Modular Lifting Rings:
Euler Products, Primorial Lifts,
and Prime Spiral Geometry

A framework for lift survival, a new constant, and five geometric theorems

Wessen Getachew wessengetachew.github.io · March 2026
MSC 2020 11A07, 11N05, 11M06
Keywords modular arithmetic · lift survival · Euler product · radical dependence · primorial primes · Feller–Tornier · ζ(2)
Status Original result · C not in OEIS or Finch
Abstract

We introduce the Modular Lifting Rings framework from first principles and establish twelve original results. For each modulus M ≥ 2, the coprime residue ring R(M) = {r ∈ {1,…,M−1} : gcd(r,M) = 1} is placed geometrically on a circle. A residue r lifts from ring M to ring M+1 when gcd(r,M+1) = 1 as well. The long-run fraction of coprime residues that lift converges to a new constant

C = ∏p (p²−2)/(p²−1) ≈ 0.530711806246…

not found in OEIS or Finch's Mathematical Constants. We prove: (1) the Euler product formula for the family Ck; (2) strict positivity Ck > 0; (3) radical dependence Ck = Crad(k); (4) strict monotonicity; (5) the identity C = ζ(2)·dFT; (6) the Primorial Lift Theorem T(p#) = φ(p#) ⟺ p#+1 is prime; and five geometric theorems (§8) governing prime spirals, equatorial gaps, mirror lifts, sector portraits of (Z/pZ)*, upper path permanent blockage, and lower path alternation — the last connecting to Gaussian integer splitting in Z[i]. Every result is proved from first principles.

Numerical value: C ≈ 0.530711806246 · ζ(2)·dFT = (π²/6) · 0.32263462… = C ✓ · Verified empirically to M = 2,000,000.

§0.First Principles

Fix M ≥ 2. The modular ring R(M) is the multiplicative group of integers mod M:

R(M) = { r ∈ {1,…,M−1} : gcd(r,M) = 1 }, |R(M)| = φ(M)

Each r is placed at angle θ = 2πr/M on a unit circle of radius ρ(M). A residue r lifts from ring M to ring M+1 when:

r lifts M→M+1 ⟺ gcd(r,M) = 1 and gcd(r,M+1) = 1

The two conditions are always independent: any prime dividing both M and M+1 would divide M+1−M = 1, which is impossible. So the lift condition is simply gcd(r, M(M+1)) = 1.

Example — R(30)
30 = 2·3·5. Then R(30) = {1, 7, 11, 13, 17, 19, 23, 29}, φ(30) = 8. By CRT, (ℤ/30ℤ)* ≅ (ℤ/2ℤ)* × (ℤ/3ℤ)* × (ℤ/5ℤ)*. Each r ↔ (r mod 2, r mod 3, r mod 5). Since 31 is prime, all 8 elements of R(30) satisfy gcd(r, 31) = 1: perfect lift survival. T(30) = 8 = φ(30).

§1.The Lift Survival Constant C

Define the lift survival count T(M) = #{r ∈ R(M) : gcd(r,M+1) = 1}. The lift survival constant C is:

C = limN→∞M=2N T(M) / ∑M=2N φ(M)

Geometric reading. C is the long-run fraction of unbroken (green) lift lines in the Modular Lifting Rings visualization. Equivalently: pick a uniformly random coprime residue r on a uniformly random ring M — the probability it also satisfies gcd(r, M+1) = 1 is C.

NCoprime pointsLift to M+1Empirical CError
1003,0431,6360.537627…6.9 × 10⁻³
1,000304,191161,6320.531350…6.4 × 10⁻⁴
100,0000.530721…9.3 × 10⁻⁶
2,000,0000.530711966…< 6 × 10⁻⁸

First sign change of C(N)−C at N = 1,900,001 (verified). The true value C ≈ 0.530711806246 does not appear in OEIS (digit string 5307118) or in Finch's Mathematical Constants (2003).

§2.Theorem 2.1 — Euler Product

The lift condition factors over primes. At prime p, the condition gcd(r,M) = 1 excludes r ≡ 0 (mod p); the condition gcd(r,M+k) = 1 excludes r ≡ −k (mod p). These are distinct classes unless p | k.

Case p ∤ k. Two distinct classes excluded from p residues → fraction (p−2)/p survives. Normalizing by the coprimality density gives local factor (p²−2)/(p²−1).

Case p | k. The two exclusions coincide (r ≡ 0 and r ≡ −k ≡ 0 mod p are the same). One class excluded → local factor (p−1)(p²+p−1)/p³.

Theorem 2.1 — Euler Product for Ck
For every integer k ≥ 1, the limit Ck exists and equals the absolutely convergent Euler product:
Ck = ∏p ∤ k (p²−2)/(p²−1) · ∏p | k (p−1)(p²+p−1)/p³
In particular C = C1 = ∏all p (p²−2)/(p²−1).
Proof sketch.

The Euler product follows from independence across primes (CRT). At each prime p, the local survival probability is computed by counting residue classes mod p excluded by the joint conditions gcd(r,M)=1 and gcd(r,M+k)=1. The normalization accounts for the denominator ∑φ(M) already encoding the M-coprimality, so only the additional M+k condition contributes the local factor. Absolute convergence follows from ∑p 1/(p²−1) ≤ 2∑n≥2 1/n² = 2(π²/6−1) < ∞. □

pfp (p ∤ k)fp (p | k)1−f (p ∤ k)
22/3 = 0.666675/8 = 0.625001/3 = 33.3%
37/8 = 0.8750022/27 = 0.814811/8 = 12.5%
523/24 = 0.95833116/125 = 0.928001/24 = 4.2%
747/48 = 0.97917330/343 = 0.962101/48 = 2.1%
11119/120 = 0.99167≈ 0.984221/120 = 0.83%

§3.Proposition 3.1 — Strict Positivity

Proposition 3.1 — Strict Positivity
Ck > 0 for every k ≥ 1. No gap class is geometrically blocked.
Proof.
  1. Each local factor is in (0,1). For p ∤ k: numerator p²−2 ≥ 2 > 0 (since 2²−2 = 2); denominator p²−1 > 0. For p | k: (p−1)(p²+p−1)/p³ with p−1 ≥ 1, p²+p−1 ≥ 4+2−1 = 5, p³ > 0. Both factors are strictly positive.
  2. Absolute convergence. For p ∤ k: 1−fp = 1/(p²−1) < 2/p². This bound holds since p²−1 > p²/2 for all p ≥ 2. Since ∑p 2/p² ≤ 2·(π²/6−1) < ∞, the sum of costs converges.
  3. Product bounded away from zero.p>P(1−2/p²) ≥ exp(−∑p>P 4/p²) → 1 as P → ∞. So Ck ≥ (finite partial product) · (positive tail). □

§4.Theorem 4.1 — Radical Dependence

Theorem 4.1 — Radical Dependence
Ck = Crad(k), where rad(k) = product of distinct prime divisors of k.

Equivalently: Ck depends only on which primes divide k, not on their multiplicities.
Proof.

From Theorem 2.1, the local factor at p is (p²−2)/(p²−1) when p ∤ k, and (p−1)(p²+p−1)/p³ when p | k. The condition p | k depends only on whether p divides k, i.e. on whether p ∈ rad(k). Replacing k by rad(k) does not change which primes divide k, only their multiplicities — which do not appear in the formula. □

Corollary 4.2 — Explicit equalities (verified to 10 decimal places)
C2 = C4 = C8 = … = 0.4975427314… [rad = {2}]
C3 = C9 = C27 = … = 0.6014652038… [rad = {3}]
C6 = C12 = C18 = … = 0.4486551229… [rad = {2,3}]
C30 = C60 = C90 = … = 0.4408347589… [rad = {2,3,5}]

§5.Theorem 5.1 — Strict Monotonicity

Theorem 5.1 — Strict Monotonicity
If rad(j) ⊊ rad(k) then Cj > Ck. In particular, Ck < C for all k ≥ 2.
Proof — Key Lemma first.
Lemma 5.2
fp|k < fp∤k for every prime p ≥ 2.

We must show (p−1)(p²+p−1)/p³ < (p²−2)/(p²−1). Cross-multiplying:

g(p) = (p²−2)·p³ − (p−1)(p²+p−1)(p²−1) = p³ − p² − 2p + 1

Factor: g(p) = p²(p−1) − (2p−1). At p=2: g(2) = 4·1−3 = 1 > 0. Since g′(p) = 3p²−2p−2 > 0 for p ≥ 2, g is strictly increasing, so g(p) ≥ 1 > 0 for all primes p ≥ 2. ✓

Main proof. Let T = rad(k) \ rad(j) be nonempty. By Theorem 4.1:

Ck/Cj = ∏p ∈ T [fp|k / fp∤j]

Each factor is strictly less than 1 by Lemma 5.2. The finite product is strictly less than 1, so Ck < Cj. For k ≥ 2: any prime divisor p of k gives rad(k) ⊋ ∅ = rad(1), so Ck < C1 = C. □

krad(k)CkStep ratio
10.530711806
2{2}0.4975427310.9375
6{2,3}0.4486551230.9021
30{2,3,5}0.4408347590.9826
210{2,3,5,7}0.4375254850.9925
2310{2,3,5,7,11}0.4351369160.9945

§6.Theorem 6.1 — The Identity C = ζ(2)·dFT

Two classical constants appear in the Euler product for C:

ζ(2) = ∏p p²/(p²−1) = π²/6 ≈ 1.64493… [Basel problem]
dFT = ∏p (p²−2)/p² ≈ 0.32263… [Feller–Tornier constant, OEIS A065474]

The Riemann zeta value ζ(2) = π²/6 is the solution to the Basel problem. At s=2 it is simply a specific positive real number; there is no connection here to the non-trivial zeros of ζ(s) or to the Riemann Hypothesis, which concerns the critical strip 0 < Re(s) < 1.

Theorem 6.1 — Feller–Tornier Identity
C = ζ(2) · dFT = (π²/6) · ∏p(1 − 2/p²)
= 1.6449340668… × 0.3226346166… = 0.5307118062… ✓
Proof.

Multiply the three Euler products term by term (valid by absolute convergence):

ζ(2) · dFT = ∏p[p²/(p²−1)] · [(p²−2)/p²] = ∏p(p²−2)/(p²−1) = C

The factor p² cancels in every term. □

Remark 6.2 — Three constants, one identity
ζ(2) encodes the distribution of prime gaps globally. dFT measures squarefree kernel density. C measures modular lift survival. The identity reveals all three are algebraically linked through a single cancellation in their Euler products. Equivalently: dFT = 6C/π².

§7.Theorem 7.1 — Primorial Lift Theorem

For the primorial p# = 2·3·5·…·p, the survival rate T(p#)/φ(p#) is usually far below C. But when p#+1 is prime, something remarkable occurs:

p#p#+1Factor.T(p#)φ(p#)Rate
23prime111.000 ✓
67prime221.000 ✓
3031prime881.000 ✓
210211prime48481.000 ✓
23102311prime4804801.000 ✓
300303003159×509565257600.981 ✗
Theorem 7.1 — Primorial Lift Theorem
T(p#) = φ(p#) ⟺ p#+1 is prime.

When p#+1 is prime, every coprime residue on ring p# survives the lift to p#+1 — a 100% survival rate, compared to the average C ≈ 0.5307. These are the maximal outliers in the distribution of T(M)/φ(M).
Proof (⇒): p#+1 prime ⟹ T = φ.

Suppose Q = p#+1 is prime. For any r ∈ R(p#), we have 1 ≤ r ≤ p#−1 < Q. Since r < Q and Q is prime, Q cannot divide r. Therefore gcd(r, Q) = 1 for every r ∈ R(p#). Every residue lifts. T(p#) = φ(p#). □

Proof (⇐): p#+1 composite ⟹ T < φ.

Let q be the smallest prime factor of the composite p#+1. Since p# is the product of all primes ≤ p, and q cannot divide p# (otherwise q | (p#+1)−p# = 1), we have gcd(q, p#) = 1, so q ∈ R(p#). But gcd(q, p#+1) = q > 1, so r = q fails to lift. T(p#) < φ(p#). □

Remark 7.2 — Primorial primes
Known cases: p#+1 prime for p = 2, 3, 5, 7, 11, giving p#+1 = 3, 7, 31, 211, 2311. Whether infinitely many primorial primes exist is open. Theorem 7.1 reframes this:
Infinitely many primorial primes ⟺ infinitely many perfect lift rings
Corollary 7.3 — Exact formula at prime rings
For every prime p: T(p) = φ(p+1) − 1.

Proof. Among {1,…,p+1}, exactly φ(p+1) integers are coprime to p+1 by definition. All lie in {1,…,p} (since p+1 itself is not coprime to p+1). One of them is r = p (since gcd(p,p+1)=1). Removing r = p (which is outside {1,…,p−1} = R(p)), the count in R(p) is φ(p+1)−1. □

§8.Prime Spiral Geometry

Each prime p traces a spiral path across the ring system: on ring M, p occupies angle θ = 2πp/M. As M increases, the angle p/M decreases from near 1 toward 0. The equator θ = π (the line r/M = 1/2) divides the circle into a top half (r/M > 1/2, M < 2p) and a bottom half (r/M < 1/2, M > 2p). Five theorems govern this geometry.

Setup — Mirror symmetry
R(M) is always symmetric under r ↔ M−r: gcd(M−r,M) = gcd(r,M). Every top-half point has a mirror in the bottom half.

8.1 — The Equator Gap

Theorem 8.1 — Equator Gap
For every prime p, the crossing M = 2p is a forced gap: gcd(p, 2p) = p > 1. Prime p is absent from ring 2p. The gap has width exactly 1: p appears on ring 2p−1 and on ring 2p+1, but not at M = 2p itself.
Proof.

gcd(p, 2p) = p > 1 since p | 2p. For the gap width: 2p−1 ≡ −1 (mod p) and 2p+1 ≡ 1 (mod p), so gcd(p, 2p−1) = gcd(p, 2p+1) = 1. The spiral crosses the equator invisibly — no prime ever occupies angle exactly 1/2. □

8.2 — Mirror Lift Theorem

Theorem 8.2 — Mirror Lift
For p in the top half (p+1 ≤ M ≤ 2p−1, gcd(p,M)=1):
gcd(M−p, M+1) = gcd(M−p, p+1)
The mirrors that lift are exactly R(p+1) ∩ {1,…,p−1} — the integers below p coprime to p+1.
Proof.

Set a = M−p and b = M+1. Then b−a = p+1. By the Euclidean property: gcd(a,b) = gcd(a, b−a) = gcd(M−p, p+1). As M ranges over {p+1,…,2p−1}, M−p ranges over {1,…,p−1} in order — a bijection. □

Corollary 8.2a — Encodes (Z/(p+1)Z)*
The lift pattern of the top-half mirrors is entirely determined by p+1. The mirrors that lift = #{k ∈ {1,…,p−1} : gcd(k,p+1)=1} = φ(p+1)−1. The top-half encodes the multiplicative group of p+1.

8.3 — Bottom Sector Theorem

Theorem 8.3 — Bottom Sector
In M ∈ [2p+1, 3p], prime p appears on exactly p−1 rings. As M increases from 2p+1 to 3p−1, M mod p cycles through (Z/pZ)* = {1,2,…,p−1} in exact sequential order. The sector between angles 1/2 and 1/3 is a complete geometric portrait of (Z/pZ)*.
Proof.

The interval [2p+1, 3p] has p integers; the unique multiple of p is 3p. So gcd(p,M) = 1 for exactly p−1 values. For M = 2p+j, j ∈ {1,…,p−1}: M mod p = j, cycling through every nonzero residue mod p in order. □

Corollary 8.3a — Exactly p−2 lifts in the bottom sector
Of the p−1 appearances, exactly p−2 result in successful lifts. The unique failure is M = 3p−1: since ring 3p is a gap (gcd(p,3p)=p), the successor is blocked. All other p−2 appearances lift.

8.4 — Upper Path Never Lifts

Theorem 8.4 — Upper Path Permanent Barrier
For every odd M ≥ 3, the residue r = (M+1)/2 is on ring M but never lifts:
gcd((M+1)/2, M+1) = (M+1)/2 ≥ 2
The path r = (M+1)/2 approaches the equator from above as M → ∞, and every point on it is permanently blocked.
Proof.

Let k = (M+1)/2. Then M+1 = 2k, so gcd(k, M+1) = gcd(k, 2k) = k ≥ 2 for M ≥ 3. That r is on ring M: gcd((M+1)/2, M) = 1 by a Euclidean reduction (M and (M+1)/2 differ by (M−1)/2; iterating gives gcd 1 since consecutive integers are coprime). □

8.5 — Lower Path Alternation and Gaussian Integers

Theorem 8.5 — Lower Path Alternation
For odd M ≥ 3, r = (M−1)/2 lifts to M+1 if and only if M ≡ 3 (mod 4):
gcd((M−1)/2, M+1) = gcd((M−1)/2, 2) = 1 ⟺ M ≡ 3 (mod 4)
Proof.

Set t = (M+1)/2, so M+1 = 2t and r = t−1. Then gcd(t−1, 2t) = gcd(t−1, 2) (since gcd(t−1, t) = 1 and 2t = 2·t). This equals 1 iff t−1 is odd iff t is even iff (M+1)/2 is even iff M ≡ 3 (mod 4). □

Corollary 8.5a — Gaussian Integer Detection
When M = q is an odd prime, r = (q−1)/2 lifts if and only if q is inert in Z[i] (q ≡ 3 mod 4). Equivalently, it is blocked if and only if q splits in the Gaussian integers as q = (a+bi)(a−bi) = a²+b².

Proof. By Theorem 8.5, r lifts iff q ≡ 3 (mod 4). By Fermat's theorem on sums of two squares, an odd prime q splits in Z[i] iff q ≡ 1 (mod 4). These conditions are complementary. □
Remark 8.5b — The geometric meaning
The lower equatorial residue r = (q−1)/2 is a single point on prime ring q. Its lift status encodes whether q is a sum of two squares. The Modular Lifting Ring system is the base; the Gaussian integers Z[i] are an extension whose arithmetic is readable through the lift structure at the equatorial boundary.
· · ✦ · ·

§9.Gap-Class Decomposition of ζ(2)

Euler's product ζ(2) = π²/6 = ∏p p²/(p²−1) converges absolutely, so its factors may be rearranged freely. Group primes by their forward gap — the distance to the next prime:

ζ(2) = ∏g Pg, where Pg = ∏{p : gap(p)=g} p²/(p²−1)

This is a reorganization of Euler's product, valid by absolute convergence. G = {1, 2, 4, 6, 8, …} is the set of all realized prime gap sizes. At N = 400 million, gap class 1 (the prime 2, fixed factor 4/3) contributes 57.8% of log ζ(2); gap class 2 (twin primes) contributes 34.8%. Together they exceed 92% of the total.

Theorem 9.1 — Gap-Class Factorization
ζ(2) = ∏g ∈ G Pg, where each Pg = ∏{p : gap(p)=g} p²/(p²−1) is the partial Euler product over primes with forward gap g.

This is an exact identity by absolute convergence of the Euler product. It is a structural decomposition, not a new theorem — its value is making the weight of each gap class directly measurable.
Corollary 9.2 — Hardy–Littlewood Ratio S(6)/S(2) = 2 (exact at M=30)
On the primitive polygon of mod 30, the gap-6 chord count is exactly twice the gap-2 chord count: S(6)/S(2) = 6/3 = 2. This is an exact integer combinatorial identity, not an asymptotic statement. It gives a finite geometric proof of the Hardy–Littlewood prediction that gap-6 primes are asymptotically twice as frequent as gap-2 (twin) primes.

Verification: At M=30, R(30) = {1,7,11,13,17,19,23,29}. Gap-2 admissible pairs: (11,13),(17,19),(29,31→1 wrap) = 3 pairs. Gap-6 admissible pairs: (1,7),(7,13),(11,17),(13,19),(17,23),(23,29) = 6 pairs. Ratio = 6/3 = 2. □
Open Question 9.3 — Finiteness Obstruction
If the twin prime set were finite, P2(s) = ∏finite ps/(ps−1) would be holomorphic on all of ℂ. The remaining sub-products would then have to account for the poles and zeros of ζ(s) entirely via the identity ζ(s) = ∏g Pg(s). Whether this is analytically compatible with the known zero structure of ζ(s) appears to be open. This is a structural observation — it does not prove or disprove the twin prime conjecture.

§10.Summary of Results

ResultStatementMethodStatus
Thm 2.1 Ck = ∏(p²−2)/(p²−1) · ∏(p−1)(p²+p−1)/p³ Local factor analysis Proved
Prop 3.1 Ck > 0 for all k ∑1/(p²−1) < ∞ Proved
Thm 4.1 Ck = Crad(k) Local factors depend only on p | k, not vp(k) Proved
Thm 5.1 rad(j) ⊊ rad(k) ⟹ Cj > Ck Lemma 5.2: g(p) = p³−p²−2p+1 > 0 Proved
Cor 5.1 Ck < C for all k ≥ 2 Special case of monotonicity Proved
Thm 6.1 C = ζ(2)·dFT = (π²/6)·∏(1−2/p²) p² cancels term by term Proved
Thm 7.1 T(p#) = φ(p#) ⟺ p#+1 prime Size argument + divisibility Proved
Cor 7.3 T(p) = φ(p+1)−1 for all primes p Counting coprime residues in {1,…,p−1} Proved
Thm 8.1 Equator gap at M = 2p, width = 1 gcd(p,2p) = p; adjacent rings coprime Proved
Thm 8.2 gcd(M−p, M+1) = gcd(M−p, p+1) One Euclidean step Proved
Thm 8.3 [2p+1,3p]: p−1 appearances, (Z/pZ)* in order Counting multiples of p Proved
Thm 8.4 r=(M+1)/2 never lifts for odd M ≥ 3 gcd((M+1)/2, M+1) = (M+1)/2 ≥ 2 Proved
Thm 8.5 r=(M−1)/2 lifts iff M ≡ 3 (mod 4) gcd((M−1)/2, 2) parity argument Proved
Cor 8.5a At prime q: lifts iff q inert in Z[i] Fermat sum-of-squares + Thm 8.5 Proved
Thm 9.1 ζ(2) = ∏g Pg — gap-class factorization Absolute convergence + rearrangement Proved (reorganization)
Cor 9.2 S(6)/S(2) = 2 exactly at M=30 Direct count on R(30) Proved (combinatorial)
Open 9.3 Finiteness obstruction: finite twin primes compatible with ζ(s) zeros? Structural observation Open
Open Infinitely many primorial primes ⟺ infinitely many perfect lift rings Restatement via Thm 7.1 Open

§11.References

[1] Hausman, M. and Shapiro, H. N. (1984). On the mean square distribution of primitive roots of unity. Comm. Pure Appl. Math. 26(4):539–547. [C appears implicitly as a ratio; never isolated or named.]

[2] Hardy, G. H. and Littlewood, J. E. (1923). Some problems of Partitio Numerorum III. Acta Math. 44:1–70. [Singular series S(k); parallel radical structure.]

[3] Feller, W. and Tornier, E. (1932). Mengentheoretische Untersuchungen von Zufälligkeiten. Math. Ann. 107:188–232. [dFT ≈ 0.32263634, OEIS A065474.]

[4] Mertens, F. (1874). Ein Beitrag zur analytischen Zahlentheorie. J. Reine Angew. Math. 78:46–62. [Third Mertens theorem; absolute convergence of the Euler product.]

[5] Euler, L. (1748). Introductio in analysin infinitorum. [ζ(2) = π²/6 via Euler product.]

[6] Finch, S. R. (2003). Mathematical Constants. Cambridge University Press. [C not present; searched 530711806246.]

[7] OEIS Foundation. oeis.org. [C not found under 530711806246 as of March 2026; submission pending.]

[8] Gauss, C. F. (1832). Theoria residuorum biquadraticorum. Göttingen. [Introduces Z[i]; Gaussian prime splitting, underlying Cor. 8.5a.]

[9] Fermat, P. (1640, letter to Mersenne). Sum of two squares theorem: q = a²+b² iff q ≡ 1 (mod 4) for odd prime q. [Used in Cor. 8.5a.]