betaFeedback

WorksMathematics

On Wagstaff primes in the kk-Pell number sequence

Abstract

A Wagstaff prime is a prime number of the form (2p+1)/3(2^{\mathfrak{p}}+1)/3, where p\mathfrak{p} is an odd prime. For an integer k≥2k\ge 2, the sequence of kk-generalized Pell numbers (Pn(k))n≥2−k\bigl(P_n^{(k)}\bigr)_{n\ge 2-k} is the linear recurrence of order kk given by Pn(k)=2Pn−1(k)+Pn−2(k)+⋯+Pn−k(k)P_n^{(k)}=2P_{n-1}^{(k)}+P_{n-2}^{(k)}+\cdots+P_{n-k}^{(k)} for n≥2n\ge 2, with initial values P2−k(k)=⋯=P0(k)=0P_{2-k}^{(k)}=\cdots=P_{0}^{(k)}=0 and P1(k)=1P_1^{(k)}=1. For k=2k=2 this is the classical Pell sequence. Motivated by the recent determination of the Wagstaff primes in the kk-Fibonacci and kk-Lucas sequences, we study the exponential Diophantine equation Pn(k)=(2p+1)/3P_n^{(k)}=(2^{\mathfrak{p}}+1)/3 in positive integers n,kn,k and an odd prime p≥3\mathfrak{p}\geq 3, with k≥2k\ge 2. We show that this equation has no solutions, that is no kk-Pell number is a Wagstaff prime. The proof combines linear forms in logarithms of algebraic numbers with reduction based on the theory of continued fractions, together with a computational verification in the remaining finite range.

Introduction

Background

The search for terms of a linear recurrence sequence that possess a prescribed arithmetic shape is a classical theme of Diophantine Number theory. Over the last two decades, the tools of Baker’s theory of linear forms in logarithms, in the explicit form due to [15] (see also [11]), combined with reduction methods such as the Baker–Davenport lemma in the version of [12] have made it possible to determine completely the members of families of sequences such as generalized Fibonacci and Lucas sequences that are, for instance, powers of two [10], repdigits [5], Mersenne numbers [6, 1], or Fermat numbers [7].

A Wagstaff prime is a prime number of the form

Wp=2p+13,W_p=\frac{2^p+1}{3},

where pp is an odd prime. The first few Wagstaff primes are 3, 11, 43, 683, 2731, 43691, …, corresponding to p=3,5,7,11,13,17,…p=3,5,7,11,13,17,\ldots (see sequence A000979 in [18]). They are named after Samuel S. Wagstaff Jr. and are closely tied to Mersenne primes through the New Mersenne Conjecture of [2]. It is not known whether there are infinitely many Wagstaff primes, and their primality has been studied both through compositeness tests [4] and through the search for probable primes [17].

Very recently, Wagstaff primes have been sought inside generalized Fibonacci-type sequences. For k≥2k\ge2, let (Fn(k))n≥2−k(F_n^{(k)})_{n\ge2-k} be the kk-generalized Fibonacci sequence, given by Fn(k)=Fn−1(k)+⋯+Fn−k(k)F_n^{(k)}=F_{n-1}^{(k)}+\cdots+F_{n-k}^{(k)} with initial values F2−k(k)=⋯=F0(k)=0F_{2-k}^{(k)}=\cdots=F_0^{(k)}=0 and F1(k)=1F_1^{(k)}=1. [16] solved the equation Fn(k)=(2p+1)/3F_n^{(k)}=(2^p+1)/3 and proved that the only Wagstaff prime appearing in any kk-Fibonacci sequence is 3, namely F4(2)=3F_{4}^{(2)} = 3. Shortly afterwards, [3] treated the kk-Lucas sequence (Ln(k))n≥2−k(L_n^{(k)})_{n\ge2-k}, whose initial values are L0(k)=2L_0^{(k)} = 2, L1(k)=1L_1^{(k)} = 1 and Lj(k)=0L_j^{(k)} = 0 for 2−k≤j≤−12-k \le j \le-1, and showed that the equation Ln(k)=(2p+1)/3L_n^{(k)} = (2p+1)/3 has only the solutions (n,k,p)∈{(5,2,5),(6,4,7)}∪{(2,k,3):k≥2}(n,k,p) \in\{(5,2,5),(6,4,7)\} \cup\{(2,k,3): k \ge2\}; in particular the Wagstaff primes 3, 11 and 43 occur in this family.

The present paper considers the same question for the kk-generalized Pell numbers, studied by [9]. For an integer k≥2k \ge2, this is the sequence (Pn(k))n≥2−k(P_n^{(k)})_{n\ge2-k} defined by the recurrence

Pn(k)=2Pn−1(k)+Pn−2(k)+⋯+Pn−k(k),n≥2,P_n^{(k)} = 2P_{n-1}^{(k)} + P_{n-2}^{(k)} + \cdots+ P_{n-k}^{(k)}, \qquad n \ge2,

with initial conditions P2−k(k)=P3−k(k)=⋯=P0(k)=0P_{2-k}^{(k)} = P_{3-k}^{(k)} = \cdots= P_0^{(k)} = 0 and P1(k)=1P_1^{(k)} = 1. For k=2k = 2 we recover the Pell numbers. Although the kk-Pell sequences resemble the kk-Fibonacci and kk-Lucas sequences, the argument cannot be transferred verbatim from the previous two papers.

Let pp be an odd prime. The main result of this paper is the following.

Main Result

Theorem 1.1. Let (Pn(k))n≥2−k(P_n^{(k)})_{n\ge2-k} be the sequence of kk-generalized Pell numbers. The Diophantine equation

Pn(k)=2p+13(1)P_n^{(k)} = \frac{2^p+1}{3} \tag*{(1)}

has no solutions in integers n≥1n \ge1, k≥2k \ge2 and primes p≥3p \ge3.

Methods

Preliminaries

The following identity, taken from [14], relates Pn(k)P_n^{(k)} to the mm-th Fibonacci number where m=2n−1m = 2n - 1. It states that

Pn(k)=F2n−1,(2)P_n^{(k)} = F_{2n-1}, \tag*{(2)}

for all 1≤n≤k+11 \le n \le k+1, while the next term satisfies Pk+2(k)=F2k+3−1P_{k+2}^{(k)} = F_{2k+3} - 1. The characteristic polynomial of the sequence (Pn(k))n≥2−k(P_n^{(k)})_{n\ge2-k} is

Ψk(x)=xk−2xk−1−⋯−x−1.\Psi_k(x) = x^k - 2x^{k-1} - \cdots- x - 1.

This polynomial is irreducible over Q[x]\mathbb{Q}[x] and has a unique real root α:=α(k)>1\alpha:= \alpha(k) > 1, with all other roots lying inside the unit circle, see [9]. Moreover, the root α\alpha satisfies

φ2(1−φ−k)<α<φ2,k≥2,where φ:=1+52,(3)\varphi^2(1-\varphi^{-k}) < \alpha< \varphi^2, \qquad k \ge2, \qquad\text{where } \varphi:= \frac{1+\sqrt{5}}{2}, \tag*{(3)}

as noted in [9]. Similarly to the case k=2k = 2, it has been shown in [9] that

αn−2≤Pn(k)≤αn−1,for all n≥1,k≥2.(4)\alpha^{n-2} \le P_n^{(k)} \le\alpha^{n-1}, \qquad\text{for all } n \ge1, k \ge2. \tag*{(4)}

For k≥2k \ge2, we define

fk(x):=x−1(k+1)x2−3kx+k−1=x−1k(x2−3x+1)+x2−1.f_k(x) := \frac{x-1}{(k+1)x^2-3kx+k-1} = \frac{x-1}{k(x^2-3x+1)+x^2-1}.

Bravo and Herrera [8] showed in Lemma 1 that

0.276<fk(α)<0.5and∣fk(αi)∣<1,0.276 < f_k(\alpha) < 0.5 \qquad\text{and} \qquad|f_k(\alpha_i)| < 1,

for all 2≤i≤k2 \le i \le k, where αi\alpha_i are the roots of Ψk(x)\Psi_k(x) inside the unit circle. This shows that fk(α)f_k(\alpha) is not an algebraic integer. Furthermore, Theorem 3.1 in [9] establishes

Pn(k)=∑i=1kfk(αi)αinand∣Pn(k)−fk(α)αn∣<12.(5)P_n^{(k)} = \sum_{i=1}^{k} f_k(\alpha_i)\alpha_i^n \qquad\text{and} \qquad \left|P_n^{(k)}-f_k(\alpha)\alpha^n\right| < \frac{1}{2}. \tag*{(5)}

for all k≥2k \ge2 and n≥2−kn \ge2-k. The first part of (5) provides a Binet-type formula for Pn(k)P_n^{(k)}, while the second inequality shows that the influence of the roots inside the unit circle is negligible.

A sharper estimate appears in Lemma 2 of [8], valid under a more restricted range of nn relative to kk. Specifically, if k≥30k \ge30 and n>1n > 1, then

∣fk(α)αn−φ2nφ+2∣<(φ2nφ+2)4φk/2,provided n<φk/2.(6)\left|f_k(\alpha)\alpha^n-\frac{\varphi^{2n}}{\varphi+2}\right| < \left(\frac{\varphi^{2n}}{\varphi+2}\right)\frac{4}{\varphi^{k/2}}, \qquad\text{provided } n < \varphi^{k/2}. \tag*{(6)}

Linear Forms in Logarithms

Let γ\gamma be an algebraic number of degree dd, and let its minimal polynomial over Z\mathbb{Z} be

u0xd+u1xd−1+⋯+ud=u0∏i=1d(x−γ(i)).u_0x^d+u_1x^{d-1}+\cdots+u_d=u_0\prod_{i=1}^{d}(x-\gamma^{(i)}).

where the coefficients uiu_i are coprime integers with u0>0u_0>0, and γ(i)\gamma^{(i)} denote the conjugates of γ\gamma. The logarithmic height of γ\gamma is defined as

h(γ)=1d(log⁡u0+∑i=1dlog⁡max⁡{∣γ(i)∣,1}).h(\gamma)=\frac{1}{d}\left(\log u_0+\sum_{i=1}^{d}\log\max\{|\gamma^{(i)}|,1\}\right).

In particular, for a rational number r/sr/s with gcd⁡(r,s)=1\gcd(r,s)=1 and s>0s>0, the logarithmic height is h(r/s)=log⁡max⁡{∣r∣,s}h(r/s)=\log\max\{|r|,s\}. The following standard properties of logarithmic heights will be applied without further reference:

(a) h(γ1±γ2)≤h(γ1)+h(γ2)+log⁡2h(\gamma_1\pm\gamma_2)\le h(\gamma_1)+h(\gamma_2)+\log2,

(b) h(γ1γ2±1)≤h(γ1)+h(γ2)h(\gamma_1\gamma_2^{\pm1})\le h(\gamma_1)+h(\gamma_2),

(c) h(γs)=∣s∣h(γ)h(\gamma^s)=|s|h(\gamma) for any integer ss.

Based on these properties, it follows from [9 Lemma 3] that

h(fk(α))<4klog⁡φ+klog⁡(k+1),for all k≥2.(7)h(f_k(\alpha))<4k\log\varphi+k\log(k+1), \qquad\text{for all } k\ge2. \tag*{(7)}

The following theorem, proved in [15] and cited in [11 Theorem 9.4], is fundamental in bounding variables in the Diophantine equation (1). Let KK be a number field of degree DD, and let γ1,…,γt∈K\gamma_1,\ldots,\gamma_t\in K be non-zero elements with integers b1,…,btb_1,\ldots,b_t. Define

B=max⁡{∣b1∣,…,∣bt∣},Γ=γ1b1⋯γtbt−1,B=\max\{|b_1|,\ldots,|b_t|\}, \qquad\Gamma=\gamma_1^{b_1}\cdots\gamma_t^{b_t}-1,
endsmid=1{ends_mid=1}

and let A1,…,AtA_1,\ldots,A_t be real numbers satisfying

Aj≥h′(γj):=max⁡{Dh(γj),∣log⁡γj∣,0.16},1≤j≤t.A_j \ge h'(\gamma_j) := \max\{D h(\gamma_j), |\log\gamma_j|, 0.16\}, \quad1 \le j \le t.

We call h′h' the modified height (with respect to the field K\mathbb{K}). With this notation, the main result in [15] implies the following estimate.

Theorem 2.1 (Theorem 9.4 in [11]). Assume that Γ\Gamma is non-zero and K\mathbb{K} is real. Then

log⁡∣Γ∣>−1.4⋅30t+3t4.5D2(1+log⁡D)(1+log⁡B)A1⋯At.\log|\Gamma| > -1.4 \cdot30^{t+3}t^{4.5}D^2(1+\log D)(1+\log B)A_1\cdots A_t.

The computations initially lead to upper bounds for the variables that are too large to be used directly. We therefore apply a Baker-Davenport type reduction to obtain smaller bounds. The details of this reduction procedure are presented in the following subsection. The precise variant we need is due to Dujella and Pethő [12].

Reduction method

For a real number rr we write ∥r∥\lVert r\rVert for the distance from rr to the nearest integer, i.e. ∥r∥:=min⁡{∣r−n∣:n∈Z}\lVert r\rVert:= \min\{|r-n|:n \in\mathbb{Z}\}.

Lemma 2.1 (Dujella, Pethő, [12]). Let τ≠0\tau\ne0, CC, β\beta and μ\mu be real numbers with C>0C > 0 and β>1\beta> 1. Let M>1M > 1 be a positive integer and suppose that r/sr/s is a convergent of the continued fraction expansion of τ\tau satisfying s>6Ms > 6M. Put

θ:=∥μs∥−M∥τs∥.\theta:= \lVert\mu s\rVert- M\lVert\tau s\rVert.

If θ>0\theta> 0, then the inequality

0<∣uτ−v+μ∣<Cβ−ℓ0 < |u\tau-v+\mu| < C\beta^{-\ell}

has no solution in positive integers u,v,ℓu,v,\ell with

log⁡(Cs/θ)log⁡β≤ℓandu≤M.\frac{\log(Cs/\theta)}{\log\beta} \le\ell \qquad\text{and}\qquad u \le M.

The following is an important fact from calculus that will be useful in the proof. If z∈Rz \in\mathbb{R} satisfies ∣z∣<12|z| < \frac{1}{2}, then

∣log⁡(1+z)∣=∣z−z22+z33−z44+⋯ ∣,<∣z∣+12(∣z∣2+∣z∣3+∣z∣4+⋯ )=∣z∣(1+∣z∣2(1−∣z∣))<32∣z∣.(8)\begin{aligned} |\log(1+z)| = \left|z-\frac{z^2}{2}+\frac{z^3}{3}-\frac{z^4}{4}+\cdots\right|, \\ &< |z|+\frac{1}{2}\left(|z|^2+|z|^3+|z|^4+\cdots\right) = |z|\left(1+\frac{|z|}{2(1-|z|)}\right)<\frac{3}{2}|z|. \tag*{(8)} \end{aligned}

Finally, we present an analytic argument which is Lemma 7 in [13].

Lemma 2.2 (Lemma 7 in [13]). If r≥1r \ge1, T>(4r2)rT > (4r^2)^r and T>x(log⁡x)rT > \frac{x}{(\log x)^r}, then

x<2rT(log⁡T)r.x < 2^rT(\log T)^r.

SageMath 10.6 is used to perform all computations in this work.

Proof of Theorem 1.1

Some necessary bounds

For every odd prime pp, equation (1) together with the lower estimate for Pn(k)P_n^{(k)} in terms of α\alpha gives

αn−2<Pn(k)=2p+13<2p+1<2p+1.\alpha^{n-2}<P_n^{(k)}=\frac{2^p+1}{3}<2^p+1<2^{p+1}.

Taking logarithms, we get (n−2)log⁡α<(p+1)log⁡2(n-2)\log\alpha< (p+1)\log2, so that n<2p+4n < 2p+4. In the other direction, the upper estimate for Pn(k)P_n^{(k)} gives

αn−1≥Pn(k)=2p+13>2p3>2p−2,\alpha^{n-1} \ge P_n^{(k)} = \frac{2^p+1}{3} > \frac{2^p}{3} > 2^{p-2},

and taking logarithms once more yields (n−1)log⁡α≥(p−2)log⁡2(n-1)\log\alpha\ge(p-2)\log2, since log⁡α<2log⁡2\log\alpha< 2\log2 from α<φ2<4\alpha< \varphi^2 < 4, we get n>p/2n > p/2. Putting the two estimates together,

p2<n<2p+4.(9)\frac{p}{2} < n < 2p+4. \tag*{(9)}

The case n≤k+1n \le k+1

Proposition 3.1. Let k≥2k \ge2, 1≤n≤k+11 \le n \le k+1 and let p≥3p \ge3 be an integer. Then Pn(k)≠(2p+1)/3P_n^{(k)} \ne(2^p+1)/3.

Proof. By (2), Pn(k)=F2n−1P_n^{(k)} = F_{2n-1}. Put W=(2p+1)/3W=(2^p+1)/3. Since p≥3p \ge3, 8∣2p8 \mid2^p, so 3W=2p+1≡1(mod8)3W=2^p+1 \equiv1 \pmod{8}. As 3⋅3≡1(mod8)3\cdot3 \equiv1 \pmod{8}, we get W≡3(mod8)W \equiv3 \pmod{8}. The Fibonacci sequence modulo 8 is periodic with period 12; one period is 0,1,1,2,3,5,0,5,5,2,7,10,1,1,2,3,5,0,5,5,2,7,1 (indices m=0,…,11m=0,\ldots,11), so Fm≡3(mod8)F_m \equiv3 \pmod{8} only if m≡4(mod12)m \equiv4 \pmod{12}, in particular only if mm is even. Since 2n−12n-1 is odd, F2n−1≢3(mod8)F_{2n-1} \not\equiv3 \pmod{8}, and therefore F2n−1≠WF_{2n-1} \ne W.

Remark 3.1. For k=2k=2 the Pell numbers satisfy Pn(2) mod 8∈{0,1,2,5,4,5,6,1}P_n^{(2)} \bmod8 \in\{0,1,2,5,4,5,6,1\} (period 8) and are never ≡3(mod8)\equiv3 \pmod{8}; so Theorem 1.1 is already complete for k=2k=2 by the same argument, without any assumption on nn. For k≥3k \ge3 the residue 3 does occur, so the analytic argument of Subsection 3.3 is needed.

We therefore have that equation (1) has no solution with n≤k+1n \le k+1 and k≥2k \ge2, and from now on we assume n≥k+2n \ge k+2.

The case n≥k+2n \ge k+2

Here n≥k+2n \ge k+2, and since k≥2k \ge2 this forces n≥4n \ge4. We first bound nn in terms of kk.

Lemma 3.1. Let (n,k,p)(n,k,p) be an integer solution of (1) with n≥k+2n \ge k+2, k≥2k \ge2 and p≥3p \ge3 prime, then

n<1.3⋅1015k6(log⁡k)3.(10)n < 1.3\cdot10^{15} k^6(\log k)^3. \tag*{(10)}

Proof. We return to (1) and use the closed formula (5). Write

Pn(k)−fk(α)αn=ek(n),P_n^{(k)}-f_k(\alpha)\alpha^n=e_k(n),

such that ∣ek(n)∣<1/2|e_k(n)|<1/2. Then

2p+13−fk(α)αn=ek(n),\frac{2^p+1}{3}-f_k(\alpha)\alpha^n=e_k(n),

thus

2p3−fk(α)αn=ek(n)−13.\frac{2^p}{3}-f_k(\alpha)\alpha^n=e_k(n)-\frac{1}{3}.

Taking absolute values on both sides yields

∣2p3−fk(α)αn∣≤∣ek(n)∣+13<12+13=56.\left|\frac{2^p}{3}-f_k(\alpha)\alpha^n\right|\le|e_k(n)|+\frac{1}{3}<\frac{1}{2}+\frac{1}{3}=\frac{5}{6}.

Thus we have

∣2p3−fk(α)αn∣<56.\left|\frac{2^p}{3}-f_k(\alpha)\alpha^n\right|<\frac{5}{6}.

Dividing through by fk(α)αnf_k(\alpha)\alpha^n yields

∣2p(3fk(α)αn)−1−1∣<56fk(α)αn<4αn.\left|2^p(3f_k(\alpha)\alpha^n)^{-1}-1\right|<\frac{5}{6f_k(\alpha)\alpha^n}<\frac{4}{\alpha^n}.

where we have used the fact that fk(α)>0.276f_k(\alpha)>0.276.

Set

∣Γ1∣:=∣2p(3fk(α)αn)−1−1∣<4αn.(11)|\Gamma_1|:=\left|2^p\left(3f_k(\alpha)\alpha^n\right)^{-1}-1\right|<\frac{4}{\alpha^n}. \tag*{(11)}

Clearly, Γ1≠0\Gamma_1\ne0 otherwise we would have 2p=3fk(α)αn2^p=3f_k(\alpha)\alpha^n. Applying a Galois automorphism σ\sigma of the Galois closure of Q(α)\mathbb{Q}(\alpha), with σ(α)=αi\sigma(\alpha)=\alpha_i for some i>1i>1, and taking absolute values on both sides, we obtain

2p=3∣fk(αi)αin∣.2^p=3\left|f_k(\alpha_i)\alpha_i^n\right|.

This is a contradiction, since the left-hand side is at least 8 while the right-hand side is less than 3.

Remark 3.2. The above argument is valid for every p≥3p\ge3.

To apply Theorem 2.1, we work in the field K:=Q(α)\mathbb{K}:=\mathbb{Q}(\alpha), so D=kD=k, and we take t=3t=3 with

γ1:=2,γ2:=3fk(α),γ3:=α,b1:=p,b2:=−1,b3:=−n.\begin{aligned} \gamma_1&:=2, & \gamma_2&:=3f_k(\alpha), & \gamma_3&:=\alpha,\\ b_1&:=p, & b_2&:=-1, & b_3&:=-n. \end{aligned}

We have h(γ1)=h(2)=log⁡2h(\gamma_1)=h(2)=\log2,

h(γ3)=h(α)<log⁡αk<2log⁡φk<1k,h(\gamma_3)=h(\alpha)<\frac{\log\alpha}{k}<\frac{2\log\varphi}{k}<\frac{1}{k},

and for γ2\gamma_2,

h(γ2)=h(3fk(α)),≤h(3)+h(fk(α)),<log⁡3+4klog⁡φ+klog⁡(k+1),<6klog⁡k.\begin{aligned} h(\gamma_2)&=h(3f_k(\alpha)),\\ &\le h(3)+h(f_k(\alpha)),\\ &<\log3+4k\log\varphi+k\log(k+1),\\ &<6k\log k. \end{aligned}

We therefore take A1:=klog⁡2A_1:=k\log2, A2:=6k2log⁡kA_2:=6k^2\log k and A3:=kA_3:=k. It remains to choose B≥max⁡{∣bi∣:i=1,2,3}B\ge\max\{|b_i|:i=1,2,3\}; by (9) we take B:=2nB:=2n.

Theorem 2.1 now gives

log⁡∣Γ1∣>−1.4⋅306⋅34.5⋅k2(1+log⁡k)(1+log⁡(2n))⋅klog⁡2⋅6k2log⁡k⋅k,>−5.4⋅1012k6(log⁡k)2log⁡n,(12)\begin{aligned} \log|\Gamma_1|&>-1.4\cdot30^6\cdot3^{4.5}\cdot k^2(1+\log k)(1+\log(2n))\cdot k\log2\cdot6k^2\log k\cdot k,\\ &>-5.4\cdot10^{12}k^6(\log k)^2\log n, \tag*{(12)} \end{aligned}

where we have used the fact that 1+log⁡k≤3log⁡k1+\log k\le3\log k and

1+log⁡(2n)=1+log⁡2+log⁡n=(1log⁡n+log⁡2log⁡n+1)log⁡n≤3log⁡n,\begin{aligned} 1+\log(2n)&=1+\log2+\log n\\ &=\left(\frac{1}{\log n}+\frac{\log2}{\log n}+1\right)\log n\\ &\le3\log n, \end{aligned}

for n≥k+2n\ge k+2, and k≥2k\ge2. Comparing (11) with (12), we obtain

nlog⁡α−log⁡4<5.4⋅1012k6(log⁡k)2log⁡n,n\log\alpha-\log4<5.4\cdot10^{12}k^6(\log k)^2\log n,

and therefore

n<1.2⋅1013k6(log⁡k)2log⁡n.n<1.2\cdot10^{13}k^6(\log k)^2\log n.

Proceeding via Lemma 2.2, fix x:=nx:=n, r:=1r:=1 and T:=1.2⋅1013k6(log⁡k)2T:=1.2\cdot10^{13}k^6(\log k)^2. Since T>x/(log⁡x)rT>x/(\log x)^r and T>(4r2)rT>(4r^2)^r, we have that

n<2(1.2⋅1013k6(log⁡k)2)log⁡(1.2⋅1013k6(log⁡k)2),<1.3⋅1015k6(log⁡k)3.\begin{aligned} n&<2(1.2\cdot10^{13}k^6(\log k)^2)\log(1.2\cdot10^{13}k^6(\log k)^2),\\ &<1.3\cdot10^{15}k^6(\log k)^3. \end{aligned}

This completes the proof of Lemma 3.1. □

We proceed to examine the two cases basing on kk.

The case k≤500k \le500

Suppose that k≤500k \le500. By Lemma 3.1, we have

n<1.3⋅1015k6(log⁡k)3<4.9⋅1033.n < 1.3 \cdot10^{15} k^6(\log k)^3 < 4.9 \cdot10^{33}.

We now refine the upper bound for nn. For this purpose, we return to (11) and write

Γ1=2p(3fk(α)αn)−1−1=eΛ1−1.\Gamma_1 = 2^{\mathfrak{p}}(3f_k(\alpha)\alpha^n)^{-1}-1=e^{\Lambda_1}-1.

We have already established that Γ1≠0\Gamma_1 \ne0, and hence Λ1≠0\Lambda_1 \ne0. Assume that n≥3n \ge3. Then

∣eΛ1−1∣=∣Γ1∣<0.5,\left|e^{\Lambda_1}-1\right|=|\Gamma_1|<0.5,

and inequality (2.7) gives ∣log⁡(1+Γ1)∣<1.5∣Γ1∣|\log(1+\Gamma_1)|<1.5|\Gamma_1|. Therefore, we obtain

Λ1=plog⁡2−nlog⁡α−log⁡(3fk(α)),\Lambda_1=\mathfrak{p}\log2-n\log\alpha-\log(3f_k(\alpha)),

which can be rearranged to

∣nlog⁡αlog⁡2−p+log⁡(3fk(α))log⁡2∣<6log⁡2⋅α−n.\left|n\frac{\log\alpha}{\log2}-\mathfrak{p}+\frac{\log(3f_k(\alpha))}{\log2}\right|<\frac{6}{\log2}\cdot\alpha^{-n}.

Applying Lemma 2.1 with the quantities

τ:=log⁡αlog⁡2,μ:=log⁡(3fk(α))log⁡2,C:=6log⁡2,β:=α.\tau:=\frac{\log\alpha}{\log2},\qquad \mu:=\frac{\log(3f_k(\alpha))}{\log2},\qquad C:=\frac{6}{\log2},\qquad \beta:=\alpha.

Here u=nu=n and v=pv=\mathfrak{p}. We set M:=9.8⋅1033M:=9.8\cdot10^{33}, which is an upper bound for 2n2n. For each k∈[2,500]k\in[2,500] we computed α=α(k)\alpha=\alpha(k), τ\tau and μ\mu to high precision and took the first convergent r/sr/s of τ\tau with s>6Ms>6M and θ>0\theta>0. If a convergent gave θ≤0\theta\le0, we passed to the next one. For k=2k=2 this is the convergent s68s_{68}, with s≈7.0157⋅1034s\approx7.0157\cdot10^{34} and θ=0.00202508\theta=0.00202508, which gives

n<log⁡((6/log⁡2)s/θ)log⁡α<100.52<101.n<\frac{\log((6/\log2)s/\theta)}{\log\alpha}<100.52<101.

Over all k∈[2,500]k\in[2,500] we found θ>4.2⋅10−4\theta>4.2\cdot10^{-4}, the smallest value occurring at k=105k=105. Writing HkH_k for the resulting bound on nn, the interval [k+2,[Hk]][k+2,[H_k]] is non-empty only for k≤86k\le86. It therefore remains to check the equation Pn(k)=(2p+1)/3P_n^{(k)}=(2^{\mathfrak{p}}+1)/3 for 2≤k≤862\le k\le86, n∈[k+2,101]n\in[k+2,101] and p∈[3,2n]\mathfrak{p}\in[3,2n]. A search in SageMath 10.6 found no solutions. We therefore turn to the case k>500k>500.

The case k>500k>500

Suppose that k>500k>500. By Lemma 3.1, we have

n<1.3⋅1015k6(log⁡k)3<φk/2,n<1.3\cdot10^{15}k^6(\log k)^3<\varphi^{k/2},

where the last inequality holds for all k>500k>500. Thus, n<φk/2n<\varphi^{k/2}, which allows us to apply inequality (2.5). Combining this with (2.4), we obtain

∣2p+13−φ2nφ+2∣=∣Pn(k)−φ2nφ+2∣,≤∣Pn(k)−fk(α)αn∣+∣fk(α)αn−φ2nφ+2∣,<12+(φ2nφ+2)4φk/2.\begin{aligned} \left|\frac{2^{\mathfrak{p}}+1}{3}-\frac{\varphi^{2n}}{\varphi+2}\right| =\left|P_n^{(k)}-\frac{\varphi^{2n}}{\varphi+2}\right|, \\ &\le\left|P_n^{(k)}-f_k(\alpha)\alpha^n\right|+\left|f_k(\alpha)\alpha^n-\frac{\varphi^{2n}}{\varphi+2}\right|, \\ &<\frac{1}{2}+\left(\frac{\varphi^{2n}}{\varphi+2}\right)\frac{4}{\varphi^{k/2}}. \end{aligned}

Therefore

∣2p3−φ2nφ+2∣<12+(φ2nφ+2)⋅4φk/2+13,=56+(φ2nφ+2)⋅4φk/2.\begin{aligned} \left|\frac{2^{p}}{3}-\frac{\varphi^{2n}}{\varphi+2}\right|<\frac{1}{2}+\left(\frac{\varphi^{2n}}{\varphi+2}\right)\cdot\frac{4}{\varphi^{k/2}}+\frac{1}{3}, \\ &=\frac{5}{6}+\left(\frac{\varphi^{2n}}{\varphi+2}\right)\cdot\frac{4}{\varphi^{k/2}}. \end{aligned}

Dividing through by (φ2nφ+2)\left(\frac{\varphi^{2n}}{\varphi+2}\right) yields

∣2p3⋅(φ+2φ2n)−1∣<56⋅(φ+2φ2n)+4φk/2<4φ2n+4φk/2.\left|\frac{2^{p}}{3}\cdot\left(\frac{\varphi+2}{\varphi^{2n}}\right)-1\right|<\frac{5}{6}\cdot\left(\frac{\varphi+2}{\varphi^{2n}}\right)+\frac{4}{\varphi^{k/2}}<\frac{4}{\varphi^{2n}}+\frac{4}{\varphi^{k/2}}.

Since 2n>n≥k+2>k/22n>n\geq k+2>k/2, then min⁡{φ2n,φk/2}=φk/2\min\{\varphi^{2n},\varphi^{k/2}\}=\varphi^{k/2} and thus we get

∣2p3⋅(φ+2φ2n)−1∣<8φk/2.\left|\frac{2^{p}}{3}\cdot\left(\frac{\varphi+2}{\varphi^{2n}}\right)-1\right|<\frac{8}{\varphi^{k/2}}.

Fix

∣Γ2∣:=∣2p3⋅(φ+2φ2n)−1∣<8φk/2.(13)|\Gamma_{2}|:=\left|\frac{2^{p}}{3}\cdot\left(\frac{\varphi+2}{\varphi^{2n}}\right)-1\right|<\frac{8}{\varphi^{k/2}}. \tag*{(13)}

Clearly Γ2≠0\Gamma_{2}\ne0 otherwise we would have

2p(φ+23)=φ2n,2^{p}\left(\frac{\varphi+2}{3}\right)=\varphi^{2n},

taking Norms NK/QN_{\mathbb{K}/\mathbb{Q}} on both sides for K=Q(5)\mathbb{K}=\mathbb{Q}(\sqrt{5}), we have

N(2p(φ+23))=N(φ2n).N\left(2^{p}\left(\frac{\varphi+2}{3}\right)\right)=N(\varphi^{2n}).

Using the fact that N(φ+2)=5N(\varphi+2)=5 and N(φ2n)=(φφ′)2n=1N(\varphi^{2n})=(\varphi\varphi')^{2n}=1, we have that

22p⋅59=1,\frac{2^{2p}\cdot5}{9}=1,

which is impossible for p≥3p\geq3.

Taking the algebraic Number field K=Q(5)\mathbb{K}=\mathbb{Q}(\sqrt{5}), we have D:=2,t=3D:=2,t=3 and

γ1:=2,γ2:=φ,γ3:=φ+23,\gamma_{1}:=2,\qquad\gamma_{2}:=\varphi,\qquad\gamma_{3}:=\frac{\varphi+2}{3},
b1:=p,b2:=−2n,b3:=1.b_{1}:=p,\qquad b_{2}:=-2n,\qquad b_{3}:=1.

Moreover, we have

h(γ1)=h(2)=log⁡2,h(γ2)=h(φ)=12log⁡φ,h(\gamma_{1})=h(2)=\log2,\qquad h(\gamma_{2})=h(\varphi)=\frac{1}{2}\log\varphi,

and

h(γ3)=h(φ+23)<12log⁡φ+log⁡12.h(\gamma_{3})=h\left(\frac{\varphi+2}{3}\right)<\frac{1}{2}\log\varphi+\log12.

Therefore, we take

A1:=2log⁡2,A2:=log⁡φ,A3:=log⁡φ+2log⁡12.A_{1}:=2\log2,\qquad A_{2}:=\log\varphi,\qquad A_{3}:=\log\varphi+2\log12.

Also, since B≥max⁡{∣bi∣:i=1,2,3}B\geq\max\{|b_{i}|:i=1,2,3\}, we take B:=2nB:=2n as before. Proceeding via Theorem 2.1 gives

log⁡∣Γ2∣>−1.4⋅306⋅34.5⋅22(1+log⁡2)(1+log⁡2n)(2log⁡2)(log⁡φ)(log⁡φ+2log⁡12)>−7.1⋅1012log⁡n.(14)\begin{aligned} \log|\Gamma_{2}|>-1.4\cdot30^{6}\cdot3^{4.5}\cdot2^{2}(1+\log2)(1+\log2n)(2\log2)(\log\varphi)(\log\varphi+2\log12) \\ &>-7.1\cdot10^{12}\log n. \tag*{(14)} \end{aligned}

Comparing (13) to (14), we get

k2log⁡φ−log⁡8<7.1⋅1012log⁡n,\frac{k}{2}\log\varphi-\log8<7.1\cdot10^{12}\log n,

thus we obtain the bound

k<3.03×1013log⁡n.(15)k<3.03\times10^{13}\log n. \tag*{(15)}

Recall the bound for nn in inequality (10). Substituting this bound in (15) yields a new bound for kk as

k<3.03⋅1013log⁡(1.3⋅1015k6(log⁡k)3)<4.5⋅1014log⁡k.(16)k<3.03\cdot10^{13}\log\left(1.3\cdot10^{15}k^{6}(\log k)^{3}\right)<4.5\cdot10^{14}\log k. \tag*{(16)}

Proceeding via Lemma 2.2, we fix x:=kx:=k, r:=1r:=1 and T:=4.5⋅1014T:=4.5\cdot10^{14}. We therefore have that T>x/(log⁡x)rT>x/(\log x)^{r}, and T>(4r2)rT>(4r^{2})^{r} which yields

k<2(4.5⋅1014)(log⁡(4.5⋅1014))<3.1⋅1016.(17)k<2(4.5\cdot10^{14})(\log(4.5\cdot10^{14}))<3.1\cdot10^{16}. \tag*{(17)}

Applying this new bound for kk into inequality (10) gives a new bound for nn as n<6.4⋅10118n<6.4\cdot10^{118}. We proceed to reduce these bounds. To do this, we revisit (13) and rewrite it as

∣Γ2∣:=∣2p⋅φ−2n⋅(φ+23)−1∣<8φk/2.\lvert\Gamma_{2}\rvert:=\left\lvert2^{p}\cdot\varphi^{-2n}\cdot\left(\frac{\varphi+2}{3}\right)-1\right\rvert<\frac{8}{\varphi^{k/2}}.

We already showed that Γ2≠0\Gamma_{2}\ne0, it follows that Λ2≠0\Lambda_{2}\ne0. Assuming k≥12k\ge12, then ∣Γ2∣<8φ−k/2<0.5\lvert\Gamma_{2}\rvert<8\varphi^{-k/2}<0.5, we obtain the inequality ∣eΛ2−1∣=∣Γ2∣<0.5\lvert e^{\Lambda_{2}}-1\rvert=\lvert\Gamma_{2}\rvert<0.5 and by inequality (8) we get ∣log⁡(1+Γ2)∣<1.5∣Γ2∣\lvert\log(1+\Gamma_{2})\rvert<1.5\lvert\Gamma_{2}\rvert. Consequently, we derive

∣plog⁡2−nlog⁡φ2+log⁡(φ+23)∣<12φk/2.\left\lvert p\log2-n\log\varphi^{2}+\log\left(\frac{\varphi+2}{3}\right)\right\rvert<\frac{12}{\varphi^{k/2}}.

which when rearranged yields

∣plog⁡2log⁡φ2−n+log⁡(φ+23)log⁡φ2∣<12log⁡φ2⋅φ−k/2.\left\lvert p\frac{\log2}{\log\varphi^{2}}-n+\frac{\log\left(\frac{\varphi+2}{3}\right)}{\log\varphi^{2}}\right\rvert<\frac{12}{\log\varphi^{2}}\cdot\varphi^{-k/2}.

Applying Lemma 2.1 with the quantities

τ:=log⁡2log⁡φ2,μ:=log⁡(φ+23)log⁡φ2,C:=12log⁡φ2,β:=φ.\tau:=\frac{\log2}{\log\varphi^{2}},\qquad\mu:=\frac{\log\left(\frac{\varphi+2}{3}\right)}{\log\varphi^{2}},\qquad C:=\frac{12}{\log\varphi^{2}},\qquad\beta:=\sqrt{\varphi}.

Let τ=[a0;a1,a2,⋯ ]=[0;1,2,1,1,2,1,6,1,8,⋯ ]\tau=[a_{0};a_{1},a_{2},\cdots]=[0;1,2,1,1,2,1,6,1,8,\cdots] be the continued fraction expansion of τ\tau. Here u=pu=p, v=nv=n and ℓ=k\ell=k. We set M:=1.28⋅10119M:=1.28\cdot10^{119}, which is an upper bound for 2n2n. The first convergent of τ\tau with s>6Ms>6M and θ>0\theta>0 is s234s_{234}, with s≈1.22⋅10120s\approx1.22\cdot10^{120} and θ≈0.3202\theta\approx0.3202. Hence

k<2log⁡((12/log⁡φ2)s/θ)log⁡φ<1165.(18)k<\frac{2\log\left((12/\log\varphi^{2})^{s/\theta}\right)}{\log\varphi}<1165. \tag*{(18)}

We repeat this reduction procedure but with a new value of M:=2.3⋅1036M:=2.3\cdot10^{36} evaluated for the reduced bound kk in (18) and find that s=s78>6Ms=s_{78}>6M with θ=0.23119\theta=0.23119. This gives k<379k<379 which contradicts our original assumption that k>500k>500 and thus the proof is complete.

Acknowledgments

The author gratefully acknowledges Makerere University for providing a conducive environment for this research. This work is dedicated to the memory of Professor Florian Luca (1969–2026), whose remarkable contributions to Number Theory profoundly inspired the author’s research in Diophantine equations and recurrence sequences. It is offered as a modest tribute to his enduring mathematical legacy.

Disclosure statement

The author reports no conflict of interest.

References

Addresses

{sup}1` Department of Mathematics, Makerere University, Kampala, Uganda.

Email: [email protected]

References

  1. [1]Altassan, A., & Alan, M. (2024). Mersenne numbers in generalized Lucas sequences. Comptes rendus de l’Académie bulgare des Sciences, 77(1), 3–10.DOI
  2. [2]Bateman, P. T., Selfridge, J. L., & Wagstaff, S. S., Jr. (1989). The new Mersenne conjecture. The American Mathematical Monthly, 96(2), 125–128.
  3. [3]Batte, H. (2026). On Wagstaff primes in the k-Lucas number sequence. arXiv preprint arXiv:2602.20387. Accepted for publication in Publicationes Mathematicae.arxiv.org/abs/2602.20387
  4. [4]Berrizbeitia, P., Luca, F., & Melham, R. (2010). On a compositeness test for (2ᵖ + 1)/3. Journal of Integer Sequences, 13(1), Article 10.1.7.
  5. [5]Bravo, J. J., & Luca, F. (2014). Repdigits in k-Lucas sequences. Proceedings–Mathematical Sciences, 124, 141–154.DOI
  6. [6]Bravo, J. J., & Gómez, C. A. (2016). Mersenne k-Fibonacci numbers. Glasnik Matematički, 51(2), 307–319.DOI
  7. [7]Bravo, J. J., & Herrera, J. L. (2020). Fermat k-Fibonacci and k-Lucas numbers. Mathematica Bohemica, 145(1), 19–32.
  8. [8]Bravo, J. J., & Herrera, J. L. (2020). Repdigits in generalized Pell sequences. Archivum Mathematicum, 56(4), 249–262.DOI
  9. [9]Bravo, J. J., Herrera, J. L., & Luca, F. (2021). On a generalization of the Pell sequence. Mathematica Bohemica, 146(2), 199–213.DOI
  10. [10]Bravo, J. J., & Luca, F. (2012). Powers of two in generalized Fibonacci sequences. Revista Colombiana de Matemáticas, 46(1), 67–79.
  11. [11]Bugeaud, Y., Mignotte, M., & Siksek, S. (2006). Classical and modular approaches to exponential Diophantine equations I. Fibonacci and Lucas perfect powers. Annals of Mathematics, 163(3), 969–1018. https://doi.org/10.4007/annals.2006.163.969
  12. [12]Dujella, A., & Pethő, A. (1998). A generalization of a theorem of Baker and Davenport. Quarterly Journal of Mathematics, 49(195), 291–306.DOI
  13. [13]Gúzman–Sanchez, S., & Luca, F. (2014). Linear combinations of factorials and S-units in a binary recurrence sequence. Annales Mathématiques du Québec 38, 169–188.
  14. [14]Kılıç, E. (2008). On the usual Fibonacci and generalized order-k Pell numbers. Ars Combinatoria, 88, 33–45.
  15. [15]Matveev, E. M. (2000). An explicit lower bound for a homogeneous rational linear form in logarithms of algebraic numbers. Izvestiya: Mathematics, 64(6), 1217–1269.
  16. [16]Rezaiguia, L., Qawaqneh, H., & Abdelouahab, M. S. (2026). On the Wagstaff prime numbers in k-Fibonacci sequences. Statistics, Optimization & Information Computing, 15(1), 396–406. https://doi.org/10.19139/soic-2310-5070-2819
  17. [17]Sikkema, D. G. (2024). On Probable Primes of Wagstaff Numbers [Bachelor’s thesis, University of Groningen]. https://fse.studenttheses.ub.rug.nl/33908/1/bMAT2024SikkemaDG.pdf
  18. [18]Sloane, N. J. A. (n.d.). The On-Line Encyclopedia of Integer Sequences, Sequence A000979. Retrieved September 19, 2026, from https://oeis.org/A000979

Paper details

vlnc:W_mudyk4zq
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents