an:06222488
Zbl 1370.11012
Knapp, Wolfgang
On Korselt's criterion for Carmichael numbers
EN
Elem. Math. 68, No. 3, 93-95 (2013).
00322508
2013
j
11A51 11A07
Carmichael numbers; Korselt's criterion
The Korselt indicator \(\kappa(n)\) for an integer \(n\geq 2\) is the product of all prime numbers \(p\) for which \(p-1\) divides \(n-1\). The Theorem in this paper claims that the proposition
``\(x^n\equiv x\pmod m\) for all integers \(x\)'' is valid for \(m=\kappa(n)\); and if it holds also for any other natural number \(m\), then it is necessary that \(m\) be a divisor of \(\kappa(n)\).
In particular, if \(n\) is a prime, then \(n\) divides \(\kappa(n)\). And if \(n\) is composite, then \(n\) divides \(\kappa(n)\) if and only if \(n\) is a Carmichael number, i.e., if and only if \(n\) is a product of distinct primes \(p\) for which \(p-1\) divides \(n-1\) -- a familiar statement of the Korselt's criterion for Carmichael numbers. The known fact that Carmichael numbers are all odd follows by the observation that \(\kappa(n)=2\) if \(n\) is even, and \(\kappa(n)\) is a multiple of 6 if \(n\) is odd. These corollaries, plus one more, are supposedly close consequences of the Theorem and are presented without details of proofs.
The readers should be informed that the hypothesis of the Theorem is missing the crucial assumption that \(x^n\equiv x\pmod {\kappa(n)}\) for all integers \(x\). And additionally, to be logically correct, the proof of the Theorem should replace the definition \(\kappa(n):=\prod_{p\in\Phi(n)}p\) by, say, \(K:=\prod_{p\in\Phi(n)}p\) and show that the assumed properties of \(\kappa(n)\) then force the identity \(K=\kappa(n)\).
Amin Witno (Amman)