Distribution of squarefree integers with double congruence conditions
Abstract
Given a positive integer with reduced residue system , we partition the set of squarefree integers according to some congruence conditions on the number of prime factors in each class , and prove two relevant equidsitribution results in this setting. We discover the bias phenomenon in some special cases, and formulate a conjecture with numerical support.
1 Introduction
Let denote the number of squarefree numbers that are and congruent to , modulo , where . In 1885 Gegenbauer[3] proved that squarefree integers have a density of among all integers, that is
Landau[5] proved in 1909 that
| (1) |
In 1974 Hooley[4] refined this result to
This is the best possible individual error term bound we have obtained so far. On the other hand, the square mean of the error term, i.e.,
are studied by many people including Valentin Blomer, Ramon M. Nunes, Pierre Le Boudec. The best estimates can be found in [6] and [1].
Here we will consider a new distribution pattern of squarefree integers. Let be an integer , , where is the Euler totient function, that is . Let be a set of representatives of . Given squarefree and coprime with M, write ,where are the prime divisors of . Define .
The first main result of this paper is an equidistribution.
Theorem 1.1.
Let , and where . Let . Define:
Then we have
Theorem 1.1 shows that are equal-distributed. We also want to consider bias among , that is, given distinct , does the sign of stablize as ? Generally, there is no bias among . However in Β§4.1 we discover that under certain conditions there should be a strict bias among . The computation there leads us to raise the following conjecture.
Conjecture.
When for all , there exists a strict bias among . That is, given distinct , the sign of always stablizes as .
The second main result of this paper studies a βfiniteβ version of .
Theorem 1.2.
Let , where . Write
Define:
Then we have
Remark.
Given distinct , under the condition , we have:
Where is the multinomial coefficient and equal to
The estimate of refines the estimate of . More precisely, let
so that . In the setting of Theorem 1.1, we choose to be the order of in the group . Let denote the set of all . We can give an additive group structure by identifying with . Then
gives a group homomorphism from to . We claim that
| (2) |
For , for some . Write , where are the prime divisors of . Recall , and . Let . For , let
Then . . Since is the order of in the group , This gives , and .
Conversely, given , decompose and introduce , , as above. Then , with , and
2 Tauberian Theorems
We need the following lemmas which connect the pole of a Dirichlet series with the order of growth of its partial sum of its coefficients.
The first lemma is due to Selberg-Delange [7]:
Let We say that a Dirichlet series has property if the Dirichlet series may be analytically continued to , and, in this domain, satifies the bound
We will say has a pole of order at . Here
If has property and there exist real numbers such that and satisfies for some We say has property
Lemma 2.1.
Given with property , we have:
where
and
Here is defined by the Taylor series coefficient:
And in the domain where is holomorphic, we set
The next Lemma is due to Delange[2]:
Lemma 2.2.
Let be a Dirichlet series with , and converges for . Suppose is holomorphic at all points of the line except at , and, in a neighborhood of in the half plane , we have:
where is a negative integer,and are holomorphic at , with . Then as
3 Proof
3.1 Congruence restriction on the number of prime divisors
Proof of Theorem 1.1.
For each , let denote a -th root of unity. Define
For each , define
| (3) |
Define
can be written as a linear combination of :
Especially, the coefficient of is for all .
Define
Note that for a Dirichlet character mod :
The second term converges absolutely for , hence in a neighborhood of we can write
Therefore we have
Where ranges over Dirichlet characters mod . For all nontrivial Dirichlet characters , converges at and . For the trivial character , has a simple pole at , that is to say for :
Hence
Therefore, has a pole of order at .
Notice that , where is bounded independently with , therefore, which converges at , hence has a pole of order at as well. In addition, the real part of the order of the pole is always unless .
Let be the Dirichlet series expansion, and let . Then
We can now apply Lemma 2.1, to , and we have
which when .
has a regular pole at with residue at , hence we have:
This gives
Remark.
If we apply Lemma 2.1 to all of in , we can refine the error term to . However the power of in the error term can never be reduced below with this method alone. Another result is that if we consider , the term cancels out and the resulting difference has to be of the form . If given specific , we may directly calculate , and would imply the sign of changes infinitely many times as . We conjecture that this is true for any distinct .
β
3.2 Restriction of the number of prime divisors
Proof of Theorem 1.2.
Let , where is in the -th component.
Define:
Then we have:
Assume . Consider the Dirichlet series expansion for . , has coefficient . On the other hand, the terms not in all contain for some . This gives
Here all summations are over prime and . We have
If , then we have
In any case, we have
Hence
Apply Lemma 2.2 to and we have:
β
4 Some possible Bias Among
Given and , we consider the sign of , where and are distinct. Weβll show that if some , this sign changes infinitely many times, and if for all , we may expect the sign to become stable as . We shall justify this phenomenon for certain specific .
4.1 for All
After re-indexing, we may assume , and . If , there would be only one element in and it makes no sense to talk about bias. So we may assume . Weβll use to denote the elements of , and let denote the subset of A with exactly one index being , and the rest being . For defined in (3), let denote the number of indices that are equal to , then has a pole of order at . Hence, among the collection of , has the second highest order pole of order after , which has a pole of order 1 at .
Given , let be the element whose -th component equals . Then
Let be the Dirichlet series expansion of and let . By Lemma 2.1 we have:
We will need the following lemma for the purpose of numerically computing .
Lemma 4.1.
We have the equality of the two following limits:
Proof.
Define
Then we have
Here converges absolutely on , hence we can exchange with .
Let
Let
We need to prove: can be extended to , and
Where , and by the mean value theorem we have:
where is a positive constant. This means converges uniformally for . Hence we can exchange the limit and valuation at . β
In particular, we have , since both and are real when .
Given , we have
| (4) |
Hence for distinct , βs term cancels out, but , terms need not. If they donβt, the sign of the leading term in the asymptotic expansion of is fixed.
Note that, for , the coefficient of in (4) is maximal for every , while for , this coefficient is minimal. We deduce the following result.
Proposition 4.2.
For all , is positive when is sufficiently large. For all , is negative when is sufficiently large.
For given , we can always justify whether or not there is a bias by numeric method. We describe the method with the following example:
Let . Then . Let . Let . Let and . Then . We have:
Where:
Denote , and , the result for the computation is as follows (where we choose , and is the -th prime number):
| 100 | 1,000 | 10,000 | 100,000 | 1,000,000 | |
|---|---|---|---|---|---|
| 1.16546 | 1.17505 | 1.1763 | 1.1766 | 1.17678 | |
| 0.744375 | 0.740081 | 0.740074 | 0.740005 | 0.739976 |
This shows that is always negative when is sufficiently large. We will write in this case.
By computing to the one million-th prime, we have calculated that , and we have the following result:
We have also computated the case for , and for all . Here is defined similarly in the above case, where
and all calculations are done to the one million-th prime.
When , let . We have the following result:
When , let . We have the following result:
When , let . We have the following result:
We conjecture that there is always a strict bias between and when and are distinct, that is, there is a strict total ordering on . Obviously, is the largest element in this ordering and is the smallest element in this ordering.
4.2 for Some
Let be the largest value among . Then the second highest order pole among would be and (by highest we mean the value of its real part being largest).
Hence by 2.1 we know that . Just like 4.1, this implies unless the coefficients happen to cancel out, we can expect the sign of to change infinitely many times.
Weβll consider the following example:
Let , , , let , and . Let and . Then .
where
As , the sign of changes infinitely many times as .
We conjecture that the sign of changes infinitely many times for all distinct , .
Acknowledgments
The author would like to thank Xi Ping for sharing the knowledge on Selberg-Delange method for this problem, and his advisor Wu Han for his guidance and advice on the writing of this paper.
References
- [1] (2014) On the distribution of squarefree integers in arithmetic progressions. External Links: 1411.2360, Link Cited by: Β§1.
- [2] (1954) GΓ©nΓ©ralisation du thΓ©orΓ¨me de ikehara. Annales scientifiques de lβΓcole Normale SupΓ©rieure 71 (3), pp.Β 213β242 (fre). External Links: Link Cited by: Β§2.
- [3] (1938) An introduction to the theory of numbers.. (English). Note: Oxford: Clarendon Press. xvi, 403 p. (1938). Cited by: Β§1.
- [4] (1975) A note on square-free numbers in arithmetic progressions. Bulletin of the London Mathematical Society 7 (2), pp.Β 133β138. External Links: Document, Link, https://londmathsoc.onlinelibrary.wiley.com/doi/pdf/10.1112/blms/7.2.133 Cited by: Β§1.
- [5] (1911) Handbuch der lehre von der verteilung der primzahlen. Monatshefte fΓΌr Mathematik und Physik 22 (1). External Links: ISSN 1436-5081, Document, Link Cited by: Β§1.
- [6] (2014) Square-free numbers in arithmetic progressions. External Links: 1402.0684, Link Cited by: Β§1.
- [7] (2015) Introduction to analytic and probabilistic number theory. Vol. 163, American Mathematical Soc.. Cited by: Β§2.
- [8] (1936) Zur additiven zahlentheorie. ii.. Mathematische Zeitschrift 40 (1), pp.Β 592β607. External Links: ISSN 1432-1823, Document, Link Cited by: Β§4.1.