Preprint
Article

This version is not peer-reviewed.

Inequalities with Some Arithmetic Functions

A peer-reviewed article of this preprint also exists.

Submitted:

02 March 2025

Posted:

03 March 2025

You are already at the latest version

Abstract
In the paper are formulated and proved some new inequalities with the classical arithmetic functions φ (of Euler) and ψ (of Dedekind).
Keywords: 
;  ;  ;  

1. Introduction

For a positive integer n = i = 1 r p i α i > 1 , let φ ( n ) and ψ ( n ) , respectively, denote the Euler and Dedekind totient function values, i.e.,
φ ( n ) = n p | n 1 1 p = i = 1 r p i α i 1 ( p i 1 ) ,
ψ ( n ) = n p | n 1 + 1 p = i = 1 r p i α i 1 ( p i + 1 ) ,
where p runs through the prime divisors of n, and for any i ( 1 i r ) p i are different primes and α i are positive integers (see, e.g., [1]).
Let
φ ( 1 ) = ψ ( 1 ) = 1 .
Another arithmetic function, which will be used, is the "core" function of n:
γ ( n ) = p | n p = i = 1 r p i .
Let us also denote
ω ( n ) = r ,
i.e., the number of the distinct prime factors of n, and
Ω ( n ) = i = 1 r α i ,
i.e., the total number of prime factors of n (see [1]).
The aim of this paper is to obtain certain new inequalities for these functions.

2. Main results

Theorem 1.
For n > 1 one has
ψ ( n ) φ ( n ) 2 ω ( n ) n γ ( n ) > 2 ω ( n ) φ ( n ) γ ( n ) .
Proof. 
The first inequality of (4) is proved in [2], where the arithmetic inequality
i = 1 r ( y i + 1 ) i = 1 r ( y i 1 ) 2 r
is used for y i 2 . By putting y i = p i ( 1 i r ) and using (1) and (2), the result follows. The second inequality of (4) follows by the classical inequality n > φ ( n ) for n 2 . □
Corollary 1.
For n > 1 ,
ψ ( n ) φ ( n ) 1 > 2 ω ( n ) γ ( n ) .
This follows by the weaker inequality in (4), by dividing both terms with φ ( n ) .
Theorem 2.
For n > 1 one has
ψ ( n ) n 1 + ω ( n ) γ ( n ) φ ( n ) n + 2 ω ( n ) γ ( n ) .
Proof. 
By (1), (2) and (3), the first inequality of (6) can be rewritten as
i = 1 r ( p i + 1 ) i = 1 r p i 1 + r i = 1 r p i = i = 1 r p i + r = γ ( n ) + r .
Relation (7) follows from the fact that
i = 1 r ( p i + 1 ) i = 1 r p i + i = 1 r p i i = 1 r p i + r
with an equality only for r = 1 . The second inequality of (6) can be rewritten as
i = 1 r p i i = 1 r ( p i 1 ) + r 2 r 0 .
For r = 1 the statement (8) is obvious, for r = 2 we obtain:
p 1 p 2 ( p 1 1 ) ( p 2 1 ) + 2 4 = p 1 + p + 2 3 > 0 .
For r 3 we have that p 1 2 , p 2 3 , p 3 , , p r 5 and
i = 1 r p i i = 1 r ( p i 1 ) + r 2 r > i = 1 r ( p i 1 ) i = 1 r 1 + 1 p i 1 2 r i = 1 r ( p i 1 ) 8 i = 4 r ( p i 1 ) 1 2 r 8 i = 4 r ( p i 1 ) 0 ,
which proves the theorem. □
We can note that the relation (6) offers an improvement of the first inequality of (4).
Corollary 2.
For n > 1
φ ( n ) n n ω ( n ) γ ( n ) .
Indeed, the second inequality of (6) can be written as
φ ( n ) n ( 2 ω ( n ) ω ( n ) ) n γ ( n ) .
Obviously, relation (10) is stronger than (9), as 2 r r r , i.e., 2 r 2 r for each r 1 . Relation (9) can be rewritten also as
n φ ( n ) γ ( n ) γ ( n ) ω ( n ) .
The following refinement of this inequality holds true:
Theorem 3.
For n > 1 so that ω ( n ) 2 ,
n φ ( n ) 1 + 2 ω ( n ) γ ( n ) γ ( n ) γ ( n ) ω ( n ) .
Proof. 
The first inequality of (11) can be written as
i = 1 r p i i = 1 r p i i = 1 r ( p i 1 ) 2 r i = 1 r ( p i 1 ) .
Obviously, for r = 1 , (12) is valid only for p 1 = 2 .
Let r = ω ( n ) 2 . As
i = 1 r p i > i = 1 r ( p i 1 ) ,
it will be sufficient to prove that
i = 1 r p i i = 1 r ( p i 1 ) > 2 r .
Since p 1 2 , p 2 3 we obtain that:
i = 1 r p i i = 1 r ( p i 1 ) 2 r 6 i = 3 r p i 2 i = 3 r ( p i 1 ) 2 r > 4 i = 3 r p i 2 r = 0 , i f r = 2 > 0 , i f r 3 .
The second inequality of (11) can be written as
γ ( n ) 2 γ ( n ) + 2 ω ( n ) γ ( n ) ω ( n ) γ ( n ) 2 ω ( n ) 2
or
γ ( n ) 2 ω ( n ) ,
which is true, because
γ ( n ) = i = 1 r p i 2 r 2 r .
Theorem 4.
For n > 1
ψ ( n ) n 2 1 + n φ ( n ) n 1 + ω ( n ) γ ( n ) .
Proof. 
From the first inequality of (11) one has
1 + n φ ( n ) 2 1 + ω ( n ) γ ( n ) ,
so, the second inequality of (13) follows.
The first inequality of (13) is due to Ch. R. Wall, but without a proof; it has been proved in [3] in the form
2 ψ ( n ) n 1 + n φ ( n ) .
Theorem 5.
For n > 1
n ω ( n ) γ ( n ) n φ ( n ) ψ ( n ) n .
Proof. 
The first inequality of (14) is exactly Corollary 2 (see (9)). The second inequality, written in the form
φ ( n ) + ψ ( n ) 2 n
is due to Ch. R. Wall [4]. □
We must mention that the relation (14) refines the first inequality of (6).
Theorem 6.
For n > 1 one has
ψ ( n ) n + 2 ω ( n ) + Ω ( n ) 2 n + 1 .
Proof. 
By using of (2) and the definitions of ω ( n ) and Ω ( n ) , one has that
ψ ( n ) n = i = 1 r p i α i 1 i = 1 r ( p i + 1 ) i = 1 r p i .
Since
i = 1 r p i α i 1 2 i = 1 r α i r = 2 Ω ( n ) ω ( n ) ,
then in order to prove (15), we have to prove that
i = 1 r ( p i + 1 ) i = 1 r p i 2 2 r 2 .
For r = 1 in (16) there is an equality, while for r = 2 one has
( p 1 + 1 ) ( p 2 + 1 ) p 1 p 2 = p 1 + p 2 + 1 2 + 3 + 1 = 6 > 4 = 2 2 .
We can prove that there is strict inequality in (16) for r 2 . Having in mind that p 1 2 , p 2 3 for each i ( 3 i r ) : p i 5 , we obtain
i = 1 r ( p i + 1 ) i = 1 r p i 2 2 r 2 12 i = 3 r ( p i + 1 ) 6 i = 3 r p i 2 2 r 2 > 6 i = 3 r ( p i + 1 ) 2 2 r 2 6 . 6 r 2 2 2 r 2 = 6 r 1 4 r 1 > 0 ,
which proves (16) and therefore (15). □
From the above proof it follows that there is equality in (15) only for n = 2 α for α = 1 , 2 , or for n being a prime number.
The second inequality of (15) follows from Ω ( n ) + ω ( n ) 2 .
Theorem 7.
Let
λ = ln 2 ln 3 .
Then
φ ( n ) n . 1 ( γ ( n ) ) 1 λ , i f   n > 1   i s   o d d 1 2 λ ( γ ( n ) ) 1 λ , i f   n > 1   i s   e v e n .
Proof. 
Let us consider the application
f 1 ( x ) = x 1 x λ
for x > 1 . Then an easy computation gives
f 1 ( x ) = 1 x λ + 1 ( ( 1 λ ) x + λ ) > 0
for 0 < λ < 1 . Thus, the function f 1 is strictly increasing. Particularly, for p 3
f ( p ) f ( 3 ) = 2 3 λ = 1
for 3 λ = 2 , i.e., it is valid only for
λ = ln 2 ln 3 = 0.63092 .
Thus, we have the inequality
p 1 p λ
for p 3 with equality only for p = 3 .
Now, it is well known that
φ ( n ) n = p | n 1 1 p ,
where p runs through the prime divisors of n. Now, for n > 1 odd, by (18) we obtain
p | n 1 1 p = p | n ( p 1 ) p | n p p | n p λ p | n p .
Thus, the first inequality of (17) follows. When n is even, then let p 1 = 2 be the least prime divisor of n. Then by (18)
φ ( n ) n = 1 2 p | n , p 3 1 1 p 1 p | n p p | n p 2 λ .
Thus, the second inequality of (17) follows as well. □
Remark 1.
As λ > 1 2 and γ ( n ) 2 , by (17) we get the weaker inequality
φ ( n ) n . 1 γ ( n ) , i f   n > 1   i s   o d d , 1 2 γ ( n ) , i f   n > 1   i s   e v e n ,
proved in [5].
Remark 2.
The number λ = ln 2 ln 3 is an irrational number. Indeed, as 0 < λ < 1 , it cannot be an integer. If it would be rational, i.e.,
ln 2 ln 3 = a b
for some integers a , b > 1 , then we would obtain 2 b = 3 a , that is impossible, as the left side is even and the right side is odd. But λ is even a transcendental number, according for the famous theorem of Gelfond–Schneider [6]. If a and b are algebraic numbers with a { 0 , 1 } and b not a rational, then a b is transcendental. In our case, 3 λ = 2 and since λ is irrational, by the above theorem, if λ would be algebraic, we would obtain a contradiction.
Theorem 8.
Let
μ = ln 4 ln 3 .
Then
ψ ( n ) n . ( γ ( n ) ) μ 1 , i f   n > 1   i s   o d d 3 2 μ ( γ ( n ) ) μ 1 ) , i f   n > 1   i s   e v e n .
Proof. 
Let us define
f 2 ( x ) = x μ x + 1
for x > 1 . For the derivative of this function, one has
f 2 ( x ) = x μ 1 ( x + 1 ) 2 ( x ( μ 1 ) + μ ) > 0 ,
as μ > 1 . This f 2 is strictly increasing and implying
f 2 ( p ) f 2 ( 3 ) = 3 μ 4 = 1
for p 3 . This implies the inequality
p + 1 p μ
for p 3 with μ satisfying 3 μ 4 , i.e., μ = ln 4 ln 3 = 1.26185 .
Now, the proof of (19) follows by applying (20) in the same manner, as in the proof of Theorem 7. □
Remark 3.
As μ 1 = 0.26185 < 1 2 , it is easy to see from (19) we get the weaker relation
ψ ( n ) n . γ ( n ) , i f   n > 1   i s   o d d , 2 γ ( n ) , i f   n > 1   i s   e v e n .
Remark 4.
From the proof of Theorem 7, it follows that there is an equality in the first part of (17) only for n = 3 k , where k 1 is integer, and for n = 2 k in the second part. Similarly, for the relation (19).
Remark 5.
As μ = 2 ln 2 ln 3 , from Remark 2 we get that μ is also a transcendental number.
Theorem 9.
For each n > 1
φ ( n ) ψ ( n ) n 2 ω ( n ) n γ ( n ) 2 .
Proof. 
When n is prime, (21) is obviously true. Let us assume that (21) is valid for some n > 1 with ω ( n ) 2 and let p 2 is not a divisor of n. Then
( n p ) 2 ω ( n p ) n p γ ( n p ) 2 φ ( n p ) ψ ( n p ) = n 2 p 2 ( ω ( n ) + 1 ) n γ ( n ) 2 φ ( n ) ψ ( n ) ( p 2 1 ) n 2 p 2 ( ω ( n ) + 1 ) n γ ( n ) 2 n 2 ω ( n ) n γ ( n ) 2 ( p 2 1 ) = n 2 p 2 ω ( n ) n γ ( n ) 2 n γ ( n ) 2 n 2 p 2 + ω ( n ) n γ ( n ) 2 p 2 + n 2 ω ( n ) n γ ( n ) 2 = ω ( n ) n γ ( n ) 2 ( p 2 2 ) + n 2 n γ ( n ) 2 > 0 .
Let p 2 be a divisor of n. Then
( n p ) 2 ω ( n p ) n p γ ( n p ) 2 φ ( n p ) ψ ( n p ) = n 2 p 2 ω ( n ) n γ ( n ) 2 p 2 φ ( n ) ψ ( n ) p 2 n 2 p 2 ω ( n ) n γ ( n ) 2 p 2 n 2 ω ( n ) n γ ( n ) 2 p 2 = 0 ,
which proves the theorem. □
Remark 6.
Let σ ( n ) denote the sum of the divisors of n. By the known inequality for n > 1
φ ( n ) σ ( n ) < n 2 ,
we get from (17) the following relation for σ ( n ) :
σ ( n ) < n . ( γ ( n ) ) 1 λ , i f   n > 1   i s   o d d , 1 2 λ ( γ ( n ) ) 1 λ ) , i f   n > 1   i s   e v e n .
As γ ( n ) n and 1 λ < 1 2 , this improves the inequaity
σ ( n ) < n n
by C. C. Lindner (see, e.g., [1]).
Remark 7.
By using the known inequality for n > 1 (see, e.g., [7])
σ ( n ) < π 2 6 ψ ( n )
and combining with relation (19), one can obtain another upper bound for σ ( n ) . For example, when n > 1 is odd, we get from (22)
σ ( n ) < π 2 6 n ( γ ( n ) ) μ 1 .
Since γ ( n ) n , a simple computation shows that the right side of (23) is less than 6 π 2 n n for n 77 , so we get an improvement of the inequality for n 9
σ ( n ) < 6 π 2 n n
due to U. Annapurna (see, [8]).

3. Conclusion

In the authors’ book [9], a lot of inequalities related to the arithmetic functions φ and ψ were given. For a brief survey of some inequalities for arithmetic functions, see paper [10].
In the present paper some new inequalities with these functions were formulated and their validity was proven.

Author Contributions

Conceptualization, J.S. and K.A.; methodology, J.S.; validation, J.S.; formal analysis, J.S. and K.A.; investigation, J.S. and K.A.; writing—original draft preparation, J.S. and K.A.; writing—review and editing, J.S. and K.A. Both authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

The authors would like to thank Dr Peter Vassilev and Dr Vassia Atanassova for technical help and proofreading.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Sándor, J.; Mitrinović, D. S.; Crstici, B. Handbook of Number Theory, Vol. 1. Springer Verlag: New York, 2006.
  2. Sándor, J. On Dedekind’s arithmetical function. Seminarul de Teoria Structurilor, Univ. of Timişoara, Romania 1988, 51, 1–15.
  3. Sándor, J. On an arithmetical inequality. Analele Stiintifice ale Universitatii Ovidius Constanta, Seria Matematica 2014, 22(1), 257–261.
  4. Wall, Ch. R. Problem B-510. Fibonacci Quarterly 1984, 22, 371.
  5. Atanassov, K. Inequalities for φ and σ functions. I. Bulletin of Number Theory and Related Topics 1991, XV, 12–14.
  6. Baker, A. Transcendental Number Theory. Cambridge University Press: Cambridge. 1975, pp. 10.
  7. Sándor, J. On the inequality σ(n) < π 2 6 ψ ( n ) . Octogon Mathematical Magazine 2008, 16(1), 295–296.
  8. Annapurna, U. Inequalities for σ(n) and φ(n). Mathematical Magazine 1972, 45(4), 187–190.
  9. Sándor, J., Atanassov, K. Arithmetic Functions. Nova Sciences: New York, 2021.
  10. Dimitrov, S. I. Inequalities involving arithmetic functions. Lithuanian Mathematical Journal 2024, 64, 421–452.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
Prerpints.org logo

Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.

Subscribe

Disclaimer

Terms of Use

Privacy Policy

Privacy Settings

© 2025 MDPI (Basel, Switzerland) unless otherwise stated