Powers in the Lucas sequence when the index is divisible by three
Mª Dolores Lópeza,, Javier Rodrigob
a
Departamento de Matemática e Informática Aplicadas a la Ingeniería Civil de la E.T.S.I. Caminos, Canales y Puertos, Universidad Politécnica de Madrid marilo.lopez@ upm.es
b
Departamento de Matemática Aplicada, E.T.S. de Ingeniería, Universidad Pontificia Comillas de Madrid
*Corresponding author: Mª Dolores López, Departamento de Matemática e Informática Aplicadas a la Ingeniería Civil de la E.T.S.I. Caminos, Canales y Puertos, Universidad Politécnica de Madrid, c/ Profesor Aranguren s/n, Madrid 28040, Spain
Abstract:
In this paper the m-powers with m≥2 included in the Lucas sequences when the index satisfies some conditions are found.
1. Introduction
The Lucas and companion Lucas sequences have been widely studied in the scientific literature as a natural extension of the so called Fibonacci sequence. They are defined inductively by: 0, =0
Q P
U , 11 , =
Q P
U , nPQ
Q P n Q
P
n PU QU
U ,2
, 1 ,
− − −
= and 0 , =2
Q P
V , P
V1P,Q = , VnP ,Q =PVnP−,1Q −QVnP−,2Q respectively, where P , are integer numbers (the Q Fibonacci sequence is the particular case of UnP,Q with P=1, Q=−1. The indices
Q
P , will be omitted when they are not necessary).
Concretely, the search of the powers included in these sequences has deserved a special attention by the number theorist. The main results at this respect for Un are:
- For odd relative primes P , , the only square in Q Un with n>2, n≠6 is 144
1 , 1 12 =
−
U
(see Ribenboim, Bremner&Tzanakis)
- For P divisible by 4, there is no squares in UnP,−1 for positive, even numbers n when
additional conditions on P are considered (Kagawa) - The only powers in 1 ,−1
n
U with n>2 are 1 ,1 8
6 =
−
U and 1 ,1 144 12 =
−
U (Bugeaud)
- The only power in Un2,−1 with n>1 is U72 ,−1 =169 (Cohn)
In this paper the case P>0, Q=−1 is studied: it is proved that, for n divisible by three and satisfying some extra conditions, the powers in Un are always attained for
1
=
P and they are 16 ,1 =8
−
U and 121 ,1 =144
−
U . To do this, the diophantine equation
m P
n u
U ,−1 = with unknowns
(
n, P,u,m)
, m≥2, is related with these diophantine equations according to the different cases:- The Catalan equation xm −y2 =1 - The Thue equation 2 xm −ym =1 - The equation 2x2 −y4 =1
- ⎟⎟
⎠ ⎞ ⎜⎜ ⎝ ⎛
k n
for the binomial coefficient.
- ep
( )
n for the exponent of the prime p in the factorization in prime numbers of the integer number n (if p is not a prime factor of n, it is assumed that ep( )
n =0).The following two identities for Un, Vn , Uk, Vk are used along the paper (see
Kagawa):
(
) (
)
kn
k k n
n k n
P U V P
U
V 4 4
2 −1 + 2 + = + 2 + when k n and k, n are positive numbers (1)
( ) (
)( )
( )
kk
k P U
V 2 − 2 +4 2 =4 −1 (2)
The paper is structured as follows: in section 2 the main result is stated and in section 3 some concluding remarks are given. Along section 2, k and n are positive integer numbers.
2. Development
To prove the result mentioned in the introduction, a lemma is needed: Lemma
If k n, then
k n U U U
GCD k
k n
, ⎟⎟
⎠ ⎞ ⎜⎜
⎝ ⎛
Proof
First it is noted that if k n, then Uk Un (see for example Ribenboim), so
k n
U U
is an integer number and the statement makes sense. Now, the following cases appear:
1) If Vk is an odd number, then by expanding (1) it is obtained that:
(
)
( )
⎟⎟⎠ ⎞ ⎜
⎜ ⎝ ⎛
+ =
+ + ⎟
⎟ ⎟
⎠ ⎞
⎜ ⎜ ⎜
⎝ ⎛ +
= − − −
−
k k
n
k k k
k n
k k
k n
k n
k n
U f V k n U P
U V k n U V k n
U 1 3 3 2 1
1
... 4 3
( )
xf is a polynomial without independent term. Now, since k k
k n U U U U
GCD , ⎟⎟
⎠ ⎞ ⎜⎜ ⎝ ⎛ ,
( )
1 1 2 , − − + = ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n k k n k k n k k n U f V k n U U U U UGCD , then ⎟⎟
⎠ ⎞ ⎜⎜ ⎝ ⎛ k k n U U U
GCD , divides to Uk,
( )
k kn
k f U
V k
n −1+
, so it divides to −1 +
( ) ( )
− = k−1 n k k k k n k V k n U f U f V k n .Since Vk is an odd number, it is satisfied that GCD
(
Vk,Uk)
=1 by identity (2) and thenk n U U U GCD k k n , ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ as desired.
2) If Vk and P are even numbers, then (1) and (2) yield respectively:
( )
⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = + + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = − − − k k n k k k k n k k k n kn g U
V k n U P U V k n U V k n U 1 2 3 3 1 2 ... 4 4 2 3 2
( )
k k k U P V 1 4 4 2 2 2 2 − = + − ⎟ ⎠ ⎞ ⎜ ⎝ ⎛Where g
( )
x is a polynomial with integer coefficients and without independent term. The last identity implies that , 12 ⎟⎠=
⎞ ⎜ ⎝ ⎛ k k U V GCD .
Now, since k k
k n U U U U
GCD , ⎟⎟
⎠ ⎞ ⎜⎜
⎝ ⎛
, k
( )
kn k k n k k n U g V k n U U U U U
GCD ⎟ +
⎠ ⎞ ⎜ ⎝ ⎛ = ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝
⎛ −1
2
, , then
⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k k n U U U
GCD , divides to
( ) ( )
1 1 2 2 − − ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = − + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ k n k k k k n k V k n U g U g V k n , so k n U U U GCD k k n , ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛
also in this case because , 1 2 ⎟⎠=
⎞ ⎜ ⎝ ⎛ k k U V GCD .
k n k k n n P U V P U V ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ + + = + + 4 2 2 4 2 2 2 2
. (3)
In an analogous way to case 1), it can be seen that
⎟ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k U U e k n U U U GCD k n , 2 2 divides to k n , so
in order to prove that
k n U U U GCD k k n , ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛
, it is sufficient to see that
⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ≤ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n e U U U GCD e k k n 2
2 , . Three new cases appear:
- If k n
is an odd number, then expanding the right side of (3) and equaling the terms with P2 +4, it is obtained that:
(
)
(
)
21 1 2 1 2 2 3 1 4 2 ... 4 2 2 3 2 ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ − − − − + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ + + + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = k n k n k k k n k k n k k n P U P U V k n V k n U U , so k n U U is
an odd number regardless 2
k
U
is an odd number or 2
k
V
is an odd number and then
⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = = ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n e U U U GCD e k k n 2
2 , 0 .
-If k n
is an even number and 2
k
V
is an odd number, then 2
k
U
is an even number and it is obtained in an analogous way than in the previous case that:
(
)
(
)
2Now, it is a known fact that ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ i k n i k n GCD k n ,
, so ⎟
⎠ ⎞ ⎜ ⎝ ⎛ ≥ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ k n e i k n
e2 2 when i is an odd
number. Since 2
k
U
is an even number, this implies that ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n e U U e k n 2
2 , so
⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ≤ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n e U U U GCD e k k n 2
2 , as desired.
-If k n and 2 k V
are even numbers, then 2
k
U
is an odd number, so
⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ≤ ≤ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ k n e U U U GCD e k k n 2
2 , 1 , as desired.
Now the theorem is stated: Theorem
If 1
2 , ⎟= ⎠ ⎞ ⎜ ⎝ ⎛ n P
GCD when n is an even number, n has not prime factors congruent with 1 modulus 4 and it has the factor three, then the only powers included in Un are
8 1 , 1 6 = −
U and U121 ,−1 =144 Proof
It is satisfied that
(
)
3 2 3 3 1 U U P U U U
U n n
n = = + , with
3 1 , 2 3 n P U U GCD n ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛
+ by the lemma.
Hence ⎟⎟
⎠ ⎞ ⎜⎜ ⎝ ⎛ +1 , 2 3 P U U
GCD n has not prime factors congruent with 1 modulus 4 since n has not prime factors congruent with 1 modulus 4 by hypothesis. It is also satisfied that
⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ +1 , 2 3 P U U GCD n
1) If , 2 1 1 3
= ⎟⎟ ⎠ ⎞ ⎜⎜
⎝ ⎛
+ P U U
GCD n
, then
(
)
n mn u
U U P
U = + =
3 2
1 implies that P2 +1=vm with
N
v∈ , so the Catalan equation xm− y2 =1 has a solution in positive integers:
(
v,P)
, a contradiction (see Mihailescu)2) If , 2 1 2
3
= ⎟⎟ ⎠ ⎞ ⎜⎜
⎝ ⎛
+ P U U
GCD n , then n must be an even number because
3 2 1 , 2 3
n P
U U
GCD n ⎟⎟=
⎠ ⎞ ⎜⎜
⎝ ⎛
+ by the lemma, so
(
)
3 2 2
3 2 3
2 1
U U
U P
P U U
U U U
Un = n = + n .
Now,
(
)
n mn u
U U
U P
P
U = + =
3 2 2
1 implies that P=wm with w∈N, P2 +1=2vm
with v∈N, because , 2 1 2 3
= ⎟⎟ ⎠ ⎞ ⎜⎜
⎝ ⎛
+ P U U
GCD n
and , 1
2 = ⎟⎟ ⎠ ⎞ ⎜⎜
⎝ ⎛
U U P
GCD n
by the lemma and
the hypothesis 1
2 , ⎟=
⎠ ⎞ ⎜ ⎝ ⎛ n
P
GCD . Then it is satisfied that 2vm−
( )
w2 m =1, so the diophantine equation 2 xm −ym =1 has a solution in positive integers:(
v,w2)
. But for3 ≥
m the only solution in positive integers to this equation is x= y=1 (see Darmon). This implies that w2 =1 and then P=wm =1 (P is a positive number). But in this case the only m-power with m≥3 in Un for the required conditions on n is
3 1
, 1
6 =8=2
−
U ,
as stated in the introduction.
For m=2 it is satisfied that 2v2 −w4 =1, so the diophantine equation 2 x2 − y4 =1 has a solution in positive integers:
(
v,w)
. But the only solution in positive integers to this equation is x= y=1 (see M. Le). This implies that P=wm =1. But in this case the only square in Un for the required conditions on n is U121 ,−1 =144=122, as stated in the introduction and the proof is finished.3. Concluding remarks
A result about the powers in the Lucas sequences Un has been stated for indices n
studies of Kagawa, Ribenboim about the squares in the Lucas sequences and of Bennett about the m-powers with m≥3 in the Lucas sequences.
A future line of work could be to relax the conditions on the index n in the theorem of section 2
References
1. M. A. Bennett, Powers in recurrence sequences: Pell equations, Transactions of the American Mathematical Society, 357 (2004), 1675–1691
2. A. Bremner and N. Tzanakis, Lucas sequences whose 12th or 9th term is a square, J. Number Theory, 107 (2004), 215-227.
3. Y. Bugeaud, M. Mignotte and S. Siksek, Classical and modular approaches to exponential Diophantine equations, I. Fibonacci and Lucas perfect powers. Annals of Math. 163 (2006), 969-1018.
4. Cohn, J. H. E. (1996). Perfect Pell powers. Glasgow Mathematical Journal 38 (1): 19–20.
5. H. Darmon and L. Merel, Winding quotients and some variants of Fermat’s Last Theorem, J. Reine Angew. Math. 490 (1997), 81–100.
6. Kagawa, T. and Terai, N.: Squares in Lucas sequences and some Diophantine equations. Manuscripta Math. 96, 195–202 (1998).
7. M. Le. On the diophantine equation