We give a characterization for the integers n≥1 such that the Fibonomial coefficient (pnn)F is divisible by p for any prime p≠2,5. Then we use it to calculate asymptotic formulas for the number of positive integers n≤x such that p∣(pnn)F. This completes the study on this problem for all primes p.
Citation: Phakhinkon Napp Phunphayap, Prapanpong Pongsriiam. Divisibility of Fibonomial coefficients in terms of their digital representations and applications[J]. AIMS Mathematics, 2022, 7(4): 5314-5327. doi: 10.3934/math.2022296
[1] | Phakhinkon Phunphayap, Prapanpong Pongsriiam . Explicit formulas for the p-adic valuations of Fibonomial coefficients II. AIMS Mathematics, 2020, 5(6): 5685-5699. doi: 10.3934/math.2020364 |
[2] | Zhi-Hong Sun . Supercongruences involving Apéry-like numbers and binomial coefficients. AIMS Mathematics, 2022, 7(2): 2729-2781. doi: 10.3934/math.2022153 |
[3] | Tingting Du, Li Wang . On the power sums problem of bi-periodic Fibonacci and Lucas polynomials. AIMS Mathematics, 2024, 9(4): 7810-7818. doi: 10.3934/math.2024379 |
[4] | Guangwei Hu, Huixue Lao, Huimin Pan . High power sums of Fourier coefficients of holomorphic cusp forms and their applications. AIMS Mathematics, 2024, 9(9): 25166-25183. doi: 10.3934/math.20241227 |
[5] | Kwang-Wu Chen, Fu-Yao Yang . Infinite series involving harmonic numbers and reciprocal of binomial coefficients. AIMS Mathematics, 2024, 9(7): 16885-16900. doi: 10.3934/math.2024820 |
[6] | Halit Orhan, Nanjundan Magesh, Chinnasamy Abirami . Fekete-Szegö problem for Bi-Bazilevič functions related to Shell-like curves. AIMS Mathematics, 2020, 5(5): 4412-4423. doi: 10.3934/math.2020281 |
[7] | Fei Hou, Bin Chen . Triple correlation sums of coefficients of θ-series. AIMS Mathematics, 2023, 8(10): 25275-25287. doi: 10.3934/math.20231289 |
[8] | Qingsheng Zhu, Changwen Xie, Jia-Bao Liu . On the impact of the digital economy on urban resilience based on a spatial Durbin model. AIMS Mathematics, 2023, 8(5): 12239-12256. doi: 10.3934/math.2023617 |
[9] | Chunli Li, Wenchang Chu . Remarkable series concerning (3nn) and harmonic numbers in numerators. AIMS Mathematics, 2024, 9(7): 17234-17258. doi: 10.3934/math.2024837 |
[10] | Rajiniganth Pandurangan, Sabri T. M. Thabet, Imed Kedim, Miguel Vivas-Cortez . On the Generalized ¯θ(t)-Fibonacci sequences and its bifurcation analysis. AIMS Mathematics, 2025, 10(1): 972-987. doi: 10.3934/math.2025046 |
We give a characterization for the integers n≥1 such that the Fibonomial coefficient (pnn)F is divisible by p for any prime p≠2,5. Then we use it to calculate asymptotic formulas for the number of positive integers n≤x such that p∣(pnn)F. This completes the study on this problem for all primes p.
The Fibonacci sequence (Fn)n≥1 is given by the recurrence relation Fn=Fn−1+Fn−2 for n≥3 with the initial values F1=F2=1. For each m≥1 and 1≤k≤m, the Fibonomial coefficient (mk)F is defined by
(mk)F=F1F2F3⋯Fm(F1F2F3⋯Fk)(F1F2F3⋯Fm−k)=Fm−k+1Fm−k+2⋯FmF1F2F3⋯Fk. |
As usual, if m=k, then the empty product F1F2⋯Fm−k is defined to be 1, and similar to the binomial coefficients, we let (mk)F=1 if k=0 and (mk)F=0 if k>m. Then it is well known that (mk)F is always an integer for every m≥1 and k≥0.
There has been some interest in the study of certain generalizations of binomial coefficients such as the Fibonomial or Lucasnomial coefficients. For instance, Marques and Trojovský [9] determined the integers n≥1 such that (3nn)F is divisible by 3. Then Ballot [1] largely extended Marques and Trojovský's results by characterizing all integers n≥1 such that (pnn)U is divisible by p for any nondegenerate fundamental Lucas sequence U and p=2,3 and for p=5,7 in the case U=F. Ballot [1] also proved that E3(x)=O(logx) and E7(x)=O((logx)3). Here and throughout this article, Ep(x) denotes the number of positive integers n less than or equal to x such that (pnn)F is not divisible by p. In other words,
Ep(x)=∑1≤n≤xp∤(pnn)F1. |
In particular, we [13,14] have recently provided an explicit formula for the p-adic valuation of certain Fibonomial coefficients, and have used it in the investigation of the integers n≥1 such that (pann)F is divisible by p for any prime p≡±2(mod5) and any integer a≥1, and also for primes p≡±1(mod5) and a=1 in terms of the sum of digit function.
In this article, we give characterizations for the integers n≥1 such that (pnn)F is divisible by p for any prime p≠2,5 in terms of the digital representation of n. Then we use it in the calculation for asymptotic formulas of Ep(x) for all primes p. This extends many results in the literature which focus only on small primes p≤7.
We organize this article as follows. In Section 2, we recall some definitions and useful results. In Section 3, we prove our main theorems and give some examples. For more information on Fibonacci numbers, Fibonomial coefficients, and generalizations, we refer the reader to some recent articles by Ballot [2,3,4], Chu and Kiliç [5], Kiliç and Akkus [7], Kiliç and Prodinger [8], Onphaeng and Pongsriiam [10,11,12], and Pongsriiam [15,16].
Throughout this article, unless stated otherwise, x is a positive real number, p is a prime, a,b,k,m,n,q,r are integers, m,n≥1, q≥2, ⌊x⌋ is the largest integer less than or equal to x, {x} is the fractional part of x given by {x}=x−⌊x⌋, amodm is the least nonnegative residue of a modulo m, and logx is the natural logarithm of x. The p-adic valuation of n, denoted by νp(n), is the exponent of p in the prime factorization of n. In addition, the order (or the rank) of appearance of n in the Fibonacci sequence, denoted by z(n), is the smallest positive integer m such that n∣Fm. Furthermore, we define sq(n) to be the sum of digits of n when n is written in base q, that is, if
n=(akak−1…a0)q=akqk+ak−1qk−1+⋯+a0 where0≤ai<qfor everyi, |
then sq(n)=ak+ak−1+⋯+a0. Next, we recall some well known and useful results for the reader's convenience.
Lemma 1. The following statements hold.
(i) n∣Fm if and only if z(n)∣m.
(ii) z(p)∣p+1 if and only if p≡±2(mod5).
(iii) z(p)∣p−1 if and only if p≡±1(mod5).
(iv) If p≠5, then gcd(z(p),p)=1.
Proof. These are well known. See, for example, in [13,Lemma 1] for more details.
We will deal with a lot of calculations involving the floor function. So it is useful to recall the following results, which will be applied throughout this article without further reference.
Lemma 2. For k∈Z and x∈R, the following statements hold.
(i) ⌊k+x⌋=k+⌊x⌋,
(ii) {k+x}={x},
(iii) ⌊x⌋+⌊−x⌋={−1,ifx∉Z;0,ifx∈Z,
(iv) 0≤{x}<1 and {x}=0 if and only if x∈Z,
(v) ⌊x+y⌋={⌊x⌋+⌊y⌋,if{x}+{y}<1;⌊x⌋+⌊y⌋+1,if {x}+{y}≥1,
(vi) ⌊⌊x⌋k⌋=⌊xk⌋ for k≥1.
Proof. These are well known and can be proved easily. For more details, see in [6,Chapter 3]. We also refer the reader to [11] for a nice application of these properties.
The next three lemmas are important tools for obtaining the characterizations of the integers n such that (pnn)F is divisible by p.
Lemma 3. [14,Corollary 13] Suppose that p≠2,5 and a, n are positive integers. If n≡0(modz(p)), then p∣(pann)F.
Lemma 4. [14,Corollary 14] Let p≠2,5, p≡±2(mod5), a,n∈N, r=panmodz(p), s=nmodz(p), A=⌊n(pa−1)pνp(n)z(p)⌋, and n≢0(modz(p)). Then the following statements hold.
(i) Assume that a is odd and p∤n. If r<s, then p∣(pann)F. If r≥s, then p∣(pann)F if and only if sp(A)≥a+12(p−1).
(ii) Assume that a is odd and p∣n. If r≠s, then p∣(pann)F. If r=s, then p∣(pann)F if and only if sp(A)≥a+12(p−1).
Lemma 5. [14,Corollary 15] Let p≠2,5, p≡±1(mod5), and A=n(p−1)pνp(n)z(p). Then p∣(pnn)F if and only if sp(A)≥p−1.
Lemma 6. Let k≥0, q≥2, 1≤a≤q−1, and 0≤b≤q−1. Then
sq(a(q−1)qk+bqk)≥b. |
Proof. When b=0, the result is obvious. So we assume that b≥1. If a=1, then we write a(q−1)qk+bqk=qk+1+(b−1)qk. If a≥2 and b≤a−1, then we write a(q−1)qk+bqk=(a−1)qk+1+(q−a+b)qk. If a≥2 and b≥a, then we write a(q−1)qk+bqk=aqk+1+(b−a)qk. In each case, sq(a(q−1)qk+bqk) is equal to, respectively, 1+b−1=b, a−1+q−a+b=q+b−1, and a+b−a=b. In any case, it is at least b.
Lemma 7. Let p≥3, p≡±2(mod5), 0≤a≤p−12, and 1≤k≤z(p)2. Then az(p)+k≡0(modp) if and only if a=p−12, z(p) is even, and k=z(p)2. In particular, if a<p−12, then az(p)+k≢0(modp).
Proof. From the assumption, we have
0<az(p)+k≤(p−12)z(p)+z(p)2=pz(p)2. |
Suppose that az(p)+k≡0(modp). Then az(p)+k=np for some 1≤n≤z(p)2. Since p≡±2(mod5), we obtain by Lemma 1 that p≡−1(modz(p)). Then k≡az(p)+k≡np≡−n(modz(p)). So there exists m∈N such that k=mz(p)−n. Therefore
z(p)2≥k=mz(p)−n≥z(p)−z(p)2=z(p)2. |
This implies that k=z(p)2, z(p) is even, m=1, and n=z(p)2. Since az(p)+k=np, we also obtain a=p−12. The converse can be verified easily. This completes the proof.
We introduce the following notation for convenience.
Definition 8. Let q and i be integers such that q≥2 and 0≤i≤q−1. We define
H(q,i)={(amam−1⋯a0)q | m∈N∪{0}, ak≤ak−1 for all1≤k≤m,and a0=i}. |
In other words, H(q,i) is the set of nonnegative integers n such that the q-adic representation of n is increasing (from the left to the right), and the last digit (the rightmost digit) is equal to i.
For example, if q=10 and i=3, then 111122233 and 11111333 are in H(10,3) but 213 and 1234 are not in H(10,3).
Definition 9. For positive integers k and q, we define
t(q,k)=⌊k(q−1)z(q)⌋. |
The next lemma is usually called stars and bars problem. Recall that if a set A has exactly n distinct elements, then the number of all possible ways in choosing m elements from A with repetitions allowed is (n+m−1m). We have the following lemma.
Lemma 10. Let k≥1, q≥2, and 1≤t≤q−1 be integers. Then
#{(akak−1⋯a1)q∈t⋃i=1H(q,i) ∣ ak≠0}=(k+t−1k). |
Proof. This is stars and bars problem. The set A is {1,2,3,…,t}. We would like to choose k elements from A with repetitions allowed. So the number of ways, as recalled above, is (t+k−1k), which proves this lemma.
Lemma 11. Let q≥2 and 1≤t≤q−1 be integers. Then
∑0≤m<qrm∈⋃ti=0H(q,i)1=(r+tr) |
Consequently,
∑0≤m<qrm∈⋃ti=0H(q,i)1=rtt!+O(rt−1), |
where the implied constant depends at most on t.
Proof. The conditions 0≤m<qr and m∈⋃ti=0H(q,i) mean that m=(arar−1⋯a1)q and 0≤ar≤ar−1≤⋯≤a1≤t. So this is also stars and bars problem. The set A is {0,1,2,…,t}. We would like to choose r elements from A with repetitions allowed. Therefore the number of ways is (t+1+r−1r)=(r+tr), which proves the first part. Next,
(r+tr)=(r+t)(r+(t−1))⋯(r+1)t!=rtt!+P(r), |
where P(r) is a polynomial in r of degree t−1 with the coefficients depending only on t. Therefore P(r)=O(rt−1) and the implied constant depends at most on t. This completes the proof.
In this section, we begin with a property of the sum of digit function. Then we use it in the study of the divisibility p∣(pnn)F in terms of the digital representation of n. After that, we determine an asymptotic formula for Ep(x).
Theorem 12. Let m≥0, q≥2, and 1≤k≤z(q)−1. Then
sq((q−1)m+t(q,k))<q−1if and only ifm∈t(q,k)⋃i=0H(q,i). |
Proof. Let H=⋃t(q,k)i=0H(q,i) and t=t(q,k). Since k<z(q), we see that t<q−1. If m=0, then m∈H(q,0)⊆H and sq((q−1)m+t)=sq(t)=t<q−1, so we are done. From this point on, we assume that m≥1. To prove this theorem, we first show that
ifm∉H,then sq((q−1)m+t)≥q−1. | (3.1) |
We prove (3.1) by induction on r where r is the number of digits in the q-adic expansion of m. For r=1, we let m=a, 1≤a≤q−1, a∉H, and write
(q−1)m+t=(a−1)q+(q−a+t). |
Observe that i∈H(q,i)⊆H for each 0≤i≤t. Since a∉H, we see that a>t which implies 0≤q−a+t≤q−1. Therefore sq((q−1)m+t)=a−1+q−a+t≥q−1. Next, let r≥1 and suppose that (3.1) holds for any m∈N such that the number of digits of m in its q-adic expansion is less than or equal to r. Assume that m=(ar+1ar⋯a1)q, ar+1≠0, 0≤ai<q for all i, and m∉H. Let m1=(arar−1⋯a1)q.
Case 1. m1∈H. If r=1, let m2=0 and if r≥2, we let m2=(ar−1ar−2⋯a1)q. Then we write (q−1)m+t as (q−1)(ar+1qr+arqr−1+m2)+t, which is equal to
(ar+1−1)qr+1+(q−ar+1+ar)qr−arqr−1+(q−1)m2+t. | (3.2) |
Since m1∈H, ar≤ai≤t for all 1≤i≤r. So we have
m2≤t(1+q+q2+⋯+qr−2)=t(qr−1−1q−1), |
m2≥ar(1+q+q2+⋯+qr−2)=ar(qr−1−1q−1). |
Therefore
(q−1)m2+t≤tqr−1and(q−1)m2+t≥arqr−1+(t−ar)≥arqr−1. |
Thus
0≤−arqr−1+(q−1)m2+t≤(t−ar)qr−1<qr. | (3.3) |
Since m∉H and m1∈H, ar+1>ar. Thus 0≤q−ar+1+ar<q. From this and from (3.2) and (3.3), we obtain that sq((q−1)m+t) is equal to
sq((ar+1−1)qr+1+(q−ar+1+ar)qr)+sq(−arqr−1+(q−1)m2+t)≥sq((ar+1−1)qr+1+(q−ar+1+ar)qr)=(ar+1−1)+(q−ar+1+ar)=q−1+ar≥q−1. |
Case 2. m1∉H. Since (q−1)m1+t<(q−1)qr+q−1<qr+1, we write (q−1)m1+t=(br+1br⋯b1)q where br+1 may be zero. Since the number of digits in the q-adic representation of m1 is less than or equal to r, we can apply the induction hypothesis on m1 to obtain
q−1≤sq((q−1)m1+t)=sq((br+1br⋯b1)q)=r+1∑i=1bi. | (3.4) |
Next we write
(q−1)m+t=(q−1)(ar+1qr+m1)+t=(q−1)ar+1qr+(br+1br⋯b1)q=(q−1)ar+1qr+br+1qr+(br⋯b1)q. |
By the above equation, Lemma 6, and (3.4), we obtain sq((q−1)m+t)=
sq((q−1)ar+1qr+br+1qr)+sq((brbr−1⋯b1)q)≥br+1+r∑i=1bi≥q−1. |
This proves (3.1). To prove the converse, assume that m∈H and let a=mmodq be the least nonnegative residue of m modulo q. Then a is the last digit of m in its q-adic expansion. Since m≥1 and m∈H, we see that 1≤a≤t and m∈H(q,a). So the possible digits in the q-adic representation of m with nonzero leading digit are 1,2,3,…,a. Therefore we can write m as
na∑i=0aqi+na+na−1∑i=na+1(a−1)qi+na+na−1+na−2∑i=na+na−1+1(a−2)qi+⋯+na+na−1+⋯+n1∑i=na+⋯+n2+1qi, | (3.5) |
where n1,n2,…,na are nonnegative integers and the empty sum is defined to be zero. So, for instance, if a−1 does not appear as a digit in the q-adic representation of m, then we let na−1=0 and the second sum in (3.5) is 0. For 0≤i≤a−1, let di=(∑i+1≤j≤anj)+1. By (3.5), m is equal to
na∑j=0aqj+na−1−1∑j=0(a−1)qda−1+j+na−2−1∑j=0(a−2)qda−2+j+⋯+n1−1∑j=0qd1+j=na∑j=0aqj+a−1∑i=1ni−1∑j=0iqdi+j=ana∑j=0qj+(a−1∑i=1(iqdini−1∑j=0qj))=a(qda−1−1q−1)+(a−1∑i=1iqdi(qni−1q−1))=1q−1(aqda−1−a+a−1∑i=1(iqdi−1−iqdi))=1q−1((a−1∑i=0qdi)−a). |
Then (q−1)m+t=∑a−1i=0qdi−a+t. Since di≥1 for all i and 0≤t−a<q−1, we see that sq((q−1)m+t) is equal to
sq(a−1∑i=0qdi)+sq(t−a)=a+t−a<q−1. |
This completes the proof.
Recall that we [14] previously gave a characterization for the divisibility p∣(pnn)F in terms of the sum of digits function. We are now ready to characterize it in terms of a digital representation. We first prove it for the prime p≡±2(mod5) in the next theorem.
Theorem 13. Let p be an odd prime, p≡±2(mod5), and n a positive integer. Then p∣(pnn)F if and only if n is not of the form
z(p)m+k where1≤k≤z(p)2andm∈t(p,k)⋃i=0H(p,i). | (3.6) |
Proof. We first assume that p∤(pnn)F. To show that n can be written as in (3.6), let k=nmodz(p), t=t(p,k), and H=⋃ti=0H(p,i).
Case 1. p∣n. We write n=paℓ where a,ℓ∈N and p∤ℓ. By Lemma 3, we obtain n≢0(modz(p)). Then by Lemma 4(ⅱ), we have
pn≡n(modz(p)) and sp=(⌊n(p−1)paz(p)⌋)=sp(⌊ℓ(p−1)z(p)⌋)<p−1. |
By Lemma 1, we know that p≡−1(modz(p)), and so n≡pn≡−n(modz(p)). Therefore z(p)∣2n and z(p)∤n. This implies
z(p)is even and nmodz(p)=z(p)2=ℓmodz(p). |
Then k=z(p)2, t=p−12, ℓ=z(p)m1+k for some m1≥0, and ⌊ℓ(p−1)z(p)⌋=(p−1)m1+t. Since sp((p−1)m1+t)<p−1, we obtain by Theorem 12 that m1∈H. In addition, we obtain that
n=ℓpa=(z(p)m1+z(p)2)pa=z(p)m+k, where |
m=m1pa+(pa−1)kz(p)=m1pa+pa−12=m1pa+t(pa−1+pa−2+⋯+1). |
Since m1∈H, so is m. Hence n is of the form (3.6).
Case 2. p∤n. This case is similar to Case 1. By Lemmas 3 and 2.4(ⅰ), we obtain
1≤nmodz(p)≤pnmodz(p) and sp(⌊n(p−1)z(p)⌋)<p−1. |
Since p≡−1(modz(p)), pn≡−n(modz(p)). Therefore nmodz(p)≤(−n)modz(p)=z(p)−(nmodz(p)). Then nmodz(p)≤z(p)2. Then
1≤k≤z(p)2, n=z(p)m+k forsomem≥0, and ⌊n(p−1)z(p)⌋=(p−1)m+t. |
Since sp(⌊n(p−1)z(p)⌋)<p−1, we obtain by Theorem 12 that m∈H. Therefore n is of the form (3.6). This proves the converse of this theorem.
For the other direction, assume that n is of the form (3.6). We still let t=t(p,k) and H=⋃ti=0H(p,i), and separate the consideration into two cases.
Case 3. k<z(p)2. Then 0≤t≤p−32. Let m=(arar−1⋯a1)p be the p-adic expansion of m. Since m∈H and 0≤t≤p−32, we see that 0≤a1≤p−32. So we obtain by Lemma 7 that
n≡z(p)m+k≡a1z(p)+k≢0(modp). | (3.7) |
Applying the fact that p≡−1(modz(p)), nmodz(p)=k, and 1≤k≤z(p)2, we obtain
npmodz(p)=(−n)modz(p)=(−k)modz(p)=z(p)−k≥k=nmodz(p). | (3.8) |
Since m∈H and ⌊n(p−1)z(p)⌋=(p−1)m+t, we obtain by Theorem 12 that
sp(⌊n(p−1)z(p)⌋)<p−1. | (3.9) |
By (3.7), (3.8), (3.9), and Lemma 4(i), we obtain p∤(pnn)F.
Case 4. k=z(p)2. Similar to Case 1, we have
nmodz(p)=z(p)2=npmodz(p) and sp(⌊n(p−1)z(p)⌋)<p−1. |
If p∤n, then we obtain by Lemma 4(i) that p∤(pnn)F. So suppose that p∣n and let a=mmodp. Since m∈H, we see that a≤t=p−12. In addition, az(p)+k≡mz(p)+k≡n≡0(modp), so we obtain by Lemma 7 that a=p−12. Since m∈H(p,a), there are r≥0 and m2∈⋃p−32i=0H(p,i) such that
m=m2pr+1+a(pr+pr−1+⋯+1)=m2pr+1+pr+1−12. |
So we have
n=z(p)(m2pr+1+pr+1−12)+z(p)2=(z(p)m2+k)pr+1. |
Since m2∈⋃p−32i=0H(p,i)⊆H, we obtain by Theorem 12 that
sp(⌊(z(p)m2+k)(p−1)z(p)⌋)=sp((p−1)m2+t)<p−1. |
In addition, if m2modp=a2, then 0≤a2≤p−32 and we obtain by Lemma 7 that
z(p)m2+k≡a2z(p)+k≢0(modp). |
Since n=(z(p)m2+k)pr+1 and z(p)m2+k≢0(modp), we obtain r+1=νp(n). In addition,
npmodz(p)=nmodz(p) and sp(⌊n(p−1)pνp(n)z(p)⌋)<p−1. |
Therefore p∤(pnn)F, by Lemma 4(ⅱ). This completes the proof.
By Theorem 13, we immediately obtain the following corollary.
Corollary 14. If p>2, p≡±2(mod5), and nmodz(p)>z(p)2, then p∣(pnn)F.
If nmodz(p)<z(p)2, then we may still have p∣(pnn)F as shown in the next corollary.
Corollary 15. Let p be an odd prime, p≡±2(mod5), p∣n, and nmodz(p)≠z(p)2. Then p∣(pnn)F.
Proof. Suppose for a contradiction that p∤(pnn)F. Then we obtain by Theorem 13 that n=z(p)m+k, 1≤k≤z(p)2, and m∈H where H=⋃t(p,k)i=0H(p,i). Since nmodz(p)≠z(p)2, k<z(p)2. This implies that t(p,k)≤p−32. Let m=(arar−1⋯a1)p be the p-adic representation of m. Since m∈H and t(p,k)≤p−32, we see that 0\leq a_1 \leq \frac{p-3}{2} . By Lemma 7, we obtain a_1 z(p) + k \not\equiv 0 \pmod{p} . Therefore
\begin{equation*} n \equiv z(p)m+k \equiv a_1 z(p) + k \not\equiv 0 \pmod{p}, \end{equation*} |
which contradicts the assumption that p\mid n . Hence the proof is complete.
Next, we give a characterization for the divisibility p \mid {pn \choose n}_F when p \equiv \pm 1 \pmod{5} .
Theorem 16. Let p be an odd prime such that p \equiv \pm 1 \pmod{5} and let n be a positive integer. Then p\mid {pn \choose n}_F if and only if n is not of the form
\begin{equation} z(p)m + k \ \ \mathit{{\text{where}\; 1\leq k \leq z(p) - 1 \;and \; m \in \bigcup\limits_{{i = 0}}^{t(p,k)}} H(p,i) .} \end{equation} | (3.10) |
Proof. Let A = \frac{n(p-1)}{p^{\nu_p (n)}z(p)} . Similar to the proof of Theorem 13, we first assume that p \nmid {pn \choose n}_F and let k = n \bmod{z(p)} . Then n = z(p)m + k for some m \geq 0 , and by Lemma 3, k \neq 0 . So 1 \leq k \leq z(p) - 1 . If remains to show that m \in \bigcup_{i = 0}^{t(p, k)} H(p, i) . Since p \equiv \pm 1 \pmod{5} , we obtain by Lemma 1 that z(p) \mid p - 1 . This implies t(p, k) = \frac{k(p-1)}{z(p)} . By Lemma 5, we have
\begin{equation*} p - 1 > s_p (A) = s_p \left(p^{\nu_p (n)} A\right) = s_p \left( \frac{n(p-1)}{z(p)} \right) = s_p ((p-1)m + t(p,k)). \end{equation*} |
By Lemma 12, m \in \bigcup_{i = 0}^{t(p, k)} H(p, i) , as required. Next, if n is of the form (3.10), then we apply Theorem 12 to obtain
\begin{equation*} s_p (A) = s_p \left( p^{\nu_p (n)} A \right) = s_p((p-1)m + t(p,k)) < p - 1, \end{equation*} |
and then use Lemma 5 to conclude that p \nmid {pn \choose n}_F . This completes the proof.
Next we apply Theorems 13 and 16 to determine an asymptotic formula for E_p (x) .
Theorem 17. Let p be an odd prime, p \equiv \pm 2 \pmod{5} , and t = t\left(p, \left\lfloor \frac{z(p)}{2} \right\rfloor \right) . Then uniformly for x \geq 2 ,
\begin{equation*} E_p (x) = \frac{(\log x)^t}{t! (\log p)^t} + O\left((\log x)^{t-1}\right), \end{equation*} |
and consequently,
\begin{equation*} \sum\limits_{\substack{1 \leq n \leq x \\ p\mid {pn \choose n}_F}} 1 = x - \frac{(\log x)^t}{t! (\log p)^t} + O\left((\log x)^{t-1}\right), \end{equation*} |
where the implied constants depend at most on p .
Proof. In this proof, the implied constants in each estimate depend at most on p . By Theorem 13, we obtain
\begin{equation} E_p (x) = \sum\limits_{\substack{1 \leq n \leq x \\ p\nmid {pn \choose n}_F}} 1 = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \sum\limits_{\substack{1 \leq n \leq x \\ n = z(p)m + k \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1 = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \sum\limits_{\substack{0 \leq m \leq \frac{x - k}{z(p)} \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1. \end{equation} | (3.11) |
For each 1 \leq k \leq \frac{z(p)}{2} , let r_k be the number of digits in the p -adic expansion of \left\lfloor \frac{x - k}{z(p)} \right\rfloor and let r = r_{\left\lfloor \frac{z(p)}{2} \right\rfloor} . Then
\begin{equation*} r_k = \left\lfloor \frac{\log \left\lfloor \frac{x - k}{z(p)} \right\rfloor }{\log p} \right\rfloor + 1 \ \ {\rm{for \;all }}\; 1 \leq k \leq \frac{z(p)}{2} . \end{equation*} |
By (3.11) and Lemma 11,
\begin{equation*} E_p (x) \leq \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \sum\limits_{\substack{0 \leq m < p^{r_k} \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1 = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \left( \frac{r_{k}^{t(p,k)}}{t(p,k)!} + O\left( r_{k}^{t(p,k) - 1}\right) \right). \end{equation*} |
In addition,
\begin{align*} E_p (x) \geq \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \sum\limits_{\substack{0 \leq m < p^{r_k - 1} \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1 & = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \left( \frac{(r_k - 1)^{t(p,k)}}{t(p,k)!} + O\left((r_k - 1)^{t(p,k) - 1}\right) \right) \\ & = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \left( \frac{r_{k}^{t(p,k)}}{t(p,k)!} + O\left(r_{k}^{t(p,k) - 1}\right) \right). \end{align*} |
Therefore
\begin{equation} E_p (x) = \sum\limits_{1 \leq k \leq \frac{z(p)}{2}} \left( \frac{r_{k}^{t(p,k)}}{t(p,k)!} + O\left(r_{k}^{t(p,k) - 1}\right) \right). \end{equation} | (3.12) |
Recall that t = t\left(p, \left\lfloor \frac{z(p)}{2} \right\rfloor \right) . Since p\equiv \pm 2 \pmod{5} , we obtain by Lemma 1 that z(p) = p + 1 or z(p) \leq \frac{p+1}{2} . If z(p) = p + 1 , then z(p) is even and for 1 \leq k \leq \frac{z(p)}{2} - 1 ,
\begin{equation*} t(p,k) \leq t\left( p, \frac{z(p)}{2} - 1\right) = \left\lfloor \frac{p-1}{2} - \frac{p - 1}{z(p)} \right\rfloor = \frac{p - 1}{2} - 1 = t - 1. \end{equation*} |
If z(p) \leq \frac{p + 1}{2} , then for 1 \leq k \leq \frac{z(p)}{2} - 1 ,
\begin{align*} t(p,k) \leq t\left( p , \left\lfloor \frac{z(p)}{2} \right\rfloor - 1\right) & = \frac{\left\lfloor \frac{z(p)}{2} \right\rfloor (p - 1) - s}{z(p)} + \left\lfloor \frac{s - (p - 1)}{z(p)} \right\rfloor \\ &\leq \frac{\left\lfloor \frac{z(p)}{2} \right\rfloor (p - 1) - s}{z(p)} + \left\lfloor \frac{(z(p) - 1) - (p - 1)}{z(p)} \right\rfloor \leq t - 1, \end{align*} |
where s = \left\lfloor \frac{z(p)}{2}\right\rfloor (p - 1) \bmod{z(p)} . In any case, t(p, k) \leq t - 1 for 1 \leq k \leq \frac{z(p)}{2} - 1 . In addition,
\begin{equation*} r_k = \frac{\log \left\lfloor \frac{x - k}{z(p)} \right\rfloor }{\log p} + O(1) = \frac{\log x}{\log p} + O(1) \ \ {\rm{for} }\; 1 \leq k \leq \frac{z(p)}{2}. \end{equation*} |
Therefore r_{k}^{t(p, k)} \ll r^{t - 1} \ll (\log x)^{t-1} for any 1 \leq k \leq \frac{z(p)}{2} - 1 . Therefore
\begin{equation*} \sum\limits_{1 \leq k \leq \frac{z(p)}{2} - 1} \left( \frac{r_k ^{t(p,k)}}{t(p,k)!} + O\left( r_k ^{t(p,k) - 1} \right) \right) = O\left((\log x)^{t-1}\right). \end{equation*} |
Thus (3.12) implies that
\begin{equation*} E_p (x) = \frac{(\log x)^t}{t! (\log p)^t} + O\left( (\log x)^{t-1} \right). \end{equation*} |
The rest is now obvious. So the proof is completes.
Theorem 18. Let p be an odd prime, p \equiv \pm 1 \pmod{5} , and t = t\left(p, z(p)-1 \right) . Then uniformly for x \geq 2 ,
\begin{equation*} E_p (x) = \frac{(\log x)^t}{t! (\log p)^t} + O\left((\log x)^{t-1}\right), \end{equation*} |
and consequently,
\begin{equation*} \sum\limits_{\substack{1 \leq n \leq x \\ p\mid {pn \choose n}_F}} 1 = x - \frac{(\log x)^t}{t! (\log p)^t} + O\left((\log x)^{t-1}\right), \end{equation*} |
where the implied constants depend at most on p .
Proof. The proof is similar to that of Theorem 17, so we omit some details, and the implied constants in the following estimates depend at most on p . We obtain by Theorem 16 that
\begin{equation} E_p (x) = \sum\limits_{1 \leq k \leq z(p) - 1} \sum\limits_{\substack{1 \leq n \leq x \\ n = z(p)m + k \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1 = \sum\limits_{1 \leq k \leq z(p) - 1} \sum\limits_{\substack{0 \leq m \leq \frac{x - k}{z(p)} \\ m \in \bigcup\limits_{i = 0}^{t(p,k)} H(p,i)}} 1. \end{equation} | (3.13) |
For each 1 \leq k \leq z(p) - 1 , let r_k be the number of digits in the p -adic expansion of \left\lfloor \frac{x - k}{z(p)} \right\rfloor and let r = r_{z(p) - 1} . Then r_k = \left\lfloor \log_p \left\lfloor \frac{x - k}{z(p)} \right\rfloor \right\rfloor + 1 for all 1 \leq k \leq z(p) - 1 . Similar to the proof of Theorem 17, we apply Lemma 11 to obtain
\begin{equation} E_p (x) = \sum\limits_{1 \leq k \leq z(p) - 1} \left( \frac{r_{k}^{t(p,k)}}{t(p,k)!} + O\left(r_{k}^{t(p,k) - 1}\right) \right). \end{equation} | (3.14) |
Recall that
\begin{equation*} t = t\left( p , z(p)-1 \right) = \left\lfloor \frac{(z(p) - 1)(p-1)}{z(p)} \right\rfloor = p - 1 + \left\lfloor - \frac{p - 1}{z(p)} \right\rfloor. \end{equation*} |
Since p\equiv \pm 1 \pmod{5} , we obtain by Lemma 1 that z(p) = p - 1 or z(p) \leq \frac{p-1}{2} . If z(p) = p - 1 , then for 1 \leq k \leq z(p) - 2 , we have
\begin{equation*} t(p,k) \leq t\left( p , z(p) - 2 \right) = \left\lfloor \frac{(z(p) - 2)(p-1)}{z(p)} \right\rfloor = p - 3 = t - 1. \end{equation*} |
If z(p) \leq \frac{p - 1}{2} , then we obtain by Lemma 2(v) that for 1 \leq k \leq z(p) - 2 ,
\begin{align*} t(p,k) \leq t\left( p , z(p) - 2 \right) = p - 1 + \left\lfloor - \frac{2(p - 1)}{z(p)} \right\rfloor &\leq p - 1 + \left\lfloor -\frac{p - 1}{z(p)} \right\rfloor + \left\lfloor - \frac{p - 1}{z(p)} \right\rfloor + 1 \\ &\leq t - 1. \end{align*} |
In any case, t(p, k) \leq t - 1 for all 1 \leq k \leq z(p) - 2 . In addition,
\begin{equation*} r_k = \log_p \left( \left\lfloor \frac{x - k}{z(p)} \right\rfloor \right) + O(1) = \log_p x + O(1) \ \ {\rm{for 1 \leq k \leq z(p)-1 }}. \end{equation*} |
Therefore r_{k}^{t(p, k)} \ll r^{t-1} \ll (\log x)^{t - 1} for any 1 \leq k \leq z(p) - 2 . From (3.14) and this observation, we obtain the desired results.
We give an example to show an application of our results. Also see [1,9,14] for the characterization of the divisibility p \mid {pn \choose n}_F when p = 2, 3, 5, 7 .
Example 19. Let n \in \mathbb{N} . Then 11 \nmid {11n \choose n}_F if and only if
\begin{equation*} n = 10m + k \ \ {\rm{where}}\; 1\leq k \leq 9 \;{\rm{ and}}\; m \in \bigcup\limits_{i = 0}^{k} H(11,i) . \end{equation*} |
In addition, 13 \nmid {13n \choose n}_F if and only if
\begin{equation*} n = 7m + k \ \ {\rm{where}}\; 1 \leq k \leq 3 m \in \bigcup\limits_{i = 0}^{2k-1} H(13,i) . \end{equation*} |
Furthermore,
\begin{equation*} E_{11} (x) = \frac{(\log x)^{9}}{9! (\log 11)^9} + O\left((\log x)^{8}\right) \ \ {\rm{and}} \ \ E_{13} (x) = \frac{(\log x)^5}{5! (\log 13)^{5}} + O\left((\log x)^4 \right). \end{equation*} |
Proof. We have z(11) = 10 , z(13) = 7 , t(11, k) = k for 1 \leq k \leq 9 , t(13, k) = 2k - 1 for 1 \leq k \leq 3 . Applying Theorems 13, 16, 17, and 18, we immediately obtain the desired results.
We give characterizations for the integers n \geq 1 such that {pn \choose n}_F is divisible by p for any prime p \neq 2, 5 in terms of the digital representation of n . We also obtain asymptotic formulas of E_p(x) for all primes p , extending many results in the literature which focus only on small primes p \leq 7 . For a future project, we may be able to extend the results to {pn \choose n}_U for any nondegenerate fundamental Lucas sequences U and any prime p .
We thank the referees for their comments which improve the quality of this paper. This project is funded by the National Research Council of Thailand (NRCT), grant number NRCT5-RSA63021-02.
The authors declare that there is no conflict of interests regarding the publication of this article.
[1] | C. Ballot, Divisibility of Fibonomials and Lucasnomials via a general Kummer rule, Fibonacci Quart., 53 (2015), 194–205. |
[2] | C. Ballot, The congruence of Wolstenholme for generalized binomial coefficients related to Lucas sequences, J. Integer Seq., 18 (2015), Article 15.5.4. |
[3] | C. Ballot, Lucasnomial Fuss-Catalan numbers and related divisibility questions, J. Integer Seq., 21 (2018), Article 18.6.5. |
[4] | C. Ballot, Divisibility of the middle Lucasnomial coefficient, Fibonacci Quart., 55 (2017), 297–308. |
[5] |
W. Chu, E. Kiliç, Quadratic sums of Gaussian q-binomial coefficients and Fibonomial coefficients, Ramanujan J., 51 (2020), 229–243. https://doi.org/10.1007/s11139-018-0023-x doi: 10.1007/s11139-018-0023-x
![]() |
[6] | R. L. Graham, D. E. Knuth, O. Patashnik, Concrete Mathematics : A Foundation for Computer Science, Second Edition, Addison–Wesley, 1994 |
[7] |
E. Kiliç, I. Akkus, On Fibonomial sums identities with special sign functions: analytically q-calculus approach, Math. Slovaca, 68 (2018), 501–512. https://doi.org/10.1515/ms-2017-0120 doi: 10.1515/ms-2017-0120
![]() |
[8] |
E. Kiliç, H. Prodinger, Closed form evaluation of sums containing squares of Fibonomial coefficients, Math. Slovaca, 66 (2016), 757–767. https://doi.org/10.3934/math.2020433 doi: 10.3934/math.2020433
![]() |
[9] | D. Marques, P. Trojovský, On divisibility of Fibonomial coefficients by 3, J. Integer Seq., 15 (2012), Article 12.6.4. |
[10] | K. Onphaeng, P. Pongsriiam, Jacobsthal and Jacobsthal-Lucas numbers and sums introduced by Jacobsthal and Tverberg, J. Integer Seq., 20 (2017), Article 17.3.6. |
[11] |
K. Onphaeng, P. Pongsriiam, Exact divisibility by powers of the integers in the Lucas sequence of the first kind, AIMS Math., 5 (2020), 6739–6748. https://doi.org/10.3934/math.2020433 doi: 10.3934/math.2020433
![]() |
[12] |
K. Onphaeng, P. Pongsriiam, Exact divisibility by powers of the integers in the Lucas sequences of the first and second kinds, AIMS Math., 6 (2021), 11733–11748. https://doi.org/10.3934/math.2021682 doi: 10.3934/math.2021682
![]() |
[13] | P. Phunphayap, P. Pongsriiam, Explicit formulas for the p-adic valuations of Fibonomial coefficients, J. Integer Seq., 21 (2018), Article 18.3.1. |
[14] |
P. Phunphayap, P. Pongsriiam, Explicit formulas for the p-adic valuations of Fibonomial coefficients Ⅱ, AIMS Math., 6 (2020), 5685–5699. https://doi.org/10.3934/math.2020364 doi: 10.3934/math.2020364
![]() |
[15] | P. Pongsriiam, Fibonacci and Lucas numbers which have exactly three prime factors and some unique properties of F_18 and L_18, Fibonacci Quart., 57 (2019), 130–144. |
[16] |
P. Pongsriiam, The order of appearance of factorials in the Fibonacci sequence and certain Diophantine equations, Period. Math. Hungar., 79 (2019), 141–156. https://doi.org/10.1007/s10998-018-0268-6 doi: 10.1007/s10998-018-0268-6
![]() |