There are two definitions of the Fermat number. The less common is a number of the form
obtained by setting
in a Fermat polynomial, the first few of which
are 3, 5, 9, 17, 33, ... (OEIS A000051).
The much more commonly encountered Fermat numbers are a special case, given by the binomial number of the
form .
The first few for
,
1, 2, ... are 3, 5, 17, 257, 65537, 4294967297, ... (OEIS A000215).
The number of digits for a Fermat number is
(1)
| |||
(2)
| |||
(3)
|
For ,
1, ..., the numbers of digits in
are therefore 1, 1, 2, 3, 5, 10, 20, 39, 78, 155, 309, 617,
1234, ... (OEIS A057755). The numbers of digits
in
for
, 1, ... are 1, 309, 381600854690147056244358827361, ...
(OEIS A114484).
Being a Fermat number is the necessary (but not sufficient) form a number
(4)
|
must have in order to be prime. This can be seen by noting that if
is to be prime, then
cannot have any odd factors
or else
would be a factorable number of
the form
(5)
|
Therefore, for a prime ,
must be a power of 2. No two Fermat
numbers have a common divisor greater than 1 (Hardy and Wright 1979, p. 14).
Fermat conjectured in 1650 that every Fermat number is prime and Eisenstein proposed as a problem in 1844 the proof that there are an infinite
number of Fermat primes (Ribenboim 1996, p. 88). At present, however, only composite Fermat numbers are known for
. An anonymous writer proposed that numbers of
the form
,
,
were prime. However,
this conjecture was refuted when Selfridge (1953) showed that
(6)
|
is composite (Ribenboim 1996, p. 88).
The only known Fermat primes are
(7)
| |||
(8)
| |||
(9)
| |||
(10)
| |||
(11)
|
(OEIS A019434), and it seems unlikely that any more will be found using current computational methods and hardware.
Factoring Fermat numbers is extremely difficult as a result of their large size. In fact, as of 2022, only to
have been completely factored. The number of factors
for Fermat numbers
for
,
1, 2, ... are 1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 4, 5, ... (OEIS A046052).
Written out explicitly, the complete factorizations are
(12)
| |||
(13)
| |||
(14)
| |||
(15)
| |||
(16)
| |||
(17)
| |||
(18)
|
(OEIS A050922). Here, the final large prime is not explicitly given since it can be computed by dividing by the other given factors.
The smallest factors of the Fermat numbers are 5, 17, 257, 65537, 641, 274177, 59649589127497217, 1238926361552897, 2424833, ... (OEIS A093179), while the largest are 5, 17, 257, 65537, 6700417, 67280421310721, 5704689200685129054721, (OEIS A070592).
The following table summarizes the properties of these completely factored Fermat numbers. Other tables of known factors of Fermat numbers are given by Keller (1983),
Brillhart et al. (1988), Young and Buell (1988), Riesel (1994), and Pomerance
(1996). A current list of the known factors of Fermat numbers is maintained by Keller.
In these tables, since all factors are of the form , the known factors are expressed
in the concise form
.
digits | factors | digits | reference | |
5 | 10 | 2 | 3, 7 | Euler 1732 |
6 | 20 | 2 | 6, 14 | Landry 1880 |
7 | 39 | 2 | 17, 22 | Morrison and Brillhart 1975 |
8 | 78 | 2 | 16, 62 | Brent and Pollard 1981 |
9 | 155 | 3 | 7, 49, 99 | Manasse and Lenstra (In Cipra 1993) |
10 | 309 | 4 | 8, 10, 40, 252 | Brent 1995 |
11 | 617 | 5 | 6, 6, 21, 22, 564 | Brent 1988 |
As of 2022,
has 6 known factors with C1133 remaining (where C
denotes a composite number with
digits),
has 4 known factors with C2391 remaining, and
has one known factor with C4880 remaining (Keller).
By the early 1980s,
was known to be composite for all
with the exceptions
, 22, 24, 28, and 31 (Riesel 1994, Crandall et al. 2003).
Young and Buell (1988) discovered that
is composite, Crandall
et al. (1995) that
is composite, and
Crandall et al. (2003) that
is composite (Crandall
1999; Borwein and Bailey 2003, pp. 7-8; Crandall et al. 2003). In 1997,
Taura found a small factor of
(Crandall et al. 2003, Keller), and a small factor
of
was also found. As of 2022, it
is known that
is composite for all
(cf. Crandall et al. 2003).
There are currently two Fermat numbers that are known to be composite, but for which no single factor is known: and
(Keller; cf. Crandall et al. 2003).
Ribenboim (1996, pp. 89 and 359-360) defines generalized Fermat numbers as numbers of the form with
even, while Riesel (1994, pp. 102 and 415) defines
them more generally as numbers of the form
.
Fermat numbers satisfy the recurrence relation
(19)
|
can be shown to be prime iff it satisfies Pépin's test.
Pépin's theorem
(20)
|
is also both necessary and sufficient.
In 1770, Euler showed that any factor of must have the form
(21)
|
where
is a positive integer. In 1878, Lucas increased
the exponent of 2 by one, showing that factors of Fermat
numbers must be of the form
(22)
|
for .
Factors of Fermat numbers are therefore Proth primes
since they are of the form
, as long as they also satisfy the additional condition
odd and
.
If
(23)
|
is the factored part of
(where
is the cofactor to be tested for primality), compute
(24)
| |||
(25)
| |||
(26)
|
Then if ,
the cofactor is a probable prime to the base
; otherwise
is composite.
In order for a polygon to be circumscribed about a circle (i.e., a constructible
polygon), it must have a number of sides given by
(27)
|
where
is nonnegative and the
are zero or more distinct Fermat primes (as stated by Gauss and first published
by Wantzel 1836). This is equivalent to the statement that the trigonometric functions
,
, etc., can be computed in terms of finite numbers
of additions, multiplications, and square root extractions iff
is of the above form.
The last
digits of
(where
is the smallest integer such that
has
digits) are eventually periodic for
, 2, ... with periods 1, 4, 20, 100, 500, 2500, ... (OEIS
A005054; Koshy 2002-2003).