Past exam of the astrophysics course of the University of Oxford / 2021 / 1

A hydrogen recombination line is a discrete spectral line produced when radiative recombination captures an electron into an excited bound level and a subsequent radiative cascade emits photons. The free-to-bound capture itself produces a continuum; the later bound-to-bound transitions produce the recombination lines. For example, H-alpha is the $n=3\to 2$ transition in the Balmer series.

For the neutral hydrogen absorbing an ionizing stellar continuum, the principal bound-free transition is photoionization from the $1s$ ground state into the continuum. Most neutral hydrogen is in its ground state, and the photoionization cross section is largest near its $13.6\mathrm{eV}$ threshold, falling approximately as ${\nu }^{-3}$ above it. Dominance of the population-weighted opacity is intended here; this does not mean that every excited-state microscopic cross section is smaller at every frequency.

In an ionization-bounded, optically thick nebula, direct radiative recombination into $1s$ emits an ionizing photon that promptly photoionizes another hydrogen atom. One ion disappears and another is created: there is no net removal of an ion from the global budget. Case B recombination therefore uses ${\alpha }_B={\sum }_{n\ge 2}{\alpha }_n$, with local recycling of ground-state recombination photons. This is the on-the-spot approximation, rather than an unconditional property of radiative recombination: if those photons escape, direct capture does produce net neutralization and a Case A treatment is appropriate.

Write $Q={\dot{N}}_{\gamma }$ for the hydrogen-ionizing photon production rate, counting only photons above the hydrogen ionization energy, not all stellar photons. In steady photoionization equilibrium, the external supply balances the volume integral of net Case B recombination:For uniform, almost completely ionized pure hydrogen, $n_e=n_p=n_H$. HenceThe final approximation requires $r_{\mathrm{in}}\ll r_S$. The Strömgren sphere also assumes spherical symmetry, a sharp ionization front, a steady source older than the recombination time, uniform number density and temperature so that ${\alpha }_r$ is constant, negligible dust absorption of ionizing photons, negligible leakage, and no important helium correction or collisional ionization. Clumping would replace $n_H^2$ by a volume average of $n_e{n}_p$.

An additional atomic input is needed: the total Case B recombination coefficient does not specify the number of H-alpha photons per net radiative recombination. Let $y_{\alpha }={\alpha }_{\mathrm{H}\alpha }^{\mathrm{eff}}/{\alpha }_r$, where ${\alpha }_{\mathrm{H}\alpha }^{\mathrm{eff}}$ is the effective recombination coefficient. At the approximately $10^4\mathrm{K}$ temperature implicit in the supplied ${\alpha }_r$, take $y_{\alpha }\simeq 0.45$. The coefficient 2.2 used later in the paper corresponds to $y_{\alpha }=1/2.2\simeq 0.455$, which we use consistently here.

For an unextinguished, fully filled Strömgren sphere with negligible central cavity, the H-alpha luminosity and observed radiative flux satisfySubstitution eliminates $r_S$ and givesConvert the angular diameter to radians: $\theta =56\times 10^{-3}/206265=2.7150\times 10^{-7}$. Also $hc/{\lambda }_{\mathrm{H}\alpha }=3.0268\times 10^{-19}\mathrm{J}$. ThusThe implied hydrogen-ionizing photon production rate is $Q\simeq 1.00\times 10^{39}{\mathrm{s}}^{-1}$. Using $y_{\alpha }=0.45$ instead gives $D\simeq 505\mathrm{pc}$. These values use the printed milliarcseconds and integrated line observed radiative flux. Assuming one H-alpha photon per recombination would give a different, unjustified distance. An appreciable cavity or a nonunit filling factor must also enter the emitting volume.

The Brackett series consists of transitions ending at $n=4$, at infrared wavelengths. They suffer much less interstellar extinction than H-alpha, so their corrected observed radiative fluxes can give a more reliable version of the same inference. At fixed measured angular size and number density, the formula has $D\propto F$: uncorrected interstellar extinction therefore makes this particular distance estimate too small. Infrared lines are weaker and need their own effective recombination coefficients; their advantage is reduced sensitivity to dust.

Let $q_{\mathrm{in}}$ be the photoionization rate per neutral hydrogen atom just outside the empty cavity. With an isotropic central source and no attenuation inside that cavity,The cross section is zero below threshold; this permits the lower integration limit zero. Pure hydrogen with hydrogen ionization fraction $\zeta$ has $n_{\mathrm{HI}}=n_H(1-\zeta )$ and $n_e=n_p=n_H\zeta$. Local photoionization equilibrium is thereforeDefine the dimensionless quantity ${\Theta }_{\mathrm{phys}}={\alpha }_r{n}_H/q_{\mathrm{in}}$. Solving the resulting quadratic equation, ${\Theta }_{\mathrm{phys}}{\zeta }_{\mathrm{in}}^2+{\zeta }_{\mathrm{in}}-1=0$, gives the physical rootThe rationalized form avoids numerical cancellation for small ${\Theta }_{\mathrm{phys}}$. Strong illumination gives ${\zeta }_{\mathrm{in}}\simeq 1-{\Theta }_{\mathrm{phys}}$; weak illumination gives ${\zeta }_{\mathrm{in}}\simeq {\Theta }_{\mathrm{phys}}^{-1/2}$.

Eliminating $Q$ with the H-alpha luminosity, $Q=2.2(4\pi D^2{F}_{\mathrm{H}\alpha }){\lambda }_{\mathrm{H}\alpha }/(hc)$, yieldsThe printed definition of $\Theta$ is inconsistent with the defined radii. The source calls $r_{\ast }$ a physical stellar radius, but its factor $(2\pi r_{\ast }D{)}^2$ gives $\Theta$ units of length to the fourth power, rather than a dimensionless value. It also omits the physically necessary cavity radius $r_{\mathrm{in}}$. For a concrete counterexample, double $r_{\mathrm{in}}$ while holding the source, $D$, and $n_H$ fixed: the incident photoionization rate falls by four, so ${\Theta }_{\mathrm{phys}}$ rises by four; the printed expression does not change. The root has the requested form with the corrected ${\Theta }_{\mathrm{phys}}$ above. There is no supplied relation allowing $r_{\mathrm{in}}$ to be replaced by $r_{\ast }$.

Retain the consistent ${\Theta }_{\mathrm{phys}}$ of d. Across a sufficiently thin layer, the neutral hydrogen column density is $N_{\mathrm{HI}}=n_H(1-{\zeta }_{\mathrm{in}})\Delta r+O(\Delta r^2)$. The direct stellar continuum acquires optical depth ${\tau }_{\nu }=N_{\mathrm{HI}}{\sigma }_{\nu }$. Define the photon-weighted second momentExpanding the attenuating exponential in the photoionization rate givesThe second cross section moment appears because the radiation both undergoes absorption and is weighted by the photoionization cross section in computing the rate.

In a local plane-parallel approximation that suppresses geometrical dilution, setThen local photoionization equilibrium becomes ${\Theta }_{\mathrm{phys}}{\zeta }^2=\Gamma (1-\zeta )$, soThis gives the requested algebraic dependence, with the corrected definition of $\Theta$. Besides $\Delta r\ll r_{\mathrm{in}}$, linear attenuation requires $N_{\mathrm{HI}}{\sigma }_{\nu }\ll 1$ over the frequencies contributing appreciably to the rate. In particular $\Gamma$ must remain positive.

The source's $\Gamma$ neglects $-2\Delta r/r_{\mathrm{in}}$, which is also first order in layer thickness. A consistent spherical first-order answer uses ${\Gamma }_{\mathrm{sph}}=\Gamma -2\Delta r/r_{\mathrm{in}}$ in the same root. Neglecting dilution relative to absorption additionally requires $n_H(1-{\zeta }_{\mathrm{in}})⟨{\sigma }^2⟩/⟨\sigma ⟩\gg 2/r_{\mathrm{in}}$. As a concrete limiting test, with negligible absorption the source's $\Gamma =1$ predicts no change, while inverse-square dilution still decreases $q$ and thus $\zeta$.