4 Discussion

Since the trapped gas is not actually accreting onto Sgr A*, it implies that dissipation of angular momentum in the central arcsecond is, at present at least, inefficient. Over the next 10⁵yrs, the density of the ISM in the central arcsecond will increase as the winds from the early-type cluster fill the region. Eventually, sufficient dissipation may result in the formation of a more standard accretion disk, turning our Galaxy into a low level Seyfert. However, if the stars in the central few parsecs formed at the same time, it is likely that another SN will explode in less than 10⁵ yrs, truncating the accumulation process and triggering another "binge and purge'' cycle.

If, as recently proposed (Paumard et al. 2001), the early-type cluster stars are primarily Luminous Blue Variables (LBVs) rather than WR stars, the shorter duration of the LBV phase would imply that the cluster has been producing massive winds for  yrs. However, the winds from LBVs are slower ( ${\sim} 200$ km s^-1) and more massive ( ${\sim} 10^{-3}~\dot M_{\odot}$ yr^-1) than WR winds, with a resulting value of $\bar n$ that is still ${\sim} 50$ times larger than the previous estimate which assumed the wind sources to be WR stars. Nonetheless, if the IRS 16 stars have entered the LBV phase only within the last few centuries, the mass of trapped gas would be sufficiently reduced to meet the X-ray observational limits. If so, Sgr A* might become significantly brighter in the next few centuries as the LBV winds continue to fill the accretion region. However, IRS 7, a red supergiant ${\simeq} 0.3$ pc from Sgr A*, has a tail that points away from IRS 16 and Sgr A*. Models (Yusef-Zadeh & Melia 1992; Dyson & Hartquist 1994) show that a GC wind of ${\sim}500$ km s^-1 is needed to produce this tail. Such a wind would require ${\sim}500$ yrs to reach IRS 7 from IRS 16, implying that the wind sources are not exclusively LBVs.

Another possibility is that the gas flows in the central few arcseconds are not in equilibrium due to an explosive event ${\sim} 10^{3-4}$yrs ago. As mentioned above, Sgr A East is a probable mixed-morphology type II supernova remnant (SNR) that appears to envelope Sgr A West, the compact HII region which contains both Sgr A* and the early-type cluster (Goss et al. 1989). When the progenitor of Sgr A East, located ${\sim} 5$ pc from Sgr A*, exploded ${\sim}10^4$ yrs ago, the strong frontal shock could have cleared the central parsec of its accumulation of wind as it swept by Sgr A* ${\sim}10^3$ yrs ago (Maeda et al. 2001). With a temperature of 2 keV and a density of ${\sim} 10$ cm^-3 (Maeda et al. 2001), the pressure of the rarified gas in the cavity of the SNR is  dyne cm^-2while the ram pressure due to the wind from an individual cluster member at a distance of $R_{\rm A}$ is on average  dyne cm^-2. Thus, after the passage of the dense frontal shock, the cluster wind would have overwhelmed the gas in the SNR cavity and refilled the central $R_{\rm A}$ over the last ${\sim}10^3$ yrs. Sgr B2, an X-ray reflection nebula near the GC, seems to have been illuminated by Sgr A* ${\sim}500$ yrs ago, suggesting that the frontal shock passed by Sgr A* at that time, dramatically increasing the emission of the blackhole (Sunyaev et al. 1993; Koyama et al. 1996; Murakami et al. 2000). The frontal shock was likely to be slow and dense enough to result in Sgr A* trapping enough gas to be accreting at near its Eddington rate for ${\sim}10^3$ yrs (Maeda et al. 2001; Baganoff et al. 2001).

Since IRS 7 is presently interacting with the wind from IRS 16, the central $R_{\rm A}$ is likely to be filled with IRS 16 winds. Thus, the rarified cavity wind from Sgr A East is unlikely to be the present source of accretion for Sgr A* as suggested by Baganoff et al. (2001). However, if IRS 7 is interacting with a wind from Sgr A* rather than IRS 16, it would be evidence of recent activity near the blackhole, suggesting that the IRS 16 winds have been reasserting themselves only in the last few centuries, if at all. A continual outflow from Sgr A* rather than IRS 16 would clear the central arcsecond of gas but would require a fast and/or dense outflow on scales greater than $R_{\rm min}$; there is no observational evidence of such a wind.

In any case, IRS 16C, a WR star only a few arcseconds from Sgr A*, is alone producing a sufficient wind to accumulate almost $10^{-5}~M_{\odot}$ yr^-1 within $R_{\rm A}$. Assuming the blackhole is accreting only a fraction of this, the upper limit on the time the wind from this star has been accumulating in the central arcsecond is 10⁴ yrs. Since there are dozens of young massive stars with heavy winds in the central parsec, it is likely that  yrs.

It is thought that Sgr A West and Sgr A* lie only slightly in front of Sgr A East since the 90 cm emission from the expanding shell of Sgr A East is seen in absorption at the location of Sgr A West (Yusef-Zadeh et al. 1999). However, a north-south tongue of emission which cuts through Sgr A West and approaches Sgr A*, can be seen. If the proposed scenario is correct, this tongue corresponds to the leading edge of the SNR and thus should expand over the next few decades as the shell moves more completely in front of Sgr A West.

It has been assumed that the early-type cluster is not displaced from Sgr A* by more than a few arcseconds along the line of sight. Proper motion and radial velocity observations (Genzel et al. 2000) support this and in fact suggest the orbits of the cluster stars are nearly Keplerian. If cluster members are on Keplerian orbits around Sgr A*, their orbital velocities might be an appreciable fraction of their wind velocities. This may add sufficient angular momentum to the winds so that a larger fraction of the gas never reaches the capture radius. However, the density of ionized gas falls off outside ${\sim} 1\hbox{$^{\prime\prime}$ }$ (Baganoff et al. 2001), suggesting that only a fraction of $\dot M_{\rm A}$ is being trapped outside the capture radius due to support from angular momentum or magnetic fields.