Superficie de Revolução

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

J. Geom.

104 (2013), 85–106


c 2013 Springer Basel
0047-2468/13/010085-22
published online April 9, 2013
DOI 10.1007/s00022-013-0149-3 Journal of Geometry

Some classification of surfaces of revolution


in Minkowski 3-space
Miekyung Choi, Young Ho Kim and Dae Won Yoon

Abstract. We study the surfaces of revolution with the non-degenerate


second fundamental form in Minkowski 3-space. In particular, we inves-
tigate the surfaces of revolution satisfying an equation in terms of the
position vector field and the 2nd-Laplacian in Minkowski 3-space. As a
result, we give some new examples of the surfaces of revolution with light-
like axis in Minkowski 3-space.

Mathematics Subject Classification (2010). 53A35, 53B30.


Keywords. Surfaces of revolution, Laplace operator, Minkowski space.

1. Introduction
Let x : M → Em be an isometric immersion of a connected n-dimensional
manifold in Euclidean m-space Em . Denote by H and Δ the mean curvature
and the Laplacian of M with respect to the Riemannian metric on M induced
from that of Em , respectively. Takahashi [7] proved that the submanifolds in
Em satisfying Δx = λx, that is, all coordinate functions are eigenfunctions of
the Laplacian with the same eigenvalue λ ∈ R are either the minimal subman-
ifolds of Em or the minimal submanifolds of hypersphere Sm−1 in Em .
As an extension of Takahashi theorem, Garay [4] studied hypersurfaces in Em
whose coordinate functions are eigenfunctions of the Laplacian, but not nec-
essarily associated to the same eigenvalue. Specifically, Garay [4] considered
hypersurfaces in Em satisfying the condition
Δx = Ax,

M. Choi was supported by the National Research Foundation of Korea Grant funded by the
Korean Government [NRF-2009-351-C00009]. Y. H. Kim was supported by Basic Science
Research Program through the National Research Foundation of Korea (NRF) funded by
the Ministry of Education, Science and Technology (2010-0007184).
86 M. Choi et al. J. Geom.

where A ∈ Mat(m, R) is an m × m-diagonal matrix, and he proved that such


hypersurfaces are minimal in Em and open pieces of either round hyperspheres
or generalized right spherical cylinders.
Related to this, Dillen et al. [3] investigated surfaces in E3 whose immersion
satisfies the condition
Δx = Ax + B, (1.1)
3
where A ∈ Mat(3, R) is a 3 × 3-matrix and B ∈ R . In other words, each coor-
dinate function is of usual 1-type in the sense of Chen [2]. For the Lorentzian
version of surfaces satisfying (1.1), Alı́as et al. [1] proved that the only such
surfaces are minimal surfaces (H = 0) and open pieces of Lorentz circular cyl-
inders, hyperbolic cylinders, Lorentz hyperbolic cylinders, hyperbolic spaces
or pseudo-spheres.
If a hypersurface M has the non-degenerate second fundamental form h, then
it is regarded as a new (pseudo-)Riemannian metric on M . So, considering
condition (1.1), we have the following question:
“Classify all hypersurfaces with the non-degenerate second fundamental form
h in Euclidean m-space or Minkowski m-space satisfying the condition
Δh x = Ax + B, A ∈ M at(m, R), B ∈ Rm , (1.2)
h
where Δ is the 2nd-Laplacian formed with h.”
For the above question, Kaimakamis and Papantoniou [5] proved that the sur-
faces of revolution with space-like axis or time-like axis in Minkowski 3-space
E31 satisfying condition (1.2) with B = 0 are minimal surface, pseudo-sphere
or hyperbolic space in E31 .
In this article, we completely classify surfaces of revolution in Minkowski 3-
space E31 satisfying condition (1.2). Also, it is to complete Kaimakamis and
Papantoniou’s classification of surfaces of revolution in E31 .

2. Preliminaries
Let E31 be Minkowski 3-space with the scalar product of index 1 given by
·, · = −dx20 + dx21 + dx22 , where (x0 , x1 , x2 ) is a standard rectangular coor-
dinate system of E31 . A vector x of E31 is said to be space-like if x, x > 0 or
x = 0, time-like if x, x < 0 and light-like or null if x, x = 0 and x = 0. A
time-like or light-like vector in E31 is said to be causal.
Now, we define a surface of revolution M in Minkowski 3-space E31 .
Let γ : I = (a, b) ⊂ R → Π be a curve in a plane Π of E31 and l a straight
line which does not intersect the curve γ. A surface of revolution M is a non-
degenerate surface revolving the curve γ around the axis l. Depending on the
axis being space-like, time-like or light-like, there are three types of motions.
If the axis l is space-like (resp. time-like), then l is transformed to the x1 -axis
or x2 -axis (resp. x0 -axis) by the Lorentz transformation. Therefore, we may
Vol. 104 (2013) Classification of surfaces of revolution 87

consider the x2 -axis (resp. x0 -axis) as the axis l if it is space-like (resp. time-
like). If the axis is light-like, we may assume that the axis is the line spanned
by the vector (1, 1, 0). Thus we consider the surfaces of revolution in E31 with
space-like, time-like or light-like axis, respectively (cf. [6]).
Case 1 The axis l is space-like.
Suppose that the profile curve γ lies in the x1 x2 -plane or x0 x2 -plane. Then the
curve γ can be represented by γ(u) = (0, f (u), g(u)) or γ(u) = (f (u), 0, g(u))
for some smooth functions f and g on an open interval I = (a, b).
On the other hand, a subgroup of the Lorentz group which fixes the vector
(0, 0, 1) is given by
⎛ ⎞
cosh v sinh v 0
⎝ sinh v cosh v 0 ⎠
0 0 1
for v ∈ R. Hence the surface M can be parameterized by
x(u, v) = (f (u) sinh v, f (u) cosh v, g(u)), f (u) > 0 (2.1)
or
x(u, v) = (f (u) cosh v, f (u) sinh v, g(u)), f (u) > 0 (2.2)
for v ∈ R.
Case 2 The axis l is time-like.
Without loss of generality, we may assume that the profile curve γ lies in the
x0 x1 -plane and so is given by γ(u) = (g(u), f (u), 0) for some positive function
f = f (u) on an open interval I = (a, b). Since a subgroup of the Lorentz group
which fixes the vector (1, 0, 0) is given by
⎛ ⎞
1 0 0
⎝ 0 cos v − sin v ⎠
0 sin v cos v
for v ∈ R, the surface of revolution M revolving γ around the axis Ox0 can be
written as
x(u, v) = (g(u), f (u) cos v, f (u) sin v)
for v ∈ R.
Case 3 The axis l is light-like.
In this case, we may suppose that the axis l is the line spanned by the vector
(1, 1, 0) and the curve γ lies in the x0 x1 -plane of the form γ(u) = (f (u), g(u), 0),
where f = f (u) is a positive function and g = g(u) is a function satisfying
p(u) = f (u) − g(u) = 0 for all u ∈ I. We notice here that a subgroup of the
Lorentz group which fixes the vector (1, 1, 0) is given by
⎛ 2 2 ⎞
1 + v2 − v2 v
⎝ v2 1 − v2
2
v⎠
2
v −v 1
88 M. Choi et al. J. Geom.

for v ∈ R. The surface of revolution M may be parameterized by


 
v2 v2
x(u, v) = f (u) + p(u), g(u) + p(u), p(u)v
2 2
for v ∈ R.
Let {x1 , x2 } be a local coordinate system of M . For the components hij (i, j =
1, 2) of the non-degenerate second fundamental form h on M , we denote by
(hij ) (resp. H) the inverse matrix (resp. the determinant) of the matrix (hij ).
By the 2nd-Laplacian Δh of the second fundamental form h on M , we mean
that it is formally defined by (cf. [5])

1 ∂
 ∂
2
Δh = −  |H| hij . (2.3)
|H| i,j=1 ∂xi ∂xj

3. Some examples
Before going into the study about surfaces of revolution in Minkowski 3-space
E31 satisfying condition (1.2), let us see some examples of surfaces in E31 .
Example 3.1. (Catenoid of the 1st kind). A catenoid of the 1st kind is param-
eterized by (Fig. 1)
x(u, v) = (u, sinh u cos v, sinh u sin v).
It is a space-like surface of revolution with time-like axis in E31 . We can check
that the 2nd-Laplacian Δh x of x satisfies
⎛ ⎞
0 0 0
Δh x = ⎝ 0 2 0 ⎠ x.
0 0 2
Example 3.2. (Catenoid of the 2nd kind). Let M be a space-like surface of
revolution with space-like axis in E31 with parametrization (Fig. 2)
x(u, v) = (cos u cosh v, cos u sinh v, u).

Figure 1 Catenoid of the 1st kind


Vol. 104 (2013) Classification of surfaces of revolution 89

Figure 2 Catenoid of the 2nd kind

Figure 3 Catenoid of the 3rd kind

It is called a catenoid of the 2nd kind in E31 . We can show that the 2nd-Lapla-
cian Δh x of x is given by
⎛ ⎞
2 0 0
Δh x = ⎝ 0 2 0 ⎠ x.
0 0 0

Example 3.3. (Catenoid of the 3rd kind). A catenoid of the 3rd kind is param-
eterized by (Fig. 3)
x(u, v) = (u, cos u cos v, cos u sin v).
It is a time-like surface of revolution with time-like axis in E31 . Furthermore,
the 2nd-Laplacian Δh x of x satisfies
⎛ ⎞
0 0 0
Δh x = ⎝ 0 2 0 ⎠ x.
0 0 2
Example 3.4. (Catenoid of the 4th kind). Consider a catenoid of the 4th kind
parameterized by (Fig. 4)
x(u, v) = (sinh u cosh v, sinh u sinh v, u).
It is a time-like surface of revolution with space-like axis in E31 . Also, it satisfies
⎛ ⎞
2 0 0
Δh x = ⎝ 0 2 0 ⎠ x.
0 0 0
90 M. Choi et al. J. Geom.

Figure 4 Catenoid of the 4th kind

Figure 5 Catenoid of the 5th kind

Example 3.5. (Catenoid of the 5th kind). Let M be a time-like surface of


revolution with space-like axis in E31 with parametrization (Fig. 5)
x(u, v) = (cosh u sinh v, cosh u cosh v, u).
It is said to be a catenoid of the 5th kind in E31 . We can easily see that the
2nd-Laplacian Δh x of x satisfies
⎛ ⎞
−2 0 0
Δh x = ⎝ 0 −2 0 ⎠ x.
0 0 0

Remark. The catenoids of the 1st, 2nd, 3rd, 4th and 5th kind are all minimal
[8].

Example 3.6. (Pseudo-sphere). A pseudo-sphere S21 (c, r) centered at c =


(0, 0, a) with the radius r given by
S21 (c, r) = {(x0 , x1 , x2 ) ∈ E31 | − x20 + x21 + (x2 − a)2 = r2 }
Vol. 104 (2013) Classification of surfaces of revolution 91

is a totally umbilical time-like surface in E31 . In this case, it is parameterized


by
x(u, v) = (r sin u sinh v, r sin u cosh v, r cos u + a)
and is a surface of revolution with space-like axis in E31 . On the other hand, it
satisfies
⎛ ⎞ ⎛ ⎞
2 0 0 0
1 1
Δh x = ⎝ 0 2 0 ⎠ x + ⎝ 0 ⎠ .
r r
0 0 2 −2a
Example 3.7. (Hyperbolic space). A hyperbolic space centered at c = (a, 0, 0)
with the radius r given by
H2 (c, r) = {(x0 , x1 , x2 ) ∈ E31 | − (x0 − a)2 + x21 + x22 = −r2 }
is a totally umbilical space-like surface. We can give a parametrization by
x(u, v) = (r cosh u + a, r sinh u cos v, r sinh u sin v).
It is a surface of revolution with time-like axis in E31 . Also, it satisfies
⎛ ⎞ ⎛ ⎞
−2 0 0 2a
1 1
Δh x = ⎝ 0 −2 0 ⎠x + ⎝ 0 ⎠.
r r
0 0 −2 0

4. Surfaces of revolution with space-like axis or time-like axis


In this section, we classify the surfaces of revolution with space-like axis or
time-like axis in E31 satisfying the condition
Δh x = Ax + B, (4.1)
3
where A = (aij ) ∈ Mat(3, R) (i, j = 1, 2, 3) and B = (b1 , b2 , b3 ) ∈ R .
Let M be a surface of revolution given by (2.1) with space-like axis, that is,
x(u, v) = (f (u) sinh v, f (u) cosh v, g(u)).
We may assume that γ is parameterized by the arc-length, i.e., it satisfies
(f  (u))2 + (g  (u))2 = 1.
Then, the components of the first fundamental form of the surface are given
by
g11 = 1, g12 = 0, g22 = −f 2 (u).
On the other hand, the unit normal vector field U of M is obtained by
U = (g  sinh v, g  cosh v, −f  ),
from which, the components of the second fundamental form of M are obtained
by
h11 = f  (u)g  (u) − f  (u)g  (u), h12 = 0, h22 = f (u)g  (u). (4.2)
92 M. Choi et al. J. Geom.

Therefore, using the data described above, the mean curvature H is given by
 
1 g
H= f  g  − f  g  − . (4.3)
2 f
We put
f  (u) = cos t, g  (u) = sin t
for some smooth function t = t(u). Since the second fundamental form h of M
is non-degenerate, the function t is non-vanishing everywhere. Furthermore,
the mean curvature H given by (4.3) becomes
 
1 sin t
H= −t − . (4.4)
2 f
By a straightforward computation, the 2nd-Laplacain Δh x of x with the help
of (2.3) and (4.2) turns out to be [5]
Δh x = (Φ1 (t(u)) sinh v, Φ1 (t(u)) cosh v, Ψ1 (t(u))), (4.5)
where
 
1 cos t t 2H cos t
Φ1 (t(u)) = − sin t − − + ,
sin t 2t t sin t
  
sin t t 2H cos t
Ψ1 (t(u)) = cos t − + .
2t t sin t
By the assumption (4.1) and the above Eq. (4.5), we get the following system
of differential equations:
Φ1 (t(u)) sinh v = a11 f sinh v + a12 f cosh v + a13 g + b1 , (4.6)
Φ1 (t(u)) cosh v = a21 f sinh v + a22 f cosh v + a23 g + b2 , (4.7)
Ψ1 (t(u)) = a31 f sinh v + a32 f cosh v + a33 g + b3 . (4.8)
From (4.6), (4.7) and (4.8), we easily deduce that a12 = a21 = a13 = a23 =
a31 = a32 = 0, b1 = b2 = 0 and a11 = a22 , i.e., the matrix A is diagonal. We
put a11 = a22 = λ and a33 = μ, λ, μ ∈ R. Then, the system reduces now to the
following equations
Φ1 (t(u)) = λf, (4.9)
Ψ1 (t(u)) = μg + b3 . (4.10)
We discuss five cases according to the constants λ and μ.
Case 1 λ = μ = 0.
In this case, combining (4.9) and (4.10) we can easily get b3 cos t = 2. So, the
function f  (u) = cos t is a constant. It is a contradiction to the second funda-
mental form to be non-degenerate. Thus, there are no surfaces of revolution
with space-like axis satisfying (4.1).
Case 2 λ = μ = 0.
From (4.9) and (4.10), we obtain
λf sin t − λg cos t = −2 + b3 cos t.
Vol. 104 (2013) Classification of surfaces of revolution 93

Differentiating the above equation with respect to u, we have


λf f  + λgg  + b3 g  = 0,
which implies
 2
2 b3
f + g+ = r2
λ
for some constant r. Thus, the surface of revolution satisfies the equation
−x20 + x21 + (x2 + bλ3 )2 = r2 , that is, M is an open piece of the pseudo-sphere
S21 (c, r) centered at c = (0, 0, − br3 ) with radius r on E31 .
Case 3 λ = 0, μ = 0.
In this case, (4.9) and (4.10) imply
−2 − λf sin t + b3 cos t = 0,
from which, by differentiating and using (4.4) we get
b3 t
2H cos t − sin t = 0, (4.11)
λf
or, equivalently
−λ sin t cos t
t = , (4.12)
P
where P = λf cos t + b3 sin t.
Differentiating Eq. (4.12), we have
λ2 sin t cos t
t = (2λf cos3 t + b3 sin t − 2b3 sin3 t). (4.13)
P3
Substituting (4.12) and (4.13) into (4.10), we get
α2 f 2 + α1 f + α0 = 0, (4.14)
where
α2 = −2λ2 b3 cos2 t,
α1 = λb3 sin t(sin2 t − 2b3 cos t),
α0 = −b23 sin2 t cos t.
Differentiating Eq. (4.14), we get
β2 f 2 + β1 f + β0 = 0, (4.15)
where
β2 = −4λ3 b3 cos2 t,
β1 = −λ2 b3 sin t cos t(− cos3 t + sin2 t cos t + 2b3 cos2 t + 2b3 ),
β0 = −2λb23 sin2 t cos2 t(− cos t + b3 ).

From (4.14) and (4.15), we obtain


(α1 β2 − α2 β1 )f + α0 β2 − α2 β0 = 0. (4.16)
94 M. Choi et al. J. Geom.

If we take the differentiation of the above equation once again, we get


γ1 f + γ0 = 0, (4.17)
where
γ0 = −2λ5 b23 sin t cos2 t(2 cos6 t − cos4 t − 8 sin2 t cos4 t + 4 sin2 t cos2 t
+ 2 sin4 t cos2 t − 4 sin2 t + 8 sin4 t + 2b3 cos5 t + 12b3 sin2 t cos3 t
−2b3 cos3 t − 6b3 sin2 t cos t),
γ1 = 2λ4 b33 sin2 cos3 t(−9 cos4 t + 3 sin2 t cos2 t + cos t + 4 sin2 t
+ 4b3 cos3 t − 6b3 sin2 t cos t).
Thus, (4.16) and (4.17) yield
(α0 β2 − α2 β0 )γ1 − (α1 β2 − α2 β1 )γ0 = 0.
By a long computation, the above equation becomes
4λ4 b53 sin3 t cos5 t(−24b3 sin2 t cos5 t − 4b3 sin2 t cos t + 4 sin2 t cos6 t
− 2b3 sin4 t cos3 t + 12b23 sin2 t cos4 t + 6b3 sin2 cos3 t − sin4 t cos4 t
+ sin2 t cos3 t + 2 sin2 t cos t + 5 cos8 t − 12b23 cos4 t + 12b23 cos6 t
+ 4b3 sin4 t cos t − 2b3 cos2 t − 14 sin2 t cos4 t + 4b3 cos3 t + 16b3 cos5 t
− 10 sin4 t cos2 t − 22b3 cos7 t + 2b3 cos4 t + 8 sin2 t − 2 cos6 t + 2 cos4 t
− 8 sin4 t − cos5 t) = 0.
Therefore, we have b3 = 0 because λ = 0. In this case, from (4.11) the mean
curvature H vanishes identically, that is, M is minimal. Thus, M is an open
piece of the catenoid of the 5th kind as time-like surface according to [8].
Case 4 λ = 0, μ = 0.
The system of Eqs. (4.9) and (4.10), in this case, takes the form
Φ1 (t(u)) = 0,
Ψ1 (t(u)) = μg + b3 . (4.18)
It follows −2 + μg cos t + b3 cos t = 0 and by differentiating we have
μ cos t −2μ2 sin t cos t
t = , t = .
μg + b3 (μg + b3 )2
From which, the first equation of (4.18) becomes
sin t
μf = .
1 − 12 cos2 t
If we differentiate once again, we have cos2 t = 1, a contradiction. Thus, in
this case, there are no surfaces of revolution satisfying (4.1).
Case 5 λ = 0, μ = 0 and λ = μ.
From (4.9) and (4.10), we have
λf sin t − μg cos t − b3 cos t = −2.
Vol. 104 (2013) Classification of surfaces of revolution 95

Differentiating the above equation, we obtain


(μ − λ) sin t cos t
t = ,
Q
2 3
((2μ − λ) sin2 t + (2λ − μ) cos2 t)t + 2t
t = , (4.19)
(λ − μ) sin t cos t
where Q = λf cos t + μg sin t + b3 sin t.
Substituting (4.19) into (4.9) and (4.10), we can easily find
Q cos t 2ν sin t cos t
+ − 2λνf sin t + 3λ − 2μ = 0, (4.20)
f Q
Q cos t 2ν sin t cos t
+ − 2ν(μg + b3 ) cos t − λ + 2μ = 0, (4.21)
f Q
where ν = μ − λ. By using μg cos t = λf sin t − b3 cos t + 2, we show that
Eq. (4.20) is equivalent to Eq. (4.21). Multiplying Qf cos t to (4.20), we have
the following algebraic equation
δ3 f 3 + δ2 f 2 + δ1 f + δ0 = 0,
where
δ3 = λ2 ν sin t, δ2 = λ(ν − λ + 2ν sin2 t),
δ1 = (ν − 3λ + ν sin2 t) sin t, δ0 = −2 sin2 t.
By [5, p. 87], we can show that either λ = μ or λ = 0 or μ = 0, a con-
tradiction. Therefore, there are no surfaces of revolution with space-like axis
satisfying (4.1).
Now, let us consider a surface of revolution M given by (2.2) with space-like
axis, that is,
x(u, v) = (f (u) cosh v, f (u) sinh v, g(u)).
We may assume that the parameter u is the arc-length of the profile curve
2 2
γ(u) = (f (u), 0, g(u)), that is, it satisfies −f  + g  = −ε(= ±1).
In this case, the components of the first and the second fundamental form of
M are given by, respectively
g11 = −ε, g12 = 0, g22 = f 2 (u),
h11 = −f (u)g (u) + f (u)g  (u),
  
h12 = 0, h22 = −f (u)g  (u).
Using the same method as above, we have the following:
If ε = −1, M is an open part of the hyperbolic space H2 (c, r) centered at
c = (0, 0, − br3 ) with radius r, or minimal surface, i.e., it is an open part of the
catenoid of the 2nd kind as space-like surface according to [8].
If ε = 1, M is an open part of the pseudo-sphere S21 (c, r) centered at c =
(0, 0, − br3 ) with radius r, or minimal surface, i.e., it is an open part of the
catenoid of the 4th kind as time-like surface.
Consequently, we have
96 M. Choi et al. J. Geom.

Theorem 4.1. Let M be a surface of revolution with space-like axis in Minkow-


ski 3-space. Then, it satisfies the condition Δh x = Ax + B for some constant
matrix A ∈ M at(3, R) and some constant vector B ∈ R if and only if one of
the the following statements holds:
1. M is an open piece of the catenoid of the 2nd kind, the 4th kind or the
5th kind.
2. M is an open piece of the pseudo-sphere S21 (c, r) for a constant vector c
and r > 0.
3. M is an open piece of the hyperbolic space H2 (c, r) for a constant vector
c and r > 0.
Let M be a surface of revolution with time-like axis given by
x(u, v) = (g(u), f (u) cos v, f (u) sin v)
for v ∈ R. We assume that the profile curve γ(u) = (g(u), f (u), 0) in the x0 x1 -
2 2
plane is of unit speed, that is, it satisfies −g  + f  = −ε(= ±1). Then the
natural frame {xu , xv } is given by
xu = (g  (u), f  (u) cos v, f  (u) sin v),
xv = (0, −f (u) sin v, f (u) cos v),
and the unit normal vector field U is obtained by
U = −(f  (u), g  (u) cos v, g  (u) sin v).
Thus, the components of the first and the second fundamental form of M are
given by, respectively
g11 = −ε, g12 = 0, g22 = f 2 (u),
h11 = f  (u)g  (u) − f  (u)g  (u), h12 = 0, h22 = f (u)g  (u).
First of all, we consider the case:
2 2
f  − g  = −1. (4.22)
In this case, the mean curvature H is given by
 
1     g
H= f g −f g + . (4.23)
2 f
From (4.22), we may put
f  (u) = sinh t, g  (u) = cosh t
for some smooth function t = t(u). Since the second fundamental form h of
M is non-degenerate, the function t is non-vanishing everywhere. Then, the
mean curvature H given by (4.23) becomes
 
1  cosh t
H= t + .
2 f
On the other hand, the 2nd-Laplacain Δh x is given by [5]
Δh x = (Ψ2 (t(u)), Φ2 (t(u)) cos v, Φ2 (t(u)) sin v), (4.24)
Vol. 104 (2013) Classification of surfaces of revolution 97

where
  
1 sinh t t 2H sinh t
Φ2 (t(u)) = cosh t + + − + ,
cosh t 2t t cosh t
  
cosh t t 2H sinh t
Ψ2 (t(u)) = sinh t + − + .
2t t cosh t
Suppose that the surface M satisfies Eq. (4.1). Then from (4.24), we obtain
the following system of ordinary differential equations:
Ψ2 (t(u)) = a11 g + a12 f cos v + a13 f sin v + b1 ,
Φ2 (t(u)) cos v = a21 g + a22 f cos v + a23 f sin v + b2 ,
Φ2 (t(u)) sin v = a31 g + a32 f cos v + a33 f sin v + b3 ,
which imply a12 = a21 = a13 = a23 = a31 = a32 = 0, b2 = b3 = 0 and
a22 = a33 . We put a22 = a33 = λ and a11 = μ. Then, the above system
reduces
Ψ2 (t(u)) = μg + b1 , (4.25)
Φ2 (t(u)) = λf. (4.26)
In order to solve the system, we split it into five cases according to the con-
stants λ and μ.
Case 1 λ = μ = 0.
From the above system of ODEs, we obtain b1 sinh t = 2. From this, the func-
tion f  (u) = sinh t is a constant. It is a contradiction to the fact that M has
the non-degenerate second fundamental form. Thus, there are no surfaces of
revolution with time-like axis satisfying (4.1).
Case 2 λ = μ = 0.
From (4.25) and (4.26), we obtain
λf cosh t − λg sinh t = 2 + b1 sinh t.
Differentiating the above equation with respect to u, we have
(λf sinh t − λg cosh t − b1 cosh t)t = 0
or, equivalently
b1 
f f  − gg  − g = 0.
λ
It follows
 2
2 b1
f − g− = r2
λ
for some constant r. Thus, M is an open part of a pseudo-sphere S21 (c, r)
centered at c = ( br1 , 0, 0) with radius r on E31 .
Case 3 λ = 0, μ = 0.
By combining (4.25) and (4.26), we get
2 − λf cosh t + b1 sinh t = 0.
98 M. Choi et al. J. Geom.

If we take the differentiation of the above equation, then we obtain


b1 cosh t 
2H sinh t − t = 0, (4.27)
λf
that is,
−λ sinh t cosh t
t = , (4.28)
R
where R = λf sinh t − b1 cosh t.
By differentiating Eq. (4.28), we get
λ2 sinh t cosh t
t = (2λf sinh3 t − b1 cosh t − 2b1 cosh3 t). (4.29)
R3
Substituting (4.28) and (4.29) into (4.25), we have
η2 f 2 + η1 f + η0 = 0,
where
η2 = −2λ2 b1 sinh2 t,
η1 = λb1 cosh t(− cosh2 t + 2b1 sinh t),
η0 = −b21 sinh t cosh2 t.
By using the same as that of Case 3 in Theorem 4.1, we can also obtain
b1 = 0,
which implies from (4.27) that the mean curvature H vanishes identically, i.e.,
the surface is minimal. Thus, M is an open part of the catenoid of the 3rd
kind as time-like surface.
Case 4 λ = 0, μ = 0.
From the system of Eqs. (4.25) and (4.26), we have
Ψ2 (t(u)) = μg + b1 ,
Φ2 (t(u)) = 0. (4.30)
From (4.30), we have 2 + μg sinh t + b1 sinh t = 0, it follows
μ sinh t 2μ2 sinh t cosh t
t = − , t = .
μg + b1 (μg + b1 )2
Therefore, the second equation of (4.30) becomes
cosh t
μf = − .
1 + 12 sinh2 t
By differentiating the above equation, cosh2 t = 0, i.e., g  = 0. It is a con-
tradiction. Thus, in this case, there are no surfaces of revolution satisfying
(4.1).
Case 5 λ = 0, μ = 0 and λ = μ.
From (4.25) and (4.26), we have
λf cosh t − μg sinh t − b1 sinh t = 2. (4.31)
Vol. 104 (2013) Classification of surfaces of revolution 99

Differentiating the above equation, we obtain

(μ − λ) sinh t cosh t
t = ,
S
2 3
((2λ − μ) sinh2 t + (λ − 2μ) cosh2 t)t + 2t
t = , (4.32)
(λ − μ) sinh t cosh t

where S = λf sinh t − μg cosh t − b1 cosh t.


Substituting (4.32) into (4.25) and (4.26), we can obtain the following equa-
tions
S sinh t 2ν sinh t cosh t
+ − 2ν(μg + b1 ) sinh t + λ − 2μ = 0,
f S
S sinh t 2ν sinh t cosh t
+ − 2λνf cosh t − 3λ + 2μ = 0, (4.33)
f S

where ν = μ − λ.
By multiplying Sf sinh t to (4.33) and using (4.31), we get

ξ3 f 3 + ξ2 f 2 + ξ1 f + ξ0 = 0,

where

ξ3 = λ2 ν cosh t, ξ2 = λ(λ − ν − 2ν cosh2 t),


ξ1 = (ν − 3λ + ν cosh2 t) cosh t, ξ0 = −2 cosh2 t.

By [5, p.91], we conclude that either λ = μ or λ = 0 or μ = 0, a contradiction.


Therefore, there are no surfaces of revolution satisfying (4.1) in this case.
Lastly, we consider ε = −1, that is, M is a space-like surface. Quite similarly
as before, we can show that M is an open part of the hyperbolic space H2 (c, r)
centered at c = ( br1 , 0, 0) with radius r, or the catenoid of the 1st kind as
space-like surface.
Consequently, we have
Theorem 4.2. Let M be a surface of revolution with time-like axis in Minkow-
ski 3-space. Then, M satisfies the condition Δh x = Ax + B for some constant
matrix A ∈ M at(3, R) and some constant vector B ∈ R if and only if one of
the following statements holds:

1. M is an open piece of the catenoid of the 1st kind or the 3rd kind.

2. M is an open piece of a pseudo-sphere S21 (c, r) for a constant vector c


and r > 0.

3. M is an open piece of a hyperbolic space H2 (c, r) for a constant vector c


and r > 0.
100 M. Choi et al. J. Geom.

5. Surfaces of revolution with light-like axis


Let M be a non-degenerate surface of revolution with light-like axis in E31
parameterized by
v2 v2
x(u, v) = (f (u) + p(u), g(u) + p(u), p(u)v), (5.1)
2 2
where p(u) = f (u) − g(u) = 0. Since M is non-degenerate, −f  (u)2 + g  (u)2
never vanish and thus p (u) = f  (u) − g  (u) = 0 everywhere. We may take the
parameter in such a way that
p(u) = −2u.
Let k(u) = f (u)+u. Then, the functions f and g in the definition of the profile
curve γ look like
f (u) = k(u) − u, g(u) = k(u) + u.
Thus, (5.1) can be rewritten as the form [6]:
x(u, v) = (k(u) − u − uv 2 , k(u) + u − uv 2 , −2uv).
Using the natural frame {xu , xv } of M , the components of the first fundamen-
tal form of the surface are
g11 = 4k  (u), g12 = 0, g22 = 4u2 .
It follows that the unit normal vector field U of M is given by
1
U=  (uk  (u) + u + uv 2 , uk  (u) − u + uv 2 , 2uv).
2 u |k  (u)|
2

From this, the components of the second fundamental form h are


−uk  (u) 2u2
h11 =  , h12 = 0, h22 =  .
u2 |k  (u)| u2 |k  (u)|
Since the second fundamental form h of M is non-degenerate, we get
uk  (u)
= 0.
k  (u)
If M is a space-like surface, that is, k  (u) > 0. By a direct computation, the
2nd-Laplacian Δh x of an isometric immersion x is obtained by

h k  (u)(k  (u) − uk  (u)) 
Δ x= k (u) − 1 − v 2 , k  (u) + 1 − v 2 , −2v
2uk  (u)2

+2 k  (u) (1, 1, 0). (5.2)
Suppose M satisfies the Eq. (4.1), that is, Δh x = Ax + B for some matrix
A = (aij ) and constant vector B. Then, (4.1) and (5.2) give the following
equations:
Vol. 104 (2013) Classification of surfaces of revolution 101


k  (u)(k  (u) − uk  (u))  
 2
k (u) − 1 − v 2 + 2 k  (u)
2uk (u)
= a11 (k(u) − u − uv 2 ) + a12 (k(u) + u − uv 2 ) − 2a13 uv + b1 , (5.3)

k  (u)(k  (u) − uk  (u))  
 2
k (u) + 1 − v 2 + 2 k  (u)
2uk (u)
= a21 (k(u) − u − uv 2 ) + a22 (k(u) + u − uv 2 ) − 2a23 uv + b2 , (5.4)

k  (u)(k  (u) − uk  (u))v

uk  (u)2
= a31 (k(u) − u − uv 2 ) + a32 (k(u) + u − uv 2 ) − 2a33 uv + b3 . (5.5)
Combining (5.3) and (5.4), we have

k  (u)(k  (u) − uk  (u))
− = (a11 + a12 − a21 − a22 )k(u)
uk  (u)2
− (a11 − a12 − a21 + a22 )u − (a11 + a12 − a21 − a22 )uv 2
− 2(a13 − a23 )uv + b1 − b2 . (5.6)
Since Eq. (5.6) is a polynomial in v with functions of u as the coefficients, we
obtain
a11 + a12 − a21 − a22 = 0, a13 − a23 = 0. (5.7)
From (5.7), Eq. (5.6) reduces now to the following equation

k  (u)(k  (u) − uk  (u))
= (a11 − a12 − a21 + a22 )u + b1 − b2 . (5.8)
uk  (u)2
Combining (5.5) and (5.8), we have
(−a11 + a12 + a21 − a22 )uv − (b1 − b2 )v
= (a31 + a32 )(k(u) − uv 2 ) + (−a31 + a32 )u − 2a33 uv − b3 . (5.9)
Since Eq. (5.9) is a polynomial with respect to v, we find
a31 + a32 = 0, −a31 + a32 = 0, a11 − a12 − a21 + a22 = 2a33 , b1 = b2 , b3 = 0,
which imply a31 = a32 = 0. Thus, Eq. (5.8) can be reduced as

k  (u)(k  (u) − uk  (u))
= 2a33 u. (5.10)
uk  (u)2
By adding (5.3) and (5.4), we obtain

k  (u)(k  (u) − uk  (u))  
 2
(k (u) − v 2 ) + 4 k  (u)
uk (u)
= 2(a11 + a12 )(k(u) − uv 2 ) − (a11 − a12 + a21 − a22 )u
− 4a13 uv + 2b1 . (5.11)
From this, we get

k  (u)(k  (u) − uk  (u))
= 2(a11 + a12 )u, a13 = 0. (5.12)
uk  (u)2
102 M. Choi et al. J. Geom.

By (5.10) and (5.12), a11 + a12 = a33 . From the above equations involving aij ,
the matrix A has the form as the following:
⎛ ⎞ ⎛ ⎞
a11 a12 0 b1
A = ⎝ −a12 a22 0 ⎠ , B = ⎝ b1 ⎠ ,
0 0 a33 0
where a11 + a12 = −a12 + a22 = a33 .
On the other hand, Eq. (5.11) with the help of (5.10) becomes

a33 uk  (u) + 2 k  (u) = a33 k(u) + 2a12 u + 2b1 . (5.13)
Thus, from (5.10) and (5.13) we have the system of differential equations as
follows:
√ 
k (u)(k (u)−uk (u))
uk (u)2 = 2a33 u, (5.14)
a33 uk  (u) + 2 k  (u) = a33 k(u) + 2a12 u + 2b1 ,
where a11 + a12 = a33 .
On the other hand, differentiating the second equation of (5.14) with respect
to u gives
k  (u)
a33 uk  (u) +  = 2a12 . (5.15)
k  (u)
Now, we deal with (5.14) and (5.15) according to three cases concerning the
constants a12 and a33 .
Case 1 a33 = 0. In this case, from the first equation in (5.14) and (5.15), we
have

k  (u)(k  (u) − uk  (u)) k  (u)
= 0,  = 2a12 . (5.16)
uk  (u)2 k  (u)
Since the second fundamental form is non-degenerate, k  (u) = 0. Hence, a
general solution k(u) of (5.16) is given by
1
k(u) = λ2 u3 + c,
3
where λ = a12 = 0 and c is a constant of integration. Substituting the function
k(u) to in the second equation of (5.14), we can easily obtain b1 = 0 and so
B = 0. Also, the matrix A becomes
⎛ ⎞
−λ λ 0
⎝ −λ λ 0 ⎠ .
0 0 0
Consequently, the surface M is parameterized by
 
1 2 3 1
x(u, v) = λ u − u − uv 2 + c, λ2 u3 + u − uv 2 + c, −2uv (5.17)
3 3
for some constants λ = 0 and c (Fig. 6).
Case 2 a33 = 0 and a12 = 0.
Vol. 104 (2013) Classification of surfaces of revolution 103

In this case, combining the first equation in (5.14) and (5.15) we get the fol-
lowing ODE:
1 
k  (u) − 2a233 u2 k  (u)2 −
k (u) = 0,
u
from which, the general solution of the equation is given by
1
k(u) = − + c,
μ2 u
where μ = a33 and c is a constant of integration. Substituting the solution k(u)
in the second equation in (5.14), we have b1 = − 21 μc and B = − 21 μc(1, 1, 0).
Also, the matrix A reduces to μI3 , where I3 denotes the identity matrix of
degree 3. Thus, M is parameterized by
 
1 1
x(u, v) = − 2 − u − uv 2 + c, − 2 + u − uv 2 + c, −2uv (5.18)
μ u μ u
for some constants μ = 0 and c (Fig. 7).
Case 3 a33 = 0 and a12 = 0.
In this case, combining the first equation in (5.14) and (5.15) yields
 
 1
k (u) + 4a12 a33 u − k  (u) = 2a233 u2 k  (u)2 . (5.19)
u
From this, k  (u) is given by
4a212 u
k  (u) = .
2a12 a33 u2 + 1
It follows that the general solution k(u) of (5.19) can be given by
  
a12  a33 2 1 
k(u) = 
ln  u + 2  du + c1 u + c2 , (5.20)
a33 2a12 4a12

Figure 6 Surface given by (5.17)

Figure 7 Surface given by (5.18)


104 M. Choi et al. J. Geom.

where c1 and c2 are constants of integration. Thus, M is parameterized by

x(u, v) = (k(u) − u − uv 2 , k(u) + u − uv 2 , −2uv), (5.21)

where k(u) is the function defined by (5.20) (Figs. 8, 9).


Similarly, we have the same conclusion as in case of time-like surface, that is,
k  (u) < 0.
Consequently, we have
Theorem 5.1. Let M be a surface of revolution with light-like axis in Minkow-
ski 3-space satisfying the condition Δh x = Ax + B for some constant matrix
A ∈ M at(3, R) and some constant vector B ∈ R. Then M is an open part of
the surface of revolution given by (5.17), (5.18) or (5.21).
Theorem 5.2. (Classification) Let M be a surface of revolution with the non-
degenerate second fundamental form h in Minkowski 3-space E31 satisfying the
condition Δh x = Ax + B for a matrix A ∈ M at(3, R) and a vector B ∈ R.
Then, M is one of the following:

Figure 8 Surface given by (5.21)

Figure 9 Graph of k(u) given by (5.20)


Vol. 104 (2013) Classification of surfaces of revolution 105

1. M is an open piece of a catenoid of the 1st kind, the 2nd kind, the 3rd
kind, the 4th kind or the 5th kind.
2. M is an open piece of a pseudo-sphere S21 (c, r) for some constant vector
c and r > 0.
3. M is an open piece of a hyperbolic space H2 (c, r) for some constant vector
c and r > 0.
4. M is open piece of the surface given by (5.17), (5.18) or (5.21).

Combining Theorem 4.1, Theorem 4.2 and the result of [1,8], we have
Theorem 5.3. Let x : M → E31 be an isometric immersion of a surface of
revolution with the non-degenerate second fundamental form h in E31 , whose
axis is space-like or time-like. Then, the following are equivalent:

1. Δx = Ax + B for a matrix A ∈ M at(3, R) and a vector B ∈ R3 .


2. Δh x = Ax + B for a matrix A ∈ M at(3, R) and a vector B ∈ R3 .
3. M is an open piece of a catenoid of the 1st kind, the 2nd kind, the 3rd
kind, the 4th kind or the 5th kind.
4. M is an open piece of a pseudo-sphere S21 (c, r) for a constant vector c
and r > 0.
5. M is an open piece of a hyperbolic space H2 (c, r) for a constant vector c
and r > 0.

References
[1] Alı́as, L.J., Ferrández, A., Lucas, P.: Surfaces in the 3-dimensional Lorentz-
Minkowski space satisfying Δx = Ax + B. Pac. J. Math. 156, 201–208 (1992)
[2] Chen, B.-Y.: Total Mean Curvature and Submanifolds of Finite Type. World
Scientific, Singapore (1984)
[3] Dillen, F., Pas, J., Vertraelen, L.: On surfaces of finite type in Euclidean 3-space.
Kodai Math. J. 13, 10–21 (1990)
[4] Garay, O.J.: An extension of Takahashi’s theorem. Geom. Dedicata 34, 105–112
(1990)
[5] Kaimakamis, G., Papantoniou, B.: Surfaces of revolution in the 3-dimensional
Lorentz-Minkowski space satisfying ΔII r = Ar. J. Geom. 81, 81–92 (2004)
[6] Ki, U.-H., Kim, D.-S., Kim, Y.H., Roh, Y.-M.: Surfaces of revolution with point-
wise 1-type Gauss map in Minkowski 3-space. Taiwan. J. Math. 13, 317–338
(2009)
[7] Takahashi, T.: Minimal immersions of Riemannian manifolds. J. Math. Soc.
Japan 18, 380–385 (1996)
[8] Woestijne, I.V.: Minimal surfaces in the 3-dimensional Minkowski space. In:
Morvan, J.M., et al. (eds.) Geometry and Topology of Submanifolds, vol. II, pp.
344–369. World Scientific, Singapore (1990)
106 M. Choi et al. J. Geom.

Miekyung Choi
Department of Mathematics Education and RINS
Gyeongsang National University
Jinju 660-701, Korea
e-mail: [email protected]

Young Ho Kim
Department of Mathematics Education
Kyungpook National University
Taegu 702-701, Korea
e-mail: [email protected]

Dae Won Yoon


Department of Mathematics Education and RINS
Gyeongsang National University
Jinju 660-701, Korea
e-mail: [email protected]

Received: 14 February 2012.


Revised: 31 January 2013.

You might also like