arXiv:1810.00227v2 [math.NT] 29 Jan 2019 1] 1] 1] 1] 1] 1] 1] 2] 2] 2] 2]ad[24]). and [23] [22], [21], [20], instance, [18], for [17], (see [16], literature [15], the [13], [12], in [10], interest great with considered aua ubr hc r utpe of multiples are which numbers natural o-eius i inside lie non-residues) td fdsrbto fqartcrsde n udai non-res quadratic and residues quadratic of distribution of instudy solution a has udai o-eius modulo non-residues) quadratic e,oemyrfrt 8,[]ad[1.Frtw hl xsm notation some fix shall we First [11]. and [9] [8], to by refer denote may one few, a let we over residue n h aersl stu for true is result same the and modulo (1) theorems. usin1. Question n n a ocueta hr r exactly are there that conclude can one and ril) oee,i ih apnta o oeprimes some for that happen might it However, article). N uteqeto.I hsatce esalase hsuigteDirichle the using this answer shall we field article, the for this formula In question. subtle ( Since nteltrtr,teeaemn aesadesdsmlrt Ques to similar addressed papers many are there literature, the In Let h tnadtcnqe naayi ubrter nwr h ab the answers theory number analytic in techniques standard The ,S p, ATACATPDYY IIH O,SBASRA N .THAN R. AND SARKAR SUBHA ROY, CHATTOPADHYAY, BIDISHA JAITRA Z p k /p mgnr udai field quadratic imaginary yn nteitra [1 interval the in lying udai eiusadnnrsde modulo non-residues and residues quadratic Abstract. S or ) modulo p Z ea d rm ubr number A number. prime odd an be k ITIUINO EIUSMODULO RESIDUES OF DISTRIBUTION Z /p nteitra [1 interval the in Q As . = Q ( Z ,S p, { ( o nodpienumber prime odd an For ,S p, a safil,tepolynomial the field, a is p p k { ∈ IIHE’ LS UBRFORMULA NUMBER CLASS DIRICHLET’S ftecongruence the if , (respectively, ) Z sa d rm,w e that see we prime, odd an is Let k ) Otherwise, . N < 1 Q p , ( 2 √ ea d rm ubr nti ril,w td h ubrof number the study we article, this In number. prime odd an be p , . . . , S ( − k ,S p, p , ? p , p ,when ), − Q Q − k ( ] yapyn h iihe’ ls ubrfruafrthe for formula number class Dirichlet’s the applying by 1], .Uigtesadr ehius ecudntase this answer not could we techniques, standard the Using ). ( − √ ,S p, 1]. N p N − a 1 ntesubset the in ( p } ( 1. k ssi ob a be to said is ,S p, ). ,S p, x = ) k 2 Introduction : 2 = ≡ k k p o all for ) p )tenme fqartcrsde (respectively, residues quadratic of number the )) a X n ie aua number natural given a and k a 2 − , o ayqartcrsde (respectively, residues quadratic many How . k p ≡ r4 oepeiey epoetefollowing the prove we precisely, More 4. or 3 − 1 p 1 (mod − 1 2 (mod 0 a p 1 + − { ∈ udai eiusa ela non-residues as well as residues quadratic hc r utpe f2o r4and 4 or 3 or 2 of multiples are which X S O a precisely has 1 k k p ( udai non-residue quadratic p − w hl epoigti ati this in fact this proving be shall (we √ fteitra [1 interval the of 1 ) 1 p , . . . , p − k log ) ( = 1 } p etesbe ossigo all of consisting subset the be ) − p emyhave may we , 1 X p } ( de modulo idues p p ssi ob a be to said is SN THE USING − 1,[] 4,[] 6,[7], [6], [5], [4], [3], [1], − 1) p , / k o-eosolutions non-zero 1 2 in1adt name to and 1 tion v usinas question ove 1)( + with − ’ ls number class t’s sflos We follows. as s 1]. modulo 1 GADURAI X ≤ Q ( p p − k ( quadratic a been has ,S p, 1) ≤ p / The . 2 p k − ) − 1) 1 > , 2 JAITRA CHATTOPADHYAY, BIDISHA ROY, SUBHA SARKAR AND R. THANGADURAI

1 Theorem 1. Let p be an odd prime. If p 3 (mod 4), then for any ǫ with 0 <ǫ< 2 , we have ≡ p 1 1 Q(p,S ) − p 2 −ǫ. 2 − 4 ≫ǫ When the prime p 1 (mod 4), we have ≡ p 1 Q(p,S )= − . 2 4 Corollary 1.1. Let p be an odd prime and let be the set of all odd integers in [1,p 1]. If R = N(p,S ) or R = Q(p, ), then for any Oǫ with 0 <ǫ< 1 , we have − 2 O 2 p 1 1 − R p 2 −ǫ, if p 3 (mod4). 4 − ≫ǫ ≡ When the prime p 1 (mod 4), we have ≡ p 1 R = − . 4 1 Theorem 2. Let p be an odd prime. If p 1, 11 (mod 12), then for any ǫ with 0 <ǫ< 2 , we have ≡ p 1 1 Q(p,S ) − p 2 −ǫ. 3 − 6 ≫ǫ When p 5, 7 (mod 12), in this method, we do not get any finer information other than in (1).≡ Corollary 1.2. Let p be an odd prime. If p 1, 11 (mod 12), then for any ǫ with 1 ≡ 0 <ǫ< 2 , we have p 1 1 − N(p,S ) p 2 −ǫ. 6 − 3 ≫ǫ Theorem 3. Let p be an odd prime. Then, for p 3 (mod 8), we have ≡ 1 p 1 Q(p,S )= − . 4 2  4  1 Also, for any 0 <ǫ< 2 , we have p 1 1 Q(p,S ) − p 2 −ǫ, if p 1 (mod4), 4 − 8 ≫ǫ ≡ and 1 p 1 1 Q(p,S ) − p 2 −ǫ; if p 7 (mod8). 4 − 2  4  ≫ǫ ≡ Corollary 1.3. Let p be an odd prime. Then, for p 3 (mod 8), we have ≡ 1 p 1 N(p,S )= − . 4 2  4  1 Also, for any 0 <ǫ< 2 , we have p 1 1 − N(p,S ) p 2 −ǫ; if p 1 (mod4), 8 − 4 ≫ǫ ≡ and 1 p 1 1 − N(p,S ) p 2 −ǫ; if p 7 (mod8). 2  4  − 4 ≫ǫ ≡ DISTRIBUTION OF RESIDUES MODULO p USING THE DIRICHLET’S CLASS NUMBER FORMULA3

Using Theorems 1 and 3, we conclude the following corollary. Corollary 1.4. Let p be an odd prime such that p 3 (mod 8). Then for any ǫ with 1 ≡ 0 <ǫ< 2 , we have 1 p 1 1 Q(p,S S ) − p 2 −ǫ. 2\ 4 − 2  4  ≫ǫ

2. Preliminaries In this section, we shall state many useful results as follows. Theorem 4. (Polya-Vinogradov) Let p be any odd prime and χ be a non-principal Dirich- let character modulo p. Then, for any integers 0 M < N p 1, we have ≤ ≤ − N χ(m) √p log p. ≤ mX=M

Let us define the following counting functions as follows. Let 1 x (2) f(x)= 1+ for all x (Z/pZ)∗ 2   p  ∈ and 1 x (3) g(x)= 1 for all x (Z/pZ)∗ 2  −  p  ∈ where · is the Legendre symbol. Then, we have p 1; if x is a (mod p), f(x)=  0; otherwise. and 1; if x is a quadratic nonresidue (mod p), g(x)=  0; otherwise. In the following lemma, we prove the “expected” result.

Lemma 1. For an integer k 1 and an odd prime p, let Sk = kI where I is the interval I = 1, 2,..., [(p 1)/k] . Then≥ { − } (p−1)/k 1 p 1 1 k m (4) Q(p,S )= − + k 2  k  2  p  p  mX=1 and hence 1 p 1 Q(p,S )= − + O(√p log p). k 2  k 

The same expressions hold for N(p,Sk) as well. 4 JAITRA CHATTOPADHYAY, BIDISHA ROY, SUBHA SARKAR AND R. THANGADURAI

Proof. We prove for Q(p,Sk) and the proof of N(p,Sk) follows analogously. Let ψk be the characteristic function for Sk which is defined as

1; if m Sk, ψk(m)= ∈  0; if m Sk. 6∈ Now, by (2), we see that

p−1 p−1 1 m Q(p,S ) = f(m)= ψ (m)f(m)= ψ (m) 1+ k k 2 k   p  mX∈Sk mX=1 mX=1 (p−1)/k 1 p 1 1 k m (5) = − + , 2  k  2  p  p  mX=1 which proves (4). Then, by Theorem 4, we get 1 p 1 Q(p,S )= − + O(√p log p). k 2  k  This finishes the proof.  Let q > 1 be a positive integer and let ψ be a non-trivial quadratic character modulo ∞ ψ(n) q. Let L(s, ψ) = be the Dirichlet L-function associated to ψ. Since ψ is a ns Xn=1 non-trivial homomorphism, L(s, ψ) admits the following Euler product expansion ψ(p) L(s, ψ)= 1  − ps  Yp∤q for all complex number s with (s) > 1. This, in particular, shows that L(s, ψ) > 0 for all real number s > 1. By continuity,ℜ it follows that L(1, ψ) 0. Dirichlet proved that L(1, ψ) = 0 in order to prove the infinitude of prime numbers≥ in an arithmetic progression. Hence,6 it follows that L(1, ψ) > 0 for all non-trivial quadratic character ψ. Since L(1, ψ) > 0, it is natural to expect some non-trivial lower bound as a function of q. This is what was proved by Landau-Siegel in the following theorem. The proof can be found in [14]. Theorem 5. Let q > 1 be a positive integer and ψ be a non-trivial quadratic character modulo q. Then for each ǫ> 0, there exists a constant C(ǫ) > 0 such that C(ǫ) L(1, ψ) > . qǫ The following lemma is crucial for our discussions. This lemma connects the sum of Legendre symbols and the Dirichlet L-function associated to Legendre symbol via the famous Dirichlet for the quadratic field. For an odd prime p, the Legendre symbol · = χ ( ) is a quadratic Dirichlet character modulo p. We also define p p · a character ( 1)(n−1)/2; if n is odd, χ4(n)= −  0; otherwise. DISTRIBUTION OF RESIDUES MODULO p USING THE DIRICHLET’S CLASS NUMBER FORMULA5

Then one can define the Dirichlet character χ4p as χ4p(n)= χ4(n)χp(n) for any odd prime p and similarly, we can define χ3p(n)= χ3(n)χp(n) for any odd prime p> 3. Clearly, χ4p and χ3p are non-trivial and real quadratic Dirichlet characters.

Lemma 2. (See for instance, Page 151, Theorem 7.2 and 7.4 in [24]) Let p > 3 be an odd prime and for any real number ℓ 1, we define ≥ (6) S(1,ℓ)= χp(m). 1≤Xm<ℓ Then we have the following equalities. (1) For a prime p 3 (mod 4), we have ≡ √p S(1,p/2) = (2 χ (2)) L(1, χ ), π − p p where L(1, χp) is the Dirichlet L-function; Also, we have √p S(1,p/3) = (3 χ (3))L(1, χ ). 2π − p p (2) For a prime p 1 (mod 4), we have ≡ √3p S(1,p/3) = L(1, χ ); 2π 3p Also, we have √p S(1,p/4) = L(1, χ ). π 4p Now, we need the following lemma, which deals with the vanishing sums of Legendre symbols. This was proved in [2]. For more such relations one may refer to [8]. Lemma 3. [2] Let p be an odd prime. Then the following equalities hold true. (p−1)/2 n (1) If p 1 (mod 4), then we have =0. ≡  p  Xn=1 ⌊p/4⌋ n (2) If p 3 (mod 8), then we have =0. ≡  p  Xn=1 ⌊p/2⌋ n (3) If p 7 (mod 8), then we have =0. ≡  p  ⌈Xp/4⌉

3. Proof of Theorem 1

Let p be a given odd prime. We want to estimate the quantity Q(p,S2). Therefore, by (5), we get (p−1)/2 1 p 1 1 2 n (7) Q(p,S )= − + . 2 2  2  2 p  p  Xn=1 Now, we consider three cases as follows. Case 1. p 1 (mod 4) ≡ 6 JAITRA CHATTOPADHYAY, BIDISHA ROY, SUBHA SARKAR AND R. THANGADURAI

(p−1)/2 n In this case, since = 0, by Lemma 3 (1), the equation (7) reduces to  p  Xn=1 p 1 Q(p,S )= − , 2 4 which is as desired.

Case 2. p 3 (mod 8) ≡ By Lemma 2 (1) and by (7), we get 1 p 1 √p Q(p,S )= − + (2 χ (2)) L(1, χ ). 2 2  2  π − p p

2 In this case, we know that p = 1. Therefore, we get   − 1 p 1 √p Q(p,S )= − +3 L(1, χ ). 2 2  2  π p 1 Let ǫ be any real number such that 0 <ǫ< 2 . Then by Theorem 5, we get

1 p 1 1 Q(p,S ) − p 2 −ǫ, 2 − 2  2  ≫ǫ as desired.

Case 3. p 7 (mod 8). ≡ 2 Since p 7 (mod 8), we know that = 1. Therefore, by Lemma 2 (1) and by (7), ≡ p we get 1 p 1 √p 1 p 1 √pL(1, χp) Q(p,S )= − + L(1, χ )= − + . 2 2  2  π p 2  2  π 1 Let ǫ be any real number such that 0 <ǫ< 2 . Then by Theorem 5 we get

1 p 1 1 Q(p,S ) − p 2 −ǫ 2 − 2  2  ≫ǫ which proves the theorem. 

4. Proof of Theorem 2

Let p be a given odd prime. We want to estimate the quantity Q(p,S3). Therefore, by (5), we get,

(p−1)/3 1 p 1 3 n (8) Q(p,S )= − + . 3 2  3  p  p  Xn=1 Now, we consider the following cases.

Case 1. p 1 (mod 12) ≡ DISTRIBUTION OF RESIDUES MODULO p USING THE DIRICHLET’S CLASS NUMBER FORMULA7

3 Note that, in this case, we have = 1. By (8) and by Lemma 2 (2), we get p 1 p 1 1 √3p Q(p,S ) − = L(1, χ χ ) 3 − 2  3  2 2π 3 p √3p C(ǫ) ≥ 4π (3p)ǫ 1 p 2 −ǫ, ≫ǫ 1 for any given 0 <ǫ< 2 in Theorem 5. Case 2. p 11 (mod 12) ≡ 3 In this case, we have, = 1. Then again by (8) and by Lemma 2 (1), we get p 1 p 1 1 √3p Q(p,S ) = − + (3 χ (3))L(1, χ ). 3 2  3  2 2π − p p Hence 1 p 1 1 Q(p,S ) − p 2 −ǫ, 3 − 2  3  ≫ǫ 1  for any 0 <ǫ< 2 in Theorem 5.

5. Proof of theorem 3 At first, using the equation (5), we note that (p−1)/4 (p−1)/4 1 p 1 1 4 m 1 p 1 1 m (9) Q(p,S )= − + = − + . 4 2  4  2 p  p  2  4  2  p  mX=1 mX=1

Case 1. p 1 (mod 4) ≡ Now, we apply Lemma 2 (2) in (9) and we get 1 p 1 1 √p Q(p,S ) = − + L(1, χ χ ). 4 2  4  2 π 4 p Hence p 1 1 Q(p,S ) − p 2 −ǫ, 4 − 8 ≫ǫ 1 for any 0 <ǫ< 2 in Theorem 5. Case 2. p 3 (mod 8) ≡ [(p−1)/4] m In this case, we apply Lemma 3 (2) which says that = 0. Hence, by (9),  p  Xn=1 we get 1 p 1 Q(p,S )= − . 4 2  4  Case 3. p 7 (mod 8) ≡ 8 JAITRA CHATTOPADHYAY, BIDISHA ROY, SUBHA SARKAR AND R. THANGADURAI

First note that by Lemma 3 (3), we have m =0. p−1 p−1  p  4

[15] L. Moser, On the equation φ(n)= π(n), Pi Mu Epsilon J., (1951), 101-110. [16] P. Pollack, The least prime quadratic non-residue in a prescribed residue class mod 4, J. Number Theory, 187 (2018), 403-414. [17] M. Szalay, On the distribution of the primitive roots mod p (in Hungarian), Mat. Lapok, 21 (1970), 357-362. [18] M. Szalay, On the distribution of the primitive roots of a prime, J. Number Theory, 7 (1975), 183-188. [19] J. Tanti and R. Thangadurai, Distribution of residues and primitive roots, Proc. Indian Acad. Sci., 123(2) (2013), 203-211. [20] E. Vegh, Primitive roots modulo a prime as consecutive terms of an arithmetic progression, J. Reine Angew. Math. 235 (1969), 185-188. [21] E. Vegh, Arithmetic progressions of primitive roots of a prime. II, ibid. 244 (1970), 108-111. [22] E. Vegh, A note on the distribution of the primitive roots of a prime, J. Number Theory, 3 (1971), 13-18. [23] E. Vegh, Arithmetic progressions of primitive roots of a prime. III, J. Reine Angew. Math., 256 (1972), 130-137. [24] S. Wright, Quadratic residues and non-residues: Selected topics, arXiv:1408.0235v7. [25] A. Weil, On the Riemann hypothesis, Proc. Nat. Acad. Sci. U.S.A, 27 (1941), 345-347.

(Jaitra Chattopadhyay, Bidisha Roy, Subha Sarkar and R. Thangadurai) Harish-Chandra Re- search Institute, HBNI, Chhatnag Road, Jhunsi, Allahabad 211019, India E-mail address, Jaitra Chattopadhyay: [email protected] E-mail address, Bidisha Roy: [email protected] E-mail address, Subha Sarkar: [email protected] E-mail address, R. Thangadurai: [email protected]