eta Carinae
Page Last Updated: May 29, 2015Update to the RXTE X-ray Light Curve
Section Last Updated: May 27, 2015While trying to make sure the X-ray modeling is consistent with the data in both cts/s units (which I did when there was only RXTE data) and flux units [ergs/s/cm^2] (which I do now since I want to compare with RXTE and Swift observations), I noticed a problem with applying the RXTE cts/s-to-flux conversion from Corcoran 2005. Subsequent to this paper, Mike fit each RXTE observation (1996-2011) with a
wabs*mekal + wabs*(mekal+mekal)
model (using the latest response function calibrations), so for comparison purposes I extracted a 2-10 keV flux light curve from these models. For the beginning of the RXTE light curve, the cts/s-to-flux result of Corcoran 2005 agrees well with this fitting. However, after phase 2.2 (0.2 after 2003.5 minimum), the results start to diverge; the cts/s-to-flux result is lower than the fit result. What Mike and I came up with as an explanation for this is that the response of the the detector changed slightly at phase 0.2 after the 2003.5 minimum, so the cts/s before and after this date are not directly compatible. This explains the arbitrary change in the hardness ratio at this phase, which we previously thought might be evidence of a mass-loss-rate change in eta Car, as in Corcoran+2010.
Here are two (quickly made) plots showing the agreement-then-divergence in the two flux calculations ('conversion' is cts/s-to-flux from Corcoran 2005; 'fit' is from the mekal model) and the X-ray light curve according to the fitting, along with Swift.
The most important result is it looks like the X-ray flux after the recovery has been the same for all three RXTE cycles; the X-ray flux post-2009.0 was not lower than the previous two cycles. Thus there is now more evidence that the mass loss rate of eta Car's primary is not changing.
Another interesting trend is the maxmimum X-ray flux before the 2009 minimum is higher, so there's a clear trend in the maximum X-ray flux increasing over all 4 observed cycles. (Previoulsy the 2003.5 max and 2009.0 max were about the same.)
This also affects how the hardness ratio is computed. I suppose we should now compare the model flux in the 2.5-3.5 keV and 6.5-7.5 keV bands, as opposed to the cts/s in those bands, across the duration of the RXTE mission.
Varying Secondary Parameters
Section Last Updated: May 29, 2015Here are, for the first time, SPH simulations where the wind speed and mass loss rate of the secondary are varied in a systematic fashion. The sims done are
Mdot_Sec = 1.4e-5, v_Sec = 3000 --> eta = 0.118 (standard case)
Mdot_Sec = 4.2e-5, v_Sec = 3000 --> eta = 0.353 (high Mdot_Sec; standard*3)
Mdot_Sec = 4.7e-6, v_Sec = 3000 --> eta = 0.039 (low Mdot_Sec; standard/3)
Mdot_Sec = 1.4e-5, v_Sec = 3500 --> eta = 0.137 (high v_Sec)
Mdot_Sec = 1.4e-5, v_Sec = 2500 --> eta = 0.098 (low v_Sec)
where eta = Mdot_Sec * v_Sec / (Mdot_Pri * v_Pri) is the wind momentum ratio -- assuming the winds are at terminal speed. The primary parameters from all these simulations are Mdot_Pri = 8.4e-5 and v_Pri = 420.
Even though they are low resolution, the changes in the wind-wind interactions are obvious. The following movies, which show density, velocity, temperature, and 1 keV X-ray source function in the orbital plane, show these changes.
standard
high Mdot-Sec
low Mdot-Sec
high v-Sec
low v-Sec
The X-ray light curves from these simulations, with the line-of-sight of i=45 and omega = 252, and line-of-sight i=90 and omega=270 are shown here.
While the first LoS is much more probable than the second, the second provides a better comparison among the model X-ray properties since this LoS is within the apastron cone. For the first LoS, the sims with a weaker secondary could mean that i=45, omega=252 is outside the apastron shock, thus exascerbating the differences between the light curves.
The change in the length of the X-ray minimum has a pretty straighforward explanation. The stronger the secondary wind, the closer the stars can be and still produce X-rays since the secondary wind accelerates to X-ray generating speeds before shocking. Once the primary wind gets close enough to the secondary star that the secondary wind shocks at low speeds, there are no more X-rays produced.
Also at play is the fact that the X-ray source moves to the other side of the primary star at periastron, i.e. an eclipse model, but the fact that the movies show the X-rays disappearing from the simulation at different times makes this eclipse effect secondary.
And the last effect is the increase in radiative cooling as the stars get closer. This should influence the exact phase at which X-rays disappear, especially given its strong v dependence (~v^4), but once there are no more X-rays generated, cooling is irrelevant.
Here is a comparison between the spectra at phase 0.93.
The different Mdots show predominantly a change in normalization/emission measure, while the different velocity spectra show slightly different spectral shapes. All these trends are as expected.
I find the different velocity sims particularly interesting since, outside the minimum, the standard and v2=2500km/s sims have approximately the same flux, while the v2=3500km/s sim produces a lower flux. At play are the change in the KE available for shocking (obviously), the change in the density of the secondary wind at the shock due to both the radial dependence of rho and the location of the shock, and the change in the opening angle of the shock, which relates to how direct/oblique the off-axis secondary wind impacts the shock, and thus what fraction of the secondary wind's KE turn into heat. It appears that the smaller opening angle of the v2=2500km/s sim, and thus its more direct shocks, compensate for the decrease in avaibale KE when compare to the standard shock, while the v2=3500km/s sim produces the lowest X-ray flux due to the most oblique shocks (even though the available KE is the highest). My prediction is that an even lower v2 sim (say 2000km/s) would produce an low X-ray flux like the 3500km/s case, so the standard and v2=2500km/s sims are around some sort of X-ray maximum.
This also means that v2=2500km/s is a viable option, given the other parameters in the sims, and should be explored more. The other parameters -- v2=3500kms/s, Mdot2=4.2e-5, Mdot2=4.7e-6 -- are not viable assuming Mdot1=8.5e-5 and v1=420km/s (as well as all the other assumptions in the SPH modeling: beta=1 velocity law, solar abundances, etc.).
Home
My Bio
