On some determinants involving Jacobi symbols
Dmitry Krachun, Fedor Petrov, Zhi-Wei Sun, Maxim Vsemirnov

TL;DR
This paper investigates conjectures on determinants with Jacobi symbol entries, proving several cases for specific moduli and primes, using character sums over finite fields to establish the results.
Contribution
The paper proves new conjectures by Sun regarding determinants with Jacobi symbol entries for specific moduli and primes, expanding understanding of their properties.
Findings
Certain determinants are zero for specific moduli, confirming conjectures.
Proved that $(10,9)_p=0$ for primes $p ot ext{divisible by }12$.
Established that $[5,5]_p=0$ for primes $p ot ext{divisible by }20$.
Abstract
In this paper we study some conjectures on determinants with Jacobi symbol entries posed by Z.-W. Sun. For any positive integer , we show that and as conjectured by Sun, where and with the Jacobi symbol. We also prove that for any prime , and for any prime , which were also conjectured by Sun. Our proofs involve character sums over finite fields.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Finite Fields Appl. 64 (2020), Article 101672.
On some determinants involving Jacobi symbols
Dmitry Krachun, Fedor Petrov, Zhi-Wei Sun, Maxim Vsemirnov
(Dmitry Krachun) St. Petersburg Department of Steklov Mathematical Institute of Russian Academy of Sciences, Fontanka 27, 191023, St. Petersburg, Russia
(Fedor Petrov) St. Petersburg Department of Steklov Mathematical Institute of Russian Academy of Sciences, Fontanka 27, 191023, St. Petersburg, Russia
(Zhi-Wei Sun) Department of Mathematics, Nanjing University, Nanjing 210093, Peopleβs Republic of China
(Maxim Vsemirnov) St. Petersburg Department of Steklov Mathematical Institute of Russian Academy of Sciences, Fontanka 27, 191023, St. Petersburg, Russia
Abstract.
In this paper we study some conjectures on determinants with Jacobi symbol entries posed by Z.-W. Sun. For any positive integer , we show that
[TABLE]
and
[TABLE]
as conjectured by Sun, where
[TABLE]
and
[TABLE]
with the Jacobi symbol. We also prove that for any prime , and for any prime , which were also conjectured by Sun. Our proofs involve character sums over finite fields.
Key words and phrases:
Determinants, Jacobi symbols, character sums over finite fields.
2020 Mathematics Subject Classification. Primary 11C20, 11T24; Secondary 11E16, 15A15.
The work is supported by the NSFC (Natural Science Foundation of China)-RFBR (Russian Foundation for Basic Research) Cooperation and Exchange Program (grants NSFC 11811530072 and RFBR 18-51-53020-GFEN-a). The third author is also supported by the Natural Science Foundation of China (grant 11971222).
1. Introduction
For an matrix over a field, we simply denote its determinant by . In this paper we study some conjectures on determinants with Jacobi symbol entries posed by Z.-W. Sun [11].
Let be an odd prime. In 2004, R. Chapman [2] determined the values of
[TABLE]
and
[TABLE]
where denotes the Legendre symbol. Chapmanβs conjecture on the evaluation of
[TABLE]
was confirmed by M. Vsemirnov [12, 13] via matrix decomposition. With this background, Z.-W. Sun [11] studied some new kinds of determinants with Legendre symbol or Jacobi symbol entries.
For any odd integer and integers and , Sun [11] introduced the notations
[TABLE]
and
[TABLE]
where denotes the Jacobi symbol. He showed that
[TABLE]
and that for any odd prime we have
[TABLE]
For and , if is relatively prime to and for some , then is called a quadratic residue modulo . If is odd and is a quadratic residue modulo , then since is a quadratic residue modulo any prime divisor of .
Now we state our first theorem.
Theorem 1.1**.**
Let be an odd integer.
(i)* If is not a quadratic residue modulo , then*
[TABLE]
(ii)* If is not a quadratic residue modulo , then*
[TABLE]
(iii)* If is not a quadratic residue modulo , then*
[TABLE]
(iv)* If is not a quadratic residue modulo , then*
[TABLE]
Combining Theorem 1.1 with (1.3), we immediately obtain the following consequence which was conjectured by Sun [11, Conjecture 4.8(ii)].
Corollary 1.1**.**
For any positive integer , we have
[TABLE]
and
[TABLE]
Actually we deduce Theorem 1.1 from the following theorems.
Theorem 1.2**.**
Let be a positive odd integer which is squarefree. For any , we have
[TABLE]
Theorem 1.3**.**
Let be a positive odd integer which is squarefree, and let . Then
[TABLE]
where the notation means that is a quadratic residue modulo .
Our following result was originally conjectured by Sun [11, Conjecture 4.8(iv)].
Theorem 1.4**.**
(i)* for any prime .*
(ii)* for any prime .*
In fact, our proof of Theorem 1.4 yields a stronger result: For each integer , we have
[TABLE]
for any prime , and
[TABLE]
for any prime .
We will prove Theorem 1.2, Theorems 1.3 and 1.1, and Theorem 1.4 in Sections 2-4 respectively.
Sun [11, Conjecture 4.8(iv)] also conjectured that for any prime . Moreover, Sun [10] conjectured that
[TABLE]
for any prime and integer , and this was confirmed by M. Stoll via two elliptic curves with complex multiplication by (see the answer in [10]).
For any prime and , we provide in Section 5 a sufficient condition for
[TABLE]
2. Proof of Theorem 1.2
Lemma 2.1**.**
Let be an odd prime and let with . Then
[TABLE]
Proof. If , then both sides of the congruence (2.1) are zero.
Below we assume and let denote the left-hand side of the congruence (2.1). As is a complete system of residues modulo , we have
[TABLE]
We may write with . For any integer , it is well known (cf. [4, p.Β 235]) that
[TABLE]
Therefore,
[TABLE]
Clearly,
[TABLE]
So, by the above, we finally obtain (2.1). β
Lemma 2.2**.**
Let be any odd prime. Then we have the congruence
[TABLE]
in the ring , where is the ring of all -adic integers, and is or [math] according as or not.
Remark 2.1*.*
For any prime , the congruence (2.3) is due to Sun [9, (1.15)]. We can easily verify that (2.3) also holds for .
Proof of Theorem 1.2. Clearly both sides of (1.5) vanish if . Below we assume and distinguish three cases.
Case 1. is an odd prime .
Define
[TABLE]
If and , then
[TABLE]
When and , for we have and (cf. [11, Remark 1.1 and Lemma 2.3]), thus
[TABLE]
Now suppose that . By Lemma 2.2,
[TABLE]
Combining this with Lemma 2.1, we obtain that
[TABLE]
Thus . Clearly .
If , then
[TABLE]
Now assume that . If neither nor is divisible by , then
[TABLE]
and
[TABLE]
hence is even. When , we also have since
[TABLE]
If , then since
[TABLE]
and
[TABLE]
So is always even, and hence as and .
Case 2. with , where are distinct primes.
By the Chinese Remainder Theorem,
[TABLE]
and hence
[TABLE]
Similarly,
[TABLE]
Thus, (1.5) holds in view of Case 1. This concludes the proof. β
3. Proofs of Theorems 1.3 and 1.1
Lemma 3.1**.**
Let be a prime. If and with and , then
[TABLE]
If , i.e., , then
[TABLE]
Remark 3.1*.*
The first assertion in Lemma 3.1 was conjectured by Z.-W. Sun [8] and confirmed by his twin brother Z.-H. Sun [7, Theorem 4.3]. The second assertion was proved by Z.-W. Sun [9, Corollary 1.3] as a consequence of (2.3) with .
Lemma 3.2**.**
Let be an odd prime and let with . Then
[TABLE]
Proof. Both sides of (3.1) vanish if . Below we assume .
Clearly,
[TABLE]
and
[TABLE]
So (3.1) holds. β
Lemma 3.3**.**
Let be any odd prime and let .
(i)* We have*
[TABLE]
(ii)* We have*
[TABLE]
Also,
[TABLE]
and
[TABLE]
Remark 3.2*.*
It is well known that any prime can be written as with . Also, for each any odd prime with can be written with (cf. [3]).
Proof of Lemma 3.3. It is easy to verify that (3.2)-(3.5) hold for . Below we assume .
(i) As , combining Lemma 2.1 and Lemma 3.1 we find that
[TABLE]
Observe that
[TABLE]
is even (since ), and its absolute value is smaller than . If with and , then . So (3.2) holds.
(ii) In light of Lemma 3.2,
[TABLE]
and
[TABLE]
On the other hand, by [1, Theorem 6.2.9] and [1, pp.β195-196],
[TABLE]
and
[TABLE]
Now we prove (3.5). Clearly, (3.5) is valid if or . Below we assume that and . Observe that
[TABLE]
By a result of Rajwade [6],
[TABLE]
Therefore (3.5) holds.
The proof of Lemma 3.3 is now complete. β
Proof of Theorem 1.3. Write with distinct primes. In light of (2.4) and Lemma 3.3(i), if fails (i.e., for some ) then
[TABLE]
Thus (1.7) holds. Note that if then for each we may write with and and hence
[TABLE]
Similarly, (1.6), (1.8) and (1.9) also hold in view of (2.4) and Lemma 3.3(ii). This concludes our proof of Theorem 1.3. β
Proof of Theorem 1.1. Suppose that , where are distinct primes and are positive integers. If with , then and hence for any we have
[TABLE]
for all . Therefore
[TABLE]
Below we assume that is squarefree. If fails, then by Theorems 1.2 and 1.3 we have
[TABLE]
for all , hence and . This proves part (i) of Theorem 1.1. Similarly, parts (ii)-(iv) of Theorem 1.1 follow from Theorems 1.2 and 1.3. This ends the proof. β
4. Proof of Theorem 1.4
Let be a prime power and let be the finite field of order . A multiplicative character on is called trivial (or principal) if for all . For a polynomial , we define the homogenous polynomial
[TABLE]
Fix a list of the elements of . For a multiplicative character on , we introduce the matrices
[TABLE]
Lemma 4.1**.**
Let be a prime power and let be a nontrivial multiplicative character on . Suppose that and . Then is singular i.e., . If the character is nontrivial with , then the matrix is singular too.
Proof.
We introduce the column vector whose coordinates are for . Let . Then, for any we have
[TABLE]
Since is a nonzero vector, the matrix is singular.
Now suppose that the degree of is and the character is nontrivial. Let and introduce the vector with coordinates for . Then for all as before. Let be the leading coefficient of the polynomial . Then
[TABLE]
Therefore is the zero vector and hence is singular. β
Motivated by Lemma 4.1, we give the following more sophisticated lemma.
Lemma 4.2**.**
Let be an odd prime power. Suppose that is not a square and is a nontrivial multiplicative character on with . Assume that and
[TABLE]
(i)* We have , in particular is singular.*
(ii)* Assume that the character with is nontrivial. Then .*
Proof.
For , set
[TABLE]
This is well defined since , The matrix has rank 2; in fact, if for some then columns and in are proportional, but columns 1 and are not proportional.
(i) Write for . It suffices to show that is the zero matrix. For , the -entry of the matric is
[TABLE]
where for any .
Now it remains to show for any the identity
[TABLE]
Clearly, for some and . Thus
[TABLE]
This proves part (i) of Lemma 4.2.
(ii) Write for , and define . (Note the slight difference between and .) The rank of is still equal to , so it suffices to show that is the zero matrix. Note that the -entry of is trivially zero if since for all . For we can repeat the computation for verbatim. Let denote the leading coefficient of . If and , then the -entry of is
[TABLE]
This is zero since is nontrivial. We are done. β
Theorem 4.1**.**
Let be an odd prime power and let with . Let be a nontrivial quadratic character on , and let
[TABLE]
with . Then
[TABLE]
If , then both and are singular, and moreover either of them has a kernel of dimension at least two.
Proof.
In view of Lemma 4.2, we only need to prove (4.5). As for all , it suffices to show that
[TABLE]
for any .
Clearly, is odd since . Recall that is the trivial character, and note that
[TABLE]
If is not a square in , then and hence (4.6) holds by the above.
Now assume that with . Since
[TABLE]
it remains to show that . Since and
[TABLE]
we have
[TABLE]
Thus as desired.
The proof of Theorem 4.1 is now complete. β
Proof of Theorem 1.4(i). Let be any prime with , and let be the quadratic character of with for all . Note that since . Clearly,
[TABLE]
Applying Theorem 4.1, we obtain that
[TABLE]
and
[TABLE]
Note that Sun stated in [S19, Remark 4.9] that if and only if . β
Let be a finite field of order . A polynomial is called a permutation polynomial if is bijective as a function on . If is a nontrivial multiplicative character on and is a permutation polynomial, then
[TABLE]
and also
[TABLE]
provided that .
Theorem 4.2**.**
Let be an odd prime power and let with . Let be a nontrivial multiplicative character on with . For the polynomial
[TABLE]
with , we have
[TABLE]
Moreover, if the character is nontrivial, then
[TABLE]
Proof.
Let . It is a classical result (cf. [5, pp.β355-357]) that the Dickson polynomial is a permutation polynomial on . For any , as , the polynomial is also a permutation polynomial on . Thus
[TABLE]
Combining this with Lemma 4.2, we immediately obtain the desired results. β
Proof of Theorem 1.4(ii). Let be any prime with . Then . Let be the quadratic character of with for all . Then since . Clearly is nontrivial and . Applying Theorem 4.2, we get that
[TABLE]
This concludes the proof. β
Note that actually our method to prove Theorem 1.4 yields a stronger result stated after Theorem 1.4 in Section 1.
5. A sufficient condition for
For an odd prime power , we let denote the quadratic multiplicative character on the finite field .
Let be a prime and let be a nonzero element of . If , then we define as an element with . When , the finite field contains an element with , and we denote such an by .
Theorem 5.1**.**
Let be a prime and let be nonnzero elements of the field . Let be or according as is or , and set
[TABLE]
Let be the number of -points on the affine curve . If , then
[TABLE]
For the sake of convenience, for an odd prime we introduce the following two polynomials over :
[TABLE]
and
[TABLE]
Lemma 5.1**.**
Let be a prime with , and let . Define
[TABLE]
Viewing as an element of , we have
[TABLE]
Consequently, if or .
Proof.
As , we have . So, the second assertion in Lemma 5.1 follows from the first one.
Now we come to prove the first assertion. With the help of (2.2), in we have
[TABLE]
(Note that if are nonnegative integers with then ) Thus,
[TABLE]
as desired.
β
Lemma 5.2**.**
Let be an odd prime and let with . For any polynomial
[TABLE]
we have
[TABLE]
Proof.
As the multiplicative group is cyclic, similar to (2.2), for each we have
[TABLE]
where we treat as when . Note also that
[TABLE]
for all integers . Thus
[TABLE]
where means that the sum is taken over all subject to the condition
[TABLE]
Write , where . Then
[TABLE]
Note that
[TABLE]
and the equality is possible only if (i.e., ) or (i.e., ). Since , if then we obtain step by step that (i.e., ).
Combining the above, we finally obtain (5.3). β
Lemma 5.3**.**
Let be an odd prime and let with . Let and set
[TABLE]
Viewing as an element of , we have
[TABLE]
Proof.
Write
[TABLE]
In view of LemmaΒ 5.2, we have
[TABLE]
Clearly,
[TABLE]
Note also that
[TABLE]
and hence
[TABLE]
Combining (5.6) with (5.7) and (5.8), we immediately obtain the desired (5.5). β
Now we study further properties of the polynomials and defined by (5.1) and (5.2). They may be viewed as truncated versions of certain hypergeometric series.
Lemma 5.4**.**
Let be an odd prime and let with .
(i)* A polynomial with satisfies the differential equation*
[TABLE]
if and only if for some .
(ii)* Suppose that . Then a polynomial with satisfies the differential equation*
[TABLE]
if and only if for some .
Proof.
It is straightforward to verify that and satisfy (5.9) and (5.10) respectively. So, the βifβ parts of (i) and (ii) are easy.
Now we prove the βonly ifβ part of (i). If a polynomial with satisfies (5.9), then there is a constant such that is a solution of (5.9) with . Thus, it suffices to show that (5.9) has no nonzero solution with . In fact, the coefficient of in is provided .
Similarly, we can show the βonly ifβ part of (ii). β
Lemma 5.5**.**
Let be a prime with . Then
[TABLE]
Proof.
Clearly, is a polynomial of degree with the leading coefficient . A direct computation based on (5.10) shows that satisfies (5.9). Now we apply LemmaΒ 5.4 and compare the leading terms of both sides. Since
[TABLE]
we immediately get the desired result. β
Proof of TheoremΒ 5.1.
Since
[TABLE]
the assumption , together with LemmaΒ 5.3 in the case , implies that . As , we have by LemmaΒ 5.5. Applying LemmaΒ 5.1 we obtain the desired result. β
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] B. C. Berndt, R. J. Evans and K. S. Williams, Gauss and Jacobi Sums , Wiley, New York, 1998.
- 2[2] R. Chapman, Determinants of Legendre symbol matrices , Acta Arith. 115 (2004), 231β244.
- 3[3] D. A. Cox, Primes of the Form x 2 + n β y 2 superscript π₯ 2 π superscript π¦ 2 x^{2}+ny^{2} , John Wiley & Sons, 1989.
- 4[4] K. Ireland and M. Rosen, A Classical Introduction to Modern Number Theory , 2nd Edition, Grad. Texts. Math., vol. 84, Springer, New York, 1990.
- 5[5] R. Lidl and H. Niederreiter, Finite Fields , 2nd Edition, Encyclopedia of Math. and its Applications, 20, Cambridge Univ. Press, Cambridge, 1997.
- 6[6] A. R. Rajwade, The Diophantine equation y 2 = x β ( x 2 + 21 β D β x + 112 β D 2 ) superscript π¦ 2 π₯ superscript π₯ 2 21 π· π₯ 112 superscript π· 2 y^{2}=x(x^{2}+21Dx+112D^{2}) and the conjectures of Birch and Swinnerton-Dyer , J. Austral. Math. Soc. Ser. A 24 (1977), 286β295.
- 7[7] Z.-H. Sun, Congruences involving Legendre polynomials II , J. Number Theory 133 (2013), 1950β1976.
- 8[8] Z.-W. Sun, Super congruences and Euler numbers , Sci. China Math. 54 (2011), 2509β2535.
