The quantum language of astrophysical radiation mechanisms

Chapter opening

The same mean spectrum can come from different optical fields. Thermal radiation, free-free emission, synchrotron radiation, inverse Compton scattering, masers, and curvature radiation can all be described by \(I_\nu\), but they do not have the same coherence time, polarization, brightness temperature, photon bunching, or cross-frequency correlations. Once a radiation mechanism is written as a photon-statistics problem, spectral indices and light curves extend naturally to \(g^{(2)}(\tau)\), polarization-resolved correlations, joint probabilities between frequency channels, and event-arrival time distributions [Brown and Twiss, 1956, Glauber, 1963, Glauber, 1963, Guerin et al., 2017, Mandel and Wolf, 1995].

From mean spectra to photon statistics

The usual starting point for a radiation mechanism is the specific intensity \(I_\nu\), the energy flux per unit time, area, frequency, and solid angle. If the source subtends solid angle \(\Omega_{\rm s}\), the observed flux density is approximately \(S_\nu\simeq I_\nu\Omega_{\rm s}\). In radio and far-infrared astronomy, the Rayleigh–Jeans form defines the brightness temperature,

(149)\[T_b=\frac{c^2 I_\nu}{2k_B\nu^2} =\frac{c^2 S_\nu}{2k_B\nu^2\Omega_{\rm s}} .\]

Here \(c\) is the speed of light, \(k_B\) is Boltzmann’s constant, and \(\nu\) is frequency. \(T_b\) is measured in K. It is the blackbody temperature that would produce the same \(I_\nu\) in the Rayleigh–Jeans limit. For nonthermal particles, masers, or fast radio bursts, it is not the material temperature. When \(\Omega_{\rm s}\) is inferred from an angular diameter, a smaller source implies a larger \(T_b\). Variability time scales, VLBI angular sizes, and scattering broadening therefore enter directly into the interpretation of the emission mechanism.

Brightness temperature connects mean intensity to source size. Dulk’s review of solar and stellar radio emission gives a useful scale: incoherent radio emission usually struggles to exceed \(10^9\) to \(10^{10}~\mathrm{K}\). Bursts well above that range need a coherent mechanism, such as plasma radiation or a cyclotron maser [Dulk, 1985]. FRBs are more extreme. A millisecond, Jy-level radio pulse at Gpc distance gives

(150)\[T_b\simeq 1.1\times10^{35}~{\rm K} \left(\frac{S_\nu}{1~{\rm Jy}}\right) \left(\frac{D_A}{1~{\rm Gpc}}\right)^2 \left(\frac{t_{\rm FRB}}{1~{\rm ms}}\right)^{-2} \left(\frac{\nu}{1~{\rm GHz}}\right)^{-2},\]

where \(D_A\) is the angular-diameter distance and \(t_{\rm FRB}\) is the pulse width. This scale is far above the limit for incoherent particle radiation. It requires many charges to radiate collectively on scales smaller than a wavelength, or at least with controlled phase over the emitting patch [Lorimer et al., 2007, Lu and Kumar, 2018, Petroff et al., 2019, Petroff et al., 2022].

Figure 50 Brightness temperature as a diagnostic of radiation mechanism. Stellar photospheres and HII regions occupy ordinary thermal temperature scales. Solar and stellar radio bursts above about 10¹⁰ K usually require coherent emission. Pulsar giant pulses and FRBs have still higher brightness temperatures, which can be explained only by coherent radiation or strong beaming geometry. The dashed line marks a common order-of-magnitude limit for incoherent radio emission.

Photon statistics enter beyond the mean spectrum through first- and second-order coherence functions. If \(E^{(+)}(t)\) is the positive-frequency electric-field operator, the normalized first-order coherence function describes phase and spectral coherence,

(151)\[g^{(1)}(\tau)= \frac{\langle E^{(-)}(t)E^{(+)}(t+\tau)\rangle} {\langle E^{(-)}(t)E^{(+)}(t)\rangle},\]

while the second-order coherence function describes intensity fluctuations and photon-arrival correlations,

(152)\[g^{(2)}(\tau)= \frac{\langle I(t)I(t+\tau)\rangle}{\langle I(t)\rangle^2}.\]

Chaotic thermal light in one spatio-temporal polarization mode has \(g^{(2)}(0)=2\). A coherent state has \(g^{(2)}(0)=1\). Real astronomical measurements usually average over many temporal, spectral, spatial, and polarization modes. The thermal bunching peak is then diluted to

(153)\[g^{(2)}_{\rm meas}(0)-1\simeq \frac{1}{M_{\rm eff}}, \qquad M_{\rm eff}\simeq M_{\rm sp}M_{\rm pol} \left(1+\frac{\Delta t}{\tau_c}\right).\]

\(M_{\rm sp}\) is the number of mixed spatial modes, \(M_{\rm pol}\) the number of polarization modes, \(\Delta t\) the electronic or software time bin, and \(\tau_c\sim1/\Delta\nu\) the coherence time set by the optical bandwidth. For \(\Delta\nu=1~\mathrm{nm}\) at \(\lambda=500~\mathrm{nm}\), \(\tau_c\sim0.8~\mathrm{ps}\). With a \(100~\mathrm{ps}\) electronic resolution, even a single-mode thermal bunching peak is reduced by more than two orders of magnitude. Narrow-band filtering, polarization selection, and spatial-mode control increase the visible contrast, but they also reduce the photon rate.

Figure 51 Representative g⁽²⁾(τ) functions for different optical fields. Single-mode thermal light gives a bunching peak of height 2 at zero delay. Multimode averaging reduces the peak. An ideal coherent state stays near 1. A narrow coherent line or stimulated component can add small oscillations or narrow peaks on top of a nearly Poisson background. The horizontal axis is normalized by the coherence time τ_c.

Figure 52 Decline of thermal bunching contrast with effective mode number. The horizontal axis is the ratio of time bin to coherence time, while M₀ denotes the number of spatial and polarization modes entering the same correlation estimator. Even if the source itself is thermal, a wide time bin, mixed polarizations, or multiple spatial modes can push g⁽²⁾(0) − 1 close to zero.

Thermal radiation, free-free emission, and radiative transfer

A blackbody field in local thermodynamic equilibrium is set by the mean occupation number

(154)\[\bar{n}_\nu= \frac{1}{\exp(h\nu/k_BT)-1}\]

where \(h\nu\) is the photon energy and \(T\) is the material temperature. Photospheres, dust, and the cosmic microwave background are close to this type of thermal state. The visible light from a stellar photosphere is not always in the Rayleigh–Jeans limit, so color temperature, effective temperature, and brightness temperature should not be used interchangeably. Thermal light is a random superposition of many independent emitters, and the electric field is well approximated as a complex Gaussian process. Under those conditions the Siegert relation gives \(g^{(2)}(\tau)=1+|g^{(1)}(\tau)|^2/M_{\rm eff}\).

Free-free emission comes from electrons accelerated in the Coulomb fields of ions. In a dilute ionized gas, the emissivity has the scaling

(155)\[j_\nu^{\rm ff}\propto n_e n_i T_e^{-1/2} \exp\!\left(-\frac{h\nu}{k_BT_e}\right) g_{\rm ff}(\nu,T_e),\]

where \(n_e\) and \(n_i\) are the electron and ion number densities, in \(\mathrm{cm^{-3}}\) or \(\mathrm{m^{-3}}\); \(T_e\) is the electron temperature; and \(g_{\rm ff}\) is the Gaunt factor, usually from order unity to tens. HII regions have \(T_e\sim10^4~\mathrm{K}\), while stellar winds and accretion flows can be hotter. After many independent collisions are added, the narrow-band field remains close to chaotic thermal light. If the observing time is much longer than \(\tau_c\), the visible bunching is diluted by Eq. (153).

Absorption and emission belong in the radiative-transfer equation,

(156)\[\frac{{\rm d}I_\nu}{{\rm d}s}=j_\nu-\alpha_\nu I_\nu, \qquad S_\nu=\frac{j_\nu}{\alpha_\nu}.\]

\(s\) is the line-of-sight distance, \(j_\nu\) is the emissivity, \(\alpha_\nu\) is the absorption coefficient, and \(S_\nu\) is the source function. In brightness-temperature form, with \({\rm d}\tau_\nu=\alpha_\nu{\rm d}s\), a uniform slab gives

(157)\[T_b=T_{\rm eff}\left(1-e^{-\tau_\nu}\right) +T_{\rm bg}e^{-\tau_\nu}.\]

\(T_{\rm eff}\) is the effective temperature corresponding to the source function, and \(T_{\rm bg}\) is the background brightness temperature. For small optical depth, \(T_b\simeq T_{\rm eff}\tau_\nu+T_{\rm bg}\). For large optical depth, \(T_b\to T_{\rm eff}\). This relation is often the quickest way to decide whether a source is thermalized, semi-transparent, or dominated by nonthermal emission [Draine, 2011, Dulk, 1985, Rybicki and Lightman, 1979].

Synchrotron radiation and inverse Compton scattering

Synchrotron radiation is emitted by relativistic electrons spiraling in a magnetic field. The characteristic frequency of one electron is

(158)\[\nu_c=\frac{3eB\sin\alpha}{4\pi m_ec}\gamma^2 ,\]

where \(e\) is the elementary charge, \(B\) is the magnetic-field strength, \(\alpha\) is the pitch angle, \(m_e\) is the electron mass, and \(\gamma\) is the Lorentz factor. For \(B=1~\mathrm{G}\), \(\gamma=10^3\), and \(\sin\alpha=1\), the characteristic frequency is \(\nu_c\simeq4.2\times10^{12}~\mathrm{Hz}\), in the far infrared. In a \(10^{-5}~\mathrm{G}\) interstellar or jet-scale magnetic field, the same electrons radiate mainly in the radio band. The single-electron spectrum is broad, and the emissivity of a real source is

(159)\[j_\nu^{\rm syn}=\int P_\nu(E,B,\alpha)\,N(E)\,{\rm d}E .\]

If the electron spectrum is \(N(E)\propto E^{-p}\), optically thin synchrotron emission gives \(I_\nu\propto\nu^{-\alpha_{\rm syn}}\), with \(\alpha_{\rm syn}=(p-1)/2\). The maximum linear polarization in a uniform magnetic field is approximately

(160)\[\Pi_{\rm syn}=\frac{p+1}{p+7/3}.\]

For \(p=2.4\), \(\Pi_{\rm syn}\simeq72\%\). Observed jets, supernova remnants, and pulsar wind nebulae often have lower polarization because magnetic-field directions, Faraday rotation, beam averaging, and multiple emission zones partly cancel the signal. When many independent electrons and turbulent cells are added, the electric field can again approach a Gaussian chaotic field. Unlike thermal light, its mean spectrum and polarization are set by the nonthermal particle distribution and magnetic geometry [Blandford and Rees, 1978, Blumenthal and Gould, 1970, Ginzburg and Syrovatskii, 1965, Zhang et al., 2012].

Inverse Compton scattering boosts low-energy photons to higher energy. In the Thomson limit, the scattered photon energy is of order

(161)\[h\nu_{\rm IC}\sim \gamma^2 h\nu_{\rm seed}.\]

For repeated scattering in a thermal electron cloud, the Compton \(y\) parameter is

(162)\[y\simeq 4\frac{k_BT_e}{m_ec^2}\max(\tau_T,\tau_T^2),\]

where \(T_e\) is the electron temperature and \(\tau_T\) is the Thomson optical depth. If \(y\ll1\), the spectrum is only weakly modified. If \(y\sim1\), a clear Comptonized continuum appears. If \(y\gg1\), Comptonization approaches saturation. In X-ray binaries, AGN coronae, and GRB photosphere models, the mean spectrum can be a mixture of synchrotron radiation, thermal seed photons, and inverse Compton components. A spectral index alone rarely identifies a unique mechanism [Gierliński et al., 1999, Lazzati et al., 2013, Sunyaev and Titarchuk, 1980].

Figure 53 Schematic normalized spectra for common continuum radiation mechanisms. A blackbody has a thermal peak and a Rayleigh–Jeans low-frequency side. Free-free emission can be nearly flat in the radio and cut off exponentially at high frequency. Synchrotron radiation produces a power law with low- and high-frequency cutoffs set by the electron distribution and opacity. Inverse Compton emission often appears as a high-energy bump. Real sources require joint fitting of these components within a radiative-transfer and geometrical model.

Masers, curvature radiation, and coherent bursts

A maser starts with negative absorption. If a spectral transition has a population inversion, its absorption coefficient can be negative, \(\alpha_\nu<0\). Along a path of length \(L\), the intensity approximately obeys

(163)\[I_\nu(L)\simeq I_\nu(0)\exp(|\tau_m|), \qquad \tau_m=\int_0^L\alpha_\nu\,{\rm d}s<0 .\]

In the unsaturated regime, the gain is exponentially sensitive to optical depth. Once the maser saturates, stimulated emission depletes the inversion, and line width, polarization, and intensity respond to pumping and geometry. Water, OH, and SiO masers can have very high brightness temperatures, but their coherence is usually shaped by many velocity-coherent paths, turbulent clumps, and polarized propagation effects. They should not be treated as simple laboratory single-mode lasers. Work by Goldreich, Keeley, Kwan, and Elitzur set the basic constraints among source size, saturation, line width, and polarization [Elitzur, 1982, Elitzur, 1992, Goldreich and Keeley, 1972, Goldreich et al., 1973, Goldreich and Kwan, 1974].

The electron-cyclotron maser is a coherent radio mechanism in strongly magnetized, low-density plasma. It requires a nonequilibrium electron distribution, such as a loss cone, and plasma conditions that allow the electromagnetic wave to escape. The cyclotron frequency is

(164)\[\nu_B=\frac{eB}{2\pi m_ec} \simeq 2.8~{\rm MHz}\left(\frac{B}{1~{\rm G}}\right).\]

If a GHz burst is fundamental or low-harmonic electron-cyclotron maser emission, the magnetic field is hundreds of gauss to several kilogauss. Highly circularly polarized radio bursts from the Sun, planets, brown dwarfs, and low-mass stars can all fit this framework [Dulk, 1985, Hallinan et al., 2008, Melrose and Dulk, 1982, Treumann, 2006].

Curvature radiation is emitted by relativistic charges moving along curved magnetic-field lines. The characteristic frequency and power of one charge are

(165)\[\nu_{\rm curv}=\frac{3c}{4\pi\rho}\gamma^3, \qquad P_{\rm curv}=\frac{2e^2c}{3\rho^2}\gamma^4 ,\]

where \(\rho\) is the curvature radius of the field line. For \(\rho=10^7~\mathrm{cm}\) and \(\gamma=300\), \(\nu_{\rm curv}\sim2\times10^9~\mathrm{Hz}\). Curvature radiation from one particle is still far too weak for pulsar radio bursts or FRBs. Coherent curvature radiation requires many charges to bunch on longitudinal scales smaller than a wavelength, so the field amplitudes add and the power can scale approximately as the square of the particle number. This idea has long been used for coherent pulsar radio emission and is also used in FRB models [Cheng and Ruderman, 1977, Lu and Kumar, 2018].

FRB observations push these mechanisms to an extreme. The Lorimer burst provided the first evidence for millisecond duration, cold-plasma dispersion, and \(T_b\sim10^{34}~\mathrm{K}\). Repeating sources and CHIME samples show that repeat activity, polarization, frequency drift, microstructure, and host environment all matter. The 2020 event from SGR 1935+2154 directly connected a Galactic magnetar to an FRB-like radio burst [Bochenek et al., 2020, CHIME/FRB Collaboration et al., 2020, Lorimer et al., 2007, Petroff et al., 2019, Petroff et al., 2022]. Many models fall into two broad families: maser-like mechanisms driven by a population inversion or shock, and antenna or curvature-like mechanisms driven by charge bunches or current sheets. Sub-microsecond structure, strong polarization, frequency drift, and multiwavelength nondetections jointly constrain the emission radius, Lorentz factor, plasma density, and radiative efficiency.

Optical natural lasers are another form of stimulated emission. In the Weigelt blobs of Eta Carinae, Fe II and O I lines can show laser or laser-like amplification under particular pumping and optical-depth conditions. Line width, spatial location, polarization, and photon correlations can test the inversion and feedback geometry [Dravins and Germanà, 2008, Johansson and Letokhov, 2004, Johansson and Letokhov, 2005]. Stimulated emission is not restricted to radio masers. If the pumping, level structure, and escape paths are favorable, optical lines can also carry statistics that are not those of ordinary thermal light.

Joint diagnostics

In real classification, the mean spectrum gives only the first layer of information. Distinguishing thermal sources, synchrotron sources, masers, and coherent bursts requires spectra, polarization, correlation functions, time scales, and angular scales in one observing vector:

(166)\[{\bf d}= \left\{ I_\nu(t),\, Q_\nu(t),U_\nu(t),V_\nu(t),\, g^{(2)}_{ab}(\tau;\nu_i,\nu_j),\, \Delta t_{\rm var},\, \Omega_{\rm s} \right\}.\]

\(Q,U,V\) are Stokes parameters; \(a,b\) label telescopes, polarization channels, or frequency channels; \(\Delta t_{\rm var}\) is the variability time scale; and \(\Omega_{\rm s}\) is an angular-size constraint. Thermal sources usually have low polarization, finite \(T_b\), and thermal bunching. Optically thin synchrotron radiation can have high linear polarization and a power-law spectrum. Masers are marked by narrow lines, high brightness temperatures, and strong polarization. FRBs are marked by extreme brightness temperature, millisecond-to-microsecond structure, dispersion, and strong polarization. Scattering and plasma propagation can reshape the time structure and spectrum, but they should not be mistaken for the source’s intrinsic emission statistics.

Figure 54 Joint diagnostic plane for radiation mechanisms. The horizontal axis is brightness temperature, the vertical axis is polarization fraction, and point area represents the order of magnitude of the visible second-order coherence contrast. Thermal dust and HII regions lie at low brightness temperature and low polarization. Synchrotron emission can be more polarized, but its brightness temperature usually remains below that of coherent bursts. Electron-cyclotron masers, molecular masers, and FRBs occupy the high-brightness-temperature region where coherent emission, pumping, or charge bunching must be explained.

In data analysis, an anomalous \(g^{(2)}\) peak has to be compared against instrumental, propagation, and source-mixing effects. Dead time, afterpulsing, time-stamp quantization, and trigger selection alter arrival-time correlations. Plasma scintillation, scattering tails, gravitational microlensing, and frequency-dependent absorption can create cross-frequency correlations. A thermal background plus a narrow maser line, or a synchrotron continuum plus a coherent pulse, can leave a smooth mean spectrum while the photon statistics are no longer described by one simple field state. The event tables, time-shift backgrounds, and covariance estimates from Chapters Detectors, clocks, and event tables and Data analysis for event tables provide the analysis layer needed to separate these cases.

The mean energy distribution gives the first classification of a radiation mechanism. Joint probabilities give the next one. Stellar photospheres are close to thermal light, and bright stars provide enough photon rate for careful calibration. The next chapter applies the same coherence functions and Fisher information to that well-controlled astronomical case: angular diameter, limb darkening, rotation, and companions all enter the same visibility curve.