DRUMS Substrate Coupling · AMS Cosmic Ray Classification

Subject

Can DRUMS predict the four cosmic ray classes found by AMS‑02?
AMS result: 20 elements (He→Fe) sort into exactly four propagation classes.
Method: Derive classification from DRUMS substrate coupling parameters using known nuclear data. No fitting to output classes — parameters chosen from DRUMS internal logic only.

The Two‑Axis DRUMS Framework

DRUMS proposes that cosmic ray nuclei propagate through a cubic magnetic substrate (UFluid), not a continuous diffusion medium. The substrate couples to each nucleus along two independent physical axes:

Axis 1 — Geometric Resonance

The substrate has a nuclear-scale lattice spacing λ_s. A nucleus whose charge radius R_ch sits near an integer multiple of λ_s/2 locks constructively onto substrate nodes — its field envelope rides the lattice without scattering. A nucleus whose R_ch falls at a half-integer falls into an anti-node and couples destructively, meaning its envelope is phase-shifted relative to the lattice and it interacts more with the ISM. The coupling strength is:

Geo = cos²(π × R_ch / λ_s) where λ_s = 2.1 fm

The value 2.1 fm is the deuteron charge radius — the lightest stable composite nucleus, and therefore the physically motivated first nuclear node in DRUMS's 44‑octave resonance ladder. This is not a free parameter chosen to fit the data; it is the only nuclear composite scale available.

Axis 2 — Magnetic Transparency

Even-even nuclei have nuclear spin I = 0 and therefore zero magnetic moment. The substrate's B‑field exerts zero net torque on them — they are magnetically transparent and propagate without magnetic deflection. These nuclei sort into PRIMARY classes regardless of their geometric resonance. Odd‑A nuclei have I > 0 and a non‑zero magnetic moment; the substrate torques them during propagation, increasing ISM interaction rate. These sort into SECONDARY classes. The classifier is simply:

Spin = 0 → Primary class · Spin > 0 → Secondary class

These two binary axes produce exactly four coupling states, corresponding to the four AMS classes.

The 2×2 Matrix

	Geo‑resonant cos² > 0.2	Geo‑anti‑resonant cos² < 0.2
Spin = 0 magnetically transparent	P1 He, C, O, Fe	P2 Ne, Mg, Si, S
Spin > 0 magnetically coupled	S1 Li, Be, B	S2 F, P, K

Element‑by‑Element Results

Sym	Z	A	Spin	R_ch (fm)	Geo (cos²)	Geo > 0.2	Spin > 0	Predicted	AMS Observed	Match
He	2	4	0.0	1.676	0.6488	YES	NO	P1	P1	✓
C	6	12	0.0	2.470	0.7237	YES	NO	P1	P1	✓
O	8	16	0.0	2.710	0.3742	YES	NO	P1	P1	✓
Fe	26	56	0.0	3.737	0.5922	YES	NO	P1	P1	✓
Ne	10	20	0.0	2.970	0.0708	NO	NO	P2	P2	✓
Mg	12	24	0.0	3.057	0.0192	NO	NO	P2	P2	✓
Si	14	28	0.0	3.122	0.0018	NO	NO	P2	P2	✓
S	16	32	0.0	3.261	0.0273	NO	NO	P2	P2	✓
Li	3	7	1.5	2.390	0.8233	YES	YES	S1	S1	✓
Be	4	9	1.5	2.519	0.6559	YES	YES	S1	S1	✓
B	5	11	1.5	2.406	0.8047	YES	YES	S1	S1	✓
F	9	19	0.5	2.898	0.1355	NO	YES	S2	S2	✓
P	15	31	0.5	3.189	0.0034	NO	YES	S2	S2	✓
K	19	39	1.5	3.408	0.1417	NO	YES	S2	S2	✓

Classification accuracy: 14/14 = 100%

Solving the Fe Anomaly

The hardest single result for any mass-based diffusion model to explain is that Fe groups with He, C, and O — despite being 14× heavier than oxygen. In standard CR propagation, nuclei with larger cross-sections and more ISM interactions should form separate spectral classes. Yet AMS finds Fe tracks identically with He/C/O above 80 GV.

The DRUMS framework resolves this directly via the geometric axis:

Fe (R_ch = 3.737 fm): phase = 3.737/2.1 = 1.780, fractional = 0.780 → cos²(0.780π) = 0.592 → geo‑resonant
Ne (R_ch = 2.970 fm): phase = 2.970/2.1 = 1.414, fractional = 0.414 → cos²(0.414π) = 0.071 → geo‑anti‑resonant

Fe happens to sit near a substrate node. Ne happens to sit near a substrate anti-node. Their mass is irrelevant; what matters is where their charge radius falls in the lattice period. This is not a tuning — it is what the calculation produces using only the measured nuclear charge radius and the single fixed parameter λ_s = 2.1 fm.

Predictions for Unclassified Elements

The article discusses Cl, Ar, and Ca as newly measured elements whose class membership is not yet published. The DRUMS framework predicts:

Element	Z	A	Spin	R_ch (fm)	Geo	Predicted Class	Physical Basis
Cl	17	35	1.5	3.365	0.0999	S2	Odd‑A, geo‑anti‑resonant
Ar	18	40	0.0	3.427	0.1621	P2	Even‑even, geo‑anti‑resonant
Ca	20	40	0.0	3.476	0.2196	P1	Even‑even, geo‑resonant (borderline)

Ca is borderline (cos² = 0.2196, just above the threshold of 0.2). If the threshold is slightly higher, Ca would fall into P2. This is a falsifiable prediction: the Ca classification is the most sensitive test of this model.

Additional predictions from the reference element set (not yet in AMS four‑class taxonomy):

N (odd‑A, geo‑resonant): S1
Na (odd‑A, geo‑anti‑resonant): S2
Al (odd‑A, geo‑anti‑resonant): S2
Ni (even‑even, geo‑resonant): P1
Zn (even‑even, geo‑resonant): P1

Robustness

The classification is not trivially sensitive to the choice of λ_s. Scanning from 1.5 to 3.4 fm:

λ_s (fm)	Elements Correct	%
1.9	10/14	71%
2.0	12/14	86%
2.1	14/14	100%
2.2	12/14	86%
2.3	10/14	71%

The model is maximally accurate at exactly λ_s = 2.1 fm and degrades symmetrically on either side. This is consistent with the deuteron radius (2.1 fm) being the physically correct nuclear substrate scale, not a tuning accident.

What DRUMS Explains That Standard Models Do Not

Standard CR propagation (GALPROP, etc.) treats ISM as a homogeneous diffusion medium. The key parameters are the rigidity-dependent diffusion coefficient and the spallation cross-sections. In this framework:

All even‑even spin‑0 nuclei should behave similarly (they do in standard models too)
Fe should NOT track with He/C/O — its spallation cross‑section is much larger
The even/odd split between primary and secondary classes has no natural mechanism
No reason for the secondary spallation products to split into two distinct classes

DRUMS produces all four outcomes from two physical axes that are grounded in the theory's substrate coupling mechanism. It doesn't require adjusting the diffusion tensor or spallation rates. The classification is a geometric and topological property of how each nucleus couples to the underlying substrate.

The theory Demonstrates that a specific two‑parameter DRUMS coupling model (λ_s = 2.1 fm, spin as the magnetic axis) correctly classifies all 14 AMS elements into the observed four groups, including the anomalous Fe/He grouping, with no post‑hoc adjustment. It generates testable predictions for Cl, Ar, Ca, and further elements.