Cahen's constant

In mathematics, Cahen's constant is defined as the value of the following infinite series of unit fractions with alternating signs:

C=∑i=0∞(−1)isi−1=11−12+16−142+11806−⋯≈0.643410546288...

(sequence A118227 in the OEIS). Here (si)i≥0 denotes Sylvester's sequence, which is defined recursively by s0=2 and

si+1=1+∏j=0isj

for i≥0. Combining these fractions in pairs leads to an alternative expansion of Cahen's constant as a series of positive unit fractions formed from the terms in even positions of Sylvester's sequence. This series for Cahen's constant forms its greedy Egyptian expansion:

C=∑i=0∞1s2i=12+17+11807+110650056950807+⋯

Cahen's constant is named after Eugène Cahen [fr], also known for the Cahen–Mellin integral, who was the first to introduce it and prove its irrationality.

Continued fraction expansion

The majority of naturally occurring mathematical constants have no known simple patterns in their continued fraction expansions. Nevertheless, the complete continued fraction expansion of Cahen's constant C is known: it is

C=[a02;a12,a22,a32,a42,…]=[0;1,1,1,4,9,196,16641,…]

where the sequence of coefficients

0,1,1,1,2,3,14,129,25298,420984147,...

(sequence A006279 in the OEIS) is defined for all n≥0 by the recurrence relation

a0=0,a1=1,an+2=an(1+anan+1).

All the partial quotients of this expansion are squares of integers. Davison and Shallit made use of the continued fraction expansion to prove that C is transcendental.

Alternatively, one may express the partial quotients in the continued fraction expansion of Cahen's constant through the terms of Sylvester's sequence. To see this, we prove by induction that 1+anan+1=sn−1. Indeed, we have 1+a1a2=2=s0, and if 1+anan+1=sn−1 holds for some n≥1, then

1+an+1an+2=1+an+1⋅an(1+anan+1)=1+anan+1+(anan+1)2=sn−1+(sn−1−1)2=sn−12−sn−1+1=sn.

As a consequence, an+2=an⋅sn−1 holds for every n≥1, from which it is easy to conclude that

C=[0;1,1,1,s02,s12,(s0s2)2,(s1s3)2,(s0s2s4)2,…].

Best approximation order

Cahen's constant C has best approximation order q−3; that is, there exist constants K1,K2>0 such that the inequality

0<|C−pq|<K1q3

has infinitely many solutions (p,q)∈ℤ×ℕ, while the inequality

0<|C−pq|<K2q3

has at most finitely many solutions (p,q)∈ℤ×ℕ. This implies that C has irrationality measure 3.

To give a proof, denote by (pn/qn)n≥0 the sequence of convergents to Cahen's constant (i.e., qn−1=an for every n≥1). It follows from an+2=an⋅sn−1and the recursion for (sn)n≥0 that

an+2an+12=an⋅sn−1an−12⋅sn−22=anan−12⋅sn−22−sn−2+1sn−12=anan−12⋅(1−1sn−1+1sn−12)

for every n≥1. As a consequence, the limits

α:=limn→∞q2n+1q2n2=∏n=0∞(1−1s2n+1s2n2)

and

β:=limn→∞q2n+2q2n+12=2⋅∏n=0∞(1−1s2n+1+1s2n+12)

both exist by basic properties of infinite products, which is due to the absolute convergence of

∑n=0∞|1sn−1sn2|.

Numerically, one can check that 0<α<1<β<2. Thus, the inequality

1qn(qn+qn+1)≤|C−pnqn|≤1qnqn+1

yields

|C−p2n+1q2n+1|≤1q2n+1q2n+2=1q2n+13⋅q2n+2q2n+12<1q2n+13

and

|C−pnqn|≥1qn(qn+qn+1)>1qn(qn+2qn2)≥13qn3

for all sufficiently large n. Therefore, C has best approximation order 3 (with K1=1 and K2=1/3), where we use that any solution (p,q)∈ℤ×ℕ to

0<|C−pq|<13q3

is necessarily a convergent to Cahen's constant.

Notes

  1. ↑ A number is said to be naturally occurring if it is not defined through its decimal or continued fraction expansion. In this sense, for example, Euler's number e=limn→∞(1+1n)n is naturally occurring.
  1. ↑ Cahen (1891).
  2. ↑ Borwein et al. (2014), p. 62.
  3. ↑ Davison & Shallit (1991).
  4. ↑ Duverney & Shiokawa (2020).
  5. ↑ Sloane, N. J. A. (ed.). "Sequence A006279". The On-Line Encyclopedia of Integer Sequences. OEIS Foundation.

References

  • Cahen, Eugène (1891). "Note sur un développement des quantités numériques, qui présente quelque analogie avec celui en fractions continues". Nouvelles Annales de Mathématiques. 10: 508–514.