Development documentation — This site follows the main branch and may differ from an installed release. View the changelog.

Photoelectron emission and lifecycle

source_mode="photo_raycast" emits particles from the first surface hit by an illumination ray. This page covers ray casting, emission amount and velocity, source charge, and the closed-PE neutral_return closure. After creation, a photoelectron uses the ordinary fixed field, Boris update, collision query, and box boundaries.

Track emission through reabsorption in the same batch

Launch rays from an illumination aperture on a box face.
Find the first mesh hit while applying box boundary conditions.
Create a photoelectron from the plasma-facing normal of the hit element.
Optionally record $-qw$ on the emission element.
Track it as an ordinary particle and hand it to the common escape/return treatment after it reaches a box boundary.
Commit emission and absorption charge to the surface at batch end.

Emission and reabsorption can occur in one batch, but the field is not refreshed between them. Net surface charge changes the field starting in the next batch.

Locate the emitting surface with illumination rays

inject_face and pos_low/pos_high define a rectangular aperture on a box face. ray_direction must point inward from the aperture and defaults to the inward face normal. For aperture area $A$ , inward normal $\mathbf n_\mathrm{in}$ , and normalized ray direction $\hat{\mathbf d}$ , the area projected perpendicular to the rays is

A_\mathrm{proj}=A\left|\hat{\mathbf d}\cdot\mathbf n_\mathrm{in}\right|.

Ray origins are uniform in the aperture. Each ray searches the first triangle hit on successive segments from its current position to the next box face.

Reached box face	Ray treatment
Nonperiodic face	Leave the box and terminate without emission
Face on `domain.periodic_axes`	Wrap to the opposite face and continue

With field_boundary.mode="periodic2", first-hit search includes periodic images and wraps the emission position into the primary cell. The hit element itself must lie inside the simulation box. A ray that still has no hit after raycast_max_bounce creates no particle. An incomplete collision-query status such as an image limit, index range, or stalled DDA fails the batch instead of silently treating the ray as a miss.

Assign emitted current to each ray

For emission current density $J_\mathrm{emit}>0$ , physical-particle charge $q\ne0$ , and global ray count $N_\mathrm{ray}$ , the macro-particle weight created by a hit ray is

w_\mathrm{hit} =\frac{J_\mathrm{emit}A_\mathrm{proj}\,\Delta t_\mathrm{batch}} {|q|N_\mathrm{ray}}.

A missed ray creates no particle, so shadowing and apparent surface area enter total emission through hit probability. w_particle and target_macro_particles_per_batch are not accepted for photo_raycast.

rays_per_batch is the sample count for the illumination integral, not the physical emission amount. Increasing it reduces $w_\mathrm{hit}$ and Monte Carlo noise in visibility and emission location. Check convergence with ray count.

Build the emitted state from the surface normal

If the stored triangle normal points along the incident ray, it is reversed to form emission normal $\mathbf n_s$ facing back toward the illumination side. Position is offset $10^{-12}$ m along $\mathbf n_s$ to avoid immediate self-intersection with the source element.

With $\sigma=\sqrt{k_\mathrm{B}T/m}$ , velocity is sampled in local basis $(\mathbf n_s,\mathbf t_1,\mathbf t_2)$ :

normal speed follows a flux-weighted half-range Maxwell distribution with normal_drift_speed;
both tangential components are zero-mean Gaussians with standard deviation $\sigma$ ;
Gaussian sampling is truncated at $6\sigma$ .

Normal velocity is positive, so a new particle leaves toward the illumination side. Its tracked orbit and common box boundaries then determine reabsorption, escape, or local reflection.

Check charge balance across emission, reabsorption, and escape

With deposit_opposite_charge_on_emit=true, source element $i$ receives

\Delta q_{i,\mathrm{emit}}=-q w.

For an electron $q<0$ , this leaves positive charge. The difference is accumulated separately as photo_emission_dq, combined with collision deposition after MPI all-reduce, and included in the same batch commit.

If the emitted particle is reabsorbed on element $j$ , ordinary absorption deposits $+qw$ there. Return to the source element cancels emission charge; return to another element transfers net charge across the surface. The current insulator model applies no subsequent lateral surface conduction.

Use the common escape and return treatment after emission

A ray hit always creates a photoelectron with weight $w_\mathrm{hit}$ . No emission-time setting multiplies the weight by an escape fraction. Surface return is an ordinary collision; an open-face event uses the same particle_boundary.ordinary_open_model treatment as other sources.

Close only photoelectrons at the injection face

Use local reflection to evaluate photoelectron surface redistribution on closed orbits:

[particle_boundary]
z_high = "open"
ordinary_open_model = "escape"

[[particles.species]]
species_key = "photoelectron"
source_mode = "photo_raycast"
inject_face = "z_high"
deposit_opposite_charge_on_emit = true
surface_charge_closure = "neutral_return"

[particles.species.boundary]
z_high = "reflect"

In this example, a z-high crossing reverses only normal velocity, preserves tangential velocity, and reintegrates the step remainder. Ambient species using the default inherit in [particles.species.boundary] follow the global open contract. All six species faces accept inherit, open, reflect, or redistributed_reflect. Closed PE requires the effective action on the face named by inject_face to be reflect or redistributed_reflect; a face on domain.periodic_axes cannot be overridden.

Ordinary reflect also preserves tangential position. Only when the in-plane position of returning photoelectrons should be uniformly redistributed, replace the baseline value above with:

[particles.species.boundary]
z_high = "redistributed_reflect"

redistributed_reflect applies ordinary reflection to velocity and, for a single face, uniformly resamples only the two in-plane coordinates over the box span excluding its end guards. This option generalizes the horizontal-position randomization used for top-boundary photoelectron return by Zimmerman et al. (2016) to any nonperiodic face and simultaneous face event. It does not add a self-consistent outer sheath. See Particle collision and boundary events for simultaneous-event rules.

Species-level reflection alone closes only the orbit. A particle that does not return by max_step remains unresolved. With surface_charge_closure="neutral_return", BEACH measures emitted charge $S<0$ and resolved absorbed charge $R<0$ globally for the batch, then multiplies deposits at measured return destinations by $S/R$ . Combined with source reaction charge, the photoelectron contribution to total surface charge is exactly zero. The approximation assigns unresolved particles the same destination distribution as resolved returns.

Raw absorbed_on_surface_C and discarded_unresolved_C are not replaced. charge_ledger.csv records the correction, neutral_return_weight_scale, and neutral_return_unresolved_fraction separately. A run stops on nonzero emission without resolved return, actual escape, soft_discard, or a charge-sign mismatch. An unresolved fraction above 5% also stops before correction because it is outside this closure’s fixed applicability range.

Full reflection is an artificial mirror at the injection face of a finite box, not a self-consistent sheath or quasineutrality solution. neutral_return is a statistical zero-net-photoelectron-current closure, not a resolved trajectory for every long-lived particle. Converge max_step, dt, ray count, and batch width until abs(weight_scale-1) and the unresolved fraction are small. See Finite-image periodic2 configuration for the integrated field, inflow, and potential-reference setup.

For coupling to an external current model, use surface_charge_closure="fixed_current" and specify target_emission_current_a and target_absorbed_current_a as separate signed surface currents. BEACH uniformly rescales the raycast emission-origin distribution and the tracked return-destination distribution independently; it never uses the small net PE current, the difference of these two large channels, as a scale denominator. If the external model injects a return VDF, keep the top face open and do not also enable neutral_return on the same or another return channel.

With [surface_current_model] model="zhao_stationary", BEACH independently calculates PE emission, escape, and return from the Zhao zero-current stationary root. Emission and return are applied to surface channels; escape is recorded as an external-boundary target that is not deposited onto the surface. The closure never uses the PE net current as a scaling denominator, so it remains stable when large currents nearly cancel. Raw escape statistics are preserved for comparison with target / applied / correction values.

See Zhao stationary closure for the zero-current root, Type A/B/C barriers, and the distinction between fixed and batch-dependent quantities.

This stability concerns total-current normalization, not the statistical accuracy of the raw spatial map. If there is one return hit, that element receives the entire return target. BEACH imposes no fixed minimum hit count, so inspect the raw ledger counts and fixed_*_weight_scale, then test the elementwise distribution across rays_per_batch, batch widths, and RNG seeds. Closure of the Zhao zero-current budget does not replace this convergence check.

The emission VDF remains the configured surface half-Maxwellian. Keep z-high open; the outward crossing test uses $\phi_m$ for Type A and 0 V for Type B/C as the outside barrier potential, together with the local crossing potential and normal kinetic energy, to split return from escape. This reflection contracts the outer turning point onto z-high: BEACH does not solve the outside sheath field, space charge, return distance, or return delay. See examples/periodic2_zhao_fixed_current.toml for the complete setup.

See Particle escape and local return to choose between particle_boundary.ordinary_open_model and closed PE.

Check convergence of photoelectron emission

Increase rays_per_batch and verify convergence of hit fraction, emitted current, and charging distribution. When reabsorption location matters, reduce dt and verify that the result is unchanged. See Inspect Output Files for the species-resolved charge balance and closed-PE correction diagnostics.

Code reference

Ray propagation, hit, emission velocity, and weight: bem_injection.f90
Source creation and emission charge difference: bem_app_config_runtime.f90