First Observed Interaction of the Circumstellar Envelope of an S-star with the Environment of Sgr A*

The SINFONI was mounted on the VLT and underwent an upgrade (for further information about the upgrade, see Kuntschner et al. 2014; Marchetti et al. 2014; Pearson et al. 2016). SINFONI operates in the NIR and provides observations in the J- (1.10–1.40 μm), H- (1.45–1.85 μm), K- (1.95–2.45 μm), and H + K-bands (1.45–2.45 μm). The output files of the ESO pipeline are in the shape of a three-dimensional data cube. This data cube consists of two spatial dimensions and one spectral dimension. The components of the data cubes are described in spaxels (pixels containing a spectrum, see Hörtner et al. 2012) rather than pixels. With SINFONI, we are able to isolate single emission lines in the H + K-band to create channel (line) maps. In comparison, the NACO ⁹ instrument works also in the J, L', and M'-bands (Lenzen et al. 2003; Rousset et al. 2003). Since dust can be traced in higher wavelengths, the L'-band setup of NACO is favored for the search of the dusty bow shock source. In both cases, we apply the usual data reduction steps, e.g., dark- and flat-field corrections. We also apply the mandatory sky correction to the adaptive optics (AO) corrected data. Additional correction steps are described in detail in Peißker et al. (2019, 2020a, 2020b, 2020c, 2020d) where the data analyzed here is also used. See Tables 1, 2, 3, 4, 5, 6 and 7, and Appendix E for a detailed overview about the data used.

We also note that a part of this data was used in Parsa et al. (2017). The authors describe the Schwarzschild precision of S2, which was independently confirmed by the Gravity Collaboration et al. (2020). The collaboration used GRAVITY, an interferometric instrument with a resolution almost one magnitude better than NACO. ¹⁰ This underlines the robustness and validity of the NACO data.

The NIR camera COMIC was installed in La Silla/Chile at the 3.6 m telescope and used the AO system ADONIS+RASOIR (Beuzit et al. 1994; Lacombe et al. 1998). It operated in the J-, H-, K-, K'-, L'-, and M-bands with two different plate scales (35 mas/pixel and 100 mas/pixel). It was optimized for L- and M-band observations and decommissioned in 2001 (Pasquini & Weilenmann 1996). For the data presented here, the exposure time was set to 10 s. The observational pattern was chosen to be sky-object-sky followed by flat and dark exposures. For combining the data, we used the shift and add algorithm to maximize the signal-to-noise ratio (S/N). This is followed by rebinning the data to smooth sharp edges caused by the resolution. The COMIC/ADONIS+RASOIR data analyzed in this work was first published in Clénet et al. (2001).

Depending on the scientific goal, a suitable frequency pass filter can improve the amount of accessible image information. In the case of an elongated source like X7, using the Lucy-Richardson algorithm (Lucy 1974) is not the most satisfying option for the L'-band NACO data. However, using a high-pass filter like the smooth-subtract algorithm can reveal the stellar component of the bow shock source in the K-band SINFONI data if the object is confused with nearby S-stars. For that, we are subtracting a smoothed version of the input image file. The size of the Gaussian that is used for the smoothing should be of the order of the related image point-spread function (PSF). The resulting smooth subtracted image should then be resmoothed with a Gaussian PSF that can be 10%–20% smaller than the image PSF. With this technique, the influence of overlapping PSF wings can be minimized.

Throughout the SINFONI data between 2005 and 2018, the source X7 can be at least partially observed. A key parameter is the FOV. Hence, the SINFONI data in 2006, 2008, 2014, 2015, and 2018 can be used for a detailed analysis. By analyzing continuum subtracted line maps, we find the length from the tip to the tail of the detected Doppler-shifted Brγ emission to be around 0.23'' in 2006. Furthermore, we measure the length of the bow shock of about 0.35'' in 2018.

Because of the high S/N of the SINFONI data in 2018 (see Appendix E) at the spatial position of the Doppler-shifted Brγ emission of X7, we use this set to investigate the velocity gradient of the bow shock source. For this purpose, we fit a Gaussian to the blueshifted Brγ line in the related spectral range (2.16 ± 0.04 μm). Afterwards, the related spaxel carrying the velocity information is copied to the same position in a new array that is as big as the input file. We manually mask the close-by source X7.1/G5 (see Ciurlo et al. 2020; Peißker et al. 2020b) and nonlinear pixel.

In Figure 1, the resulting velocity gradient is shown. We find a difference from the tip to the tail along the projected bow shock structure of ≈190 ± 20 km s⁻¹. Considering a spatial pixel scale of 12.5 mas in the H + K-band in the highest plate-scale setting of SINFONI and the measured projected bow shock length of 349 mas, we get a linear gradient of ≈7.1 ± 1.0 km s⁻¹/pixel for 2018. Furthermore, we find several prominent emission lines that indicate the presence of ionized gas (see Table 8). In several data sets, a H₂ emission triplet can be found (Appendix B, Figure 10). Due to crowding and therein the resulting possibility of confusion, we limit the analysis of the H₂ emission triplet to the data of 2018.

Zoom In Zoom Out Reset image size

L'-band observation of the bow shock source X7 in the close distance of the S-cluster shows that the object is always one of the most prominent sources in the close vicinity of the SMBH (see Figure 2). The L'-band brightness and elongated shape of X7 underlines the unique character of the object.

After 2002, X7 becomes increasingly brighter than most of its nearby stars like, e.g., S1, S2, S61, and S71. The bow shock shape is clearly noticeable in the NACO L'-band (green circles indicate its position in Figure 2). After 2010, the source shows a more elongated shape with an approximate projected length of 333 mas in 2016. This is almost three times as much as the L'-band dust emission detected with NACO in 2002 (112 mas). Compared with the line emission area of the SINFONI data in 2006, 2008, 2009, 2013, 2014, 2015, and 2018, we find matching values for the gaseous emission (for a detailed list, see Table 9). Hence, the size of the projected area of the ionized gas coincides with the L' continuum dust emission of X7.

Since we observe a clear increasing projected size of X7 between 2002 and 2018, we investigate the L'-band data of 1999 to see if this trend is also observable in data before 2002. For this purpose, we use the investigated COMIC/ADONIS+RASOIR L'-band data in 1999 by Clénet et al. (2001). We apply a high-pass filter to reduce the influence of overlapping PSF. Afterwards, we use a Gaussian that is about 50% in size of the initial smoothing kernel (Appendix D, Figure 12) on the resulting high-pass filtered image. In addition to some prominent members of the S-cluster, we identify at the expected position of S50 a spherical L'-band emission several magnitudes above the noise level. By comparing the closest NACO L'-band data to verify the COMIC/ADONIS+RASOIR identifications of 1999, we find matching positions for almost all stars/features.

Mužić et al. (2010) reported that the stellar counterpart of X7 could be associated with the S-cluster star S50. In Figure 12 (right side), we present the orbit plots of S50 based on the analysis presented in Ali et al. (2020). Throughout the available data covering the related spatial area, we find without confusion that the bow shock source X7 is moving along with S50 (see Figure 3 and Appendix A, Figure 9) until 2009. S2 (K-band) and S65 (L-band) as the two brightest and therefore most prominent members of the S-cluster can always be observed in the same FOV as S50. Hence, we are use these two S-stars for a photometric analysis to investigate the magnitude of S50 and X7 in various bands (see Table 10). In combination with the published System for High Angular Resolution Pictures (SHARP) data (Schödel et al. 2002), we find a constant K-band magnitude of S50 of m_K ≈ 16 mag. We find a similar magnitude with the NACO (VLT) data of 2007 and SINFONI (VLT) data of 2019. Based on the data that covers almost 20 years, we conclude that the S-cluster member S50 does not show a variable K-band emission. However, this is not the case for the L'-band continuum emission that seems to vary between 2008 and 2018. We will elaborate on this point in detail in Section 3.5.

Zoom In Zoom Out Reset image size

Based on the L'-band observations, we find a noticeable elongation of the source X7 that becomes increasingly prominent after 2009 (please see Figure 2). Comparing the NACO L'-band images with the SINFONI line maps, we find that the symmetric distribution of gas and dust cannot be observed after 2009. Compared to the SINFONI data between 2006–2008 and 2010–2018, the data shows a rather compact gas emission in 2009 (see Figure 3). Whereas the data of 2006–2008 shows a symmetrical gas-to-dust distribution with respect to S50 and X7, we find that this symmetry of the S50-X7 system is broken for the observations between 2010 and 2018. Furthermore, we observe that the distance of the gaseous front of X7 (i.e., head) is increasing year by year with respect to S50 (Figure 4).

Zoom In Zoom Out Reset image size

In contrast, the back (or tail) of the Brγ gas emission does not show a comparable behavior compared to the head after 2009. As previously described, this leads to an asymmetric distribution of the gas around the central stellar source S50. This broken symmetry between the shell and the star can also be observed in the NACO L'-band data (see Figure 5).

Zoom In Zoom Out Reset image size

Hence, the data implies that the gas and dust shell starts to detach in projection after 2010. This process can be tracked throughout the available NACO and SINFONI data beginning in 2009 and is indicated in Figure 4. Furthermore, we find that the intensity maximum of the dust is located at a distance of less than 13 mas to the position of S50 (Figure 2). As a result, the tail of X7 gets increasingly brighter when comparing the data between 2002 and 2018.

The photometry was done in the H-, K-, L'-, and M'-bands. As shown in Figure 2, Figure 5, Figure 12, and Figure 4 the dusty bow shock of X7 becomes elongated between 1999 and 2018. After 2007, the projected elongated size of the L'-band emission exceeds a spatial coverage of two PSF (≈0.20'') and we categorize the source in a front- (i.e., head) and back-part (i.e., tail). For this analysis, we focus on the tail of X7 since deriving the emission area of the faint L'-band head magnitude is not without confusion. For the photometric analysis, we use S65 because of its well-known stable magnitude of about 10.96 mag (Hosseini et al. 2020). For the magnitude of X7, we use the peak emission of the L'-band dust emission (see Figure 2). The magnitude of X7 is derived from the peak intensity and can be related to the tail of the source after 2008. For every data set, a 1 pixel aperture is used. No background subtraction is applied because of the high S/N that exceeds several orders of magnitude the intensity of the surroundings. The fit presented in Figure 6 can be categorized into two different results:

• 1.  
  A constant magnitude of X7 before 2007,
• 2.  
  A variable magnitude of the tail after 2007.

Regarding point 1, the COMIC/ADONIS+RASOIR and NACO L'-band data between 1999 and 2006 does not show a variation in magnitude. Additionally, point 2 underlines a slightly variable L'-band tail magnitude of the bow shock at the K-band position of S50. These variations of the L'-band magnitude of X7 coincide with the discontinuous shape evolution that is observed in the Brγ line maps (see Figure 3 and Figure 5). By investigating several data sets of the GC that cover individual bands, we find an increasing flux toward higher bands (from the H- to M-band, see Table 10) for X7. Using the magnitude values, we derive the SED with a two-component fit for the emission of S50 (H and K) and X7 (L' and M'). This indicates a dust-dominated emission source with a multiwavelength appearance. Since the commonly observed dust temperature in the GC is about 200 K (Cotera et al. 1999), the derived envelope temperature of 450 K must be heated up by the internal stellar source S50.

Zoom In Zoom Out Reset image size

From the survey of X7 over two decades with all publicly available SINFONI and NACO data, we have shown that the shape of the bow shock does change over time on a significant level. Even when we consider different weather and background scenarios, the findings presented here underline a dynamical star-envelope setup. As shown in Table 9 and Figure 3, the shape of X7 undergoes a transition: we find an almost constant position angle and magnitude with a linear increasing bow shock size both in gas and dust until 2009. Based on Mužić et al. (2010), this setup for X7 is expected because S50 as the stellar counterpart is located close to the front tip of the bow shock X7. As theoretically described by Wilkin (1996, 2000) and observed by Mužić et al. (2010), we can confirm that the S-star S50 is always located at the position of the maximum peak intensity of the observed L'-band emission of the bow shock X7. This L'-band intensity peak can be found close to the apex of the bow shock at a distance of ${R}_{0}=\sqrt{\tfrac{{\dot{m}}_{{\rm{w}}}{v}_{{\rm{w}}}}{{\rm{\Omega }}{\rho }_{{\rm{a}}}{v}_{{\rm{a}}}^{2}}}\approx 2.5\times {10}^{15}$ cm (Mužić et al. 2010) until 2009. Here, ${\dot{m}}_{{\rm{w}}}$ describes the mass-loss rate of the star, v_w is the stellar-wind velocity, Ω a dimensionless parameter to control the shape of the bow shock (Ω = 4π for an isotropic stellar wind), ρ_a is the density of the ambient medium, and v_a the relative stellar velocity in a nonstationary medium.

Between 2009.47 and 2010.49, we observe a discontinuous process since the Brγ- and L'-band size is decreased by almost 30% compared to the observation in 2009.26 (NACO). After 2010, not only is the Brγ- and L'-band continuum size expanding, but also the position of S50 seems to change with respect to the shell. Hence, R₀ is not a fixed value anymore and seems to change year by year. Because the stellar position with respect to its dusty envelope does not follow any simple stationary model, we will speculate about some possible interpretations.

Henney & Arthur (2019) discussed dust and bow waves as a possibility for asymmetric shapes. Considering a possible "rip-point" (where the shell gets detached from the star) harbors the problem that these processes (including the trajectories of the dust grains) take up to several thousand years as proposed by Henney & Arthur (2019). We have shown that the gas distribution coincides with the dust emission (see Table 9 and Figure 5). In 2008, we find a matching size of the emission of about 230 mas. The NACO data of 2009.26 seems to follow the linear evolution of the observed emission size in 2008. For the SINFONI data of 2009.47, we observe a source size that is unexpected. Because of these timescales, we see a reduced chance of the possibility of dust and bow waves as a suitable explanation for the discontinuous evolution.

Another possibility are projection effects. Considering the possibility that S50 could maybe not be related to X7 at all and just moves on a random orbit that coincides in projection with X7 opens a new set of questions. In the following, we independently discuss these questions ignoring the already complete discussion of Mužić et al. (2010), where the authors exclude the possibility of a random encounter based on the matching proper motion of S50 and X7.

The most obvious one is regarding the statistical robustness of a randomized orbit that is oriented along the trajectory of X7 over time. As derived by Sabha et al. (2012) and Eckart et al. (2013), the probability for such an event is of the order of 10⁻⁴ to a few percent for a consecutive observation of 3 yr. The probability for the outer region of the S-cluster should therefore be in a comparable range since we observe S50 along with X7 between 2002 and 2009 (NACO) and 1999 with COMIC/ADONIS+RASOIR (Appendix D, Figure 12).

As shown in Figure 2 and Figure 5, the shifted L'-band intensity maximum toward the tail is followed by the projected position of S50. Based on the derived L'-band magnitude year by year, the temperature of X7 is always well above 200K which can only be achieved by an internal heating source. Hence, we conclude that the tail of X7 gets heated up by S50. Alternatively, a wind that originates southwest of the position of Sgr A* could be responsible for the increased tail emission in 2018. However, this does not explain the Wilkinoide (Wilkin 1996) bow shock between 2002 and 2009 that is observed throughout the NACO and SINFONI data. In combination with the continuum and line emission data of 2006 and 2008 (see Figure 5), we will not discuss the possibility of another wind coming from the southeast any further, especially considering the observed footprint of a wind that originates at the position of Sgr A* or IRS16 in the mini-cavity (see, e.g., Lutz et al. 1993).

A more suitable explanation of the observed gas and dust emission of X7 is forward scattering explained by the Mie theory. This scatter mechanism describes dust grains as an emitter with the mentioned forwarded scattering. Single and multiscattering events occur where, e.g., dust emits and transmits stellar light, which is reemitted by close-by grains. As long as S50 is embedded in the dusty shell X7, the ionized and blueshifted Brγ emission is symmetrically distributed following the aligned dust grains. After 2009, the peak emission of the L'-band emission can be observed closer to the tail of X7, whereas the gaseous tip becomes more prominent. ¹¹

Overall, we conclude that a projection scenario that describes a random encounter between S50 and X7 is highly unlikely but not excluded.

Besides the observed decreased projected source size in 2009–2010, we find that the position angle (with respect to Sgr A*) is increasing faster as the shell of S50 is aligned toward the SMBH (Table 9). Even though a change in the position angle is expected since the proper motion of the X7/S50 system is directed toward the north (Mužić et al. 2010), the gas and dust shell points to/aligns with a position 0.45'' north of the SMBH (see Figures 2 and 5) in 2018. Comparing the position angle of 2006 and 2018 shows a growth of about 40%. If S50 were located close to the position of the tip of the bow shock at a distance R₀, a growth by around 12% would be expected in 2018. However, assuming the chance of reading uncertainties, the position angle of 60° in 2018 between Sgr A* and X7 marks a lower limit. The observations and the measured properties suggest distinguishing the description of X7 in pre 2009.26 and post 2009.46 since the object shows a discontinuous development as a function of time. Summing up the observational results leads to two assumptions: either X7 is a tidally stretched object ¹² where the head is on its way toward Sgr A* (A), or the dust and gas shell seems to be ripped apart by an unknown interaction (B).

• (A) The trajectory of the head, as shown in Figure 4, shows a clear trend toward Sgr A*. The distance between the SMBH and the gaseous head of X7 decreased by around 20% over almost two decades. Taking into account the proper motion of the S50/X7 system, this is expected. Even though a clear trend can be observed, projection effects could also play a role because of the orbit of S50 (see Appendix D, Figure 12). Studying the projected positions of the head, tail, and S50 with respect to Sgr A* (Figure 4), implies that the R.A. distance of the head remains almost constant. If the head were attracted to Sgr A*, we would not observe a preserved dusty shell of X7 because the front would simply accelerate toward the SMBH with respect to S50 and the tail. Hence, the shape of the Brγ emission in 2018 might be explained by the forward (and backward) single- and multiscattered stellar light of S50. If upcoming observations can confirm the observed decoupling of the head from S50 and its tail, it might trigger the flaring activity of Sgr A* above the statistical level (Witzel et al. 2012). Please consider Appendix D (Figure 11) for a possible outlook.
• (B) As discussed before, the Brγ line map of 2009 (Figure 3) but also the size of the L'-band continuum detection (Table 9 and Figure 2) marks a noticeable step in the discontinuous evolution of X7. Adding the growth of the position of the angle of X7, the increasing distance between the head and S50, and the relative position of the shell and the S-star to the calculation creates the assumption that we observe a dissolving event. Since the overall shape of the dust shell as observed with NACO seems to be preserved even though a clearly increased elongation can be observed, it is safe to assume that the shell remains intact. Hence, clear evidence for the scenario of a destroyed shell cannot be given.

Considering the observational results discussed here leads to the problem of the ongoing spatial misplacement of S50 with respect to X7 and the growing position angle. We elaborate on this in the following subsections.

Recently, Vorobyov et al. (2020) modeled the behavior of the gas and dust features of protoplanetary disks that move with a supersonic motion in a dense ambient medium. Considering the Brγ emission in Figure 3 in 2009 in combination with the related L'-band emission size (Table 9), we conclude that there might be a prominent decoupling of gas and dust as discussed by Henney & Arthur (2019). As discussed, the timescales of the cited work does not fit the observation. Hence, the observations suggest the presence of a disturbing event. We speculate that this event has been caused by the close fly-by (in projection) of S33, which would at least partially explain the almost compact Brγ line map emission in 2009 and the discontinuous evolution of the projected L'-band size of X7 (see Table 9). A critical parameter of this speculative scenario is the three-dimensional distance and therefore the position of S50/X7 and S33 with respect to each other.

To give an estimate of the three-dimensional distance between S33 and S50, we use the related proper motion (v_t ) and line-of sight (LOS) velocity (v_r ). For v_r , we use a lower limit of around 500 km s⁻¹ (Mužić et al. 2010). For deriving an LOS velocity, an averaged value of the observed H₂Q(1) ¹³ and H₂Q(3) ¹⁴ absorption line is used. Hence, for the v_t of S50 we derive a value of around 350 km s⁻¹ in 2018 (see Appendix B (Figure 10) and Table 8). This velocity estimate results in an approximate three-dimensional velocity of ${({{v}_{r}}^{2}+{{v}_{t}}^{2})}^{-1/2}\,\approx 600$ km s⁻¹. This results in an approximate distance d toward Sgr A* of d_S50 ≈ 0.047 pc ≈ 1.19''. From Ali et al. (2020), we use the three-dimensional position of S33 based on their presented orbit plots. We find that the three-dimensional distance of S33 in 2009 with respect to Sgr A* is about 12''. Because the three-dimensional distance of S50 with respect to Sgr A* is a lower limit, we set the distance of S33 to S50 at about 0.01'' or 120 au.

Considering the derived three-dimensional distance between S33 and S50, the modeled interaction between an intruder and the host star with an envelope as presented in Vorobyov et al. (2020) could be a possibility. A detailed model should answer the question about the stellar-wind interaction with the ambient wind (Yusef-Zadeh et al. 2020) but is beyond the scope of this work.

Furthermore, it should be mentioned that O'Gorman et al. (2015) and Wallström et al. (2017) presented Atacama Large Millimeter/submillimeter Array observations that do not show a symmetrical dust/gas distribution of the envelope related to the host star (which happens to be in both cases a giant). Wallström et al. (2017) observed a so-called "spur," which describes an asymmetric gas feature related to the host star. This spur could be compared to the dust and gas shell X7 of S50. Wallström et al. (2017) argue that this spur might be created by a sporadic eruption event of the host star. Nevertheless, Zajaček et al. (2020) recently modeled the depletion of red giants and showed that the detached and shocked envelope of the host star can suffer from the interaction with Sgr A*. Even though Schartmann et al. (2018) used stellar winds to model the S2 pericenter passage, it is shown that the presence of a SMBH results in an asymmetric mass distribution. If the gas/dust shell got detached and its length scale increased beyond the stellar Hill radius, the gravitational influence of Sgr A* would dominate the evolution of X7 as was described by Eckart et al. (2012) and numerically modeled by Zajaček et al. (2014).

From the multiwavelength analysis with NACO and SINFONI in the H-, K-, L-, and M-bands, and the modeled SED, we find that the X7/S50 system consists of a stellar component in combination with the internally heated dusty envelope (Figure 7).

Zoom In Zoom Out Reset image size

Comparing the SINFONI Brγ-line map of 2006 and 2018 with the NACO L'-band continuum observations shows the gas- to dust-component ratio is around 1:2–1:3, which are typical values for HAe/Be or T-Tauri stars (Mannings & Sargent 2000). The weak H₂-absorption lines (Appendix B, Figure 10) underline the possibility of observing a young stellar object (YSO) as discussed in Mužić et al. (2010). The theoretical modeling of the dust and gas of X7 strengthen the possibility of a YSO.

Additionally, Rivinius et al. (1997) reported wind variations for early-B hypergiants with mass-loss rates of several 10⁻⁶ M_⊙ yr⁻¹. These variations are also investigated by Muratorio et al. (2002). In both cases, the P-Cygni profile of highly excited [Fe iii] multiplets/lines are indications of a complex wind interaction with the stellar source. Even if we do not find a prominent P-Cygni profile in the spectrum, a nondetection can be explained by the high sky emission line variations that lead to over/undersubraction effects as shown by Davies (2007). Finding a P-Cygni feature would increase the complexity of the X7 system since there would be wind-wind-accretion processes that should be a part of the mentioned model. The wind launched at the position of Sgr A* would be accompanied by stellar winds of S50. Therefore, the S50 dust and gas accretion would be influenced by the aforesaid wind-wind process.

Furthermore, the origin of the excited [Fe iii] lines is still not clear (Peißker et al. 2019, 2020b) even though we speculate the detection could be linked to the area and the Brγ bar (Schödel et al. 2011; Peißker et al. 2020c). However, Wolf & Stahl (1985) mentioned that higher excited [Fe iii] lines could have been pumped by He i lines. In the spectrum of X7, we find a strong blueshifted He i line at 2.058 μm ¹⁵ with a matching LOS velocity. Hence, we consider the pumping of the forbidden Fe lines as a possible explanation. For the sake of completeness, we note that all of the four most prominent emission lines in the present K-band spectrum in Figure 1 is accompanied by a less intense line that is related to the source X7.1/G5. In addition, we do find a redshifted H₂ line (about 650 km s⁻¹) at 2.228 μm (transition v = 1–0 S(0)). Because of the direction of the Doppler-shifted H₂ line, this emission might be related to another species.

From the results shown here and the discussed scenarios, we conclude that the stellar source of X7 can be associated without any doubt with the S-cluster star S50, which confirms the analysis of Mužić et al. (2010). As implied by the H₂ absorption lines, the LOS velocity of the star is blueshifted. Hence, the Doppler-shifted direction of the stellar LOS velocity matches the emission lines of the surrounding envelope, which also shows a blueshifted motion.

The shape of the bow shock in 2002 is almost spherical and Wilkinoide. With the presented COMIC/ADONIS+RASOIR data of 1999, we find evidence that earlier L'-band data than 2002 confirm the trend of a "growing" dusty envelope.

The two distinct observed processes, the LOS velocity, and the star/envelope evolution underline the prominent dynamical process that highlights the uniqueness of the X7/S50 system.

Along the X7/S50 source, we observe a strong and prominent velocity gradient in 2018. Considering the existence of a formed wind at the position of Sgr A* or IRS16, we assume that this might be the origin of the gradient. In 2009, it seems that the envelope starts to interact with the nearby S-cluster star S33 since we trace indications of this possible interaction in the same year (Figure 3). The L' NACO data shows that the tail of X7 becomes brighter between 2010 and 2018. We predict that this gain in brightness will likely continue in the future. We also speculate that the ongoing interaction of S33 and Sgr A* with the shell of S50 could lead to the partial destruction of the bow shock.

As we have observed and presented in Figure 2 , and also listed in Table 9, the shell of S50 is pointing in projection above Sgr A*. As proposed by Wardle & Yusef-Zadeh (1992), strong stellar winds arising from the IRS16 complex are responsible for the creation of the mini-cavity. The authors discussed an observed 2.217 μm emission line at the position of the mini-cavity (see also Lutz et al. 1993), which can most likely be related to the [Feiii] multiplet observed in several dusty sources west of Sgr A* (Ciurlo et al. 2020; Peißker et al. 2020b). The ionized iron multiplet can also be observed for X7/S50 as shown in Figure 1. If we exclude the possibility of a wind arising at the position of Sgr A*, the excitation of iron as well as the position angle (Table 9) could be linked to stellar winds from IRS16. The SMBH would be responsible for refocusing the wind (Figure 8) and sources leaving the "slip stream of Sgr A*" would suffer from this interaction. This dynamical evolution of the gaseous and dusty shell of the X7/S50 system underlines the need for a constant survey of the GC region in various bands.

Zoom In Zoom Out Reset image size

If a wind responsible for the alignment and evolution of the X7/S50 system is indeed arising at the position of Sgr A*, the apparent change in the position of the angle with respect to Sgr A* is unexpected. Since we clearly observe the evolution of the elongation of the X7/S50 system, it may be explained by a temporarily active wind phase of Sgr A* as indicated by Morris & Serabyn (1996). Speculatively, this could contribute to the "paradox of youth" (Ghez et al. 2003) where star formation is "allowed" for a short period of time. Nevertheless, in combination with the X7 proper motion (Mužić et al. 2010) directed toward the north, the alignment angle of X7/S50 may have been induced to the system before 2009. After 2009, the wind activity may have been decreased, while the position angle increased (Table 9) because of the proper motion of the X7/S50 system.

NIR and MIR instruments will play a key role in investigating the evolution of the X7/S50 system. The prominent detection of X7 in the L'- and M'-bands promises successful observations with Mid-Infrared Instrument (MIRI; James Webb Space Telescope, see Bouchet et al. 2015; Ressler et al. 2015; Rieke et al. 2015), Mid-infrared ELT Imager and Spectrograph (METIS; ELT, see Brandl et al. 2018), and Multi-AO Imaging Camera for Deep Observations (MICADO; ELT, see Trippe et al. 2010). MIRI and METIS will be able to finalize the investigation of the possible clumpiness of X7, which could be used for theoretical models (e.g., the filling factor, see Peißker et al. 2020c). With a more accurate result, we will be able to precisely determine the density and therefore the mass of the dusty shell. Furthermore, we are able to search for more complex emission lines in the local LOS interstellar medium (ISM), for example, NH₃. Additionally, gas emission lines, e.g., CO and HCN, can provide a more detailed description of the nature of the X7/S50 system. These gas- and ice-absorption lines can also be used as an additional probe for a stellar disk and a possible YSO. Moultaka et al. (2006) and Moultaka et al. (2009) showed that these lines are useful in determining local extinction values for the ISM (see also Schödel et al. 2010; Peißker et al. 2020c).

Even if we have shown S50 could be associated with the stellar counterpart of X7, a hidden star at a distance of R₀ from the apex of the bow shock should be detectable with MICADO (see the simulated view of the GC with MICADO in Davies & Genzel 2010).

As we have presented in Figure 2, investigating the GC with a wider FOV in the mentioned bands should also reveal more (elongated) sources that might be suffering from the wind that is formed at the position of Sgr A* or at IRS16. We conclude that the upcoming observations of the GC with the ELT will be able to manifest the dynamical influence of the nuclear wind. We can safely assume the X7/S50 system will not be the only source in the GC that is undergoing a dynamical influence. Yusef-Zadeh et al. (2017a) have already shown that YSOs with bipolar outflows can be observed in the environment of the SMBH. Even though we cannot finally answer the question about the nature of the X7/S50 system, we see some weak traces that point toward its YSO nature. If the theoretical models reveal matching parameters of the X7/S50 system with a YSO, the origin of these sources is still not clear. However, the implication of a population of YSOs promises an important cornerstone in the investigation of the direct vicinity of the nearest SMBH that resides in our galaxy.