myXCLASS

creating single, synthetic spectra

The myXCLASS function is the core of XCLASS and models a spectrum by solving the radiative transfer equation for an isothermal object in one dimension, called detection equation [Stahler_Palla_2005]. Here, the user can define, if LTE is assumed, i.e. if the source function is given by the Planck function of an excitation temperature [2] or, if a Non-LTE description within the escape probability method is used instead. The myXCLASS function is designed to describe line-rich sources which are often dense, where LTE is a reasonable approximation. Additionally, a Non-LTE description requires collision rates which are available only for a few molecules, see Sect. “Non-LTE description of molecules”. Details of the myXCLASS function are described in Sect. “myXCLASS”.

Example call of the myXCLASS function:

>>> from xclass import task_myXCLASS
>>> import os

# get path of current directory
>>> LocalPath = os.getcwd() + "/"

## define path and name of molfit file
>>> MolfitsFileName = LocalPath + "files/12C.molfit"

## define path and name of obs. xml file
>>> ObsXMLFileName = LocalPath + "files/SMF_obs.xml"

# call myXCLASS function
>>> spectra, log, te, IntOpt, JobDir = task_myXCLASS.myXCLASSCore(MolfitsFileName = MolfitsFileName, \
                                                                  ObsXMLFileName = ObsXMLFileName)

What is myXCLASS?

The myXCLASS function is able to model a spectrum with an arbitrary number of molecules where the contribution of each molecule is described by an arbitrary number of components. These could be identified with clumps, hot dense cores etc. The myXCLASS function offers the possibility to describe the position and shape of each component in a simple 1-d, see Sect. “Simple 1-d description”, or in an advanced 3-d description, see Sect. “Stacking” and Sect. “Sub-beam description”. Additionally, myXCLASS is able to take the local and non-local overlap between different lines into account, see Sect. “Local-overlap of neighboring lines”.

In general, the myXCLASS function assumes, that all components are aligned in a user-defined stack along the line-of-sight, where the background and the observer are located at the ends, respectively. The solution of the radiative transfer equation for components which are located in a layer (i=1) closest to the background is [3],

(1)\[\begin{split}T_{\rm mb}(\nu) = \sum_{m,c \in i} &\Bigg[\eta \left(\theta^{m,c}\right) \left[S^{m,c}(\nu) \left(1 - e^{-\tau_{\rm total}^{m,c}(\nu)}\right) + I_{\rm bg} (\nu) \left(e^{-\tau_{\rm total}^{m,c}(\nu)} - 1\right) \right] \Bigg] \\ &+ \left(I_{\rm bg}(\nu) - J_\mathrm{CMB} \right),\end{split}\]

where the sums go over the indices \(m\) for molecule, and \(c\) for component, respectively. In the following we will briefly describe each term in Eq. (1).

The beam filling (dilution) factor \(\eta(\theta^{m,c})\) of molecule \(m\) and component \(c\) in Eq. (1) for a source with a Gaussian brightness profile, see below, and a Gaussian beam is given by [4]

(2)\[\eta(\theta^{m,c}) = \frac{(\theta^{m,c})^2}{\theta_t(\nu)^2 + (\theta^{m,c})^2},\]

where \(\theta^{m,c}\) and \(\theta_t\) represents the source and telescope beam full width half maximum (FWHM) sizes, respectively. The sources beam FWHM sizes \(\theta^{m,c}\) (in arcsec) for the different components are defined by the user in the molfit file, described in Sect. “The molfit file”. Additionally, the myXCLASS program assumes for single dish observations (indicated by Inter_Flag = False), that the telescope beam FWHM size is related to the diameter of the telescope by the diffraction limit

(3)\[\theta_t(\nu) = \left(1.22 \cdot \frac{\lambda}{D}\right) \cdot \xi = \left(1.22 \cdot \frac{c_{\rm light}}{\nu \, D}\right) \cdot \xi,\]

where \(D\) describes the diameter of the telescope, \(c_{\rm light}\) the speed of light, and \(\xi = 3600 \cdot 180 /\pi\) a conversion factor to get the telescope beam FWHM size in arcsec. For interferometric observations (indicated by Inter_Flag = True, the user has to define the interferometric beam FWHM size directly using parameter TelescopeSize. In contrast to single dish observations the myXCLASS program assumes a constant interferometric beam FWHM size for the whole frequency range.

In general, the brightness temperature of radiation temperature \(J(T, \nu)\) is defined as

(4)\[J(T, \nu) = \frac{h \, \nu}{k_B}\frac{1}{e^{h \, \nu / k \, T} - 1}.\]

The expression \(J_\mathrm{CMB}\) describes the radiation temperature Eq. (4) of the cosmic background \(T_{\rm cbg}\) = 2.7 K, i.e. \(J_\mathrm{CMB} \equiv J(T_{\rm cbg}, \nu)\).

In Eq. (1), the expression \(S^{m,c}(\nu)\) represents the source function and is according to Kirchhoff’s law of thermal radiation given by

(5)\[\begin{split} S^{m,c}(\nu) &= \frac{\epsilon_l^{m,c}(\nu) + \epsilon_{\rm dust}^{m,c}(\nu) + \epsilon_{\rm free-free}^{m,c}(\nu) + \epsilon_{\rm sync}^{m,c}(\nu)}{\kappa_l^{m,c}(\nu) + \kappa_{\rm dust}^{m,c}(\nu) + \kappa_{\rm free-free}^{m,c}(\nu) + \kappa_{\rm sync}^{m,c}(\nu)} \\ &= \frac{\kappa_l^{m,c}(\nu) \, J(T_\mathrm{ex}^{m,c}, \nu) + \kappa_{\rm dust}^{m,c}(\nu) \, J(T_{\rm dust}^{m,c}, \nu) + \kappa_{\rm free-free}^{m,c}(\nu) \, J(T_{\rm free-free}^{m,c}, \nu) + \epsilon_{\rm sync}^{m,c}(\nu)}{\kappa_l^{m,c}(\nu) + \kappa_{\rm dust}^{m,c}(\nu) + \kappa_{\rm free-free}^{m,c}(\nu) + \kappa_{\rm sync}^{m,c} (\nu)} \\ &= \left(1 - \delta_{\gamma, 0}\right) \\ & \quad \times \Bigg[ \frac{\tau_l^{m,c}(\nu) \, J(T_\mathrm{ex}^{m,c}, \nu) + \tau_{\rm dust}^{m,c}(\nu) \, J(T_{\rm dust}^{m,c}, \nu) + \tau_{\rm free-free}^{m,c}(\nu) \, J(T_{\rm free-free}^{m,c}, \nu) + \epsilon_{\rm sync}^{m,c}(\nu) \cdot l}{\tau_l^{m,c}(\nu) + \tau_{\rm dust}^{m,c}(\nu) + \tau_{\rm free-free}^{m,c}(\nu) + \tau_{\rm sync}^{m,c}(\nu)}\Bigg]\\ & \quad + \delta_{\gamma, 0} \, J(T_\mathrm{ex}^{m,c}, \nu),\end{split}\]

where \(\epsilon_{i}^{m,c}(\nu)\) and \(\kappa_{i}^{m,c}(\nu)\) are the emission and absorption coefficients for line, dust, free-free, and synchrotron, respectively. Additionally, the optical depth is given by \(\tau^{m,c}_i(\nu) = \int \kappa^{m,c}_i(\nu) \, ds\), which assumes that molecules and continuum contributions are well mixed. In older versions, the background temperature could only be defined as the measured continuum offset, which corresponds to the beam-averaged continuum brightness temperature. At the same time, the dust, as agent of line attenuation, was described by column density and opacity. This is practical, because the observable \(T_{\rm bg}\) is used, but does not constitute a self-consistent and fully physical description. Therefore, we now use optionally either a physical (\(\gamma \equiv 1\), defined by the input parameter setting t_back_flag = False) or phenomenological (\(\gamma \equiv 0\), defined by the input parameter setting t_back_flag = True) description of the background indicated by the Kronecker delta \(\delta_{\gamma, 0}\), i.e. \(S^{m,c}(\nu) \equiv J(T_\mathrm{ex}^{m,c}, \nu)\) for \(\gamma \equiv 0\). Note, if \(\gamma \equiv 0\), the definition of the dust temperature \(T_d^{m,c} (\nu)\), Eq. (33), is superfluous.

The total optical depth \(\tau_{\rm total}^{m,c}(\nu)\) of each molecule \(m\) and component \(c\) is defined as the sum of the optical depths \(\tau_l^{m,c}(\nu)\) of all lines of each molecule \(m\) and component \(c\) plus the dust optical depth \(\tau_{\rm dust}^{m,c}(\nu)\) (see Sect. “Dust contribution”), the free-free optical depth \(\tau_{\rm free-free}^{m,c}(\nu)\) (see Sect. “Free-free contribution”), and the synchrotron optical depth \(\tau_{\rm sync}^{m,c}(\nu)\) (see Sect. “Synchrotron contribution”), i.e.

(6)\[\tau_{\rm total}^{m,c}(\nu) = \tau_l^{m,c}(\nu) + \tau_{\rm dust}^{m,c}(\nu) + \tau_{\rm free-free}^{m,c}(\nu) + \tau_{\rm sync}^{m,c}(\nu)\]

The optical depth \(\tau_t^{m,c}(\nu)\) for a transition \(t\) of molecule \(m\) and component \(c\) is described as [5]

(7)\[\tau_t^{m,c}(\nu) = \Bigg[\frac{c_{\rm light}^2}{8 \pi \nu^2} \, A_{ul}^t \, N_{\rm tot}^{m,c} \, \frac{g_u^t \, e^{-E_l^t/k_B \, T_{\rm ex}^{m,c}}}{Q \left(m, T_{\rm ex}^{m,c} \right)} \left(1 - e^{-h \, \nu^t /k_B \, T_{\rm ex}^{m,c}} \right) \cdot \phi^{m,c,t}(\nu)\Bigg],\]

where \(\phi^{m,c,t}(\nu)\) indicates a line profile function, described in Sect. “Line profile function”. The Einstein \(A_{ul}\) coefficient [6], the energy of the lower state \(E_l\), the upper state degeneracy \(g_u\), and the partition function [7] \(Q \left(m, T_{\rm ex}^{m,c} \right)\) of molecule \(m\) are taken from the embedded SQLite3 database. In addition, the values of the excitation temperatures \(T_{\rm ex}^{m,c}\) and the column densities \(N_{\rm tot}^{m,c}\) for the different components and molecules are taken from the user defined molfit file.

If local-overlap of lines is not taken into account (LocalOverlapFlag = False) the optical depth \(\tau_l^{m,c}(\nu)\) of all lines for each molecule \(m\) and component \(c\) is described as

\[\tau_l^{m,c}(\nu) = \sum_t \tau_t^{m,c}(\nu),\]

where the sum with index \(t\) runs over all spectral line transitions of molecule \(m\) within the given frequency range. The calculation procedure including local-overlap is described in Sect. “Local-overlap of neighboring lines”.

The beam-averaged continuum background temperature \(I_{\rm bg} (\nu)\) is parameterized as

(8)\[I_{\rm bg} (\nu) = T_{\rm bg} \cdot \left(\frac{\nu}{\nu_{\rm min}} \right)^{T_{\rm slope}} + J_\mathrm{CMB} + I_{\rm bg}^{\rm file} (\nu)\]

to allow the user to define the continuum contribution for each frequency range, individually. Here, \(\nu_{\rm min}\) indicates the lowest frequency of a given frequency range. \(T_{\rm bg}\) and \(T_{\rm slope}\), defined by the user, describe the background continuum temperature and the temperature slope, respectively. Additionally, XCLASS offers the possibility to describe the background continuum, i.e. the continuum seen by the layer close to the cosmic microwave background, by a so-called background file \(I_{\rm bg}^{\rm file} (\nu)\).

Finally, the last term \(J_\mathrm{CMB}\) in Eq. (1) describes the OFF position for single dish observations (defined by Inter_Flag = False) where we have an intensity caused by the cosmic background \(J_\mathrm{CMB}\). For interferometric observations, the contribution of the cosmic background is filtered out and has to be subtracted as well.

Hence, the layers, which do not belong to the largest distance parameter, have to be considered in an iterative manner where the solution of the radiative transfer equation for a certain distance \(i>1\) can be expressed as

(9)\[ T_{\rm mb}^i(\nu) = \sum_{m,c \in i} \left(S^{m,c}(\nu) - T_{\rm mb}^{i-1}(\nu)\right) \, \left(1 - e^{-\tau_{\rm total}^{m,c}(\nu)}\right) + T_{\rm mb}^{i-1}(\nu) e^{-\tau_{\rm total}^{m,c}(\nu)},\]

where \(m\) and \(c\) indicates the index of the current molecule and component, respectively. Here, the expression \(T_{\rm mb}^{\rm i=1}(\nu)\), described by Eq. (1), represents the spectrum caused by layers which are closest to the background, including the beam-averaged continuum background temperature \(I_{\rm bg}^{\rm core} (\nu)\). The contribution by other components (\(i > 1\)) is considered by first calculating \(T_{\rm mb}^{\rm i=1}(\nu)\) and then use this as new continuum for lines at distance \(i\).

By fitting all species and their components at once, line blending and optical depth effects are taken into account. The modeling can be done simultaneously with isotopologues (and higher vibrational states) of a molecule assuming an isotopic ratio stored in the so-called iso ratio file, see Sect. “The iso ratio file”. Here, all parameters are expected to be the same except the column density which is scaled by one over the isotopic ratio for each isotopologue. Additionally, it is assumed that radiation emitted by all isotopologues of a molecule in a component interact with all other isotopologues, but the radiation emitted in one component does not interact with other molecules or with the same molecule in different components, i.e. their intensities are added, except local-overlap is taken into account, see Sect. “Local-overlap of neighboring lines”.

In order to correctly take instrumental resolution effects into account in comparing the modeled spectrum with observations, myXCLASS integrates the calculated spectrum over each channel. Thereby, myXCLASS assumes that the given frequencies \(\nu\) describe the center of each channel, respectively. The resulting value is than given as

(10)\[T_{\rm mb}(\nu) = \frac{1}{\Delta_c} \int_{\nu-\frac{\Delta_c}{2}}^{\nu+\frac{\Delta_c}{2}} T_{\rm mb}(\tilde{\nu}) d\tilde{\nu},\]

where \(\Delta_c\) represents the width of a channel. Due to the complexity of Eqn. (1), (9) the integration in Eq. (10) can not be done analytically. Therefore, myXCLASS performs a piecewise integration of each component and channel using the trapezoidal rule and summaries the resulting values to get the final value used in Eq. (10). In order to reduce the computation effort myXCLASS determines the minimal line width of all components of a certain distance. If a channel contains no transition, myXCLASS determines the intensities at \(\nu \pm \frac{\Delta_c}{2}\) and evaluates Eq. (10) by applying the trapezoidal rule. But, if transitions are included, myXCLASS re-samples the corresponding channel, whereby the number \(n_\nu\) of additional frequency points is given by

\[n_\nu = (m \cdot 3) \cdot \min \left[1, {\rm int} \left[\frac{\Delta_c}{\sigma_{\rm min}} \right] \right],\]

where \(m\) represents the number of transitions contained in the current channel and \(\sigma_{\rm min} = \frac{\Delta v_{\rm min}}{2 \, \sqrt{2 \, \ln \, 2} \, c_{\rm light}} \, \nu\). If local-overlap is taken into account, see Sect. “Local-overlap of neighboring lines”, \(\Delta v_{\rm min}\) indicates the minimal line width of all components of the current layer, otherwise \(\Delta v_{\rm min}\) describes the line width of the current component. Here, the function int(\(x\)) gives the integer part of \(x\). Finally, myXCLASS computes the intensities at these frequencies and evaluates Eq. (10).

The molfit file

Within the molfit file the user defines both which molecules (or radio recombination lines (RRLs), continuum contributions) are taken into account and how many components are used. The definition of parameters for a molecule (or RRLs, continuum contributions) starts with a line describing the name of the molecule [8] (or RRLs, continuum contributions), followed by the number of components \(N\). The following \(N\) lines describe the parameters for each components, separately. The number of columns and their meaning depends on the modeled contribution (molecule, RRL, or continuum contribution). Generally, all parameters have to be separated by blanks, comments are marked with the % character.

Molecules

For molecules, the user has to define for each component the excitation temperature \(T_{\rm ex}\) in K (T_rot), the column density \(N_{\rm tot}\) in cm\(^{-2}\) (N_tot), and the velocity offset (relative to v\(_{\rm LSR}\)) in km s\(^{-1}\) (V_off). Depending of the used source description, see Sect. “Source description”, the size of the source \(\theta^{m,c}\) in arcsec (s_size) and their offset relative to the beam center is required as well. Furthermore, the selected line profile requires one or two velocity widths (FWHM) \(\Delta \nu\) in km s\(^{-1}\) (V_width). Additionally, the stacking of the components along the line-of-sight requires the cf-flag (CFFlag), see Sect. “Simple 1-d description”, or the distance parameter (distance), see Sect. “Stacking”. Finally, the column (keyword) to describe so-called keywords, indicating the line profile function, Non-LTE description etc. Note, different keywords have to be separated by “_” character. For example, the composed keyword Non-LTE_Voigt indicates a Non-LTE description and the usage of a Voigt line profile for the corresponding component.

Example of a molfit file describing two molecules CS\(_{v=0}\) and HCS\(^+_{v=0}\) with two components, respectively:

% Number of molecules = 2
% s_size:  T_rot:    N_tot:  V_width:   V_off:  CFFlag:   keyword:
CS;v=0;   3
   48.470  300.00  3.91E+17      2.86  -20.564        c
   21.804  320.00  6.96E+17      8.07   30.687        c
   81.700  208.00  1.46E+17      5.16  -10.124        c
HCS+;v=0;   2
% s_size:  T_rot:    N_tot:  V_width:   V_off:  CFFlag:   keyword:
           150.00  1.10E+18      5.00   -0.154        f
           200.00  2.20E+17      3.10   -2.154        f

Limiting the number of transitions

XCLASS offers the possibility to limit the number of transitions of molecules which enter Eq. (1). Therefore, the user has to add the following command words to the beginning of the molfit file:

  • "%%MinNumTransitionsSQL": defines the minimum number of transitions, i.e. XCLASS takes only those molecules into account which have at least this number of transition in the given frequency ranges.

  • "%%MaxNumTransitionsSQL" describes the max. number of transitions (in conjunction with "%%TransOrderSQL"). Here, "0" means all transitions are included.

  • "%%TransOrderSQL" (in conjunction with "%%MaxNumTransitionsSQL") defines the order of transition. (E.g. "%%MaxNumTransitionsSQL = 10" and "%%TransOrderSQL = 1" means, that only the first 10 transitions with the lowest lower energy are taken into account.)

  • "%%MaxElowSQL" defines the upper limit of the lower energy of a transition.

  • "%%MingASQL" describes the lower limit of gA (upper state degeneracy * Einstein A coefficient) of a transition.

Please note, for recombination lines only the command words "%%MinNumTransitionsSQL" and "%%MaxNumTransitionsSQL" are taken into account.
Example for a molfit file containing limiting transition parameters:
%==============================================================================
%
% define transition parameters:
%
%==============================================================================
%%MinNumTransitionsSQL = 1      % min. number of transitions
%%MaxNumTransitionsSQL = 0      % max. number of transitions,
                                % =0): all transitions are taken into account
%%TransOrderSQL = 1             % order of transitions:
                                %  (=1): by lower energy,
                                %  (=2): by gA,
                                %  (=3): gA/E_low^2, else by trans. freq.
%%MaxElowSQL = 1.0e+06          % max. lower energy, i.e. upper limit of
                                %  the lower energy of a transition
%%MingASQL = 0.0                % minimal intensity, i.e. lower limit of
                                %  gA of a transition


%==============================================================================
%
% define molecules and their components:
%
%==============================================================================
% s_size:   T_rot:    N_tot:  V_width:   V_off:  CFFlag:   keyword:
CS;v=0;   3
    48.470  300.00  3.91E+17      2.86  -20.564        c
    21.804  320.00  6.96E+17      8.07   30.687        c
    81.700  208.00  1.46E+17      5.16  -10.124        c
HCS+;v=0;   2
% s_size:   T_rot:    N_tot:  V_width:   V_off:  CFFlag:   keyword:
            150.00  1.10E+18      5.00   -0.154        f
            200.00  2.20E+17      3.10   -2.154        f

The iso ratio file

In order to reduce the number of model parameters, XCLASS offers the possibility to use a so-called iso ratio file. Here, XCLASS assumes, that both molecules (isotopologue and iso master molecule) are described by the same number of components, where the source size, the rotation temperature, the line width, and the velocity offset are identical. The corresponding column densities of the isotopologue \(\left({\rm N}_{\rm tot}^{\rm (isotopologue)}\right)\) are scaled by the so-called iso ratio, i.e.

\[{\rm N}_{\rm tot}^{\rm (isotopologue)} = {\rm N}_{\rm tot}^{\rm (iso \,\, master)} / {\rm (iso \, \, ratio)}.\]

The iso ratio file contains three columns separated by tabs or at least two blank characters, where the first two columns indicates the isotopologue and the corresponding iso master molecule, respectively. The third column defines the ratio for both molecules. Comments are marked with a % character, i.e. all characters on the right side of a % are ignored.

Example for an iso ratio file:

% isotopologue:         molecule:        ratio:
HC-33-S+;v=0;           HCS+;v=0;          75.0
HC-34-S+;v=0;           HCS+;v=0;          22.5
CS;v=4;                 CS;v=0;             2.3
CS;v=3;                 CS;v=0;             2.1
CS;v=2;                 CS;v=0;             2.1
CS;v=1;                 CS;v=0;             2.0
CS-34;v=0;              CS;v=0;            22.5
CS-34;v=1;              CS;v=0;            22.5
CS-33;v=1;              CS;v=0;            75.0
CS-33;v=0;              CS;v=0;            75.0

In the example described above we define in the first line an iso ratio of 75.0 between HC-33-S+;v=0; and HCS+;v=0;. Assuming that HCS+;v=0; is described with one component and with a column density of \(9 \cdot 10^{16}\) cm\(^{-2}\) the column density of HC-33-S+;v=0; is

\[{\rm N}_{\rm tot} = 9 \cdot 10^{16} \, {\rm cm}^{-2} / 75.0 = 1.2 \cdot 10^{15} \, {\rm cm}^{-2}.\]

Globally defined iso-ratios

In addition, the user can defined so-called globally defined iso ratios, e.g. (\(^{12}\)C / \(^{13}\)C). This defined ratio is multiplied to all ratios of isotopologues in the iso ratio file which contain e.g. \(^{13}\)C raised to the power of occurrences of the globally defined ion in each isotopologue. For example, the ratio (\(^{12}\)C / \(^{13}\)C) is set globally to 45. Additionally, we set the ratio of (CH\(_3\)CN,v=0 / \(^{13}\)CH\(_3 ^{13}\)CN,v=0) to 3.0. The final iso ratio used by XCLASS for (CH\(_3\)CN,v=0 / \(^{13}\)CH\(_3 ^{13}\)CN,v=0) is then given by \(3.0 \times 45^2 = 6075\). (Here, the exponent \(2\) is caused by the two \(^{13}\)C ions). Additionally, globally defined iso-ratios can be combined as well. In order to determine the used iso ratio for e.g. (\(^{12}\)C\(^{32}\)S)/(\(^{13}\)C \(^{33}\)S) we just have to multiply the globally defined is ratios for (\(^{12}\)C / \(^{13}\)C) and (\(^{32}\)S / \(^{33}\)S).

Globally defined ratios in the iso ratio file are indicated by the phrase “GLOBAL__” and can not be used alone without other iso ratios, e.g.

% isotopologue:         molecule:           ratio:
GLOBAL__C-13            C-12                45.0
HO-18;v=0;              OH;v=0;             500
HDO-18;v=0;             HDO;v=0;            500
C-13-H3C-13-N;v=0;      CH3CN;v=0;          1.0

Fitting iso ratios in the iso ratio file

The myXCLASSFit function offers the possibility to optimize the ratios between isotopologues as well. For that purpose, the user has to add two additional columns on the right indicating the lower and the upper limit of a certain ratio.

Example for an iso ratio file where the ratios are optimized by the myXCLASSFit function:

% isotopologue:         molecule:        ratio:     lower:     upper:
CS;v=4;                 CS;v=0;             2.3        0.1      500.0
CS;v=3;                 CS;v=0;             2.1        0.1      500.0
CS;v=2;                 CS;v=0;             2.1        0.1      500.0
CS;v=1;                 CS;v=0;             2.0        0.1      500.0
CS-33;v=1;              CS;v=0;            75.0        0.1      500.0
CS-33;v=0;              CS;v=0;            75.0        0.1      500.0
HC-33-S+;v=0;           HCS+;v=0;          75.0        0.1      500.0
HC-34-S+;v=0;           HCS+;v=0;          22.5        0.1      500.0

If the lower and upper limit are equal or if the lower limit is higher than the upper limit, the ratio is kept constant and is not optimized by the myXCLASSFit function. Note, if either the iso master molecule or a corresponding isotopologue has no transition within at least one given frequency range, the myXCLASSFit function does not optimize the corresponding iso-ratio. For example, if the isotopologue HNC-13;v=0; (used in the example described above) has no transition in at least one given frequency range, the given iso-ratio (here 60) is kept constant. Additionally, if a iso master molecule has no transition within at least one given frequency range, the iso-ratios to all of its isotopologues are kept constant.

Local-overlap of neighboring lines

XCLASS offers the possibility to take the local overlap of neighboring (molecules and recombination) lines into account. Here, we follow the derivation described by [Cesaroni_Walmsley_1991] and define an average source function \(S_l\) at frequency \(\nu\) and distance \(l\)

(11)\[S_l (\nu) = \frac{\varepsilon_l (\nu)}{\alpha_l (\nu)} = \frac{\sum_{c, t \in l} \tau_t^c (\nu) \, S_\nu (T_{{\rm rot}}^c)}{\sum_{c, t \in l} \tau_t^c (\nu)},\]

where \(\varepsilon_l\) represents the emission and \(\alpha_l\) the absorption function, respectively. Additionally, \(T_{\rm rot}^c\) indicates the excitation temperature and \(\tau_t^c\) the optical depth of transition \(t\) and component \(c\). For each frequency channel, we take all components \(c\) and transitions \(t\) into account, which belongs to the current distance and whose Doppler-shifted transitions frequencies are located within a range of 5\(\Delta v_{\rm max}\), where \(\Delta v_{\rm max}\) describes the largest line width of all components of the current layer. The iterative treatment of components at different distances, takes non-local effects into account as well. In the optically thin limit, Eq. (11) is equal to the traditional approach of convolving a line with several Gaussians. But it better describes situations in which both optically thin and optically thick lines are present: photons emitted from the optically thin transition are locally absorbed by the optically thick emission. The intensities of the lines do not simply add up like in the optically thin limit, but the intensity at the overlapping frequencies is mainly described by the optically thick emission.

Please note, that the myXCLASS function can not compute the intensity and corresponding optical depth for each component if the local overlap is taken into account. Only the intensities and optical depths at each distance can be determined.

Source description

The position and size of each component in the molfit file can be defined in two different ways: A simple 1-d description with a very simplistic geometrical structure offers a fast but not very accurate and flexible description with a minimum of parameters, wheres the advanced 3-d description is more accurate but requires more parameters and more computational effort. The different source descriptions can be applied to all kinds of components, i.e. to molecules, RRLs, and continuum contributions.

../_images/Sketch__Emission.png

Fig. 2 Sketch of a distribution of core layers within the Gaussian beam of the telescope (black ring). Here, we assume three different core components \(1\), \(2\), and \(3\), centered at the middle of the beam with different source sizes, excitation temperatures, velocity offsets (relative to v\(_{\rm LSR}\)) etc. indicated by different colors. Additionally, we assume that all core components have the same distance to the telescope, i.e. all core layers are located within a plane perpendicular to the line of sight. Furthermore, we assume that this plane is located in front of a background layer \(4\) with homogeneous intensity \(I_{\rm bg}^{\rm core}(\nu)\) over the whole beam. Core components do not interact with each other radiatively.

Simple 1-d description

For the 1-d description the molfit file has to contain column (CFFlag) indicating the core and foreground flag and must not contain column (distance) describing the stacking parameter, see Sect. “Stacking”. The 1-d assumption imposes a very simplistic geometrical structure where we recognize two classes of components:

One, the core objects (in earlier implementations called emission component), consists of an ensemble of objects centered at the middle of the beam. These could be identified with clumps, hot dense cores etc. which overlaps but do not interact either because they do not overlap in physical or in velocity space. For computational convenience, they are assumed to be centered in the beam, as shown in Fig. 2. It is also assumed that the dust emission emanates (partly) from these components. Their intensities are added, weighted with the beam filling factor, see Eq. (2). Note, all core components are contained in the layer which is closest to the background.

The second class, foreground objects (in earlier implementations called absorption components), are assumed to be in layers in front of the core components. In the current 1-d implementation, they would have a beam-averaged intensity of the core sources as background, and would fill the whole beam. Examples for such structures would be source envelopes in front of dense cores, or foreground components along the line-of-sight.

As shown in Fig. 2, we assume that core components do not interact with each other radiatively, i.e. one core layer is not influenced by the others, if local-overlap is not taken into account. But the core layers may overlap to offer the possibility to model sources consisting of several molecules and compounds. For core components, Eq. (1) has to be slightly modified [9], so that the solution of the radiative transfer equation for core layers is [10],

(12)\[\begin{split}T_{\rm mb}^{\rm core}(\nu) = \sum_m \sum_c &\Bigg[\eta \left(\theta^{m,c}\right) \left[S^{m,c}(\nu) \left(1 - e^{-\tau_{\rm total}^{m,c}(\nu)}\right) + I_{\rm bg}^{\rm core} (\nu) \left(e^{-\tau_{\rm total}^{m,c}(\nu)} - 1\right) \right] \Bigg] \\ &+ \left(I_{\rm bg}^{\rm core} (\nu) - J_\mathrm{CMB} \right),\end{split}\]

where the sums go over the indices \(m\) for molecule, and \(c\) for (core) component, respectively. The different terms contained in Eq. (12) are described in the previous sections.

../_images/Sketch__Absorption.png

Fig. 3 Sketch of a distribution of core and foreground layers within the Gaussian shaped beam of the telescope (black ring). Here, we assume three different core components \(2a\), \(2b\), and \(2c\) located in a plane perpendicular to the line of sight which lies in front of the background layer \(1\) with intensity \(I_{\rm bg}^{\rm core}(\nu)\), see Eq. (8). The foreground layers \(3\) and \(4\) are located between the core layers and the telescope along the line of sight (black dashed line). Here, each component is described by different excitation temperatures, velocity offsets etc. indicated by different colors. The thickness of each layer is described indirectly by the total column density \(N_{\rm tot}^{m,c}\), see Sect. “Optical Depth”. For each foreground layer we assume a beam filling factor of 1, i.e. each foreground layer covers the whole beam.|

In contrast to core layers, foreground components may interact with each other, as shown in Fig. 3, where absorption takes places only, if the excitation temperature for the absorbing layer is lower than the temperature of the background.

Hence, the solution of the radiative transfer equation for foreground layers can not be given in a form similar to Eq. (12). Foreground components have to be considered in an iterative manner similar to Eq. (9). The solution of the radiative transfer equation for foreground layers can be expressed as

(13)\[\begin{split}T_{\rm mb}^{\rm fore}(\nu)_{m,c=1} &= \eta \left(\theta^{m,c=1} \right) \left(S^{m,c=1}(\nu) - T_{\rm mb}^{\rm core}(\nu)\right) \, \left(1 - e^{-\tau_{\rm total}^{m,c=1}(\nu)}\right) + T_{\rm mb}^{\rm core}(\nu) \\ T_{\rm mb}^{\rm fore}(\nu)_{m,c=i} &= \eta \left(\theta^{m,c=i} \right) \left(S^{m,c=i}(\nu) - T_{\rm mb}^{\rm fore}(\nu)_{m,c=(i-1)}\right) \, \left(1 - e^{-\tau_{\rm total}^{m,c=i}(\nu)}\right) + T_{\rm mb}^{\rm fore}(\nu)_{m,c=(i-1)},\end{split}\]

where \(m\) indicates the index of the current molecule and \(i\) represents an index running over all foreground components \(c\) of all molecules. Additionally, we assume that each foreground component covers the whole beam, i.e. \(\eta \left(\theta^{m,c=i} \right) \equiv 1\) for all foreground layer [11]. Thus, Eq. (13) simplifies to

(14)\[\begin{split}T_{\rm mb}^{\rm fore}(\nu)_{m,c=1} &= \Big[ S^{m,c=1}(\nu) \left(1 - e^{-\tau_{\rm total}^{m,c=1}(\nu)}\right) + T_{\rm mb}^{\rm core}(\nu) e^{-\tau_{\rm total}^{m,c=1}(\nu)}\Big] \\ T_{\rm mb}^{\rm fore}(\nu)_{m,c=i} &= \Big[S^{m,c=i}(\nu) \left(1 - e^{-\tau_{\rm total}^{m,c=i}(\nu)}\right) + T_{\rm mb}^{\rm fore}(\nu)_{m,c=(i-1)} e^{-\tau_{\rm total}^{m,c=i}(\nu)}\Big],\end{split}\]

where \(T_{\rm mb}^{\rm core}(\nu)\) describes the core spectrum, see Eq. (12), including the beam-averaged continuum background temperature \(I_{\rm bg}^{\rm core} (\nu)\). For foreground lines the contribution by other components is considered by first calculating the contribution of core objects and then use this as new continuum for foreground lines reflecting the fact that cold foreground layers are often found in front of hotter emission sources. The myXCLASS function assumes, that the cosmic background describes together with the core components one end of a stack of layers. Additionally, the foreground components are located between this plane and the telescope, see Fig. 3. The total column density \(N_{\rm tot}^{m,c}\) depends on the abundance of a certain molecule and on the thickness of a layer containing the molecule. The order of components along the line of sight is defined by the occurrence of a certain foreground component in the molfit file.

Stacking

../_images/Stacking.png

Fig. 4 Sketch of components located at different distances along the line of sight.

In addition to the simple 1-d description described in the previous section, XCLASS offers the possibility to define the order of each component along the line-of-sight by a so-called stacking or distance parameter which replaces column CFFlag in the molfit file, i.e. the column CFFlag is replaced by column distance right after column V_off. The stacking parameter can be interpreted as distance of a component to the observer where the component(s) with the largest distance parameter is (are) closest to the background while those components with the smallest distance parameter are closest to the observer. Additionally, the components which have the same distance parameter (i.e. within a difference of \(< 10^{-9}\)) are located in the same plane perpendicular to line-of-sight. In contrast to foreground components, described in the previous section, which can describe only one molecule at a certain distance, the stacking formalism offers the possibility to describe an arbitrary number of clouds consisting of an arbitrary number of molecules, recombination lines, or continuum contributions (like dust or synchrotron) along the line-of-sight.

Note, the distance parameter has no unit, can not be optimized by MAGIX, and must not be used in conjunction with the old cf-flag formalism, described in the previous section.
In the following example SO\(_2\) is described by four and CH\(_3\)CN with two components where the first components of SO\(_2\) and CH\(_3\)CN are located close to the background, i.e. both components are located in the same plane perpendicular to the line of sight and are illuminated by the background continuum Eq. (8). The second and third component of SO\(_2\) are located in another plane perpendicular to the line of sight, which is closer to the observer. Both components “see” a background which is caused by the background plus the first components of SO\(_2\) and CH\(_3\)CN. The fourth component of SO\(_2\) and the second component of CH\(_3\)CN are located in front of the plane at 500, where the second component of SO\(_2\) is closer to the observer.
SO2;v=0;   4

% s_size:  T_rot:   N_tot:   V_width:   V_off: distance:  keyword:
%[arcsec]     [K]   [cm-2]     [km/s]   [km/s]        []

   500.00   50.00  1.9E+14       2.86     1.22     600.0
   500.00  190.00  3.9E+15       4.16    -1.61     500.0
   500.00  250.00  3.9E+16       8.55    -5.75     500.0
   500.00   80.00  3.9E+14       1.23    -1.82     300.0

CH3CN;v=0;   2

% s_size:  T_rot:   N_tot:   V_width:   V_off: distance:  keyword:
%[arcsec]     [K]   [cm-2]     [km/s]   [km/s]        []

   500.00   50.00  2.2E+15       2.86     6.01     600.0
   500.00   80.00  8.5E+14       1.23    -2.82     200.0

Sub-beam description

../_images/sub-beam_description.png

Fig. 5 Sketch of a sub-beam description for a single frequency channel. Here, the gray grid indicates the model grid, the black ellipse the beam, and the red dot the center of the beam, respectively. Additionally, the filled circles indicate different components, where the gray scale represent the location of a component along the line-of-sight, i.e. light gray describe components close the cosmic microwave background, while dark gray represents components close the observer. The projected width \(w\) and height \(h\) of the beam are described by Eq. (15).

Besides to the aforementioned component stacking XCLASS offers the possibility to take the shape and the position of each component perpendicular to the line-of-sight into account as well, see Fig. 4. The so-called sub-beam description, see Fig. 5, is used whenever an offset position of a component is defined, or if a component at a smaller distance is smaller than a component at larger distance, or if local-overlap is taken into account. Each component can have a different size and center position. In the current version only circle and square-shaped components can be described. Additionally, elliptical rotated beam sizes, see Sect. “Elliptical rotated Gaussians”, can be defined as well, but the frequency dependence of single dish beam sizes is ignored, i.e. the beam size is assumed to be constant for the whole frequency range [12]. For the sub-beam description, XCLASS samples the elliptical rotated beam by an rectangular linear grid, called model grid, where the number of pixels along the x- and y-direction is defined by the user. The size of each model pixel is defined by the beam, where we assume that the grid along the x- and y-axis is three times the width \(w\) and height \(h\) of the beam, respectively, described by

(15)\[\begin{split}w &= \sqrt{\left({\tt BMIN} \cdot \cos {\tt BPA} \right)^2 + \left({\tt BMAJ} \cdot \sin {\tt BPA} \right)^2},\\ h &= \sqrt{\left({\tt BMIN} \cdot \sin {\tt BPA} \right)^2 + \left({\tt BMAJ} \cdot \cos {\tt BPA} \right)^2}.\end{split}\]

Here, \({\tt BMIN}\) and \({\tt BMAJ}\) represents the minor and major axis and \({\tt BPA}\) the position angle of the beam. Furthermore, XCLASS identifies for each component the model pixels, which are covered by the corresponding component. While the identification of all pixels covered by a component is trivial for square components, the Bresenham’s circle algorithm [Bresenham_1977] is used for circular components. After this, XCLASS computes the spectra for each model pixel based on the corresponding component configuration. Afterwards, XCLASS convolves the calculated intensity map for each frequency channel with the elliptical rotated Gaussian beam using the FFTW library [13] [Padua_2011]. Finally, XCLASS uses the spectrum at the beam center for the final output. The intensities and optical depths for each component / distance are computed in the same way.

For beam-centered components only the diameter of the component (in arcsec) is required which has to be defined in the first column s_size of the molfit file, e.g.

SO2;v=0;   4

% s_size:  T_rot:   N_tot:   V_width:   V_off: distance:  keyword:
%[arcsec]     [K]   [cm-2]     [km/s]   [km/s]        []

    50.00   50.00  1.9E+14       2.86     1.22     600.0
   110.00  190.00  3.9E+15       4.16    -1.61     500.0
Components whose center are not identical to the center of the beam require three additional parameters: The column s_off_x and s_off_y describe the x- and y-coordination of the component center position relative to the beam center (in arcsec), respectively. Additionally, column (keyword) has to contain the keyword circleoff.
In the following example, we describe SO\(_2\) with two components, where the first component has a size (diameter) of 50 arcsec, and a relative offset of 1.0 and 1.45 arcsec, respectively. The second component has a larger diameter of 200 arcsec and is centered at the middle of the beam.
SO2;v=0;   2

% s_size: s_off_x: s_off_y: T_rot: N_tot: V_width:   V_off: distance: keyword:
%[arcsec] [arcsec] [arcsec]    [K] [cm-2]   [km/s]   [km/s]        []

    50.00      1.0    1.45   50.00  1.e14     2.86     1.22     600.0 circleoff
   200.00                    90.00  3.e15     4.16    -1.61     500.0

Additionally, the user can use keywords box and boxoff to define quadratic components. For keyword box, XCLASS assumes that the corresponding component is located at the center of the beam, where column s_size indicates the length (in arcsec) of the square. Similar to keyword circleoff the keyword boxoff defines a square which center (in arcsec) is defined by columns s_off_x and s_off_y, respectively.

Non-LTE description of molecules

In addition to the local thermodynamic equilibrium (LTE) description, XCLASS offers the possibility to treat molecules in Non-LTE as well using the RADEX routines [van_der_Tak_et_al_2007]. In the following we briefly describe the underlying formalism, using [van_der_Tak_et_al_2007] and the RADEX manual [14].

In general, collisions between atoms or molecules can cause transitions from any state \(i\) to any other state \(j\). In dense environments these collisions take place so often that the occupation numbers \(n_i\) and \(n_j\) are thermally distributed, i.e. they are described by the Boltzmann distribution where the Boltzmann temperature is identical to the kinetic temperature. If the density is very low, the radiative transitions can become more frequent than collisional transitions which requires a Non-LTE description.

In order to calculate the level populations we assume that for every level \(i\) the rate at which atoms/molecules are being (de-)excited out of level \(i\) is equal to the rate at which level \(i\) is being re-populated by (de-)excitation from other levels:

\[\frac{d \, n_i}{dt} = \sum_{j \neq i}^N n_j P_{ji} - n_i \sum_{j \neq i}^N P_{ij} = 0,\]

where \(P_{ji}\) describes the formation rate coefficient of level \(i\), while \(P_{ij}\) is the corresponding destruction rate coefficient which is given by

\[\begin{split}P_{ij} = \begin{cases} A_{ij} + B_{ij} \, u_{ij} + C_{ij}, & (i > j)\\ B_{ij} \, u_{ij} + C_{ij}, & (i < j).\\ \end{cases}\end{split}\]

Thus we get

(16)\[\begin{split}\frac{d \, n_i}{dt} &= \sum_{j > i} \left[ n_j \, A_{ji} + \left(n_j \, B_{ji} - n_i \, B_{ij}\right) u_{ji} \right] \\ & \quad - \sum_{j<i} \left[ n_i \, A_{ij} + \left(n_i \, B_{ij} - n_j \, B_{ji} \right) u_{ij} \right] \\ & \quad + \sum_{j \neq i} \left[n_j \, C_{ji} - n_i \, C_{ij} \right] = 0,\end{split}\]

where \(A_{ij}\) indicates the probability for spontaneous emission, \(B_{ji} \, u_{ij}\) the probability for absorption, and \(B_{ij} \, u_{ij}\) the probability for induced emission. Additionally, the local radiative energy density \(u_{ij}\) is given by

\[u_{ij} = \int_0^\infty \, \phi_{ij} \left(\nu\right) \, I_\nu \, d \nu,\]

where \(I_\nu\) represents the radiation field and \(\phi_{ij} \left(\nu\right)\) the line profile function for transition \(i \rightarrow j\). Furthermore, the \(C_{ij}\) describe the collision rates per second per molecule of the species of interest. They have to depend on the density of the collision partner, so that they can be expressed as \(C_{ij} = K_{ij} \, n_{\rm col}\), where \(n_{\rm col}\) indicates the density of the collision partner, which is in most cases H\(_2\). (XCLASS offers the possibility to define up to seven different collision partners for each distance, see Sect. “The collision partner file”.) The collisional rate coefficients \(K_{ij}\) (described in cm\(^3\) s\(^{-1}\)) are the velocity-integrated collision cross sections,

\[K_{ij} = \left(\frac{8 \, k_B \, T_{\rm kin}}{\pi \mu}\right)^{-1/2} \left(\frac{1}{k_B \, T_{\rm kin}}\right)^2 \, \int \sigma \, E \, e^{-E / k_B T_{\rm kin}} dE,\]

where \(E\) represents the collision energy, \(T_{\rm kin}\) the kinetic temperature, and \(\mu\) the reduced mass of the system. If collisions dominate \((n_j \, C_{ij} \gg A_{ij} + B_{ij} u_{ij})\), the upward rates are obtained through detailed balance

\[K_{ji} = K_{ij} \, \frac{g_i}{g_j} \, e^{-h \nu / k_B \, T_{\rm kin}},\]

The \(K_{ij}\) depend on the temperature through the relative velocity of the colliding molecules and possibly also through the collision cross sections directly. The collision rates can be interpreted as the “collisional analogs” to the Einstein relations. [15]

For some cases, Eq. (16) can be expressed by the Boltzmann distribution of a specific temperature: If collisions dominate \((n_j \, C_{ij} \gg A_{ij} + B_{ij} u_{ij})\) the level populations are described by

\[\frac{n_j}{n_i} = \frac{C_{ij}}{C_{ji}} = \frac{g_j}{g_i} e^{-\Delta E_{ji} / k_B T_{\rm kin}},\]

where \(\Delta E_{ji} = E_j - E_i\) indicates the energy difference between both levels and \(T_{\rm kin}\) the kinetic temperature.

If radiation dominates \((n_j \, C_{ij} \ll A_{ij} + B_{ij} u_{ij})\) the level populations can be expressed as

\[\frac{n_j}{n_i} = \frac{A_{ji} + B_{ji} \, u_{ij} }{B_{ij} \, u_{ij}} = \frac{g_j}{g_i} e^{-\Delta E_{ji} / k_B T_{\rm rad}},\]

where \(u_{ij} = B_\nu (T_{\rm rad})\) was assumed in the last step. For a given temperature, one can define the critical number density

\[N_{\rm crit} = \frac{A_{ij}}{C_{ij}}\]

as the density above which the collisions are so frequent, that LTE can be used, while below which there are substantial deviations from LTE.

But in many cases \((n_j \, C_{ij} \approx A_{ij} + B_{ij} u_{ij})\) and one defines an excitation temperature through

\[\frac{n_j}{n_i} = \frac{g_j}{g_i} e^{-\Delta E_{ji} / k_B T_{\rm ex}},\]

Assuming that \(u_{ij} = B_\nu (T_{\rm bg})\) , with \(T_{\rm bg}\) = 2.73 K representing the temperature of the cosmic microwave background, we can describe the excitation temperature \(T_{\rm ex}\) as

(17)\[T_{\rm ex} = -\frac{\Delta E_{ji}}{k_B \, \ln\left[\frac{A_{ij} + C_{ij} \, \exp \left(-\Delta E_{ji} / k_B \, T_{\rm kin}\right) \, \left(\exp\left(\Delta E_{ji} / k_B \, T_{\rm bg}\right) - 1\right)}{A_{ij} \, \exp \left(\Delta E_{ji} / k_B \, T_{\rm bg}\right) + C_{ij} \, \left(\exp\left(\Delta E_{ji} / k_B \, T_{\rm bg}\right) - 1\right)} \right]}\]

Note, that different lines will have different excitation temperatures.

But, in general, the radiation field is not described by the Planck function. Additionally, Eq. (16) is a local equation (to be solved at each location separately), but it has a global character due to the dependency on \(u_{ij}\) and \(u_{ji}\) which can be solved, usually with some simplifying assumptions.

The problem is how to “decouple” the radiative transfer calculations from the calculations of the level populations. A popular approach for this is the so-called escape probability method. The basic idea is to invent a factor that determines the chance that a photon at some position in the cloud can escape the system. This probability \(\beta\) depends only on the optical depth \(\tau\) and is related to the intensity within the medium, ignoring background radiation and any local continuum through

\[u_{ij} = S_{ij} \, \left(1 - \beta \right),\]

where \(S_{ij}\) describes the source function

\[S_{ij} = \frac{n_i \, A_{ij}}{n_j \, B_{ji} - n_i \, B_{ij}}.\]

The source function \(S_{ij}\) is independent of \(\nu\), at least over the small frequency range across the transition \(i \rightarrow j\) and has a single value for each line. (\(S_{ij} = B_{\nu_{ij}}(T)\) for LTE).

Note, that if all the photons escape, then \(\beta = 1\) and \(u_{ij}\) is the blackbody radiation field intensity. If no photons escape from the cloud, then \(u_{ij}\) is the source function.

../_images/EscapeProbability__RADEX.png

Fig. 6 Escape probability \(\beta\) as a function of optical depth \(\tau\) for three different geometries: uniform sphere (solid line), expanding sphere (dotted line) and plane-parallel slab (dashed line). (Taken from [van_der_Tak_et_al_2007]).

There are detailed relations between \(\beta\) and \(\tau\) for specific geometrical assumptions, see Fig. 6. The expression derived for an expanding spherical shell, the so-called Sobolev or large velocity gradient (LVG) approximation [Sobolev_1960] , [Castor_1970] , [Elitzur_1992]. This method is also widely applied for moderate velocity gradients, to mimic turbulent motions. RADEX uses the formula by [Mihalas_1978] and [Boland_De-Jong_1982] for this geometry:

(18)\[\beta = \langle e^{- \tau} \rangle = \frac{1}{\tau} \int_0^\tau e^{-\tau'} d \tau' = \frac{1 - e^{-\tau}}{\tau}.\]

For a homogeneous slab is found:

(19)\[\beta = \frac{1 - e^{-3 \tau}}{3 \tau}.\]

Also for a turbulent medium an escape probability has been estimated:

\[\beta = \frac{1}{\tau \, \sqrt{\pi \ln \left(\tau / 2 \right)}}.\]

Finally for a uniform sphere

(20)\[\beta = \frac{1.5}{\tau} \left[1 - \frac{2}{\tau^2} + \left( \frac{2}{\tau} + \frac{2}{\tau^2}\right) \, e^{-\tau} \right].\]

RADEX uses this last formula to estimate the excitation and radiation field in the following way. (For high optical depth, only the first term of the formula is retained; at low optical depth, a power series approximation is used.) As a first guess the level populations in the optically thin case (or for LTE) are calculated; this then gives the optical depth and hence the escape probability, from which the new level populations can be directly calculated. The program iterates this procedure to find a consistent level population and optical depth, and computes all line strengths for that solution.

The optical depth \(\tau_0\) in the line at line center is related to the populations by

\[\tau_0 = \frac{c^3}{8 \, \pi \, \nu_{ij}^3} \frac{A_{ij} \, N_{\rm mol}}{(\sqrt{\pi} / 2 \, \sqrt{ \ln 2}) \Delta V} \, \left[x_i \frac{g_j}{g_i} - x_j \right],\]

where \(N_{\rm mol}\) is the total column density, \(\Delta V\) the full width at half-maximum of the line profile in velocity units, and \(x_l\) the fractional population of level \(l\). The formalism is analogous to the LVG method, with the global \(n/(dV/dR)\) replaced by the local \(N/\Delta V\).

The calculations proceed as follows. A first guess of the populations of the molecular energy levels is produced by solving statistical equilibrium in the optically thin case. The only radiation taken into account is the unshielded background radiation field; internally produced radiation is not yet available. The solution for the level populations allows calculation of the optical depths of all the lines, which are then used to re-calculate the molecular excitation. The new calculation treats the background radiation in the same manner as the internally produced radiation. The program iteratively finds a consistent solution for the level populations and the radiation field. When the optical depths of the lines with \(\tau > 10^{-2}\) are stable from one iteration to the next to a given tolerance (default 10\(^{-6}\)), the program writes output and stops.

RADEX includes some limitations. First of all only one molecule is treated at a time, so that the effects of line overlap are not taken into account. In special cases, overlap between lines of the same molecule may influence their excitation, for example the hyperfine components of HCN. By taking local-overlap into account, see Sect. “Local-overlap of neighboring lines”, XCLASS is able to overcome this limitation.

For certain molecules under certain physical conditions (especially low density and/or strong radiation field), population inversions occur, which cause negative optical depth and hence nonlinear amplification of the incoming radiation [Elitzur_1992]. This phenomenon, known as “maser” action, requires non-local treatment of the radiative transfer, in particular a fine sampling of directions. Generally, the escape probability approximation is justified until the masers saturate, which occurs at \(\tau \approx -1\). In practice, the computed intensities of lines with \(\tau \leq -0.1\) are not as accurate as those of other lines, and the intensities of lines with \(\tau \leq -1\) should be disregarded altogether. If many lines have negative optical depths, the intensities of non-maser lines may also be affected. RADEX may well be used to predict which lines of a molecule may display maser action under certain physical conditions. Note however that lines may be masers even if \(\tau > 0\) according to RADEX, for example “Class II” CH\(_3\)OH masers which are pumped by infrared radiation.

XCLASS uses the RADEX routines to compute the optical depths (multiplied with the user-defined line profile function) and the corresponding excitation temperatures for all transitions of a certain molecule and component, where the kinetic temperature, column density, line width and velocity offset of the corresponding component are used as input for RADEX. For components which are closest to the background, XCLASS uses the beam-averaged continuum background temperature \(I_{\rm bg} (\nu)\), Eq. (8), in conjunction with dust, free-free, and synchrotron continuum contributions as radiation field for RADEX. For components with smaller distance parameters, a radiation field caused by components at larger distances is used as input for the RADEX routines. So non-local effects are taken into account as well which partly removes the limitation of RADEX.

In order to describe a component of a certain molecule in Non-LTE the keyword Non-LTE has to be added to column keyword for the corresponding component. In the following example, we describe the first component in Non-LTE:

CH3OH;v=0;   1
% s_size:  T_rot:   N_tot:  V_width:   V_off:  CFFlag:  keyword:
%[arcsec]     [K]   [cm-2]    [km/s]   [km/s]    [c/f]
    48.47   50.00  3.9E+16      2.86    -2.53        c   Non-LTE

But, the Non-LTE description requires the collision rates for the corresponding molecules and the density of the collision partners as well.

The collision partner file

These parameters are described by the so-called collision partner file, where the first line indicates the used approximation for the escape probability \(\beta\) (Geometry = 1 for an uniform sphere described by Eq. (20), Geometry = 2 for an expanding spherical shell, Eq. (18), and Geometry = 3 for a homogeneous slab, Eq. (19)). The second line describes the number of molecules, which have at least one component described in Non-LTE. The following lines describe for each molecule the name of the molecule [16], and the path and name of the so-called molecular data file, see below. Thereafter, two lines are required for each collision partner: One line describes the name of the collision partner (H2, p-H2, o-H2, e, H, He, and H+), see below. The second line indicates the initial density (in cm\(^{-3}\)), the lower and upper limit of the density (used for fitting), and the corresponding distance in the molfit file, respectively. If no distance is specified, the corresponding density is used for all distances where the corresponding molecule is described in Non-LTE. Additionally, two collision partners with a fixed ratio between both densities can be defined as well, see below.

In the following example, we use the escape probability approximation for an expanding spherical shell (Geometry = 2) and describe two molecules CH\(_3\)OH-A and CO, with collision rates stored in files "a-ch3oh.dat" and "co.dat", respectively. For CH\(_3\)OH-A, we use two collision partners: H\(_2\) (H2) and electrons (e). As mentioned above, XCLASS offers the possibility to define the density of collision partners for different distances and to fit these densities (here, between 1 cm\(^{-3}\) and 10\(^5\) cm\(^{-3}\)) as well using the myXCLASSFit function. Here, H\(_2\) has a initial density of 10\(^2\) cm\(^{-3}\) at distance 10\(^{10}\) and a density of 10\(^4\) cm\(^{-3}\) at distance 10\(^{9}\). Additionally, we define a global density, i.e. a density for all distances, of 5 cm\(^{-3}\) for electrons. For CO, H\(^+\) is used as collision partner with an initial global density of 10 cm\(^{-3}\) for all distances.

2                    % Geometry (=1: sphere, =2: LVG, =3: slab)
2                    % number of molecules in molfit file described in Non-LTE
CH3OH;v=0;A          % 1st molecule (with at least one component in Non-LTE)
a-ch3oh.dat          % path and name of file containing molecular data
3                    % number of collision partners for CH3OH;v=0;A
H2                   % first collision partner for CH3OH;v=0;A
1.e2 1.0 1.e5 1.e10  % density of 1st coll. partner (cm-3) at distance 1.e10
H2                   % second collision partner for CH3OH;v=0;A
1.e4 1.0 1.e5 1.e09  % density of 1st coll. partner (cm-3) at distance 1.e09
e                    % third collision partner for CH3OH;v=0;A
5.0  1.0 1.e5        % density of 2nd coll. partner (cm-3) (for all distances)
CO;v=0;              % 2nd molecule (with at least one component in Non-LTE)
co.dat               % path and name of file containing molecular data
1                    % number of collision partners for CO
H+                   % first collision partner (for CO)
1.e1 1.0 1.e5        % density of 1st coll. partner (cm-3) (for all distances)
In general, the RADEX routines can deal with up to seven different collision partners which can be defined for each distance used in the corresponding molfit file: H\(_2\) (H2), p-H\(_2\) (p-H2), o-H\(_2\) (o-H2), electrons (e), H-atoms (H), He (He), and H\(^+\) (H+). The user has to make sure that the collisional data file contains collision rate coefficients for all partners for which densities are given here. In most cases, using H\(_2\) as only partner is a good default. When H\(_2\) is given as collision partner for CO, a thermal average of ortho and para H\(_2\) is taken.
In order to define two collision partners with a fixed ratio between the respective densities, the user has to extend the name of the first collision partner by two additional columns separated by blanks. Here, the first two columns specify the names of the two collision partners respectively, while the third column describes the ratio \(r\) between the two densities. The density defined in the following line is interpreted as the summed density of both partners, i.e.
\[n_{\rm cp1} + n_{\rm cp2} = n_{\rm total}.\]

The density \(n_{\rm cp1}\) of collision partner 1 is defined as

(21)\[n_{\rm cp1} = \frac{1}{1 + r} \cdot n_{\rm total},\]

while the density \(n_{\rm cp2}\) of collision partner 2 is given as

(22)\[n_{\rm cp2} = \frac{r}{1 + r} \cdot n_{\rm total}.\]

As in the non-fixed case, the density can be adjusted within the given range and/or also defined for a given distance.

2                    % Geometry (=1: sphere, =2: LVG, =3: slab)
1                    % number of molecules in molfit file described in Non-LTE
CH3OH;v=0;A          % 1st molecule (with at least one component in Non-LTE)
a-ch3oh.dat          % path and name of file containing molecular data
1                    % number of collision partners for CH3OH;v=0;A
p-H2 o-H2 3.0        % collision partners for CH3OH;v=0;A
1.e2 1.0 1.e5 1.e10  % density of 1st coll. partner (cm-3) at distance 1.e10
In the example described above, we use para- and ortho-H\(_2\) with a fixed ratio of 3.0 and a total density of 100 cm\(^{-3}\). Using Eqn. (21), (22) the densities of para- and ortho-H\(_2\) are 25 and 75 cm\(^{-3}\), respectively. Please note that here the line with a fixed density ratio counts only as one, although two collision partners are used.
: Files with molecular data (term energies, statistical weights, Einstein coefficients, and rate coefficients for collisional de-excitation). For many molecules of astrophysical interest, the LAMDA database [17] provides data files in the required format. For molecules for which no collisional data exist, users need to make their own files. The databases at NASA-JPL and in Cologne provide spectroscopic data for many molecules. For collisional rate coefficients, consult NASA-GISS for the calculations of Sheldon Green, or the literature in the BASECOL [18] and CASSIS databases for other work. Unfortunately, collisional data do not exist for all molecules of (potential) astrophysical interest.

If you need to make your own molecular data file, here is a detailed description of the required file format. The format can be used for all molecules, whether linear (HCO\(^+\)) or not (H\(_2\)O). The lines that start with an exclamation mark (!) are not read by the program.

% Lines 1 - 2: molecule name
% Lines 3 - 4: molecular weight (a.m.u.)
% Lines 5 - 6: number of energy levels (NLEV)
% Lines 7 - (7 + NLEV): level number, level energy (cm-1), statistical
               weight. These numbers may be followed by additional
               info such as the quantum numbers, which are however
               not used by the program. The levels must be listed
               in order of increasing energy.
% Lines (8+NLEV) - (9+NLEV): number of radiative transitions (NLIN)
% Lines (10+NLEV - (10+NLEV+NLIN): transition number, upper level,
               lower level, spontaneous decay rate (s-1). These numbers may
               be followed by additional info such as the line frequency and
               upper state energy, which is however not used by the program.
% Lines (11+NLEV+NLIN) - (12+NLEV+NLIN): number of collision partners
% Lines (13+NLEV+NLIN) - (14+NLEV+NLIN): collision partner ID and reference.
               Valid identifications are: 1=H2, 2=para-H2, 3=ortho-H2,
               4=electrons, 5=H, 6=He, 7=H+.
% Lines (15+NLEV+NLIN) - (16+NLEV+NLIN): number of transitions for which
                                         collisional data exist (NCOL)
% Lines (17+NLEV+NLIN) - (18+NLEV+NLIN): number of temperatures for which
                                         collisional data exist
% Lines (19+NLEV+NLIN) - (20+NLEV+NLIN): values of temperatures for which
                                         collisional data exist
% Lines (21+NLEV+NLIN) - (21+NLEV+NLIN+NCOL): transition number, upper level,
               lower level; rate coefficients (cm3 s-1) at each temperature.
               RADEX interpolates between rate coefficients in the specified
               temperature range.

Outside this range, it assumes the collisional de-excitation rate coefficients are constant with T, i.e.  it uses rate coefficients specified at the highest T (400 K in this case) also for higher temperatures, and similarly at temperatures below the lowest value (10 K in this case) for which rate coefficients were specified.

Radio recombination lines (RRLs)

In addition to molecules myXCLASS can deal with radio recombination lines (RRLs) as well. The description can be done in LTE as well as in Non-LTE.

RRLs in LTE

The optical depths of RRLs (in LTE) is given by [19]

(23)\[\begin{split}\tau_{{\rm RRL}, \nu} &= \int \kappa_{n_1, n_2, \nu}^{\rm ext} \, ds \\ &= \frac{\pi \, h^3 \, e^2}{(2 \pi \, m_e \, k_B)^{3/2} \, m_e \, c} \cdot {\rm EM} \cdot \frac{n_1^2 \, f_{n_1, n_2}}{T_e^{3/2}} \, \exp \left[\frac{Z^2 \, E_{n_1}}{k_B T_e} \right] \, \left(1 - e^{-h \nu_{n_1, n_2} / k_B T_e} \right) \, \phi_\nu,\end{split}\]

where the LTE source function \(S_{{\rm RRL}, \nu}^{\rm LTE}\) can be written in terms of the brightness temperature Eq. (4) as

(24)\[S_{{\rm RRL}, \nu}^{\rm LTE} = \frac{\tau_{{\rm RRL}, \nu} \cdot J(T_e)}{\tau_{{\rm RRL}, \nu}}.\]

The terms \(n_1\), \(f_{n_1, n_2}\), \(E_{n_1}\), and \(\nu_{n_1, n_2}\) are taken from the embedded database whereas the emission measure EM (in cm\(^{-6}\) pc), electronic temperature \(T_e\) (in K), source size \(\theta^{m,c}\) (in arcsec) (or other parameters related to source description), line width \(\Delta v\) (in km/s), and velocity offset \(v\) (in km/s) used for the line profile function \(\phi_\nu\) are taken from the user-defined molfit file.

The parameters which are used for the description of radio recombination lines (RRLs) are quite similar to the parameters used for molecules. Here, the kinetic temperature is interpreted as electronic temperature (in K) and the column density as emission measure EM (in pc cm\(^{-6}\)). RRLs of hydrogen are marked as RRL-H, of Helium as RRL-He, of Carbon as RRL-C, of Nitrogen as RRL-N, and of Oxygen as RRL-O. In the following example we describe RRLs of hydrogen at two distances with a Gaussian line shape:

RRL-H    2
% s_size:     T_e:   EM_RRL:  V_width:   V_off:  distance:
500.00000    6.9e4   5.5E+16       1.3      2.0      2.e15
500.00000    2.3e4   2.3E+12       2.5      2.0      1.e15

RRLs in Non-LTE

../_images/RRL__Non-LTE__SH-coefficients.png

Fig. 7 Departure coefficients \(b_n\) derived by [Storey_Hummer_1995] for Hn\(\alpha\) lines for an electronic temperature of \(T_e\) = 10 000 K, and different densities.

In the previous case the Saha-Boltzmann distribution was used, see Sect. “Radio recombination lines, to determine the electron population. In the Non-LTE case, the radiative transitions are dominating over collisional transitions. In order to describe these departures one introduces for each electronic level \(n\) a so-called departure coefficient \(b_n\) relating the electron population in Non-LTE \(N_n\) with the LTE case \(N_n^*\),

\[N_n = b_n \, N_n^*.\]

These departure coefficients have been derived by [Storey_Hummer_1995], see Fig. 7, and depend on both collisional and radiative processes. XCLASS uses their tabulated coefficients [20] for optical thin and optical thick (default) HII-regions for all RRLs. Due to the fact that the departure coefficients depends on the electron density, the Non-LTE description of the RRLs requires an additional parameter, the electron density \(N_e\) (in cm\(^{-3}\)) for each component. Using the departure coefficients the Non-LTE source function \(S_{{\rm RRL}, \nu}^{\rm Non-LTE}\) can be expressed as

(25)\[S_{{\rm RRL}, \nu}^{\rm Non-LTE} = \frac{\tau_{{\rm RRL}, \nu} \cdot b_{n_2} \cdot J(T_e)}{\tau_{{\rm RRL}, \nu} \cdot b_{n_1} \cdot \beta},\]

where

(26)\[\beta = \frac{1 - \left(b_{n_2} / b_{n_1}\right) \, e^{-h \, \nu / k_B \,T_e}}{1 - e^{-h \, \nu / k_B \,T_e}}.\]

For \(b_{n_1} =b_{n_2} = 1\) Eq. (25) reduces to the LTE expression (24).

The hot plasma in HII regions gives rise to the emission of thermal bremsstrahlung which causes a continuum opacity which will be described in Sect. “Free-free contribution”. If the input parameter t_back_flag indicates a phenomenological continuum description (t_back_flag = True, \(\gamma \equiv 0\)) the free-free contribution to the continuum is neglected. For a physical description, i.e. (t_back_flag = False, \(\gamma \equiv 1\)) this contribution is taken into account, see Sect. “Free-free contribution”.

The description of radio recombination lines in Non-LTE in the molfit ifle requires two additional parameters: the electron density (in cm\(^{-3}\)), which is given in a column right after the emission measure (N_e), and the keyword non-lte in an additional column at the end of the line. In order to use departure coefficients [Storey_Hummer_1995] for optical thin lines the column keyword has to contain keyword thin as well. In the next example, we describe the first component of the example mentioned above in Non-LTE, while the second component is described still in LTE:

RRL-H    2
% s_size:   T_e:  EM_RRL:   N_e:   V_width:   V_off:  distance:   keyword:
500.00000  6.9e4  5.5E+16  5.1e9        1.3      2.0      2.e15    non-lte
500.00000  2.3e4  2.3E+12               2.5      2.0      1.e15

Line profile function

../_images/LineProfiles.png

Fig. 8 Comparison of Gaussian, Lorentz, and their convolution (Voigt) line profiles. Each profile has the same area. The full widths at half-intensity are 20 for the Gaussian and Lorentz profiles, and 35 for the resulting Voigt profile. (Taken from [Lines_2002]).

XCLASS offers the possibility to use four different line profile functions to describe the line profile of each component which are briefly described in the following.

Lorentzian line profile

The normalized, i.e. \(\int_0^{\infty} \phi^{m,c,t}_{\rm Lorentz}(\nu) \, d\nu\) = 1, (Cauchy-) Lorentzian profile can be written as

\[\phi^{m,c,t}_{\rm Lorentz}(\nu) = \frac{1}{\pi} \, \frac{\delta^{m,c,t}}{(\nu - \nu_0^{m,c,t})^2 + (\delta^{m,c,t})^2},\]

where

(27)\[\nu_0^{m,c,t} = \nu^t \cdot \left(1-\frac{\left(\delta {\rm v}_{\rm offset}^{m,c} + {\rm v}_{\rm LSR}\right)}{c_{\rm light}} \right),\]

describes the Doppler-shifted transition frequency \(\nu^t\) and \(\delta^{m,c,t} = f_L / 2\) the half-width at half-maximum which is related to the full width at half maximum \(f_L\) by

(28)\[f_L = 2 \, \delta^{m,c,t} = \frac{\Delta {\rm v}_{\rm width-Lorentz}^{m,c}}{c_{\rm light}} \cdot \nu_0^{m,c,t}.\]

Additional, \(\Delta {\rm v}_{\rm width-Lorentz}^{m,c}\) and \(\delta {\rm v}_{\rm offset}^{m,c}\) indicates the line width and offset of molecule \(m\) and component \(c\), respectively and \({\rm v}_{\rm LSR}\) the source velocity.

In order to describe a component with a Lorentzian line shape the keyword Lorentz has to be added to column keyword for the corresponding component. The line width (defined in column V_width) is now interpreted as Lorentz line width \(\Delta {\rm v}_{\rm width-Lorentz}^{m,c}\). In the following example, we describe the first component with a Lorentzian line shape, while the second component uses a Gaussian line shape:

SO2;v=0;   2

% s_size:  T_rot:   N_tot:  V_width:   V_off:  CFFlag:  keyword:
%[arcsec]     [K]   [cm-2]    [km/s]   [km/s]    [c/f]
    48.47   50.00  3.9E+16      2.86    -2.53        c   Lorentz
    29.08   68.44  2.0E+17      4.02     0.22        f

Gaussian line profile

In order to take broadening caused by the thermal motion of the gas particles and micro-turbulence into account a normalized Gaussian line profile, i.e. \(\int_0^{\infty} \phi^{m,c,t}_{\rm Gauss}(\nu) \, d\nu\) = 1, for a spectral line \(t\) can be assumed:

\[\phi^{m,c,t}_{\rm Gauss}(\nu) = \frac{1}{\sqrt{2 \pi} \, \sigma^{m,c,t}} \cdot e^{-\frac{\left(\nu - \nu_0^{m,c,t} \right)^2} {2 (\sigma^{m,c,t})^2}},\]

where \(\nu_0^{m,c,t}\) indicates the Doppler-shifted transition frequency described by Eq. (27) and \(\sigma^{m,c,t}\) the standard deviation which is related to the full width at half maximum \(f_G\) by

(29)\[f_G = \left(2 \, \sqrt{2 \, \ln 2}\right) \cdot \sigma^{m,c,t} = \frac{\Delta {\rm v}_{\rm width-Gauss}^{m,c}}{c_{\rm light}} \cdot \left(\nu^t + \delta \nu_{\rm LSR}^{m,c,t} \right).\]

Similar to the Lorentzian line profile \(\Delta {\rm v}_{\rm width-Gauss}^{m,c}\) and \(\delta {\rm v}_{\rm offset}^{m,c}\) indicates the line width and offset of molecule \(m\) and component \(c\), respectively and \({\rm v}_{\rm LSR}\) the source velocity.

The myXCLASS function assumes a Gaussian line shape for all components by default. (In order to explicitly describe a component with a Gaussian line shape, the keyword Gauss has to be added to column keyword for the corresponding component.)

Voigt profile

The Voigt profile is a convolution of the Gaussian and the Lorentzian function, i.e.

\[V(x, \sigma, \gamma) = \int_{-\infty}^{\infty} G(x; \sigma) \, L(x - x', \gamma) \, dx'.\]

The integral can not be solved analytically, but the Voigt function can be identified as real part of the Feddeeva function

\[w(x) := e^{-z^2} \, {\rm erfc}(i \, z) = {\rm erfcx}(i \, z) = e^{-z^2} \, \left(1 + \frac{2 \, i}{\sqrt{\pi}} \cdot \int_0^t e^{t^2} dt \right),\]

with argument

\[z = \frac{x + i \gamma}{\sigma \, \sqrt{2}},\]

i.e.

\[V(x, \sigma, \gamma) = \frac{{\rm Re}\left[z\right]}{\sqrt{2 \pi}}.\]

Due to the fact that the computation of the Voigt profile is computational quite expensive, XCLASS uses the pseudo-Voigt profile which is an approximation of the Voigt profile \(V(x)\) using a linear combination of a Gaussian \(G(x)\) and a Lorentzian line profile function \(L(x)\) instead of their convolution.

The mathematical definition of the normalized, i.e. \(\int_0^{\infty} \phi^{m,c,t}_{\rm pseudo-Voigt}(\nu) \, d\nu\) = 1, pseudo-Voigt profile is given by

\[\phi^{m,c,t}_{\rm pseudo-Voigt}(\nu) = \eta \cdot L (\nu, f) + (1 - \eta) \cdot G (\nu, f) \; \text{with} \; 0 < \eta < 1,\]

where \(f_L\) and \(f_G\) represents the Lorentzian Eq. (28) and Gaussian Eq. (29) full width at half maximum (FWHM), respectively. There are several possible choices for the \(\eta\) parameter. XCLASS use the simple formula [Thompson_et_al_1987]

\[\eta = 1.36603 \, \left(\frac{f_L}{f}\right) - 0.47719 \, \left(\frac{f_L}{f}\right)^2 + 0.11116 \, \left(\frac{f_L}{f}\right)^3,\]

where

\[f = \left[f_G^5 + 2.69269 \, f_G^4 \, f_L + 2.42843 \, f_G^3 \, f_L^2 + 4.47163 \, f_G^2 \, f_L^3 + 0.07842 \, f_G \, f_L^4 + f_L^5\right]^{1/5},\]

which is accurate to 1%.

In the molfit file, the Voigt line profile requires three additional columns (V_width_G, V_width_L, and keyword indicating the Gaussian line width (in km s\(^{-1}\)), the Lorentz line width (in km s\(^{-1}\)) and keyword Voigt, respectively) for each component which is described with a Voigt line profile, while column V_width is not needed. In the following example, the first component of SO\(_2\) is described with a Voigt line profile, while the second and third component use a Lorentzian and a Gaussian line shape, respectively.

SO2;v=0;   3

% s_size: T_rot: N_tot: V_width: V_width_G: V_width_L:   V_off: CFFlag: keyword:
%[arcsec]    [K] [cm-2]   [km/s]     [km/s]     [km/s]   [km/s]   [c/f]
      50.00  50.00 3.E+16                2.86       1.12    -2.53       c Voigt
      50.00  99.24 6.E+17     1.02                           3.22       f Lorentz
      50.00  68.44 2.E+17     4.02                           0.22       f

For RRLs, there is a contribution due to the broadening of the electronic levels in regions with high electron density. This is known as pressure broadening and it produces a Lorentzian profile. This broadening can be specified by the half-width at half-maximum height, \(f_L\), which has been computed numerically using analytic expressions that approximate the results derived from classical and semi-classical theory [Gee_et_al_1976] for the different quantum number ranges of the RRLs

Table 1 Analytic approximations of the pressure broadening for different ranges of electronic levels. References: (a) [Brocklehurst_Seaton_1972], (b) [Walmsley_1990] (c) [Strelnitski_et_al_1996].

Range of electronic levels

Pressure broadening, \(f_L\)

n \(>\) 100

1.9 \(\times\) 10\(^{-8} \, n^{4.4} \, N_ e / \left(Z^2 \, T_e^{0.1}\right)^a\)

30 \(<\) n \(<\) 100

6.7 \(\times\) 10\(^{-9} \, n^{4.6} \, N_ e / \left(Z^2 \, T_e^{0.1}\right)^b\)

n \(<\) 30

8.0 \(\times\) 10\(^{-10} \, n^{5} \, N_ e / \left(Z^2 \, T_e^{0.1}\right)^c\)

In order to use the expressions described in Tab. “Analytic approximations of the pressure broadening for different ranges of electronic levels. References: (a) [Brocklehurst_Seaton_1972], (b) [Walmsley_1990] (c) [Strelnitski_et_al_1996].” for the pressure broadening for a specific component of a RRL, the Lorentzian line width has to be set to a negative value, e.g.

RRL-H   1

% s_size: T_rot: N_tot: V_width: V_width_G: V_width_L:   V_off: CFFlag: keyword:
%[arcsec]    [K] [cm-2]   [km/s]     [km/s]     [km/s]   [km/s]   [c/f]
    50.00  50.00 3.E+16                2.86       -1.0    -2.53       c Voigt

Horn profile

../_images/Horn-line-profile.png

Fig. 9 Synthetic Horn line-profiles, for three different Horn H parameters. The full width at zero level is \(\Delta \nu\) = 8 km/s in all cases.

Horn line profiles [21] [Potter_et_al_2004] are like those encountered in envelopes of stars. They are parameterized by the transition frequency \(\nu_0^{m,c,t}\), the line width and the Horn to Center ratio. The aspect of the profile varies from parabola (as obtain in optically thick lines) for Horn/Center = -1 to flat-topped lines (unresolved optically thin lines) for Horn/Center = 0 and double peaked profiles (resolved optically thin lines) for Horn/Center > 0. Note, a Horn/Center parameter \(H < (-1)\) is not allowed. The profile is symmetric, normalized, i.e. \(\int_0^{\infty} \phi^{m,c,t}_{\rm Horn}(\nu) \, d\nu\) = 1, and is described by

(30)\[\begin{split}\phi^{m,c,t}_{\rm Horn}(\nu) = \begin{cases} \frac{1}{f_H} \, \frac{1 + 4 \, H \, \left((\nu - \nu_0^{m,c,t}) / f_H \right)^2}{1 + H/3}, & {\rm if} \quad \left|\nu - \nu_0^{m,c,t}\right| \leq f_H / 2\\ 0, & {\rm otherwise},\\ \end{cases}\end{split}\]

where \(H\) represents the dimensionless Horn/Center parameter and \(f_H\) the full width at half maximum

(31)\[f_H = \frac{\Delta {\rm v}_{\rm width-Horn}^{m,c}}{c_{\rm light}} \cdot \nu_0^{m,c,t},\]

where \(\nu_0^{m,c,t}\) again represents the Doppler-shifted transition frequency described by Eq. (27).

The Horn line profile requires three additional columns (V_width_Horn_W, V_width_Horn_H, and keyword indicating the Horn line width \(\Delta {\rm v}_{\rm width-Horn}^{m,c}\) (in km s\(^{-1}\)), the Horn H parameter (dimensionless) and keyword Horn, respectively) for each component which is described with a Horn line profile, while column V_width is not needed. In the following example, the first component of SO\(_2\) is described with a Horn line profile.

SO2;v=0;   1

% s_size: T_rot: N_tot: V_width_Horn_W: V_width_Horn_H: V_off: CFFlag: keyword:
%[arcsec]    [K] [cm-2]          [km/s]              [] [km/s]   [c/f]
    50.00  50.00 3.E+16            2.86            1.12  -2.53       c Horn

Contributions to continuum

../_images/ContinuumContributions.png

Fig. 10 Sketch of different continuum contributions of Arp 220-West. (Taken from [Downes_Eckart_2007]).

XCLASS offers the possibility to describe the dust, free-free, and synchrotron contribution to the continuum, see Fig. 10, self-consistently, where the different contributions are modeled by a set of parameters. These parameters can be defined globally for each frequency range or in the molfit file for each distance. For globally defined parameters, the myXCLASS function applies the continuum parameters only to the component with the largest beam filling factor, see Eq. (2), for each distance to avoid an overestimation.

Dust contribution

The dust optical depth \(\tau_d^{m,c}(\nu)\) takes the dust attenuation into account and is given by

(32)\[\begin{split}\tau_d^{m,c}(\nu) &= \tau_{d, {\rm ref}}^{m,c} \cdot \left(\frac{\nu}{\nu_{\rm ref}} \right)^{\beta^{m,c}} + \tau_d^{\rm file}(\nu) \\ &= \left(N_H^{m,c} \cdot \kappa^{m,c}_{\nu_{\rm ref}} \cdot m_{H_2} \cdot \frac{1}{\zeta_{\rm gas-dust}}\right) \cdot \left(\frac{\nu}{\nu_{\rm ref}} \right)^{\beta^{m,c}} + \tau_d^{\rm file}(\nu).\end{split}\]

Here, \(N_H^{m,c}\) describes the hydrogen column density (in cm\(^{-2}\)), \(\kappa^{m,c}_{\nu_{\rm ref}}\) the dust mass opacity for a certain type of dust (in cm\(^{2}\) g\(^{-1}\)) [Ossenkopf_Henning_1994], and \(\beta^{m,c}\) the spectral index. Additionally, \(\nu_{\rm ref}\) = 230 GHz indicates the reference frequency for \(\kappa^{m,c}_{\nu_{\rm ref}}\), \(m_{H_2}\) the mass of a hydrogen molecule, and \(1 / \zeta_{\rm gas-dust}\) the dust to gas ratio which is set here to (1/100) [Hildebrand_1983]. The equation is valid for dust and gas well mixed. Furthermore, the expression \(\tau_d^{\rm file}(\nu)\) represents the dust optical depth described in an ASCII file, which path and name is defined by the input parameter DustFileName.

In Eqn. (1) and (9), the dust continuum contribution is described through the source function \(S^{m,c}(\nu)\), Eq. (5), by an effective dust temperature \(T_d^{m,c}\) (in K) (through \(J(T_d^{m,c}, \nu)\)).
The myXCLASS function offers three different ways to define the dust parameters \(N_H^{m,c}\), \(\kappa^{m,c}_{\nu_{\rm ref}}\), \(\beta^{m,c}\), and \(T_d^{m,c}\):

On the one hand the dust parameters \(N_H^{m,c}\), \(\kappa^{m,c}_{\nu_{\rm ref}}\), and \(\beta^{m,c}\) can be defined globally for all components in the molfit file by the input parameters N_H, kappa_1300, and beta_dust of the myXCLASS function, while the dust temperature is set to the excitation temperature of the corresponding component. For the myXCLASSFit, myXCLASSMapFit, myXCLASSMapRedoFit, LineID, and CubeFit function the dust parameters can be defined globally in an obs. xml file as well, see Sect. “Observational xml file”.

Additionally, the dust parameters can be described in the molfit file for different distances, while the dust continuum has to be marked as cont-dust. Each component is described by the size of the dust component s_size (in arcsec), the dust temperature T_cont_dust (in K), the hydrogen column density nHcolumn_cont_dust (in cm\(^{-2}\)), the dust mass opacity kappa_cont_dust (in g cm\(^{-2}\)), the spectral index beta_cont_dust, and the distance parameter distance. In the example described below, we define four different dust components with different dust temperatures at different distances

cont-dust    4
% s_size: T_cont_dust: nHcolumn_cont_dust: kappa_cont_dust: beta_cont_dust: distance:
500.00000       50.000             3.9E+16             0.42             2.0    4.e15
500.00000       10.000             1.2E+20             0.42             2.0    3.e15
500.00000       32.000             5.3E+12             0.42             2.0    2.e15
500.00000        5.000             7.1E+18             0.42             2.0    1.e15

In order to be backward compatible with older XCLASS releases, the user can define the dust parameters for each component of a molecule as well. Here, the dust temperature \(T_d^{m,c}\) is parameterized as offset to the corresponding excitation temperature \(T_{\rm ex}^{m,c}\) as

(33)\[\begin{split}T_d^{m,c} (\nu) &= T_{\rm ex}^{m,c} + \Delta T_d^{m,c}(\nu) \\ &= T_{\rm ex}^{m,c} + T_{d, {\rm off}}^{m,c} \cdot \left(\frac{\nu}{\nu_{\rm min}} \right)^{T_{d, {\rm slope}}^{m,c}},\end{split}\]

where \(T_{d, {\rm off}}^{m,c}\) and \(T_{d, {\rm slope}}^{m,c}\) can be defined by the user for each component in the molfit. If \(T_{d, {\rm off}}^{m,c}\) and \(T_{d, {\rm slope}}^{m,c}\) are not defined for a certain component, we assume \(T_d^{m,c} (\nu) \equiv T_{\rm ex}^{m,c} (\nu)\) for all components. For a physical (\(\gamma \equiv 1\)) description of the background intensity, see Eq. (5), the user can define the dust opacity, Eq. (32), and dust temperature, Eq. (33), for each component. In order to define a dust temperature for each component which is not identical to the excitation temperature \(T_{\rm ex}\) of the corresponding component, the molfit file has to contain two additional columns between columns V_off and CFFlag (or column distance), describing the parameters \(T_{\rm d, off}^{m,c}\) (T_doff) and \(T_{\rm d, slope}^{m,c}\) (T_dSlope), respectively, and keyword t-dust-inline in column (keyword), e.g.

CS;v=0;    3
% s_size: T_rot: N_tot: V_width: V_off: T_doff: T_dSlope: CFFlag: keyword:
%[arcsec]    [K] [cm-2]   [km/s] [km/s]     [K]        []   [c/f]
    48.47  50.00   3e16     2.86  -2.56     3.0       0.0       c t-dust-inline
    40.10  56.53   2e18     4.21  -7.31     3.0       0.0       c t-dust-inline
    29.09  68.44   2e17     4.02   0.21     2.5       1.0       c t-dust-inline

Additionally, the myXCLASS program offers the possibility to define \(N_H\), \(\kappa_{\nu_{\rm ref}}\) and \(\beta\) for each component of a molecule as well. In order to define these parameters individually for each component, the molfit file has to contain three additional columns on the left side of column CFFlag (or column distance), indicating parameters \(N_H^{m,c}\) (nHcolumn), \(\kappa^{m,c}_{\nu_{\rm ref}}\) (kappa), and \(\beta^{m,c}\) (beta), and keyword nh-dust-inline in column (keyword), e.g.

CS;v=0;    3
% s_size: T_rot: N_tot: V_width: V_off: nHcolumn:  kappa: beta: CFFlag: keyword:
%[arcsec]    [K] [cm-2]   [km/s] [km/s]    [cm-2] [cm2/g]    []   [c/f]
    48.47  50.00   3e16     2.86  -2.56      3e24    0.42   2.0       c nh-dust-inline
    56.53  20.00   2e18     4.21  -7.31      2e24    0.42   2.1       c nh-dust-inline
    68.44  90.00   2e17     4.02   0.21      3e21    0.42   1.9       c nh-dust-inline
In order to define \(T_{d, {\rm off}}^{m,c}\), \(T_{d, {\rm slope}}^{m,c}\), \(N_H\), \(\kappa_{\nu_{\rm ref}}\), and \(\beta\) for each component the columns defining the dust temperature\(T_{d, {\rm off}}^{m,c}\) (T_doff), \(T_{d, {\rm slope}}^{m,c}\) (T_dslope) have to be given before the hydrogen column density, kappa and beta, i.e. in the order of s_size, T_rot, N_tot, V_width, V_off, T_doff, T_dslope, nHcolumn, kappa, beta, and CFFlag. Additionally, the column (keyword) has to contain the keyword t-dust-inline_nh-dust-inline.

Free-free contribution

The optical depth of the free-free contribution in terms of the classical electron radius \(r_e = (\alpha \, \hbar \, c) / (m_e \, c^2) = \alpha^2 \, a_0\) is given by [22]

(34)\[\begin{split}\tau_{\rm ff} &= \frac{4}{3} \, \left(\frac{2 \pi}{3} \right)^{(1/2)} \, r_e^3 \, \frac{Z_i^2 m^{3/2} \, c^5}{\sqrt{k_B \, T} \, h \, \nu^3} \, \left(1 - e^{-\frac{h \nu}{k_B T}} \right) \, \langle g_{\rm ff} \rangle \cdot {\rm EM} \\ &= 1.13725 \cdot \left(1 - e^{-\frac{h \nu}{k_B T}} \right) \cdot \langle g_{\rm ff} \rangle \cdot \left[ \frac{T}{\rm K}\right]^{-\frac{1}{2}} \, \left[ \frac{\nu}{\rm GHz}\right]^{-3} \, \left[ \frac{\rm EM}{{\rm pc} \, {\rm cm}^{-6}}\right].\end{split}\]

where the electron temperature \(T_e\) and the emission measure EM has to be defined by the user. Additionally, XCLASS uses the tabulated thermal averaged free-free Gaunt coefficients [23] \(\langle g_{\rm ff} \rangle\) derived by [Van_Hoof_et_al_2015], which include relativistic effects as well, see Fig. 11.

../_images/Gaunt-coefficients.png

Fig. 11 Relativistic thermally averaged free-free Gaunt factors \(\langle g_{\rm ff} \rangle\) for different temperatures \(T_e\) as function of \(u = \hbar \, \omega / k_B \, T_e\), derived by [Van_Hoof_et_al_2015].

In the molfit file contributions of free-free transitions to the continuum are indicated by the phrase cont-free-free. Each free-free component is described by the size of the component s_size (in arcsec) the electronic temperature Te_ff (in K) the emission measure EM_ff (in pc cm\(^{-6}\)) and the distance. In the following example, we describe the free-free contribution at three different distances:

cont-free-free    3
% s_size:    Te_ff:     EM_ff:   distance:
500.00000     9.0e4     6.3E+5       9.e15
500.00000     4.2e4     3.9E+7       4.e15
500.00000     1.0e4     2.9E+3       1.e15

Similar to the dust contribution, the free-free continuum parameters can be defined globally, by using the input parameters Te_ff and EM_ff. For the myXCLASSFit, myXCLASSMapFit, myXCLASSMapRedoFit, LineID, and CubeFit function the free-free parameters can be defined globally in an obs. xml file as well, see Sect. “Observational xml file”.

Bound-free contribution

In one of the next releases XCLASS will be able to describe the bound-free contribution to the continuum as well. Free-bound transitions contribute significantly to the NIR-emission of HII-regions. At NIR-frequencies the free-bound radiation dominates over free-free emission. Assuming LTE conditions, we can use Saha’s equation and obtain the Milne relation

\[\frac{\sigma_{\rm fb}}{\sigma_{\rm bf}} = \frac{2 \, \left( h \, \nu\right)^2}{\left(m \, c^2\right) \, \left(m \, v^2\right)} \, \frac{g_n}{g_e \, g_+}.\]

[Karzas_Latter_1961] obtained for hydrogenic atoms the cross section

\[\sigma_{\rm bf} = \frac{512 \, \pi^7 \, m \, e^{10} \, Z^4}{3 \sqrt{3} \, c \, \omega^3 \, h^6 \, n^5} \, g\left(\omega, n, l, Z\right),\]

with a Gaunt factor \(g\left(\omega, n, l, Z\right)\) depending on the radial and angular quantum numbers \((n, l)\). This Gaunt factor is close to 1.

Integrating the number of emitted photons \(N_+ \, N_e \, \sigma_{\rm fb} \, f(v) \, v \, dv\) times \(h \, \nu \, \delta(\nu - \xi/h - m v^2 / 2)\) over the electron velocity distribution gives

\[\begin{split}\epsilon_{(n,\nu)} d \nu &= N_e \, N_i \, \frac{A}{\left(k_B \, T\right)^{3/2}} \, \frac{\overline{g}}{n^3} \, \exp\left[-\frac{h \nu}{k_B T} + \frac{\chi}{n^2 \, k_B \, T}\right] d \nu\\ A &= \frac{16 \sqrt{2 \pi} Z^2 \, e^6 \, \sqrt{m}}{3 \sqrt{3} m^2 c^3} \chi, \\ \chi &= \frac{2 \pi^2 \, m \, e^4 \, Z^2}{h^2},\end{split}\]

which is the emissivity for radiative recombinations to states of radial quantum number \(n\). The spectral emissivity for all recombinations is the sum over all transitions to ionization energies \(\chi/n^2\) less than the photon energy. So

(35)\[\epsilon_\nu = \sum_{n \ge k}^\infty \epsilon_{(n,\nu)}, \quad \text{with} \; k = \sqrt{\chi/h \, \nu}\]

is the free-bound emissivity for hydrogenic ions of charge \(Z\). The sum included in Eq. (35) can be evaluated using the approximation of [Brussaard_Van_de_Hulst_1962].

In the molfit file contributions of bound-free transitions to the continuum are described by the free-free parameter, see Sect. “Free-free contribution” and can be computed only in conjunction with the free-free contribution.

Synchrotron contribution

Synchrotron radiation (also known as magnetobremsstrahlung radiation) is the electromagnetic radiation emitted when charged particles are accelerated radially, i.e., when they are subject to an acceleration perpendicular to their velocity. If the particle is non-relativistic, then the emission is called cyclotron emission otherwise the emission is called synchrotron emission. In astronomy, synchrotron radiation is produced by fast electrons moving through magnetic fields and has a characteristic polarization.

The emissivity \(\epsilon_{\rm sync}^{m,c}(\nu)\), see Eq. (5), of a power-law electron energy distribution \(N(E) = \kappa E^{-p}\) (expressed in electron m\(^{-3}\) GeV\(^{-1}\)) in the case of a random magnetic field is [24]

\[\epsilon_{\rm sync}^{m,c}(\nu) = \frac{J(\nu)}{4 \pi} \cdot \frac{1}{2 \, k_B \, \nu^2 / c^2},\]

where

(36)\[\begin{split}J(\nu) &= \frac{\sqrt{3} \, e^3 \, B \, \kappa}{4 \, \pi \, \epsilon_0 \, c \, m_e} \cdot \left(\frac{3 \, e \, B}{2 \, \pi \, \nu \, m_e^3 \, c^4} \right)^{(p - 1)/2} \cdot a(p) \\ &= 2.344 \times 10^{-25} \, a(p) \, B^{(p + 1) / 2} \, \kappa' \, \left(3.217 \times 10^{17}\right)^{(p - 1)/2} \,\, \text{W} \, \text{m}^{-3} \, \text{Hz}^{-1},\end{split}\]

with

\[a(p) = \frac{\sqrt{\pi}}{2} \, \frac{\Gamma \left(\frac{p}{4} + \frac{19}{12}\right) \, \Gamma \left(\frac{p}{4} - \frac{1}{12}\right) \, \Gamma \left(\frac{p}{4} + \frac{5}{4}\right)}{(p + 1) \, \Gamma \left(\frac{p}{4} + \frac{7}{4}\right)}.\]

The absorption coefficient \(\kappa_{\rm sync}^{m,c} (\nu)\), see Eq. (5), for a random magnetic field is

(37)\[\begin{split}\kappa_{\rm sync}^{m,c} (\nu) \equiv \chi_\nu &= \frac{\sqrt{3} \, e^3 \,c}{8 \, \pi^2 \, \epsilon_0 \, m_e} \cdot \kappa \cdot B^{(p + 2) / 2} \cdot \left(\frac{3 \, e}{2 \, \pi \, m_e^3 \, c^4} \right)^{p/2} \cdot b(p) \, \nu^{-(p + 4) / 2} \\ &= 20.9 \, \kappa' \, B^{(p + 2) / 2} \, \left(5.67 \times 10^{9}\right)^p \, b(p) \, \nu^{-(p + 4) / 2} \,\, \text{m}^{-1},\end{split}\]

where

\[b(p) = \frac{\sqrt{\pi}}{8} \, \frac{\Gamma \left(\frac{3p + 22}{12}\right) \, \Gamma \left(\frac{3p + 2}{12}\right) \, \Gamma \left(\frac{p + 6}{4}\right)}{\Gamma \left(\frac{p + 8}{4}\right)}.\]

The description of synchrotron radiation (see Sect. “Synchrotron contribution”) in the molfit file is indicated by cont-sync. Here, each component defines the size of the component s_size (in arcsec) the energy of the electrons kappa_sync (in electron m\(^{-3}\) GeV\(^{-1}\)), the strength of the magnetic field B_sync (in Gauss), the spectral energy index p_sync (dimensionless), the thickness of the slab l_sync (in AU), and the distances distance. In the following example, we model the synchrotron contribution at two different distances:

cont-sync    2
% s_size:   kappa_sync:   B_sync:   p_sync:   l_sync:   distance:
500.00000         9.0e4    6.3E-7       3.3     1.e12        2.e9
500.00000         5.0e4    6.3E-7       3.3     1.e12        1.e9

Similar to the dust contribution, the synchrotron continuum parameters can be defined globally, by using the input parameters kappa_sync, B_sync, p_sync, and l_sync. For the myXCLASSFit, myXCLASSMapFit, myXCLASSMapRedoFit, LineID, and the CubeFit function the synchrotron parameters can be defined globally in a obs. xml file as well, see Sect. “Observational xml file”.

Phenomenological continuum description

In addition to the the different continuum contributions described above, XCLASS offers the possibility to describe the continuum of a given distance by phenomenological functions. Currently, the following functions are included:

  • Standing wave (ContPhenFuncID = 1):

    \[c(\nu) = A_1 \, \sin \left( (\nu \cdot x) + \phi_1 \right) + A_2 \, \cos \left( (\nu \cdot x) + \phi_2 \right),\]

    where \(A_{1, 2}\) and \(x\) describe scale parameters whereas \(\phi_{1,2}\) indicate phase shifts. Here, \(A_1\) and \(A_2\) are defined by parameters ContPhenFuncParam1 and ContPhenFuncParam2, respectively. Parameter ContPhenFuncParam3 indicates the scale parameter \(x\) and the phase shifts \(\phi_{1,2}\) are described by parameters ContPhenFuncParam4 and ContPhenFuncParam5, respectively.

  • Planck function (ContPhenFuncID = 2):

    \[c(\nu) = \frac{h \, \nu}{k_B}\frac{1}{e^{h \, \nu / k \, T} - 1},\]

    with temperature \(T\) described by input parameter ContPhenFuncParam1. The other input parameters ContPhenFuncParam2 - ContPhenFuncParam5 have to be set to zero.

  • beam-averaged continuum background temperature (ContPhenFuncID = 3):

    (38)\[c(\nu) = T_{\rm bg} \cdot \left(\frac{\nu}{\nu_{\rm ref}} \right)^{T_{\rm slope}},\]

    where \(\nu_{\rm ref}\) (described by input parameter ContPhenFuncParam1) indicates the reference frequency (in MHz) for \(T_{\rm bg}\) (defined by input parameter ContPhenFuncParam2). The temperature slope (unitless) \(T_{\rm slope}\) is defined by input parameter ContPhenFuncParam3. If input parameter ContPhenFuncParam4 is set to zero Eq. (38) is applied to all frequency ranges. But the user can restrict the application of Eq. (38) to a given frequency range by using input parameters ContPhenFuncParam1 and ContPhenFuncParam4, where ContPhenFuncParam1 defines the first and ContPhenFuncParam4 the last frequency, respectively. The other input parameter ContPhenFuncParam5 must always be set to zero.

  • Anomalous Microwave Emission (AME) (ContPhenFuncID = 4):

    \[c(\nu) = {A_{\rm AME}} \cdot \mathrm{exp} \left[-\frac{1}{2} \cdot \left[ \frac{ \ln \left(\frac{\nu }{{\nu_{\rm AME}}} \right)}{W_{\rm AME}}\right]^2 \right],\]

    where \(A_{\rm AME}\) (described by input parameter ContPhenFuncParam2) indicates the the maximum flux intensity (in K) due to AME, \(\nu_{\rm AME}\) (described by input parameter ContPhenFuncParam3) the correspondent frequency (in MHz) for that maximum, and \(W_{\rm AME}\) (described by input parameter ContPhenFuncParam4) the width (unitless) of the distribution on the log-log plane. For more details see e.g. [Stevenson_2014] and [Poojon_2024].

  • CMB Emission (ContPhenFuncID = 5):

    \[c(\nu) = \frac{x^2 \, \mathrm{e^x}}{(\mathrm{e}^x - 1)^2} \cdot \Delta {T_{\rm CMB}},\]

    where \(x = \left( h \, \nu / k_B \, T_{\rm CMB} \right)\) and \(\Delta T_{\rm CMB}\) (described by input parameter ContPhenFuncParam2) models the CMB anisotropies (in K). Here \(T_{\rm CMB}\) is fixed to 2.72548 K. For more details see e.g. [Poojon_2024].

Note, the aforementioned continuum function \(c(\nu)\) is added to the background intensity of the corresponding distance.

In order to use a phenomenological description of the continuum in the molfit file, the phrase cont-phen has to be used. Each component is described by the source size s_size (in arcsec) (or other parameters related to source description), the function-id func-id, the input parameters p_1, p_2, p_3, p_4, and p_5, representing the input parameters ContPhenFuncParam1 - ContPhenFuncParam5 and the distance distance. In the example describe below, we use an expression for a standing wave to describe the continuum phenomenologically at a given distance:

cont-phen    1
% s_size:   func-id:   p_1:   p_2:   p_3:   p_4:   p_5:   distance:
500.00000          1    0.2    0.2    0.2    0.2    0.2       1.e20

Similar to the dust contribution, the phenomenological continuum parameters can be defined globally, by using the input parameters ContPhenFuncID, ContPhenFuncParam1, ContPhenFuncParam2, ContPhenFuncParam3, ContPhenFuncParam4, and ContPhenFuncParam5. For the myXCLASSFit, myXCLASSMapFit, myXCLASSMapRedoFit, LineID, and CubeFit function the phenomenological parameters can be defined globally in a obs. xml file as well, see Sect. “Observational xml file”.

Derivations

Detection Equation

In absence of scattering, the radiative transfer equation

\[\frac{\mathrm{d} I_{\nu}}{\mathrm{d} s} = -\kappa_{\nu}(s) I_{\nu} + \epsilon_{\nu}(s)\]

describes the propagation of radiation which passes through a medium. During the propagation photons are absorbed and emitted indicated by the emission and absorption coefficients \(\epsilon_\nu\) and \(\kappa_{\nu}\), respectively.

The optical depth \(\tau_{\nu}(s)\) which measures the distance in units of the mean free path, is given by

(39)\[\tau_{\nu}(s) = \int_{s'=0}^{s'=s} \kappa_{\nu} \, d s'.\]

By using the source function \(S_{\nu} = \frac{\epsilon_{\nu}}{\kappa_{\nu}}\), the radiative transfer equation can be expressed as

(40)\[\frac{\mathrm{d} I_{\nu}}{\mathrm{d} \tau} = I_{\nu}(\tau) + S_{\nu}(\tau).\]

Within the local thermodynamic equilibrium (LTE) the source function is described by Planck’s law for black body radiation \(B_{\nu}(T)\). Integrating Eq. (40) over \(\tau\) leads to

\[I_{\nu}(s) = I_{\nu}(0) \, e^{-\tau_{\nu}} + \int_{\tau'=0}^{\tau'=\tau} S_{\nu}(\tau'_{\nu}) e^{\tau'_{\nu} - \tau_{\nu}} d \tau'_{\nu}.\]

Assuming a constant source function, i.e. constant emission and absorption coefficients through the medium, the transfer equation can be written as

(41)\[I_{\nu}(s) = I_{\nu}(0) \, e^{-\tau_{\nu}} + S_{\nu} \left(1 - e^{-\tau_{\nu}} \right).\]

Additionally, we have to take into account, that the different components may not cover the whole beam, i.e. that the background behind a certain component might contribute as well. So, we have to extend Eq. (41) by introducing the beam filling factor \(\eta\), Eq. (2), which describes the fraction of the beam covert by a component

(42)\[I_{\nu}(s) = \eta \, \left[I_{\nu}(0) \, e^{-\tau_{\nu}} + S_{\nu} \left(1 - e^{-\tau_{\nu}} \right)\right] + \left(1 - \eta\right) \, I_{\nu}(0).\]

Here, the term \(\eta \, I_{\nu}(0) \, e^{-\tau_{\nu}}\) indicates the attenuated radiation from the background \(I_{\nu}(0)\). The second term \(\eta \, S_{\nu} \left(1 - e^{-\tau_{\nu}} \right)\) describes the self attenuated radiation emitted by a certain component. Finally, the last term \(\left(1 - \eta\right) \, I_{\nu}(0)\) represents the contribution of the background which is not covert by a component.

In real observations, we do not measure absolute intensities but only differences of intensities, i.e. we have to subtract the OFF-position from Eq. (42) as well, where we have an intensity caused by the cosmic background \(J_\mathrm{CMB}\). So, we achieve

(43)\[I_{\nu}(s) = \eta \, \left[I_{\nu}(0) \, e^{-\tau_{\nu}} + S_{\nu} \left(1 - e^{-\tau_{\nu}} \right)\right] + \left(1 - \eta\right) \, I_{\nu}(0) - J_\mathrm{CMB}.\]

Beam Filling Factor

../_images/Gaussians.png

Fig. 12 a)~A single two-dimensional Gaussian function, Eq. (44), with \(\sigma_x = \sigma_y = 1.9\) and \(\mu_x = \mu_y = 0\) representing a Gaussian beam of a telescope. b)~Cut through the Gaussian function described in a) at half height. c)~A single two-dimensional Gaussian function, with \(\sigma_x = \sigma_y = 1.1\) and \(\mu_x = 0.5\) and \(\mu_y = 0.9\) representing a Gaussian beam of a point source. d)~Cut through the Gaussian function described in a) at half height. e)~Convolved Gaussian of telescope and point`source as given by Eq. (46). f)~Cut through the convolved Gaussian function described in e) at half height.

The derivation of the beam filling factor (2) starts with the normalized two-dimensional Gaussian function

(44)\[g(x, y, \sigma_x, \sigma_y, \mu_x, \mu_y) = \frac{1}{\sqrt{2 \, \pi \, \left(\sigma_x^2+\sigma_y^2\right)}} \, e^{-\left(\frac{\left(x-\mu_x\right)^2}{2 \, \sigma_x^2}+\frac{\left(y - \mu_y\right)^2}{2 \, \sigma_y^2}\right)},\]

where \(\sigma_x^2\) and \(\sigma_y^2\) describe the variances and \(\mu_x\) and \(\mu_y\) the center along the \(x\) and \(y\) axis, respectively.

Observing a Gaussian shaped extended source with a telescope is described by a convolution of two two-dimensional Gaussian functions:

(45)\[\begin{split}\left(g_1 * g_2 \right) &= \int_{-\infty}^\infty \, \int_{-\infty}^\infty \, g_1(x - u, y - v, \sigma_{x,1}, \sigma_{y,1}, \mu_{x, 1}, \mu_{y, 1}) \cdot g_2(u, v, \sigma_{x,2}, \sigma_{y,2}, \mu_{x,2}, \mu_{y, 2}) \, du \, dv\\ &= \frac{1}{2 \pi \sqrt{\left(\sigma_{x,1}^2 + \sigma_{x,2}^2\right) \left(\sigma_{y,1}^2 + \sigma_{y,2}^2\right)}} \, e^{-\left(\frac{\left(\mu_{x,1} + \mu_{x,2} - x\right)^2}{2 \, \left(\sigma_{x,1}^2 + \sigma_{x,2}^2\right)} + \frac{\left(\mu_{y,1} + \mu_{y,2} - y\right)^2}{2 \, \left(\sigma_{y,1}^2 + \sigma_{ y,2}^2\right)}\right)}.\end{split}\]

Assuming that \(g_1\) describes the telescope with \(\mu_{x,1} = \mu_{y, 1} \equiv 0\) and that telescope and extended source are described by non-elliptical Gaussians, i.e. \(\sigma_{x,1} = \sigma_{y,1} \equiv \sigma_1\) and \(\sigma_{x,2} = \sigma_{y,2} \equiv \sigma_2\), Eq. (45) can be simplified to [25]

(46)\[\left(g_1 * g_2 \right) = \frac{1}{2 \pi \left(\sigma_1^2 + \sigma_2^2\right)} \, e^{-\frac{\left(x - \mu_{x,2}\right)^2 + \left(y - \mu_{y,2}\right)^2}{2 \left(\sigma_1^2 + \sigma_2^2\right)}}.\]

The FWHM of the resulting Gaussian is given by

(47)\[{\rm FWHM} = 2 \sqrt{2 \, \log 2} \sqrt{\sigma_1^2 + \sigma_2^2}\]

which describe an area of

(48)\[\begin{split}A_{\rm conv}^{\rm FWHM} &= \pi \cdot 2 \, \log 2 \cdot \left(\sigma_1^2 + \sigma_2^2\right) \\ &= \frac{\pi}{4} \cdot \left(\theta_1^2 + \theta_2^2\right).\end{split}\]

In the last line we used the relation between the variances \(\sigma_{1,2}\) and the user defined FWHM of telescope (\(\theta_1 \equiv \theta_t\)) and source size (\(\theta_2 \equiv \theta^{m,c}\)) which is given by

(49)\[\theta_i = 2 \, \sqrt{2 \, \log 2} \cdot \sigma_i.\]

The beam filling factor Eq. (2) is defined as ratios of areas

(50)\[\eta_{g_1,g_2} = \frac{A_{\rm source}^{\rm FWHM}}{A_{\rm conv}^{\rm FWHM}} = \frac{\frac{\pi \, \theta_2^2}{4}}{\frac{\pi}{4} \cdot \left(\theta_1^2 + \theta_2^2\right)} = \frac{\theta_2^2}{\left(\theta_1^2 + \theta_2^2\right)},\]

which is completely independent of the position \(\mu_x\) and \(\mu_y\) of the extended source within the telescope beam.

If more extended sources are observed with the telescope, we have to convolve the already convolved Gaussian Eq. (46) with a further two-dimensional Gaussian function with variance \(\sigma_{x,3} = \sigma_{y,3} \equiv \sigma_3\) and center \(\mu_{x,3}\) and \(\mu_{y,3}\). We get

\[\left(\left(g_1 * g_2 \right) * g_3 \right) = \left(g_1 * g_2 * g_3 \right) = \frac{1}{2 \pi \left(\sigma_1^2 + \sigma_2^2 + \sigma_3^2\right)} \, e^{\left(-\frac{\left(\mu_{x,2} + \mu_{x,3} - x\right)^2 + \left(\mu_{y,2} + \mu_{y,3} - y\right)^2}{2 \, \left(\sigma_1^2 + \sigma_2^2 + \sigma_3^2\right)}\right)}.\]

The FWHM of the resulting Gaussian is given by

(51)\[{\rm FWHM} = 2 \sqrt{2 \, \log 2} \sqrt{\sigma_1^2 + \sigma_2^2 + \sigma_3^2},\]

which describe an area of

(52)\[\begin{split}A_{\rm conv}^{\rm next} &= \pi \cdot 2 \, \log 2 \cdot \left(\sigma_1^2 + \sigma_2^2 + \sigma_3^2\right) \\ &= \frac{\pi}{4} \cdot \left(\theta_1^2 + \theta_2^2 + \theta_3^2\right).\end{split}\]

In the last line we used again the relation between the variances and FWHM, Eq. (49). Now, we can again define a beam filling factor similar to Eq. (2) and get

(53)\[\eta_{g_1,g_2,g_3} = \frac{\frac{\pi \, \theta_2^2}{4}+\frac{\pi \, \theta_3^2}{4}}{\frac{\pi}{4} \cdot \left(\theta_1^2 + \theta_2^2 + \theta_3^2\right)} = \frac{\theta_2^2+\theta_3^2}{\left(\theta_1^2 + \theta_2^2 + \theta_3^2\right)}.\]

When we observe more than three extended sources with a telescope we have to iteratively convolve the resulting Gaussian with a further two-dimensional Gaussian function and we can generalize Eq. (53) to

(54)\[\eta_{\rm general} = \frac{\sum_{i > 1} \theta_i^2}{\left(\sum_i \theta_i^2\right)} = \frac{\sum_{i > 1} \theta_i^2}{\theta_t^2 + \sum_{i > 1} \theta_i^2},\]

where \(\theta_1 \equiv \theta_t\) describes the FWHM of the Gaussian shaped beam of the telescope, defined by the diffraction limit, Eq. (3).

Optical Depth

../_images/Sketch__OpticalDepth.png

Fig. 13 Transitions between the lower \(l\) and the upper \(u\) level with the corresponding Einstein coefficients.

In order to derive Eq. (7) we consider a system which involves radiative transitions between a lower \(l\) and an upper \(u\) level only. As shown in Fig. 13 , the lower level has an energy \(E_l\) and the upper level an energy \(E_u > E_l\). With

(55)\[h \, \nu_{u,l} = E_u - E_l\]

describing the energy difference between these two levels we can express the emissivity due to spontaneous radiative decay as

(56)\[\epsilon_{l,u,\nu} = \frac{h \, \nu}{4 \, \pi} \, n_u \, A_{u,l} \, \phi_{l,u} (\nu),\]

where \(A_{u,l}\) describes the Einstein A-coefficient, or radiative decay rate for the transition from the lower \(l\) to the upper \(u\) level. The expression \(1 / A_{u,l}\) gives the averaged time, that a quantum mechanical system can stay in level \(u\) before radiatively decaying to level \(l\), where we assume no collisional (de-)excitation. The expression \(\phi_{l,u} (\nu)\) describes the line profile of the transition of photons of frequency \(\nu\) and is normalized to 1, i.e. \(\int_0^\infty \phi(\nu) \, d \nu = 1\).

Similar to Eq. (56) we can write the extinction coefficient, which describes the radiative excitation from the lower to the upper level

(57)\[\kappa_{l,u,\nu}^{\rm ext} = \frac{h \, \nu}{4 \, \pi} \, n_l \, B_{l,u} \, \phi_{l,u} (\nu),\]

where \(B_{l,u}\) describes the Einstein B-coefficient for extinction.

In addition to spontaneous emission and extinction we have to take the stimulated emission into account, which can be described by adding a negative opacity contribution to Eq. (57):

(58)\[\kappa_{l,u,\nu} = \frac{h \, \nu}{4 \, \pi} \, \left(n_l \, B_{l,u} - n_u \, B_{u,l}\right) \, \phi_{l,u} (\nu),\]

where \(B_{u,l}\) represents the Einstein B-coefficient for stimulated emission. So, for \(n_l B_{l,u} < n_u \, B_{u,l}\) we get laser (maser) emission.

The different Einstein coefficients are related to each other by the Einstein relations:

(59)\[\begin{split}A_{u,l} &= \frac{2 \, h \, \nu^3}{c_{\rm light}^2} \, B_{u,l},\\ g_l \, B_{l,u} &= g_u \, B_{u,l}.\end{split}\]

Following Eq. (39), the differential optical depth \(\tau_{\nu}\) is defined as

(60)\[\begin{split}d \tau_{\nu} = \kappa_{\nu} \, ds &= \left(\frac{h \, \nu}{4 \, \pi} \, \left(n_l \, B_{l,u} - n_u \, B_{u,l}\right) \, \phi_{l,u} (\nu) \right) \, ds \\ &= \left(\frac{c_{\rm light}^2}{8 \, \pi \, \nu^2} \, A_{u,l} \, \left(n_l \, \frac{g_u}{g_l} - n_u \right) \, \phi_{l,u} (\nu) \right) \, ds,\end{split}\]

where we used the Einstein relations Eqn. (59) in the last line. By assuming LTE conditions and therefore Boltzmann population distribution

\[\frac{n_u}{n_l} = \frac{g_u}{g_l} \, e^{(-E_u - E_l)/k_B T_{\rm ex}} = \frac{g_u}{g_l} \, e^{-h \, \nu_{u,l}/k_B T_{\rm ex}}\]

we can rewrite Eq. (60) by using Eq. (55)

(61)\[d \tau_{\nu} = \left(\frac{c_{\rm light}^2}{8 \, \pi \, \nu^2} \, A_{u,l} \, n_u \, \left(e^{h \, \nu_{u,l} / k_B \, T_{\rm ex}} - 1 \right) \, \phi_{l,u} (\nu) \right) \, ds.\]

Finally, we have to integrate along the line of sight and obtain the optical depth \(\tau_{\nu}\)

(62)\[\tau_{\nu} = \frac{c_{\rm light}^2}{8 \, \pi \, \nu^2} \, A_{u,l} \, N_u \, \left(e^{h \, \nu_{u,l} / k_B \, T} - 1 \right) \, \phi_{l,u} (\nu),\]

where \(N_u = \int n_u \, ds\) describes the column density of a certain molecule in the upper state.

In order to express Eq. (62) in terms of the total column density \(N_{\rm tot} = \sum_{i=0}^\infty n_i\) we start again with the Boltzmann population distribution

(63)\[N_{\rm tot} = \sum_{i=0}^\infty n_i = n_j \, \sum_{i=0}^\infty \frac{n_i}{n_j} = n_j \, \sum_{i=0}^\infty \frac{g_i}{g_j} \, e^{(-E_i+E_j)/k_B T_{\rm ex}} = \frac{n_j}{g_j} \, e^{E_j/k_B T_{\rm ex}} \cdot Q \left(T_{\rm ex} \right),\]

where we used the partition function \(Q(T_{\rm ex})\), which is defined as sum over all states

(64)\[Q(T_{\rm ex}) = \sum_{i=0}^\infty g_i \, e^{-E_i/k_B \, T_{\rm ex}}.\]

For our purpose we rewrite Eq. (63) in terms of \(N_u\) and \(N_l\):

(65)\[N_{\rm tot} = \frac{N_u}{g_u} \, e^{E_u/k_B T_{\rm ex}} \cdot Q \left(T_{\rm ex} \right).\]

Inserting this expression into Eq. (62) gives

(66)\[\begin{split}\tau_{\nu} &= \frac{c_{\rm light}^2}{8 \, \pi \, \nu^2} \, A_{u,l} \, \left[\frac{N_{\rm tot} \, g_u}{Q \left(T_{\rm ex} \right)} \, e^{-E_u/k_B T_{\rm ex}} \right] \, \left(e^{h \, \nu_{u,l} / k_B \, T} - 1 \right) \, \phi_{l,u} (\nu) \\ &= \frac{c_{\rm light}^2}{8 \, \pi \, \nu^2} \, A_{u,l} \, N_{\rm tot} \, \frac{g_u \, e^{-E_l/k_B T_{\rm ex}}}{Q \left(T_{\rm ex} \right)} \, \left(1 - e^{-h \, \nu_{u,l} / k_B \, T} \right) \, \phi_{l,u} (\nu),\end{split}\]

where we used Eq. (55) to achieve Eq. (7) for a single line of a certain component and molecule.

Radio recombination lines

In order to get the linear absorption coefficient we consider a system involving radiative transitions between to levels indicated by the quantum numbers \(n_1\) and \(n_2\), respectively. The corresponding level energies \(E_{n_1}\) and \(E_{n_2}\) are related by the transition frequency \(\nu_{n_1, n_2}\) with

\[h \, \nu_{n_1, n_2} = E_{n_2} - E_{n_1},\]

where we assumed that \(E_{n_1} < E_{n_2}\). Now, we can write the extinction coefficient, which describes the radiative excitation from the lower (\(n_1\)) to the upper (\(n_2\)) level

(67)\[\kappa_{n_1, n_2, \nu}^{\rm ext} = \frac{h \, \nu_{n_1, n_2}}{4 \pi} \, N_{n_1} \, B_{n_1, n_2} \, \phi_{n_1, n_2} (\nu),\]

where \(B_{n_1, n_2}\) represents the Einstein B coefficient for extinction and \(\phi_{n_1, n_2} (\nu)\) describes the normalized line profile (in units of Hz\(^{-1}\)) of the transition of photons of frequency \(\nu_{n_1, n_2}\).

In addition we take the stimulated emission into account, which can be described by adding a negative opacity contribution to (67):

(68)\[\kappa_{n_1, n_2, \nu}^{\rm ext} = \frac{h \, \nu_{n_1, n_2}}{4 \pi} \, \left(N_{n_1} \, B_{n_1, n_2} - N_{n_2} \, B_{n_2, n_1}\right) \, \phi_{n_1, n_2} (\nu),\]

where \(B_{n_2, n_1}\) indicates the Einstein B coefficient for stimulated emission.

Now we use the Boltzmann equation

\[\frac{N_{n_1}}{N_{n_2}} = \frac{g_{n_2}}{g_{n_1}} \, \exp\left[-\frac{h \, \nu_{n_1, n_2}}{k_B \, T_e} \right]\]

and the well-known Einstein relation

\[g_{n_1} \, B_{n_1, n_2} = g_{n_2} \, B_{n_2, n_1}\]

to get

(69)\[\kappa_{n_1, n_2, \nu}^{\rm ext} = \frac{h \, \nu_{n_1, n_2}}{4 \pi} \, N_{n_1} \, B_{n_1, n_2} \left(1 - \exp\left[-\frac{h \, \nu_{n_1, n_2}}{k_B \, T_e} \right] \right) \, \phi_{n_1, n_2} (\nu),\]

where the term in squared brackets describes the correction of the stimulated emission in thermodynamic equilibrium.

Now we can rewrite the stimulated transition coefficient in terms of the oscillator strength

(70)\[\begin{split}f_{n_1, n_2} &= -\frac{g_{n_2}}{g_{n_1}} f_{n_2, n_1} \\ &= \frac{m_e \, c \, h \, \nu_{n_1, n_2}}{4 \pi^2 \, e^2} B_{n_1, n_2} \\ &\approx n_1 \, M_{\Delta n} \left(1 + 1.5 \frac{\Delta n}{n_1} \right).\end{split}\]

In the last step we applied Menzel’s approximation [Menzel_1968] for the absorption oscillator strength with \(M_{\Delta n}\) = 0.190775, 0.026332, 0.0081056, 0.0034917, 0.0018119, and 0.0010585 for \(\Delta n = n_2 - n_1\) = 1, 2, 3, 4, 5, and 6.

Finally, we use the Saha-Boltzmann equation to relate the number density of atoms in the lower level \(n_1\) to the population of the electrons and ions (\(N_e\) and \(N_i\)) of the unbound states of hydrogenic atoms

(71)\[N_{n_1} = \frac{N_e \, N_i}{T_e^{3/2}} \, \frac{n_1^2 \, h^3}{(2 \pi \, m_e \,k_B)^{3/2}} \, \exp \left[\frac{Z^2 \, E_{n_1}}{k_B \, T_e} \right],\]

where the energy of the lower state is given by

(72)\[E_{n_1} = -\frac{R_M}{n_1^2}.\]

Note, \(E_{n_1}\) is the energy of level \(n_1\) below the continuum. The level energy is negative, and the energy of level \(n_1\) below the continuum is then positive.

In Eq. (72), the Rydberg constant \(R_M\) is defined as

\[R_M = \frac{2 \pi^2 \, m_e \, e^4}{h^3 \, c} \cdot \left[ 1 + \frac{m_e}{M}\right]^{-1},\]

where \(M\) represents the mass of the atom. The database contains entries for Hydrogen, Helium, Carbon, Nitrogen, Oxygen, and Sulfur.

The transition frequency \(\nu\) is given as

\[\nu_{n, n + \Delta n} = R_M \cdot c \cdot \left[\frac{1}{n^2} - \frac{1}{(n + \Delta n)^2} \right],\]

where \(\Delta n\) describes the RRL, i.e. \(\Delta n\) = 1 indicates \(\alpha\)-, \(\Delta n\) = 2 \(\beta\), \(\Delta n\) = 3 \(\gamma\), and \(\Delta n\) = 4 \(\delta\)-transitions, respectively.

Substituting (71) and (70) into (69), gives

\[\kappa_{n_1, n_2, \nu}^{\rm ext} = \frac{\pi \, h^3 \, e^2}{(2 \pi \, m_e \, k_B)^{3/2} \, m_e \, c} \, n_1^2 \, f_{n_1, n_2} \, \phi_\nu \, \frac{N_e \, N_i}{T_e^{3/2}} \exp \left[\frac{Z^2 \, E_{n_1}}{k_B T_e} \right] \left(1 - e^{-h \nu_{n_1, n_2} / k_B T_e} \right).\]

In order to get the optical depth we have to integrate the linear absorption coefficient along the line of sight and get

(73)\[\begin{split}\tau_\nu &= \int \kappa_{n_1, n_2, \nu}^{\rm ext} \, ds \\ &= \frac{\pi \, h^3 \, e^2}{(2 \pi \, m_e \, k_B)^{3/2} \, m_e \, c} \cdot {\rm EM} \cdot \frac{n_1^2 \, f_{n_1, n_2}}{T_e^{3/2}} \, \exp \left[\frac{Z^2 \, E_{n_1}}{k_B T_e} \right] \, \left(1 - e^{-h \nu_{n_1, n_2} / k_B T_e} \right) \, \phi_\nu,\end{split}\]

where the emission measure \(EM\) is defined as

(74)\[EM \equiv \int N_e \, N_i \, ds.\]

The terms \(n_1^2 \, f_{n_1, n_2}\), \(E_{n_1}\), and \(\nu_{n_1, n_2}\) are taken from the embedded database whereas the parameters EM (in cm\(^{-6}\) pc), \(T_e\) (in K), \(\Delta v\) (in km/s), and \(v\) (in km/s) used for the line profile function \(\phi_\nu\) are taken from the user-defined molfit file.

In order to derive Eq. (25) we start with the Boltzmann distribution describing the relative population of two quantum levels in terms of volume densities:

(75)\[\frac{N_{n_2}}{N_{n_1}} = \frac{g_{n_2}}{g_{n_1}} \, e^{-h \nu / k_B T},\]

where \(g_{n_i}\) indicates the statistical weight of the level \(i\). Here we used a single temperature \(T\) to characterize the statistical state of bound and unbound levels of the atoms.

For the Non-LTE case Eq. (75) has to be extended by introducing two correction factors which describe differences between the LTE and the Non-LTE behavior.

(76)\[\frac{N_{n_2}}{N_{n_1}} = \frac{b_{n_2}}{b_{n_1}} \, \frac{g_{n_2}}{g_{n_1}} \, e^{-h \nu / k_B T} = \frac{g_{n_2}}{g_{n_1}} \, e^{-h \nu / k_B T_{\rm ex}},\]

where we introduced a new excitation temperature \(T_{\rm ex}\). Now, we start with the expression for the line absorption coefficient Eq. (69), and multiply Eq. (69) by \(b_{n_1}\) to correct for the population available to absorb photons and insert a second factor, involving the ratio \(b_{n_2} / b_{n_1}\), to correct for the Non-LTE amount of stimulated emission:

(77)\[\kappa_{n_1, n_2, \nu}^{\rm ext, Non-LTE} = \kappa_{n_1, n_2, \nu}^{\rm ext, LTE} \, b_{n_1} \, \beta,\]

where the factor \(\beta\) is given by

\[\beta = \left[\frac{1 - \left(b_{n_2} / b_{n_1}\right) \, e^{-h \nu / k_B T}}{1 - e^{-h \nu / k_B T}} \right].\]

Integrating Eq. (77) along the line of sight gives Eq. (23).

The Non-LTE emission coefficient is by definition

\[j_{n_1, n_2, \nu}^{\rm Non-LTE} = \kappa_{n_1, n_2, \nu}^{\rm ext, LTE} \, b_{n_2} \, B_\nu(T),\]

which leads to the expression of the source function for RRLs in Non-LTE (25)

\[S_{{\rm RRL}, \nu}^{\rm Non-LTE} = \frac{j_{n_1, n_2, \nu}^{\rm Non-LTE}}{\kappa_{n_1, n_2, \nu}^{\rm ext, Non-LTE}} = \frac{\tau_{{\rm RRL}, \nu}^{\rm LTE} \cdot b_{n_2} \cdot J(T_e)}{\tau_{{\rm RRL}, \nu}^{\rm LTE} \cdot b_{n_1} \cdot \beta}.\]

Free-free continuum

The cross section for the emission of one photon is

(78)\[d\sigma_{\rm ff} = \frac{16 \, \pi}{3 \sqrt{3}} \alpha^3 \, Z_1^2 \, Z_2^2 \, \left(\frac{\hbar}{m \, v}\right)^2 \, g\left(\eta_i, \eta_f \right) \frac{d\nu}{\nu},\]

where \(\alpha\) describes the fine structure constant, \(Z_1\) the charge of the hydrogenic ion, \(Z_2\) the charge, \(m\) the mass and \(v\) the velocity of the electron. Additionally, the factor \(g\left(\eta_i, \eta_f \right)\) indicates the Gaunt factor [Gaunt_1930] as the ratio of the quantum mechanical cross section to the semiclassical result of Kramers and Wentzel. The Gaunt factor was derived by [Grant_1958] using Coulomb wave functions for initial and final states of free electron in the form

(79)\[g\left(\eta_i, \eta_f \right) = \frac{\eta_i \, \eta_f}{\left|\eta_i - \eta_f\right|} \, \frac{\sqrt{3} \, \pi \, \left|\Delta \left(i\eta_i, i\eta_f\right)\right|}{\left(\exp\left[2 \pi \eta_f \right] -1\right) \, \left(1 - \exp\left[-2 \pi \eta_i \right] \right)}\]

with the complex function

\[\Delta \left(i\eta_i, i\eta_f\right) = \left(F(1 - i \eta_i, -i \eta_f, 1, z)\right)^2 - \left(F(1 - i \eta_f, -i \eta_i, 1, z)\right)^2\]

expressed by the hypergeometric functions \(F\). The parameter \(\eta_i\) and \(\eta_f\) are related to the electron velocities by

\[\begin{split}\eta_{f,i} &= \alpha \, Z_1 \, Z_2 \, \frac{c}{v_{f,i}}, \\ z &= -4 \frac{\eta_i \, \eta_f}{\xi^2}, \\ \xi &= \eta_f - \eta_i,\end{split}\]

where the indices \(i,f\) refer to the initial \(i\) and final \(f\) state of the electron. In addition, Eq. (79) assumes a point-like scattering object why it is strictly correct only for fully ionized hydrogenic ions like H\(^+\), He\(^{++}\), …In many cases the electron can be assumed to obey a thermal distribution of temperature \(T = T_e\) and the free-free emission of a dilute thermal plasma is obtained from the Maxwellian average of the Gaunt factor by

(80)\[\langle g_{\rm ff} \rangle = \int_0^\infty g\left(\eta_i, \eta_f \right) e^{-y} dy\]

with the dimensionless kinetic energy of the electrons after scattering scaled to the mean electron energy \(\langle E \rangle\)

\[y = \left(m \, v_i^2 / 2 - h \nu\right) / \langle E \rangle.\]

The spectral emissivity depends on the frequency through the Gaunt factor times the exponential \(\exp\left[-h \nu / k_B T\right]\) which follows since only electrons with initial energies in excess of \(h \nu\) contribute to the emissivity. The free-free emissivity is given by

(81)\[\epsilon_\nu = n_e \, n_i \, \int_{v_0}^\infty \left(\frac{{\rm d} \sigma_{\rm ff}}{{\rm d} \nu}\right) \frac{h \nu}{4 \pi} \, v_i f(v_i) dv_i,\]

where \(v_i f(v_i)\) describes the initial (thermal) velocity distribution.

Combining Eqn. (78), (80), and (81) gives

(82)\[\epsilon_\nu = n_e \, n_i \, \frac{8}{3} \, \left(\frac{2 \pi}{3} \right)^{(1/2)} \, \frac{e^6 \, Z_i^2}{\left(m \, c^2\right)^{3/2}} \, \frac{1}{\sqrt{k_B T}} \, e^{-\frac{h \nu}{k_B T}} \, \langle g_{\rm ff} \rangle.\]

Integrating Eq. (82) along the line of sight one gets the optical depth for thermal electrons in terms of the classical electron radius \(r_e = (\alpha \, \hbar \, c) / (m_e \, c^2) = \alpha^2 \, a_0\)

\[\begin{split}\tau_{\rm ff} &= \frac{4}{3} \, \left(\frac{2 \pi}{3} \right)^{(1/2)} \, r_e^3 \, \frac{Z_i^2 m^{3/2} \, c^5}{\sqrt{k_B \, T} \, h \, \nu^3} \, \left(1 - e^{-\frac{h \nu}{k_B T}} \right) \, \langle g_{\rm ff} \rangle \cdot {\rm EM} \\ &= 1.13725 \cdot \left(1 - e^{-\frac{h \nu}{k_B T}} \right) \cdot \langle g_{\rm ff} \rangle \cdot \left[ \frac{T}{\rm K}\right]^{-\frac{1}{2}} \, \left[ \frac{\nu}{\rm GHz}\right]^{-3} \, \left[ \frac{\rm EM}{{\rm pc} \, {\rm cm}^{-6}}\right].\end{split}\]

Synchrotron continuum

We start with

\[\begin{split}J(\nu) &= \frac{\sqrt{3} \, e^3 \, B \, \kappa}{4 \, \pi \, \epsilon_0 \, c \, m_e} \, \left(\frac{3 \, e \, B}{2 \, \pi \, \nu \, m_e^3 \, c^4} \right)^{(p - 1)/2} \, a(p) \\ &= 2.344 \times 10^{-25} \, a(p) \, B^{(p + 1) / 2} \, \kappa' \, \left(3.217 \times 10^{17}\right)^{(p - 1)/2} \, \text{W} \, \text{m}^{-3} \, \text{Hz}^{-1},\end{split}\]

with

\[a(p) = \frac{\sqrt{\pi}}{2} \, \frac{\Gamma \left(\frac{p}{4} + \frac{19}{12}\right) \, \Gamma \left(\frac{p}{4} - \frac{1}{12}\right) \, \Gamma \left(\frac{p}{4} + \frac{5}{4}\right)}{(p + 1) \, \Gamma \left(\frac{p}{4} + \frac{7}{4}\right)}.\]

Here, the energy of the electrons in measured in GeV and \(N(E)\) is expressed in electron m\(^{-3}\) GeV\(^{-1}\).

The absorption coefficient \(\chi_\nu\) for a random magnetic field is

\[\begin{split}\chi_\nu &= \frac{\sqrt{3} \, e^3 \,c}{8 \, \pi^2 \, \epsilon_0 \, m_e} \, \kappa \, B^{(p + 2) / 2} \, \left(\frac{3 \, e}{2 \, \pi \, m_e^3 \, c^4} \right)^{p/2} \, b(p) \, \nu^{-(p + 4) / 2} \\ &= 20.9 \, \kappa' \, B^{(p + 2) / 2} \, \left(5.67 \times 10^{9}\right)^p \, b(p) \, \nu^{-(p + 4) / 2} \, \text{m}^{-1},\end{split}\]

where

\[b(p) = \frac{\sqrt{\pi}}{8} \, \frac{\Gamma \left(\frac{3p + 22}{12}\right) \, \Gamma \left(\frac{3p + 2}{12}\right) \, \Gamma \left(\frac{p + 6}{4}\right)}{\Gamma \left(\frac{p + 8}{4}\right)}.\]

Here, \(N(E)\) is expressed in electron m\(^{-3}\) GeV\(^{-1}\).

The source function is

\[S_{\nu,{\rm sync}} = \frac{J(\nu)}{4 \pi \chi_\nu}.\]

Elliptical rotated Gaussians

An arbitrary rotated elliptical 2D-Gaussian can be written as

(83)\[f(x,y) = \frac{1}{V} \cdot A \cdot \exp \left[ -\left( a(x - x_0)^2 + 2 b (x - x_0) (y - y_0) + c (y - y_0)^2 \right) \right],\]

where the coefficient \(A\) describes the amplitude, \(x_0\), \(y_0\) the center and \(\sigma_x\), \(\sigma_y\) are the \(x\) and \(y\) spreads of the blob. Additionally, \(V\) indicates the volume of the Gaussian and is given by

(84)\[V = \int_{-\infty}^{\infty} \int_{-\infty}^{\infty} f(x,y) \, dx \, dy = 2 \pi A \sigma_{x} \sigma_{y}.\]

The coefficients \(a\), \(b\), and \(c\) are

\[\begin{split}a &= {\frac {\cos^{2} \theta} {2 \sigma_{x}^{2}}} + {\frac{\sin^{2} \theta} {2 \sigma_{y}^{2}}}, \\ b & = -{\frac{\sin 2\theta } {4 \sigma_{x}^{2}}} + {\frac{\sin 2 \theta } {4 \sigma_{y}^{2}}}, \\ c &= {\frac{\sin^{2} \theta }{2 \sigma_{x}^{2}}} + {\frac{\cos^{2} \theta }{2\sigma_{y}^{2}}} .\end{split}\]

where the blob is rotated by a clockwise angle \(\theta\).

Footnotes

References

[Boland_De-Jong_1982]

Boland, W. & De~Jong, T. 1982, The Astrophysical Journal, 261, 110

[Bresenham_1977]

Bresenham, J. 1977, Commun. ACM, 20, 100–106

[Brocklehurst_Seaton_1972]

Brocklehurst, M. & Seaton, M. 1972, Monthly Notices of the Royal Astronomical Society, 157, 179

[Brussaard_Van_de_Hulst_1962]

Brussaard, P. & Van~de Hulst, H. 1962, Reviews of Modern Physics, 34, 507

[Castor_1970]

Castor, J.~I. 1970, Monthly Notices of the Royal Astronomical Society, 149, 111

[Cesaroni_Walmsley_1991]

Cesaroni, R. & Walmsley, C. 1991, Astronomy and Astrophysics, 241, 537

[Downes_Eckart_2007]

Downes, D. & Eckart, A. 2007, Astronomy & Astrophysics, 468, L57

[Elitzur_1992] (1,2)

Elitzur, M. 1992, Dordrecht, Netherlands: Kluwer Academic Publishers Astrophysics and Space Science Library., 170

[Gaunt_1930]

Gaunt, J.~A. 1930, Philosophical Transactions of the Royal Society of London. Series A, Containing Papers of a Mathematical or Physical Character, 229, 163

[Gee_et_al_1976]

Gee, C., Percival, L., Lodge, J., & Richards, D. 1976, Monthly Notices of the Royal Astronomical Society, 175, 209

[Grant_1958]

Grant, I. 1958, Monthly Notices of the Royal Astronomical Society, 118, 241

[Hildebrand_1983]

Hildebrand, R. 1983, Molster, FJ, Dorschner, J., Henning, T., & Mutschke, H.

[Karzas_Latter_1961]

Karzas, W. & Latter, R. 1961, Astrophysical Journal Supplement, vol. 6, p. 167, 6, 167

[Lines_2002]

Lines, R.~R. 2002, Their Physics and Astronomical Applications ed MA Gordon and RL Sorochenko (Astrophysics and Space Science Library, Vol. 282); Dordrecht

[Menzel_1968]

Menzel, D.~H. 1968, Nature, 218, 756

[Mihalas_1978]

Mihalas, D. 1978, San Francisco: WH Freeman

[Ossenkopf_Henning_1994]

Ossenkopf, V. & Henning, T. 1994, Astronomy and Astrophysics, 291, 943

[Padua_2011]

Padua, D. 2011, Encyclopedia of parallel computing (Springer Science & Business Media)

[Poojon_2024] (1,2)

Poojon, P., Chung, A., Hoang, T., et al. 2024, ApJ, 963, 88

[Potter_et_al_2004]

Potter, C., Habing, H., & Olofsson, H. 2004, Journal of the British Astronomical Association, 114, 168

[Sobolev_1960]

Sobolev, V.~V. 1960, Moving envelopes of stars (Harvard University Press)

[Stahler_Palla_2005]

Stahler, S. & Palla, F. 2005, Stahler, SW & Palla, F, 3, 9

[Stevenson_2014]

Stevenson, M.~A. 2014, ApJ, 781, 113.

[Storey_Hummer_1995] (1,2,3)

Storey, P. & Hummer, D. 1995, Monthly Notices of the Royal Astronomical Society, 272, 41

[Strelnitski_et_al_1996]

Strelnitski, V.~S., Ponomarev, V.~O., & Smith, H.~A. 1996, Astrophysical Journal, 470, 1118

[Thompson_et_al_1987]

Thompson, P., Cox, D., & Hastings, J. 1987, Journal of Applied Crystallography, 20, 79

[van_der_Tak_et_al_2007] (1,2,3)

van der Tak, F.~F.~S., Black, J.~H., Schöier, F.~L., {Jansen}, D.~J., & {van Dishoeck}, E.~F. 2007, Astronomy and Astrophysics, 468, 627

[Van_Hoof_et_al_2015] (1,2)

Van~Hoof, P., Ferland, G.~J., Williams, R., et~al. 2015, Monthly Notices of the Royal Astronomical Society, 449, 2112

[Walmsley_1990]

Walmsley, C.~M. 1990, Astronomy and Astrophysics, Supplement, 82, 201