Classification of harmonic homomorphisms between Riemannian three-dimensional unimodular Lie groups

Abdelkader, Zagane; Nada, Osamnia; Zegga, Kaddour

doi:10.1108/AJMS-01-2022-0010

Purpose

The purpose of this study is to classify harmonic homomorphisms ϕ : (G, g) → (H, h), where G, H are connected and simply connected three-dimensional unimodular Lie groups and g, h are left-invariant Riemannian metrics.

Design/methodology/approach

This study aims the classification up to conjugation by automorphism of Lie groups of harmonic homomorphism, between twodifferent non-abelian connected and simply connected three-dimensional unimodular Lie groups (G, g) and (H, h), where g and h are two left-invariant Riemannian metrics on G and H, respectively.

Findings

This study managed to classify some homomorphisms between two different non-abelian connected and simply connected three-dimensional uni-modular Lie groups.

Originality/value

The theory of harmonic maps into Lie groups has been extensively studied related homomorphism in compact Lie groups by many mathematicians, harmonic maps into Lie group and harmonics inner automorphisms of compact connected semi-simple Lie groups and intensively study harmonic and biharmonic homomorphisms between Riemannian Lie groups equipped with a left-invariant Riemannian metric.

1. Introduction

The theory of harmonic maps is old and rich and has gained a growing interest in the past decade (see Ref. [1] and others). The theory of harmonic maps into Lie groups has been extensively studied related homomorphism in compact Lie groups by many mathematicians (see for examples [2]), in particular, harmonic maps into Lie groups [3] and harmonic inner automorphisms of compact connected semi-simple Lie groups in Ref. [4] and intensively study harmonic and biharmonic homomorphisms between Riemannian Lie groups equipped with a left-invariant Riemannian metric in Ref. [5].

The investigations described here are motivated by the paper [6], the author studied the classification, up to conjugation by an automorphism of Lie groups, of harmonic and biharmonic maps f : (G, g₁) → (G, g₂), where G is non-abelian connected and simply connected three-dimensional unimodular Lie group, f is a homomorphism of Lie group and g₁, g₂ are two left-invariant Riemannian metrics. The Lie group is unimodular if every left Haar measure is a right Haar measure and vice versa. It is known that G is unimodular if and only if | det Ad_x| = 1 for all x ∈ G if and only if the tracead(X) = 0 for all X in its Lie algebra $g$ if and only if $g$ is unimodular.

There are five non-abelian connected and simply connected three-dimensional unimodular Lie groups, the nilpotent Lie group (or the Heisenberg group), the special unitary group SU(2), the universal covering group $\tilde{P S L} (2, R)$ of the special linear group, the solvable Lie groups Sol and the universal covering group ${\tilde{E}}_{0} (2)$ of the connected component of the Euclidean group, for more detail, see Ref. [7].

In this paper, we aim the classification up to conjugation by an automorphism of Lie groups of harmonic homomorphism, between two different non-abelian connected, and simply connected three-dimensional unimodular Lie groups ϕ : (G, g) → (H, h), where g and h are two left-invariant Riemannian metrics on G and H, respectively.

2. Preliminaries

Let φ : (M, g) → (N, h) be a smooth map between two Riemannian manifolds with m = dim M and n = dim N. We denote by ∇^M and ∇^N the Levi-Civita connexions associated, respectively, to g and h and by T^φN the vector bundle over M pull-back of TN by φ. It is a Euclidean vector bundle and the tangent map of φ is a bundle homomorphism dφ : TM → T^φN. Moreover, T^φN carries a connexion ∇^φ pull-back of ∇^N by φ and there is a connexion on the vector bundle End(TM, T^φN) given by

(\nabla_{X} A) (Y) = \nabla_{X}^{ϕ} A (Y) - A (\nabla_{X}^{M} Y), X, Y \in Γ (T M), A \in Γ (E n d (T M, T^{φ} N)) .

The map φ is called harmonic if it is a critical point of the energy

E (φ) = \frac{1}{2} \int_{M} | d φ^{2} | v_{g} .

The corresponding Euler-Lagrange equation for the energy is given by the vanishing of the tension field

τ (φ) = t r \nabla d φ = \sum_{i = 1}^{m} (\nabla_{e_{i}} d φ) e_{i},

where ${(e_{i})}_{i = 1}^{m}$ is a local frame of orthonormal vector fields.Let (G, g) be a Riemannian Lie group, i.e., a Lie group endowed with a left-invariant Riemannian metric. If $g = T_{e} G$ is its Lie algebra and $< {, >}_{g} = g (e)$ ⁠, then there exists a unique bilinear map $A : g \times g \to g$ called the Levi-Civita product associated with $(g, < {, >}_{g})$ given by the formula:

2 < A_{u} v, {w >}_{g} = < {[u, v]}^{g}, {w >}_{g} + < {[w, u]}^{g}, {v >}_{g} + < {[w, v]}^{g}, {u >}_{g} .

A is entirely determined by the following properties

for any $u, v \in g, A_{u} v - A_{v} u = {[u, v]}^{g}$ ⁠,
for any $u, v, w \in g, < A_{u} v, {w >}_{g} + < v, A_{u} {w >}_{g} = 0$ ⁠.

If we denote by u^ℓ the left-invariant vector field on G associated with $u \in g$ then the Levi-Civita connection associated with (G, g) satisfies $\nabla_{u^{ℓ}} v^{ℓ} = {(A_{u} v)}^{ℓ}$ ⁠, the couple $(g, <, > g)$ defines a vector denoted $U^{g}$ by

< U^{g}, v > g = t r (a d_{v}), f o r a n y v \in g .

One can deduce easily that, for any orthonormal basis ${(e_{i})}_{i = 1}^{m}$ of $g$ ⁠,

U^{g} = \sum_{i = 1}^{m} A_{e_{i}} e_{i} .

Note that $g$ is unimodular if and only if $U^{g} = 0$ ⁠.

Let φ : (G, g) → (H, h) be a Lie group homomorphism between two Riemannian Lie groups. The differential $ξ : g \to h$ of φ at e is a Lie algebra homomorphism. There is a left action of G on Γ(T^φH) given by

(a . X) (b) = T_{φ (a b)} L_{φ (a^{- 1})} X (a b), a, b \in G, X \in Γ (T^{φ} H) .

A section X of T^φH is called left-invariant if, for any a ∈ G, a.X = X. For any left-invariant section X of T^φH, we have for any a ∈ G, X(a) = (X(e))^ℓ(φ(a)). Thus the space of left-invariant sections is isomorphic to the Lie algebra $h$ ⁠. Since φ is a homomorphism of Lie groups, g and h are leftinvariant, one can see easily that τ(φ) is left invariant and hence φ is harmonic if and only if τ(φ)(e) = 0. Now, one can see easily that

τ (ξ) ≔ τ (φ) (e) = U^{ξ} - ξ (U^{g}),

where

U^{ξ} = \sum_{i = 1}^{m} B_{ξ (e_{i})} ξ (e_{i}),

where B is the Levi-Civita product associated with $(h, < {, >}_{h})$ and ${(e_{i})}_{i = 1}^{m}$ is an orthonormal basis of $g$ ⁠. So we get the following proposition.

Proposition 2.1.

Let ϕ : G → H be a homomorphism between two Riemannian Lie groups. Then ϕ is harmonic if only if τ(ξ) = 0, where $ξ : g \to h$ is the differential of ϕ at e. The classification of harmonic homomorphisms will be done up to conjugation.

Two homomorphisms between Euclidean Lie algebras:

ξ_{1} : (g, < {, >}_{g}) \to (h, < {, >}_{h}) a n d ξ_{2} : (g, < {, >}_{g}) \to (h, < {, >}_{h})

are conjugate if there exists two isometric automorphisms $φ_{1} : (g, < {, >}_{g}) \to (g, < {, >}_{g})$ and $φ_{2} : (h, < {, >}_{h}) \to (h, < {, >}_{h})$ such that

φ_{2} ° ξ_{1} = ξ_{2} ° φ_{1} .

(1.1)

Proposition 2.2.

Let $ξ : (g, < {, >}_{g}) \to (h, < {, >}_{h})$ be a homomorphism between unimodular Euclidean Lie algebras, the following formula was established in [5]

< τ (ξ), {X >}_{h} = t r_{g} (ξ^{*} ° a d_{X} ° ξ) \forall X \in h

(1.2).

where $ξ^{*} : h \to g$ is given by

< ξ^{*} U, {V >}_{g} = < U, ξ {V >}_{h}, f o r V \in g a n d U \in h .

(1.3).

3. Riemanian three-dimensional unimodular Lie groups G

Definition 3.1.

The Heisenberg group Nil

The nilpotent Lie group Nil known as Heisenberg group, whose Lie algebra will be denoted by $n$ ⁠. We have

N i l = \{(\begin{matrix} 1 & a & c \\ 0 & 1 & b \\ 0 & 0 & 1 \end{matrix}), w i t h a, b, c \in R\}

and

n = \{(\begin{matrix} 0 & x & z \\ 0 & 0 & y \\ 0 & 0 & 0 \end{matrix}), w i t h x, y, z \in R\} .

The Lie algebra $n$ has a basis {X, Y, Z}, where $X = (\begin{matrix} 0 & 1 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix})$ ⁠, $Y = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & 0 & 0 \end{matrix})$ and $Z = (\begin{matrix} 0 & 0 & 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix})$ ⁠, where the non-vanishing Lie bracket is [X, Y] = Z.

Proposition 3.1.

[7]

Any left-invariant metric on Nil is equivalent up to automorphism to a metric whose associated matrix is of the form

< {, >}_{n} = (\begin{matrix} ρ & 0 & 0 \\ 0 & ρ & 0 \\ 0 & 0 & 1 \end{matrix}), w h e r e ρ > 0 .

(3.1)

Definition 3.2.

The solvable Lie group Sol

The solvable Lie group Sol whose Lie algebra will be denoted by $s o l$ ⁠. We have $s o l = R^{2} ⋊_{ι} R$ where $ι (t) = (\begin{matrix} t & 0 \\ 0 & - t \end{matrix})$ ⁠. We can choose a basis {X, Y, Z} of $s o l$ ⁠, where $X = ((\begin{matrix} 1 \\ 0 \end{matrix}), 0)$ ⁠, $Y = ((\begin{matrix} 0 \\ 1 \end{matrix}), 0)$ and $Z = ((\begin{matrix} 0 \\ 0 \end{matrix}), 1)$ ⁠.

and the non-vanishing Lie brackets are [Z, X] = X and [Y, Z] = Y. The Lie group of the solvable Lie algebra $s o l = R^{2} ⋊_{ι} R$ is the solvable Lie group Sol, which is the semi-direct product $R^{2} ⋊_{Θ} R$ ⁠, where $t \in R$ acts on $R^{2}$ by $Θ (t) = (\begin{matrix} e^{t} & 0 \\ 0 & e^{- t} \end{matrix})$ ⁠.

Proposition 3.2.

[7]

Any left-invariant metric on $S o l = R^{2} ⋊_{Θ} R$ is equivalent up to automorphism to a metric whose associated matrix is of the form

< {, >}_{s o l} = (\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & ν \end{matrix}), w h e r e ν > 0,

(3.2)

Or

< {, >}_{s o l} = (\begin{matrix} 1 & 1 & 0 \\ 1 & μ & 0 \\ 0 & 0 & ν \end{matrix}), w h e r e ν > 0 a n d μ > 1 .

(3.3)

Definition 3.3.

The solvable Lie group ${\tilde{E}}_{0} (2)$

The solvable Lie group ${\tilde{E}}_{0}$ whose Lie algebra will be denoted by $e_{0} (2)$ ⁠, where $e_{0} (2) = R^{2} ⋊ s o (2)$ ⁠. We can choose a basis {X, Y, Z} of $e_{0} (2)$ where $X = ((\begin{matrix} 1 \\ 0 \end{matrix}), (\begin{matrix} 0 & 0 \\ 0 & 0 \end{matrix}))$ ⁠, $Y = ((\begin{matrix} 0 \\ 1 \end{matrix}), (\begin{matrix} 0 & 0 \\ 0 & 0 \end{matrix}))$ ⁠, $Z = ((\begin{matrix} 0 \\ 0 \end{matrix}), (\begin{matrix} 0 & - 1 \\ 1 & 0 \end{matrix}))$ and the non-vanishing Lie brackets are [Z, X] = Y, [Y, Z] = X.

The Lie algebra $e_{0} (2) = R^{2} ⋊ s o (2)$ is Lie algebra of the Lie group $E_{0} (2) = R^{2} ⋊ S O (2)$ ⁠.

The group E₀(2) is not simply connected. The unique simply connected Lie group corresponding to the Lie algebra $e_{0} = R^{2} ⋊ s o (2)$ is universal covering group ${\tilde{E}}_{0} (2)$ of E₀(2).

The group ${\tilde{E}}_{0} (2)$ is the semi-direct product $C ⋊_{R}$ ⁠, where (z, t).(z′, t′) = (z + z′e^2iπt, t + t′) has a faithful matrix representation in $G L (3, C)$ by

(z, t) \mapsto (\begin{matrix} e^{2 i π t} & z & 0 \\ 0 & 1 & 0 \\ 0 & 0 & e^{t} \end{matrix}),

where $z \in C$ and $t \in R$ ⁠.

Proposition 3.3.

[7]

Any left-invariant metric on ${\tilde{E}}_{0} (2)$ is equivalent up to automorphism to a metric whose associated matrix is of the form

(\begin{matrix} 1 & 0 & 0 \\ 0 & ϱ & 0 \\ 0 & 0 & σ \end{matrix}), w h e r e σ > 0 a n d 0 < ϱ \leq 1 .

(3.4)

4. Harmonic homomorphisms between Sol and Nil

The following result gives a complete classification of harmonic homomorphisms between $s o l$ equipped with the left-invariant metric defined in (3.2) or (3.3) and $n$ equipped with the left-invariant metric defined in (3.1).

Theorem 4.1.

A homomorphism from $s o l$ to $n$ is conjugate to $ξ : s o l \to n$ ⁠, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}) w i t h a, b, c \in R .

Proof.

The basis of $s o l$ is {X, Y, Z} where [Z, X] = X, and [Y, Z] = Y.

The basis of $n$ is {E, F, H} with [E, F] = H. we put

\begin{aligned} ξ & : X \mapsto a_{1} E + b_{1} F + c_{1} H \\ Y \mapsto a_{2} E + b_{2} F + c_{2} H \\ Z \mapsto a_{3} E + b_{3} F + c_{3} H . \end{aligned}

Thus, we obtain

\{\begin{cases} [ξ X, ξ Y] = ξ [X, Y] = 0 \\ [ξ X, ξ Z] = ξ [X, Z] = - ξ X \\ [ξ Y, ξ Z] = ξ [Y, Z] = ξ Y \end{cases} \Leftrightarrow a_{1} = b_{1} = c_{1} = a_{2} = b_{2} = c_{2} = 0 .

Theorem 4.2.

Let $ξ : s o l \to n$ a homomorphism, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}),

(4.1)

the Lie algebra $s o l$ equipped with the left-invariant metric defined in (3.2) or (3.3) and $n$ equipped with the left-invariant metric defined in (3.1). Then

τ (ξ) = \frac{b c}{ν} E - \frac{a c}{ν} F .

(4.2)

Proof.

We have

a d_{E} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 1 & 0 \end{matrix}), a d_{F} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ - 1 & 0 & 0 \end{matrix}) a n d a d_{H} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix}) .

Using formula (1.3) where $U \in n$ and $V \in s o l$ ⁠, we obtain

ξ^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ ρ a & ρ b & c \end{matrix}) .

Using formula (1.2), a simple calculation gives us

< τ (ξ), {E >}_{n} = t r (ξ^{*} ° a d_{E} ° ξ) = \frac{b c}{ν},

< τ (ξ), {F >}_{n} = t r (ξ^{*} ° a d_{F} ° ξ) = \frac{- a c}{ν}

and

< τ (ξ), {H >}_{n} = t r (ξ^{*} ° a d_{H} ° ξ) = 0

Corollary 4.1.

$ξ : (s o l, < {, >}_{s o l}), \to (n < {, >}_{s o l})$ is harmonic if and only if (a = b = 0 or c = 0).

Theorem 4.3.

A homomorphism from $n$ to $s o l$ is conjugate to $ξ_{i = 1,2} : n \to s o l$ ⁠, where

ξ_{1} = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & 0 \end{matrix}) w i t h a, b \in R,

(4.3)

Or

ξ_{2} = (\begin{matrix} a_{1} & a_{2} & 0 \\ b_{1} & b_{2} & 0 \\ 0 & 0 & 0 \end{matrix}) w i t h a_{i}, b_{i} \in R f o r i = 1,2 .

(4.4)

Proof.

The basis of $s o l$ is {X, Y, Z}, where [Z, X] = X and [Y, Z] = Y, the basis of $n$ is {E, F, H} with [E, F] = H, then we can suppose

E \mapsto a_{1} X + b_{1} Y + c_{1} Z,

F \mapsto a_{2} X + b_{2} Y + c_{2} Z

and

H \mapsto a_{3} X + b_{3} Y + c_{3} Z .

Thus we obtain

\{\begin{cases} ξ [E, F] = [ξ E, ξ F] = ξ H \\ ξ [E, H] = [ξ E, ξ H] = 0 \\ ξ [F, H] = [ξ E, ξ H] = 0 \end{cases} \Leftrightarrow (c_{3} = a_{1} = b_{1} = c_{1} = a_{2} = b_{2} = c_{2} = 0 o r a_{3} = b_{3} = c_{3} = c_{1} = c_{2} = 0) .

Theorem 4.4.

Let $ξ_{1}, ξ_{2} : n \to s o l$ be homomorphisms, where ξ₁ and ξ₂ are defined in formulas (4.3) and (4.4), the Lie algebra $s o l$ is equipped with the left-invariant metric defined in formula (3.2). Then

τ (ξ_{1}) = (a^{2} - b^{2}) Z

(4.5)

and

τ (ξ_{2}) = \frac{((a_{1}^{2} - b_{1}^{2}) + (a_{2}^{2} - b_{2}^{2}))}{ρ} Z .

(4.6)

Proof.

We have

a d_{X} = (\begin{matrix} 0 & 0 & - 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix}), a d_{Y} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & 0 & 0 \end{matrix}) a n d a d_{Z} = (\begin{matrix} 1 & 0 & 0 \\ 0 & - 1 & 0 \\ 0 & 0 & 0 \end{matrix}) .

For the homomorphism ξ₁, using formula (1.3), where $V \in n$ and $U \in s o l$ ⁠, we obtain

ξ_{1}^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ a & b & 0 \end{matrix}) .

Using formula (1.2), a simple calculation gives us

< τ (ξ_{1}), {X >}_{s o l} = t r (ξ_{1}^{*} ° a d_{X} ° ξ_{1}) = 0,

< τ (ξ_{1}), {Y >}_{s o l} = t r (ξ_{1}^{*} ° a d_{Y} ° ξ_{1}) = 0

and

< τ (ξ_{1}), {Z >}_{s o l} = t r (ξ_{1}^{*} ° a d_{Z} ° ξ_{1}) = a^{2} - b^{2} .

For the homomorphism ξ₂, we have

ξ_{2}^{*} = (\begin{matrix} a_{1} & b_{1} & 0 \\ a_{2} & b_{2} & 0 \\ 0 & 0 & 0 \end{matrix}) .

By using formula (1.2), we obtain

< τ (ξ_{2}), {X >}_{s o l} = t r (ξ_{2}^{*} ° a d_{X} ° ξ_{2}) = 0,

< τ (ξ_{2}), {Y >}_{s o l} = t r (ξ_{2}^{*} ° a d_{Y} ° ξ_{2}) = 0

and

< τ (ξ_{2}), {Z >}_{s o l} = t r (ξ_{2}^{*} ° a d_{Z} ° ξ_{2}) = \frac{((a_{1}^{2} - b_{1}^{2}) + (a_{2}^{2} - b_{2}^{2}))}{ρ} .

Corollary 4.2.

$ξ_{1} : (n, < {, >}_{n}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if a = ±b.

$ξ_{2} : (n, < {, >}_{n}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if $a_{1}^{2} + a_{2}^{2} = b_{2}^{2} + b_{1}^{2}$ ⁠.

Theorem 4.5.

Let $ξ_{1}, ξ_{2} : n \to s o l$ be homomorphisms, where ξ₁ and ξ₂ are defined in formulas (4.3), (4.4) and the Lie algebra $s o l$ is equipped with the left-invariant metric defined in formula (3.3). Then

τ (ξ_{1}) = (a^{2} - μ b^{2}) Z,

(4.7)

τ (ξ_{2}) = \frac{((a_{1}^{2} - μ b_{1}^{2}) + (a_{2}^{2} - μ b_{2}^{2}))}{ρ} Z .

(4.8)

Proof.

By using formula (1.3), where $V \in n$ and $U \in s o l$ ⁠, we obtain:

For ξ₁

ξ_{1}^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ a + b & a + μ b & 0 \end{matrix}) .

Using formula (1.2), we get

< τ (ξ_{1}), {X >}_{s o l} = t r (ξ_{1}^{*} ° a d_{X} ° ξ_{1}) = 0,

< τ (ξ_{1}), {Y >}_{s o l} = t r (ξ_{1}^{*} ° a d_{Y} ° ξ_{1}) = 0

and

< τ (ξ_{1}), {Z >}_{s o l} = t r (ξ_{1}^{*} ° a d_{Z} ° ξ_{1}) = a^{2} - μ b^{2} .

For ξ₂, we have

ξ_{2}^{*} = (\begin{matrix} a_{1} + b_{1} & a_{1} + μ b_{1} & 0 \\ a_{2} + b_{2} & a_{2} + μ b_{2} & 0 \\ 0 & 0 & 0 \end{matrix}),

furthermore

< τ (ξ_{2}), {X >}_{s o l} = t r (ξ_{2}^{*} ° a d_{X} ° ξ_{2}) = 0,

< τ (ξ_{2}), {Y >}_{s o l} = t r (ξ_{2}^{*} ° a d_{Y} ° ξ_{2}) = 0

and

< τ (ξ_{2}), {Z >}_{s o l} = t r (ξ_{2}^{*} ° a d_{Z} ° ξ_{2}) = \frac{((a_{1}^{2} - μ b_{1}^{2}) + (a_{2}^{2} - μ b_{2}^{2}))}{ρ} .

Corollary 4.3.

$ξ_{1} : (n < {, >}_{n}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if $a = \pm \sqrt{μ} b$ ⁠.

$ξ_{2} : (n < {, >}_{n}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if $a_{1}^{2} + a_{2}^{2} = \sqrt{μ} (b_{2}^{2} + b_{1}^{2})$ ⁠.

5. Harmonic homomorphisms between Sol and ${\tilde{E}}_{0} (2)$

The following result gives a complete classification of harmonic homomorphisms between $s o l$ equipped with the left-invariant metric defined in (3.2), (3.3) and $e_{0} (2)$ equipped with the left-invariant metric defined in (3.4).

Theorem 5.1.

Any homomorphism from $s o l$ to $e_{0} (2)$ is conjugate to $ξ : s o l \to e_{0} (2)$ ⁠, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}), s u c h t h a t a, b, c \in R .

Proof.

The basis of $s o l$ is {X, Y, Z} where [Z, X] = X, [Y, Z] = Y and the basis of $e_{0} (2)$ is {A, B, C} with [A, B] = 0, [C, A] = B and [B, C] = A, we suppose

ξ (X) = a_{1} A + b_{1} B + c_{1} C,

ξ (Y) = a_{2} A + b_{2} B + c_{2} C

and

ξ (Z) = a_{3} A + b_{3} B + c_{3} C .

Thus we obtain

\{\begin{cases} [ξ X, ξ Y] = ξ [X, Y] = 0 \\ [ξ X, ξ Z] = ξ [X, Z] = - ξ X \\ [ξ Y, ξ Z] = ξ [Y, Z] = ξ Y \end{cases} \Leftrightarrow (a_{1} = b_{1} = c_{1} = a_{2} = b_{2} = c_{2} = 0) .

Theorem 5.2.

Let $ξ : s o l \to e_{0} (2)$ be a homomorphism, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}),

(5.1)

and $s o l$ equipped with the left-invariant metric defined in (3.2) or in (3.3), then

τ (ξ) = \frac{1}{ν} (- ϱ b c A + a c B + (ϱ - 1) a b C) .

(5.2)

Proof.

We have

a d_{A} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & - 1 \\ 0 & 0 & 0 \end{matrix}), a d_{B} = (\begin{matrix} 0 & 0 & 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix}), a d_{C} = (\begin{matrix} 0 & - 1 & 0 \\ 1 & 0 & 0 \\ 0 & 0 & 0 \end{matrix}) .

By using formula (1.3), where $U \in e_{0} (2)$ and $V \in s o l$ ⁠, we obtain

ξ^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ a & ϱ b & σ c \end{matrix}) .

Use formula (1.2), we get

< τ (ξ), {A >}_{n} = t r (ξ^{*} ° a d_{A} ° ξ) = \frac{- ϱ b c}{ν},

< τ (ξ), {B >}_{n} = t r (ξ^{*} ° a d_{B} ° ξ) = \frac{a c}{ν}

and

< τ (ξ), {C >}_{n} = t r (ξ^{*} ° a d_{C} ° ξ) = \frac{(ϱ - 1) a b}{ν} .

Corollary 5.1.

$ξ : (s o l, < {, >}_{s o l}), \to (e_{0} (2) < {, >}_{e_{0} (2)})$ is harmonic if and only if (ϱ = 1 and c = 0) or (a = b = 0).

Theorem 5.3.

A homomorphism from $e_{0} (2)$ to $s o l$ is conjugate to $ξ : e_{0} (2) \to s o l$ ⁠, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}), s u c h t h a t a, b, c \in R .

Proof.

$ξ : e_{0} (2) \to s o l$ ⁠, we have

A \mapsto a_{1} X + b_{1} Y + c_{1} Z,

B \mapsto a_{2} X + b_{2} Y + c_{2} Z

and

C \mapsto a_{3} X + b_{3} Y + c_{3} Z .

Thus we obtain

\{\begin{cases} [ξ A, ξ B] = ξ [A, B] = 0 \\ [ξ A, ξ C] = ξ [A, C] = - ξ B \\ [ξ B, ξ C] = ξ [B, C] = ξ A \end{cases} \Leftrightarrow (a_{1} = b_{1} = c_{1} = a_{2} = b_{2} = c_{2} = 0) .

Theorem 5.4.

Let $ξ : e_{0} (2) \to s o l$ be a homomorphism, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}),

(5.3)

and $s o l$ equipped with the left-invariant metric defined in (3.2). Then

τ (ξ) = \frac{1}{σ} (- a c X + b c Y + (a^{2} - b^{2}) Z) .

(5.4)

Proof.

By using formula (1.3) where $V \in e_{0} (2)$ and $U \in s o l$ ⁠, we obtain

ξ^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ a & b & ν c \end{matrix}) .

By direct calculation and we use formula (1.2), we obtain

< τ (ξ), {X >}_{n} = t r (ξ^{*} ° a d_{X} ° ξ) = \frac{- a c}{σ},

< τ (ξ), {Y >}_{n} = t r (ξ^{*} ° a d_{Y} ° ξ) = \frac{b c}{σ}

and

< τ (ξ), {Z >}_{n} = t r (ξ^{*} ° a d_{Z} ° ξ) = \frac{(a^{2} - b^{2})}{σ} .

Corollary 5.2.

$ξ : (e_{0} (2) < {, >}_{e_{0} (2)}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if (c = 0 and a = ±b or a = b = 0).

Theorem 5.5.

Let $ξ : e_{0} (2) \to s o l$ be a homomorphism, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}) .

(5.5)

Where $s o l$ equipped with left-invariant metric define in (3.3).

Then

τ (ξ) = \frac{1}{σ} (- (a + b) c X + μ b c Y + (a^{2} - μ b^{2} + a b) Z) .

(5.6)

Proof.

by a similar calculation, we get $ξ^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ a + b & μ b & ν c \end{matrix})$ ⁠.Using formula (1.2), a direct calculation gives us

< τ (ξ), {X >}_{n} = t r (ξ^{*} ° a d_{X} ° ξ) = - \frac{(a + b) c}{σ},

< τ (ξ), {Y >}_{n} = t r (ξ^{*} ° a d_{Y} ° ξ) = \frac{μ b c}{σ}

and

< τ (ξ), {Z >}_{n} = t r (ξ^{*} ° a d_{Z} ° ξ) = \frac{a^{2} - μ b^{2} + a b}{σ} .

Corollary 5.3.

$ξ : (e_{0} (2) < {, >}_{e_{0} (2)}) \to (s o l, < {, >}_{s o l})$ is harmonic if and only if (a = b = 0) or (b = c = 0).

6. Harmonic homomorphisms between Nil and ${\tilde{E}}_{0} (2)$

The following result gives a complete classification of harmonic homomorphisms between $n$ equipped with the left-invariant metric defined in (3.1) and $e_{0} (2)$ equipped with the left-invariant metric defined in (3.4).

Theorem 6.1.

A homomorphism from $e_{0} (2)$ to $n$ is conjugate to $ξ : e_{0} (2) \to n$ ⁠, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}), s u c h t h a t a, b, c \in R .

Proof.

The basis of $e_{0} (2)$ is {A, B, C} with [A, B] = 0, [C, A] = B, [B, C] = A and the basis of $n$ is {E, F, H} with [E, F] = H. Suppose that

A \mapsto a_{1} E + b_{1} F + c_{1} H,

B \mapsto a_{2} E + b_{2} F + c_{2} H,

and

C \mapsto a_{3} E + b_{3} F + c_{3} H .

Thus, we obtain

\{\begin{cases} [ξ A, ξ B] = ξ [A, B] = 0 \\ [ξ A, ξ C] = ξ [A, C] = - ξ B \\ [ξ B, ξ C] = ξ [B, C] = ξ A \end{cases} \Leftrightarrow (a_{1} = b_{1} = c_{1} = a_{2} = b_{2} = c_{2} = 0) .

Theorem 6.2.

Let $ξ : e_{0} (2) \to n$ a homomorphism, where

ξ = (\begin{matrix} 0 & 0 & a \\ 0 & 0 & b \\ 0 & 0 & c \end{matrix}) .

(6.1)

Then

τ (ξ) = \frac{b}{c} ν E - \frac{a c}{ν} F .

(6.2)

Proof.

We have $a d_{E} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 1 & 0 \end{matrix}), a d_{F} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ - 1 & 0 & 0 \end{matrix}),$ and $a d_{H} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix})$ ⁠.

using formula (1.3), where $U \in n$ and $V \in e_{0} (2)$ ⁠, we get

$ξ^{*} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & 0 \\ ρ a & ρ b & c \end{matrix})$ ⁠.

Using formula (1.2), a simple calculation gives us

< τ (ξ), {E >}_{n} = t r (ξ^{*} ° a d_{E} ° ξ) = \frac{b c}{σ}

and

< τ (ξ), {F >}_{n} = t r (ξ^{*} ° a d_{F} ° ξ) = - \frac{a c}{σ} .

Corollary 6.1.

$ξ : (e_{0} (2), < {, >}_{e_{0} (2)}) \to (n < {, >}_{n})$ is harmonic if and only if (a = b = 0 or c = 0).

Theorem 6.3.

A homomorphism from $n$ to $e_{0} (2)$ is conjugate to $ξ_{i} : n \to e_{0} (2)$ ⁠, with i = 1, 2, 3 where

ξ_{1} = (\begin{matrix} a & b & 0 \\ c & d & 0 \\ 0 & 0 & 0 \end{matrix}), w h e r e a, b, c, d \in R,

ξ_{2} = (\begin{matrix} 0 & a & 0 \\ 0 & b & 0 \\ 0 & c & 0 \end{matrix}) w h e r e a, b, c \in R

and

ξ_{3} = (\begin{matrix} a & d & 0 \\ b & \frac{c d}{a} & 0 \\ c & \frac{b d}{a} & 0 \end{matrix}) w i t h b, c, d \in R a n d a \in R^{*} .

Proof.

The basis of $e_{0} (2)$ is {A, B, C} such that [A, B] = 0, [C, A] = B, [B, C] = A and the basis of $n$ is {E, F, H} with [E, F] = H. We put

E \mapsto a_{1} A + b_{1} B + c_{1} C,

F \mapsto a_{2} A + b_{2} B + c_{2} C

and

H \mapsto a_{3} A + b_{3} B + c_{3} C .

Thus, we obtain

\{\begin{cases} [ξ E, ξ F] = ξ [E, F] = ξ H \\ [ξ E, ξ H] = 0 \\ [ξ F, ξ H] = ξ [F, H] = 0 \end{cases} \Leftrightarrow \{\begin{cases} a_{3} = b_{3} = c_{3} = 0 \\ c_{1} = c_{2} = 0 \end{cases} o r \{\begin{cases} a_{3} = b_{3} = c_{3} = 0 \\ a_{1} = b_{1} = c_{1} = 0 \end{cases} o r \{\begin{cases} a_{3} = b_{3} = c_{3} = 0 \\ c_{1} \times a_{2} = b_{2} \times a_{1} \\ b_{1} \times a_{2} = c_{2} \times a_{1} \end{cases}

Theorem 6.4.

Let $ξ_{i} : n \to e_{0} (2)$ be homomorphisms, where ${(ξ_{i})}_{i = 1,2,3}$ are defined in (Theorem 5.3.), then

τ (ξ_{1}) = (ϱ - 1) \frac{a c + b d}{ρ} C .

τ (ξ_{2}) = \frac{1}{ρ} (- ϱ b c A + a c B + a b (ϱ - 1) C) .

τ (ξ_{3}) = \frac{- ϱ b c}{ρ} (1 + \frac{d^{2}}{a^{2}}) A + \frac{1}{ρ} (a c + \frac{b^{2} d}{a}) B + \frac{1}{ρ} (\frac{c d}{a} (ϱ d - 1) + a b (ϱ - 1)) C .

Proof.

We have $a d_{A} = (\begin{matrix} 0 & 0 & 0 \\ 0 & 0 & - 1 \\ 0 & 0 & 0 \end{matrix}), a d_{B} = (\begin{matrix} 0 & 0 & 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{matrix}),$ and $a d_{C} = (\begin{matrix} 0 & - 1 & 0 \\ 1 & 0 & 0 \\ 0 & 0 & 0 \end{matrix})$ ⁠.

By using formula (1.3), where $U \in e_{0} (2)$ and $V \in n$ ⁠, we obtain

ξ_{1}^{*} = (\begin{matrix} a & ϱ c & 0 \\ b & ϱ d & 0 \\ 0 & 0 & 0 \end{matrix}) .

We use formula (1.2), we obtain

< τ (ξ_{1}), {A >}_{e_{0} (2)} = t r (ξ^{*} ° a d_{A} ° ξ_{1}) = 0,

< τ (ξ_{1}), {B >}_{e_{0} (2)} = t r (ξ^{*} ° a d_{B} ° ξ) = 0,

and

< τ (ξ_{1}), {C >}_{e_{0} (2)} = t r (ξ^{*} ° a d_{C} ° ξ) = (ϱ - 1) \frac{a d + b d}{ρ} .

For ξ = ξ₂, we have $ξ_{2}^{*} = (\begin{matrix} 0 & 0 & 0 \\ a & ϱ b & σ c \\ 0 & 0 & 0 \end{matrix})$ ⁠.

We use formula (1.2), we obtain

< τ (ξ_{2}), {A >}_{e_{0} (2)} = t r (ξ_{2}^{*} ° a d_{A} ° ξ_{2}) = - \frac{1}{ρ} ϱ b c,

< τ (ξ_{2}), {B >}_{e_{0} (2)} = t r (ξ_{2}^{*} ° a d_{B} ° ξ_{2}) = \frac{1}{ρ} a c

and

< τ (ξ_{2}), {C >}_{e_{0} (2)} = t r (ξ_{2}^{*} ° a d_{C} ° ξ_{2}) = \frac{1}{ρ} a b (ϱ - 1) .

For ξ = ξ₃, we have

ξ_{3}^{*} = (\begin{matrix} a & b ϱ & c σ \\ d & ϱ \frac{c d}{a} & σ \frac{b d}{a} \\ 0 & 0 & 0 \end{matrix}) .

(6.3)

We use formula (1.2), we obtain

< τ (ξ_{3}), {A >}_{e_{0} (2)} = t r (ξ_{3}^{*} ° a d_{A} ° ξ_{3}) = \frac{- ϱ b c}{ρ} (1 + \frac{d^{2}}{a^{2}}),

< τ (ξ_{3}), {B >}_{e_{0} (2)} = t r (ξ_{3}^{*} ° a d_{B} ° ξ_{3}) = \frac{1}{ρ} (a c + \frac{b d^{2}}{a})

and

< τ (ξ_{3}), {C >}_{e_{0} (2)} = t r (ξ_{3}^{*} ° a d_{C} ° ξ_{3}) = \frac{ϱ - 1}{ρ} (\frac{c d^{2}}{a} + a b) .

Corollary 6.2.

$ξ_{1} : (n < {, >}_{n}) \to (e_{0} (2), < {, >}_{e_{0} (2)})$ ⁠, is harmonic if and only if (ϱ = 1 or ac + bd = 0).

$ξ_{2} : (n < {, >}_{n}) \to (e_{0} (2), < {, >}_{e_{0} (2)})$ ⁠, is harmonic if and only if (b = c = 0 or ϱ = 1, and c = 0).

$ξ_{3} : (n < {, >}_{n}) \to (e_{0} (2), < {, >}_{e_{0} (2)})$ ⁠, is harmonic if and only if (b = c = 0 or c = d = 0 and ϱ = 1).

The authors thank the referee for many useful suggestions and corrections which improved the first version.

Osamnia Nada and Zegga Kaddour are contributed equally to this work.

References

1.

Baird

P

,

Wood

JC

.

Harmonic morphisms between Riemannian manifolds

.

Oxford Science Publications

;

2003

.

Google Scholar

Crossref

2.

Dai

YJ

,

Shon

M

,

Urakawa

H

.

Harmonic maps between into Lie groups and homogeneous spaces

.

Differ Geom Appl

.

1997

;

7

:

143

-

69

.

Google Scholar

Crossref

3.

Uhlenbeck

K

.

Harmonic maps into Lie groups (classical solutions of the chiral model)

.

J Differ Geom

.

1989

;

30

:

1

-

50

.

Google Scholar

Crossref

4.

Park

JS

.

Harmonic inner automorphisms of compact connected semisimple Lie groups

.

Tohoku Math J

.

1990

;

42

:

80

-

91

.

Google Scholar

Crossref

5.

Boucetta

M

,

Ouakkas

S

.

Harmonic and biharmonic homomorphisms between Riemannian Lie groups

.

J Geom Phys

.

2017

;

116

:

64

-

80

.

Google Scholar

Crossref

6

Boubekour

S

,

Boucetta

M

.

Harmonic and biharmonic homomorphisms between Riemannian three dimensional unimodular Lie groups

.

J Geom Phys

.

2021

;

164

:

104

-

78

.

Google Scholar

7.

Ha

KY

,

Lee

JB

.

Left invariant metrics and curvatures on simply connected three-dimensional Lie groups

.

Math Nachar

.

2009

;

282

:

868

-

98

.

Google Scholar

Crossref

2022

Zagane Abdelkader, Osamnia Nada and Kaddour Zegga

Published in the Arab Journal of Mathematical Sciences. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/legalcode

Classification of harmonic homomorphisms between Riemannian three-dimensional unimodular Lie groups

1. Introduction

2. Preliminaries

3. Riemanian three-dimensional unimodular Lie groups G

4. Harmonic homomorphisms between Sol and Nil

5. Harmonic homomorphisms between Sol and ${\tilde{E}}_{0} (2)$

6. Harmonic homomorphisms between Nil and ${\tilde{E}}_{0} (2)$

References

Email Alerts

Cited By

Classification of harmonic homomorphisms between Riemannian three-dimensional unimodular Lie groups Open Access

1. Introduction

2. Preliminaries

3. Riemanian three-dimensional unimodular Lie groups G

4. Harmonic homomorphisms between Sol and Nil

5. Harmonic homomorphisms between Sol and Ẽ0(2)

6. Harmonic homomorphisms between Nil and Ẽ0(2)

References

Email Alerts

Recommended for you

Cited By

Sharing Unavailable

Classification of harmonic homomorphisms between Riemannian three-dimensional unimodular Lie groups

5. Harmonic homomorphisms between Sol and ${\tilde{E}}_{0} (2)$

6. Harmonic homomorphisms between Nil and ${\tilde{E}}_{0} (2)$