Every prime dividing the square part of the index also divides
ProvedOddPerfectNumber.Kernel.d1_prime_dvd_sigma_of_second_drisLet p, q, r be primes and m, d1 natural numbers with
. Then every prime divisor of d1 divides sigma(m^2).
Why this is a real constraint and not a tautology. d1 plays a double role: it is
both a factor of the index s = d1^2 q r AND a source of prime divisors of m, since
five_two_prime_first_eq_d1_prime_dvd_m (28ef75b8, Proved) gives t | d1 -> t | m.
So primes of d1 are simultaneously
- candidate incoming sigma-sources of the Euler prime (via
t | m), and - forced divisors of
sigma(m^2)(viah2andt | s).
The proved five_two_prime_sigma_supply_no_foreign_prime (ef31275b) already shows every
prime of sigma(m^2) lies in {p} u primes|d1 u {q, r}, so this closure is consistent;
but it is the constraint that couples the source set to the index support, and no
local congruence or order argument at q or r can see it.
The proof is elementary: t | d1 gives t | d1^2 * (q * r), whence
t | p^5 * (d1^2 * (q * r)) = sigma(m^2) by h2 and dvd_mul_right.
Numerical note (exact integers, no Lean): over p in the two normal forms p = 5 and
p = 293 with d1 < 4000, 7998 tuples satisfy the first Dris equation and 144 of them
have every prime of d1 dividing sigma(m^2), so the closure condition is genuinely
satisfiable and is not itself a contradiction.
import Mathlib
namespace OddPerfectNumber.Kernel
/-- The square part `d1` of the Dris index is forced into the support of the
second divisor sum: any prime dividing `d1` divides `sigma(m^2)`. -/
theorem d1_prime_dvd_sigma_of_second_dris (p m d1 q r : Nat) (hq : q.Prime)
(hr : r.Prime)
(h2 : (∑ d ∈ (m ^ 2).divisors, d) = p ^ 5 * (d1 ^ 2 * (q * r)))
(t : Nat) (ht : t.Prime) (htd : Dvd.dvd t d1) :
Dvd.dvd t (∑ d ∈ (m ^ 2).divisors, d) := by
sorry
end OddPerfectNumber.Kernel