of accretion processes.

Findings

Detected QPOs with frequencies 1.85-2.88 Hz and harmonics.

Estimated black hole mass between 5.14 and 7.83 solar masses.

Observed high flux corresponding to near Eddington accretion rates.

Abstract

We present the results of Target of Opportunity (ToO) observations made with AstroSat of the newly discovered black hole binary MAXI J1535-571. We detect prominent C-type Quasi-periodic Oscillations (QPOs) of frequencies varying from 1.85 Hz to 2.88 Hz, along with distinct harmonics in all the AstroSat observations. We note that while the fundamental QPO is seen in the 3 - 50 keV energy band, the harmonic is not significant above ~ 35 keV. The AstroSat observations were made in the hard intermediate state, as seen from state transitions observed by MAXI and Swift. We attempt spectral modelling of the broadband data (0.7-80 keV) provided by AstroSat using phenomenological and physical models. The spectral modelling using nthComp gives a photon index in the range between 2.18-2.37 and electron temperature ranging from 21 to 63 keV. The seed photon temperature is within 0.19 to 0.29 keV. The high flux in 0.3 - 80 keV band corresponds to a luminosity varying from 0.7 to 1.07 L_Edd assuming the source to be at a distance of 8 kpc and hosting a black hole with a mass of 6 M$_{\odot}$. The physical model based on the two-component accretion flow gives disc accretion rates as high as ~ 1 $\dot{m}_{Edd}$ and halo rate ~ 0.2 $\dot{m}_{Edd}$ respectively. The near Eddington accretion rate seems to be the main reason for the unprecedented high flux observed from this source. The two-component spectral fitting of AstroSat data also provides an estimate of a black hole mass between 5.14 to 7.83 M$_{\odot}$.

Tables

Table 1. Table 1: QPO parameters for the 15 15 15 orbits data of AstroSat/LAXPC in the full energy band of 3 − 80 3 80 3-80  keV. These results are from modelling the power spectrum after dead time correction and Poisson noise subtraction.

MJD	Orbit	Mean rate	QPO	Q-factor	Sigb	amplitude	Harmonic	Q-factor	Sigb	amplitude	BLN	LFN	red. ${\chi }^2$
		(cts/s)	(Hz)		($\sigma$)	(rms %)	(Hz)		($\sigma$)	(rms %)	(rms %)	(rms %)	(${\chi }^2$/dof)
58008.24	10585	4479	2.21 $\pm$ 0.02	9.20	11.78	8.23 $\pm$ 0.49	4.37 $\pm$ 0.03	5.75	10.30	4.10 $\pm$ 0.37	11.3 $\pm$ 1.44	6.56 $\pm$ 1.02	0.98 (115.7/118)
58008.53	10589	5337	2.65 $\pm$ 0.01	8.54	13.76	8.51 $\pm$ 0.41	5.24 $\pm$ 0.04	5.88	10.63	3.68 $\pm$ 0.29	10.7 $\pm$ 1.25	7.03 $\pm$ 0.69	0.99 (116.7/118)
58008.82	10594	5398	2.54 $\pm$ 0.01	8.75	09.22	8.42 $\pm$ 0.63	4.95 $\pm$ 0.05	4.54	07.75	4.32 $\pm$ 0.63	9.83 $\pm$ 2.03	7.37 $\pm$ 0.96	0.89 (105.1/118)
58009.02	10597	5618	2.88 $\pm$ 0.01	10.28	14.10	8.56 $\pm$ 0.42	5.68 $\pm$ 0.04	4.89	10.69	4.00 $\pm$ 0.39	9.91 $\pm$ 1.22	7.29 $\pm$ 0.64	1.39 (164.0/118)
58009.55	10604	5593	2.56 $\pm$ 0.01	9.48	09.89	8.46 $\pm$ 0.66	5.02 $\pm$ 0.04	4.29	08.38	4.28 $\pm$ 0.53	10.0 $\pm$ 1.63	6.86 $\pm$ 1.03	1.70 (200.3/118)
58009.77	10608	5590	2.02 $\pm$ 0.02	8.41	07.58	8.44 $\pm$ 0.80	3.99 $\pm$ 0.06	3.21	07.30	4.85 $\pm$ 0.90	9.99 $\pm$ 2.24	6.50 $\pm$ 1.33	0.93 (110.0/118)
58010.03	10612	5642	1.85 $\pm$ 0.01	9.73	12.53	8.22 $\pm$ 0.46	3.72 $\pm$ 0.03	4.76	10.92	4.53 $\pm$ 0.39	11.2 $\pm$ 1.76	6.14 $\pm$ 1.48	0.83 ( 98.0/118)
58010.49	10618	5708	2.08 $\pm$ 0.01	11.55	09.87	8.42 $\pm$ 0.64	4.15 $\pm$ 0.03	5.60	10.09	4.39 $\pm$ 0.40	11.2 $\pm$ 1.35	6.23 $\pm$ 1.08	1.09 (128.6/118)
58010.92	10625	5910	2.20 $\pm$ 0.01	9.56	11.38	8.35 $\pm$ 0.42	4.33 $\pm$ 0.03	5.85	09.08	4.17 $\pm$ 0.46	10.8 $\pm$ 1.61	6.61 $\pm$ 1.17	0.91 (107.3/118)
58011.13	10628	5923	2.17 $\pm$ 0.01	9.43	13.71	8.44 $\pm$ 0.48	4.30 $\pm$ 0.02	7.16	11.80	3.87 $\pm$ 0.28	11.1 $\pm$ 1.08	6.61 $\pm$ 0.72	1.15 (135.2/118)
58011.66	10636	6011	2.31 $\pm$ 0.01	9.62	08.61	8.73 $\pm$ 0.77	4.53 $\pm$ 0.03	7.07	07.27	3.85 $\pm$ 0.48	10.4 $\pm$ 2.12	6.80 $\pm$ 1.31	0.68 ( 80.8/118)
58011.87	10639	6035	2.21 $\pm$ 0.01	9.60	09.85	8.62 $\pm$ 0.56	4.38 $\pm$ 0.03	6.53	08.39	3.95 $\pm$ 0.45	10.5 $\pm$ 1.76	6.87 $\pm$ 1.12	0.78 ( 92.3/118)
58012.24	10644	6189	2.59 $\pm$ 0.01	8.63	15.52	8.86 $\pm$ 0.42	5.09 $\pm$ 0.03	5.19	13.88	3.94 $\pm$ 0.31	9.93 $\pm$ 1.53	8.03 $\pm$ 0.73	1.66 (197.6/119)
58012.67	10651	6282	2.46 $\pm$ 0.02	9.11	08.88	8.82 $\pm$ 0.69	4.83 $\pm$ 0.04	5.81	07.51	3.98 $\pm$ 0.49	9.56 $\pm$ 2.25	7.21 $\pm$ 1.26	0.69 ( 81.0/118)
58013.02	10656	6338	2.11 $\pm$ 0.01	9.59	10.04	8.89 $\pm$ 0.58	4.20 $\pm$ 0.03	4.82	11.83	4.46 $\pm$ 0.49	10.1 $\pm$ 2.34	7.45 $\pm$ 1.17	1.35 (159.4/118)

Table 2. Table 2: Energy dependent QPO parameters of MAXI J1535 − - 571 from AstroSat/LAXPC . Power spectra have been modelled in frequency-rms space. The parameters corresponding to the fundamental component are indicated with (f) and those corresponding to the harmonic component with (h).

MJD	E-band	QPO(f)	FWHM(f)	Norm(f)	Q(f)	sig(f)a	rms(f)	QPO(h)	FWHM(h)	Norm(h)	Q(h)	sig(h)a	rms(h)
	keV	Hz	Hz			($\sigma$)	(%)	Hz	Hz			($\sigma$)	(%)
	3 – 20	2.22 $\pm$ 0.01	0.23 $\pm$ 0.02	0.018	9.64	11.05	8.09 $\pm$ 0.46	4.43 $\pm$ 0.03	0.60 $\pm$ 0.08	0.0018	7.38	9.84	5.09 $\pm$ 0.42
58008.24	20 – 35	2.19 $\pm$ 0.01	0.21 $\pm$ 0.03	0.021	10.42	7.95	8.44 $\pm$ 0.82	${4.57}_{-0.10}^{+0.14}$	${0.78}_{-0.34}^{+0.62}$	0.0018	5.85	3.53	4.71 $\pm$ 1.62
	35 – 50	${2.22}_{-0.05}^{+0.07}$	${0.31}_{-0.16}^{+0.12}$	0.010	7.16	4.47	7.12 $\pm$ 2.53	-	-	-	-	-
	50 – 80	-	-	-	-	-	-	-	-	-	-	-
	3 – 20	2.89 $\pm$ 0.01	0.29 $\pm$ 0.02	0.016	9.96	11.78	8.61 $\pm$ 0.45	5.72 $\pm$ 0.03	${0.87}_{-0.09}^{+0.11}$	0.0013	6.57	12.76	4.17 $\pm$ 0.36
58009.02	20 – 35	2.88 $\pm$ 0.01	0.22 $\pm$ 0.02	0.023	13.09	9.91	9.09 $\pm$ 0.69	${5.86}_{-0.09}^{+0.10}$	${0.82}_{-0.32}^{+0.50}$	0.0019	7.14	4.08	5.02 $\pm$ 1.43
	35 – 50	2.87 $\pm$ 0.03	${0.19}_{-0.09}^{+0.06}$	0.014	15.10	4.28	6.61 $\pm$ 2.20	-	-	-	-	-
	50 – 80	-	-	-	-	-	-	-	-	-	-	-
	3 – 20	2.61 $\pm$ 0.01	0.28 $\pm$ 0.01	0.016	9.32	15.18	8.58 $\pm$ 0.35	5.14 $\pm$ 0.02	${0.84}_{-0.09}^{+0.11}$	0.0013	6.11	13.82	4.19 $\pm$ 0.29
58012.24	20 – 35	2.59 $\pm$ 0.01	0.26 $\pm$ 0.02	0.020	9.96	11.30	9.27 $\pm$ 0.57	5.35 $\pm$ 0.10	${1.09}_{-0.30}^{+0.45}$	0.0016	4.90	5.16	5.24 $\pm$ 1.05
	35 – 50	${2.58}_{-0.06}^{+0.04}$	${0.24}_{-0.21}^{+0.07}$	0.0099	10.75	4.75	6.14 $\pm$ 1.93	-	-	-	-	-
	50 – 80	-	-	-	-	-	-	-	-	-	-	-

Table 3. Table 3: Spectral parameters resulting from a fit with the model Tbabs × \times constant × \times (gaussian + + edge × \times nthComp) . † † \dagger denotes broadband modelling with SXT and LAXPC , whereas other observations are modelled using data from LAXPC instrument alone. Unabsorbed flux values in 0.3 − 80 0.3 80 0.3-80  keV are also included in the Table.

MJD	Orbit	$\Gamma$	kTe (keV)	kTbb (keV)	norm	flux ($0.3-80$ keV)	$\tau$	${\chi }_{red}^2$
						(erg cm-2 s-1)		(${\chi }^2$/dof)
58008.24 $†$	10585	${2.254}_{-0.005}^{+0.004}$	${37.8}_{-2.3}^{+2.7}$	${0.21}_{-0.04}^{+0.02}$	${34.3}_{-1.5}^{+2.6}$	$7.14\times {10}^{-8}$	1.640 $\pm$ 0.011	1.31 (1181/895)
58008.53 $†$	10589	${2.346}_{-0.005}^{+0.004}$	100.0 (fixed)	${0.19}_{-0.04}^{+0.03}$	${43.5}_{-0.3}^{+0.3}$	$8.68\times {10}^{-8}$	0.720 $\pm$ 0.004	1.25 (1127/896)
58008.82 $†$	10594	${2.316}_{-0.003}^{+0.003}$	100.0 (fixed)	${0.19}_{-0.05}^{+0.05}$	${35.3}_{-0.4}^{+0.3}$	$8.63\times {10}^{-8}$	0.739 $\pm$ 0.004	1.27 (1140/896)
58009.02	10597	${2.374}_{-0.005}^{+0.006}$	${63.0}_{-6.9}^{+10.2}$	${0.20}_{-0.009}^{+0.008}$	${34.7}_{-8.3}^{+2.2}$	$7.17\times {10}^{-8}$	1.010 $\pm$ 0.008	1.14 (311/271)
58009.55 $†$	10604	${2.327}_{-0.003}^{+0.003}$	100.0 (fixed)	${0.22}_{-0.03}^{+0.01}$	${43.4}_{-5.5}^{+5.2}$	$9.21\times {10}^{-8}$	0.732 $\pm$ 0.004	1.41 (1268/896)
58009.77	10608	${2.243}_{-0.004}^{+0.005}$	${39.1}_{-2.3}^{+2.9}$	${0.24}_{-0.008}^{+0.006}$	${25.2}_{-8.3}^{+4.1}$	$7.29\times {10}^{-8}$	1.612 $\pm$ 0.009	1.15 (312/271)
58010.03	10612	${2.189}_{-0.004}^{+0.005}$	${24.2}_{-0.6}^{+0.8}$	${0.29}_{-0.12}^{+0.03}$	${21.6}_{-7.1}^{+0.5}$	$7.41\times {10}^{-8}$	2.380 $\pm$ 0.010	1.09 (296/271)
58010.49 $†$	10618	${2.233}_{-0.002}^{+0.005}$	${32.6}_{-1.2}^{+1.3}$	${0.21}_{-0.07}^{+0.01}$	${38.2}_{-0.3}^{+0.3}$	$9.39\times {10}^{-8}$	1.856 $\pm$ 0.012	1.26 (1131/895)
58010.92 $†$	10625	${2.261}_{-0.005}^{+0.006}$	${33.0}_{-1.5}^{+1.9}$	${0.24}_{-0.04}^{+0.04}$	${38.1}_{-0.5}^{+0.5}$	$9.32\times {10}^{-8}$	1.798 $\pm$ 0.014	1.09 (945/866)
58011.13 $†$	10628	${2.238}_{-0.005}^{+0.005}$	${21.2}_{-0.5}^{+0.6}$	${0.19}_{-0.04}^{+0.04}$	${31.6}_{-1.8}^{+2.2}$	$9.43\times {10}^{-8}$	2.514 $\pm$ 0.012	1.19 (1071/895)
58011.66 $†$	10636	${2.287}_{-0.004}^{+0.009}$	${41.2}_{-2.4}^{+2.9}$	${0.20}_{-0.05}^{+0.02}$	${46.1}_{-0.3}^{+0.2}$	$1.03\times {10}^{-7}$	1.504 $\pm$ 0.010	1.34 (1202/895)
58011.87	10639	${2.281}_{-0.004}^{+0.005}$	${36.5}_{-2.1}^{+2.5}$	${0.22}_{-0.008}^{+0.006}$	${31.6}_{-6.5}^{+2.4}$	$7.96\times {10}^{-8}$	1.645 $\pm$ 0.009	1.29 (351/271)
58012.24 $†$	10644	${2.327}_{-0.005}^{+0.005}$	${26.1}_{-0.8}^{+1.0}$	${0.22}_{-0.003}^{+0.004}$	${52.0}_{-0.4}^{+0.4}$	$1.03\times {10}^{-7}$	2.038 $\pm$ 0.010	1.21 (1091/895)
58012.67 $†$	10651	${2.316}_{-0.005}^{+0.003}$	${54.4}_{-4.4}^{+4.8}$	${0.19}_{-0.01}^{+0.03}$	${53.9}_{-2.5}^{+0.4}$	$1.10\times {10}^{-7}$	1.188 $\pm$ 0.009	1.51 (1353/895)
58013.02	10656	${2.249}_{-0.004}^{+0.005}$	${23.6}_{-0.58}^{+0.81}$	${0.29}_{-0.005}^{+0.005}$	${28.4}_{-8.6}^{+0.71}$	$8.48\times {10}^{-8}$	2.317 $\pm$ 0.012	1.05 (285/271)

Table 4. Table 4: Fit parameters from the model TBabs × \times constant(gaussian + + edge × \times highecut × \times two-component-flow) applied on AstroSat data.

MJD	58011.13	58012.24
Orbit	10628	10644
Mass ($M_{\odot }$)	${6.31}_{-0.86}^{+0.68}$	${6.47}_{-1.33}^{+1.36}$
Shock location ($r_g$)	${16.3}_{-0.99}^{+0.80}$	${15.9}_{-0.8}^{+2.1}$
Halo rate (${\dot{m}}_{Edd}$)	${0.203}_{-0.022}^{+0.019}$	${0.15}_{-0.02}^{+0.02}$
Disc rate (${\dot{m}}_{Edd}$)	${0.977}_{-0.157}^{+0.109}$	${0.78}_{-0.20}^{+0.15}$
Cutoff Energy (keV)	${20.0}_{-1.4}^{+1.2}$	${18.9}_{-0.89}^{+1.1}$
Fold Energy (keV)	${78.6}_{-7.8}^{+8.0}$	${85.2}_{-4.7}^{+5.8}$
${\chi }^2/dof$	1.15 (1031.32/891)	1.19 (1063.99/891)

Table 5. Table 5: Observations with AstroSat : ObsID T01_191T01_9000001536 between 12 th superscript 12 th 12^{\mathrm{th}}  Sep till 17 th superscript 17 th 17^{\mathrm{th}}  Sep, 2017. Fifteen orbits chosen are bold-faced.

Orb-#	Exp Time (s)	Start Time (UTC)	End Time (UTC)
10584	496.953770123	2017-09-12 05:32:33	2017-09-12 06:04:44
10585	4941.77365309	2017-09-12 05:55:48	2017-09-12 07:50:36
10586	5357.744978	2017-09-12 07:35:51	2017-09-12 09:36:54
10588	7382.02297796	2017-09-12 09:25:32	2017-09-12 12:50:29
10589	4583.00317103	2017-09-12 12:45:50	2017-09-12 14:49:28
10590	4031.87576006	2017-09-12 14:36:24	2017-09-12 16:33:09
10592	3851.26237635	2017-09-12 16:21:05	2017-09-12 18:16:50
10593	3531.06603971	2017-09-12 18:08:39	2017-09-12 20:00:38
10594	3292.33470635	2017-09-12 19:50:12	2017-09-12 21:43:51
10595	2970.22884139	2017-09-12 21:33:21	2017-09-12 23:28:27
10596	3249.72885065	2017-09-12 22:59:54	2017-09-13 01:12:44
10597	3264.45310393	2017-09-13 00:39:54	2017-09-13 03:02:45
10598	4230.74553183	2017-09-13 02:39:27	2017-09-13 04:45:33
10599	4552.78409085	2017-09-13 04:28:08	2017-09-13 06:28:26
10600	5069.68189324	2017-09-13 06:13:33	2017-09-13 08:13:57
10603	12317.7855291	2017-09-13 08:06:35	2017-09-13 13:29:21
10604	3694.92741311	2017-09-13 13:21:28	2017-09-13 15:11:56
10606	3628.15777545	2017-09-13 15:04:25	2017-09-13 16:55:33
10607	3874.79451605	2017-09-13 16:42:45	2017-09-13 18:39:45
10608	3442.76126461	2017-09-13 18:31:27	2017-09-13 20:23:45
10609	3070.35393815	2017-09-13 20:14:13	2017-09-13 22:05:06
10610	3307.57441911	2017-09-13 21:48:02	2017-09-13 23:50:57
10611	3153.40418664	2017-09-13 23:31:51	2017-09-14 01:35:17
10612	4134.9651357	2017-09-14 00:46:03	2017-09-14 03:25:09
10613	4389.06476232	2017-09-14 03:03:45	2017-09-14 05:08:50
10614	4761.01065524	2017-09-14 04:47:38	2017-09-14 06:52:10
10615	5219.04197027	2017-09-14 06:39:20	2017-09-14 08:38:08
10617	8634.5027135	2017-09-14 08:25:20	2017-09-14 12:07:52
10618	4120.10051683	2017-09-14 11:56:09	2017-09-14 13:51:28
10619	3991.58977488	2017-09-14 13:39:03	2017-09-14 15:34:31
10621	3821.43898252	2017-09-14 15:23:18	2017-09-14 17:18:31
10622	3667.28994137	2017-09-14 17:09:19	2017-09-14 19:02:57
10623	3609.30743047	2017-09-14 18:49:47	2017-09-14 20:47:03
10624	3379.89339639	2017-09-14 20:31:08	2017-09-14 22:30:13
10625	3406.84135797	2017-09-14 22:09:55	2017-09-15 00:14:39
10626	3357.2935165	2017-09-14 23:43:14	2017-09-15 01:59:46
10627	4507.45603695	2017-09-15 01:13:12	2017-09-15 03:48:27
10628	4542.56313907	2017-09-15 03:10:26	2017-09-15 05:31:16
10629	4904.77152288	2017-09-15 05:22:25	2017-09-15 07:15:57
10632	12804.8367192	2017-09-15 07:02:51	2017-09-15 12:31:09
10633	4078.40076538	2017-09-15 12:18:56	2017-09-15 14:14:48
10635	3657.04443797	2017-09-15 14:06:59	2017-09-15 15:57:55
10636	3488.04376683	2017-09-15 15:50:31	2017-09-15 17:41:02
10637	3765.87370964	2017-09-15 17:28:37	2017-09-15 19:25:07
10638	3336.03761904	2017-09-15 19:15:16	2017-09-15 21:08:22
10639	3148.43251339	2017-09-15 20:56:25	2017-09-15 22:53:34
10640	3081.41305334	2017-09-15 22:37:55	2017-09-16 00:38:52
10641	3561.31836302	2017-09-16 00:03:14	2017-09-16 02:22:36
10642	4247.31112233	2017-09-16 01:51:34	2017-09-16 04:11:31
10643	4544.92289703	2017-09-16 03:54:46	2017-09-16 05:54:00
10644	4903.1540284	2017-09-16 05:46:24	2017-09-16 07:40:10
10646	8637.82817281	2017-09-16 07:33:24	2017-09-16 11:09:57
10647	3889.49490765	2017-09-16 11:03:05	2017-09-16 12:54:18
10648	3992.12343246	2017-09-16 12:41:34	2017-09-16 14:37:30
10650	3518.57945908	2017-09-16 14:29:53	2017-09-16 16:20:18
10651	3424.59817114	2017-09-16 16:13:03	2017-09-16 18:04:24
10652	3822.22094573	2017-09-16 17:50:32	2017-09-16 19:48:56
10653	3282.38882317	2017-09-16 19:36:53	2017-09-16 21:30:31
10654	3198.43616311	2017-09-16 21:16:46	2017-09-16 23:15:04
10655	3446.71049657	2017-09-16 22:41:34	2017-09-17 01:00:46
10656	3721.80164763	2017-09-17 00:29:54	2017-09-17 02:46:05
10657	2967.15269032	2017-09-17 02:21:23	2017-09-17 03:42:35

Table 6. Table 6: Swift Public archive. Selected datasets used in our analysis have bold-faced ObsIds.

ObsId	Start Time (UTC)	XRT Exp (s)
00770431000	2017-09-02 19:44:52	1615.85
00770431001	2017-09-02 22:11:56	3387.90
00770502000	2017-09-03 08:44:02	75.03
00770656000	2017-09-04 13:29:33	15.59
00770656001	2017-09-04 17:10:06	988.79
00770656002	2017-09-05 05:17:20	4504.50
00010264002	2017-09-06 00:38:57	449.22
00010264003	2017-09-08 06:38:57	4948.52
00771371000	2017-09-08 07:58:07	69.12
00010264004	2017-09-11 06:32:57	1089.60
00010264005	2017-09-12 06:16:57	899.60
00010264007	2017-09-13 00:08:57	889.45
00010264006	2017-09-14 22:17:57	1074.60
00010264008	2017-09-15 09:15:57	984.63
00010264009	2017-09-16 02:44:57	1059.78
00088245001	2017-09-17 04:14:57	2223.08
00010264010	2017-09-18 04:10:56	1104.60
00010264011	2017-09-19 08:52:57	989.38
00088245002	2017-09-20 23:10:57	954.61
00088245003	2017-09-21 00:46:57	1014.18
00010264012	2017-09-22 15:15:57	1189.20
00010264013	2017-09-23 02:20:56	1009.60
00010264014	2017-09-24 02:25:57	599.20
00773893000	2017-09-24 09:54:14	620.93
00010264015	2017-09-25 21:28:56	899.48
00010264016	2017-09-26 18:12:56	909.21
00010264017	2017-09-27 16:37:57	859.22
00010264018	2017-09-28 18:01:57	1044.60
00010264019	2017-09-29 17:56:57	964.65
00010264020	2017-09-30 17:56:57	889.61
00010264021	2017-10-01 17:46:56	874.61
00010264022	2017-10-02 17:44:56	1109.60
00010264023	2017-10-03 17:44:22	1004.60
00010264024	2017-10-04 17:40:57	964.60
00010264025	2017-10-05 15:53:27	604.61
00010264026	2017-10-06 11:06:57	1154.60
00010264027	2017-10-07 04:35:57	1024.60
00010264028	2017-10-08 05:53:56	974.60
00010264029	2017-10-09 10:51:57	1059.61
00010264030	2017-10-10 12:03:57	994.06
00010264031	2017-10-11 07:32:57	854.62
00088245004	2017-10-22 23:40:57	2088.03
00088246001	2017-10-24 23:28:57	1528.19
00010490007	2017-12-31 06:36:57	109.75
00010491001	2017-12-31 06:39:31	869.95
00010492001	2017-12-31 06:41:06	864.96
00010493001	2017-12-31 06:42:39	915.76
00010494001	2017-12-31 06:44:12	877.77
00010495001	2017-12-31 06:45:41	885.02
00010496001	2017-12-31 06:47:05	724.70
00010497001	2017-12-31 06:48:28	847.41
00010491002	2018-01-19 08:05:57	1119.61
00010264032	2018-03-17 19:13:28	279.61
00010264033	2018-03-18 10:47:57	1370.34
00010264034	2018-03-24 21:16:57	1859.22
00010264035	2018-03-26 13:28:47	1409.22
00010264036	2018-03-28 13:23:57	2015.28
00010264037	2018-03-30 02:03:57	1910.27

Full text

††pubyear: 2019††pagerange: AstroSat view of MAXI J1535$-$571: broadband spectro-temporal features–B

1 Introduction

X-ray binaries (XRBs) are astrophysical systems that are luminous in X-rays. An XRB hosts two stars of which one is either an accreting neutron star or black hole and the other is the mass donor which is referred to as the secondary star. XRBs are classified into low mass X-ray binaries (LMXBs) or high mass X-ray binaries (HMXBs) based on the mass of the secondary star (Seward & Charles, 2010, and references therein). HMXBs tend to be persistently bright, whereas most of the LMXBs are transient in nature. LMXBs can remain in quiescence (below detection limit of the monitoring instruments) for a very long time and suddenly enter a transient phase exhibiting an outburst in the flux emitted. This increase in flux extends over a period of weeks to months, followed by a slower decay. Several hitherto unknown sources have been detected during such outbursts. Catalogs of black hole binaries discovered till 2016 can be found in Corral-Santana et al. (2016); Tetarenko et al. (2016). Prompt follow up observations, in different energy bands are required to understand the nature of the source and deduce the binary properties.

Emission from black hole binaries can be of thermal and non-thermal origins. The Keplerian accretion disc (Shakura & Sunyaev, 1973) around a black hole is considered to be composed of concentric regions with different temperature, each radiating thermally and its total spectrum is approximated by a multi-color disk blackbody model (Mitsuda et al., 1984). Inverse-Comptonisation (Sunyaev & Titarchuk, 1980; Titarchuk, 1994; Tanaka & Lewin, 1995) of the soft photons from the disc by a hot corona contributes to the higher energy thermal or non-thermal emission from the X-ray binary. The study of black hole binaries helps us to understand the nature of accretion processes in astrophysics. We can also infer the rate of mass accretion from the disc and corona in the source observed and understand their relative contributions to the flux. Spectral and temporal analysis gives astronomers an idea about the state evolution of the source during an outburst. Besides this, multi-wavelength observations are useful in studying the mechanism of outflows and jet ejection from these sources.

A plot of the hardness ratio versus the total flux is called the hardness intensity diagram (HID) (Homan et al., 2001) where hardness is the ratio of flux in hard band to that in soft band. It is generally observed that the HID of a complete outburst follows a ‘q’ profile (Homan & Belloni, 2005; Nandi et al., 2012; Motta et al., 2012; Nandi et al., 2018; Radhika et al., 2018). Using HID and spectro-temporal parameters, one can classify the outburst periods of the black hole binaries into low hard state (LHS), hard intermediate state (HIMS), soft intermediate state (SIMS) and soft state (Homan et al., 2001; Homan & Belloni, 2005; Remillard & McClintock, 2006; Nandi et al., 2012).

Apart from the HID, the rms intensity diagram (RID) and the hardness rms diagram (HRD) (Motta et al., 2012; Belloni & Motta, 2016) can also be used for state classification. Besides this, the presence of low frequency QPOs (LFQPOs) can indicate the state of a black hole binary system. C-type QPOs (Casella et al., 2005) are usually observed in LHS and HIMS whereas the presence of A and/or B-type QPOs correspond to SIMS (Radhika et al., 2016; Belloni & Motta, 2016, and references therein).

The X-ray transient source MAXI J1535$-$571 was detected independently by MAXI (Monitor of All-sky X-ray Image) and Swift/BAT on Sep $2$, $2017$ (Negoro et al. (2017a) and Kennea et al. (2017)) at (RA, DEC) = ($15~{}35~{}19.73,-57~{}13~{}48.1$). Negoro et al. (2017b) suggested that it could be a black hole LMXB, based on the observed rapid random X-ray variability and column density. The unprecedented large flux levels prompted a flurry of observations made by different observatories as detailed below. A strong radio counterpart was detected using ATCA data with spectral powerlaw index of $0.09$ $\pm$ $0.03$. This suggested emission from a compact radio jet (Russell et al., 2017a). Swift/XRT data in the rising phase of the outburst revealed C-type QPOs at $4.00\pm 0.05$ Hz, $3.67\pm 0.05$ Hz and $2.50\pm 0.02$ Hz (Russell et al., 2017b), and evolution of QPOs from $0.21\pm 0.03$ Hz to $3.22\pm 0.04$ Hz (Mereminskiy & Grebenev, 2017; Mereminskiy et al., 2018) using Swift and INTEGRAL data. Shang et al. (2018) also observed QPOs in the range between $0.44$ to $6.48$ Hz using Swift/XRT data in the period MJD 58004 to 58050. MAXI and Swift monitoring of the outburst as it reached its peak in about $20$ days since detection showed state changes (Nakahira et al., 2017; Kennea, 2017; Shidatsu et al., 2017). Negoro et al. (2018a) reported that the source flux dropped sharply by MJD 58200 (by Mar $23,~{}2018$) below detection limit in MAXI, while the source was in the soft state. X-ray and Radio detections with state changes were nevertheless reported beyond this too (Russell et al., 2018a, b; Parikh et al., 2018; Negoro et al., 2018b).

Nakahira et al. (2018) provided the details of monitoring with MAXI $2-20$ keV energy band, and the state changes till fading of the source. Xu et al. (2018) described the NuSTAR spectra considering the disc reflection features while modelling the broad Fe-K line and Compton hump in $20-50$ keV. They deduced a black hole spin $a$ > 0.84, inner disc radius $R_{in}<2.01~{}r_{ISCO}$, and a lamp-post height $h=7.2^{+0.8}_{-2.0}~{}r_{\rm g}$ for a compact coronal illuminating source. Tao et al. (2018) made use of the XRT data from the entire outburst phase to examine the hardness based state changes, and evolution of spectral properties. Baglio et al. (2018) discussed about the possibility of using rapid variability in mid-IR in order to study the disc-jet connection. Recently, Stiele & Kong (2018) have done detailed timing and spectral analysis of this source using data from XRT, XMM-Newton and NICER observatories. Also, Stevens et al. (2018) detected LFQPO of $5.72^{+0.04}_{-0.06}$ Hz in the average power spectrum corresponding to the NICER observations from MJD 58016 to 58026, when the source was in SIMS. Huang et al. (2018) reported on timing and spectral analysis using data from HXMT and suggested that MAXI J1535$-$571 could be a high inclination system. Analysis of NICER spectrum in the energy range $2.3-10$ keV by Miller et al. (2018) also indicated a high inclination of $67.4\pm 0.8^{\circ}$. Besides this, Miller et al. (2018) determined that the source spins at near maximal rate of $0.994\pm 0.002$. Though there are estimates on the inclination and spin of the system, the distance and mass of the black hole are yet to be estimated. In this work, we have used our physical model to estimate the mass of MAXI J1535$-$571.

AstroSat ToO observations were triggered on MAXI J1535$-$571, in the second week after its detection. AstroSat is India’s first satellite, dedicated to astronomy. The broadband coverage and high sensitivity of LAXPC and SXT onboard AstroSat, enabled simultaneous spectral and temporal studies of the source. We studied the nature and energy dependence of LFQPOs and the broadband spectrum using these observations. We modelled the broadband spectrum in the energy range $0.7-80$ keV phenomenologically as well as in the framework of the two-component accretion flow (Chakrabarti & Titarchuk, 1995; Chakrabarti & Mandal, 2006; Iyer et al., 2015). This allowed us to compute the accretion rates for the source and also to estimate the mass of the black hole. In order to augment the 6 days of AstroSat observations, we also used data from observations with Swift/XRT and MAXI.

In §2, we have given details on the observation and steps for data reduction. We have discussed the outburst profile and HID of this outburst in §3. Detailed temporal analysis was done with AstroSat/LAXPC and the results are presented in §4. We undertook a broadband spectral analysis using both SXT and LAXPC data and the results from phenomenological as well as physical modelling are presented in §5. Finally, we have discussed the results and have concluded in §6.

2 Observation and Data Reduction

AstroSat provides a unique platform for simultaneous observations, over a wide X-ray band from $0.3$ keV to $100$ keV, via its suite of co-aligned X-ray instruments – Soft X-ray Telescope (SXT) (Singh et al., 2016, 2017), Large Area X-ray Proportional Counter (LAXPC) (Yadav et al., 2016b; Antia et al., 2017) and Cadmium Zinc Telluride Imager (CZTI) (Vadawale et al., 2016). We exploit the detection capabilities of SXT in the soft X-rays to obtain spectra in the lower energies and LAXPC enables us to study spectra upto 80 keV. We also make use of the temporal resolution ($10~{}\mu$s) of LAXPC to search for QPOs across a wide range of frequencies.

The ToO observations of AstroSat were scheduled over $62$ slots, each of about a few ks exposure (Table 5), intermittently between Sep $12$, $2017$ (MJD $58008.23$) till Sep $17$, $2017$ (MJD $58013.15$). These were the days when the source was reported by monitoring observations to be softening compared to the initial hard state though the spectrum below 10 keV remained non-thermal (Nakahira et al., 2017; Kennea, 2017; Palmer et al., 2017). We made use of data from SXT and LAXPC, which operate respectively in $0.3$ keV to $8$ keV and $3$ keV to $80$ keV bands, for the timing and spectral analyses. The AstroSat level 1 ToO data were obtained from the ISSDC data dissemination archive111https://gads.issdc.gov.in/astro-gads/. Table 5 lists details of all the AstroSat observations of MAXI J1535$-$571, including orbit numbers and exposure time.

The Swift mission, as mentioned earlier, continuously monitored the source from its first detection of the outburst. We made use of the XRT data in the $0.3$ keV to $10$ keV band, in order to augment the AstroSat data used in the analysis. The XRT data were obtained from the HEASARC public archive222https://heasarc.gsfc.nasa.gov/cgi-bin/W3Browse/w3browse.pl. Details of all publicly available Swift monitoring data are provided in Table 6.

We also used data from MAXI333http://maxi.riken.jp/mxondem/ to plot the outburst flux profile and the HID. Besides these X-ray instruments, we used radio results from the Australia Telescope Compact Array (ATCA) and Atacama Large Millimeter/submillimeter Array (ALMA) to study the radio flux levels during the outburst. Note that the radio results used are the ones available in Russell et al. (2017a); Russell et al. (2017b); Tetarenko et al. (2017). There are more observations reported by ALMA (Russell et al., 2018a). However, those results are not yet published.

2.1 AstroSat Data Reduction

Firstly, we carried out temporal and spectral analysis of the source MAXI J1535$-$571 using AstroSat data. The capabilities of LAXPC for timing and spectroscopic studies of various sources are already demonstrated by Yadav et al. (2016a); Misra et al. (2017); Verdhan Chauhan et al. (2017). We followed these papers and the instructions444http://astrosat-ssc.iucaa.in/?q=laxpcData provided along with the LAXPC analysis software (LaxpcSoft555http://www.tifr.res.in/$\sim$astrosat_laxpc/LaxpcSoft.html) released on May 19, 2018. The details of how the responses and background spectra are generated are given by Antia et al. (2017).

The three LAXPC detectors (LAXPC10, LAXPC20 and LAXPC30) were operated in the normal or Event Analysis (EA) mode, in all those observations between Sep $12$ – Sep $17$, $2017$. Of the $62$ sets of data in Table 5, we used $15$ for our analysis, sampling a few from the several orbits of observation per day. Selected orbit numbers have been bold-faced in the table. As will be evident from the results in §4 and §5, the temporal and spectral parameters do not evolve indicating that the selected observations are enough to characterise the source.

Of the three LAXPC detectors, data from LAXPC10, and in that, the eventlist subset corresponding to the longest available good time interval slot has been used for analysis. The lightcurve has been extracted in the $3-80$ keV energy range with a resolution of $0.005$ s which corresponds to power spectra ranging upto a frequency of $100$ Hz. The background subtracted mean count rate for the LAXPC10 observations varied from around $4400$ cts/s to $6300$ cts/s. We also generated lightcurves in the energy bands $3-20$ keV, $20-35$ keV, $35-50$ keV and $50-80$ keV for selected observations to study the energy dependent evolution of the features in the power spectrum. Further, the source and background energy spectra were generated. Out of the three layers of LAXPC, we used data from the top layer only. Moreover, only single events were considered for generating the spectra. These options were selected to have a minimal undesired bump in the spectral residuals at around 33 keV. A systematic error of $2\%$ has been added while creating the background spectrum. XSPEC v 12.9.1 has been used for spectral analysis and modelling.

SXT is an instrument capable of X-ray imaging and spectroscopy in the $0.3-8$ keV energy range. SXT has a focusing telescope and a CCD detector (Singh et al., 2017). Extraction of level $2$ data from SXT has been done using the sxtpipeline tool. The SXT data are available in the Fast Windowed (FW) mode. The extracted cleaned event file is used to generate energy spectra with XSELECT V2.4d. The image count rate at the central pixels goes as high as 369 cts/s which crosses the pile-up limit of 344 cts/s for FW mode as indicated in the AstroSat Handbook666http://www.iucaa.in/$\sim$astrosat/AstroSat_handbook.pdf. To avoid pile-up effects, an annular region centred at the source RA and DEC, and of $1^{\prime}$ inner radius and $4^{\prime}$ outer radius is chosen to extract the spectrum. In Figure 1, we have shown an example of choosing source region excluding the central $1^{\prime}$ from SXT data. It is evident from the figure that the largest radius for the annulus possible in this mode is $4^{\prime}$. The average count rate of the entire image is 175 cts/s and it reduces to 101 cts/s after filtering the region using the annulus as mentioned above. We used the background, response and ARF files provided by the SXT instrument team777http://www.tifr.res.in/$\sim$astrosat_sxt/dataanalysis.html for the analysis.

2.2 Swift/XRT Data Reduction

Among all the Swift observations of the source in the public archive, the ones that we selected for analysis are indicated by bold-faced ObsIds in Table 6. All XRT observations from 02 Sep 2017 to 10 Mar 2018 with an exposure greater than 800 s were considered for this analysis. For temporal analysis, lightcurves, power spectra and hardness ratios of all datasets were obtained. XRT data extraction was done using steps outlined in the XRT reduction manual888https://swift.gsfc.nasa.gov/analysis/xrt_swguide_v1_2.pdf. HEAsoft v 6.21 was used to reduce the data with CALDB updated till Jan 2018. All XRT data used were taken in the Windowed Timing (WT) mode. Initial reduction was done using the ftool xrtpipeline with event grades between $0-2$ chosen. The cleaned event files so obtained were used for further analysis.

For generating the power spectra, lightcurves with a bin size of 0.005 s were made using photons in the energy range $0.5-10$ keV and $3-10$ keV (for compatibility with AstroSat results). All observations with an average count rate $>$ 100 cts/s were treated for pile-up effects. We selected the source extraction region using an annulus of inner radius $12.5^{\prime\prime}$ and outer radius $50^{\prime\prime}$, centred on the source position. Data from the region within the inner radius of the annulus was discarded to correct for pile-up. The power spectra were generated from these pile-up corrected lightcurves as explained in §4.2. Lightcurves with a time bin of 1 s in the bands $0.5-4$ keV and $4-10$ keV were created to compute the hardness ratio between these bands. In the next section, we present the outburst profile and HID.

3 2017 outburst profile and Hardness Intensity

Figure 2 shows the MAXI outburst profile in $2-20$ keV, XRT light curve ($0.3-10$ keV), radio detections made using ATCA & ALMA, MAXI  HR1 ($4-10$ keV / $2-4$ keV) and MAXI  HR2 ($10-20$ keV / $4-10$ keV), respectively, from top to bottom. MJDs corresponding to AstroSat observations are shaded in all the panels. It can be noticed that the MAXI and XRT flux vary similarly. Radio flux of the order of a few hundred mJy was observed just before the observations by AstroSat. It is to be noted that the radio activity observed in black hole binaries are strongly correlated with spectral states (Radhika et al., 2016), and thus the radio results confirm that MAXI J1535$-$571 was making a transition to HIMS in that period. HR1 decreased from a high value of around 1.0 at the beginning of the outburst to below 0.1 as the flux reduced considerably. Meanwhile HR2 remained low throughout the outburst except during the initial few days.

Figure 3 shows the Hardness Intensity Diagram (HID) produced from the same MAXI data points used in Figure 2. The hardness ratio ($6-20$ keV / $2-6$ keV) is plotted against MAXI $2-20$ keV band flux. The uncertainty on the MAXI flux increased as the outburst activity dropped beyond MJD $\sim\>58100$ and that in turn caused the significant spread in the HID regime of flux $\leq\>2$ ph cm*-2* s*-1* and hardness $\leq\>0.1$. The LAXPC observations were made just before the outburst peak, when the source was in HIMS with a hardness around $0.3$. In this figure, LHS is represented by red circles, HIMS by golden filled circles, SIMS by turquoise filled inverted triangles, and soft state is represented using cyan triangles. The states are marked as per the classification done by Tao et al. (2018).

The blue shaded region representing AstroSat observations, falls in HIMS of the outburst. Beyond the HIMS, the source hardness decreased as can be seen from Figure 3. In this phase starting from MJD 58015, the source entered SIMS (Huang et al., 2018). The HID for this source does not follow the exact ‘q’ - shape that is exhibited by canonical black hole binaries (Belloni et al., 2005; Nandi et al., 2012; Radhika et al., 2018). Instead, we noticed an increase in hardness ratio after the source reached its peak indicating a transition back to the HIMS. This continued for a few days (MJD 58049 to 58051) before the hardness ratio started decreasing again. Though there is no XRT observation following this until MJD 58137, the decrease in hardness ratio from MAXI HID indicates that the source had gone back to SIMS in this period. After that, both flux and hardness ratios decreased considerably. Tao et al. (2018) have classified this period as soft state (from MJD 58137 to 58230). In §4 we present the procedure and results of temporal analysis of this outburst.

4 Timing Analysis and Results

4.1 Timing Analysis

Initially, we generated 1 s lightcurves from LAXPC and obtained the count rates corresponding to each observation. We did not observe any signature of structured variability in these lightcurves. Then we created $0.005$ s resolution lightcurves from LAXPC10 data sets to generate the power spectra. We divided each lightcurve in the intervals of $8192$ time bins, and then created power density spectra (hereafter PDS) for each interval. The final PDS for each lightcurve was generated after averaging these individual PDS. The averaged PDS was binned by a factor of $1.05$ in frequency space. We have fitted the PDS with the combination of $4$  Lorentzians representing a broad band limited noise (BLN; a zero centered Lorentzian), a low frequency noise (LFN), QPO and its upper harmonics. We took the QPO frequency as the centroid of the Lorentzian component fitting the QPO. QPOs were identified by fitting a multi-Lorentzian model to the entire power spectrum and searching for features with quality factor,  $Q=\frac{\nu}{\Delta\nu}>3$ and $Significance=\frac{Norm}{Err_{norm}}>3$, where $\nu$ is Lorentzian centroid frequency, $\Delta\nu$ is FWHM and $Err_{norm}$ is the negative error in norm.

Further, we subtracted the dead-time ($\tau_{d}$) affected Poisson noise level (Agrawal et al., 2018). Besides altering the noise level, the dead-time effects can also change the source rms (van der Klis, 1989). Therefore, we applied correction on rms as prescribed by Bachetti et al. (2015) to account for the effect of dead-time. The dead-time corrected rms ($rms_{in}$) is calculated as $rms_{in}=rms_{det}/(1-\tau_{d}\>r_{det})$, where $rms_{det}$ is the detected rms and $r_{det}$ is the detected count rate. We modelled the resultant PDS in frequency-rms space using multiple Lorentzians as mentioned before. The QPO rms amplitude in percentage was computed by finding the area under the Lorentzian corresponding to the QPO, in the frequency range between 0.05 to 100 Hz. RMS calculations were done for the full band ($3-80$ keV) as well as for the $3-10$ keV using LAXPC10 data in the $0.05-100$ Hz frequency range. Similarly, we computed rms for the XRT data in the range $0.5-10$ keV as well as for the $3-10$ keV band in order to compare with LAXPC10 results. Total rms of each PDS was obtained by adding the rms of individual features in quadrature.

Figure 4 shows the power spectrum corresponding to the observation on MJD 58008.24. This was obtained from a lightcurve in the range $3-80$ keV. The power spectrum in frequency-rms space was modelled with multiple Lorentzians corresponding to fundamental QPO , its harmonic, BLN and LFN. The fundamental frequency was found to be $2.21\pm 0.02$ Hz with an FWHM of $0.24\pm 0.02$ Hz. The second harmonic was found at $4.37\pm 0.03$ Hz with an FWHM of $0.76\pm 0.10$ Hz. The percentage rms of the fundamental frequency was $8.23\pm 0.49$ and for the second harmonic, it was $4.10\pm 0.37$. The BLN contributed $11.30\pm 1.44$% rms and the LFN had rms of $6.56\pm 1.02$%. The $\chi^{2}/dof$ for this fit was 0.98 (115.7/118). Unless mentioned explicitly, all the errors are computed using $\Delta\chi^{2}=1.0$ (68% confidence level) for all the observations (see Tables 1 and 2). The detection of a QPO and its harmonic in this PDS encouraged us to do an energy dependent analysis of the power spectra, and the details are presented in §4.3.

4.2 Relation between flux and RMS

We generated the power spectra from XRT lightcurves (pile-up corrected) in the $0.5-10$ keV energy range, using the ftool powspec, considering 8192 bins per interval. The dead-time correction factor for XRT is $\sim$ 1 and so it does not affect the power spectrum. Exposure correction was not applied as it can affect the Poissonian nature of the noise power. The power spectrum in frequency-rms space was integrated from 0.05 to 100 Hz to compute the rms value. The evolution of fractional rms variability of the XRT observations with respect to hardness and intensity is shown in Figure 5. Hardness was calculated as the ratio of flux in $4-10$ keV to the flux in $0.5-4$ keV. Total rms for each PDS was calculated using the imodel command of QDP999https://heasarc.gsfc.nasa.gov/ftools/others/qdp/qdp.html. Here hollow circles represent LHS, filled circles represent HIMS, inverted triangles stand for SIMS and hollow triangles indicate soft state. The arrows have colours assigned as in Figure 3 based on the state at the tail of the arrow. Motta et al. (2012); Belloni & Motta (2016) show the general trend of an HRD corresponding to a canonical black hole binary outburst where the rms during LHS varies from $\sim 20$ to $40$%, HIMS from $\sim 10$ to $20$% and below $\sim 5$% in the softer states. Figure 5a shows the HRD for MAXI J1535$-$571 with XRT data from which it was evident that at the beginning of the outburst in LHS, the hardness (0.56) and rms (around 35%) are both at a comparatively larger value. In the HIMS, the rms value was observed to be in the range 10 to 20%. As the outburst progressed, the rms value decreased to around 6% in SIMS. AstroSat observations are indicated by the blue shaded portion. These observations were made just before the drop in rms values. The observation from soft state shown in the HRD has a significantly high value of rms ($\sim 17$%) as compared to other sources (see Nandi et al., 2012; Motta et al., 2012; Radhika et al., 2018).

From Figure 5b, we can see that in the LHS, though the source had highest rms ($\sim 35$%), the flux was below $\sim 400$ cts/s. In HIMS, we observe count rates in between 600 and 1500 cts/s and rms within $\sim 10$ to $25$%. RID for other sources (Muñoz-Darias et al., 2011) also have higher count rates in the intermediate states as compared to the hard state. The rms and hardness of the AstroSat observation period suggests that the source was in HIMS as indicated by the golden arrows and filled circles. This is consistent with the spectral state classification by Tao et al. (2018). The RID shows dip in rms values in the SIMS. Besides this, the intensity peaks (above 2000 cts/s) during the time when the rms has lowest values ($\sim 6$%).

4.3 Temporal Properties

As detailed above, we modelled all the 15 LAXPC power spectra with multiple Lorentzians and it was observed that the frequency of the fundamental QPO varies from $1.85$ Hz to $2.88$ Hz. The quality factor (Q) for the QPOs are around 9.0 and the significance was also high ($\sim$ 10.0). The QPO rms amplitude was near 8.5% for all the LAXPC observations. The harmonics have Q-factor around 5.0, significance around 10.0 and rms varying from 3.68 to 4.85%. Low frequency noise (LFN) with rms around 6.5% and band limited noise (BLN) with rms around 10% are also present in the power spectra. These QPO parameters are tabulated in Table 1. There was no indication of increase/decrease of fundamental frequency with time. Based on Casella et al. (2005); Belloni & Motta (2016), we have classified the QPOs as C-type as they are in the frequency range from $\sim 2-3$ Hz, have $Q\sim 9$ and rms of around 8.5%, and appear along with strong band limited flat noise component. The strong QPOs observed in this source could be possibly due to the oscillation of post-shock corona (hereafter PSC; see Das et al. 2014).

Figure 6 consists of the energy dependent Power Density Spectra (PDS) obtained with AstroSat/LAXPC for the following energy ranges: $3-20$ keV, $20-35$ keV, $35-50$ keV and $50-80$ keV, for MJD 58008.24 (Orbit 10585), 58009.02 (Orbit 10597) and 58012.24 (Orbit 10644). Table 2 lists the energy dependent QPO parameter variations for MJDs 58008.24, 58009.02 and 58012.24. It is evident that the fundamental QPO appears in the range $3-50$ keV, while the second harmonic was significant only in the $3-35$ keV band. No QPOs were seen in $50-80$ keV band. A low frequency noise (LFN) and a band limited noise (BLN) also contributed to the power in energy range $3-20$ keV.

The evolution of QPO centroid frequency, total rms and QPO rms are presented in Figure 7. Here, XRT results are shown in red while LAXPC results are shown in blue. In order to compare results, QPO parameters were calculated in $3-10$ keV for both XRT and LAXPC. Note that, the XRT PDS in the $0.5-3$ keV band had no QPO like features and these photons piled-up to bring down the QPO rms values in $3-10$ keV band. Therefore, we applied pile-up correction as mentioned in §2.2 in order to estimate the correct rms in $3-10$ keV XRT data. It was observed that the QPO frequency did not evolve except for the first four Swift observations. The first QPO of 0.44 Hz was detected with XRT data on MJD 58004.29 when the source was in the LHS. Beyond that, the fundamental frequency was almost constant during the HIMS as indicated by both XRT and LAXPC observations. The total rms for the LAXPC observations was around $\sim$ $16.2\pm 1.1$% and the corresponding QPO rms remained almost constant at $7.1\pm 0.52$%. The total rms in XRT PDS during the HIMS was around $15\pm 2.4$%, and the QPO rms was found to have an average value of $6.9\pm 1.3$%. The larger uncertainties on width and normalisation of the Lorentzians resulted in larger errors on rms values. This is due to the lower photon statistics of XRT as compared to the high LAXPC count rate in $3-10$ keV band. Only three QPOs ($\sim 4.3$ Hz) were detected between MJD 58014 and MJD 58040 with XRT data. During this period the source was in SIMS (Tao et al., 2018) with very high fluxes (see Figure 2). The total integrated rms in $3-10$ keV was larger during LHS and HIMS of the outburst than during the SIMS. As the source underwent a transition back to HIMS during MJD 58049 to 58051, the rms observed then was around 18%. We also detected the maximum QPO frequency ($\sim 6.72$ Hz) from the source on MJD 58050.98. QPOs were not observed beyond this date upto MJD 58200. The presence of C-type QPOs and their frequencies indicated that the source was likely to be in HIMS during AstroSat observations. In §5, we have presented the results of spectral analysis and modelling with the AstroSat data. Further, we intended to estimate the accretion parameters of the source and constrain the mass of the black hole. It was with this motivation that we decided to use the two-component flow model (see §5.3).

5 Spectral Analysis and Results

5.1 Spectral Analysis

Initially, we modelled the SXT spectrum of MJD 58008.24 (Orbit 10585) with Tbabs (Wilms et al., 2000), diskbb and powerlaw. This gave an inner disc temperature of 0.23 keV and a powerlaw photon index of 2.31 with a $\chi^{2}/dof$ of $782.4/612=1.27$. Energy spectrum from SXT in the range $0.7-7$ keV was then modelled with thermally comptonized continuum - nthComp (Zdziarski et al. (1996), Życki et al. (1999)). We used gain correction with a fixed slope of 1 using the gain fit command to flatten the residuals at $1.8$ and $2.2$ keV. Even after that, residuals were slightly on the higher side at around $1.7$ keV and below $0.8$ keV. The modelling with nthComp gave a $\chi^{2}/dof$ of $866/613=1.41$.

Then we modelled the corresponding LAXPC10 spectrum in the range $3-80$ keV initially with diskbb + powerlaw and then with nthComp. The fit with Tbabs $\times$ (diskbb + powerlaw) gave poor results at a $\chi^{2}_{red}\sim 6$ and it was found that the diskbb component was insensitive to the fit. The fit with Tbabs $\times$ (nthComp) was better and gave $\chi^{2}_{red}$ of $504/274=1.84$. Further, the inclusion of a gaussian to account for the iron line at around 6.6 keV and an instrumental Xenon edge around 33 keV improved the fit resulting in a $\chi^{2}_{red}$ of $272.8/270=1.01$. The strong iron line indicated the possible presence of a continuum reflection component in the spectra. In order to check this, we introduced a reflection component (ireflect) and found that the relative reflection coefficient was too low ($0.03-0.08$). Moreover, the $\chi^{2}/dof=272.61/269=1.01$, did not improve after including reflection component. Therefore, we inferred that a reflection continuum was not required to model the energy spectra of the source using data from AstroSat observations. The fit results implied that when we combine SXT and LAXPC data, it was better to model with nthComp than using the phenomenological model containing diskbb and powerlaw.

A 2% systematic error in each bin was incorporated in the spectral response as specified in the SXT user manual101010http://www.tifr.res.in/$\sim$astrosat_sxt/dataanalysis.html and by Leahy & Chen (2019). The systematic error for spectral data from LAXPC detectors was also reported to be within 2% (Antia et al., 2017). Therefore, the combined spectral fit with both SXT and LAXPC were considered to have a 2% systematic error in order to achieve a good estimate of the spectral fit parameters. Once good independent fits were obtained, we attempted for combined fitting of energy spectra with SXT and LAXPC data using nthComp. The gaussian line width was frozen to 0.6 keV for the broadband modelling. This gave a reasonable $\chi^{2}_{red}$ of $1181/895=1.31$ with a photon index of $2.254_{-0.005}^{+0.004}$, electron temperature of $37.8_{-2.37}^{+2.74}$ keV and seed photon temperature of $0.21_{-0.04}^{+0.02}$ keV on MJD 58008.24. Unless mentioned explicitly, all the errors are computed using $\Delta\chi^{2}=1.0$ (68% confidence level) for all the observations (see Table 3). Further, we carried out spectral analysis for all the observations under consideration with nthComp and the results are presented in §5.2.

The hydrogen column density used for this source so far by other investigators for spectral fitting are $(4.44\pm 1.1)\times 10^{22}$ atoms cm*-2* (Russell et al., 2018b; Parikh et al., 2018) and $4.05\times 10^{22}$ atoms cm*-2* (Miller et al., 2018). Russell et al. (2018b) arrived at this value for $N_{H}$ (with 25% error) using three observations from the declining phase of the outburst, when the source was undergoing soft to hard state transition. Tao et al. (2018); Shang et al. (2018) have left $N_{H}$ parameter free. It should be noted that the HEASARC $N_{H}$ calculator111111https://heasarc.gsfc.nasa.gov/cgi-bin/Tools/w3nh/w3nh.pl gives hydrogen column density of only $1.46\times 10^{22}$ atoms cm*-2*. The $N_{H}$ value can change with states of the source due to variation in radio activity (jets/outflows). As AstroSat observations of the source were made when the radio flux varied, we attempted to determine $N_{H}$ density independently from the AstroSat spectra. The combined fits with SXT and LAXPC using free $N_{H}$ parameter resulted in $N_{H}$ values within a narrow range of $2.52\times 10^{22}$ to $2.67\times 10^{22}$ atoms cm*-2*. As our spectral modelling indicated minimal scatter of the $N_{H}$ value, the Tbabs component was left free throughout our analysis.

5.2 Spectral Properties

As detailed in §5.1, the appropriate model to fit the energy spectra (in the energy range $0.7-80$ keV) is Tbabs $\times$ constant(gaussian $+$ edge $\times$ nthComp). In Figure 8, we have shown the broadband spectrum from MJD 58011.13 fitted with the above model based on nthComp. The photon index for this fit is $2.238\pm 0.005$ with a $\chi^{2}_{red}=1.19~{}(1071/895)$ indicating that the source was in HIMS. The results of spectral modelling and the $\chi^{2}_{red}$ values for all observations are included in Table 3. The entries indicated with a $\dagger$ symbol correspond to broadband observations combining SXT and LAXPC data. All other entries in this table correspond to LAXPC observations alone. The model based on nthComp satisfactorily fits all these observations. We find that the photon index ($\Gamma$) values vary from $2.18$ to $2.37$ only and did not show much evolution within the six days of AstroSat observation. The value of electron temperature, $kT_{e}$ is in the range between $21.2-63$ keV indicating a minimum high energy cut-off of $\sim 60$ keV ($3~{}kT_{e}$). For a few observations, we were unable to constrain $kT_{e}$ from the fit and in those cases we fixed it to a high value of 100 keV (see Table 3). The seed photon temperature parameter $kT_{bb}$, which specifies the low energy roll over varied between $0.19$ to $0.29$ keV. The norm remained in the range 21.6 to 53.9. A Xenon edge in the range $32-34$ keV was required in all cases. The gaussian line energy around 6.6 keV had width fixed to 0.6 keV. The total flux in the $0.3-80$ keV during the AstroSat observations was high and it varied from $7.14\times 10^{-8}$ to $1.10\times 10^{-7}$ erg cm*-2* s*-1*. Assuming the source to be at 8 kpc distance as considered by Tao et al. (2018), we computed the luminosity to be in the range between $5.46~{}\times~{}10^{38}$ to $8.38~{}\times~{}10^{38}$ erg s*-1*. For a black hole of mass 6 $M_{\odot}$, the luminosity when expressed in terms of Eddington Luminosity is $0.70$ to $1.07$ $L_{Edd}$. Using the $\Gamma$ and $kT_{e}$ parameters, we also computed the optical depth ($\tau$) of the system based on the formula given by Zdziarski et al. (1996). $\tau$ was found to be in the range $0.72-2.5$.

As mentioned in §4.3, the presence of strong QPOs in the power spectrum helps us to conclude that the corona is oscillating in nature. The lack of evolution in the QPO frequency and the tight range of spectral photon indices both point to the fact that the corona is of almost constant size during the AstroSat observations. In the next section, we present the details of spectral modelling with two-component accretion flow and determine the shock location which indicates the size of the PSC. We also attempt to constrain the mass of the source using the same model.

5.3 Spectral modelling: two-component accretion flow

Further, we have used the two-component accretion flow (Chakrabarti & Titarchuk, 1995; Chakrabarti & Mandal, 2006) to model the broadband energy spectra ($0.7-80$ keV) using spectral data from SXT and LAXPC. In this model, a Keplerian disc is situated at the equatorial plane of the binary system and a sub-Keplerian halo component resides on top and bottom of the Keplerain disc. The inner boundary condition of the accretion flow around a black hole demands a supersonic flow and the rotating supersonic flow faces a virtual centrifugal barrier close to the horizon. This triggers a shock discontinuity in the accretion flow (Fukue, 1987; Chakrabarti, 1989) and the flow becomes sub-sonic after passing through a shock. Here, the kinetic energy gets converted into thermal energy and this ‘hot’ region beyond the shock is referred to as the post-shock corona (PSC). The Keplerian disc produces multi-colour blackbody photons and a significant fraction of these photons intercept the ‘hot’ PSC. The high energy photons are formed due to the inverse Comptonisation of soft photons from the Keplerian disc in the PSC. This model has five parameters (see Iyer et al., 2015) namely the mass of the black hole ($M$), the shock location ($x_{s}$), the sub-Keplerian halo rate ($\dot{m}_{h}$), Keplerian disc accretion rate ($\dot{m}_{d}$) and free normalisation constant. Here, mass of the black hole is expressed in solar mass unit ($M_{\odot}$), radial distance is expressed in Schwarzschild radii ($r_{g}=2GM/c^{2}$), where $G$ is gravitational constant and $c$ is the speed of light in vacuum. The accretion rate is expressed in unit of Eddington rate ($\dot{m}_{Edd}$).

This model is implemented in XSPEC as a table model (Iyer et al., 2015). In Figure 9, we have presented a fit of combined spectrum of SXT+LAXPC data corresponding to MJD 58011.13 (Orbit number 10628) using two-component model along with an additional highecut component of XSPEC. The gaussian and edge components used in our earlier model was required here as well. The disc and halo accretion rates are respectively $0.977_{-0.157}^{+0.109}$ and $0.203_{-0.022}^{+0.019}$ $\dot{m}_{Edd}$. The corresponding shock location is $16.3_{-0.99}^{+0.80}~{}r_{g}$. The mass of the black hole obtained by modelling this particular observation with the two-component flow model is $6.31_{-0.86}^{+0.68}$ $M_{\odot}$. In Table 4, we have presented the results of broadband spectral fitting using two-component accretion flow model (at 90% confidence level) for two representative cases and observed that the spectral state of the source did not change much. Model fitting for the two observations considered suggests that the possible mass of the BH source is in the range $5.14-7.83~{}M_{\odot}$ and the relative accretion rate (see Table 4) indicates the source to be in intermediate state (Nandi et al., 2018). We have chosen two broadband ($0.7-80$ keV) observations (MJD 58011.13 and 58012.24) for two-component flow modelling as the QPO frequencies in these cases are not the same. We expect that the shock location which is an estimate of the size of the PSC decreases with an increase in QPO frequency. The results of physical modelling and temporal modelling indicates the same, as we have a shock location of $\sim 16.3~{}r_{g}$ and QPO frequency of $\sim 2.17$ Hz on MJD 58011.13, while the shock location on MJD 58012.24 is only $\sim 15.9~{}r_{g}$ corresponding to a higher QPO frequency of $\sim 2.59$ Hz (see Tables 1 and 4). Also, the relative disc to halo accretion rates in both cases are similar, indicating that the source has not undergone a state transition during the AstroSat observations.

This model has a free normalisation which is a function of inclination angle and distance to the source. The normalisation constant shifts the scale of the overall luminosity of the source. The mass and accretion rates are independent quantities in the model and both determine the source luminosity as well as the spectral characteristics. In fact, for a given mass, the change in accretion rates determines the spectral evolution during the outburst. Hence, only a specific range of mass can model all the observed data. We believe that the observed data may include some contributions from the base of the jet, particularly in intermediate states. This is not considered in the current model. Also, as discussed in §6, the spin parameter of the black hole is not taken into account in this model. With the inclusion of these contributions in the model, the estimated accretion parameters may differ slightly.

6 Discussion and Conclusions

The MAXI detection of a bright X-ray transient lead to a large follow up observation campaign with multiple observatories following the outburst. These observations point to the black hole binary nature of this source, which shows state transitions similar to a canonical black hole binary. AstroSat made a few observations in the intermediate state of the source. In this paper, we have studied the outbursting source MAXI J1535$-$571 using four different instruments namely MAXI, Swift/XRT, AstroSat/SXT and AstroSat/LAXPC. Using MAXI, we have studied the outburst profile (Figure 2) and the HID (Figure 3) of the entire outburst. This source showed a slightly unusual HID evolution with extremely high flux seen in the intermediate states before it dimmed considerably during its transition towards the soft state. The source has gone through LHS, HIMS, SIMS, HIMS, SIMS and soft state in the rising phase. We have not considered data beyond MJD 58200 due to weak activity of the source leading to large errorbars in counts, though Tao et al. (2018) have classified the source to be in soft state from MJD 58137 to 58230.

We made use of observations from XRT to study the state transitions based on the HRD and RID. From the HRD (Figure 5a) it is evident that the hardness decreases as the states evolve and rms is comparatively lower during the SIMS. The RID (Figure 5b) suggests that the rms is larger during the harder states as expected and is low during the SIMS which corresponds to the highest count rate. The AstroSat observations of the source shaded in Figures 5a and 5b falls in the HIMS region.

XRT and LAXPC data were used to carry out timing analysis of the source. We detected C-type QPOs in the range $1.85-2.88$ Hz along with second harmonics during the rising phase of the outburst with LAXPC data. The power spectra also showed band limited noise and low frequency noise. We confirm that this period corresponds to the HIMS phase of the outburst as classified by Tao et al. (2018); Huang et al. (2018). Using HXMT data, Huang et al. (2018) have detected QPOs upto 100 keV, though LAXPC showed no evidence of QPOs beyond 50 keV. This is probably due to the fact that the LAXPC has only an effective area in the range 4100 cm2 to 2200 cm2 as energy changes from 60 keV to 80 keV (Antia et al., 2017), whereas HXMT has a higher effective area of 5100 cm2 in the $20-250$ keV energy range (Li et al., 2018). QPOs are thought to be signatures of physical oscillations in the accreting flow and many reasons have been proposed for explaining the presence of such oscillations. Magneto-rotational instabilities (MRIs) in the accretion disc (Machida & Matsumoto, 2008) can be a reason for QPOs. During the process of accretion, when the magnetic energy accumulated in an accretion disc at a particular stage is released, the disc becomes weakly magnetized and axisymmetric. The magnetic energy again gets amplified due to MRI and the cycle repeats. The period of these cycles results in low frequency peaks around 4 to 8 Hz. LFQPOs in black hole binaries can possibly be generated due to these magnetic cycles. Lense-Thirring effects in the accretion disc (Ingram et al., 2009) may also be a reason for QPOs. In this case, the rotation of the central black hole causes frame-dragging in the vicinity of the black hole. This in turn, causes the orbits of particles in the accretion disc to precess and the associated precession frequency is found to be close to observed QPO frequencies. In this work, we consider the two-component accretion flow framework, under which an oscillating PSC (Chakrabarti et al., 2008) can cause these QPOs.

It is evident from the energy dependent analysis of power spectra that the QPO rms exceeded $\sim$8% in $3-20$ keV energy band. It increased up to $\sim$9% in the $20-35$ keV range and in the $35-50$ keV energy band the QPO rms decreased slightly to $6-7$%. Rodriguez et al. (2002) have reported similar energy dependence of LFQPOs for GRS 1915+105. Rao Jassal et al. (2016) have studied the connection between the QPOs and the spectral components and have shown that a small in-phase variation of the spectral index with the QPO could reproduce the energy dependence of QPO rms. This is consistent with the origin of LFQPOs due to oscillation of PSC. For a given QPO frequency, the PSC will oscillate with respect to an average position. During oscillation, as PSC expands the spectral index decreases (i.e. harder spectrum) and when the PSC shrinks the spectral index increases. This behaviour is same as the change in spectral index due to the change in the shock location as shown in Mandal & Chakrabarti (2005).

Further, SXT+LAXPC combined broadband spectra were analysed to examine the source characteristics. Initially, the spectra were modelled with a thermally Comptonised continuum (nthComp) along with a Xenon edge and an iron line (gaussian). The seed photon temperature was consistently in $0.19$ to $0.29$ keV range. The high energy cut off was at least $\sim 60$ keV as inferred from the electron temperature ($kT_{e}$). The photon index of the energy spectrum was in a narrow range of $2.18-2.37$ throughout the observations with AstroSat. This along with the lack of QPO frequency evolution indicates that the oscillating post shock corona was almost constant in size during this period. The phenomenological modelling implies that the spectrum was completely dominated by thermal Comptonisation. The absence of a separate disc component in the model, suggests that disc was either truncated away as indicated by the low seed photon temperature or was buried within the corona. The soft photons supplied by the disc enters into the corona which has a moderate optical depth ($0.72<\tau<2.5$) and produces the strong comptonised component.

Moreover, as the relative reflection coefficient was not significant in the energy spectra of HIMS observations with AstroSat, we expected that the iron abundance must be high to get a strong iron line feature. In order to verify the results, we modelled the SXT + LAXPC broadband spectra of HIMS with relxillCp (García et al., 2014), which includes Comptonisation and reflection taking into consideration the iron abundance. The value of reflection coefficient obtained from relxillCp was $0.07\pm 0.02$ and the iron abundance was in the range between $1.2-1.7$ times solar abundance. The larger than solar abundance levels of iron caused the strong iron lines in the energy spectra. The iron abundance is comparable to the value $1.4_{-0.1}^{+0.3}$ times solar abundance, as obtained by Xu et al. (2018), during LHS (MJD 58003). The equivalent width (EW) of the iron line was $\sim 0.11$ keV, which falls in the range between 0.05 keV to 0.3 keV, typical for black hole binaries (Gilfanov, 2010).

Besides this, we modelled the combined SXT + LAXPC energy spectra using the two-component accretion flow model. The results from Table 4 show that the Keplerian-disc accretion rate is around the Eddington accretion limit. Similar rates are seen in XTE J1859$+$226 (Nandi et al., 2018) during its HIMS, but not usually seen in other black hole binaries (see Iyer et al., 2015; Radhika et al., 2018). The near Eddington accretion rates (see Table 4) observed in the intermediate states seem to be the reason for the excessive brightening of the source with luminosity as high as 1.07 $L_{Edd}$ (see Table 3 and §5.2). High accretion rates can trigger outflows to escape from the accretion stream, which can be observationally seen as radio emissions (Radhika & Nandi, 2014; Radhika et al., 2016). The large absorption columns reported (as compared to the expected galactic column from HI maps) also point to the presence of some outflows. Baglio et al. (2018) report wild infra-red flickering during this period, consistent with erratic outflows.

The spectral modelling with two-component flow suggests a compact corona of almost constant size ($\sim 16~{}r_{g}$). As a result, the C-type QPOs formed due to the oscillations of the corona are in a narrow range of frequencies as observed from LAXPC data. It should be noted that, the model is applicable only for a stationary black hole. The effect of black hole spin on the radiation spectrum is that the spectrum gets harder (Mandal & Mondal, 2018) as the black hole spins faster. Hence, inclusion of black hole spin will change the estimated value of the size of PSC for a given data set. In future, we aim to model the spectral data including the black hole spin.

Finally, physical modelling of the energy spectra from AstroSat observations based on the two-component accretion flow model suggests that the source hosts a black hole of mass $5.14-7.83$ $M_{\odot}$, which is similar to other Galactic black hole binaries (Corral-Santana et al., 2016; Tetarenko et al., 2016; Nandi et al., 2018). In summary, we have characterised the spectro-temporal properties of the source MAXI J1535$-$571 using AstroSat data during its 2017 outburst with the help of MAXI and Swift results.

Acknowledgement

Authors are thankful to the reviewer for his/her valuable suggestions and comments that helped to improve the quality of the manuscript. We acknowledge the clarifications that we got from Dr. Matteo Bachetti regarding dead-time effects. We are thankful to the Swift help desk for the information on pile-up and dead-time effects in XRT. This research has made use of data obtained through the High Energy Astrophysics Science Archive Research Center (HEASARC) online service, provided by the NASA/Goddard Space Flight Center. This publication uses the data from the AstroSat mission of the Indian Space Research Organisation (ISRO), archived at the Indian Space Science Data Centre (ISSDC). This work has used the data from the Soft X-ray Telescope (SXT) developed at TIFR, Mumbai, and the SXT POC at TIFR is thanked for verifying & releasing the data and providing the necessary software tools. This work has also used the data from the LAXPC Instruments developed at TIFR, Mumbai, and the LAXPC POC at TIFR is thanked for verifying & releasing the data. We thank the AstroSat Science Support Cell hosted by IUCAA and TIFR for providing the LaxpcSoft software which we used for LAXPC data analysis. AN thanks GD, SAG; DD, PDMSA and Director, URSC for encouragement and continuous support to carry out this research.

Appendix A Observations with AstroSat

Appendix B Observations with Swift