Draft version December 11, 2024

Typeset using LATEX preprint style in AASTeX631

Kinematics of Cataclysmic Variables in the Solar Neighborhood in the Gaia Era

Remziye Canbay ,1 Tansel AK ,2, 3 Selçuk Bilir ,2 Faruk Soydugan ,4, 5 and Zeki Eker 6

1 Istanbul University, Institute of Graduate Studies in Science, Programme of Astronomy and Space Sciences, 34116, Istanbul, Turkey
2 Istanbul University, Faculty of Science, Department of Astronomy and Space Sciences, 34119, Istanbul, Turkey
3 Istanbul University, Observatory Research and Application Center, Istanbul University, 34119, Istanbul, Turkey
arXiv:2412.06882v1 [astro-ph.GA] 9 Dec 2024

4 Faculty of Sciences, Department of Physics, Canakkale Onsekiz Mart University, Canakkale, Turkey
5 Astrophysics Research Center and Ulupınar Observatory, Canakkale Onsekiz Mart University, Canakkale, Turkey
6 Akdeniz University, Faculty of Sciences, Department of Space Sciences and Technologies, 07058, Antalya, Turkey

ABSTRACT
Using high-precision astrometric data from Gaia DR3 and updated systemic velocities from the litera-
ture, kinematical properties of cataclysmic variables (CVs) were investigated. By constraining the data
according to the total space velocity error and Galactic population class, a reliable sample of data was
obtained. Non-magnetic CVs located in the thin disk have been found to have a total space velocity
dispersion of σν = 46.33 ± 4.23 km s−1 , indicating that the thin disk CVs with a mean kinematical
age of τ = 3.95 ± 0.75 Gyr are much younger than the local thin disk of the Galaxy with τ ∼6-9
Gyr. Total space velocity dispersions of non-magnetic CVs belonging to the thin disk component of
the Galaxy were found to be σν = 47.67 ± 3.94 and σν = 44.43 ± 4.33 km s−1 for the systems below
and above the orbital period gap, respectively, corresponding to kinematical ages of τ = 4.19 ± 0.71
and τ = 3.61 ± 0.74 Gyr. γ velocity dispersions of the thin disk CVs below and above the gap were
obtained σγ = 27.52 ± 2.28 and σγ = 25.65 ± 2.44 km s−1 , respectively. This study also shows that the
orbital period is decreasing with increasing age, as expected from the standard theory. The age-orbital
period relation for non-magnetic thin disk CVs was obtained as dP/dt = −2.09(±0.22) × 10−5 sec
yr−1 . However, a significant difference could not be found between the γ velocity dispersions of the
systems below and above the gap, which were calculated to be σγ = 27.52 ± 2.28 and σγ = 25.65 ± 2.44
km s−1 , respectively.

Keywords: Star: Cataclysmic binaries, Galaxy: Stellar dynamics and kinematics: Galaxy: Solar neigh-
borhood

1. INTRODUCTION of their complex evolutionary pathways, their variabil-

Cataclysmic variables (CVs) are a class of binaries ity, and the information they provide about binary star
consisting of a white dwarf primary and a secondary, evolution and accretion processes (Warner 1995; Hellier
typically a late-type main sequence star. The secondary 2001; Giovannelli 2008; Knigge 2011).
fills its Roche lobe and transfers material to the white The fundamental nature of CV systems is intrinsically
dwarf through an accretion disk, leading to a variety of linked to their binary configuration. Mass transfer from
phenomena such as novae, dwarf novae, and nova-like the donor star to the white dwarf is driven by the an-
variables with irregular changes in brightness due to the gular momentum loss reducing the distance between the
thermal instability of the accretion disk (Warner 1995; primary and the secondary and causing the Roche lobe
Hellier 2001; Takata et al. 2022). CVs are a major fo- overflow of the secondary star (Knigge 2011). Magnetic
cus of interest in the field of stellar astrophysics because braking and gravitational radiation could be the two
mechanisms to maintain such an angular momentum
loss associated with the decreasing of the system size.
Corresponding author: Remziye Canbay It has been well established that gravitational wave ra-
[email protected] diation leads to angular momentum loss in CVs at short
orbital periods, and some form of magnetic braking at
2 Canbay et al.

longer periods (Rappaport et al. 1983; Schreiber et al. brightness (Pretorius et al. 2007). Despite this, the age
2016). Then, the mass transfer, which could be spo- distribution of CVs remains unaffected by brightness-
radic or continuous, leads to the rich phenomenology related selection biases (Kolb 2001), as the age of a CV
associated with CVs, including outbursts and periodic does not impact its mass transfer rate for a given orbital
variability (Patterson 1984). period.
The number of CVs appears to decrease in the orbital The modelled population of Galactic CVs based on
period range of 2 < Porb (h) < 3, which is known as the the standard formation and evolution model predicted
period gap. The standard evolutionary model of CVs that systems above the period gap have ages younger
successfully explains why such a period gap exists. In than 1.5 Gyr, with an average age of 1 Gyr for them,
addition, population synthesis studies based on the stan- while the mean age of the ones below the period gap is
dard model predict an accumulation of systems below 3-4 Gyr, with all ages above about 1 Gyr (Kolb & Stehle
the period gap, as the CVs spend most of their lifetime 1996; Ritter & Burkert 1986). Kolb & Stehle (1996) pre-
below the gap (Kolb & Baraffe 1999; King & Schenker dicted dispersions of the γ velocities to be σ(γ) ≃ 15 and
2002; Littlefair et al. 2008). Indeed, SDSS revealed σ(γ) ≃ 30 km s−1 for the systems above and below the
many systems near the period minimum (Gänsicke et al. period gap, respectively, where γ is the centre of mass
2009), though only a handful of systems past the pe- radial velocity of a CV. They used the empirical rela-
riod minimum, known as period bouncers, have been tion between age t and space velocity dispersion σ(t)
seen (Inight et al. 2023). However, there is a disagree- of field stars in the solar neighborhood (Wielen 1977;
ment between observed and predicted minimum orbital Wielen et al. 1992), σ(t)= k 1 + k 2 t1/2 (k 1 and k 2 are
period values, 76-82 (see Kalomeni et al. 2016, and ref- constants), to estimate the dispersions of γ velocities.
erences therein) and 65-70 minutes (Knigge et al. 2011; According to Kolb (2001), the age difference of systems
McAllister et al. 2019), respectively. Another problem above and below the orbital period gap should be mainly
is about white dwarf masses. Single white dwarf masses due to time spent evolving from the post-common en-
are smaller than those of white dwarfs in CVs (see Zoro- velope (post-CE) orbit into contact. Kolb’s (2001) an-
tovic & Schreiber 2020, and references therein), while other prediction is that the γ velocity dispersions of CVs
the white dwarf mass in CVs does not depend on the are σ(γ) ≃ 27 and σ(γ) ≃ 32 km s−1 for the systems
orbital period (McAllister et al. 2019; Pala et al. 2022). above and below the gap, respectively, if magnetic brak-
In addition, the observed space density of these systems ing does not operate in the detached phase during the
is in agreement only with the lower limit of the predic- evolution.
tions of the standard model (Canbay et al. 2023; Ro- However, observed γ velocities for the systems col-
driguez et al. 2024, and references therein). In a recent lected from published radial velocity studies do not show
study, Schaefer (2024) investigated the orbital period such a difference between velocity dispersions of the CVs
changes of CVs and found that nearly half of the ob- above and below the period gap (van Paradijs et al.
jects have increasing orbital periods. This is impossible 1996). Precise γ velocity measurements for only four
in the magnetic braking model which is used as the angu- dwarf novae above the period gap performed by North
lar momentum loss mechanism for the objects above the et al. (2002) suggest a velocity dispersion of ∼8 km s−1 .
orbital period gap in the standard evolutionary model Ak et al. (2010) estimated kinematical ages of the non-
of CVs. Thus, Schaefer (2024) claimed that the mag- magnetic CVs as 5.01±1.48 and 3.65±1.34 Gyr for the
netic braking model fails in explaining the orbital period systems below and above the gap. Although their dif-
changes of these systems. King & Lasota (2024), how- ference is in agreement with that predicted by Kolb &
ever, proposed a counterargument that short-term phe- Stehle (1996), these ages are older than expected. Ak
nomena can result in the observed period changes. Ad- et al. (2010) also calculated γ velocity dispersions of σγ
ditional angular momentum loss mechanisms were also = 26±4 and σγ = 30±5 km s−1 for non-magnetic sys-
introduced (Andronov et al. 2003; Taam & Spruit 2001; tems above and below the gap, respectively. The obser-
Schenker et al. 2002; Willems et al. 2005, 2007; Knigge vational difference between these γ velocity dispersions
2011; El-Badry et al. 2022). Note that evolutionary sce- is smaller than that expected from the modelled popu-
narios of magnetic CVs may be different from scenarios lation study of CVs made by Kolb & Stehle (1996). It is
of non-magnetic CVs (see the references in Belloni et al. interestingly in agreement with the above prediction of
2020). Kolb (2001). Ak et al. (2015) increased the number of
CV evolution models are often tested through pho- systems with γ velocities by collecting data from the lit-
tometric data sets, though these are typically skewed erature and performing a new kinematic analysis. Their
due to selection biases, particularly those influenced by difference from Ak et al. (2010) was not only the num-
 Kinematic Properties of CVs 3

ber of systems used but also estimating kinematic pa- (µα cos δ, µδ ), trigonometric parallaxes (ϖ), and the ra-
rameters for CVs belonging to the thin disk component dial velocities. The systems classified as CV with known
of the Galaxy, which constitute a more homogeneous γ velocities have been compiled from the literature,
group. Ak et al. (2015) estimated mean kinematical while astrometric data were retrieved from the Gaia
ages of the non-magnetic thin disk CVs as 3.40±1.03 DR3 (Gaia Collaboration et al. 2023). Orbital periods
and 3.90±1.28 Gyr for the systems below and above the were collected from Ritter & Kolb (2003, Edition 7.24),
period gap. This estimation is not in agreement with Downes et al. (2001) and literature.
that found by Ak et al. (2010) for the systems below the
gap. Ak et al. (2015) calculated γ velocity dispersions of 2.1. Systemic Velocities
σγ = 24.95±3.46 and σγ = 26.60±4.48 km s−1 for non-
The space velocity of a star is computed using its ra-
magnetic thin disk systems below and above the gap,
dial velocity with respect to the Sun, which is measured
respectively, which are very similar to each other. They
from Doppler shifts of spectral lines. The systemic ve-
also found γ velocity dispersions of σγ = 32.10±4.41 and
locities (γ) of binary stars are used and calculated from
σγ = 25.92±4.03 km s−1 for magnetic and non-magnetic
the radial velocities. To predict the systemic velocity of
thin disk CVs, respectively, corresponding mean kine-
a CV, the function Vr (ϕ) = γ + K1,2 sinϕ is preferred as
matical ages of 5.64±1.39 and 3.69±1.22 Gyr.
their orbit is circular. Here, ϕ is the orbital phase, γ is
The advent of the Gaia mission (Gaia Collaboration
the centre of mass radial velocity of a CV, K1,2 represent
et al. 2016) has revolutionized our understanding of stel-
semi-amplitudes of radial velocity variation and 1 and
lar populations in the Milky Way. With its unprece-
2 denote primary (white dwarf) and secondary (donor)
dented astrometric precision, Gaia has provided de-
components of the system, respectively. Systemic veloc-
tailed and accurate trigonometric parallax/proper mo-
ity γ is generally found by a non-linear least-square fit
tion measurements, and photometric data for millions
of this function to the observed radial velocities.
of stars, including CVs. This wealth of data has opened
The first collection for systemic velocities of CVs from
new avenues for studying the kinematics of CVs, al-
the literature was created by van Paradijs et al. (1996)
lowing for more accurate distance determinations and
for CVs known up until 1994. Their study was followed
motion analyses within the solar neighborhood (Gaia
by Ak et al. (2010) who collected systemic velocities
Collaboration et al. 2018, 2021, 2023). These data can
of CVs from the literature up to the year 2007 in a
be used to refine the space velocity dispersion of CVs
similar style. Ak et al. (2015) increased the number
and compare them with the kinematics of field stars
of systemic velocities of CVs and analysed their data.
and other stellar populations. The kinematic ages of
Ak et al. (2010, 2015) used distances estimated from
CVs can be estimated by comparing their space velocity
PLCs (Period-Luminosity-Colours) relations of Ak et al.
dispersions with those of other stellar populations with
(2007) and Özdönmez et al. (2015), respectively.
known ages. Older populations tend to have higher ve-
In this study, the systemic velocity data for 455 cata-
locity dispersions due to the cumulative effects of grav-
clysmic variables have been collected from the literature,
itational interactions with other stars and molecular
as well as checking previous compilations of van Paradijs
clouds (Wielen 1977). The centre of mass radial velocity
et al. (1996) and Ak et al. (2010, 2015). We adopted
of a CV (γ velocity) is the next important parameter to
very similar criteria defined by Ak et al. (2010) when
estimate its space velocity and it must be collected from
collecting γ velocities. Thus, if there is more than one
the literature.
measurement of γ velocity for a system, we inspected
In this paper, we calculate the space velocities and de-
the studies and found the one with the radial velocity
rive the γ velocity dispersions of CV groups according
curve including more data points that are well spread
to different orbital period regimes and Galactic popu-
over the orbital phase. The γ velocities obtained during
lations to test the model predictions. To do this, we
superoutbursts of SU UMa-type dwarf novae were not
have collected γ velocities of these objects from the lit-
taken into account. We adopted the γ velocity recom-
erature and used precise equatorial coordinate, trigono-
mended by a researcher when more than one γ velocity
metric parallax and proper motion data from Gaia data
determination was presented in a study.
release 3 (Gaia DR3, Gaia Collaboration et al. 2023).
It should be noted that motions in the accretion disk
or the matter stream falling on the disk from the sec-
2. DATA ondary component of the system probably affect the
Calculating the spatial velocity components of stars spectral lines. Since Doppler shifts are measured from
requires the equatorial (α, δ) and Galactic (l, b) coor- these lines, γ velocities determined from them may not
dinates of the objects, the proper motion components be reliable. Although radial velocities measured from
4 Canbay et al.

absorption lines represent the system best, they are

weak lines and thus can not be observable in all CVs
(North et al. 2002), as they are most likely produced in
the secondary star’s atmosphere. Furthermore, there are
cases (e.g. Schwope et al. 2011) in which the donor star’s
spectral line traces the irradiated surface of the donor
star, not the centre of mass. That affects the amplitude
of the radial velocity curve, but not the γ velocity.
To understand the impact of using measurements from
emission and absorption lines on the results, we com-
pared γ velocity measurements from emission (γems ) and
absorption (γabs ) lines. We found 44 CVs, for which
γems and γabs measurements are both available. These
γ values are compared in Figure 1. As shown in Figure
1, the systemic velocities calculated from emission and
absorption lines are consistent within 1σ errors, in gen-
eral. The average difference of γems and γabs values is
⟨γems − γabs ⟩ = +17.4 ± 25.6 km s−1 , where the error
is the standard deviation of individual differences. This
average value is in agreement with those calculated by
van Paradijs et al. (1996), Ak et al. (2010) and Ak et al.
(2015), +2.5±13.8, +0.57±14.3 and -3.5±22.6 km s−1 ,
respectively, within errors. We conclude that meaning-
ful statistics can be obtained from the observed systemic
velocities in our sample. Based on these comparisons, al- Figure 1. (a) Comparison of systemic velocities measured
though γ velocity measurements obtained from absorp- from emission (γems ) and absorption (γabs ) lines for 44 CVs
tion lines were preferred during the compilation, we also in the sample. Solid black line and blue dotted lines represent
collected γems velocities to construct our data sample. γ the linear fit and 1σ limits, respectively. The red solid line
velocities and associated uncertainties are listed in Ta- in the upper panel is the one-to-one line. (b) Distribution of
ble 1. Figure 3a shows an error histogram of γ veloc- velocity differences. Red and black lines show zero level and
ities, where intermediate polars (IPs) and polars (Ps) mean value of velocity differences, respectively.
are shown separately. The median value and standard
deviation of γ velocity errors are 5.0 and 6.2 km s−1 for IPs, hereafter referred to as mCVs) and non-magnetic
all CVs, respectively. systems (hereafter referred to as non-mCVs) depending
on their classification in the catalogues (e.g. Ritter &
Kolb 2003) and their studies in the literature.
2.2. Astrometric Data
Since the examined systems are located from about
Equatorial coordinates (α, δ), proper-motion compo- 45 pc to 6 kpc as seen in Figure 2, the trigonometric
nents (µα cos δ, µδ ) and trigonometric parallaxes (ϖ) parallax data may be affected by some biases. To over-
for 455 CVs with known systemic velocities were re- come this issue, Bailer-Jones et al. (2021) recalculated
trieved from the Gaia DR3 (Gaia Collaboration et al. the distances and their errors for all stars with known
2023) database. 10,832 objects were identified as Gaia trigonometric parallaxes using Bayesian statistics
CVs in this catalogue. To do this, we matched and a Galaxy model. Thus, we matched our catalogue
our catalogue with the Gaia DR3 catalogue and also with Bailer-Jones et al. (2021) using the Gaia DR3
found the Gaia ID for each CV. For these systems, IDs to obtain precise distances and distance errors for
phot g mean flux over error, fG /δfG , > 8, more than the 432 CVs in our sample. Figure 2 shows a comparison
eight “good” astrometric observations, as character- of distances calculated from Gaia DR3 (Gaia Collabora-
ized by Gaia DR3, astrometric excess noise<2, and tion et al. 2023) database and Bailer-Jones et al. (2021)
ϖ > 0.1 mas were considered. Although we have col- parallaxes. As can be seen from Figure 2, distances
lected γ velocities of 455 systems from the literature, found from the two methods are in good agreement up to
only 432 of them have Gaia data as a result of the above about 2 kpc, while distances obtained from Bailer-Jones
criteria used in matching with the Gaia DR3 database. et al. (2021) are smaller than those of Gaia database
We classified CVs in the sample as magnetic (Polars and
 Kinematic Properties of CVs 5

Figure 2. Comparison of distances calculated from Gaia

DR3 (Gaia Collaboration et al. 2023) database and Bailer-
Jones et al. (2021) parallaxes for 432 CVs in the sample. The
distribution of differences is given in the lower panel.

for beyond about 2 kpc. As distances of almost all sys-

tems in the sample are smaller than 2 kpc and they are
Figure 3. Distribution of γ velocity errors (a), relative par-
estimated from precise trigonometric parallaxes, we con- allax errors (b) and relative proper motion errors (c) of the
cluded that our sample is reliable concerning distances. CV sample. P and IP denote polars and intermediate polars,
Gaia distances of V909 Sgr, CQ Vel and V373 Sct are respectively.
larger than those estimated by Bailer-Jones et al. (2021).
An inspection shows that G > 18.5 mag and | b |< 10◦ .5 CVs in our sample are listed in Table 1. Histograms of
for these systems, where b is the Galactic latitude. Thus, relative parallax errors and proper motion errors of the
disagreement between distances obtained from the two CV sample are shown in Figure 3b and 3c, respectively,
methods for these systems may be due to the use of con- where IPs and Ps are shown separately. As magnetic
stant Galactic length scale in Bailer-Jones et al. (2021) systems are relatively faint objects, CVs close to the
(see also Canbay et al. 2023). Indeed, Bilir et al. (2006a) Sun could be detected in general. As magnetic systems
and Yaz Gökçe et al. (2015) showed in their star counts tend to be faint, those in the solar neighborhood tend
for star fields at the same Galactic latitude but different to be predominantly detected. Nearby systems tend to
Galactic longitudes that the length scale values of the have small parallax errors, though even faint systems
thin disk are in a wide range and that the length scale beyond a few hundred pc become hard to detect. The
is a function of Galactic longitude. median value and standard deviation of relative paral-
In this study, to minimize the bias in the trigonometric lax errors are calculated at 0.028 and 0.070 for all CVs
parallax data of the 432 systems, the distances and dis- in the sample, respectively. The median value and stan-
tance errors provided by Bailer-Jones et al. (2021) were
used. Proper motion components and distances of the
6 Canbay et al.

dard deviation of relative proper motion errors are 0.080

and 0.143 for all CVs, respectively.

2.3. Galactic Space Velocities

The space velocities relative to the Sun were calcu-
lated using the algorithms and transformation matri-
ces developed by Johnson & Soderblom (1987). Equa-
torial coordinates (α, δ), proper-motion components
(µα cos δ, µδ ), distances (dBJ , Bailer-Jones et al. 2021),
and systemic velocities (γ) in Table 1 were used as the
main input data. The transformation matrices of John-
son & Soderblom (1987) use the right-handed system
notation. Consequently, the space velocity components
U , V , and W represent the velocity vector of a star
relative to the Sun. U is directed toward the Galactic
Figure 4. The differential rotation corrections dU (a) and
Center (l = 0o , b = 0o ), V towards the Galactic rotation dV (b) applied to the U and V space velocity components
direction (l = 90o , b = 0o ), and W towards the North of the 432 CVs in the sample as a function of the Galactic
Galactic Pole (b = 90o ), where l and b are the Galactic longitude .
longitude and latitude, respectively.
As the distances of CVs in the sample are in a wide also displayed. We determined from this histogram
range from about 45 pc to 6 kpc, a correction for the that the total space velocity errors corresponding to
differential rotation of the systems around the Galactic 50%, 68%, 90% and 95% of the sample are Serr = 6.05
centre is required. We applied the procedure of Mihalas km s−1 , Serr = 9.08 km s−1 , Serr = 16.02 km s−1 and
& Binney (1981) to the distribution of the CVs and es- Serr = 20.01 km s−1 , respectively.
timated the first-order Galactic differential rotation cor- In order to increase the reliability of space velocities to
rections for the U and V space velocity components of indicate kinematical ages, the dispersion is expected to
the sample. The ranges of these corrections are −122 < be larger than the propagated errors. Mean values and
dU < 118 and −8 < dV < 12 km s−1 for U and V space dispersions of the space velocity components and their
velocity components, respectively. The W velocity is dispersions, total space velocity dispersions, and kine-
not affected in this first-order approximation. Variation matical ages for the CV groups classified according to
of the differential correction values applied to the U and the total space velocity error are listed in Table 3. The
V space velocity components of the CVs in the sample space velocity dispersion of the sample increases as stars
with Galactic longitude is shown in Figure 4. As ex- with large space velocity errors are taken into account,
pected, U is affected more than the V component. The which leads to unreliable calculation of group ages. The
high values for the U component show that corrections U − V and W − V diagrams of the four groups separated
for differential Galactic rotation cannot be ignored. In according to their total space velocity errors are shown
addition, the Galactic space velocity components were in Figure 6a and b, while Figure 6c and d displays the
reduced to the Local Standard of Rest (LSR) by adding U − V and W − V diagrams of non-magnetic systems,
the Sun’s space velocity to the velocity components of polars and IPs. Figure 6a and b reveals that systems
the CVs. The space velocity of the Sun is adopted with large space velocity errors have larger space ve-
as (U, V, W )⊙ = (8.83 ± 0.24, 14.19 ± 0.34, 6.57 ± 0.21) locity components, while magnetic systems were mostly
km s−1 (Coşkunoǧlu et al. 2011). Galactic space veloc- found in lower space velocity regions of the diagrams
ity components corrected for the LSR are listed in Table compared to non-magnetic systems. We removed sys-
2. tems with Serr > 16.02 km s−1 to increase the reliability
Uncertainties of the space velocity components were of the analysis, reducing the number of CVs in the final
determined through the propagation of uncertainties sample to 385.
from the input data, employing the algorithm provided
by Johnson & Soderblom (1987). p Uncertainties of the
total space velocities Serr = Uerr 2 + V 2 + W 2 were 2.4. Velocity Dispersions and Kinematical Ages
err err
also computed. A histogram depicting the total space Estimating the kinematical age of subsamples ex-
velocity errors of 432 CVs was constructed and shown tracted from a large sample is important for determining
in Figure 5, where their cumulative distribution was the kinematical properties of a group of objects. The
 Kinematic Properties of CVs 7

Table 1. Equatorial coordinates (α, δ), type (mCV or non-mCV), orbital period (Porb ), trigonometric parallax (ϖ), distance
from ϖ, proper motion components (µα cos δ, µδ ), and distance from Bailer-Jones et al. (2021) (dBJ ), and systemic velocity (γ)
of CVs in the sample. The reference for γ velocity is given in the last column. The table can be obtained electronically.
ID Star Name αJ2000 δJ2000 Type Porb ω dϖ µα cos δ µδ dBJ γ Ref
(hh:mm:ss) (dd:mm:ss) (d) (mas) (pc) (mas yr−1 ) (mas yr−1 ) (pc) (km s−1 )
001 WW Cet 00 11 27.77 -11 28 43.10 non-mCV 0.176 4.541±0.030 220±2 12.458±0.031 13.510±0.022 218±1 5.8±3.7 (001)
002 1RXS J001538.2+263656 00 15 38.23 +26 36 56.81 non-mCV 0.102 1.804±0.107 554±33 22.390±0.098 -10.733±0.103 558±35 -36±3 (002)
003 V513 Cas 00 18 14.91 +66 18 13.65 non-mCV 0.216 1.167±0.030 857±22 6.034±0.030 -1.398±0.032 835±20 -35.9±0.4 (003)
004 V592 Cas 00 20 52.24 +55 42 16.20 non-mCV 0.115 2.145±0.017 466±4 -7.133±0.014 -14.501±0.015 460±3 21±14 (004)
005 FL Psc 00 25 11.11 +12 17 12.38 non-mCV 0.057 6.330±0.103 158±3 -70.239±0.130 -37.349±0.085 158±3 -17±3 (005)
... ... ... ... ... ... ... ... ... ... ... ... ....
... ... ... ... ... ... ... ... ... ... ... ... ....
... ... ... ... ... ... ... ... ... ... ... ... ....
428 V378 Peg 23 40 04.33 +30 17 47.70 non-mCV 0.139 1.014±0.021 986±21 -5.711±0.024 -5.160±0.014 949±17 -3.4±2 (139)
429 1RXS J234015.8+764207 23 40 20.65 +76 42 10.50 mCV 0.154 1.788±0.052 559±16 19.682±0.054 2.412±0.050 552±15 -21±8 (033)
430 HX Peg 23 40 23.69 +12 37 41.71 non-mCV 0.200 1.656±0.031 604±11 -14.502±0.036 6.804±0.023 586±11 29±6 (140)
431 V630 Cas 23 48 51.90 +51 27 39.10 non-mCV 2.564 0.381±0.037 2623±257 -0.863±0.032 -4.095±0.031 2409±247 -87.3±2.3 (141)
432 Gaia 14ade 23 50 52.01 +28 58 59.50 non-mCV 0.054 0.419±0.251 2389±1431 -5.602±0.236 -5.783±0.142 1907±583 -1.3±0.4 (008)

(001) Ringwald et al. (1996), (002) Thorstensen et al. (2016), (003) Szkody et al. (2013), (004) Huber et al. (1998), (005)
Thorstensen et al. (2016), ..., (008) Szkody et al. (2018), (133) Halpern & Thorstensen (2015), (139) Ringwald & Velasco (2012),
(140) Ringwald (1994), (141) Orosz et al. (2001)

Table 2. Sun-centred rectangular Galactic coordinates (X, Y , and Z), space velocity components corrected for the LSR (ULSR ,
VLSR , and WLSR ) and T D/D values for CVs in the sample. The table can be obtained electronically.
ID Star Name l b dBJ X Y Z ULSR VLSR WLSR T D/D
(o ) (o ) (pc) (pc) (pc) (pc) (km s−1 ) (km s−1 ) (km s−1 )
001 WW Cet 90.01 -71.74 218±1 0 68 -207 -11.12±0.09 21.50±1.16 2.87±3.51 0.009
002 1RXS J001538.2+263656 113.09 -35.57 558±35 -178 418 -325 -31.61±2.60 -53.68±3.32 -3.45±2.54 0.070
003 V513 Cas 119.60 3.65 835±20 -411 724 53 -13.41±0.52 -26.84±0.43 -4.26±0.24 0.009
004 V592 Cas 118.60 -6.91 460±3 -219 401 -55 6.51±6.65 39.09±12.20 -25.28±1.70 0.069
005 FL Psc 112.91 -50.07 158±3 -39 93 -121 68.23±1.31 12.56±1.78 7.23±2.31 0.027
... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ...
428 V378 Peg 105.13 -30.10 949±17 -214 792 -476 20.46±0.75 13.39±1.67 -3.78±1.03 0.007
429 1RXS J234015.8+764207 118.77 14.41 552±15 -257 469 137 -40.00±3.92 -25.83±6.82 -6.78±2.01 0.014
430 HX Peg 97.23 -46.67 586±11 -51 399 -426 21.73±0.71 61.80±4.12 8.33±4.38 0.293
431 V630 Cas 113.10 -10.21 2409±247 -930 2181 -427 5.49±2.25 -61.39±2.09 -20.23±3.91 0.207
432 Gaia 14ade 107.29 -32.04 1907±583 -481 1545 -1012 33.12±15.48 14.60±1.10 -23.60±7.28 0.022

basis of kinematic age determination lies in calculating

reliable space velocity dispersions and systemic velocity
dispersions of the star groups under investigation. By
determining the group space velocity dispersions, age
can be determined using the age-space velocity disper-
sion relationship. In this study, the age-space velocity
dispersion relationship described by Cox (2000) is uti-
lized:

   
3 τ
σν3 (τ ) = 3
σv,τ =0 + σV δ2 Tδ exp −1 (1)
2 Tδ

Here, σν,τ =0 represents the total space velocity disper-

sion at zero age and is generally assumed to be 10 km s−1
(Cox 2000). σV describes the rotation curve, typically
taken as ≈2.95. Tδ is a timescale of 5 Gyr, and δ2 is
Figure 5. (a) Cumulative distribution and (b) histogram of the diffusion coefficient of 3.7 × 10−6 km s3 yr−1 . σν
the total space velocity errors of 432 CVs in the sample. and τ are the total space velocity dispersion and the
kinematic age of the considered CV group, respectively.
8 Canbay et al.

Table 3. Mean space velocity components (⟨ULSR ⟩, ⟨VLSR ⟩ and ⟨WLSR ⟩) with their errors, space velocity dispersions (σU , σV
and σW ), total space velocity dispersions (σν ), kinematical ages (τ ) and γ velocity dispersions (σγ ) for CV groups. N is the
number of systems.

Group N ⟨ULSR ⟩ ⟨VLSR ⟩ ⟨WLSR ⟩ σU LSR σV LSR σW LSR σν τ σγ

−1 −1 −1 −1 −1 −1 −1
(km s ) (km s ) (km s ) (km s ) (km s ) (km s ) (km s ) (Gyr) (km s−1 )

Preliminary sample of 432 CVs

−1
Serr ≤ 6.05 (km s ) 214 0.76±1.98 -5.46±2.09 -3.28±1.60 35.47±1.24 29.35±1.35 21.02±1.12 50.61±2.15 4.72±0.39 29.22±1.24
−1
Serr ≤ 9.08 (km s ) 291 0.32±2.59 -5.25±2.71 -2.16±2.05 36.15±1.80 29.89±1.84 22.37±1.48 51.97±2.97 4.97±0.54 30.00±1.72

Serr ≤ 16.02 (km s−1 ) 385 1.08±3.65 -6.52±3.55 -2.75±2.78 38.66±2.80 29.90±2.67 23.64±2.31 54.29±4.51 5.40±0.83 31.34±2.61
Serr ≤ 20.01 (km s−1 ) 407 1.61±3.98 -6.97±3.94 -2.88±3.06 39.52±3.24 31.36±3.27 23.63±2.75 55.71±5.36 5.66±0.98 32.16±3.10

All Systems 432 2.30±5.01 -10.01±4.98 -3.70±3.68 44.03±5.85 40.79±7.62 26.22±5.02 65.50±10.84 7.43±1.92 37.82±6.25

Final sample of 385 CVs with Serr ≤ 16.02 km s−1

T D/D ≤ 0.1 307 -0.74±0.53 -4.46±3.44 -0.33±2.54 30.84±2.74 20.47±2.63 15.72±2.18 40.21±4.38 2.89±0.72 23.22±2.52

0.1 < T D/D ≤ 1 47 10.47±4.14 -10.19±4.04 -6.18±3.31 50.78±3.34 40.71±2.65 30.03±2.27 71.68±4.83 8.51±0.83 41.38±2.79
1 < T D/D ≤ 10 13 -9.84±3.46 -18.54±4.55 -23.86±3.52 56.38±2.44 50.22±3.85 34.33±2.19 82.94±5.06 10.37±0.80 47.89±2.92

T D/D > 10 18 15.39±4.52 -21.49±3.63 -21.15±5.19 78.33±2.25 74.93±2.37 61.67±2.89 124.71±4.36 16.03±0.50 72.00±2.52

All CVs 385 1.08±3.65 -6.52±3.55 -2.75±2.78 38.66±2.80 29.90±2.67 23.64±2.31 54.29±4.51 5.40±0.83 31.34±2.61

T D/D ≤ 1 354 0.75±3.61 -5.22±3.52 -1.10±2.64 34.37±2.83 24.23±2.64 18.39±2.21 45.90±4.46 3.87±0.79 26.50±2.58

Final sample of 354 CVs with Serr ≤ 16.02 km s−1 and T D/D ≤ 1

Porb < 2.15 (h) 139 0.39±3.62 -7.74±3.50 -1.06±2.98 35.09±2.61 25.22±2.54 19.58±2.07 47.44±4.19 4.15±0.75 27.39±2.42
Porb > 3.18 (h) 188 1.16±3.54 -3.21±3.39 -2.13±2.31 33.45±2.93 22.83±2.51 16.49±2.25 43.73±4.47 3.49±0.78 25.25±2.58

mCVs 49 0.71±5.05 -6.38±4.20 4.31±3.18 33.64±3.48 21.84±2.96 14.60±2.82 42.68±5.37 3.31±0.92 24.64±3.10
non-mCVs 305 0.76±3.38 -5.03±3.41 -1.98±2.56 34.49±2.64 24.59±2.57 18.78±2.08 46.33±4.23 3.95±0.75 26.75±2.44

Porb < 2.15 (h) (non-mCVs) 123 1.06±3.42 -7.67±3.47 -1.98±2.95 34.90±2.41 25.98±2.43 19.47±1.96 47.67±3.94 4.19±0.71 27.52±2.28

Porb > 3.18 (h) (non-mCVs) 161 0.75±3.30 -2.95±3.20 -2.73±2.19 33.92±2.79 22.97±2.38 17.20±2.09 44.43±4.22 3.61±0.74 25.65±2.44

0.030 < Porb (d) ≤ 0.072 77 -3.71±3.59 -10.57±3.32 -0.30±2.54 36.44±2.61 25.05±2.48 20.18±1.83 48.61±4.04 4.36±0.73 28.06±2.34

0.072 < Porb (d) ≤ 0.110 75 4.84±3.63 -3.33±3.97 -0.45±3.53 33.25±2.58 25.60±2.89 19.35±2.16 46.21±4.44 3.93±0.79 26.68±2.56
0.110 < Porb (d) ≤ 0.170 86 -0.45±3.77 -6.43±3.50 -1.62±2.55 33.03±3.18 24.03±2.80 20.12±2.54 45.53±4.94 3.81±0.87 26.29±2.85

0.170 < Porb (d) ≤ 0.280 70 3.20±3.48 -3.33±3.58 -2.98±2.37 35.02±2.85 23.18±2.46 14.04±2.18 44.28±4.35 3.59±0.76 25.57±2.51

0.280 < Porb (d) ≤ 0.370 24 0.53±3.37 3.57±2.76 5.28±1.81 35.78±2.45 18.50±1.60 11.62±1.59 41.92±3.33 3.18±0.57 24.20±1.93
0.370 < Porb (d) ≤ 0.550 14 -0.37±3.56 0.87±3.70 -7.76±1.86 31.53±3.57 16.75±2.94 14.18±2.09 38.42±5.08 2.60±0.81 22.18±2.93

0.550 < Porb (d) 08 -0.43±3.85 -11.87±2.94 -0.50±2.56 — — — — — —

−1
Final sample Non-magnetic CVs with Serr ≤ 16.02 km s

T D/D ≤ 0.1 263 -0.60±3.32 -4.52±3.33 -1.27±2.2.46 30.87±2.56 20.82±2.53 15.94±2.04 40.50±4.14 2.94±0.69 23.38±2.39

0.1 < T D/D ≤ 1 42 9.25±3.77 -8.23±3.91 -6.37±3.18 50.87±3.11 40.79±2.73 30.77±2.22 72.10±4.70 8.58±0.80 41.63±2.71
1 < T D/D ≤ 10 8 -8.69±4.02 -48.60±4.87 -11.62±3.67 65.65±2.52 23.11±3.84 37.64±1.77 79.12±5.00 9.76±0.81 45.68±2.89

T D/D > 10 15 17.63±4.83 -14.14±3.49 -22.78±5.04 82.03±2.33 78.64±2.19 61.41±2.57 129.17±4.10 16.54±0.46 74.58±2.36
All non-mCVs 328 1.30±3.46 -6.51±3.45 -3.16±2.70 39.17±2.65 30.10±2.61 23.57±2.17 54.73±4.31 5.48±0.79 31.60±2.49
−1
Final sample Non-Magnetic CVs with Serr ≤ 16.02 km s and T D/D ≤ 1

0.030 < Porb (d) ≤ 0.072 66 -4.05±3.24 -10.92±3.06 -1.51±2.54 35.53±2.35 25.64±2.22 20.10±1.69 48.21±3.65 4.29±0.66 27.83±2.11
0.072 < Porb (d) ≤ 0.110 69 5.27±3.60 -3.32±4.13 -0.93±3.42 33.46±2.48 26.39±2.89 19.47±2.11 46.85±4.35 4.04±0.78 27.05±2.51

0.110 < Porb (d) ≤ 0.170 72 0.20±3.49 -5.54±3.20 -3.16±2.25 33.40±3.06 23.71±2.71 20.88±2.15 45.97±4.62 3.88±0.82 26.54±2.67
0.170 < Porb (d) ≤ 0.280 57 2.90±3.17 -3.02±3.50 -3.99±2.29 35.37±2.51 23.72±2.52 14.49±2.23 44.98±4.20 3.71±0.74 25.97±2.42

0.280 < Porb (d) ≤ 0.370 22 0.52±3.41 3.31±2.78 5.26±1.86 37.32±2.45 18.46±1.66 12.14±1.63 43.37±3.38 3.43±0.59 25.04±1.95

0.370 < Porb (d) ≤ 0.550 12 -0.25±2.82 2.13±3.39 -8.31±1.92 33.89±3.11 15.23±2.25 14.35±2.22 39.83±4.43 2.83±0.73 23.00±2.55
0.550 < Porb (d) 07 -7.66±4.00 -16.10±3.25 0.10±2.83 — — — — — —
 Kinematic Properties of CVs 9

The relation between the total dispersion of space ve- proposed by Bensby et al. (2003, 2005). Their kine-
locity vectors σν and the dispersion of space velocity matic criteria were used to determine the Galactic pop-
components is described as: ulation types of the CVs in our sample. This method
assumes that the space velocity components relative to
σν2 = σU
2 2
+ σV 2
+ σW (2) the Local Standard of Rest (LSR), (U, V, W )LSR , exhibit
a Gaussian-like distribution for the thin disk, thick disk,
After calculating σν2 from the dispersions of velocity and halo populations in the Galaxy:
components using Equation (2), the kinematic age of
a certain CV group can be determined using Equation !
(1). Additionally, under the assumption that CVs are U2 (VLSR − vi,a )2 W2
Pi (U, V, W ) = k exp − LSR2 − 2 − LSR
2
isotropically distributed, the γ velocity dispersion σγ 2σi,U 2σi,V 2σi,W
can also be calculated from the definition σγ2 = 31 σν2 (3)
(Wielen et al. 1992), allowing for comparison with the- where
1
oretical predictions. Total space velocity dispersions, k= (4)
(2π)3/2 σi,U σi,V σi,W
kinematic ages and γ velocity dispersions of CV groups
are listed in Table 3. σU , σV , and σW are the velocity dispersions, with values
for the thin disk (i = D) being 35, 20, and 16 km s−1 ,
2.5. Spatial Distributions respectively; for the thick disk (i = T D), 67, 38, and
35 km s−1 ; and for the halo (i = H), 160, 90, and 90
Distributions of the systems in the final sample of 385
km s−1 (Bensby et al. 2003). The asymmetric drift νa
CVs according to equatorial and Galactic coordinates
values for the thin disk, thick disk, and halo populations
are shown in Figures 7a and 7b. There are 278 and
are -15, -46, and -220 km s−1 , respectively (Bensby et al.
107 CVs with positive and negative declination values,
2003).
respectively, while the numbers of systems with b ≥ 0◦
To determine whether the systems belong to a par-
and b < 0◦ are 221 and 164, respectively.
ticular Galactic population, the probabilities given by
We computed Sun-centred rectangular Galactic coor-
Equation 3 are calculated and multiplied by the pop-
dinates (X towards Galactic Center, Y Galactic rota-
ulation fractions (X) near the Sun. X represents the
tion, Z North Galactic Pole) to investigate the spatial
population ratios of stars near the Sun, with values for
distribution of the CV sample. Figure 8 displays the pro-
the thin disk, thick disk, and halo being XD = 0.9385,
jected positions on the Galactic plane (X, Y ) and the
XTD = 0.06, and XH = 0.0015, respectively (e.g., Buser
Galactic plane perpendicular to it (X, Z). Polars and
et al. 1999; Bilir et al. 2006b,c, 2008; Cabrera-Lavers
IPs are shown separately in Figure 8. Median values and
et al. 2007). The relative probabilities of the thick disk
standard deviations of the X, Y and Z coordinates are -
compared to the thin disk and halo are determined by
82±503, 104±509 and 50±331 pc for all CVs in the final
the following equations:
sample, respectively, -82±504, 111±461 and 73±337 pc
for non-magnetic CVs, and -97±500, 58±727 and 2±287
pc for magnetic CVs. These values show that magnetic TD XT D PT D TD XT D PT D
= × = × (5)
systems have a smaller dispersion in their spatial dis- D XD PD H XH PH
tribution. Figure 8 reveals that most of the systems Bensby et al. (2003) classified stars with T D/D ≤ 0.1 as
(85%) are located closer than about 1 kpc. The median high-probability thin disk stars. According to this classi-
distance of the final sample is 570 pc. These median fication, the stars selected have a likelihood of belonging
values reveal that the sample is well within the Galactic to the thin disk that is 10 times higher than the likeli-
disk in the solar neighborhood. Figure 8 also shows that hood of belonging to the thick disk. Conversely, stars
magnetic systems are located near the Sun, in general. with T D/D > 10 are considered high-probability thick
This is expected since only close magnetic CVs can be disk stars. Stars within the ranges 0.1 < T D/D ≤ 1
detected, as they are relatively faint systems. and 1 < T D/D ≤ 10 are classified as low-probability
thin disk and low-probability thick disk stars, respec-
2.6. Population Types tively.
As the velocity dispersions of object groups from var- For 385 CVs in the final sample, the results based on
ious components of the Galaxy can bias the results, it the kinematic population separation by Bensby et al.
is crucial to determine which Galactic component an (2003, 2005) indicate that 307 systems fall within the
object belongs to. A method for distinguishing between T D/D ≤ 0.1 range, 47 systems within the 0.1 <
Galactic population classes based on kinematic data was T D/D ≤ 1 range, 13 systems within the 1 < T D/D ≤
10 Canbay et al.

Figure 6. U − V and W − V diagrams of CV groups according to their total space velocity errors Serr (a and b), and U − V
and W − V diagrams of magnetic (polars (P) and intermediate polars (IP) and non-magnetic systems in the sample of 432 CVs
(c and d).

10 range, and 18 systems within the T D/D > 10 range (2008), it is concluded that the systems in the sample
(Table 3). Population types (T D/D values) of CVs in are mostly members of the thin disk population, as ex-
the sample are listed in Table 2. This classification pected from the spatial distribution of CVs in the sample
shows that the percentage of high and low probability (Figure 8). The total space velocity dispersions of the
thin disk CVs in the final sample is 92%, while 8% of CV populations are listed in Table 3 together with their
the final sample are the high- and low-probability thick errors. The Toomre diagram of CVs in the final sam-
disk CVs. ple is shown in Figure 9. In this diagram, systems are
Local densities of the Galactic components can be coloured according to their population types.
obtained from different objects in the solar neighbor-
hood. For example, Chen et al. (2001) and Siegel et al. 2.7. Orbital Periods
(2002) found 6.5–13% and 6–10%, respectively, for the An orbital period is the most precisely determined pa-
local space density of the thick disk (see also Karaali rameter for a CV. The orbital period histogram of 385
et al. 2004; Bilir et al. 2006c, 2008; Cabrera-Lavers et al. CVs in the final sample is presented in Figure 10, where
2007, and references therein). Chen et al. (2001) and the orbital period gap between 2.15 and 3.18 hr (Knigge
Siegel et al. (2002) used field stars from The Sloan Dig- 2006; Knigge et al. 2011) makes the period histogram
ital Sky Survey (SDSS, York et al. 2000) and Kapteyn bimodal. Ps, IPs, and non-mCVs are shown separately
selected areas, respectively. These values are in agree- in Figure 10. Orbital period histograms plotted accord-
ment with the thick disk percentage found in this study. ing to population types are shown in Figure 11, where
Considering also the local thick-to-thin disk density nor- the period gap is very prominent for CV groups with
malization of ρthick /ρthin = 12% and local halo-to-thin T D/D ≤ 0.1 and 0.1 < T D/D ≤ 1, high and low-
disk density normalization of 0.05% given by Jurić et al. probability thin disk stars. Hence, as they constitute a
homogeneous sample, we decided to study the kinemati-
 Kinematic Properties of CVs 11

Figure 7. Distribution of 385 CVs in the sample according

to their equatorial (a) and Galactic (b) coordinates. In the
upper panel, the red curve indicates the Galactic plane and
the green star symbol indicates the Galactic center.

cal properties of high and low-probability thin disk CVs

only, for which the total number of systems is 354.
Space velocity dispersions of all CVs and non-
magnetic systems belonging to the thin disk component Figure 8. The spatial distribution of 385 CVs in the sample
with respect to the Sun. X, Y and Z are Sun-centred rect-
of the Galaxy are given in Table 3 for below and above
angular Galactic coordinates. P and IP denote polars and
the orbital period gap. Total space velocity disper- intermediate polars, respectively.
sions of the thin disk CVs below and above the gap are
47.44±4.19 and 43.73±4.47 km s−1 , respectively, where
errors are standard deviations of individual space veloc-
ity errors. When magnetic CVs are removed from the
sample, total space velocity dispersions of non-magnetic
CVs belonging to the thin disk component are found to
be 47.67±3.94 and 44.43±4.33 km s−1 for the systems
below and above the gap, respectively. The total space
velocity dispersion of all magnetic CVs in the thin disk
component is estimated at 42.68±5.37 km s−1 . The to-
tal space velocity dispersion of polars and IPs in the
thin disk were calculated separately and found to be
42.76±5.77 and 41.70±4.91 km s−1 , respectively.
The sample of thin disk CVs including 354 systems
were also investigated for total space velocity dispersion
differences between the various orbital period regimes.
To perform this investigation, the sample was divided
into smaller subsamples of almost a similar number of Figure 9. Toomre diagram of the final CV sample of 385
CVs at different period ranges. Since the number of sys- objects. Systems are colored according to their population
tems is being smaller for periods longer than 0.28 d, the types described in the text.
12 Canbay et al.

2.8. On the Selection Biases and Completeness of the

Sample
Photometric data sets used in testing CV evolution
models are affected by selection biases. The primary
bias comes from the brightness-related selection effects
(Pretorius et al. 2007).
Because it is easier to obtain spectra of brighter stars,
our sample is skewed towards brighter systems. This
results in our sample mostly containing long-period ob-
jects that are close to the Sun, in general, since long-
period CVs are brighter than those with short orbital
Figure 10. Histogram of orbital periods of the final CV periods. Indeed, Table 3 shows that the number of sys-
sample including 385 CVs. The shaded area represents the
tems with P < 2.15 h and P > 3.18 h are 139 and 188,
orbital period gap. The vertical axis is given on a logarithmic
scale since the numbers of polars (P) and intermediate polars respectively. It seems that there is no crucial differ-
(IP) are small compared to non-magnetic systems. ence between the numbers of faint and brighter systems
in the sample. In addition, most systems are located
within 1 kpc of the Sun (see Figure 8). Because there
is a statistically significant number of systems across
orbital periods and magnetic nature (Figure 10), we ex-
pect there are enough systems across all sub-groups for
a meaningful kinematic analysis.
A comparison of the orbital period histograms of the
samples used in Pala et al. (2020, hereafter referred to
as P20), (Rodriguez et al. 2024, hereafter referred to
as R24) and this study is shown in Figure 12. The
orbital period histogram of the catalogue of Ritter &
Kolb (2003) is also shown in the figure. Note that R24
and P20 used volume-limited samples. Figure 12 reveals
that the samples in P20 and R24 are biased towards the
short-period (fainter) systems, while the sample in this
study better represents the CVs with known orbital pe-
riods.
Spatial distribution and Galactic positions of the ob-
jects can also produce a somewhat small bias in the final
sample. This bias was tested using Figure 7 and Figure
8, where positions of CVs are shown according to their
equatorial and Galactic coordinates, and Sun-centred
rectangular Galactic coordinates, respectively. Median
values and standard deviations of the Sun-centred rect-
angular Galactic coordinates X, Y and Z are -82±503,
104±509 and 50±331 pc for all CVs in the final sample,
respectively, while median value and standard deviation
of Galactic latitudes are 4◦ .72 ± 31◦ .68. These values
Figure 11. Histogram of orbital periods of the final CV show that there is no considerable bias according to the
sample. Systems are coloured according to their population spatial distribution of CVs in our study.
types described in the text. Observational studies of CVs suffer from strong selec-
tion biases not only due to brightness and spatial distri-
number of systems is also smaller for the period regimes bution but also due to different phenomenology. It can
above 0.28 d. The total space velocity dispersions, γ be found a detailed discussion of these biases in P20 (see
velocity dispersions, and kinematical ages are given in also R24). When considering brightness-dependent se-
Table 3 for all the thin disk CVs and non-magnetic thin lection biases, which is probably the strongest one, it is
disk CVs grouped into various orbital period intervals. worth paying some more attention to the studies of P20,
 Kinematic Properties of CVs 13

Canbay et al. (2023, hereafter referred to as C23) and P20 and C23, and compared them with the distribution
R24, who estimated space densities of CVs from Gaia of systems in our sample. Expected numbers of objects
DR2, Gaia DR3 and X-ray surveys, respectively. P20 can be estimated by utilizing the exponential density
+0.6
and R24 obtained space densities of 4.8−0.8 ×10−6 pc−3 equation of the thin disk population:
−6 −3
and 3.7(±0.7)×10 pc from their volume-limited CV
samples, respectively, which are in good agreement. 
|z + z0 |
 
(x − R0 )

Their samples include objects within the sphere with D(x, z) = n × exp − × exp −
H h
a radius of 150 pc with the Sun at the centre. P20 (6)
point out that their volume-limited sample is 77±10% where n is the normalized density of CVs in the solar
complete and it must contain 12 more CVs to obtain neighborhood, z and z0 are the distances of the CV being
a 100% complete sample. They give their final space studied and the Sun from the Galactic plane (z0 = 27±4
density estimation as 4.8+0.6
−0.8 ×10
−6
pc−3 by using this pc, Chen et al. 2000), respectively, H and h scale height
assumed 100% complete sample of CVs. and scale length of CVs, R0 solar distance from the
On the contrary, C23 did not use a volume-limited Galactic centre (R0 = 8 kpc, Majewski 1993). x is the
sample. Instead, they limited their sample by defining planar distance of the object from the Galaxy centre and
the completeness limits of the data for certain absolute calculated as,
magnitude intervals. The median value of CV distances
in their sample is 989 pc, with distances extending to 1/2
x = R02 + (z/ tan b)2 − 2R0 (z/ tan b) cos l

5-6 kpc from the Sun. This median value is about 6.5 (7)
times larger than the upper distance limit of the volume-
where l and b are the Galactic longitude and latitude of
limited samples mentioned here. The space density of
+1.3 the object. Since the scale length of CVs is not known,
CVs estimated by C23 is 6.8−1.1 ×10−6 pc−3 , which is
we could estimate the number of CVs from the space
surprisingly in good agreement with those given by P20
densities for l=0◦ and b= 90◦ , for the vertical distances
and R24. This value is even higher than the estimates
from the Galactic plane. Thus, the exponential density
given in the aforementioned studies. It should be noted
equation takes the following form,
that P20 and R24 assumed a scale height of H = 280
pc in their studies, while C23 simultaneously estimated 
|z + z0 |

∼
D(z) = n × exp − . (8)
the space density and scale height (H = 375 pc). H
Since the main purpose of this work is to study the
We estimated the theoretical number of stars for 100 pc
general kinematic properties of CVs rather than focus
intervals up to 1200 pc distance. Since these distances
on 100% completeness, the larger sample of C23 is most
were designed for b = 90◦ and z = d×sin b, we get z = d.
appropriate for this work.
Centroids of distance intervals of 100 pc are defined as
White dwarf kicks can be considered to add some bias
d∗ = z ∗ = (d31 + d32 )/2. In this study, we take d1 = 0 pc.
to the space velocity dispersions. Simulations performed
Then, we determined space densities for each selected
by El-Badry & Rix (2018) show that a modest velocity
distance using space densities in the solar neighborhood
kick with typical velocities of ∼0.75 km s−1 is possible
and scale heights given by P20 and C23. The volume el-
during white dwarf formation, probably due to asym-
ement increases with increasing distance, thus, we used
metric mass-loss. Adding this small velocity to the total
the equation Vi = (4/3)πd3i to calculate the volume cor-
space velocity dispersion will affect the age. For exam-
responding to the distance di . The expected number of
ple, the final sample of 385 CVs in Table 3 has a total
CVs was calculated by multiplying the volume element
space velocity dispersion of 54.29±4.51 km s−1 , corre-
and its related density. The estimated numbers of CVs
sponding to a mean kinematical age of 5.40±0.83 Gyr.
for 100 pc intervals in the direction (l, b) = (0◦ , 90◦ )
An increase of 0.75 km s−1 in this total space velocity
were found for 4.8×10−6 pc−3 and H = 280 pc from P20
dispersion has a very small effect on the total dispersion
and 6.8×10−6 pc−3 and H=375 pc from C23. Cumula-
and age calculations, resulting in 54.04±4.52 km s−1 and
tive numbers of CVs estimated in this way are plotted
5.53±0.83 Gyr. We conclude that the possible bias from
against distance from the Galactic plane (z) in Figure
the white dwarf kicks can be ignored, unless substantial
13. For comparison, we used the CVs with Serr ≤ 16.02
evidence for larger WD kicks during CV formation are
km s−1 in this study. Since unit absolute magnitudes
presented in the future.
(∆MG =1) must be considered in order to estimate the
In order to test the completeness of our sample, we
Galactic model parameters, the absolute magnitude in-
found the expected distributions of the cumulative num-
terval of the sample must be determined. As our calcu-
ber of CVs from the space density estimations given by
lations show that G-magnitude changes between 0 and
14 Canbay et al.

11 mag, the effective unit absolute magnitude interval is

11 mag for the sample. Thus, CV numbers were normal-
ized by dividing them by 11. The calculated cumulative
number of CVs for our sample are shown in Figure 13.
Figure 13 reveals that our sample is not complete due
to brightness-related selection effects, compared to the
expected numbers of CVs from the space densities in
P20 and C23. A statistical study based on this sample
would be biased due to brightness-related selection ef-
fects. However, such effects can not create strong bias
on our sample since we have sufficient systems across
orbital periods and magnetic nature. All we need in
a kinematical analysis is to have a sample including as
many as objects possible, and precisely measured pa-
rameters.

3. DISCUSSIONS
A sample of 432 CVs with systemic velocities and as-
trometric data was collected from literature and Gaia
DR3 (Gaia Collaboration et al. 2023), respectively. In
order to minimize biases in the trigonometric parallax
data of the Gaia DR3, we have used recalculated paral-
laxes in Bailer-Jones et al. (2021) to obtain precise dis-
tances and their errors of the CV sample. We also lim-
ited the preliminary sample of 432 CVs with Serr =16.02
km s−1 , obtaining a final sample of 385 CVs. There
are 1,108 systems with distances calculated from Gaia
DR3 database and Bailer-Jones et al. (2021) parallaxes
in the catalogue of Ritter & Kolb (2003). By taking into
account distance errors, the Ritter & Kolb (2003) cata-
logue includes 42 CVs within the sphere with a radius
of 150 pc with the Sun at the centre, 182 CVs within
300 pc, 390 CVs within 500 pc and 810 CVs within 1
kpc. Our sample of 385 CVs contains 26 CVs (62% of
the Ritter & Kolb (2003) catalogue) within 150 pc, 97
CVs (53%) within 300 pc, 179 CVs (46%) within 500 pc Figure 12. Comparison of orbital period histograms of the
and 332 CVs (41%) within 1 kpc. Although our sample samples used in Rodriguez et al. (2024) (a), Pala et al. (2020)
is not complete for a brightness-related study, kinemat- (b) and this study (c). Orbital period histogram of the cata-
logue of Ritter & Kolb (2003) is shown in the bottom panel.
ical studies can be very weakly affected by brightness-
The shaded area represents the orbital period gap.
related biases (Kolb 2001) (see section 2.8). The median
distance of the final sample is 570 pc. The present sam-
⟨ULSR ⟩ = 1.08 ± 3.65 km s−1 , ⟨VLSR ⟩ = −6.52 ± 3.55 km
ple of this study occupies a much larger space than that
s−1 and ⟨WLSR ⟩ = −2.75 ± 2.78 km s−1 . Space veloc-
in previous studies (van Paradijs et al. 1996; Ak et al.
ity dispersions of this sample were calculated σULSR =
2010, 2015). We believe that this data sample is the
38.66 ± 2.80 km s−1 , σVLSR = 29.90 ± 2.67 km s−1 and
most reliable than ever used for kinematical analysis of
σWLSR = 23.64 ± 2.31 km s−1 km s−1 . Total space
CVs.
velocity dispersion of all CVs in the final sample is
Objects in the final sample are located well within
σν = 54.29 ± 4.51 km s−1 , indicating a γ velocity disper-
the Galactic disk in the solar neighborhood, as median
sion of σγ = 31.34±2.61 km s−1 and a mean kinemat-
values and standard deviations of their X, Y and Z co-
ical age of τ =5.40±0.83 Gyr. These values and scatter
ordinates are -82±503 pc, 104±509 pc and 50±331 pc,
in Figure 6 indicate that there are both younger and
respectively (Figure 8). The mean space velocities with
older systems in the final sample. Thus, in order to ob-
respect to LSR of the final sample were found to be
 Kinematic Properties of CVs 15

27.5% and 10%. As can be seen that these percentages

are not in agreement with those found from the different
objects in the solar neighborhood. Although they found
that the most of the systems in their sample are mem-
bers of the thin disk component of the Galaxy, Ramsay
et al. (2018) concluded that a significant number of AM
CVn stars are likely awaiting to be discovered. We con-
clude that our sample better represents the population
profile of the solar neighbourhood.
The total space velocity dispersions of the CVs in
T D/D ≤ 0.1, 0.1 < T D/D ≤ 1, 1 < T D/D ≤ 10
and T D/D > 10 ranges were calculated as 40.21±4.38,
71.68±4.83, 82.94±5.06 and 124.71±4.36 km s−1 , re-
Figure 13. Comparison of the cumulative number of CVs spectively. The γ velocity dispersions corresponding to
estimated for 100 pc intervals. The numbers were estimated these groups are 23.22±2.52, 41.38±2.79, 47.89±2.92
using the space densities and scale heights given by Pala et al.
and 72.00±2.52 km s−1 , respectively, while kinemat-
(2020) and Canbay et al. (2023). Red circles represent the
data sample in this study. ical ages are 2.89±0.72, 8.51±0.83, 10.37±0.80, and
16.03±0.50 Gyr, respectively. It is clear that CV groups
in Table 3 according to population types exhibit an in-
tain homogeneous object groups and investigate their
crease in dispersions and ages from high probability thin
kinematical properties, it is necessary to form various
disk stars to high probability thick disk stars, as ex-
sub-groups of the final sample.
pected.
3.1. Groups According to Population Types A wide range of ages for the thin and thick disk com-
ponents of the Galaxy is found in various studies. Liu
The Toomre diagram in Figure 9 reveals that kine-
& Chaboyer (2000) found the ages of the thin and thick
matical properties of stars are related to their Galac-
disks to be 9.7±0.6 Gyr and t ∼12.5±1.5 Gyr, respec-
tic populations, as the distribution of objects belonging
tively, while del Peloso et al. (2005) and Grundahl et al.
to different Galactic population components shows very
(2008) estimated 8.8±1.7 and 8-9 Gyr for age of the thin
explicit differences. Thus, it is very important to deter-
disk. In a recent study, Kilic et al. (2017) derived the
mine the Galactic population types of stars in a kine-
ages of 6.8–7.0 Gyr and 8.7±0.1 Gyr for local thin and
matical analysis. We found that about %92 of CVs in
thick disks from the white dwarf luminosity functions,
the final sample belong to the thin disk component of
respectively. Kinematical ages estimated for CV popu-
the Galaxy, while ∼%8 of them belong to the thick disk.
lations are not in accord with those found in previous
Local densities derived from different objects in the
studies based on field stars and clusters. High probabil-
solar neighborhood have a wide range of values. Chen
ity thin disk CVs were found much younger than local
et al. (2001) and Siegel et al. (2002) found 6.5–13%
thin disk. Since white dwarfs in CVs are more mas-
and 6–10%, respectively, for the local space density of
sive than single white dwarfs (see Zorotovic & Schreiber
the thick disk (see also Karaali et al. 2004; Bilir et al.
2020, and references therein), CVs could be younger
2006c, 2008; Cabrera-Lavers et al. 2007, and references
than single white dwarfs due to faster evolution. How-
therein). Jurić et al. (2008) gives the local thick-to-thin
ever, high probability thick disk CVs were found much
disk density normalization ρthick /ρthin = 12%, while
older than local thick disk and this idea does not work
local halo-to-thin disk density normalization is 0.05%.
for high probability thick disk CVs, probably due to
Thus, we concluded that the number density rates of
their number in our sample being only 13 and the re-
thin and thick disk CVs in our study are in agreement
sults for thick disk CVs are unreliable.
with the above values. A similar result can be found
Population analysis and Figure 8 show that almost all
in Ramsay et al. (2018) who used Gaia DR2 parallaxes
of the CVs in the final sample are systems belonging
and proper motions to determine the expected cumula-
to the thin-disk component of the Galaxy. The number
tive distribution of AM CVn stars (a subgroup of CVs
of high and low-probability thin disk CVs in the final
including two white dwarfs in the binary system) and
sample is 354, which constructs a homogeneous sample
compare it with the distribution of a Galactic disk pop-
for kinematical analyses. Thus, high and low-probability
ulation. In their Table B.1, the numbers of thin disk,
thin disk CVs were also considered for further analyses.
thick disk and thin/thick objects are 25, 11 and 4, re-
spectively, corresponding to the percentages of 62.5%,
16 Canbay et al.

3.2. Kinematics of Magnetic and Non-magnetic CVs analysis based on orbital period ranges could be used to
Derived kinematical parameters are listed in Table 3 test evolutionary scenarios.
for magnetic and non-magnetic CVs in the thin disk In order to investigate kinematical differences between
component of the Galaxy. Since the number of mCVs is subgroups of orbital period ranges, systems were first
only 49, they were not divided into sub-groups. There divided into two subgroups according to the orbital pe-
are only 4 polars and 1 intermediate polar belonging to riod gap. Total space velocity dispersions of the thin
the thick disk component of the Galaxy in the sample disk CVs below and above the gap are 47.44±4.19 and
with Serr < 16.02 km s−1 . 43.73±4.47 km s−1 , respectively, corresponding to kine-
The total space velocity dispersion of all thin disk matical ages of 4.15±0.75 and 3.49±0.78 Gyr. When
mCVs is 42.68±5.37 km s−1 , corresponding to a mean magnetic CVs are removed from the sample, total space
kinematical age of 3.31±0.92 Gyr. This parameter was velocity dispersions of non-magnetic CVs belonging to
calculated for non-mCVs as 46.33±4.23 km s−1 , with a the thin disk component are 47.67±3.94 and 44.43±4.33
mean kinematical age of 3.95±0.75 Gyr. It should be km s−1 for the systems below and above the gap, respec-
noted that Ak et al. (2015) calculated these parameters tively, corresponding to kinematical ages of 4.19±0.71
for mCVs and non-mCVs in the thin disk component of and 3.61±0.74 Gyr.
the Galaxy, as done in this study. They estimated total According to population synthesis models of Galactic
space velocity dispersions of mCVs and non-mCVs as CVs based on the standard formation and evolution the-
55.60±7.63 and 44.90±6.97 km s−1 , respectively, with ory, the mean age of systems below the gap is 3-4 Gyr,
mean kinematical ages of 5.64±1.39 and 3.69±1.22 Gyr. while the CVs above the period gap are younger than
A comparison shows that differences in dispersions and 1.5 Gyr (Kolb & Stehle 1996; Ritter & Burkert 1986).
ages between mCVs and non-mCVs in this study are not It is concluded that this age difference is mainly due to
as large as found in Ak et al. (2010, 2015). Moreover, the time spent evolving from the post-CE phase into the
dispersions of mCVs are larger than that of non-mCVs contact phase Kolb (2001). The mean kinematical age
in Ak et al. (2010, 2015), contrary to this study. It calculated in this study for the systems below the gap
was expressed that evolutionary scenarios of magnetic is in agreement with the upper limit of the theoretical
CVs may be different from scenarios of non-magnetic prediction, while the mean kinematical age for the CVs
CVs (see the references in Belloni et al. 2020). Thus, a above the gap is not in agreement with the prediction.
considerable difference is expected between total space Note that the difference between the mean ages of sys-
velocity dispersions of mCVs and non-mCVs. However, tems below and above the period gap is not as large as
the data sample in this study does not show a large expected from the population models of Galactic CVs
difference between total space velocity dispersions, cor- based on the standard formation and evolution theory.
respondingly ages, of mCVs and non-mCVs. Polars and Our results show that the dispersion of CVs below the
intermediate polars represent different phases of CV evo- gap is larger than that of those above the gap. Note that
lution. Thus, we also calculated the total space veloc- this is at most a 2σ result and the measured quantity is
ity dispersions of polars and intermediate polars in the the total space velocity dispersion, not the age. Thus,
thin disk as 42.76±5.77 and 41.70±4.91 km s−1 , respec- expected and observational γ velocity dispersions of sys-
tively, with the ages of 3.32±0.99 and 3.14±0.83 Gyr. tems below and above the gap should be also compared.
The kinematic properties of these magnetic CV groups CVs in the sample were grouped according to orbital
belonging to the thin disk component of the Galaxy do period ranges to investigate the kinematic properties of
not show a significant difference, although we believe these groups. The total space velocity dispersions and
that their numbers in our sample are still small. corresponding kinematical ages of CVs in these period
ranges are listed in Table 3. The continuous decrease
of the dispersion and age from shorter to longer orbital
3.3. Groups According to Orbital Periods periods is remarkable. The results emphasize the impor-
tance of detecting stellar populations in kinematic anal-
An orbital period is the most precisely determined pa-
yses for CV samples. Since the sample including only
rameter of a CV, and it is related to the mass of the sec-
non-magnetic thin disk CVs is supposed to be the most
ondary star (Kalomeni et al. 2016). As the secondary
homogeneous sample, these objects were also grouped
mass and orbital period are age-related quantities, kine-
according to orbital period ranges. The total space ve-
matical parameters derived for the orbital period ranges
locity dispersions and corresponding kinematical ages of
could be meaningful. Since the orbital period evolu-
CVs in these period ranges are listed in Table 3. The
tion of CV samples has been predicted by the standard
same trend, decreasing total space velocity dispersions
theory of evolution of these systems, any kinematical
 Kinematic Properties of CVs 17

from shorter to longer periods, is clearly present in these

data as well.
Thus, we can conclude that the orbital period is de-
creasing with increasing age. The mean kinematical age
increases from 2.83±0.73 to 4.29±0.66 Gyr, while the
orbital period decreases from 0.550 to 0.030 d for the
non-magnetic thin disk systems. This is not in agree-
ment with the age-period relation in Ak et al. (2010),
but it is similar to the relation presented in Ak et al.
(2015). The age-orbital period relations found in this
study are shown in Figure 15a and b for all thin disk CVs
and non-magnetic thin disk CVs, respectively, where red
solid lines represent linear fits. Decreasing rates in Fig- Figure 14. Comparison of age-velocity dispersion relations
of CVs and single white dwarfs. Blue circles represent the
ure 15a and b are dP/dt = −1.82(±0.18)×10−5 sec yr−1
relation obtained from Cheng et al. (2019), while red circles
and dP/dt = −2.09(±0.22)×10−5 sec yr−1 , respectively. show the values found for thin-disk CVs in this study.
Decreasing orbital periods with increasing age is ex-
pected from the standard evolution model of CVs (see
Kalomeni et al. 2016, and references therein). Assum-
ing that the observed period changes in Schaefer (2024)
are due to evolutionary processes, an inspection of CVs
with decreasing orbital period in Table 1 of Schaefer
(2024) shows that mean period decrease of WD-MS sys-
tems with orbital periods 0.0567 ≤ P (d) ≤ 0.618 is
dP/dt = −4.13(±0.44) × 10−6 sec yr−1 . Note that
this period range is consistent with the sample in this
study. The period change rate in our study is about 5
times higher than that of the orbital periods in Schaefer
(2024). However, it is not clear if the observed orbital
period changes in Schaefer (2024) originated from evo-
lutionary processes, since King & Lasota (2024) claims
that the observed period changes must be originated
from short–term phenomena and long-term evolution of
CVs can be constrained only by the observations with
timescales longer than ∼ 3 × 105 yr.
Age-velocity dispersion relation of CVs in the thin-
disk component of the Galaxy can be compared with
that of isolated white dwarfs given by Cheng et al.
(2019, hereafter referred to as C19) (see also Raddi et al.
(2022)) in Figure 14. For comparison, we have taken the
σU , σV , σW and age values of isolated white dwarfs in Figure 15. Orbital period-kinematic age relations found for
C19. Raddi et al. (2022) compared their dispersion-age all CVs (a) and non-magnetic CVs (b) located in the thin-
relation with that of C19 in their Figure 11, in which disk component of the Galaxy. Red solid lines represent the
the relations in C19 are extrapolated in order to remove linear fits.
the thick disk contribution for ages higher than ∼7 Gyr.
We used these extrapolated relations to obtain the total dwarfs. However, considering that the age-dispersion re-
space velocity values and ages in C19. Using the ex- lation in C19 includes both the thin and thick disk CVs,
trapolated relations, we calculated total space velocity while the relation in this study represents only thin-disk
dispersions (σν ) of single white dwarfs. Figure 14 shows systems, disagreement between the two relations should
that the age-dispersion relation of CVs is not in agree- be less than in Figure 14.
ment with that of isolated white dwarfs, although the
youngest CVs (long-period systems) are partly in agree-
ment. It seems that CVs are younger than isolated white
18 Canbay et al.

3.4. The γ Velocity Dispersions the period gap, while there is only a 3σ disagreement
The γ velocity dispersions of CV sub-groups can be for the CVs below the gap.
used to test predictions of the standard formation and
evolution theory of CVs. The γ velocity dispersions of 4. CONCLUSIONS
mCVs and non-mCVs belonging to the thin disk com- With its most precise distances and proper motions
ponent of the Galaxy are very similar, σγ = 24.64 ± 3.10 data from Gaia DR3, and systemic velocities collected
km s−1 and σγ = 26.75±2.44 km s−1 , respectively. Note from the literature, the most reliable ever data sample
that this is a 1σ result. of cataclysmic variables was constructed for their kine-
When it is calculated only for thin disk non-mCVs, matical analysis. The sample is comprised of systems
γ velocity dispersions of the systems below and above well within the Galactic disk in the Solar neighborhood.
the orbital period gap are found σγ = 27.52 ± 2.28 km Total space velocity dispersion of all CVs in the final
s−1 and σγ = 25.65 ± 2.44 km s−1 , respectively (Table sample is σν = 54.29 ± 4.51 km s−1 , indicating a γ ve-
3). Dispersions of the γ velocities are expected to be locity dispersion of σγ = 31.34±2.61 km s−1 and a mean
σ(γ) ≃ 30 km s−1 and σ(γ) ≃ 15 for the systems be- kinematical age of τ =5.40±0.83 Gyr.
low and above the period gap, respectively, according to High probability thin disk CVs were found to be much
the standard formation and evolution theory based on younger than the local thin disk. Age of the local thin
disrupted magnetic braking (Kolb & Stehle 1996). Ob- disk was derived 6.8–7.0 Gyr from the white dwarf lu-
servations can not reveal such a large difference between minosity functions (Kilic et al. 2017). CVs could be
the γ velocity dispersions of the systems located below younger than single white dwarfs due to faster evolu-
and above the gap. tion, as CV white dwarfs are more massive compared to
It is clear that the results of the analysis in this study single white dwarfs.
do not meet the expectations of the theory with re- We have also concluded that there is no considerable
spect to the γ velocity dispersions. Also, previous ob- difference between total space velocity dispersions of
servational kinematical studies do not satisfy the the- magnetic and non-magnetic CVs. In addition, a mean-
ory with respect to gamma velocities. On the other ingful difference could not be found between polars and
hand, Kolb (2001) predicts that the γ velocity disper- IPs in the thin disk when comparing total space velocity
sions are σ(γ) ≃ 27 and σ(γ) ≃ 32 km s−1 for the sys- dispersions and kinematical ages.
tems above and below the gap, respectively, if magnetic Total space velocity dispersions of non-magnetic CVs
braking does not operate in the detached phase during belonging to the thin disk component of the Galaxy
the evolution. Although observational results in Table 3 were found to be 47.67±3.94 and 44.43±4.33 km s−1
do not fully match this prediction, while there is just a for the systems below and above the period gap, respec-
3σ disagreement, it is clear that the difference of the dis- tively, corresponding to kinematical ages of 4.19±0.71
persions is smaller, as expected in Kolb’s study (2001). and 3.61±0.74 Gyr. The mean kinematical age calcu-
Similar results were obtained in previous studies (Ak lated for the systems below the gap is in agreement with
et al. 2010, 2015). Another possible explanation of the the upper limit of the theoretical prediction, while there
present result may come from nova outbursts. We know is no agreement with the theoretical age prediction for
that a longer life for a CV means more dynamical inter- the CVs above the gap. In addition, the difference be-
action with massive objects and irregular components of tween the mean ages of systems below and above the
the Galactic potential. Thus, an older CV group must period gap could not be found as large as expected from
have a higher σγ than younger object groups. However, the population models of Galactic CVs.
if an object has undergone repeatedly to nova explo- It is also found in this study that the orbital period
sions with possibly asymmetric envelope ejection, then is decreasing with increasing age. The age-orbital pe-
its space velocity may additionally be affected by the cu- riod relations were found to be dP/dt = −1.82(±0.18) ×
mulative effect of these nova explosions. Kolb & Stehle 10−5 s yr−1 for all thin disk CVs, and dP/dt =
(1996) states that even such a single intrinsic evolution- −2.09(±0.22) × 10−5 s yr−1 for non-magnetic thin disk
ary event affects quadratically the dispersion of γ ve- CVs. Such a trend is expected from the standard evo-
locity. As a result, σγ values of older and younger CV lution model of CVs.
groups can be very similar. Not to mention, the increase A comparison of age-velocity dispersion relations of
in the γ velocity dispersions of CVs with the decreasing CVs in the thin disk component of the Galaxy with that
orbital period is clear (Table 3). Although this trend is of isolated white dwarfs given by Cheng et al. (2019, see
expected from the theory, calculated values do not ver- also Raddi et al. (2022)) shows that CVs are younger
ify the ones predicted by Kolb (2001) for the CVs above than isolated white dwarfs. As presented here, this dif-
 Kinematic Properties of CVs 19

ference between isolated WDs and CVs is just over a The authors thank the anonymous referee for his/her
1σ result, but after considering the previous point that helpful comments that improved the quality of the
would bring this difference closer, it is likely that there manuscript. This study has partly been supported
would be no significant difference in age between isolated by the Scientific and Technological Research Coun-
WDs and CVs. cil (TÜBİTAK) 119F072. This work has been sup-
γ velocity dispersions of the non-magnetic thin disk ported in part by Istanbul University: project number
CVs below and above the orbital period gap were ob- NAP-33768. This study is a part of the PhD the-
tained σγ = 27.52 ± 2.28 km s−1 and σγ = 25.65 ± 2.44 sis of Remziye Canbay. This research made use of
km s−1 , respectively. We could not find a large differ- NASA’s (National Aeronautics and Space Adminis-
ence between the γ velocity dispersions of the systems tration) Astrophysics Data System and the SIMBAD
below and above the gap, contrary to the prediction of Astronomical Database, operated at CDS, Strasbourg,
the standard formation and evolution theory Kolb & France and the NASA/IPAC Infrared Science Archive,
Stehle (1996). This small difference between the disper- which is operated by the Jet Propulsion Laboratory,
sions is in agreement with Kolb (2001), in which it is California Institute of Technology, under contract with
expressed that the dispersion difference is small if mag- the National Aeronautics and Space Administration.
netic braking does not operate in the detached phase This study used data from the European Space Agency
during the evolution. Note that the kick imparted by (ESA) mission Gaia (https://www.cosmos.esa.int/gaia)
nova outbursts could be responsible for this result, as and, processed by the Gaia Data Pro-
well. cessing and Analysis Consortium (DPAC,
https://www.cosmos.esa.int/web/gaia/dpac/consortium).
Funding for the DPAC was provided by national institu-
tions, particularly institutions participating in the Gaia
ACKNOWLEDGMENTS Multilateral Agreement.

REFERENCES
Ak, T., Bilir, S., Ak, S., Coşkunoğlu, K. B., & Eker, Z. Bilir, S., Karaali, S., Güver, T., Karataş, Y., & Ak, S. G.
2010, NewA, 15, 491, doi: 10.1016/j.newast.2009.11.007 2006c, Astronomische Nachrichten, 327, 72,
Ak, T., Bilir, S., Ak, S., & Retter, A. 2007, NewA, 12, 446, doi: 10.1002/asna.200510480
doi: 10.1016/j.newast.2007.01.002 Buser, R., Rong, J., & Karaali, S. 1999, A&A, 348, 98
Ak, T., Bilir, S., Özdönmez, A., et al. 2015, Ap&SS, 357, Cabrera-Lavers, A., Bilir, S., Ak, S., Yaz, E., &
72, doi: 10.1007/s10509-015-2245-3 López-Corredoira, M. 2007, A&A, 464, 565,
Andronov, N., Pinsonneault, M., & Sills, A. 2003, ApJ, 582, doi: 10.1051/0004-6361:20066475

358, doi: 10.1086/343030 Canbay, R., Bilir, S., Özdönmez, A., & Ak, T. 2023, AJ,
165, 163, doi: 10.3847/1538-3881/acbead
Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M.,
Chen, B., Stoughton, C., Smith, J. A., et al. 2001, ApJ,
Demleitner, M., & Andrae, R. 2021, AJ, 161, 147,
553, 184, doi: 10.1086/320647
doi: 10.3847/1538-3881/abd806
Chen, Y. Q., Nissen, P. E., Zhao, G., Zhang, H. W., &
Belloni, D., Schreiber, M. R., Pala, A. F., et al. 2020,
Benoni, T. 2000, A&AS, 141, 491,
MNRAS, 491, 5717, doi: 10.1093/mnras/stz3413
doi: 10.1051/aas:2000124
Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410,
Cheng, S., Cummings, J. D., & Ménard, B. 2019, ApJ, 886,
527, doi: 10.1051/0004-6361:20031213
100, doi: 10.3847/1538-4357/ab4989
Bensby, T., Feltzing, S., Lundström, I., & Ilyin, I. 2005,
Coşkunoǧlu, B., Ak, S., Bilir, S., et al. 2011, MNRAS, 412,
A&A, 433, 185, doi: 10.1051/0004-6361:20040332 1237, doi: 10.1111/j.1365-2966.2010.17983.x
Bilir, S., Cabrera-Lavers, A., Karaali, S., et al. 2008, PASA, Cox, A. N. 2000, Allen’s astrophysical quantities
25, 69, doi: 10.1071/AS07026 del Peloso, E. F., da Silva, L., Porto de Mello, G. F., &
Bilir, S., Karaali, S., Ak, S., Yaz, E., & Hamzaoğlu, E. Arany-Prado, L. I. 2005, A&A, 440, 1153,
2006a, NewA, 12, 234, doi: 10.1016/j.newast.2006.10.001 doi: 10.1051/0004-6361:20053307
Bilir, S., Karaali, S., & Gilmore, G. 2006b, MNRAS, 366, Downes, R. A., Webbink, R. F., Shara, M. M., et al. 2001,
1295, doi: 10.1111/j.1365-2966.2006.09891.x PASP, 113, 764, doi: 10.1086/320802
20 Canbay et al.

El-Badry, K., Conroy, C., Fuller, J., et al. 2022, MNRAS, Knigge, C., Baraffe, I., & Patterson, J. 2011, ApJS, 194, 28,
517, 4916, doi: 10.1093/mnras/stac2945 doi: 10.1088/0067-0049/194/2/28
El-Badry, K., & Rix, H.-W. 2018, MNRAS, 480, 4884, Kolb, U. 2001, in Astrophysics and Space Science Library,
doi: 10.1093/mnras/sty2186 Vol. 264, The Influence of Binaries on Stellar Population
Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. Studies, ed. D. Vanbeveren, 321,
2016, A&A, 595, A1, doi: 10.1051/0004-6361/201629272 doi: 10.1007/978-94-015-9723-4 24
Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. Kolb, U., & Baraffe, I. 1999, MNRAS, 309, 1034,
2018, Astronomy and Astrophysics, 616, A1, doi: 10.1046/j.1365-8711.1999.02926.x
doi: 10.1051/0004-6361/201833051 Kolb, U., & Stehle, R. 1996, MNRAS, 282, 1454,
—. 2021, A&A, 649, A1, doi: 10.1051/0004-6361/202039657 doi: 10.1093/mnras/282.4.1454
Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al. Littlefair, S. P., Dhillon, V. S., Marsh, T. R., et al. 2008,
2023, A&A, 674, A1, doi: 10.1051/0004-6361/202243940 MNRAS, 388, 1582,
Gänsicke, B. T., Dillon, M., Southworth, J., et al. 2009, doi: 10.1111/j.1365-2966.2008.13539.x
MNRAS, 397, 2170, Liu, W. M., & Chaboyer, B. 2000, ApJ, 544, 818,
doi: 10.1111/j.1365-2966.2009.15126.x doi: 10.1086/317231
Giovannelli, F. 2008, Chinese Journal of Astronomy and Majewski, S. R. 1993, ARA&A, 31, 575,
Astrophysics Supplement, 8, 237 doi: 10.1146/annurev.aa.31.090193.003043
McAllister, M., Littlefair, S. P., Parsons, S. G., et al. 2019,
Grundahl, F., Clausen, J. V., Hardis, S., & Frandsen, S.
MNRAS, 486, 5535, doi: 10.1093/mnras/stz976
2008, A&A, 492, 171, doi: 10.1051/0004-6361:200810749
Mihalas, D., & Binney, J. 1981, Galactic astronomy.
Halpern, J. P., & Thorstensen, J. R. 2015, AJ, 150, 170,
Structure and kinematics (W.H. Freeman & Co.)
doi: 10.1088/0004-6256/150/6/170
North, R. C., Marsh, T. R., Kolb, U., Dhillon, V. S., &
Hellier, C. 2001, Cataclysmic Variable Stars
Moran, C. K. J. 2002, MNRAS, 337, 1215,
Huber, M. E., Howell, S. B., Ciardi, D. R., & Fried, R.
doi: 10.1046/j.1365-8711.2002.05795.x
1998, PASP, 110, 784, doi: 10.1086/316192
Orosz, J. A., Thorstensen, J. R., & Kent Honeycutt, R.
Inight, K., Gänsicke, B. T., Schwope, A., et al. 2023,
2001, MNRAS, 326, 1134,
MNRAS, 525, 3597, doi: 10.1093/mnras/stad2409
doi: 10.1046/j.1365-8711.2001.04682.x
Johnson, D. R. H., & Soderblom, D. R. 1987, AJ, 93, 864,
Özdönmez, A., Ak, T., & Bilir, S. 2015, NewA, 34, 234,
doi: 10.1086/114370
doi: 10.1016/j.newast.2014.07.012
Jurić, M., Ivezić, Ž., Brooks, A., et al. 2008, ApJ, 673, 864,
Pala, A. F., Gänsicke, B. T., Breedt, E., et al. 2020,
doi: 10.1086/523619
MNRAS, 494, 3799, doi: 10.1093/mnras/staa764
Kalomeni, B., Nelson, L., Rappaport, S., et al. 2016, ApJ,
Pala, A. F., Gänsicke, B. T., Belloni, D., et al. 2022,
833, 83, doi: 10.3847/1538-4357/833/1/83
MNRAS, 510, 6110, doi: 10.1093/mnras/stab3449
Karaali, S., Bilir, S., & Hamzaoǧlu, E. 2004, MNRAS, 355, Patterson, J. 1984, ApJS, 54, 443, doi: 10.1086/190940
307, doi: 10.1111/j.1365-2966.2004.08319.x Pretorius, M. L., Knigge, C., O’Donoghue, D., et al. 2007,
Kilic, M., Munn, J. A., Harris, H. C., et al. 2017, ApJ, 837, MNRAS, 382, 1279,
162, doi: 10.3847/1538-4357/aa62a5 doi: 10.1111/j.1365-2966.2007.12461.x
King, A., & Lasota, J.-P. 2024, arXiv e-prints, Raddi, R., Torres, S., Rebassa-Mansergas, A., et al. 2022,
arXiv:2406.03948, doi: 10.48550/arXiv.2406.03948 A&A, 658, A22, doi: 10.1051/0004-6361/202141837
King, A. R., & Schenker, K. 2002, in Astronomical Society Ramsay, G., Green, M. J., Marsh, T. R., et al. 2018, A&A,
of the Pacific Conference Series, Vol. 261, The Physics of 620, A141, doi: 10.1051/0004-6361/201834261
Cataclysmic Variables and Related Objects, ed. B. T. Rappaport, S., Verbunt, F., & Joss, P. C. 1983, ApJ, 275,
Gänsicke, K. Beuermann, & K. Reinsch, 233, 713, doi: 10.1086/161569
doi: 10.48550/arXiv.astro-ph/0110186 Ringwald, F. A. 1994, MNRAS, 270, 804,
Knigge, C. 2006, MNRAS, 373, 484, doi: 10.1093/mnras/270.4.804
doi: 10.1111/j.1365-2966.2006.11096.x Ringwald, F. A., Thorstensen, J. R., Honeycutt, R. K., &
Knigge, C. 2011, in Astronomical Society of the Pacific Smith, R. C. 1996, AJ, 111, 2077, doi: 10.1086/117943
Conference Series, Vol. 447, Evolution of Compact Ringwald, F. A., & Velasco, K. 2012, NewA, 17, 108,
Binaries, ed. L. Schmidtobreick, M. R. Schreiber, & doi: 10.1016/j.newast.2011.07.007
C. Tappert, 3, doi: 10.48550/arXiv.1108.4716 Ritter, H., & Burkert, A. 1986, A&A, 158, 161
 Kinematic Properties of CVs 21

Ritter, H., & Kolb, U. 2003, A&A, 404, 301, Taam, R. E., & Spruit, H. C. 2001, ApJ, 561, 329,
doi: 10.1051/0004-6361:20030330 doi: 10.1086/322331
Takata, J., Wang, X. F., Kong, A. K. H., et al. 2022, ApJ,
Rodriguez, A. C., El-Badry, K., Suleimanov, V., et al. 2024,
936, 134, doi: 10.3847/1538-4357/ac8100
arXiv e-prints, arXiv:2408.16053,
Thorstensen, J. R., Alper, E. H., & Weil, K. E. 2016, AJ,
doi: 10.48550/arXiv.2408.16053
152, 226, doi: 10.3847/1538-3881/152/6/226
Schaefer, B. E. 2024, ApJ, 966, 155, van Paradijs, J., Augusteijn, T., & Stehle, R. 1996, A&A,
doi: 10.3847/1538-4357/ad31a9 312, 93
Schenker, K., King, A. R., Kolb, U., Wynn, G. A., & Warner, B. 1995, Cataclysmic variable stars, Vol. 28
Wielen, R. 1977, A&A, 60, 263
Zhang, Z. 2002, MNRAS, 337, 1105,
Wielen, R., Dettbarn, C., Fuchs, B., Jahreiss, H., &
doi: 10.1046/j.1365-8711.2002.05999.x
Radons, G. 1992, in The Stellar Populations of Galaxies,
Schreiber, M. R., Zorotovic, M., & Wijnen, T. P. G. 2016, ed. B. Barbuy & A. Renzini, Vol. 149, 81–92
MNRAS, 455, L16, doi: 10.1093/mnrasl/slv144 Willems, B., Kolb, U., Sandquist, E. L., Taam, R. E., &
Schwope, A. D., Horne, K., Steeghs, D., & Still, M. 2011, Dubus, G. 2005, ApJ, 635, 1263, doi: 10.1086/498010
A&A, 531, A34, doi: 10.1051/0004-6361/201016373 Willems, B., Taam, R. E., Kolb, U., Dubus, G., &
Sandquist, E. L. 2007, ApJ, 657, 465, doi: 10.1086/510479
Siegel, M. H., Majewski, S. R., Reid, I. N., & Thompson,
Yaz Gökçe, E., Karaali, S., Duran, Ş., et al. 2015, PASA,
I. B. 2002, ApJ, 578, 151, doi: 10.1086/342469
32, e012, doi: 10.1017/pasa.2015.12
Szkody, P., Everett, M. E., Dai, Z., & Serna-Grey, D. 2018, York, D. G., Adelman, J., Anderson, John E., J., et al.
AJ, 155, 28, doi: 10.3847/1538-3881/aa9ede 2000, AJ, 120, 1579, doi: 10.1086/301513
Szkody, P., Albright, M., Linnell, A. P., et al. 2013, PASP, Zorotovic, M., & Schreiber, M. R. 2020, Advances in Space
125, 1421, doi: 10.1086/674170 Research, 66, 1080, doi: 10.1016/j.asr.2019.08.044