NONTHERMAL PROPERTIES OF SUPERNOVA REMNANT G1.9+0.3

G1.9+0.3 has been known as a potentially young shell type Galactic supernova remnant (SNR) of very small angular size (Green & Gull 1984). Recently, the interest in this SNR was revived by Reynolds et al. (2008, 2009) who analyzed the expansion rate of the object and deduced an age t_SN of about 100 yr, which makes it the youngest known SNR in the Galaxy. Although the expansion rate was derived by a comparison of radio observations in 1985 and Chandra observations in 2007, this rate has been confirmed very soon thereafter by independent radio observations (Green et al. 2008; Murphy et al. 2008).

According to Reynolds et al. (2008), the line-free X-ray emission has a pure synchrotron origin which clearly indicates that effective particle acceleration takes place, at least for electrons. There are also arguments, like the bilateral symmetry of the X-ray synchrotron emission suggesting a roughly uniform ambient magnetic field, that favor a type Ia origin for G1.9+0.3. Finally, the distance estimate d = 8.5 kpc is based on an analysis of the absorption toward G1.9+0.3.

During the survey of the inner Galaxy by H.E.S.S. in very high energy γ-rays, no emission was reported from the direction to G1.9+0.3 (Aharonian et al. 2006). Therefore, one can derive an upper limit at the level of 2% of the Crab flux above 200 GeV.

For the purpose of a more general study of such an unusual object regarding its nonthermal properties, it is of interest to describe it by a kinetic theory of cosmic ray (CR) acceleration in SNRs, coupled with the gas dynamics of the thermal plasma, as given by Berezhko et al. (1996) and Berezhko & Völk (1997). This model assumes spherical symmetry, although the assumption is later relaxed. Similar models on almost the same physical basis have recently been developed by two other groups (Kang & Jones 2006; Zirakashvili & Ptuskin 2009), whose calculations very well confirm these earlier results.

The kinetic description allows a corresponding analysis of the nonthermal evolution of an SNR at a very early phase. This assumes that the plasma physics underlying especially the temporal dependence of the magnetic field amplification process can be extrapolated to such an early evolutionary phase, where the dynamical behavior of the ejecta plays an essential role. Such a study is reported here. It is combined with a discussion about the influence of the assumption of a smaller distance on the TeV γ-ray  flux.

Following Reynolds et al. (2008), it is assumed that G1.9+0.3 is a Type Ia supernova (Type Ia SN) which expands into a uniform interstellar medium (ISM). Specifically, the object is assumed to eject a Chandrasekhar mass M_ej = 1.4 M_☉ with a total hydrodynamic explosion energy E_SN = 10⁵¹ erg. During an initial period, the ejecta material has a broad distribution in velocity v. The fastest part of these ejecta is described by a power law dM_ej/dv ∝ v^2−k with k = 7 (e.g., Chevalier 1982).

The ISM mass density ρ₀ = 1.4m_pN_H, which is usually characterized by the hydrogen number density N_H, is an important parameter which strongly influences the expected SNR dynamics and nonthermal emission; here m_p denotes the proton mass.

Following Reynolds et al. (2008), a distance d = 8.5 kpc is adopted for the main part of the paper. The observed shock size R_s = 2 pc and shock speed V_s = 14,000 km s⁻¹ are then used to determine the SNR age t_SN and the ISM number density N_H for the given source distance d.

As reviewed earlier (Völk 2004; Berezhko 2005, 2008) and elaborated most recently in detail in Berezhko et al. (2009), the key parameters of the theoretical model (proton injection flux density, given by a constant injection parameter η ≪ 1 times the thermal particle flux density into the shock, electron–proton ratio below the synchrotron cooling range, later denoted as K_ep, and magnetic field amplification) can be estimated in a semi-empirical way from a fit of the theoretical solution to the observed synchrotron emission spectrum if the characteristics of the initial explosion and the relevant astronomical parameters are known. For this purpose, the nonlinear aspects of the kinetic description are required.

Of the above-mentioned processes, the magnetic field amplification (Bell 2004) is the least well understood. It is connected with the strong nonlinear excitation of magnetic field fluctuations in the shock precursor by the accelerating energetic particles (McKenzie & Völk 1982; Lucek & Bell 2000; Bell & Lucek 2001; Bell 2004). Therefore, it is assumed here that these fluctuations lead to Bohm diffusion of the energetic particles in this amplified field. Such a bootstrap mechanism can be approximately justified by the results of recent particle simulations (Reville et al. 2008). In addition, the amplification process is strongly dissipative, as shown by hydromagnetic and kinetic simulations (Bell 2004; Zirakashvili et al. 2008; Reville et al. 2008; Niemiec et al. 2008; Riquelme & Spitkovsky 2009; Ohira et al. 2009). Therefore, it must be accompanied by strong gas heating within the precursor region r>R_s due to wave dissipation, adopted in the present model in the form (∂e_g/∂t)_diss = −α_Hc_A∂P_c/∂r with α_H = 1 (Berezhko et al. 1996; Berezhko & Völk 1997), where e_g is the gas thermal energy density, P_c denotes the energetic particle pressure, and where, in a second bootstrap mechanism, $c_{\rm{A}}=B(4\pi \rho)^{-1/2}$ is the Alfvén velocity in the amplified field B (see below). The value α_H = 1 corresponds to the assumption that the Alfvén wave field excited within the precursor reaches amplitudes which are very much smaller than the maximal amplitudes which could be reached if the whole work −c_A∂P_c/∂r done by CRs went into wave excitation. In such a case, this work goes almost completely into gas heating due to the wave damping. The gas thermal pressure just ahead of subshock is in this case considerably larger than the pressure of the magnetic field (Berezhko 2008). Therefore, the subshock can be treated approximately as a pure gas shock. This approximation will later be re-examined through the approximate inclusion of the amplified field and its associated turbulent gas motions in the subshock dynamics.

The magnitude of field amplification in all young SNRs is such that a non-negligible fraction of the shock energy ρ₀V²_s is converted into magnetic field energy (e.g., Berezhko 2008). In fact, a time dependent, amplified upstream magnetic field strength

$$\begin{equation} B_0(t)=B_0(t_\mathrm{SN})[V_\mathrm{s}(t)/V_\mathrm{s}(t_\mathrm{SN})]^{\delta } \end{equation} \tag{ 1 }$$

is used here, where the theory parameter B₀(t_SN) is the rms field strength at the present epoch t_SN and is estimated by a comparison of the theoretically calculated synchrotron spectrum with the observed one (see below). Such a form of the time dependence of the amplified magnetic field with δ ≈ 1 is consistent with the interior field strengths estimated from observational results for a number of young SNRs (e.g., Völk et al. 2005). Since Bell (2004) even estimated a dependence B₀(t) ∝ V^3/2_s(t) for the field amplification due to the nonresonant streaming instability alone, the case of δ = 3/2 will also be examined here. The radial dependence of the rms magnetic field strength B(r, t) in the shock precursor is then modeled by B(r, t) = B₀(t)ρ(r, t)/ρ₀, where ρ₀ is the far upstream (interstellar) density. In the overall conservation relations for momentum and energy of the system, the magnetic field strength in the upstream ISM is B_ISM < B₀. Thus, B₀/B_ISM is the field amplification factor by the accelerating energetic particles alone. An analysis performed for a number of young SNRs shows (e.g., Berezhko 2008) that the field strength B₀, required to fit the observed synchrotron spectrum, is well within the range expected from theoretical estimates (e.g., Bell 2004; Pelletier et al. 2006).

The observed X-ray morphology of SNR G1.9+0.3 agrees with the theoretical expectations regarding the morphology of ion injection and the corresponding morphology of magnetic field amplification for a Type Ia SN (Völk et al. 2003). It is therefore consistent with a correction for the spherically symmetric solution by a renormalization factor f_re ≈ 0.2 of the energy density of nuclear particles, like in the case of SN 1006 (Berezhko et al. 2009). In this case, the electron–proton ratio K_ep, calculated under the assumption of spherical symmetry, should be increased by a factor 1/f_re.

The calculated evolution of the gas dynamical variables of G1.9+0.3 is shown in Figure 1. The observed shock radius R_s and shock speed V_s are fitted at the age t_SN = 80 yr and for an ISM hydrogen number density N_H = 0.018 cm⁻³ (Figure 1(a)). Since the remnant is in the free expansion phase, it is approximately consistent with an analytical self-similar solution R_s ∝ t^4/7 (Chevalier 1982). Note that an explosion model with an exponential ejecta velocity profile gives slightly larger values of the age, t_SN = 100 yr, and of the ISM density, corresponding to a hydrogen number density of about N_H ≈ 0.03 cm⁻³, for the same assumed distance of d = 8.5 kpc (Reynolds et al. 2008).

Zoom In Zoom Out Reset image size

The calculated radius R_c of the contact discontinuity (CD) and the CD speed V_c are also shown in Figure 1(a). One can see that the ratio R_c/R_s is rather small. At the current epoch R_c/R_s ≈ 0.9.

To obtain a good fit for the observed synchrotron spectrum (see below), first of all a proton injection rate η = 10⁻³ is required. It leads to a nonlinear modification of the shock which, at the current age of t = 80 yr, has a total compression ratio σ ≈ 4.6 and a subshock compression ratio σ_s ≈ 3.6 (Figure 1(b)). In addition, an electron–proton ratio K_ep ≈ 5 × 10⁻³ and an upstream amplified magnetic field strength at the current epoch B₀(t_SN) ≈ 100 μG are required. This implies a downstream magnetic field strength of B_d ≈ 460 μG.

All the above quantities are almost the same for the two different time dependences of the magnetic field B₀(t) within the evolutionary period t = t_SN ± 100 yr. The only difference between these cases is a slightly more rapid increase of the shock modification (characterized by the shock compression ratio σ(t)) in the case δ = 3/2 compared with the case δ = 1 due to the increase of the Alfvénic Mach number M_A ∝ V_s/B₀ ∝ V^1−δ_s ∝ t^3(δ−1)/7.

To explain these results it is noted that, as in other, similar cases of strong, modified shocks, the existing measurements permit an estimate of the three parameters of the theoretical model. This takes into account the following influence of the parameters on the synchrotron spectrum. (1) Since the high-energy electrons undergo strong synchrotron losses, and since during the acceleration process they dominate the nonthermal electron pressure P^e_c, the flux of X-ray synchrotron emission which they produce νS_ν is approximately proportional to the total energy flux of nonthermal electrons F_e ∝ P^e_cV_sR²_s ∝ K_epP_cV_sR²_s and is only weakly sensitive to the magnetic field strength B₀. The spectrum of high-energy protons N(p), which give the main contribution to the total CR pressure P_c ∼ ρV²_s, is only weakly sensitive to the injection parameter η in the case of a modified shock. Therefore, the fit of the observed X-ray flux mainly determines the value K_ep of the electron–proton ratio. The spectrum of accelerated electrons is calculated in absolute numbers. It is expressed in terms of K_ep, due to the presumably dominant dynamical role of protons. (2) Values α>0.5 of the radio spectral index α = −dln S_ν/dln ν, as observed in young SNRs, require a modified shock with σ_s < 4 < σ (This also implies a curved electron spectrum that hardens toward high frequencies.) The value of α is mainly determined by the subshock compression ratio σ_s, which in turn is determined mainly by the proton injection rate η. Therefore, the fit of the measured spectral shape of the radio synchrotron emission gives mainly the required value of the injection parameter η. (3) Finally, since the radio emission flux value $S_{\nu }\propto K_{\rm{ep}}\nu ^{-\alpha }B_\mathrm{d}^{\alpha +1}$ is strongly dependent upon the magnetic field strength, its value B_d is derived from the fit to the observed amplitude of the radio spectrum. Thus, three measured characteristics of the synchrotron spectrum—X-ray flux, shape, and amplitude of the radio emission—make it possible to obtain an estimate of the three relevant "theory parameters" B_d, K_ep, and η even though this is not a simple three-step procedure but an iterative procedure, minimizing the combined χ²-value (see Berezhko et al. 2009, for details).

Note that the magnetic field value B_d can also be estimated by another, independent method, namely from a fit of the observed spatial fine structure of the X-ray emission. In all the cases where such measurements exist, both methods give consistent values of B_d (e.g., Völk et al. 2005). Unfortunately, the fine structure of the X-ray emission is not determined yet for G1.9+0.3. Therefore, this consistency check cannot be made at present.

The uncertainties of the estimated values of η, B_d, and K_ep depend upon the quality of the measurements of the synchrotron spectrum and can be rather small, as was recently demonstrated for the case of SN 1006 (Berezhko et al. 2009). In the case under consideration, it is about 20% for B_d and K_ep and 30% for η.

It should also be noted that due to the very low age of the SNR and the low ISM density the expected thermal X-ray emission is far below the observed X-ray flux, which is therefore completely dominated by the nonthermal component.

The required proton injection rate η = 10⁻³ is considerably higher than the critical value η_* ≈ 10⁻⁴, which separates a nonlinearly modified shock with η>η_* from an unmodified state, resulting from very low injection rates η < η_* (see Berezhko & Ellison 1999, Equation (38)). The relatively weak shock modification is the result of the very large magnetic field B₀, which leads to strong gas heating within the precursor region r>R_s.

In order to check the sensitivity of the results to the adopted value of the parameter α_H, calculations with 0.02 ⩽ α_H < 1 were performed. It turned out that even for α_H = 0.02, the shock properties are not very far from the case α_H = 1: the required far upstream magnetic field value is B₀ = 125 μG, and the shock compression ratios are σ = 6.3 and σ_s = 3.99.

It is noted here that the results of Vladimirov et al. (2008) suggest a stronger sensitivity of the shock properties to the value of α_H. The reason is that in the considerations of these authors, the parameter α_H determines not only the gas heating through the term −α_Hc_A∂P_c/∂r, but also the upstream magnetic field generation through a complementary term (α_H − 1)c_A∂P_c/∂r. Such an approach implies that magnetic field amplification takes place only due to resonant excitation of Alfvén waves and, in addition, according to such a quasi-linear expression. In contrast, the present model allows that the magnetic field is also amplified nonresonantly (Bell 2004; Pelletier et al. 2006). Secondly, the determination of the amplified field B₀(t_SN) in the present semi-empirical model uses the synchrotron observations of the source. For both these reasons, the magnetic field amplification and the gas heating are not connected by a simple relation.

3.1. Subshock Dynamics with Amplified B-field

3.2. Charged Particle and γ-ray Spectra

With the renormalization f_re = 0.2, the nuclear CRs inside G1.9+0.3 SNR contain (Figure 1(c))

$$\begin{equation} E_\mathrm{c}\approx 0.0025E_\mathrm{SN} \approx 3\times 10^{48}\ \mbox{erg}. \end{equation} \tag{ 2 }$$

The volume-integrated (or overall) CR spectrum

$$\begin{equation} N(p,t)=16\pi ^2p^2 \int _0^{\infty }dr r^2 f(r,p,t) \end{equation} \tag{ 3 }$$

has, for the case of protons, almost a pure power-law form N ∝ p^−γ over a wide momentum range from 0.1m_pc up to the cutoff momentum p_max ≈ 3 × 10⁶m_pc (Figure 2). This value p_max ∝ R_sV_sB₀ is limited mainly by the finite size and speed of the shock, its deceleration, and the adiabatic cooling effect in the downstream region (see Berezhko 1996, for details). As pointed out above, particle diffusion is approximated by Bohm diffusion in the amplified magnetic field B, cf. Equation (1). It is important to note that the calculated value of p_max is therefore an upper limit, because it is assumed that up to the cutoff all particles "see" the amplified field everywhere.

Zoom In Zoom Out Reset image size

Consequently, G1.9+0.3 represents the youngest SNR where the accelerated proton spectrum extends up to the so-called knee energy. Such a maximum proton energy indeed appears required to describe the overall CR spectrum for energies up to 10¹⁷ eV (Berezhko & Völk 2007).

The shape of the overall electron spectrum N_e(p) deviates from that of the proton spectrum N(p) at high momenta p>p_l ∼ 10³m_pc on account of the synchrotron losses during the electron residence time in the downstream region (Figure 2). Within the momentum range p_l < p < p^e_max, the electron spectrum is considerably steeper, N_e ∝ p⁻³, due to synchrotron losses taking place in the downstream region after the acceleration at the shock front. The maximum electron momentum p^e_max ≈ 10⁵m_pc corresponds closely to the result obtained by equating the synchrotron loss time and the acceleration time.

Figure 3 illustrates the consistency of the synchrotron spectrum, calculated for the above-mentioned best set of parameters with the observed spatially integrated spectra.

Zoom In Zoom Out Reset image size

As mentioned above, values α>0.5 of the radio spectral index α = −dln S_ν/dln ν, as observed in young SNRs, require a curved electron spectrum that hardens toward higher energies, as predicted by nonlinear shock acceleration theory. To have α = 0.62 in the radio range, as observed for G1.9+0.3 (Green et al. 2008), requires efficient CR acceleration with a proton injection rate η = 10⁻³ which leads to the required shock modification, and also leads to the high magnetic field value above. As is clear from Figure 3, a good fit of the observed X-ray energy flux can only be achieved due to the softening of the synchrotron spectrum for ν ≳ 10¹⁵ Hz, which is due to the strong synchrotron losses of electrons with momenta p>p_l ≈ 700m_pc. Using the known dependence p_l ∝ B⁻²_dt⁻¹, at t = t_SN one can immediately estimate the required value of the interior magnetic field value B_d ≈ 500 μG, consistent with the above determination from the radio spectrum. The calculated hard power-law X-ray spectrum continues almost up to 50 keV, making this source in principle attractive to be observed with Suzaku and INTEGRAL and the future Astro-H X-ray instrument.

A very important question for every young SNR is whether the existing data indeed unavoidably require efficient proton acceleration accompanied by strong magnetic field amplification. In order to explore the alternative possibility, Figure 3 presents in the dotted curve a synchrotron spectrum which corresponds to a hypothetical leptonic scenario with a proton injection rate so small (η ≪ 10⁻⁴) that the accelerated nuclear CRs do not produce any significant shock modification and therefore also no magnetic field amplification. This corresponds to the test particle limit, when the distribution function of shock accelerated electrons has the form

$$\begin{equation} f_\mathrm{e}= A p^{-4}\exp (-p/p_\mathrm{max}), \end{equation} \tag{ 4 }$$

where the amplitude A and the value of the cutoff momentum p_max are determined by the fit to the observed synchrotron spectrum for a given interior magnetic field value B_d. Since magnetic field amplification is not expected in this case, the downstream magnetic field cannot be larger than the MHD-compressed ISM field B_ISM ≈ 5 μG. The maximal possible downstream field B_d ≈ 20 μG is adopted which corresponds to the minimal number of accelerating electrons, and therefore the γ-ray  emission produced by these electrons is also minimal. The synchrotron spectrum for the leptonic test particle scenario in Figure 3 corresponds to the maximal electron energy ^e_max = p^e_maxc = 6 TeV, determined by the Chandra observation. There are two differences in the synchrotron spectra, corresponding to these two scenarios. The high-injection scenario leads to a soft radio spectrum S_ν ∝ ν^−α with power-law index α = 0.62, whereas in the test particle case α = 0.5. On the other hand, the two spectra behave essentially differently at X-ray frequencies ν ≳ 10¹⁸. This demonstrates that only in the high-injection case with its high, amplified magnetic field value B_d ≈ 460 μG the spectrum S_ν(ν) has a smooth cutoff, consistent with the observations (see Figure 3). In the test particle case, the spectrum S_ν(ν) has too sharp a cutoff to be consistent with the observations.

If G1.9+0.3 was indeed a Type Ia SN, then the explosion parameters E_SN, M_ej, and k are known. Assuming the value d = 8.5 kpc for the source distance and using the age t_SN and the ambient gas number density N_H from fits to the observed astronomical parameters size and expansion rate, and also using the "theory parameter" values η, K_ep, and B₀, estimated from the fit to the synchrotron spectrum, one can predict the γ-ray flux for the assumed source distance d.

In Figure 4, the calculated γ-ray spectral energy distributions (SEDs) due to π⁰-decay and inverse Compton (IC) collisions are presented for the source distance d = 8.5 kpc together with the sensitivities of the Fermi and H.E.S.S. instruments. For the modified shock, consistent with the observed synchrotron emission, the expected total TeV γ-ray SED is _γF_γ ≈ 3 × 10⁻³ eV cm⁻² s⁻¹. Such a flux is too low for an H.E.S.S. detection in ∼50 hr by a factor of the order of 30. The π⁰-decay flux is only about 6% of the IC γ-ray  flux as a result of the low gas density. Since according to recent estimates (Porter et al. 2006), the Galactic interstellar optical and infrared radiation fields in the inner Galaxy are considerably higher than previously thought, for this region the calculation of the IC flux was performed on the basis of those estimates. The higher radiation field leads to an increase of the IC gamma-ray flux at energies _γ < 1 TeV by an order of magnitude compared with a standard interstellar radiation field in the solar neighborhood (e.g., Drury et al. 1994). Nevertheless, as indicated above, the expected TeV emission flux is still far below the H.E.S.S. sensitivity.

Zoom In Zoom Out Reset image size

As can be seen from Figure 4, the TeV γ-ray  flux expected in the unmodified leptonic scenario considerably exceeds the H.E.S.S. sensitivity, corresponding to ∼50 hr of observation time. Since the region of the Galactic center was already explored by H.E.S.S. for times of more than 100 hr without detection of G1.9+0.3, this purely leptonic test particle scenario should be rejected as in all similar cases of Type Ia SNe (Völk et al. 2008).

3.3. Dependence on the Assumed Source Distance

It is, however, to be noted that the expected γ-ray  flux is very sensitive to the assumed source distance d. Therefore, SNR G1.9+0.3 could be a potential γ-ray  source if the actual distance was lower than 8.5 kpc. Qualitatively, the dependence of the expected γ-ray  flux on distance can be understood if one takes into account that the π⁰-decay γ-ray  energy flux _γF_γ ∝ M_swe_c/d² is proportional to the mass of gas, M_sw = 4πR³_sρ₀/3, swept up by the SN shock, and to the energy density e_c of the CRs producing γ-rays  of given energy. Since for high acceleration efficiency e_c is proportional to the shock kinetic energy density ρ₀V²_s, one can write

$$\begin{equation} \epsilon _{\gamma } F_{\gamma } \propto N_\mathrm{H}^2V_\mathrm{s}^2R_\mathrm{s}^3/d^2. \end{equation} \tag{ 5 }$$

For fixed explosion energy E_SN, the distance d and ISM density N_H are connected by the relation

$$\begin{equation} N_\mathrm{H}\propto d^{-7}, \end{equation} \tag{ 6 }$$

because in the free expansion phase the SNR radius R_s ∝ d is determined by the expression R_s ∝ N^−1/7_Ht^4/7 (Chevalier 1982), where the SNR age t ∝ R_s/V_s is fixed if the angular size and angular expansion speed are known as in our case of G1.9+0.3.

Taking also into account that for a fixed angular expansion rate of the object V_s ∝ d, one obtains

$$\begin{equation} \epsilon _{\gamma } F_{\gamma } \propto d^{-11}. \end{equation} \tag{ 7 }$$

According to this relation, a mere 30% reduction of the source distance leads to an increase of the expected γ-ray  flux by a factor of more than 10. This is illustrated in Figure 4, where γ-ray  spectra are presented that were calculated for the distance value d = 5.6 kpc. In this case, the shock velocity and size could be fitted at the same age t_SN = 80 yr and for an ISM hydrogen number density N_H = 0.2 cm⁻³. A similar fit for the radio and X-ray data as in Figure 4 could be achieved with an electron–proton ratio K_ep = 8 × 10⁻⁴ and a downstream magnetic field strength B_d = 670 μG. It is clear that G1.9+0.3 could be visible in TeV γ-rays  by future instruments like the Cherenkov Telescope Array (CTA), if the actual distance was not larger than d = 5.6 kpc.

The theoretical, spatially integrated radio synchrotron flux slowly increases with time, as can be seen in Figure 5, essentially due to the rapidly increasing total number of accelerated electrons in the increasing SNR volume ∝R³_s.

Zoom In Zoom Out Reset image size

The X-ray synchrotron flux is expected to be nearly constant in time (Figure 5). This is mainly due to the strong synchrotron cooling of the highest energy electrons which produce the X-ray synchrotron emission.

The TeV γ-ray  flux is expected to increase with time as well (Figure 5), mainly due to the increase of overall number of CRs with energy above 10 TeV.

3.4. Renormalization of the Nuclear Particle Spectra

The existing data for G1.9+0.3, when analyzed within the framework of the nonlinear kinetic theory of CR production in SNRs described above, are consistent with a type Ia explosion in a rarefied medium at a distance of d = 8.5 kpc, whose nuclear CR spectrum reaches the energy of the "knee" in the observed Galactic CR spectrum at the present epoch. This conclusion also concerns the derived strong magnetic field amplification. A test particle, purely leptonic gamma-ray scenario, is inconsistent with existing TeV gamma-ray observations with the H.E.S.S. telescope array. However, the data set is not complete enough to unequivocally determine the value of the source distance. Since the expected γ-ray  flux is very sensitive to the distance d, F_γ ∝ d⁻¹¹, a detection of the γ-ray  flux from G1.9+0.3, as improbable as it may be, would yield the distance. However, in the case when G1.9+0.3 is located near the Galactic center (d = 8.5 kpc), the expected TeV γ-ray  energy flux is so low, _γF_γ ≈ 5 × 10⁻¹⁵ erg cm⁻² s⁻¹, that it is not detectable with present instruments. It is clear that for a distance that was not larger than d = 5.6 kpc G1.9+0.3 could be visible with future instruments like CTA which can be assumed to have a sensitivity of ∼1 mcrab at a few 100 GeV (Bernlöhr 2009).

We are indebted to Dr. Stephen Reynolds for providing us the X-ray spectra for G1.9+0.3 from Chandra in physical units. This work has been supported in part by the Russian Foundation for Basic Research (grants 07-02-00221, 10-02-00154), Federal Agency of Science and Innovations (contract 02.740.11.0248), Program of PRAS No. 16, and by the Leading Scientific Schools of Russia (project 3526.2010.2). E.G.B. and L.T.K. acknowledge the hospitality of the Max-Planck-Institut für Kernphysik, where part of this work was carried out.