MNRAS 519, 5263–5270 (2023) https://doi.org/10.

1093/mnras/stac3795
Advance Access publication 2022 December 26

The optical spectral features of 27 Fermi blazars

Bing-Kai Zhang ,1,2 ‹ Wei-Feng Tang,1 Chun-Xiao Wang,1,2 Qi Wu,1,2 ‹ Min Jin,1,3 Ben-Zhong Dai 4,5

and Feng-Rong Zhu3

1 Department of Physics, Fuyang Normal University, Fuyang 236037, China
2 Key Laboratory of Functional Materials and Devices for Informatics, Anhui Higher Education Institutes, Fuyang 236037, China
3 Department of Applied Physics, Southwest Jiaotong University, Chengdu 610031, China
4 Key Laboratory of Astroparticle Physics, Yunnan Province, Kunming 650091, China
5 Department of Physics, Yunnan University, Kunming 650091, China

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

Accepted 2022 December 20. Received 2022 December 1; in original form 2022 October 21

ABSTRACT
Spectral variation accompanied by flux variability is a commonly observed phenomenon for blazars. In order to investigate the
optical spectral feature of blazars further, we collected the long-term optical V- and R-band data of 27 blazars (14 BL Lacaerte
objects, BL Lacs, and 13 flat-spectrum radio quasars, FSRQs) and calculated their optical spectral indices. The results show
that the spectral indices vary with respect to brightness for all of these blazars. In general, the optical spectrum progressively
becomes flatter (or steeper), when the brightness increases. However the spectrum changes more and more slowly, until it tends
to stability. In other words, the source becomes bluer (or redder) and then gradually stabilizes when it brightens, behaviours that
are referred to as bluer-stable-when-brighter (BSWB) and redder-stable-when-brighter (RSWB), respectively. A total of 13 of
the 14 BL Lacs show the BSWB behaviour, the exception being AO 0235+164. In contrast, most of the FSRQs (10 out of 13)
exhibit the RSWB trend. It is confirmed that blazars follow the two universal optical spectral behaviours, namely BSWB and
RSWB. The model of two constant-spectral-index components can explain the optical spectral features well both qualitatively
and quantitatively. The results illustrate that the optical emission is composed mainly of two stable-colour components, namely
less variable thermal emission and highly variable synchrotron radiation. Furthermore, in most cases, the thermal component of
BL Lacs is redder than that of synchrotron radiation, whereas the opposite is true for FSRQs.
Key words: methods: data analysis – galaxies: active – (galaxies:) BL Lacertae objects: general – (galaxies:) quasars: general.

ultrarelativistic protons or photopion production followed by pion

1 I N T RO D U C T I O N
decay and subsequent cascade (e.g. Mücke et al. 2003; Murase et al.
Blazars are a subclass of active galactic nuclei (AGNs). According to 2012). However, the emission mechanism remains an open question.
their optical emission/absorption line features, they can be divided Blazars exhibit large and rapid variations on a variety of time-
into two subclasses: flat-spectrum radio quasars (FSRQs) and BL scales from years down to even shorter than an hour, and their
Lacaerte objects (BL Lacs). Blazars have been observed in all flux variations are always accompanied by colour variations, that
wavebands from the radio to gamma-rays. Their spectral energy is, by optical spectral variations. The study of the colour variations
distributions display a typical structure, with two main peaks located is helpful for understanding the emission mechanisms. Many works
in a low-energy band, from the radio to the X-ray, and in the high- have investigated colour variation behaviours in the optical band (e.g.
energy gamma-ray band (e.g. Falomo, Pian & Treves 2014). Vagnetti, Trevese & Nesci 2003; Gu et al. 2006; Rani et al. 2010;
It is widely accepted that the low-energy component arises mainly Ikejiri et al. 2011; Bonning et al. 2012; Wierzcholska et al. 2015;
from synchrotron radiation emitted by ultrarelativistic electrons in Zhang et al. 2015; Mao & Zhang 2016; Li et al. 2018; Meng et al.
the jet. In addition, the accretion disc, dusty torus and broad-line 2018; Gaur et al. 2019; Safna et al. 2020; Negi et al. 2022; Zhang,
region (BLR) also make significant contributions to the low-energy Zhao & Wu 2022, and references therein). Generally, there are two
component (e.g. Ghisellini et al. 2019). The high-energy peak can be main types of colour variation behaviours. One type is commonly
explained with leptonic models, in which the high-energy gamma- abbreviated to bluer-when-brighter (BWB), which indicates that the
rays are generated by the inverse Compton scattering of the low- source becomes bluer (i.e. spectrum becomes flatter) as the brightness
energy photons by the same ultrarelativistic electrons in the jet increases (e.g. Takalo & Sillanpää 1989; Raiteri et al. 2015; Dai
(e.g. Dermer & Schlickeiser 1993; Sokolov, Marscher & McHardy et al. 2009; Gupta et al. 2019). The other type is generally referred
2004; Böttcher et al. 2013), or with hadronic models, in which the to as redder-when-brighter (RWB), which indicates that the source
high-energy emission is dominated by the synchrotron radiation of becomes redder (i.e. the spectrum becomes steeper) as the brightness
increases (e.g. Villata et al. 2006; Raiteri et al. 2017).
Previous results showed that most FSRQs exhibit RWB trends,
 E-mail: zhangbk [email protected] (BZ); [email protected] (QW) and BL Lacs tend to display BWB behaviours. However, other

© 2022 The Author(s)

Published by Oxford University Press on behalf of Royal Astronomical Society
5264 B. K. Zhang et al.
Table 1. The characteristics of 27 Fermi blazars.
20
Source Class N AV Behaviour V band

3C 66A BL Lac 352 0.231 BSWB R band

AO 0235+164 BL Lac 182 0.219 RSWB 15
S5 0716+714 BL Lac 206 0.085 BSWB

Flux (mJy)
PKS 0735+178 BL Lac 34 0.094 BSWB
OJ 287 BL Lac 491 0.077 BSWB
Mrk 421 BL Lac 555 0.042 BSWB 10
H 1219+305 BL Lac 14 0.056 BSWB
W Com BL Lac 315 0.064 BSWB
H 1426+428 BL Lac 14 0.034 BSWB
Mrk 501 BL Lac 448 0.052 BSWB 5
1ES 1959+650 BL Lac 131 0.474 BSWB

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

PKS 2155–304 BL Lac 296 0.060 BSWB
BL Lac BL Lac 672 0.901 BSWB
0
1ES 2344+514 BL Lac 80 0.580 BSWB 0 500 1000 1500 2000 2500 3000
CTA 26 FSRQ 24 0.253 RSWB MJD-55000 (days)
PKS 0420–014 FSRQ 51 0.345 RSWB
PKS 0736+01 FSRQ 130 0.375 SWB
OJ 248 FSRQ 193 0.092 RSWB
Ton 599 FSRQ 198 0.054 RSWB
PKS 1222+216 FSRQ 422 0.063 RSWB
3C 273 FSRQ 277 0.057 BSWB
2.5
3C 279 FSRQ 479 0.078 BSWB V band
PKS 1510–08 FSRQ 352 0.275 RSWB
B2 1633+382 FSRQ 340 0.033 RSWB R band
3C 345 FSRQ 41 0.036 RSWB 2.0
CTA 102 FSRQ 304 0.198 RSWB
Flux (mJy)

3C 454.3 FSRQ 531 0.289 RSWB

1.5

special phenomena also exist, and there seems to be a lack of unique

behaviours for blazars. Meanwhile, BWB and RWB trends cannot 1.0
always well describe the characteristics of the colour variations
of blazars (e.g. Zhang et al. 2015). Therefore, segmented linear
functions have been used in an attempt to characterize the colour 0.5
variation features (e.g. Yuan et al. 2017; Fan et al. 2018; Sarkar et al.
2019; Safna et al. 2020; Xiong et al. 2020; Fang et al. 2022). It is,
0 500 1000 1500 2000 2500 3000
however, difficult to explain the abrupt inflection. Recently, Zhang
et al. (2022) found two universal optical spectral behaviours for MJD-55000 (days)
blazars with monitoring data from the Small and Moderate Aperture
Research Telescope System (SMARTS). In the low state, the optical Figure 1. The optical V- and R-band light curves for BL Lac object 1ES
spectrum gradually becomes flatter or steeper, but more and more 1959+650 (upper panel) and FSRQ B2 1633+382 (lower panel).
slowly as the brightness increases, and then tends to stabilize in the
high state. These two phenomena are abbreviated as the bluer-stable-
when-brighter (BSWB) and redder-stable-when-brighter (RSWB) to 2018 using the 1.54-m Kuiper and the 2.3-m Bok telescopes,
behaviours, respectively. These authors also successfully con- and all of the observed data are publicly available at the website.1
structed a non-linear function that well described the quantita- Details of the observation and data analysis are given in Smith et al.
tive optical behaviour of blazars and gave a reasonable physical (2009).
explanation. In order to have enough data points and ensure the quality of the
The purpose in this work is to further investigate the optical spectral behaviours of blazars, we selected objects that have at least
spectral features with a different large sample of Fermi blazars. In 10 pairs of quasi-simultaneous VR photometry observations. Thus,
Section 2, the blazar sample and their light curves in the optical V 27 blazars survive, namely 14 BL Lacs and 13 FSRQs. Most of them
and R bands are introduced. Next, the optical spectral indices are have hundreds of pairs of VR photometry observations, and only a
calculated, and the spectral behaviours are described in Section 3, few sources have fewer than 100 pairs. These sources as well as
followed by a detailed discussion in Section 4 and brief conclusions their optical types and the numbers of observation pairs are given in
in the final section. columns 1–3 of Table 1, respectively.
The magnitudes of each source are converted into flux densities
after the Galactic interstellar reddening correction. The values of
2 DATA C O L L E C T I O N Galactic extinction, Aλ , of the V band are taken from the NASA
We collected optical V- and R-band photometry data from the
long-term monitoring program carried out at Steward Observatory,
University of Arizona. The observations were performed from 2008 1 http://james.as.arizona.edu/∼psmith/Fermi

MNRAS 519, 5263–5270 (2023)

 Spectral features of blazars 5265

Index Index
Fitted curve − 0.5 Fitted curve
99% UCL 99% UCL
− 0.5
99% LCL 99% LCL
Spectral index

Spectral index
− 1.0

− 1.0
− 1.5

α = 2.67ln(0.81 - 0.84/F R ) α = 2.67ln(0.53 + 0.05/F R )

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

− 2.0
− 1.5
4 6 8 10 12 14 16 0.5 1 1.5 2 2.5

Flux (mJy) Flux (mJy)

Figure 2. The optical spectral index against the flux of the optical R band (black dots) for BL Lac object 1ES 1959+650 (left-hand panel) and FSRQ B2
1633+382 (right-hand panel). The red solid line indicates the result of the spectral behaviour fitted with equation (2). The green dashed and blue dash–dotted
lines denote the 99 per cent confidence level bands of the fitted curve. Note that owing to the observation resolution, almost all of the dots are distributed on
horizontal lines with different values, and some of the dots overlap.

Index Index
− 0.5 Fitted curve
Fitted curve
99% UCL 99% UCL
− 0.5
99% LCL 99% LCL
Spectral index

Spectral index

− 1.0

− 1.0
− 1.5

α = -1.38 + 0.05F R α = -1.07 - 0.32F R

− 2.0
− 1.5
4 6 8 10 12 14 16 0.5 1 1.5 2 2.5

Flux (mJy) Flux (mJy)

Figure 3. As Fig. 2, except that equation (3) is used instead of equation (2).

Extragalactic Database (NED2 ), and are listed in column 4 of Table 1. where ν represents the frequency, and Fν represents the flux. We
Fig. 1 gives two examples, which illustrate the optical VR light curves obtained the spectral index for each V and R pair.
for BL Lac object 1ES 1959+650 and FSRQ B2 1633+382. The We also explored the relationship of the optical spectral index α
light curves show violent and large amplitude variations, and exhibit with the flux FR . As a typical example, the left-hand panel of Fig. 2
similar variation trends in the V and R bands. plots the spectral index versus optical R-band flux of the BL Lac
object 1ES 1959+650, which shows how the spectral index varies
with the flux. From the figure, it can be distinctly seen that the value
3 O P T I C A L S P E C T R A L B E H AV I O U R of the optical spectral index gradually gets larger as the flux increases,
which means that the optical spectrum becomes flatter (harder) when
Because the optical spectrum follows a power law and there are
the flux increases, in the sense that the source becomes bluer when it
only two optical-band (V and R) measurements, we determined the
brightens. However, the spectral index and the flux follow a distinct
spectral index by the formula
non-linear relationship. It is clear that the spectral index increases
log Fν2 − log Fν1 step by step as the flux increases, but more and more slowly, and
α= , (1)
log ν2 − log ν1 finally tends to stability. This phenomenon is the typical BSWB
trend.
The right-hand panel of Fig. 2 displays another representative
2 https://ned.ipac.caltech.edu/ example: the opposite spectral behaviour of FSRQ B2 1633+382.

MNRAS 519, 5263–5270 (2023)

5266 B. K. Zhang et al.
are 10 and 284 lower than those of equation (3), which suggests that
Index
− 0.5 equation (2) is the preferred model to describe the spectral behaviour.
Fitted curve
The optical spectral features of other objects are all investigated,
99% UCL
and the behaviours are listed in column 5 of Table 1 and also displayed
99% LCL
in Figs 4–8 for BL Lac objects and in Figs 9–12 for FSRQs. It can
be seen that the variation trends of the spectral index α with flux FR
Spectral index

exhibit BSWB or RSWB trends, and do not seem to follow linear

relationships. Therefore, we fit the spectral behaviour by means of
− 1.0 equations (2) and (3) for each source, and only superimpose the
fitting results with equation (2) on the corresponding figures. The
χ 2 -values fitted by equation (2) for 20 sources are lower than or
equal to those fitted by equation (3). For 24 sources, equation (2) has
α = 2.67ln(0.74 - 0.28/F R ) BICs lower than or equal to those for equation (3). Only three sources
have higher BICs with equation (2) than with equation (3). White’s

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

− 1.5 test for equation (2) shows that there are 19 sources with p-values
4 6 8 10 12 14 16 18
> 0.01, which means that the heteroscedasticity is not statistically
significant. In general, the non-linear equation (2) can well describe
Flux (mJy) the optical spectral features and is superior to the linear equation (3).

Figure 4. As Fig. 2, but for BL Lac object 3C 66A.

4 DISCUSSION
The features of the optical spectral behaviour of 27 blazars have
When the source brightens, the value of the spectral index decreases
been explored, including 14 BL Lacs and 13 FSRQs. Generally,
gradually; that is, the optical spectrum progressively becomes steeper
they follow the BSWB or RSWB trends well. Only one exception,
(softer), but more and more slowly, and then eventually tends to
PKS 0736+01 exhibits, on average, the stable-when-brighter (SWB)
stabilize. In other words, the colour becomes redder, and tends to
trend. In the past, the BWB or RWB trends were widely used
stability when the source brightens. This behaviour is the represen-
to describe the spectral characteristics of blazars. However, an
tative RSWB trend.
increasing number of hints indicate that BWB and RWB trends
Although the BSWB and RSWB exhibit opposite trends in low
are insufficient to represent the spectral characteristics of blazars.
states, they share a common feature that the spectrum tends to
It has been observed that, as the source brightens and exceeds a
stabilize when the source becomes brighter. To characterize the
certain value, the BWB or RWB trend is subsequently followed by
relationship between the spectral index α and the R-band flux FR
a saturation phenomenon (e.g. Villata et al. 2006; Zhang et al. 2015,
quantitatively, we employ a non-linear function,
2021; Sarkar et al. 2019; Safna et al. 2020; Xiong et al. 2020; Otero-
α = 2.67 ln(a + b/FR ), (2) Santos et al. 2022). That is, the colour remains stable when the source
brightens above a certain level (SWB). Recently, Zhang et al. (2022)
proposed by Zhang et al. (2022), to fit the curve of α versus FR , found two new universal optical spectral behaviours, namely BSWB
with the two free parameters a and b. The fitted results as well as the and RSWB trends, and inferred that both BWB and RWB as well as
99 per cent confidence level bands are superimposed on Fig. 2 with the SWB trend can be regarded as special cases of the BSWB and
dashed and dash–dotted lines. The χ 2 -values are 6.58 and 315.69, RSWB trends, respectively. The results of this work further confirm
respectively. Both of the fitted results with equation (2) are consistent the existence of two spectral behaviours in the optical region, namely
with the trends of the spectral behaviours. the BSWB and RSWB trends.
To compare with the linear fitting, we fitted the curves of Fig. 2 The previous results suggest that most FSRQs exhibit a RWB
with the equation trend, while most BL Lacs objects show BWB behaviour (e.g.
α = a + bFR , (3) Safna et al. 2020, and references therein). Furthermore, there exist
exceptional cases; for example, all six BL Lacs observed by the Rapid
with the two free parameters a and b. The results are plotted in Eye Mounting (REM) telescope during 2005 April to 2012 June show
Fig. 3. The χ 2 -values are 16.77 and 599.27, respectively. The fitted no BWB trends according to the data (Sandrinelli, Covino & Treves
results with equation (3) do not seem to be as good as those in 2014), and another example is that only one out of 29 FSRQs in the
Fig. 2. We carried out White’s test to diagnose whether there is Sloan Digital Sky Survey stripe 82 region exhibits the RWB trend
heteroscedasticity in the fitting process. For equation (3), White’s test (Gu & Ai 2011). Zhang et al. (2022) analysed the spectral behaviours
gives p-values of 3.89 × 10−7 and 1.0 × 10−12 for 1ES 1959+650 of 53 Fermi blazars, and their results showed that the vast majority
and B2 1633+382, respectively. For equation (2), the corresponding of FSRQs exhibit RSWB trends; however, BL Lacs show no clear
p-values are 0.74 and 0.25. The test means that at the 1 per cent preference between the BSWB and RSWB trend in their sample.
significance level, the fitting with equation (3) provides evidence of When we examined the subclass of blazars in detail in this work,
heteroscedasticity, while the fitting with equation (2) indicates that we found that 13 out of the 14 BL Lacs obey the BSWB trend,
there is no heteroscedasticity. White’s test suggests that the fitting and only one exception, BL Lac object AO 0235+164, follows the
results with the non-linear equation (2) are better than those with the RSWB trend (see Figs 4–8). For the 13 FSRQs, there are 10 objects
linear equation (3). following the RSWB trend and two (3C 273 and 3C 279) following
For the two equations, we also calculate the Bayesian information the BSWB trend (see Figs 9–12). This means that BL Lacs are more
criteria (BICs), which is a popular statistical approach for a compar- inclined to exhibit the BSWB trend, and FSRQs are more inclined to
ative evaluation among different models. The model with the lower exhibit the RSWB trend. It seems that the spectral behaviour of an
BIC value is preferred. For these two blazars, the BICs of equation (2) object is independent of whether it is a TeV source.

MNRAS 519, 5263–5270 (2023)

 Spectral features of blazars 5267
0.0
Index − 0.5 Index Index
Fitted curve Fitted curve Fitted curve
99% UCL 99% UCL 99% UCL
− 1.5
99% LCL 99% LCL − 0.5 99% LCL

− 1.0
Spectral index

Spectral index

Spectral index
− 2.0
− 1.0

− 2.5 − 1.5
− 1.5
α = 2.67ln(0.39 + 0.01/F R ) α = 2.67ln(0.62 - 0.33/F R ) α = 2.67ln(0.74 - 0.10/F R )
− 3.0
− 2.0 − 2.0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 10 20 30 40 50 60 70 1.2 1.3 1.4 1.5 1.6 1.7

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

Figure 5. As Fig. 2, but for BL Lac objects (a) AO 0235+164, (b) S5 0716+714, and (c) PKS 0735+178.

0.0 − 0.8
− 0.5
Index Index Index
Fitted curve Fitted curve Fitted curve
99% UCL 99% UCL 99% UCL

− 1.0 99% LCL 99% LCL − 1.0 99% LCL

Spectral index

Spectral index

Spectral index
− 0.5

− 1.5 − 1.2

− 2.0 − 1.0 − 1.4

α = 2.67ln(0.58 - 0.10/F R ) α = 2.67ln(0.84 - 2.07/F R ) α = 2.67ln(0.81 - 0.24/F R )

− 2.5
− 1.6
2 4 6 8 10 12 14 16 10 20 30 40 50 60 70 1.2 1.4 1.6 1.8 2 2.2

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Figure 6. As Fig. 2, but for BL Lac objects (a) OJ 287, (b) Mrk 421, and (c) H 1219 + 305.

− 0.5
Index Index Index
Fitted curve − 2.6 Fitted curve Fitted curve
− 1.5
99% UCL 99% UCL 99% UCL

− 1.0 99% LCL 99% LCL 99% LCL

Spectral index

Spectral index

Spectral index

− 2.8

− 1.5

− 2.0
− 3.0
− 2.0

α = 2.67ln(0.63 - 0.17/F R ) α = 2.67ln(0.41 - 0.11/F R ) α = 2.67ln(0.64 - 2.01/F R )

− 3.2
− 2.5
2 3 4 5 6 7 1.2 1.3 1.4 1.5 1.6 1.7 1.8 12.5 13 13.5 14 14.5 15 15.5 16 16.5 17

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Figure 7. As Fig. 2, but for BL Lac objects (a) W Com, (b) H 1426+428, and (c) Mrk 501.

In general, the relationship between spectral indices and fluxes month time-scales. Fig. 13 shows an example of the 12-day short-
can be well fitted by equation (2). In order to assess the fitting term spectral behaviour of B2 1633+382 during MJD 58281–58293.
quality, we examined the fitted results source by source. On closer The results demonstrate that the sources obey the BSWB or RSWB
inspection, the spectral feature can generally be fitted well in both trend on both long-term and short-term time-scales. Furthermore, the
high and low states. For instance, it can be seen that equation (2) sources have been checked during different stages, and the results
is a good description of the spectral behaviour of 1ES 1959+650 show that the sources follow the law of BSWB or RSWB well
whether the source is in a brighter or darker state, that is, whether whether during the outburst or descent stage. Fig. 14 shows the
the flux has a high value or lower value (see Fig. 2). Only three spectral behaviour of 1ES 1959+650 when it is in an outburst stage
sources, CTA 102, 3C 454.3 and OJ 287, cannot be well fitted in their during MJD 55600–56300.
very-high states. For each source, the spectral behaviours have been Several mechanisms have been proposed to interpret the spectral
examined on different time-scales. They have been investigated not behaviour. The two different spectral behaviours of BWB and
only on several-year time-scales, but also on several-day to several- RWB are explained by different models, and even for the same

MNRAS 519, 5263–5270 (2023)

5268 B. K. Zhang et al.

Index Index Index

Fitted curve
− 2.0 Fitted curve − 1.0 Fitted curve
− 0.5 99% UCL 99% UCL 99% UCL
99% LCL 99% LCL 99% LCL

Spectral index
Spectral index

Spectral index
− 2.5 − 1.5

− 1.0
− 3.0
− 2.0

α = 2.67ln(0.70 - 0.20/F R ) α = 2.67ln(0.37 - 0.37/F R ) α = 2.67ln(0.70 - 1.01/F R )

− 3.5
− 2.5
− 1.5
5 10 15 20 25 30 35 40 45 50 10 20 30 40 50 60 4.6 4.8 5 5.2 5.4 5.6 5.8 6 6.2

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

Figure 8. As Fig. 2, but for BL Lac objects (a) PKS 2155–304, (b) BL Lac, and (c) 1ES 2344+514.

Index Index Index

Fitted curve 0.0 Fitted curve Fitted curve
99% UCL 99% UCL 99% UCL
− 1.5
99% LCL 99% LCL 99% LCL

− 1.0 − 0.5
Spectral index

Spectral index

Spectral index
− 2.0
− 1.0

− 1.5 − 1.5
− 2.5
α = 2.67ln(0.53 + 0.11/F R ) α = 2.67ln(0.62 + 0.06/F R ) α = 2.67ln(0.46 - 0.00/F R )

− 2.0
0.6 0.8 1 1.2 1.4 1.6 0.5 1 1.5 2 2.5 3 2 4 6 8 10 12

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Figure 9. As Fig. 2, but for FRSQs (a) CTA 26, (b) PKS 0420–014, and (c) PKS 0736+01.

Index − 0.5 Index Index

Fitted curve Fitted curve Fitted curve
0.0 99% UCL 99% UCL
0.5 99% UCL
99% LCL 99% LCL 99% LCL

− 1.0
Spectral index

Spectral index

Spectral index

− 0.5
0.0

− 1.0
− 1.5
− 0.5
− 1.5
α = 2.67ln(0.55 + 0.16/F R ) α = 2.67ln(0.58 + 0.01/F R ) α = 2.67ln(0.75 + 0.57/F R )

− 2.0 − 1.0
− 2.0
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 0 2 4 6 8 10 12 2 3 4 5 6 7 8 9 10

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Figure 10. As Fig. 2, but for FRSQs (a) OJ 248, (b) Ton 599, and (c) PKS 1222+216.

behaviour, there exist different explanations. For the BWB trend, mechanism. They proposed a new model with two constant-spectral-
different models have been proposed, such as fresh electron injection index components, and inferred a non-linear formula, which can
(Kirk, Rieger & Mastichiadis 1998), a one-component synchrotron explain both the BSWB and RSWB trends well, not only qualitatively
contribution (Fiorucci, Ciprini & Tosti 2004), inter-band time delay but also quantitatively.
(Wu et al. 2007), and a two-component contribution (Fiorucci et al. In this study, the fact that both BSWB and RSWB behaviours can
2004; Ikejiri et al. 2011). Meanwhile, the RWB behaviour has been be well fitted by the same non-linear formula (equation 2) implies
interpreted as a strong contribution of blue thermal emission from that both of them can be explained by the model of two constant-
the accretion disc (Villata et al. 2006; Rani et al. 2010; Bonning spectral-index components. The emission of the optical band consists
et al. 2012; Isler et al. 2017), or changes in the Doppler factor of a weak and relatively stable thermal component, and a strong and
(Raiteri et al. 2017). Zhang et al. (2022) suggested that the BWB highly variable synchrotron emission component, both of which have
and RWB trends are special cases of the BSWB and RSWB trends constant spectral indices. In the low state, the spectral index tends
in low states, respectively, and that both of them arise from the same to be close to the spectral index of thermal emission, while in the

MNRAS 519, 5263–5270 (2023)

 Spectral features of blazars 5269
0.0
0.5 Index − 0.5 Index Index
Fitted curve Fitted curve Fitted curve
99% UCL 99% UCL 99% UCL
99% LCL 99% LCL − 0.5 99% LCL
0.0 − 1.0
Spectral index

Spectral index
Spectral index
− 1.0
− 0.5 − 1.5

− 1.5
− 1.0 − 2.0

α = 2.67ln(0.89 - 1.26/F R ) α = 2.67ln(0.57 - 0.05/F R ) α = 2.67ln(0.63 + 0.05/F R )

− 2.0
− 1.5 − 2.5
22 24 26 28 30 32 34 36 0 5 10 15 20 25 2 4 6 8 10 12

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

Figure 11. As Fig. 2, but for FRSQs (a) 3C 273, (b) 3C 279, and (c) PKS 1510–08.

Index Index Index

− 0.5
Fitted curve Fitted curve Fitted curve
− 0.5 99% UCL −1 99% UCL 99% UCL
99% LCL 99% LCL 99% LCL
− 1.0
Spectral index

Spectral index

Spectral index
− 1.0

− 1.5
−2
− 1.5

− 2.0
− 2.0
α = 2.67ln(0.57 + 0.02/F R ) α = 2.67ln(0.41 + 0.11/F R ) α = 2.67ln(0.49 + 0.32/F R )
−3
− 2.5
− 2.5
0.4 0.6 0.8 1 1.2 1.4 1.6 0 20 40 60 80 100 120 0 5 10 15 20 25

Flux (mJy) Flux (mJy) Flux (mJy)

(a) (b) (c)

Figure 12. As Fig. 2, but for FRSQs (a) 3C 345, (b) CTA 102, and (c) 3C 454.3.

− 1.0
Index Index

Fitted curve Fitted curve

− 0.5
Spectral index

Spectral index

− 1.5

− 1.0

α = 2.67ln(0.50 + 0.09/F R ) α = 2.67ln(0.82 - 0.90/F R )

− 2.0
0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 6 8 10 12 14 16

Flux (mJy) Flux (mJy)

Figure 13. As Fig. 2, but for FRSQ B2 1633+382 during MJD 58281– Figure 14. As Fig. 2, but for BL Lac object 1ES 1959+650 during an
58293. outburst stage between MJD 55600 and MJD 56300.

high state, because the emission begins to be dominated by the more colour of the synchrotron emission component is bluer than that of the
variable non-thermal component, the index progressively approaches thermal component, so most BL Lacs exhibit the BSWB trend. For
that of the non-thermal component. This mechanism naturally creates FSRQs, however, the colour of the synchrotron emission component
the BSWB or RSWB trend. is redder than that of the thermal component, and as a result most of
BL Lacs have less thermal emission contributions than FSRQs; them show the RSWB trend. However, there are several exceptions.
therefore, their colours of the thermal component are much redder For example, AO 0235+164, as a BL Lac object, is more like an
than those of FSRQs. Moreover, in most cases, for BL Lacs, the FSRQ in many physical characteristics (Chen & Jiang 2001), and it

MNRAS 519, 5263–5270 (2023)

5270 B. K. Zhang et al.
has a much more luminous accretion disc than other BL Lacs, and can Fang Y., Zhang Y., Chen Q., Wu J., 2022, ApJ, 926, 91
even be comparable to FSRQs (Ghisellini et al. 2010; Paliya et al. Fernandes S., Patino- ˜ Álvarez V. M., Chavushyan V., Schlegel E. M., Valdés
2021), so it follows the RSWB trend. The situation for the FSRQ J. R., 2020, MNRAS, 497, 2066
3C 279 is just the opposite, which implies that its thermal emission Fiorucci M., Ciprini S., Tosti G., 2004, A&A, 419, 25
Gaur H. et al., 2019, MNRAS, 484, 5633
is relatively weak. As another special case, the source FSRQ 3C
Ghisellini G., Tavecchio F., Foschini L., Ghirlanda G., Maraschi L., Celotti
273 has a strong blue component from the accretion disc (Fernandes
A., 2010, MNRAS, 402, 497
et al. 2020; Li et al. 2020), but it exhibits a slight BSWB trend, which Ghisellini G. et al., 2019, A&A, 627, A72
may be due to its bluer synchrotron radiation. Whether it is a BL Lac Gu M. F., Ai Y. L., 2011, A&A, 528, A95
object or an FSRQ, an object’s spectral behaviour ultimately depends Gu M. F., Lee C.-U., Pak S., Yim H. S., Fletcher A. B., 2006, A&A, 450, 39
on the quantitative comparison between the spectral index values of Gupta A. C. et al., 2019, AJ, 157, 95
two components, namely the thermal emission component from the Ikejiri Y. et al., 2011, PASJ, 63, 639
accretion disc and the synchrotron radiation component from the jet. Isler J. C. et al., 2017, ApJ, 844, 107
Kirk J. G., Rieger F. M., Mastichiadis A., 1998, A&A, 333, 452
Li X.-P., Luo Y.-H., Yang H.-T., Yang H.-Y., Yang C., Cai Y., 2018, Res.

Downloaded from https://academic.oup.com/mnras/article/519/4/5263/6961056 by guest on 25 March 2025

5 CONCLUSIONS Astron. Astrophys., 18, 150
Li Y.-R., Zhang Z.-X., Jin C., Du P., Cui L., Liu X., Wang J.-M., 2020, ApJ,
In sum, we have explored the optical spectral features of 27 blazars 897, 18
and confirmed that blazars follow the universal BSWB or RSWB Mao L., Zhang X., 2016, Ap&SS, 361, 345
trend. The results suggest that FSRQs prefer the RSWB trend, while Meng N., Zhang X., Wu J., Ma J., Zhou X., 2018, ApJS, 237, 30
BL Lacs favour the BSWB trend. These two trends can be described Mücke A., Protheroe R. J., Engel R., Rachen J. P., Stanev T., 2003, Astropart.
quantitatively by the same non-linear formula. The model with Phys., 18, 593
two constant-spectral-index components can successfully interpret Murase K., Dermer C. D., Takami H., Migliori G., 2012, ApJ, 749, 63
the optical spectral features both qualitatively and quantitatively, Negi V., Joshi R., Chand K., Chand H., Wiita P., Ho L. C., Singh R. S., 2022,
including the differences between BL Lac objects and FSRQs. MNRAS, 510, 1791
Otero-Santos J. et al., 2022, MNRAS, 511, 5611
Paliya V. S., Domı́nguez A., Ajello M., Olmo-Garcı́a A., Hartmann D., 2021,
AC K N OW L E D G E M E N T S ApJS, 253, 46
Raiteri C. M. et al., 2015, MNRAS, 454, 353
We are grateful to the referee for his/her valuable suggestions. Raiteri C. M. et al., 2017, Nature, 552, 374
This work has been supported by the National Natural Science Rani B. et al., 2010, MNRAS, 404, 1992
Foundation of China (grant no. U1831124) and by the Key Lab- Safna P. Z., Stalin C. S., Rakshit S., Mathew B., 2020, MNRAS, 498, 3578
oratory of Functional Materials and Devices for Informatics (grant Sandrinelli A., Covino S., Treves A., 2014, A&A, 562, 79
no. 2020XXGN01). We thank Steward Observatory of University Sarkar A. et al., 2019, ApJ, 887, 185
of Arizona for the long-term optical monitoring photometry data. Smith P. S., Montiel E., Rightley S., Turner J., Schmidt G. D., Jannuzi B. T.,
The Steward Observatory spectropolarimetric monitoring program 2009, preprint (arXiv:0912.3621)
is supported by Fermi Guest Investigator grants NNX08AW56G, Sokolov A., Marscher A. P., McHardy I. M., 2004, ApJ, 613, 725
Takalo L. O., Sillanpää A., 1989, A&A, 218, 45
NNX09AU10G, NNX12AO93G, and NNX15AU81G.
Vagnetti F., Trevese D., Nesci R., 2003, ApJ, 590, 123
Villata M. et al., 2006, A&A, 453, 817
DATA AVA I L A B I L I T Y Wierzcholska A., Ostrowski M., Stawarz Ł., Wagner S., Hauser M., 2015,
A&A, 573, A69
All of the optical data used in this work are publicly available at the Wu J., Zhou X., Ma J., Wu Z., Jiang Z., Chen J., 2007, AJ, 133, 1599
website http://james.as.arizona.edu/∼psmith/Fermi. Xiong D. et al., 2020, ApJS, 247, 49
Yuan Y.-H. et al., 2017, A&A, 605, A43
Zhang B.-K., Zhou X.-S., Zhao X.-Y., Dai B.-Z., 2015, Rev. Astron.
REFERENCES Astrophys., 15, 1784
Bonning E. et al., 2012, ApJ, 756, 13 Zhang B.-K., Jin M., Zhao X.-Y., Zhang L., Dai B.-Z., 2021, Rev. Astron.
Böttcher M., Reimer A., Sweeney K., Prakash A., 2013, ApJ, 768, 54 Astrophys., 21, 186
Chen Y. J., Jiang D. R., 2001, A&A, 376, 69 Zhang B.-K., Zhao X.-Y., Wu Q., 2022, ApJS, 259, 49
Dai B. Z. et al., 2009, MNRAS, 392, 1181
Dermer C. D., Schlickeiser R., 1993, ApJ, 416, 458
Falomo R., Pian E., Treves A., 2014, A&AR, 22, 73
Fan X.-L. et al., 2018, ApJ, 856, 80 This paper has been typeset from a TEX/LATEX file prepared by the author.

MNRAS 519, 5263–5270 (2023)