Robopol: optical polarization monitoring of an
unbiased sample of blazars
Emmanouil Angelakis,
Max-Planck-Institut für Radioastronomie, Auf dem Huegel 69, Bonn 53121, Germany
T. Hovatta, D. Blinov. V. Pavlidou, S. Kiehlmann, I. Myserlis, M. Böttcher
and the RoboPol collaboration
gamma-ray loud:
2FGL: ➡ integrated
➡F
photon flux F (> 100 MeV) (> 100 MeV) > 2 × 10−8 cm−2 s−1,
➡ we
exclude the galactic plane: |b| 10◦
➡ non-biasing
cuts: 62 GL sources gamma-ray quiet:
OVRO monitoring:
➡ intrinsic
➡ mean
modulation, m>= 0.05 flux density, S>= 0.06 Jy ➡ non-biasing
cuts: 15 GQ sources season 2013
season 2014
the
all
gles
and
he
ble
les
ved
nd
een
ion
ble
ver
ed
inen
ow
on
me
ver
han
ingles
ow
me
ean
an
vales
gue
an
vaue
in
stiits
ies.
in
der-
CDF
all
from June 2013 survey
intrinsic polarization fraction p0 . The orange triangles indicate the sources
fraction that30switched from the GQ sample to the GL in the 3FGL catalogue.
Our
analysis
is focused
on the behavior of the polarization
GL:
0.064
(+0.009-0.008)
-p and
its variability for GL and GQ sources. We first examine the 25
4median
ANALYSIS
polarization
fraction p̂ of each source computed from mea0.032(+0.02-0.011)
- GQ:
surements with p/σ p ≥ 3. This quantity has the advantage that it GL20sources classified as “bzb” are found at systematically lower
Our analysis
focused
on the behavior of the polarization fraction
Pavlidou
et al.is2014,
MNRAS.442.1693P
(median 0.308) as opposed to “bzq” sources that are lois very straight forward to define and compute. However, it clearly redshifts
15
p and its variability for GL and GQ sources. We first examine the
only characterizes sources during their stages of significant polar- cated clearly farther with a median of around 0.867, in accordance
median polarization fraction p̂ of each source computed from meaGL fitted(e.g.
ization, ignoring non-detections and the associated cycles of low with10 what systematic studies of blazar samples have shown
GQ fitted
p/σ p ≥ 3. This quantity has the advantage thatMassaro
et al. 2009). The GQ sources on the other hand
are alsurements
with
it
polarization. For this reason, we also include a realistic analysis most5 uniformly distributed over a broad range of redshifts
GLreaching
is
very
straight
forward
to
define
and
compute.
However,
it
clearly
GQ
which accounts for limited sampling, measurement uncertainties, up to 3.18. Hence, their cosmological distance cannot explain
– at
only
characterizes
sources
during
their
stages
of
significant
polar0
Angelakis
et al. Submitted toanalysis
MNRASto least0.00
and Ricean bias, by applying
a maximum-likelihood
not alone0.05
– their gamma-ray
silence. Their
is
0.10
0.15
0.20 median
0.25 redshift
0.30
ization,
ignoring
non-detections
and
the
associated
cycles
of
low
p0 the positions of the two GQ
compute the intrinsic mean polarization fraction p0 and its associ- around 0.5. The orange triangles mark
polarization.
For this reason,
we (Sect.
also include
a realistic
analysis
that appeared in the 3FGL (Acero et al. 2015).
ated intrinsic modulation
index m
3.2), together
with uncerCDF
hat
!
"
!
" pα−1 1 − p β−1
PDF p; α,
= polarization
!
"
(1)
interval,
as βis the
B α, β fraction):
0.0
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
!
" β−1
α−1
p̂
If the! parameters
" pa, β of1this
− pdistribution are known, the intrinsic
PDF
p; α,itsβ modulation
=
!
"
(1)
mean and
B index
α, β are then given by
1.0
α
If
p0the
= parameters a, β of this distribution are known, the intrinsic
(2)
α+β
0.8
mean and its modulation index are then given by
and
α
#
p0 = √
(2)
Var
α
β
α
+
β
α+β
0.6
-3
p-value
~
10
=
· !
mp =
.
(3)
" !
"
p0
α
α+ β 2 α+ β+1
and
#
√
0.4
with Var Var
the variance
αβ
α + βof the distribution.
= advantage
· ! of this
m p =An essential
" 2 approach
!
is" .that it provides(3)
p0
α
GL
α +distribution:
β
α+ β+1
assuming
a
power
law
estimates of both uncertainties and, when appropriate, upper limits. 0.2
GQ
The method
been applied
in cases with at least 3 data points
with
Var ⟨p
thehas
variance
of the
distribution.
± only
0.008
-outGL:
0⟩ ~ 0.092 p
RBPLJ1624+5652
ofAn
which
at leastadvantage
2 had /σ pof≥ this
3. All
the results
ouritanalysis
essential
approach
is ofthat
provides RBPLJ1638+5720
0.0
3.
are
shown
in
Table
0.00
0.05
0.10
0.15
0.20
0.25
estimates
of
both
uncertainties
and,
when
appropriate,
upper
limits.
- GQ: ⟨p0⟩ ~ 0.031 ± 0.008
p0
The method has been applied only in cases with at least 3 data points
40
out of which at least 2 had p/σ p ≥ 3. All the results of our analysis
Figure 1. The cumulative distribution function of the median polarization
are
in Table 3.
35 for the GL (black) and GQ samples (blue lines). Lower: same for the
fraction
4 shown
ANALYSIS
GL:
• highly variable, strong jet dominance due to:
- high degree of Doppler boosting (e.g.
Savolainen et al. 2010; Lister et al. 2015)
- frequent occurrence of impulsive events of
particle acceleration
mildly relativistic shock,
particle acceleration (DSA or reconnection)
B-field compression
• optical from smaller volumes hence higher
polarization
shock →
GQ:
high energyreconnection)
e- particle acceleration (DSA/magn.
small volume
B-"ield compression
high polarization
low energy e- large volume
low polarization
νFν
• objects with:
- less extreme Doppler boosting or
- impulsive episodes less efficient,
• optical from larger volumes hence lower high-energy epolarization
small volume
→ high polarization
-47*10
-6)
for
forfor
BL
GLGL
Lac
only
&GL
GQ:
: ρ only
=ρ−0.3
=: −0.2
ρ (p-value:
= −0.5
(p-value:
(p-value:
2.8*10
0.02)
)
Angelakis et al. Submitted to MNRAS
νFν
high-energy esmall volume
→ high polarization
low-energy e large volume
→ low polarization
MORE of cells
less variable p
optical in
HSP blazars
optical in
LSP blazars
LESS cells
less variable p
ν
Two types
Blinov et al., MNR
+ Kiehlmann et a
Blinov et al. MNRAS 453, 1669
Blinov et al. in preparation
νFν
high-energy esmall volume
→ high polarization
low-energy e large volume
→ low polarization
optical in
HSP blazars
turbulent field is dominant
stochastic rotations
optical in
LSP blazars
helical field is dominant
deterministic rotations
close to gamma activity
ν
9
The optical polarization of GL and GQ blazars
5
16
RBPLJ1751+0939
The optical polarization of GL and GQ blazars
5
RBPLJ1653+3945
16
RBPLJ1751+0939
RBPLJ1653+3945
14
14
4
2
10
Counts
3
8
12
Counts
12
Counts
Counts
4
3
2
6
2
−50
0
2
0
50
−50
EVPA (deg)
0
0
50
−50
EVPA (deg)
1.0
1.0
RBPLJ1751+0939
8
4
1
0
10
6
4
1
0
0
50
−50
EVPA (deg)
RBPLJ1653+3945
0.6
0.6
0.6
0.6
CDF
CDF
0.8
CDF
0.8
CDF
0.8
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
0
50
EVPA (deg)
0.0
−50
0
EVPA (deg)
50
50
1.0
RBPLJ1751+0939
0.8
−50
0
EVPA (deg)
1.0
RBPLJ1653+3945
0.0
9
0.0
−50
0
50
EVPA (deg)
0.0
−50
0
50
EVPA (deg)
Figure 7. The distribution of EVPA for a close-to-uniform case (highly random, RBPLJ1653+3945 left) and one case far-from-uniform
(low randomness,
Figure 7. The distribution
of EVPA for a close-to-uniform case (highly random, RBPLJ1653+3945 left) and one case far-from-uniform (low randomness,
RBPLJ1751+0939 right). Upper row: The distribution of EVPA. Lower row: The cumulative distribution function of the EVPA
for those two cases
line)
RBPLJ1751+0939
right).(solid
Upper
row: The distribution of EVPA. Lower row: The cumulative distribution function of the EVPA for those two cases (solid line)
and the one of uniform distribution (dashed line). There are 46 data angle measurements for RBPLJ1751+0939 and 51 for and
RBPLJ1653+3945.
the one of uniform distribution (dashed line). There are 46 data angle measurements for RBPLJ1751+0939 and 51 for RBPLJ1653+3945.
0.037 ± 0.010
high-synchrotron peaked (LSP, if Log(νs ) < 14, ISP if 14 ≤
Log(νs ) < 15 and HSP if Log(νs ) ≥ 15, respectively). Then we selected 0.1 as the limiting value of χ2red for a source to be considered
as non-uniform. We then found that: 11/14 (79%) LSP, 7/14 (50%)
ISP and 3/8 (38%) HSP sources, have χ2red below 0.1. Despite the
small number statistics, this result indicates that HSP sources are
more likely to have a preferred and less variable EVPA than LSP
sources.
Second, the green markers in Fig. 6 show the mean χ2red in
each of five synchrotron-peak frequency bins. The vertical errorbars show the spread of the values in the bin (1σ). A linear fit to
the binned data – the green dashed line – gave a significant slope of
0.037 ± 0.010.
We conclude that the randomness of the EVPA depends on the
synchrotron peak frequency. The EVPA of HSP sources is concentrated around preferred directions. The EVPA of LSP sources, on
the other hand, is more variable and less likely to have a preferred
direction. In § 5 we argue that these two findings may indeed be
evidence for a helical structure of the magnetic field.
4.6
Polarization and source variability
Depending on the mechanism producing the variability, it is likely
that the degree of polarization relates to the degree of variability
at different bands. Here we examine the role that the radio and the
optical modulation indices may play.
In Fig. 8 we plot the median polarization fraction versus the
variability amplitude at 15 GHz from Richards et al. (2014), as that
ρ = 0.34 (p-value: 0.044)
Richards et al. (2011). As it is seen
there, the two are
correlated
high-synchrotron
peaked
(LSP, if Log(νs ) < 14, ISP if 14 ≤
with Spearman’s test to be giving a Log(ν
ρ ∼ 0.35
p-value
about s ) ≥ 15, respectively). Then we se15aand
HSP of
if Log(ν
s ) <and
3 × 10−4 . The GQ sources have preferentially
radio
modulation
lected 0.1 low
as the
limiting
value of χ2red for a source to be considered
indices as it was already found by Richards
et
al.
(2011).
However,
as non-uniform. We then found that: 11/14 (79%) LSP, 7/14 (50%)
the GQ sources have average polarization
fractions
that are
low
even have χ2red below 0.1. Despite the
ISP and
3/8 (38%)
HSP
sources,
compared to GL sources with comparable
modulation
small radio
number
statistics,indices.
this result indicates that HSP sources are
more likely
have a preferred
In Fig. 9 we examine the dependence
of thetopolarization
frac- and less variable EVPA than LSP
sources.
tion on the variability amplitude of
the R-band flux density. In
the green markers
the upper panel we plot the observed Second,
median polarization
frac- in Fig. 6 show the mean χ2red in
of five synchrotron-peak
tion p̂ and the R-band flux densityeach
modulation
index m S . In thisfrequency bins. The vertical errorshow
of the values
in the bin (1σ). A linear fit to
case Spearman’s ρ, when includingbars
both
GL the
andspread
GQ sources,
is
binned data
– the significant
green dashed line – gave a significant slope of
around 0.38 with a p-value of 10−4the
, indicating
a rather
0.010.
correlation. Similarly, in the lower 0.037
panel ±we
show the maximumWe conclude
thatmthe
randomness of the EVPA depends on the
likelihood intrinsic mean polarization fraction
p0 and the
S which
peak
frequency.
The EVPA of HSP sources is concengave a Spearman’s ρ ≈ 0.38 with a synchrotron
p-value of 8 ×
10−4
. Again, GQ
trated around
preferred
directions. The EVPA of LSP sources, on
sources are systematically less polarized
on average
than sources
other hand, is more variable and less likely to have a preferred
with comparable optical modulationthe
indices.
direction.
§ 5 we on
argue
that these two findings may indeed be
Finally, in Fig. 10 we examine whether p0Independs
the amevidence
a helical
structure of the magnetic field.
plitude of its variability quantified through
thefor
intrinsic
polarization
modulation index m p . Spearman’s test gave a ρ of around −0.31
with a significance of p-value 0.013.
4.6 Polarization and source variability
We conclude that the variability amplitude, in both radio and
optical flux density, affects the mean
observedonpolarization.
Withproducing the variability, it is likely
Depending
the mechanism
comparable Spearman’s test results,that
thethe
higher
polarization
is as- relates to the degree of variability
degree
of polarization
sociated with stronger variability inateither
the bands.
opticalHere
or theweradio
different
examine the role that the radio and the
light curves. Finally, there is also aoptical
weak modulation
indication that
stronger
indices
may play.
variability in optical polarization associates
(on average)
8 we plotwith
thelower
median polarization fraction versus the
In Fig.
polarization although of lower significance.
Nevertheless,
these
cor- from Richards et al. (2014), as that
variability
amplitude at
15 GHz
Richards et al. (2011). As it is seen there, the two are correlated
with Spearman’s test to be giving a ρ ∼ 0.35 and a p-value of about
3 × 10−4 . The GQ sources have preferentially low radio modulation
indices as it was already found by Richards et al. (2011). However,
the GQ sources have average polarization fractions that are low even
compared to GL sources with comparable radio modulation indices.
In Fig. 9 we examine the dependence of the polarization fraction on the variability amplitude of the R-band flux density. In
the upper panel we plot the observed median polarization fraction p̂ and the R-band flux density modulation index m S . In this
case Spearman’s ρ, when including both GL and GQ sources, is
around 0.38 with a p-value of 10−4 , indicating a rather significant
correlation. Similarly, in the lower panel we show the maximumlikelihood intrinsic mean polarization fraction p0 and the m S which
gave a Spearman’s ρ ≈ 0.38 with a p-value of 8 × 10−4 . Again, GQ
sources are systematically less polarized on average than sources
with comparable optical modulation indices.
Finally, in Fig. 10 we examine whether p0 depends on the amplitude of its variability quantified through the intrinsic polarization
modulation index m p . Spearman’s test gave a ρ of around −0.31
with a significance of p-value 0.013.
We conclude that the variability amplitude, in both radio and
optical flux density, affects the mean observed polarization. With
comparable Spearman’s test results, the higher polarization is associated with stronger variability in either the optical or the radio
light curves. Finally, there is also a weak indication that stronger
variability in optical polarization associates (on average) with lower
polarization although of lower significance. Nevertheless, these cor-
Angelakis et al. Submitted to MNRAS
0.30
0.40
ρ = 0.38 (p-value: 3*10-4)
0.35
GL
GQ
0.25
ρ = 0.35 (p-value: 3*10-4)
GL
GQ
0.30
0.25
0.15
p̂
p̂
0.20
0.20
0.15
0.10
0.10
0.05
0.05
0.00
0.0
0.2
0.4
0.6
mS
0.8
1.0
1.2
0.00
0.0
0.1
0.2
0.3
0.4
m15,Richards
0.5
0.6
0.7
0.8
et el. 2014
0.25
GL
GQ
0.20
stronger variability causes higher
polarization
➡
the more variable sources are less
polarized
p0
0.15
➡
0.10
0.05
ρ = −0.31 (p-value: 0.013)
0.00
0.0
0.5
1.0
mp
1.5
2.0
Angelakis et al. Submitted to MNRAS
1.0
1.6
weak1.4indication: GL
GQ
ρ = 0.3
(p-value: 0.016 ~ 2.5σ)
1.2
0.8
3.0
KS test: p-value: 0.255
1.0
2.5
CDF
0.6
3.0
0.4
0.2
mS
0.6
0.5
0.5
0.4
0.0
0.0
0.2
0.5
1.0
1.5
0.4
0.6
mp
0.8
mS
1.0
1.2
1.4
1.6
0.0
0.0
0.1
0.2
0.3
0.4
m15
0.5
0.6
0.7
0.8
no correlation
0.0
0.0
2.0
1.5
1.0
0.2
0.0
0.0
2.0
1.5
1.0
GL
GQ
0.8
GL
GQ
GL: 2σ upp. limit
GL: 2σ upp. limit
2.5
mp
mp
2.0
GL
GQ
GL: 2σ upp. limit
GL: 2σ upp. limit
0.1
0.2
0.3
0.4
m15
0.5
0.6
0.7
0.8
3.0
Figure 11. The intrinsic modulation index
-3) m p for the GL and GQ samples.
ρ
=
0.43
(p-value:
10
GL
In cases
2.5 where the m p was not available 2σ upper limits have been included
GQ
instead.
GL: 2σ upp. limit
GQ: 2σ upp. limit
mp
2.0
3.0
1.5
2.5
mp
3.0 polarization variability amplitude mp
the
increase with cosmic distance GL
➡ weak
GQ
dependence of polarization
variability
GL: 2σ upp. limit
GL: 2σ upp. limit
in2.0optical variability and no in radio
mp
0.5
1.0
1.5
2.0
2.5
1.5
1.0
z
0.5
0.5
and GQ are indistinguishable for mp
2.5
0.5
1.5
0.0
1.00.0
➡ GL
➡
GL
GQ
GL: 2σ upp. limit
GQ: 2σ upp. limit
1.0
2.0
Figure 13. The R-band intrinsic modulation index m S versus that at 15
GHz, m 15 .
Angelakis et al. Submitted to MNRAS
➡ GL
more polarized than GQ sources
➡ the
polarization decreases with the synchrotron peak frequency
➡ the
polarization spread decreases with the synchrotron peak frequency
-
mildly relativistic shock in a jet with helical and a turbulent component may
explain the observations
➡ the
EVPA clearly shows a preferred direction for HSP sources ➡ the
EVPA is randomly orientated in LSP sources
-
???
➡ LSPs
-
dominate the deterministic rotations with small lags to the gamma activity possibly with plasmoids traveling in regions with helical field
➡ stronger
➡
variability in optical or radio causes higher polarization
the more variable the polarization the less polarized
➡ GL
and GQ are indistinguishable for mp