On polar relative normalizations of ruled surfaces

Ioanna-Iris Papadopoulou and Stylianos Stamatakis

Aristotle University of Thessaloniki
Department of Mathematics
arXiv:1711.11035v1 [math.DG] 29 Nov 2017

GR-54124 Thessaloniki, Greece

e-mail: [email protected]

Abstract

This paper deals with skew ruled surfaces in the Euclidean space E3 which are
equipped with polar normalizations, that is, relative normalizations such that
the relative normal at each point of the ruled surface lies on the corresponding
polar plane. We determine the invariants of a such normalized ruled surface
and we study some properties of the Tchebychev vector field and the support
vector field of a polar normalization. Furthermore, we study a special polar
normalization, the relative image of which degenerates into a curve.
Key Words: Ruled surfaces, Polar normalizations, Tchebychev vector field,
Pick invariant
MSC 2010 : 53A25, 53A05, 53A15, 53A40

1 Introduction
In 1989 F. Manhart introduced the one-parameter family of relative normalizations
(a)
y of a hypersurface with non-vanishing Gaussian curvature K e in the Euclidean
e a , a R and
space En+1 which are characterized by the support functions (a) q = |K|
called Manharts normalizations (see [2]).
G. Stamou and A. Magkos in [9] and G. Stamou, St. Stamatakis and I. Delivos in [10]
studied ruled surfaces in the Euclidean space E3 which are equipped with Manharts
normalizations. Later, S. Stamatakis and I. Kaffas studied in [5] the asymptotic
relative normalizations of a ruled surface , that is, relative normalizations such
that the relative normals at each point P of lie on the corresponding asymptotic
plane of .
Following this idea the authors introduced in [7] three special relative normalizations:

1. the central normalizations, i.e, relative normalizations such that the relative
normals at each point P of lie on the corresponding central plane,

1
 2. the polar normalizations, i.e, relative normalizations such that the relative
normals at each point P of lie on the corresponding polar plane and finally

3. the right normalizations, that is relative normalizations of whose relative im-

ages are also ruled surfaces with the additional property that their generators
are parallel to those of . Some of these relative normalizations degenerate
into a curve.

The central and the right normalizations were studied thoroughly in [7] and [8],
respectively. In this paper we will study the polar normalizations.

2 Preliminaries
A brief discussion of some definitions, results and formulae of relative Differential
Geometry of surfaces and Differential Geometry of ruled surfaces in the Euclidean
space E3 appears in this section. We refer the reader to [3] and [4].
In the three-dimensional Euclidean space E3 let = (U, x) be a ruled C r -surface
of nonvanishing Gaussian curvature, r 3, defined by an injective C r -immersion
x = x(u, v) on a region U := I R (I R open interval) of R2 . We introduce the
so-called standard parameters u I, v R of , such that

x(u, v) = s(u) + v e(u), (2.1)

and
|e| = |e | = 1, hs , e i = 0,
where the differentiation with respect to u is denoted by a prime and h , i denotes
the standard scalar product in E3 .. Here : s = s(u) is the striction curve of and
the parameter u is the arc length along the spherical curve e = e(u).
The distribution parameter (u) := (s , e, e ), the conical curvature (u) := (e, e , e )
and the function (u) := cot , where (u) := (e, s ) is the striction of ( 2 <
2 , sign = sign ), are the fundamental invariants of and determine uniquely
the ruled surface up to Euclidean rigid motions. We also consider the central
normal vector n(u) := e and the central tangent vector z(u) := e n. It is known
that the vectors of the moving frame D := {e, n, z} of fulfil the following equations
[3, p. 280]
e = n, n = e + z, z = n. (2.2)
Then we have
s = e + z. (2.3)
We denote partial derivatives of a function (or a vector-valued function) f in the
coordinates u1 := u, u2 := v by f/i , f/ij etc. Then from (2.1) and (2.3) we take

x/1 = e + v n + z, x/2 = e, (2.4)

2
and thus the unit normal vector (u, v) to is given by
nvz
= , where w := 2 + v2.
w
Let I = gij dui duj and II = hij dui duj , i, j = 1, 2 be the first and the second
fundamental form of , respectively, where
g11 = w 2 + 2 2 , g12 = , g22 = 1, (2.5)
w2 + v 2
h11 = , h12 = , h22 = 0. (2.6)
w w
The Gaussian curvature K(u,e e v) of are given by
v) and the mean curvature H(u,
(see [3])
2 2 2
e = w + v + .

Ke = , H (2.7)
w4 2w 3
A C s -relative normalization of is a C s -mapping y = y(u, v), 1 s < r, defined on
U, such that
rank({x/1 , x/2 , y}) = 3, rank({x/1 , x/2 , y /i }) = 2, i = 1, 2, (u, v) U. (2.8)
The pair (, y) is called a relatively normalized ruled surface in R3 and the straight
line issuing from a point P in the direction of y is called the relative normal of
(, y) at P . The pair = (U, y) is called the relative image of (, y).
The support function of the relative normalization y is defined by q(u, v) := h, yi
(see [1]). For q = 1, we have y = , that is, the normalization is the Euclidean one.
Due to (2.8), q never vanishes on U. Conversely, when a support function q is given,
the relative normalization y of the ruled surface is uniquely determined and can
be expressed in terms of the moving frame D as follows [5, p. 179]:
y = y1 e + y2 n + y3 z, (2.9)
where
q/1 + q/2 ( w 2 + v) 2 q w 2 v q/2 v q + w 2 q/2
y1 = w , y2 = , y3 = .
2 w w
(2.10)
For the coefficients Gij (u, v) of the relative metric G(u, v) of (, y), which is indefi-
nite, we have
Gij = q 1 hij . (2.11)
Then, because of (2.6), the coefficients of the inverse relative metric tensor are
wq w2 + v 2
G(11) = 0, G(12) = , G(22) = w q . (2.12)
2

For a function (or a vector-valued function) f we denote by G i f the covariant

derivatives in the direction of ui , both with respect to the relative metric. The
coefficients Aijk (u, v) of the Darboux tensor are given by
Aijk := q 1 h, G G
k j x/i i.

3
Then, by using the relative metric tensor Gij for raising and lowering the indices,
the Pick invariant J(u, v) of (, y) is defined by
1
J := Aijk Aijk .
2
As we proved in [7] (see equation (2.2)) the Pick invariant is calculated by


3 w 2 q/2 + v q n 2  2 2

 2
o
J= w qv + 2q/1 + q/2 w + v q (v ) .
2 2 w 3 q
(2.13)
The relative shape operator has the coefficients Bij (u, v) given by

y /i =: Bij x/j . (2.14)

Then, for the relative curvature K(u, v) and the relative mean curvature H(u, v) of
(, y) we have
j
B11 + B22
K := det Bi , H := . (2.15)
2
We conclude this section by mentioning that, among the surfaces of E3 with negative
Gaussian curvature the ruled surfaces are characterized by the relation

3H J 3S = 0 (2.16)

(see [6]), where S(u, v) is the scalar curvature of the relative metric G of such a
surface , which is defined formally as the curvature of the pseudo-Riemannian
manifold (, G).

3 Polar normalizations
We concentrate now on the main topic of this paper, namely the polar normalizations
of a skew ruled surface , i.e., relative normalizations such that the relative normal
at each point P of lies on the corresponding polar plane {P ; n, z}. In [7] it was
shown that the support function of y is of the form

q = f (V ), (3.1)

where f (V ) is an arbitrary C 2 -function of

Z
v
V = arctan du. (3.2)

By means of (2.9), (2.10), (3.1) and (3.2) we deduce that the arising relative nor-
malization, i.e., the polar normalization of the given ruled surface is
q qv qv + q
y= n z, (3.3)
w w

4
where the dot denotes the differentiation with respect to V . Then, from (2.2), (2.4),
(2.14) and (3.3), we take the coefficients Bij of the relative shape operator of a polar
normalization:
(w 2 + v)(q + q)
B11 = ,
w3
1 n o
2 3 2 3
B1 = 3 qv qv + [q ( + 1) + q] + v [q (v + v + ) + q (v + )] ,
w
(q + q)
B21 = ,
w3
2 (q + q)
B22 = .
w3
Hence, by using (2.15) and (2.7b), we obtain the relative curvature K and the
relative mean curvature H:
(q qv) (q + q) e (q + q) .
K = , H=H (3.4)
w4
From (3.4a) we deduce that the relative curvature K of a polar normalization van-
ishes identically iff
q qv = 0 or q + q = 0,
or, equivalently, iff
V
q = ce v , c R or q = c1 cos V + c2 sin V, c1 , c2 R, c21 + c22 6= 0.

We reject the first support function since it leads to a non polar normalization. Thus
we have the following

Proposition 3.1. Let E 3 be a polar normalized ruled surface. The relative

curvature K of (, y) vanishes identically iff the support function is of the form

q = c1 cos V + c2 sin V, c1 , c2 R, c21 + c22 6= 0.

By taking (2.7b) and (3.4b) into consideration we arrive at

Proposition 3.2. Let E 3 be a polar normalized ruled surface. (, y) is relatively

minimal (H = 0) iff one of the following holds true
(a) the support function is of the form

q = c1 cos V + c2 sin V, c1 , c2 R, c21 + c22 6= 0,

(b) (, y) is a polar normalized right helicoid ( = c R and = = 0).

We notice that both the relative curvature K and the relative mean curvature H
vanish identically iff the support function is of the form

q = c1 cos V + c2 sin V, c1 , c2 R, c21 + c22 6= 0. (3.5)

5
By using (2.7b) and (2.13) we find the Pick invariant
!
e q
JEU K 3H
J = (qv + q) + , (3.6)
v q

where
v 3 + 2 ( ) v + 2
JEU K = 3v (3.7)
2 2 w 3
is the Pick invariant of the Euclidean normalization. The Pick invariant vanishes
identically iff
JEU K 3H e q
qv + q = 0 or + = 0,
v q
or, equivalently, iff

the support function is of the form

V v
q = c1 e , c1 R , or

is not a right helicoid and the support function is of the form

[
V v 3 + 2 ()v+ 2 ]
[v 2 + v+ 2 (+)]
q = c2 e , c2 R , or

is a right helicoid.

We reject the two support functions since they are not polar. Hence, we deduce

Proposition 3.3. Let E 3 be a polar normalized ruled surface. The Pick in-
variant J of (, y) vanishes identically iff is a right helicoid.

From (2.16), (3.4b), (3.6) and (3.7) we evaluate the scalar curvature of the relative
metric
1 n  


S = 2 3 q 2 w 4 + 2 v 2 + 2 + 2 v + 2 w 2 + 2 + v q 2
2 w q


o
+ q 2 2 v + v 2 2 q w 2 + v + 2 q .

4 The Tchebychev vector field and the support

vector field of a polar normalization
In [5] it was shown that the coordinate functions of the Tchebychev vector T (u, v)
of (, y), which is defined by
1
T := T m x/m , where T m := Aim ,
2 i

6
are given by
w 2 q/2 + v q 2 w 2 q/1 + q ( 2 v 2 ) T 1 (w 2 + v 2 )
T1 = , T2 = + .
w 2 2 w
Hence, by using (3.1) and (3.2), we deduce that the coordinate functions of the
Tchebychev vector of a polar normalization are
qv + q q (2vw 2 2 2 v + w 2 ) 2 3 q
T1 = , T2 = . (4.1)
w 2 2 w
The divergence divI T of T with respect to the first fundamental form I of , which
initially reads (see [5])
(w T i )/i
divI T = , (4.2)
w
becomes, on account of (4.1),
 h i
I 1

div T = 2 3 2w 2q 3v 2 + 2 2
2 w
nh i 
2 2 2 2

o
+ v + (2v + ) q 2 w + + v q .

The rotation curlI T of T with respect to the first fundamental form I of , which
initially reads (see [5])

I
(g12 T 1 + g22 T 2 )/1 (g11 T 1 + g12 T 2 )/2
curl T = , (4.3)
w
becomes, by taking (2.5) and (4.1) into consideration,
 h i
I 1
curl T = 3 2 2 qv 2 (2v + ) + 2 q 4 ( + 1) v 2 + (2 + ) v + 2
2 w
n h i o
+ q 4v + ( + ) (2v + ) q (2 v + )
h i h i
2 2 2 4
+ v 2 q v + 3 q v + q qv (2 v + ) + 2 q ( + 1) + q .

Analogously we calculate the divergence and the rotation of T with respect to the
relative metric of :
G 1 n 2 4 


div T = 2 3 q w + 2 v 2 2 2 v + 2 q 2 w 2 + v + 2
w q
 

o
+ q q 2 2 v + v 2 2 q w 2 + v + 2 ,
curlG T = 0.
Last relation agrees with
 
G q e 1/4
T = ln ,x where qAF F = |K|
qAF F
(see [6]). So, we have

7
Proposition 4.1. Let E 3 be a polar normalized ruled surface. The rotation of
the Tchebychev vector field with respect to the relative metric of vanishes identi-
cally and its potential is given by
wq
(u, v) = ln 1/2 + c, c R.
||

Now let
1 G1 
Q := ,x (4.4)
4 q
be the support vector Q(u, v) of (, y), which is introduced in [5]. By taking (2.12),
(3.1) and (3.2) into consideration we find that the coordinate functions of the support
vector field of a polar normalization are
q q
Q1 = , Q2 = . (4.5)
4wq 4wq
By means of (2.7b), (4.2) and (4.5), we find the divergence divI Q of Q with respect
to the first fundamental form I of
2
e q q q .
divI Q = H
2q 2
Hence, we derive

Proposition 4.2. Let E 3 be a polar normalized ruled surface. The support vec-
tor field is incompressible with respect to the first fundamental form of (divI Q = 0)
iff
(a) the support function is of the form
q = c2 ec1 V , c1 R, c2 R , or
(b) is a right helicoid.

By taking (2.5), (4.3) and (4.5) into account we deduce that the rotation curlI Q of
Q with respect to the first fundamental form I of is
I q 2 + q (qv + q)
curl Q = .
4w 2q
By taking (2.11), (4.2) and (4.5) into consideration we find the divergence divG Q
of Q with respect to the relative metric of
1 n   2 

divG Q = q q v + 2
(2v +
) 2 q w 2
+
v + 2

4w 3 q 2

o
+ q q w 2 + 2 + v .

By using (2.6), (2.11), (3.1), (3.2), (4.3) and (4.5) we have the rotation curlG Q of
Q with respect to the relative metric of
curlG Q = 0,
which agrees with the relation (4.4). Thus, we have

8
Proposition 4.3. Let E 3 be a polar normalized ruled surface. The rotation of
the support vector field with respect to the relative metric of vanishes identically
and its potential is given by
1
(u, v) = + c, c R.
4q

5 A special polar normalization

In this section we will study the support function of the form (3.5), which arises
when the relative curvature K or the relative mean curvature H vanishes identically
(see Sec. 3). By using (3.3) the corresponding relative normalization takes the form
 Z  Z   Z  Z 
y = c1 cos du c2 sin du n c2 cos du + c1 sin du z,

i.e., the relative normalization degenerates into a curve with curvature

1
= R
R

|c1 cos du c2 sin du |

and torsion

= R
R
.
c1 cos du c2 sin du
Since
1
= ,

we deduce that y is a curve of constant slope iff is a ruled surface of constant
slope.
By means of (3.6) and (3.7) we find the Pick invariant of this normalization:
 R
R
 Z h
3 c2 cos du + c1 sin du n 2 2


J= cos du c 2 v + 2c 1 v c 2
2 2 w (c1 cos V + c2 sin V )
i Z h io



+ (c2 + c1 ) + sin du c1 v 2 2c2 v c1 2 (c1 + c2 ) .

Then by using (2.4) and (4.1) we deduce the Tchebychev vector

 Z  Z 
w v
T = 2 (c1 cos V + c2 sin V ) (2v + ) e + c2 cos du + c1 sin du n
2
 Z  Z 
+ c2 cos du + c1 sin du z.

Finally, by taking (2.4) and (4.5) into consideration we derive the support vector

c1 sin V c2 cos V
Q= (vn + z) .
4w (c1 cos V + c2 sin V )

9
References
[1] F. Manhart, Eigentliche Relativspharen, die Regelflachen oder Ruckungs-
flachen sind, Anz. Osterreich. Akad. Wiss. Math.-Natur. Kl., 125 (1988), 3740

[2] F. Manhart, Relativgeometrische Kennzeichnungen Euklidischer Hyper-

spharen, Geom. Dedicata, 29 (1989), 193207

[3] H. Pottmann and J. Wallner, Computational Line Geometry, Springer-

Verlag, New York, 2001

[4] P. A. Schirokow and A. P. Schirokow, Affine Differentialgeometrie, B. G.

Teubner Verlagsgesellschaft, Leipzig, 1962

[5] S. Stamatakis and I. Kaffas, Ruled surfaces asymptotically normalized, J.

Geom. Graph., 17 (2013), 177191

[6] S. Stamatakis, I. Kaffas and I.-I. Papadopoulou, Characterizations of

ruled surfaces in R3 and of hyperquadrics in Rn+1 via relative geometric invari-
ants, J. Geom. Graph., 18 (2014), 217223

[7] S. Stamatakis and I.-I. Papadopoulou, On ruled surfaces relatively nor-

malized, Beitr. Algebra Geom., 58 (2017), 591605

[8] S. Stamatakis and I.-I. Papadopoulou, Ruled surfaces right normalized,

arXiv:1706.07277 [math.DG]

[9] G. Stamou and A.Magkos, Regeleflachen relativgeometrisch behandelt,

Beitr. Algebra Geom., 45 (2004), 209215

[10] G. Stamou, S. Stamatakis and I.Delivos, A relative-geometric treatment

of ruled surfaces, Beitr. Algebra Geom., 53 (2012), 297309

10