1910.03609.pdf

arXiv:1910.03609v2 [astro-ph.HE] 11 Nov 2019

Mon. Not. R. Astron. Soc.

000

, 1–24 (2019)

Printed 12 November 2019

(MN L

X style file v2.2)

Investigating the multiwavelength behaviour of the flat

spectrum radio quasar CTA 102 during 2013–2017

F. D’Ammando

⋆

, C. M. Raiteri

, M. Villata

, J. A. Acosta-Pulido

, I. Agudo

A. A. Arkharov

, R. Bachev

, G. V. Baida

, E. Ben ́ıtez

, G. A. Borman

W. Boschin

, V. Bozhilov

, M. S. Butuzova

, P. Calcidese

, M. I. Carnerero

D. Carosati

, C. Casadio

, N. Castro-Segura

, W.-P. Chen

, G. Damljanovic

A. Di Paola

, J. Echevarr ́ıa

, N. V. Efimova

, Sh. A. Ehgamberdiev

, C. Espinosa

A. Fuentes

, A. Giunta

, J. L. G ́omez

, T. S. Grishina

, M. A. Gurwell

, D. Hiriart

H. Jermak

, B. Jordan

, S. G. Jorstad

, M. Joshi

, G. N. Kimeridze

E. N. Kopatskaya

, K. Kuratov

, O. M. Kurtanidze

, S. O. Kurtanidze

A. L ̈ahteenm ̈aki

, V. M. Larionov

, E. G. Larionova

, L. V. Larionova

C. L ́azaro

, C. S. Lin

, M. P. Malmrose

, A. P. Marscher

, K. Matsumoto

B. McBreen

, R. Michel

, B. Mihov

, M. Minev

, D. O. Mirzaqulov

, S. N. Molina

J. W. Moody

, D. A. Morozova

, S. V. Nazarov

, A. A. Nikiforova

M. G. Nikolashvili

, J. M. Ohlert

, N. Okhmat

, E. Ovcharov

, F. Pinna

T. A. Polakis

, C. Protasio

, T. Pursimo

, F. J. Redondo-Lorenzo

, N. Rizzi

G. Rodriguez-Coira

, K. Sadakane

, A. C. Sadun

, M. R. Samal

, S. S. Savchenko

E. Semkov

, L. Sigua

, B. A. Skiff

, L. Slavcheva-Mihova

, P. S. Smith

, I. A. Steele

A. Strigachev

, J. Tammi

, C. Thum

, M. Tornikoski

, Yu. V. Troitskaya

I. S. Troitsky

, A. A. Vasilyev

, O. Vince

(the WEBT Collaboration), T. Hovatta

S. Kiehlmann

, W. Max-Moerbeck

, A. C. S. Readhead

, R. Reeves

, T. J. Pearson

(the OVRO Team), T. Mufakharov

, Yu. V. Sotnikova

, and M. G. Mingaliev

Accepted 2019 October 02

ABSTRACT

We present a multiwavelength study of the flat-spectrum radi

o quasar CTA 102 during 2013–2017. We

use radio-to-optical data obtained by the Whole Earth Blaza

r Telescope, 15 GHz data from the Owens

Valley Radio Observatory, 91 and 103 GHz data from the Atacam

a Large Millimeter Array, near-

infrared data from the Rapid Eye Monitor telescope, as well a

s data from the

Swift

(optical-UV and

X-rays) and

Fermi

(

rays) satellites to study flux and spectral variability and t

he correlation between

flux changes at different wavelengths. Unprecedented

-ray flaring activity was observed during 2016

November–2017 February, with four major outbursts. A peak fl

ux of (2158

63)

−

ph cm

−

, corresponding to a luminosity of (2.2

0.1)

erg s

−

, was reached on 2016 December 28.

These four

-ray outbursts have corresponding events in the near-infra

red, optical, and UV bands,

with the peaks observed at the same time. A general agreement

between X-ray and

-ray activity is

found. The

-ray flux variations show a general, strong correlation with

the optical ones with no time

lag between the two bands and a comparable variability ampli

tude. This

-ray/optical relationship

is in agreement with the geometrical model that has successf

ully explained the low-energy flux and

spectral behaviour, suggesting that the long-term flux vari

ations are mainly due to changes in the

Doppler factor produced by variations of the viewing angle o

f the emitting regions. The difference in

behaviour between radio and higher energy emission would be

ascribed to different viewing angles of

the jet regions producing their emission.

Key words:

galaxies: nuclei – galaxies: jets – galaxies: individual: CTA 102 – gamma

rays: general – radiation mechanisms: non-thermal

⋆

E-mail: dammando@ira.inaf.it

2019 RAS

D’Ammando, Raiteri, Villata, et al.

1 INTRODUCTION

Blazars are an extreme class of active galactic nuclei (AGN)

whose bright and violently variable non-thermal radiation

across the entire electromagnetic spectrum is ascribed to

the presence of a collimated relativistic jet closely align

to our line of sight (e.g., Blandford & Rees 1978). This pe-

culiar setting implies a strong amplification of the rest-fr

ame

radiation because of Doppler boosting, together with a con-

traction of the variability time-scales, and a blueshift of

the

frequencies.

The relativistic jets of blazars are able to transport a

huge amount of power away from the central engine in the

form of radiation, kinetic energy, and magnetic fields. When

this power is dissipated, the particles emit the observed ra

diation, showing the typical double-hump spectral energy

distribution (SED) of blazars. The first peak of the SED,

usually observed between radio and X-rays, is due to the

synchrotron radiation from relativistic electrons, while

the

second peak, usually observed from X-ray up to TeV en-

ergies, is commonly interpreted as inverse Compton (IC)

scattering of seed photons, either internal or external to t

jet, by highly relativistic electrons. However, the nature

this second hump is a controversial issue and other models

involving hadronic and lepto-hadronic processes have been

proposed (e.g., B ̈ottcher et al. 2013).

Blazars are traditionally divided into flat-spectrum ra-

dio quasars (FSRQ) and BL Lac objects (BL Lacs), based

on the presence or not, respectively, of broad emission line

(i.e., equivalent width

A) in their optical and UV spec-

trum (e.g., Stickel et al. 1991). Recently, a new classificat

ion

was proposed based on the luminosity of the broad-line re-

gion (BLR) in Eddington luminosity (Ghisellini et al. 2011)

sources with

BLR

Edd

higher or lower than 5

−

being

classified as FSRQ or BL Lacs, respectively, in agreement

with a transition of the accretion regime from efficient to

inefficient between the two classes.

Blazar emission shows strong and unpredictable vari-

ability over all the electromagnetic spectrum, from the ra-

dio band to

rays, with time-scales ranging from minutes

to years. Long-term observations of blazars during differ-

ent activity states provide an ideal laboratory for investi

gating the emission mechanisms at work in this class of

sources. In this paper we present multifrequency observa-

tions of the blazar CTA 102 during 2013–2017. CTA 102

(also known as 4C +11.69) is an FSRQ at redshift

= 1

037

(Schmidt 1965). Flaring activity in the optical band has bee

observed from this source in 1978 (Pica et al. 1988), 1996

(Katajainen et al. 2000), and 2004 (Osterman Meyer 2009).

However, simultaneous

-ray observations were not available

for those events. The source was detected for the first time

rays by the

Compton Gamma Ray Observatory

in 1992

with both the EGRET (Hartman et al. 1999) and COMP-

TEL (Blom et al. 1995) instruments. Unfortunately, no op-

tical observations were available during the

-ray detection

(Villata et al. 1997). On the other hand, during the

Fermi

era a remarkable outburst was simultaneously observed in

2012 September–October in near-infrared (near-IR) and op-

tical bands by the Whole Earth Blazar Telescope

(WEBT)

and

rays by the Large Area Telescope (LAT) on board

http://www.oato.inaf.it/blazars/webt

Fermi Gamma-ray Space Telescope

. Correlated variability

in the two energy bands suggested a co-spatial origin of the

optical and

-ray emitting regions during the flaring activity

(Larionov et al. 2016).

In 2016 November, CTA 102 entered a new very-high-

activity state in

rays, as observed by

Fermi

-LAT, reaching

a daily flux higher than 1

−

ph cm

−

on 2016 De-

cember 16 (Ciprini et al. 2016). This flaring activity con-

tinued for a few weeks in

rays (e.g., Bulgarelli et al.

2016; Xu et al. 2016). A significant increase of activity was

observed over the entire electromagnetic spectrum (e.g.,

Calcidese et al. 2016; Ojha et al. 2016; Righini et al. 2016).

In particular, an extreme optical and near-IR outburst oc-

curred in 2016 December, with a brightness increase up to

six magnitudes with respect to the faint state of the source

(Raiteri et al. 2017). In Raiteri et al. (2017) we explained

the flux and spectral variations in optical, near-IR, and rad

bands by means of an inhomogeneous curved jet with differ-

ent jet regions changing their orientation, and hence their

Doppler factors, in time. Alternative theoretical scenari

have been proposed to explain the 2016–2017 flaring be-

haviour of CTA 102. According to Casadio et al. (2019) the

outburst was produced by a superluminal component cross-

ing a recollimation shock, while for Zacharias et al. (2017,

2019) it was due to ablation of a gas cloud penetrating the

relativistic jet in a leptonic or hadronic scenario.

The radio-to-optical and

-ray emission are produced

by two different mechanisms (i.e. synchrotron and IC emis-

sion in leptonic models), although related to the same rela-

tivistic electron population. Therefore, the

-ray variability

can be used as a further test to verify the geometrical model

that we proposed to explain the low-energy flux variability

in CTA 102 during 2013–2017. In the geometrical scenario,

the

-ray and optical radiation are produced in the same jet

region, therefore the

-ray and optical fluxes undergo the

same Doppler beaming and should be linearly correlated.

In this paper we present a multiwavelength analysis of

the CTA 102 emission from radio to

rays between 2013

January 1 and 2017 February 9, in particular during the

bright flaring activity occurred during 2016 November–2017

February. The radio-to-optical observations performed in

the framework of a campaign led by the WEBT, already pre-

sented in Raiteri et al. (2017), are complemented by the At-

acama Large millimeter/Submillimeter Array (ALMA) at 91

and 103 GHz, the Owens Valley Radio Observatory (OVRO)

data at 15 GHz, the Rapid Eye Mount (REM) near-IR data,

and a detailed analysis of data collected by the

Neil Gehrels

Swift Observatory

(optical–UV and X rays) and

Fermi

(

rays) satellites. The data set used in this paper is the riche

in terms of number of data points and broad-band coverage

presented in literature for the period considered here.

Sun constraints prevented us to have observations from

optical and near-infrared WEBT observatories and

Swift

satellite after 2017 February 9, not allowing us to inves-

tigate the connection between the

-ray flaring activity ob-

served in 2017 March–April (see e.g., Shukla et al. 2018) and

the emission from near-IR to X-rays. After that period the

infrared-to-X-ray coverage is insufficient to adequately te

the geometrical model and to investigate the connection be-

tween low-energy and

-ray emission.

The paper is organized as follows. In Section 2 and 3

we present

Fermi

-LAT and

Swift

data analysis and results,

2019 RAS, MNRAS

000

, 1–24

MWL behaviour of CTA 102 during 2013–2017

Figure 1.

Integrated flux light curve of CTA 102 (upper panel),

spectral slope (middle panel), curvature parameter (botto

panel) obtained in the 0.1–300 GeV energy range during 2013

January 1–2017 February 9 (MJD 56293–57793) with 30-d time

bins. The open symbols refer to results obtained with

fixed to

0.07 (see the text for details).

respectively, whereas in Section 4 we report on the radio-

to-optical observations. Multifrequency flux and spectral

variability are discussed in Sections 5 and 6, respectively

The application of the geometrical model by Raiteri et al.

(2017) to the

-ray, optical, and radio variability is discussed

in Section 7. We discuss the previous results and draw

our conclusions in Section 8. Throughout this paper, we

assume the following cosmology:

= 71 km s

−

Mpc

−

Ω

= 0

27, and Ω

= 0

73 in a flat Universe (Ade et al.

2016).

FERMI

-LAT DATA: ANALYSIS AND

RESULTS

The

Fermi

-LAT is a pair-conversion telescope operating

from 20 MeV to

300 GeV. Further details about the

Fermi

-LAT are given in Atwood et al. (2009).

The LAT data used in this paper were collected from 2013

January 1 (MJD 56293) to 2017 February 9 (MJD 57793).

During this time, the LAT instrument operated almost

entirely in survey mode. The Pass 8 data (Atwood et al.

2013), based on a complete and improved revision of the

entire LAT event-level analysis, were used. The analysis

was performed with the

ScienceTools

software package ver-

sion v11r5p3. Only events belonging to the ‘Source’ class

(

evclass=128

evtype=3

) were used. We selected only events

within a maximum zenith angle of 90 deg to reduce con-

tamination from the Earth limb

rays, which are produced

by cosmic rays interacting with the upper atmosphere. The

spectral analysis was performed with the instrument re-

sponse functions

P8R2

SOURCE

V6

using a binned maximum-

likelihood method implemented in the Science tool

gtlike

Isotropic (‘iso

source

v06.txt’) and Galactic diffuse emission

(‘gll

iem

v06.fit’) components were used to model the back-

ground (Acero et al. 2016)

. The normalization of both com-

ponents was allowed to vary freely during the spectral fit-

ting.

We analysed a region of interest of 20

◦

radius centred

at the location of CTA 102. We evaluated the significance of

the

-ray signal from the source by means of a maximum-

likelihood test statistic (TS) defined as TS = 2

(log

log

), where

is the likelihood of the data given the model

with (

) or without (

) a point source at the position of

CTA 102 (e.g., Mattox et al. 1996). The source model used

gtlike

includes all the point sources from the 3FGL cata-

logue that fall within 30

◦

of CTA 102. We also included new

candidates within 10

◦

of CTA 102 from a preliminary eight-

year point source list (FL8Y

). The spectra of these sources

were parametrized by a power law (PL), a log-parabola

(LP), or a super exponential cut-off, as in the catalogues.

A first maximum-likelihood analysis was performed over

the whole period to remove from the model the sources hav-

ing TS

25. A second maximum-likelihood analysis was

performed on the updated source model. In the fitting proce-

dure, the normalization factors and the spectral parameter

of the sources lying within 10

◦

of CTA 102 were left as free

parameters. For the sources located between 10

◦

and 30

◦

from our target, we kept the normalization and the spec-

tral shape parameters fixed to the values from the 3FGL

catalogue.

Integrating over 2013 January 1–2017 February 9 the

fit with an LP model,

dN/dE

∝

E/E

−

log(

E/E

)

, as in

the 3FGL and FL8Y catalogues, results in TS = 125005 in

the 0.1–300 GeV energy range, with an integrated average

flux of (93.8

0.6)

−

ph cm

−

, a spectral slope

= 2.16

0.01 at the reference energy

= 308 MeV,

and a curvature parameter around the peak

= 0.07

0.01. The corresponding apparent isotropic

-ray luminosity

is (5.1

0.1)

erg s

−

. As a comparison in the 3FGL

catalogue, covering the period 2008 August 4–2012 July 31,

the integrated average flux is (16.1

0.5)

−

ph cm

−

, and the spectrum is described by an LP with a spectral

slope

= 2.34

0.03 at the reference energy

= 308

MeV, and a curvature parameter around the peak

= 0.13

0.02. This indicates a moderate change of the average

ray spectrum during the period studied here, in which the

flux is a factor of approximately six higher than the first

four years of LAT operation.

Fig. 1 shows the

-ray flux (top panel) and spectral pa-

rameters (middle panel: spectral slope; bottom panel: cur-

vature parameter) evolution of CTA 102 for the period 2013

January 1–2017 February 9 using an LP model and 30-d

time bins. For each time bin, the spectral parameters of both

CTA 102 and all sources within 10

◦

from it were left free to

vary. For the time bins in which the fit results in a TS

300 for CTA 102, the statistics is not enough for obtaining

a detailed characterization of the spectrum with complex

http://fermi.gsfc.nasa.gov/ssc/data/access/lat/

BackgroundModels.html

https://fermi.gsfc.nasa.gov/ssc/data/access/lat/fl8y

2019 RAS, MNRAS

000

, 1–24

D’Ammando, Raiteri, Villata, et al.

Figure 2.

Upper panel: Integrated flux light curve of CTA 102

obtained in the 0.1–300 GeV energy range during 2013 January

1–2017 February 9 with five-day time bins. The arrow refers to

upper limit on the source flux. Upper limits are computed when

10. Different outbursts are labelled with an identification

number in the plot. Bottom panel: X-ray light curve in the 0.3

–

10 keV energy range obtained by

Swift

-XRT (see Section 3 for

details).

spectral models, therefore we run again the likelihood anal

ysis using an LP model with the curvature parameter fixed

to the value obtained integrating over the entire period (i.

= 0.07). The

-ray spectrum of CTA 102 shows a remark-

able variability on monthly time-scale, with a spectral slo

between 1.88 and 2.97 (the average spectral slope is

2.30

0.09), and a curvature parameter between 0.04 and

0.26 (the average curvature parameter is

= 0.13

0.04),

although for the latter the uncertainties are relatively la

rge.

For investigating the

-ray variability on different time-

scales, we have produced a

-ray light curve for the entire

period with five-day time bins (Fig. 2). For each time bin,

the spectral parameters of CTA 102 and all sources within

◦

of it were frozen to the values resulting from the likeli-

hood analysis over the respective monthly time bin. When

10, 2

upper limits were calculated. Six peaks corre-

sponding to periods with fluxes higher than 2

−

ph cm

−

were observed in 2013 April 4–8 (MJD 56386–56390; I),

2014 October 21–25 (MJD 56951–56955; II), 2015 December

26–30 (MJD 57382–57386; III), 2016 February 19–23 (MJD

57437–57441; IV), 2016 August 22–26 (MJD 57622–57626;

V), and 2016 December 30–2017 January 3 (MJD 57752–

57756; VI), with an increase of the flux of a factor between

2.5 and 14 with respect to the average flux estimated during

2013–2017.

Finally, we have produced a

-ray light curve with one-

day and 12-h time bins for the period of high activity, i.e.

2016 November 11–2017 February 9 (MJD 57703–57793), as

shown in Fig. 3. In the analysis of the sub-daily light curves

we fixed the flux of the diffuse emission components at the

Figure 3.

Integrated flux light curve of CTA 102 obtained by

Fermi

-LAT in the 0.1–300 GeV energy range during 2016 Novem-

ber 11–2017 February 9, with one-day time bins (top panel), a

12-h time bins (bottom panel).

value obtained by fitting the data over the entire period

analysed in this paper. For each time bin, the spectral pa-

rameters of both CTA 102 and all sources within 10

◦

of it

were frozen to the values resulting from the likelihood anal

sis in the monthly time bins. The peak flux of the daily light

curve dates 2016 December 28 (MJD 57750), with a flux of

(2158

63)

−

ph cm

−

, corresponding to a

-ray

luminosity (2.2

0.1)

erg s

−

. A similar peak flux was

observed on 12-h time-scales, (2200

111)

−

ph cm

−

in the second bin of 2016 December 28, corresponding

to a

-ray luminosity (2.2

0.1)

erg s

−

. These val-

ues are among the highest

-ray luminosities ever measured

for blazars, comparable to what was observed for 3C 454.3

(Abdo et al. 2011) and S5 0836+710 (Orienti et al. 2019).

As a comparison, in 2012 the

-ray flux of CTA 102 reached

a peak flux of

∼

−

ph cm

−

(Larionov et al. 2016).

The search for variability on very short time-scale in

rays is beyond the scope of this paper. Rapid variability on

time-scale of minutes was observed in 2016 December, with

a peak flux of

∼

3.5

−

ph cm

−

(Gasparyan et al.

2018; Shukla et al. 2018; Meyer et al. 2019).

NEIL GEHRELS SWIFT OBSERVATORY

DATA: ANALYSIS AND RESULTS

The

Neil Gehrels Swift Observatory

satellite (Gehrels et al.

2004) carried out 73 observations of CTA 102 between 2013

March 24 (MJD 56436) and 2017 January 18 (MJD 57771).

The observations were performed with all three instruments

on board: the X-ray Telescope (XRT; Burrows et al. 2005,

0.2–10.0 keV), the Ultraviolet/Optical Telescope (UVOT;

2019 RAS, MNRAS

000

, 1–24

MWL behaviour of CTA 102 during 2013–2017

Roming et al. 2005, 170–600 nm), and the Burst Alert Tele-

scope (BAT; Barthelmy et al. 2005, 15–150 keV).

The hard X-ray flux of this source turned out to be

below the sensitivity of the BAT instrument for such short

exposures and therefore the data from this instrument will

not be used. Moreover, the source is not present in the

Swift

BAT 105-month hard X-ray catalogue (Oh et al. 2018).

The XRT data were processed with standard procedures

(

xrtpipeline v0.13.3

), filtering, and screening criteria by

using the

HEAsoft

package (v6.22). The data were collected

in photon counting mode in all the observations. The source

position in detector coordinates was optimized for each ob-

servation by means of

XIMAGE

. The source extraction region

is centred on these coordinates. The source count rate in

some observations is higher than 0.5 count s

−

: these ob-

servations were checked for pile-up and a correction was

applied following standard procedures (e.g., Moretti et al

2005). To correct for pile-up we excluded from the source

extraction region the inner circle of three-pixel radius by

considering an annular region with outer radius of 30 pixels

(1 pixel

∼

2.36 arcsec). For the other observations source

events were extracted from a circular region with a radius of

20 pixels. Background events were extracted from a circular

region with radius of 50 pixels far away from bright sources.

Ancillary response files were generated with

xrtmkarf

, and

account for different extraction regions, vignetting and po

int

spread function corrections. We used the spectral redistri

bution matrices v014 in the calibration data base main-

tained by

HEASARC

. We fitted the spectrum with an ab-

sorbed PL using the photoelectric absorption model

tbabs

(Wilms et al. 2000), with a neutral hydrogen column den-

sity fixed to its Galactic value in the source direction (

= 2.83

−

; Kalberla et al. 2005). The results of the

fit are reported in Table A1 and the 0.3–10 keV fluxes are

shown in Fig. 2 in comparison to the

-ray light curve ob-

tained by

Fermi

-LAT.

The X-ray flux (0.3–10 keV) varied between 7.6

−

and 68.1

−

erg cm

−

and the photon index between

1.14 and 1.85, with an average value of

= 1.40

0.15.

In Fig. 4 we plotted the XRT photon index as a function of

flux in the 0.3–10 keV energy range. A harder-when-brighter

spectral trend has been observed in X-rays in several blazar

(e.g., Krawczynski et al. 2004; D’Ammando et al. 2011;

Raiteri et al. 2012; Hayashida et al. 2015; Aleksic et al.

2015), although not always present also in the same

source (e.g. Hayashida et al. 2012; Aleksic et al. 2017;

Carnerero et al. 2017). This behaviour is usually related to

the competition between acceleration and cooling processe

acting on relativistic electrons. No spectral hardening wi

increasing flux is observed either in the entire period or in

the high-activity period alone for CTA 102. This suggests

that a change in the electron energy distribution is not the

main driver of the long-term variability in this energy band

However, at the peak of the X-ray activity the photon index

is harder than the average value observed over 2013–2017.

During the

Swift

pointings, the UVOT instrument ob-

served CTA 102 in all its optical (

, and

) and UV

(

2, and

2) photometric bands (Poole et al. 2008;

Breeveld et al. 2010). We analysed the data using the

https://heasarc.gsfc.nasa.gov/docs/heasarc/caldb/

Figure 4.

Swift

/XRT photon index as a function of the 0.3–10

keV unabsorbed flux. The red points highlight the high-activ

ity

period.

uvotsource

task included in the

HEAsoft

package (v6.22).

Source counts were extracted from a circular region of five-

arcsec radius centred on the source, while background count

were derived from a circular region of 20 arcsec radius in

a nearby source-free region. Observed magnitudes are re-

ported in Table A2. An increase of 4.5–5.5 mag with respect

to the faint state of the source was observed in the UVOT

bands, with a range of values:

= 11

44–16.86,

= 12

56–

17.24,

= 11

78–16.27,

1 = 11

33–16.18,

2 = 11

36–

16.17,

2 = 11

52–16.46.

Following Raiteri et al. (2010, 2011), to obtain de-absorbe

flux densities we used the count rate to flux density conver-

sion factors CF and amount of Galactic extinction

for

each UVOT band that have been obtained by folding the

quantities of interest with the source spectrum and effectiv

areas of UVOT filters and are reported in Larionov et al.

(2016).

Besides correcting flux densities for Galactic extinction,

we also subtracted the thermal emission contribution due

to the accretion disc and BLR according to the model by

Raiteri et al. (2014).

4 OPTICAL-TO-RADIO OBSERVATIONS

CTA 102 has been monitored by the GLAST-AGILE Sup-

port Program (GASP) of the WEBT in the optical, near-

infrared and radio bands since 2008. Optical-to-radio GASP

WEBT data collected in 2013–2017 have been presented in

Raiteri et al. (2017). That data set is here complemented

with data at 15 GHz by the OVRO telescope, at 91 and 103

GHz by ALMA, in the near-IR by the REM telescope, in the

optical by the

Swift

satellite. The optical photometric ob-

servations used in this paper were acquired at the following

observatories: Abastumani (Georgia), AstroCamp (Spain),

2019 RAS, MNRAS

000

, 1–24

D’Ammando, Raiteri, Villata, et al.

Belogradchik (Bulgaria), Calar Alto (Spain), Campo Imper-

atore (Italy), Crimean (Russia), Kitt Peak (USA), Lowell

(USA; 70 cm, DCT and Perkins telescopes), Lulin (Tai-

wan), Michael Adrian (Germany), Mt. Maidanak (Uzbek-

istan), New Mexico Skies (USA), Osaka Kyoiku (Japan),

Polakis (USA), Roque de los Muchachos (Spain; Liverpool,

NOT and TNG telescopes), ROVOR (USA), Rozhen (Bul-

garia; 200 and 50/70 cm telescopes), San Pedro Martir

(Mexico), Sirio (Italy), Skinakas (Greece), Steward (USA;

Kuiper, Bok, and Super-LOTIS), St. Petersburg (Russia),

Teide (Spain), Tien Shan (Kazakhstan), Tijarafe (Spain),

Tucson (USA), Valle d’Aosta (Italy), and Vidojevica (Ser-

bia).

Near-IR data were collected within the WEBT project

in the

, and

bands at the Campo Imperatore,

Lowell (Perkins) and Teide observatories. These data are

here complemented with observations performed by REM

(Zerbi et al. 2001; Covino et al. 2004), a robotic telescope

located at the ESO Cerro La Silla observatory (Chile), in

the period 2016 November 15–December 11. All raw near-IR

frames obtained with the REM telescope were reduced fol-

lowing standard procedures. Instrumental magnitudes were

obtained via aperture photometry and absolute calibration

has been performed by means of secondary standard stars in

the field reported by 2MASS

. The data presented here were

obtained as Target of Opportunity observations triggered b

the

-ray flaring activity of the source (PI: F. D’Ammando).

Observed magnitudes are reported in in Table A3.

Optical and near-IR flux densities were dereddened fol-

lowing the prescriptions of the NASA/IPAC Extragalactic

Database

(NED) and corrected for the thermal emission

contribution according to the model of Raiteri et al. (2017)

Radio and mm observations were done at the Mets ̈ahovi

Radio Observatory (Finland) at 37 GHz, and at the 30-

m IRAM telescope (Spain) and the Sub-millimeter Array

(Hawaii, USA) at 230 GHz. We also added the ALMA data

collected at 91 and 103 GHz (Band 3) during 2013–2017 and

included in the ALMA calibrator source catalogue

As part of an ongoing blazar monitoring program, the

OVRO 40-Meter Telescope has observed CTA 102 at 15 GHz

regularly since the beginning of 2009. The OVRO 40-Meter

Telescope uses off-axis dual-beam optics and a cryogenic re-

ceiver with 2 GHz equivalent noise bandwidth centred at

15 GHz. Atmospheric and ground contributions as well as

gain fluctuations are removed with the double switching

technique (Readhead et al. 1989) where the observations are

conducted in an ON–ON fashion so that one of the beams

is always pointed on the source. Until 2014 May the two

beams were rapidly alternated using a Dicke switch. Since

2014 May, when a new pseudo-correlation receiver replaced

the old receiver, a 180 deg phase switch is used. Relative cal

ibration is obtained with a temperature-stable noise diode

compensate for gain drifts. The primary flux density calibra

tor is 3C 286, with an assumed value of 3.44 Jy (Baars et al.

1977); DR21 is used as secondary calibrator source. Details

of the observation and data reduction schemes are given in

Richards et al. (2011). Observations acquired at 15 GHz by

http://www.ipac.caltech.edu/2mass/

http://ned.ipac.caltech.edu/

https://almascience.eso.org/alma-data/calibrator-ca

talogue

OVRO were used in this paper together with those obtained

within the WEBT project. The radio flux at 15 GHz varied

between 2.85 and 3.88 Jy during 2013–2017.

Radio-to-optical light curves collected by WEBT, REM,

ALMA, and OVRO will be compared to the

-ray light curve

obtained by

Fermi

-LAT in the Section 5.

5 MULTIFREQUENCY FLUX VARIABILITY

Variability studies of radio-to-

-ray emission from blazars

can provide important insights into the physics of the jet

and the mechanisms at work in these sources. Flares can be

explained e.g. by a shock propagating downstream the jet

and/or by variations of the Doppler factor, which depends

on the bulk Lorentz factor of the relativistic plasma in the j

and on the viewing angle. During flares, blazars usually show

greater variability amplitudes in the high-energy part of

the spectrum than in the low-energy one (e.g., Wehrle et al.

1998; Raiteri et al. 2012). However, some sources increased

their brightness by hundreds of times also in infrared and

optical bands. In this context, the 2016–2017 outburst of

CTA 102 presented here is one of the best cases to study.

We quantify the observed variability of the CTA 102 jet

emission in the different energy bands through the variabil-

ity amplitude, calculated as the ratio of maximum to mini-

mum flux. Values in the

-ray band are based on the light

curve with five-day time bins. Near-infrared-to-UV fluxes

are corrected for Galactic extinction. Moreover, the contr

bution of the thermal emission from the disc, BLR, and torus

is removed to properly consider only the jet contribution to

the flux. In Table 1 and Fig. 5, we report the variability

amplitude estimated over the period 2013 January 1–2017

February 9 in the different bands. The variability amplitude

may depend on the sampling of the light curves at the differ-

ent frequencies. In the case of CTA 102, observations were

available in all energy bands at the peak of the flare, making

the values obtained representative of the increase of activ

ity

of the source. The variability amplitude shows a rising tren

with increasing frequency in the radio-to-optical range an

it declines in the UV. The X-ray band has a small variability

amplitude, in particular if compared to the

-ray one. This

can be related to the lower energies of the electrons pro-

ducing X-rays with respect to those producing

rays. The

similar variability amplitude at 230 GHz and in the X rays

may be a hint that they are produced by the same electron

population in the same jet region through the synchrotron

and IC emission mechanism, respectively, as found e.g. for

BL Lacertae (Raiteri et al. 2013).

Simultaneous flux variations at low and high energies

indicate that their emission comes from the same region

of the jet and that the same electrons produce both the

synchrotron and IC fluxes (e.g., Fossati et al. 2008). How-

ever, flaring events in the optical band with no counter-

part at high energies were observed in some blazars (e.g.,

Chatterjee et al. 2013; D’Ammando et al. 2013). Strong

variability characterizes the emission over the entire ele

ctro-

magnetic spectrum of CTA 102 in 2016–2017, in particular

during the period 2016 November 11–2017 February 8 (MJD

57003–57792), making this an ideal target to investigate th

connection of the flux behaviour observed in

rays with the

flux behaviour at lower energies.

2019 RAS, MNRAS

000

, 1–24

MWL behaviour of CTA 102 during 2013–2017

Table 1.

Variability amplitude estimated over the period 2013 Janua

ry 1–2017 February 9 in the different energy bands. Values are

corrected for Galactic extinction and the thermal emission

contribution. For the minimum and maximum flux density the MJ

D at which

the value is collected is reported in parenthesis.

Band

Minimum Flux density

Maximum Flux density

Variability

amplitude

(erg cm

−

)

(erg cm

−

)

rays 5.91

−

(MJD 57177–57181) 3.02

−

(MJD 57752–57756)

5110

X-ray

7.60

−

(MJD 57386)

6.81

−

(MJD 57759)

1.95

−

(MJD 56958)

4.92

−

(MJD 57751)

253

3.41

−

(MJD 56958)

6.90

−

(MJD 57752)

202

2.55

−

(MJD 56958)

5.95

−

(MJD 57751)

233

2.50

−

(MJD 56958)

3.37

−

(MJD 57752)

135

2.98

−

(MJD 56893)

7.20

−

(MJD 57750)

2416

1.97

−

(MJD 56874)

7.71

−

(MJD 57750)

3920

2.20

−

(MJD 56785)

7.76

−

(MJD 57750)

3523

3.62

−

(MJD 56785)

8.01

−

(MJD 57750)

2210

6.18

−

(MJD 56879)

6.96

−

(MJD 57752)

1126

8.59

−

(MJD 56879)

7.10

−

(MJD 57752)

827

1.09

−

(MJD 56879)

5.73

−

(MJD 57751)

526

230 GHz

2.25

−

(MJD 56832)

1.62

−

(MJD 57754)

103 GHz

1.63

−

(MJD 56837)

6.93

−

(MJD 57649)

4.3

91 GHz

1.51

−

(MJD 56837)

5.84

−

(MJD 57649)

3.9

37 GHz

7.92

−

(MJD 56839)

1.89

−

(MJD 57671)

15 GHz

4.28

−

(MJD 56651)

5.99

−

(MJD 57708)

1.5

Figure 5.

Variability amplitude vs frequency in the different en-

ergy bands estimated over the period 2013 January 1-2017 Feb

ru-

ary 9. Values are corrected for Galactic extinction and the t

hermal

emission contribution.

In Fig. 6, we compare the

-ray light curve obtained by

Fermi

-LAT with 12-h time bins during the highest activity

period, 2016 November 11–2017 February 8 (MJD 57003–

57792, corresponding to the outburst VI in Fig. 2), to the

infrared-to-X-ray light curves. All fluxes are normalized t

the maximum value observed in the considered period in

order to compare when and how much the flux increased

in the different energy bands. In the

-ray light curve we

can see four major outbursts peaked on 2016 December 15

(MJD 57737; F1), 2016 December 22 (MJD 57744; F2), 2016

December 27 (MJD 57749; F3), and 2017 January 4 (MJD

57757; F4). The third outburst appears more prominent and

shows a larger increase with respect to the others. These

four outbursts have corresponding events in optical, with

the peaks observed at the same time. In

, and

bands

the sampling is sparse and only two of the four peaks are

observed. A high X-ray flux has been observed in the period

that covers the

-ray flares F1 and F2, and two X-ray peaks

are evident at the time of the

-ray flares F3 and F4.

If we compare the

-ray and optical (

band, the best

sampled band) fluxes normalized to the respective lowest

values observed during 2016 November 11–2017 February 9

(Fig. 7), it is evident that the four main flares occurred at

the same time. A similar amplitude has been observed in the

two bands, except for the first flare. In particular, the same

variability amplitude has been observed during the main

peak. On the other hand, not all the events observed in the

optical band have a counterpart in

rays and vice versa.

In addition to the “sterile” optical flare

occurred around

2016 December 1 (MJD 57723), no significant optical activ-

ity corresponds to the increase of

-ray activity peaked on

2017 January 24 (MJD 57777) and 2017 February 6 (MJD

57790) (“orphan” flares). This indicates that the interpre-

tation of the source variability must be more complicated

than the fair overall optical-

correlation would suggest.

Cross-correlation analysis between flux variations in

different bands can allow us to determine whether the

emissions come from the same region of the jet or

not. We used the discrete correlation function (DCF;

Edelson & Krolik 1988; Hufnagel & Bregman 1992) to ana-

With “sterile flare” and “orphan flare” we mean a flare observed

in the optical band with no counterpart at

-ray frequencies, and

a flare that is observed at the high energies only, respective

ly.

2019 RAS, MNRAS

000

, 1–24

D’Ammando, Raiteri, Villata, et al.

Figure 6.

Multifrequency light curve normalized to the maximum value

observed for the period 2016 November 11–2017 February 8

(MJD 57003–57792) in the following energy bands (from top to

bottom):

rays (100 MeV–300 GeV), X-rays (0.3–10 keV),

, and

. Filled triangles: WEBT data; open triangles: UVOT data; op

en squares: REM data. Main

-ray outbursts are labelled as

F1, F2, F3, and F4 in the top panel.

lyze cross-correlations. Correlation produces positive p

eaks

in the DCF and is strong if the peak value approaches or

even exceeds one. The DCF between the

-ray and the opti-

cal (

-band) light curves over the whole 2013–2017 period is

displayed in Fig. 8, showing a main peak compatible with no

time lag. When comparing the

-ray light curve with 12-h

time bin with the optical light curve with one-hour binning

in the high-activity period of Fig. 7, the DCF shows again a

main peak compatible with no time lag, with DCF

= 0.94

(Fig. 9). This indicates strong correlation between

-ray and

optical emission, with no evidence of delay between the flux

variations in the two bands, in agreement with the results

presented in Larionov et al. (2017). We determine the uncer-

tainty in this result by performing 1000 Monte Carlo sim-

ulations according to the “flux redistribution/random sub-

set selection” technique (Peterson et al. 1998; Raiteri et a

2003), which tests the importance of sampling and data er-

rors. Among the 1000 simulations, we obtain that 78.3 per

cent of simulations (

) give a time lag between 0 and 0.6

d, as shown in the inset of Fig. 9. This is compatible with

no delay between optical and

-ray emission. The secondary

DCF peaks in Fig. 9 are due to the multi-peaked structure

of the outburst.

Although the UV data collected by

Swift

-UVOT are

sparser than the optical ones, we can recognize a similar

behaviour in the

-ray and UV light curves with similar

increase of flux during the flaring period when normalized

to the maximum value (Fig. 10), even though the variability

amplitude is smaller in UV with respect to the

-ray band.

The sparse X-ray data in 2013–2015 indicate a low flux,

in keeping with the relatively low activity observed also in

rays (see Fig. 2). The comparison between the

-ray and

the X-ray light curves during the high-activity period show

a general agreement, with the

-ray peaks corresponding to

high X-ray fluxes (Fig. 11). Cross-correlation of the

-ray

(12-h time bins) and X-ray light curves in the high-activity

period shows a good correlation with a time lag compati-

ble to zero within the DCF bin size of six days (Fig. 12).

The different sampling of the light curves does not allow

a more detailed comparison. However, the X-ray variability

amplitude appears much smaller than at

rays.

The correlation between the radio-mm fluxes on one

side and the optical and

-ray fluxes on the other side is

rather puzzling. In general, they present a different behav-

ior, but sometimes they share common features. As one can

see in Fig. 13, the mm data at 230 GHz show a steady

flux increase starting from the end of 2015 and culminat-

ing with a prominent outburst peaking on 2017 January 1

(MJD 57754). A comparison with the

-ray light curve re-

veals that the end of 2015 also marks the beginning of the

activity in this band, with the major peak observed at the

same time of the 230 GHz one.

A steady flux increase starting from the end of 2015 is ob-

served also at 91–103 GHz, is marginally detectable at 37

2019 RAS, MNRAS

000

, 1–24

MWL behaviour of CTA 102 during 2013–2017

Figure 7.

Comparison of the

Fermi

-LAT

-ray light curve with

12-h time bins (top panel) and

-band light curve (bottom

panel) normalized to the lowest value observed in the period

2016

November 11–2017 February 8. Main

-ray outbursts and “or-

phan” flares are labelled as F1, F2, F3, F4, and orphan in the

top panel. The “sterile” flare is labelled as sterile in the bo

ttom

panel.

Figure 8.

DCF between the

-ray fluxes obtained with five-day

time bins and the

-band flux densities with 1 day binning over

the whole 2013–2017 period.

GHz, and vanishes at 15 GHz. Enhanced activity is present

at the beginning of 2014 at 230, 91–103, and 37 GHz, when

the light curves at both lower and higher frequencies appear

rather flat.

The light curves at 91–103, 37, and 15 GHz present

peaks in 2016 October–November. The increase of delay

of the radio peaks going to lower frequencies is in agree-

ment with synchrotron self-absorption opacity effects. The

decrease of the flux variation amplitudes towards lower fre-

quencies is expected, if the radio emission at lower frequen

Figure 9.

DCF between the

-ray fluxes obtained with 12-h time

bins and the

-band flux densities with one-hour binning during

the 2016–2017 flaring period. The inset shows the result of cr

oss-

correlating 1000 Monte Carlo realizations of the two data se

according to the “flux redistribution/random subset select

ion”

technique.

Figure 10.

Multifrequency light curve normalized to the max-

imum value observed for the period 2016 November 11–2017

February 8 (MJD 57003–57792) in the following energy bands

(from top to bottom):

rays (100 MeV–300 GeV),

2, and

1 bands.

cies comes from more external and extended regions of the

jet, in average less aligned to our line of sight. In and aroun

the same period, the sampling at 230 GHz is poor. The clos-

est events at

-ray and optical frequencies date 2016 August

(flare V in Fig. 2), and, before, there is a stronger

-ray flare

in 2016 March (flare IV) during a seasonal gap in the optical

light curve.

The DCF between the

-ray and 15 GHz fluxes (see Fig.

14) shows a small peak at a negative time lag of about two

2019 RAS, MNRAS

000

, 1–24

D’Ammando, Raiteri, Villata, et al.

Figure 11.

Fermi

-LAT

-ray light curve with 12-h time bins in

the 0.1–300 GeV energy range (top panel) and X-ray light curv

in the 0.3–10 keV energy range (bottom panel) in the period 20

November 11–2017 February 9.

Figure 12.

DCF between the

-ray fluxes obtained with 12-h

time bins and the X-ray flux.

months, indicating radio variations preceding the

-ray ones.

This signal comes from the cross-match of the 15 GHz flare

peaking at the beginning of 2016 November with the

-ray

acme two months later. A stronger DCF maximum occurs at

time lag of about 250 d, indicating radio variations follow-

ing the

-ray ones by about eight to nine months. This sug-

gests that linking the 2016 October–November radio flare

observed from 103 to 15 GHz with the 2016 March

-ray

flare is more likely. Therefore, a delayed radio outburst at

15 GHz can be expected also some months after the 2017

January main flare observed in the

-ray, optical, and 230

GHz band.

However, extending the OVRO light curve at 15 GHz up

to 2018 July (see Fig. A1), a radio outburst is visible only in

Figure 13.

Multifrequency light curve for the period 2013 Jan-

uary 1–2017 February 9 in the following energy bands (from to

to bottom):

rays (100 MeV–300 GeV; 5-d time bins, in units

of 10

−

ph cm

−

;

Fermi

-LAT data), 230 GHz (in units of

Jy; triangles: SMA data, squares: IRAM data), 91 and 103 GHz

(in units of Jy; ALMA data), 37 GHz (in units of Jy; Mets ̈ahovi

data), 15 GHz (in units of Jy; OVRO data).

2018 February, more than one year after the 2017

-ray, opti-

cal and mm peaks. Its peak flux is only a few percent greater

than the maximum flux reached in 2016 October–November.

On the basis of the delay between

-ray emission and ra-

dio emission at 15 GHz found here and usually observed in

blazar objects (e.g., Pushkarev et al. 2010; Fuhrmann et al.

2014), it is unlikely that the 2018 radio outburst at 15 GHz

is related to the main flaring activity observed at the begin-

ning of 2017 in

-ray, optical, and at 230 GHz. Therefore,

we conclude that the

-ray flaring activity in 2017 has no

(delayed) counterpart in the 15 GHz light curve.

In the framework of the geometrical model, the differ-

ence in behaviour between radio and higher energy emission

would be ascribed to different viewing angle (with conse-

quent different Doppler boosting of the emission) of the jet

regions producing their emission. The extent of misalign-

ment between the emitting jet regions can be inferred from

the corresponding light curves, as will be shown in Sect. 7.

6 MULTIFREQUENCY SPECTRAL

VARIABILITY

Figure 15 shows the broad-band SED of CTA 102 in three

brightness states with near-contemporaneous data in the op

tical, UV, X-ray, and

-ray bands. The optical-to-radio data

set presented in Raiteri et al. (2017) has been complemented

with data at 15 GHz by the OVRO telescope, at 91 and 103

GHz by ALMA, and in the near-IR by the REM telescope, in

the optical-to-X-rays by the

Swift

satellite, and at

rays by

the

Fermi

satellite. In addition to SMA, Mets ̈ahovi, ALMA,

2019 RAS, MNRAS

000

, 1–24

MWL behaviour of CTA 102 during 2013–2017

Figure 14.

DCF between the

-ray fluxes obtained with five-day

time bins and the 15 GHz flux densities.

and OVRO data, radio observations collected by RATAN-

600 radio telescope (Mingaliev et al. 2001) from 1.0 to 21.7

GHz have been considered. Because of the longer variability

time-scales of the radio light curves, radio data are includ

also if taken within a few days from the reference epoch.

The highest state corresponds to MJD = 57761 (2017

January 8). Optical data were acquired at the Tijarafe Ob-

servatory and fairly overlap with the UVOT data. The in-

termediate state dates MJD = 57635 (2016 September 4),

and comes from the declining phase of the small flare pre-

ceding the big outburst (flare V in Fig. 2). The optical data

are from the Mt. Maidanak Observatory and are in satis-

factorily agreement with the UVOT data. For the low state,

we chose MJD = 57364 (2015 December 8), about one year

before the culmination of the big outburst. Optical data are

from the St. Petersburg Observatory; the

-band point ap-

pears a bit fainter than the corresponding UVOT value, but

they agree within errors. For the LAT data the spectra are

extracted on 5-day time-scale as done in Fig. 2, except for

the high state in which the daily LAT spectrum is included.

The intermediate- and the low-state SED become very

close in the UV, where the emission contribution of the big

blue bump peaks in coincidence of the Ly

αλ

1216 broad emis-

sion lines. The X-ray spectral shape in the intermediate sta

is softer than expected, but it is affected by large errors. Fi

nally, the peak of the IC component shifts at a much higher

energy in the high state, confirming the result found for the

synchrotron component by Raiteri et al. (2017).

7 GEOMETRICAL MODEL APPLIED TO

-RAY, OPTICAL, AND RADIO DATA

In Raiteri et al. (2017) we interpreted the long-term vari-

ability of the CTA 102 synchrotron flux in terms of variation

of the Doppler factor because of changes of the viewing an-

gle of the jet-emitting regions. The intrinsic flux is assume

to be constant on time-scales of months or longer in the rest

frame, while flast flares can be due to intrinsic, energetic

processes. From the observed multiwavelength light curves

we derived how the jet moves, i.e. how the regions emit-

ting at different frequencies align with respect to the line

Figure 15.

Broad-band SED of CTA 102 in three different bright-

ness states labelled with their MJD. In the optical-UV bands

ground-based data are shown as open circles, while UVOT data

are displayed with squares. The X-ray spectra are represent

ed by

PL fits, while the

Fermi

-LAT spectra are plotted according to LP

models; in both cases we took the uncertainties on the parame

ters

into account.

of sight. Support to this twisting jet scenario comes from

both observations and theory. Examples of helical jet struc

tures and wobbling motion have been observed with high

angular resolution images in the radio band in both extra-

galactic and Galactic sources (see e.g., Agudo et al. 2007,

2012; Perucho et al. 2012; Fromm et al. 2013; Britzen et al.

2017, 2018; Miller-Jones et al. 2019).

In numerical magnetohydrodynamics (MHD) simulations

of relativistic jets in 3D, instabilities can develop which

distort the jet itself and produce wiggled structures

(Nakamura et al. 2001; Mignone et al. 2010). Moreover, or-

bital motion in a binary black hole system or a warped ac-

cretion disc can lead to jet precession, which modifies the

jet orientation with respect to the line of sight (Liska et al

2018).

Raiteri et al. (2017) performed their analysis on radio–

optical data, concentrating on the light curves at 37 and 230

GHz, and in the

band. In the following we investigate the

outcome of the proposed geometrical model when applied

to both higher and lower frequencies. We examine the

-ray

and 15 GHz flux variability; the X-ray data are too sparse

for a meaningful analysis. We consider the

-ray light curve

from 2013 January 1 (MJD 56293), with a time bin of five

days before 2016 November 11 (MJD 57703) and 12 h after.

The

-ray light curve is compared to the optical (

band)

and radio (15 GHz) light curves in Fig. 16.

As in Raiteri et al. (2017), we modelled the optical long-

term trend with a cubic spline interpolation through the

data binned with a variable time interval, which shortens as

the flux rises: ∆

= ∆

when

F > F

min

, where the

exponent 2 applies to a continuous jet and

= 1

7 is the

spectral index in the

band. We used the same binning to

obtain the spline in the

-ray band, while for 15 GHz case

we adopted a 30-d bin because of the smoother behaviour,

as done by Raiteri et al. (2017) for the 37 and 230 GHz light

curves.

2019 RAS, MNRAS

000

, 1–24