nonnegligible height comparing RwD, the released energy iS decreased in the shock at top of accretion column and the temperature of the accretion column is reduced. Simulta-neously, the temperature distribution changes and, with even lower a the energy input by gravity overcomes the radiation cooling. Then, the temperature peak emerges in the middle of the accretion column and the prominent temperature peak is formed close to the WD surface with fu11y Iow a. These temperature distribution more easily changes in more massive WD because more massive WD causes faster accretion flow which leads lower density following the cbntinuum equation and easier propagation of the physical
state of the top of the accretion column. The non-equilibrium between ion and electron is also enhanced with lower a and more massive WD. Influence of'abundance giving on the accretion column structure is limited because the abundance of IPs are lower than O.6 Zo in general although higher abundance enhances the radiation cooling and shortens the accretion column. As described above, we have revealed that the accretion column structure is significantly changed with a and its height may become to be equivalent to the WD radius. Then we need to consider the magnetic dipolar geometry.
Secondly, : e have taken into account the magnetic dipolar geometry, i-n- otdher w. ords, replaced the continuum equation with pvS = a which includes the cross-section change.
The cross-section change lead to the nozzle effect, which converts the thermal energy into that of the bulk motion.As the cylindrical post-shock accretion coluMn, the dipolar
accretion column occupy•the two phases divided by 'h nv RwD. In low a phase, the
accretion column height relates with a as h or a-O'i5 while in high a phase relates as h oc a-i similar to the cylindrical. In high a phase, the structure of the dipolar accretion column except its height and density dose not depend on a and is identical to that of the cylindrical. On the other hand, in low a phase, the effect of extend of the cross-section o'f the accretion column which causes the nozzle effect and simple attenuation of the density emerges and therefore, the density becomes Iower than that of the cylindrical.At the same time, the density distribution of the dipolar accretion column'depends on a although the cylindrical dose not. The temperature dist,ribution is also lower than that of the cylindrical because the nozzle effect reduces the temperature. As a result, the temperature peak emerging in the low a phase of the cylindrical disappear. On the other hand, since the energy conversi'on between ion and electron is suppressed by the reduction of the density and the faster fiow, non-equilibrium area significantly extends
and may extend in the bulk of the accretion celumn. Furthermore, in extreme case
such as massive WD of 1.4 Mo and low a of O.OOI g cm-2 s-i, the prominent electron temperature peak emerges close to the WD surface similar to the averaged temperature of the cylindrical. Then, we have obtained the post-shock accretion column structure involving the difference of a, magnetic dipolar geometry and equilibrium between ion and electron.Next, in order to compare observations to our accretion column model, we have con-structed the spectrum model. After the post-shock dipolar accretion column was divided into one hundred components and spectra of the each components were calculated with SPEX involving the ionization non-equilibrium of iron ion, the all spectra were summed up to gqin the entire spectrum of the accretion column. As a decreases and, therefore, the accretion column temperature reduces, the hard X-ray continuum reduces regardless of the WD mass. As for line emissions, for somewhat light WD as below 1 Mo, He-like and H-like iron Kcu lines become more intense and weaker, respectively, as a decreases•
On the other hand, for massive WD, both He-like and H-like iron Kcy lines become more intense as a decreases because most iron ions is fu11y ionized with higher a due to very high temperature. The 3360 spectra have be calculated with loga = -4, -3.75, -3.5, -3.25, -3,
11.2. FUTUREPROSPECT 137
-2.75, -2.5, -2.25, -2, -1.5, -1, -O.5, O, O.5, 1, 1.5 and 2, MwD between O.4 and 1.4 per O.05 MwD, and Z between O.1 and 1 Zo per O.1 Zo. And then, the spectra have interpolated with spline function for a and linear function for MwD and Z, and are eventually added
in XSPEC as Acrad model.
We have observed high mass accretion IP V1223 Sagittarii and low mass accretion IP EX Hydrae with SzLzakzL satellite which have large effective area and good energy resolution in iron K-shell energy band, and high detection sensitivity in hard X-ray band.
Before application of the Acrad model to these SzLzak2L observation, we have measured the accretion column height of V1223 Sagiitarii which had not been measuted with the refiected X-ray. And we have obtained strong restriction of the accretion column height We have applied the Acrad model to the SzLiaku observations of V1223 Sagittarii and EX Hydrae. Then we have estimated various parameters of V1223 Sagittarii as loga > O, M.. - O.83Å}8182 M., Z - O.24Å}8I8g Z., N. = 9.0Å}glg Å~ lo22 cm-2, Ar. ,. = 12oÅ}gg Å~
1022 cm-2, Ccp == O.39Å}8.'6g and h < O.036 RwD. Our MwD and Z are consistent with past results based on the Cropper model, which indicate the exactness of the Acrad model because the estimated a is suficiently high and the post-shock accretion column structure is completely identical to the Cropper model. The accretion column height expected by our accretion column model using the obtained parameters is also consistent with our direct measurement. Similarly we have estimated parameters of EX Hydrae as -3.5 <
loga < -2.5, MwD = O.83Å}O.03 Mo, Z == O.56Å}O.04 Zo, NH == 3.8Å}l:2 Å~ 1022 cm-2,
NH pc = 268t6i5 Å~ 1022 cm-2, Ccp = O.37Å}8:,i9 and O.73 < h < 1.2 RwD. Our MwD
corresponds to that measured by the binary motional method. These result shows thatdifference of a must be considered for WD mass estimation with X-ray and made up
longstanding problem of the'WD mass in EX Hydrae. The spectral difference between V1223 Sagittarii and EX Hydrae is cause by only difference of a. On the other hand, assuming a > 1, the WD is estimated at O.46Å}8:85 Mo, which is consistent with results based on Cropper model and means that the Acrad model is correct. As for the accretion column height, our calculation with estimated parameters and the observation is well consistentr As above, we have succeeded in the establishment of the WD mass estimation method using X-ray.11.2 FutureProspect
'
Our accretion column model and its Spectral model may revise the WD masses of most IPs then we need to apply Acrad model to many observations in order to re-evaluate the WD mass function in IPs. Furthermore, we will involve the cyclotron radiation into our model to extend the range of application to polars.
Appendix
Discovery of spin-modulated Florescent Iron Kdv line
We detect a fluorescent iron Kor emission line in V1223 Sagittarii, whose central energy is discovered to be modulated with the WD rotation for the first time in magnetic-CVs (Hayashi et al., 2011). Detailed spectral analysis indicates that the line comprises of a stable 6.4 keV component and a red-shifted component, the Iatter of which appears only around the rotational intensity-minimum phase. The equivalent width (EVV) of the former stable component rv80 eV together with the measured st indicates the major reflector is the WD surface, and the shock height is not more than 7oro of the WD radius. severe constraint in non-eclipsing IPs. The red-shifted iron line component can be interpreted as emanating from the pre-shock accretion flow via fiuorescence. Its EW (28Å}S eV) and the central energy (6.30Å}gl8g keV) at the intensity-minimum phase are consistent with this mterpretatlon.
11.3 Analysis and Results
11•3.1 TimingAnalysis t
In order to evaluate the period of the WD spin, we first carried out an FFT analysis with the barycentric corrected light curve of the XIS, and identified a rough spin period. We theh performed an epoch-folding analysis and produced a periodogram, which is shown in Fig. 11.1. As a results, we got the period of 745.7 Å} 1.1 sec, which is consistent with 745.6
o8,
co
88
z*S
88
o
I l l 1 l 1 l l
745.7 (s)
749
745 750
Spin Period (s)755
Figure 11.1: Periodogram calculated from background subtracted XIS light curve in the O.5-11.5 keV energy band. The horizontal and vertical axes show trial periods and x2, respectively. The x2 values are evaluated with light curves with 31 bin/cycle. The trial period step is O.Ol s. The maximum x2 is obtained at the period 745.7 s.
139
140
'
sec (Jablonski & Steiner, light curves folded at the
:I]
g:
2.s3
i
'i}3g )t2.5E
• g,
U3
1.2 ,.g O.7 O.6 '
g'.2'
Figure 11.2: Folded light is scale