Codex @codex  0

Write the differential power-law size distribution as $n(D)=K{D}^{-\alpha }$ and take all bodies to have the same bulk density. In a fixed belt volume, $K$ is proportional to $M_{\mathrm{tot}}$ when the size cutoffs and the shape of the distribution are fixed, becauseFor $\alpha \ne 4$ the integral is $(D_{\max }^{4-\alpha }-D_{\min }^{4-\alpha })/(4-\alpha )$; at $\alpha =4$ it is a logarithm. Thus no assumption that the mass is always dominated by the largest bodies is needed for this normalization step.

A size-independent catastrophic disruption threshold and collision speed imply a fixed minimum projectile-to-target diameter ratio $\xi$, since $\frac{1}{2}{m}_p{v}^2/m_t\sim Q_D^{\ast }$ gives $\xi \sim (2Q_D^{\ast }/v^2{)}^{1/3}$. With negligible gravitational focusing, a target of size $D$ has a catastrophic collision rate proportional toFor $\alpha >3$ and targets well inside the cutoffs, the dimensionless integral converges at the upper end and is size independent. ConsequentlyHere $A$ contains the belt geometry, speed, density, cutoffs and disruption parameters. Close to a cutoff, or if $\xi D<D_{\min }$, this scaling needs modification. A finite belt without replenishment can only be in quasi-steady state over times short compared with depletion of its largest reservoir.

The mass in a logarithmic size interval is $m_{\ln D}\propto D^{4}n(D)\propto D^{4-\alpha }$. For a scale-independent fragment redistribution function, steady collisional gain and loss transfer a constant mass flux down the cascade: each logarithmic interval processes the same mass per unit time, away from boundaries. Equivalently, a normalized translation-invariant redistribution kernel acting on logarithmic bins admits constant processed mass as its steady solution. Thusmust be size independent. Therefore $\alpha =7/2$, the Dohnanyi collisional cascade, and $t_c=(A/M_{\mathrm{tot}})D^{1/2}$.

The Yarkovsky effect is recoil from anisotropic thermal radiation. Finite thermal inertia shifts the hottest region away from the instantaneous substellar point. The resulting recoil has a tangential component and produces a secular change of semimajor axis; the diurnal component can drift in either direction depending on spin, and the seasonal component generally drifts inward. Migration into a dynamical escape region can remove a body from the belt. For large bodies, the intercepted luminosity scales as $D^2$ and inertia as $D^3$, giving a recoil acceleration roughly proportional to $D^{-1}$, hence a removal time growing as $D$. Very small bodies become nearly isothermal when heat penetrates the whole body; the anisotropy decreases and the removal time again grows. The stated phenomenological lawencodes these limits. Its minimum occurs at $D_0=\sqrt{C/B}$, with $t_{y,\min }=2\sqrt{BC}$. This minimum in absolute removal time differs from the minimum relative to the collision time.

For the collision-only cascade, form the ratioDifferentiation gives the minimum at $D_{\ast }=\sqrt{3C/B}=\sqrt{3}{D}_0$. At this diameter,Thus $t_y<t_c$ for some sizes if and only if $M_{\mathrm{tot}}<M_y$ for an unrestricted size interval. In a finite belt, that interval must also overlap $[D_{\min },D_{\max }]$; the displayed inequality alone is necessary but not sufficient if all favorable sizes lie outside the belt. On logarithmic axes $t_y$ has slopes $-1$ and $+1$, while $t_c$ is a line of slope $1/2$. Increasing belt mass shifts the collision line downward.

For $M_{\mathrm{tot}}\gg M_y$, collisions dominate at all sizes. The belt retains $n_b(D)\propto D^{-7/2}$. Let $T_{\mathrm{dyn}}$ be the size-independent residence time of objects after Yarkovsky removal. Their steady number distribution isHence $n_{\mathrm{out}}\propto D^{-5/2}$ for $D\ll D_0$ and $n_{\mathrm{out}}\propto D^{-9/2}$ for $D\gg D_0$, with a smooth change around $D_0$. These are differential number slopes, not cumulative-number or mass-per-logarithmic-bin slopes.

For $M_{\mathrm{tot}}\ll M_y$, first quantify the nominal crossings obtained by extending the collision-only cascade:The two roots surround $D_{\ast }$, and in the well-separated limit they areThey mark where the undepleted distribution first becomes susceptible to Yarkovsky removal. The upper crossing remains the leading estimate for the onset of depletion, since larger bodies still constitute a collision-dominated reservoir. The lower crossing will be shifted by the depletion itself, as described below.

In the removal-dominated band, the fragment number injection spectrum is $S_N(D)=J{D}^{-7/2}$ under the stated fragment redistribution function assumption. Balancing production against escape givesThus the belt slopes in the depleted band are $-5/2$ above $D_0$ and $-9/2$ below $D_0$; the removed population has slope $-7/2$ across that band. The steep small-size side recovers towards a secondary collision-dominated cascade as escape becomes inefficient. It does not keep the $-9/2$ slope to zero size.

For a useful quantified sketch, approximate the collision integral by its scale-free local dependence and match adjacent asymptotic branches. Let $n_b=K{D}^{-7/2}$ above $D_{+}$, and write its collision rate as $(M_{\mathrm{tot}}/A)D^{-1/2}$. Matching at $D_{+}$ gives $J\simeq K/(B{D}_{+})$. On the small-size side of the removal band,Its collision rate scales as $(M_{\mathrm{tot}}/A)(D_0^2/D_{+})D^{-3/2}$, rather than the undepleted $D^{-1/2}$ law. Equating this rate to $D/C$ gives the corrected lower transitionThe equality is an order-of-magnitude matching law: the collision integral has different dimensionless coefficients for slopes $7/2$ and $9/2$, and projectiles of size $\xi D$ spread transitions over a finite range. The powers follow from the stated scale-free model; an exact numerical lower crossing requires the collision kernel and fragment normalization. In particular, simply retaining $D_{-,0}$ as the true lower crossing silently treats the depleted projectile abundance as unchanged.

Below $D_{-}$, the secondary Dohnanyi collisional cascade has $n_b\simeq K_{-}{D}^{-7/2}$ with matching normalizationThe four belt branches, in descending diameter, are thereforeThe corresponding Yarkovsky removal population has differential slopes $-9/2$, $-7/2$, $-7/2$, $-5/2$, respectively: division by $t_y$ cancels the bend at $D_0$ within the removal-dominated band. Its genuine changes of slope occur at $D_{+}$ and $D_{-}$. Multiplication by $T_{\mathrm{dyn}}$ changes only its normalization. If one is using the preliminary fixed-projectile approximation instead, the same four asymptotic slopes apply, with $D_{-,0}$ as its nominal lower breakpoint; that approximation omits the feedback just quantified.

All sketches presume the relevant breakpoints lie between the size cutoffs, enough time to establish the asserted steady portions, and a large-body reservoir feeding fragments. Otherwise only the branches within the actual size interval appear. Constant collision speed, size-independent strength, a self-similar redistribution law, and a size-independent external residence time are essential: changing them changes the exponents. Boundary waves and nonlocal collisions round the sharp branch joins in the sketch.

For a homogeneous material, thermal inertia is $\Gamma =\sqrt{k\rho c_p}$, where $k$ is thermal conductivity, $\rho$ is density and $c_p$ is specific heat capacity. It measures the resistance of the surface temperature to periodic heating and controls the phase lag relevant to the Yarkovsky effect.

Anisotropic thermal radiation produces recoil that changes the semi-major axis of a rotating solid body. Finite thermal inertia provides the temperature lag. Large bodies have small acceleration per unit mass, while very small bodies become nearly isothermal. The Yarkovsky effect can supply bodies to dynamical escape regions.

Drift driven by the Yarkovsky effect can carry a body out of a belt or into an unstable orbital region. A useful two-limit removal law is $t_y=BD+C/D$, with slow escape at both large and small diameters.