Area of Torricelli's Trumpet or Gabriel's Horn,

Sum of the Reciprocals of the Primes, Factorials of

Negative Integers
Sinisa Bubonja
30.12.2016.
Abstract
In our previous work [1], we defined the method for computing general limits
of functions at their singular points and showed that it is useful for calculating
divergent integrals, the sum of divergent series and values of functions in their
singular points. In this paper, we have described that method and we will use
it to calculate the area of Torricelli's trumpet or Gabriel's horn, the sum of the
reciprocals of the primes and factorials of negative integers.

Contents
1 Introduction

2 Method for Computing General Limits of Functions at Their Singular Points

2
3 Area of Torricellis Trumpet or Gabriels Horn

4 Sum of the Reciprocals of the Primes

5 Factorials of Negative Integers

Introduction

Divergent series and divergent integrals have appeared in mathematics for a long
time. Mathematicians have devised various means of assigning finite values to such
1

series and integrals, although intuition suggests that the answer is infinity or it does
not exist. Method for computing general limits of functions at their singular points,
discovered in our previous work [1], will permit us to use the method of partial sums
for calculating sums of divergent series and Newton - Leibniz formula for calculating divergent integrals, which is the new and surprising result. We also showed that
our method is the strongest method around for summing divergent series and it is
superior to other known methods; for more details we refer the reader to [2]. As for
prerequisites, the reader is expected to be familiar with real and complex analysis in
one variable.
In Section 2 we describe the method for computing general limits of functions at their
singular points and show how that method may be used for assigning finite values to
divergent series and divergent integrals. In this section, we present definitions and
theorems with proofs because paper, [1] where the method is discovered, is not written in English.
In Sections 3, 4 and 5 we have compiled some of the standard facts on an area of
Torricelli's trumpet or Gabriel's horn, the sum of the reciprocals of the primes and
factorials of negative integers, respectively. In those sections, we assign finite values
to an infinite area of Torricelli's trumpet or Gabriel's horn, the sum of divergent
series of the reciprocals of the primes and Gamma function at their singular points,
respectively. Gamma function extends factorials to real and even complex numbers.
The gamma function is undefined for zero and negative integers, from which we can
conclude that factorials of negative integers do not exist.

Method for Computing General Limits of Functions at Their Singular Points

Definition
2.1 Let f be a function and has a series expansion about the point a
S
C . We will denote by limD
za f (z) the general limit of function f at point a and
define
limD
za f (z) = c,
where c is constant term of any series expansion of f about a.
Example 2.1 The series expansions of sin z, cos z and ez at infinity are same these
functions and we considered that constant terms of their series expansions are 0. By
previous definition,
D

lim sin z = lim cos z = lim ez = 0.

Example 2.2 Let us find the general limit of Riemann zeta function as z approaches
1
1. The Laurent series expansions of a function (z) about z = 1 is the series z1
+
1
1
1
2
3
4
5
1 (z 1) + 2 2 (z 1) 6 3 (z 1) + 24 4 (z 1) + O((z 1) ), where is EulerMascheroni constant and n is the nth Stieltjes constant. By previous definition,
D

lim (z) = .

z1

Definition 2.2 Let f be a function and has a series expansion about the point .
D
D
We will denote by limD
z() f (z), limz()+ f (z) and limz() f (z) the general
limit, upper general limit and lower general limit of a function f(z) as z approaches
to over radial line l, = {r ei |r R}, [0, 2), respectively, and define
D

lim

f (z) =
+

z()

z+e

lim

limi f (z) =

z()

limi f (z) =

ze

r+

r(0)

f (z) =

f (r ei ) = lim f (r ei ),
+

lim

f (r ei ) = lim f (r ei ),

lim

r(0)

D
D
D
D
1
lim f (z) = ( lim + f (z) + lim f (z)) = lim f (r ei ).
z()
r(0)
z()
2 z()

Definition 2.3 Let f be a function and has a series expansion about the point a
D
D
C. We will denote by limD
za() f (z), limza()+ f (z) and limza() f (z) the general
limit, upper general limit and lower general limit of a function f(z) as z approaches
to a over radial line l,a = {a + r ei |r R}, [0, 2), respectively, and define
D

lim + f (z) =

za()
D

lim f (z) =

za()

limi0 f (z) =

za+e
D

r0(0)

r0

limi0 f (z) =

zae

lim + f (a + r ei ) = lim+ f (a + r ei ),
D

lim f (a + r ei ) = lim f (a + r ei ),

r0(0)

r0

D
D
D
D
1
lim f (z) = ( lim + f (z) + lim f (z)) = lim f (a + r ei ).
r0(0)
za()
za()
2 za()

Definition 2.4 Let f be a function and has a pole of order m N at a C

Define
Z 0
D
limz+ Pn (z) =
Pn (z)dz,
1

limD
z Pn (z) =

Pn (z)dz,
0

limD
z(0) Pn (z)

1
=
2
3

Pn (z)dz,
1

limD
z0+ Pn
limD
z0 Pn
limD
z0(0) Pn

1

z
1

z
1
z

Z
=

Pn

1

Pn

=
1

1
2

z
1

1
dz,
z2

1
dz,
z2

z
1 1
2 dz,
Pn
z
z

lim f (z) = lim F1 (z) + c0 ,

za()

za()

where Pn (z) = an z n + an1 z n1 + + a1 z + a0 is polynomial of degree n 0, c0 is a

constant term and F1 (z) is the principal part of a Laurent series expansion of f at a.
Example 2.3 Let us find the general limit of a Riemann zeta function as z approaches 1 over radial line l0,1 , where l0,1 is real axis. The Laurent series expansions
1
+ 1 (z 1) + 21 2 (z 1)2 61 3 (z
of a function (z) about z = 1 is the series z1
1
1)3 + 24
4 (z 1)4 + O((z 1)5 ), where is Euler-Mascheroni constant and n is the
nth Stieltjes constant. By previous definitions,
Z
D
D
D
1 1 1 1
1
1
+ = lim
+ =
dr + = 0 + = .
lim (z) = lim
r0(0) 1 + r 1
z1(0)
z1(0) z 1
2 1 r r2
P

Example 2.4 Let us find the sum of divergent series

Thus, by previous definition,

X
n=1

1 = lim

m+

m
X

n=1

mdm =

1 = lim m =

n=1

m+

1 = 1+1+1+1++1+.

1
.
2

P
k
k
k
Example 2.5 Let us find the sum of divergent series
3 k + 4k +
n=1 n = 1 + 2 +
P
m
k
+ mk + , where k is positive integer. By Faulhabers formula,
n=1 n =


P
k
1
n k+1
Bn mk+1n since B1 = 12 , where Bn denotes the nth Bernoulli
n=0 (1)
k+1
n


P k
Pm
Pk
D
1
k
n k+1
number. Therefore n=1 n = limD
Bn mk+1n ) =
m+
n=1 n = limm+ ( k+1
n=0 (1)
n




R 0 1 Pk
R
Pk
0
1
n k+1
n k+1
(
Bn mk+1n )dm = k+1
Bn 1 mk+1n dm =
n=0 (1)
n=0 (1)
n
n
1 k+1

(1)k+2n



Bn
Pk
Pk
Pk
(1)k
(1)k
k+1
k+1
1
1
n k+1
k+1
B
=

B
=
(
)=
n
n
n=0 (1)
n=0
n=0
n
k+2n
k+1
n
k+2n
k+1
n k+2n
(1)k
k+1

have

Bk+1 by recurrence equation for Bernoulli numbers and previous definition. We

nk =

n=1

Bk+1
k+1

since k N, because the odd Bernoulli numbers are zero.

Theorem 2.1 If f is a function and has a pole of order 1 at a C

a constant term of a Laurent series expansion of f at a, then

and if c0 is

lim f (z) = c0 , [0, 2).

za()

c1
D
D
Proof: By previous definition, limD
za() f (z) = limza() F1 (z)+c0 = limza() za +
R
R 1 1
c1
c1
1 1 c1
c0 = limD
1 dr + c0 = 2e
dr + c0 = 0 + c0 = c0 ,
i
r0(0) a+rei a + c0 = 2 1
rei r2
r3
1
where F1 (z) is the principal part of a Laurent series expansion of f at a. Similarly
we can prove that the theorem holds for a = . 

Example 2.6 Let us find the general limit of a Gamma function, denoted by (z),
as z approaches 0 over radial line l0,0 , where l0,0 is real axis. The Laurent series
1
(6 2 + 2 )z+ 16 z 2 ( 3
expansions of a function (z) about z = 0 is the series z1 + 12

+ () ()) +
z ( + +
() ()) +
z ( +
2

() ()+ () ()+ () ())+O(z ), where is Euler-Mascheroni constant

and () (z) is the nth derivative of the digamma function. By previous theorem,
D

lim (z) = lim (z) = .

z0(0)

z0

Definition
2.5 Let f be a function and has a series expansion about the point a
S
C and does not have a pole at a.
D

lim + f (z) = c ( lim f (z) = c)

za()

za()

if limza()+ f (z) (limza() f (z)) is infinite or does not exist, where c is a constant
term of any series expansion of f about a; otherwise
D

lim + f (z) =

za()

lim + f (z) ( lim f (z) =

za()

za()

lim f (z)).

za()

Example
Let us find P
the sum of the harmonic series which are divergent. We
P 2.7
m
D
D
1
1
=
lim
have
m+
n=1 n = limm+ Hm , where Hm is harmonic number.
n=1 n
Therefore, by previous definition,

X
1
= ,
n
n=1

where is Euler-Mascheroni constant, because the series expansions of a function Hm

1
1
1
1
1 7
12m
about m = is the series ( ln( m1 )) + 2m
2 + 120m4 252m6 + O(( m ) ), where
ln(z) is natural logarithm.
5

P
Example 2.8 Let us find the sum of divergent
series
n=1 (n 1)!
Pm= 0! + 1! +
P
D
2! + 3! + + (n 1)! + . We have
(n

1)!
=
lim
m+
n=1 (n 1)! =
n=1
D
m
limm+ (1) m!!(m 1)+!(2) + 1, where n!! is the double factorial function and
!n is subfactorial function. Therefore, by previous definition,

X
(n 1)! 0.697175 + 1.15573 i,
n=1

because the constant term of a series expansion of function (1)m m!!(m1)+!(2)+

1 about are 1+ (1,1)
= 0.69717488323506606876547868191955159531717543095436951732...+
e
1.1557273497909217179100931833126962991208510231644158204... i, where (a, z)
is the incomplete gamma function.
R +
Example 2.9 Let us find the finite value of divergent integral 0 sin xdx. We have
R +
sin xdx = ( cos x)|+
= limD
0
x+ ( cos x)+cos 0. Thus, by previous definition,
0
Z +
sin xdx = 0 + 1 = 1,
0

because the series expansions of cos x at are cos x = cos x + 0.

R +
Example 2.10 Let us find the finite value of divergent integral 0 ln x sin xdx,
R +
where ln x is natural logarithm. We have 0 ln x sin xdx = (Ci(x)ln x cos x)|+
=
0
D
D
limx+ (Ci(x) ln x cos x) limx0 (Ci(x) ln x cos x) = 0 = , where Ci(x)
is cosine integral and is Euler-Mascheroni constant. Thus, by previous definition,
Z +
ln x sin xdx = 0 = ,
0

because the series expansions of Ci(x) ln x cos x at are cos x(ln x1 + O(( x1 )7 )) +
cos x(( x1 )2 + x64 120
+O(( x1 )7 ))+sin x( x1 x23 + x245 +O(( x1 )7 ))+O(( x1 )9 )ib 21 arg(x)
c+0
x6

1 2
1 4
and the series expansions of Ci(x) ln x cos x at 0 are + 4 x (2 ln x 1) + 96 x (1
1
4 ln x) + 4320
x6 (6 ln x 1) + O(x7 ).
S
Theorem 2.2 If f is a function and has a pole of order m N at a C and if
c0 is a constant term of a Laurent series expansion of f at a, then limD
za f (z) is a
D
mean value of general limits limza() f (z), [0, 2).
Proof: Let us first prove that the theorem holds for a C. By previous defi1
nitions and the first mean value theorem for definite integrals, limD
za f (z) = 2
R 2
R
R
2
2 1
D
D
1
1
limD
za() f (z)d = 2 0 limza() F1 (z)d + c0 = 2 0 2 (limza()+ F1 (z) +
0
R
2 1
D
D
1
i
i
limD
za() F1 (z))d+c0 = 2 0 2 (limr0+ F1 (a+re )+limza() F1 (a+re ))d+
6

R 2
P1
P1
D
1
i
k
i
k
c0 = 2
c
(a+re
a)
+lim
0 12 (limD
+
k
r0
r0+
k=m ck (a+re a) )d+
k=m
R
R
R
P
P
+
2
1
1
1
1
1
i k
c0 = 2
0 21 ( k=m ck (rei )k dr + 1
k=m ck (re ) dr)d + c0 = 2
R
R 2 1 P
P
k
k
2 ik
1
1+(1) ik
1+(1)
1 1
e d+c0 = 0+c0 = c0 ,
d+c0 = 2
2 1
k=n ck k+1 e
k=n ck k+1
0
0 2
where F1 (z) is the principal part of a Laurent series expansion of f at a. Similarly
we can prove that the theorem holds for a = . 

Area of Torricellis Trumpet or Gabriels Horn

Torricellis Trumpet, also called Gabriels Horn, a mathematical figure that stretched
to infinity but was not infinitely big is the surface of revolution obtained by rotating
the graph of the function f (x) = x1 on the interval [1, ) around the x-axis. Using
integration, it is possible to find the surface area A:
r
r
Z +
Z +
Z +
h 1 0 i2
1
1
1
1
A = 2
1+
dx = 2
1 + 4 dx 2
dx = +.
x
x
x
x
x
1
1
1
R +

1
x

Let us find the finite value of divergent integral A = 2 1

1 + x14 dx. We have
q


+
q
 q 1 +1x2 sinh1 (x2 )
1
R + 1 q
+1x2 sinh1 (x2 )

D
x4
x4
1
1
1

2 1 x 1 + x4 dx = 2
2 x4 + 1  . Thus, limx+ 2

2 x4 +1
2 x4 +1
1
q

 q 1 +1x2 sinh1 (x2 ) q

x4
1
1
1
1

+
1
lim
2

+
1
= 12 ln(4)2( 21 sinh1 (1)
x1
2
x4
2
x4
2 x4 +1
0.70996, where ln(z) is natural logarithm and sinh1 (z) is the inverse hyper q 1 +1x2 sinh1 (x2 ) q

x4
1
1

bolic sine function, because the series expansions of 2

+
1

2
x4
2 x4 +1
1 )
2

at are (2 ln( x1 ) + 12 ln(4))

4x4

32x8

+ O(( x1 )11 ). This gives

A = 0.70995...
.

Sum of the Reciprocals of the Primes

The sum of the reciprocals of all prime numbers diverges. This was proved by Leonhard Euler in 1737, and strengthens Euclids 3rd-century-BC result that there are
infinitely many prime
numbers. We will
by pn nth prime
Let us find
P+
Pdenote
Pm number.

D
1
1
1
the sum of series n=1 pn . We have n=1 pn = limm+ n=1 pn = limD
s1 P (s),
X 1
where P (s)
is the prime zeta function. For s close to 1, P (s) has the
ps
p is prime
7

P
(n)
expansion P (1 + ) = ln  + C + O(), where  0 and C = +
n=2 n ln (n) =
M = 0.261497212... 0.577215664... = 0.315718452..., where M is MeisselMertens constant, is Euler-Mascheroni constant, (n) is the Mobius function,
P+ (n)
is the Riemann zeta function and ln(z) is natural logarithm. Therefore, n=1 p1n =
D
limD
0 P (1 + ) = lim0 ( ln  + C + O()) = C because the series expansions of a
function ln  + C + O() about  = 0 is the series (C + ln()) + O (1 ). This gives
+
X
1
= M = 0.315718452...
p
n=1 n

Factorials of Negative Integers

The gamma function was first introduced by Leonhard Euler in his goal to generalize
factorial to non integer values. The (complete) gamma function (z) =
R + the
z1 x
x
e
dx is defined to be an extension of the factorial to complex and real
0
number arguments. It is analytic everywhere except at z = 0, 1, 2, ..., where it has
a poles of order 1. It is related to the factorial by (n + 1) = (n)! as special case of
functional equation (z + 1) = z(z). Gamma function is not the only solution of the
previous functional equation. Let us find the factorials of negative numbers as the
general limit of a gamma function as z approaches n, where n are positive integers.
We have (n)! = limD
zn (z) = c(n), where c(n) denote the constant term of the
Laurent series expansion of a function (z) about z = n. This gives
(1)! = 1 + ,
3
,
4 2
11
(3)! = + ,
36 6
25

(4)! =
,
288 24
..., where is Euler-Mascheroni constant.
(2)! =

References
[1] Sinisa Bubonja, General Method for Summing Divergent Series. Determination
of Limits of Divergent Sequences and Functions in Singular Points, Preprint,
viXra:1502.0074
8

[2] Sinisa Bubonja, General Method for Summing Divergent Series Using Mathematica and a Comparison to Other Summation Methods, Preprint, viXra:1511.0247
[3] G. H. Hardy, Divergent series, Oxford at the Clarendon Press (1949)
[4] Bruce C. Brendt, Ramanujans Notebooks, Springer-Verlag New York Inc. (1985)
[5] John Tucciarone, The Development of the Theory of Summable Divergent Series from 1880 to 1925, Archive for History of Exact Sciences, Vol. 10, No. 1/2,
(28.VI.1973), 1-40 1, 2
2