deepdive

From a Single Slice to the Volume Integral

Why individually negligible per-event probabilities, summed over a 3D volume of ~10²² scatterers and a broad spectrum, become macroscopically relevant in P(t).

Last updated: 2026-07-26

The second structural reason that a “vanishingly small” elementary interaction can nonetheless yield a measurable output has nothing to do with quantum coherence — treated separately in coherent enhancement and the N² law — and everything to do with counting. A converter is not a single scatterer but a macroscopic body containing on the order of $10^{22}$ to $10^{23}$ constituents, each exposed to a flux that is itself spread over a broad range of energies and directions. The output is the sum over all of them. This chapter makes the aggregation argument explicit: the passage from a single spatial slice to the full volume integral, the law-of-large-numbers logic that turns many rare events into a stable expectation, and the reason that “one event is negligible” is a category error when the quantity of interest is a many-body sum.

The category error

Consider the shape of the objection. A single interaction deposits an energy or momentum so small that, taken alone, it is utterly negligible — say a per-event probability $p_1 \ll 1$ of doing anything useful at all. The inference “therefore the device does nothing” treats the aggregate as if it inherited the smallness of the individual. But the aggregate is a sum, and a sum of many small numbers is not small. If $N$ independent scatterers each contribute an expected amount $\mu_1$ , the expected total is

\langle S \rangle = \sum_{i=1}^{N} \mu_i = N\mu_1 ,

by linearity of expectation, regardless of how tiny each $\mu_1$ is. The individual being negligible and the sum being negligible are simply different statements; conflating them is the category error. With $N \sim 10^{22}$ , a per-constituent contribution smaller than any laboratory scale can still aggregate to a macroscopic total, because $10^{22}$ is an enormous multiplier. The correct question is never “is one event negligible?” — it almost always is — but “what is $N\mu_1$ , and does it clear the threshold of measurability?”

The law of large numbers as a stabiliser

Aggregation does more than sum the mean; it also suppresses the relative fluctuation, which is what makes the total a reliable quantity rather than a lucky accident. Model each constituent as contributing an independent random amount with mean $\mu_1$ and variance $s^2$ . The total $S$ has mean $N\mu_1$ and, for independent contributions, variance $N s^2$ , so its standard deviation grows only as $\sqrt{N}$ . The relative spread therefore shrinks:

\frac{\sigma_S}{\langle S \rangle} = \frac{\sqrt{N}\,s}{N\mu_1} = \frac{1}{\sqrt{N}}\,\frac{s}{\mu_1} \;\xrightarrow[N\to\infty]{}\; 0 .

This is the weak law of large numbers in the guise the framework needs: as $N$ grows, the aggregate output concentrates ever more tightly around its expectation. A device built from $10^{22}$ scatterers has a relative fluctuation of order $10^{-11}$ from this counting alone, so the output is not a noisy trickle of rare events but a smooth, essentially deterministic quantity. The rarity of the individual event and the steadiness of the collective are, once again, entirely compatible — indeed the second follows from the first through the sheer size of $N$ .

Riemann sum to integral

The master formula is the continuum expression of exactly this sum. Begin with the discrete picture: partition the converter into small cells labelled by position, and partition the incident field into small bins of energy $\Delta E$ and direction $\Delta\theta$ . The contribution of one cell in one $(E,\theta)$ bin is a product of the flux in that bin, the coupling, the directed transfer, the transport, and the efficiency — a single small term. The total output is the sum of all such terms over every cell and every bin:

P \;\approx\; \sum_{\text{cells}}\;\sum_{E_j,\,\theta_k} \Phi(E_j,\theta_k)\,\sigma(E_j)\,\Delta p\,v_{\text{ph}}\,\eta_{\text{conv}}\;\Delta E\,\Delta\theta .

As the bins are refined, $\Delta E, \Delta\theta \to 0$ , this Riemann sum converges to the integral, and the sum over cells becomes an integral over the spectrum weighted by the density of scatterers. What remains is precisely the master formula,

P(t) = \iint \Phi(E,\theta,t)\,\sigma(E)\,\Delta p\,v_{\text{ph}}\,\eta_{\text{conv}}\;dE\,d\theta ,

in which the integral sign is not decoration but a bookkeeping device that performs the aggregation automatically. Reading the equation this way dissolves much confusion: the integral is the “sum over many rare events,” carried to the continuum limit in the spectral variables, and the density of scatterers enters through the factors it multiplies. The full construction of this passage is set out in full derivation, and the parallel aggregation across distinct interaction channels in multi-channel energy flux.

Surface versus volume: 2D against 3D

A crucial quantitative point distinguishes converters that harvest at a surface from those that harvest throughout a volume, and it is a matter of dimensional scaling. A surface interaction accumulates over an area, so the number of participating scatterers grows as $L^2$ with the linear size $L$ . A volume interaction accumulates over the whole body, so the count grows as $L^3$ . For a device of characteristic size $L$ and constituent spacing $d$ , the two counts are

N_{\text{surf}} \sim \left(\frac{L}{d}\right)^{2}, \qquad N_{\text{vol}} \sim \left(\frac{L}{d}\right)^{3},

so the volume count exceeds the surface count by the factor $L/d$ , which for a centimetre-scale body with sub-nanometre spacing is of order $10^{7}$ . This is why a field that couples only weakly per constituent but penetrates the bulk is treated on a fundamentally different footing from one that interacts only at the boundary: the available $N$ , and hence the aggregate $N\mu_1$ , is larger by many orders of magnitude. A weakly interacting, deeply penetrating flux is in this sense the ideal candidate for volume aggregation, precisely because its weakness lets it reach every one of the $L^3$ scatterers rather than being absorbed at the surface. The framework keeps this distinction explicit by integrating over a volume density rather than a surface density, and any concrete estimate must state which regime it is in.

What the aggregation argument does and does not license

The discipline of the framework requires stating the limits of this reasoning as plainly as its power. Aggregation legitimately converts a tiny per-event quantity into a potentially macroscopic total, and the law of large numbers legitimately makes that total steady. What it does not do is manufacture energy or evade thermodynamics: every term in the sum is still bounded by the conservation laws and the efficiency ceiling, so aggregating many terms cannot produce more than the flux delivers. Nor does aggregation rescue a mechanism whose per-event contribution is genuinely zero — multiplying zero by $10^{22}$ is still zero. The argument is therefore a necessary corrective to a specific fallacy, not a universal solvent: it shows that smallness of the individual event is not by itself a valid objection, while leaving intact the requirement that the individual event be non-zero and thermodynamically permitted. A concrete instance of carrying an aggregation estimate through to numbers, with every factor flagged as illustrative, is given in worked example. Read together with the coherence argument, the two chapters answer the two most common versions of the “it’s too small” objection: coherence can raise the per-target coupling, and aggregation can raise the per-device count, and neither violates a conservation law.