WorksMathematics
Classification of real Lie algebras of dimension six
Abstract
We summarize a coordinate-free classification of six-dimensional real Lie algebras up to isomorphism. Instead of solving the overdetermined Jacobi system of 120 quadratic equations on 90 structure constants modulo , the method uses the Levi–Malcev decomposition, Mostow's nilradical constraints, and the Skjelbred–Sund cohomological central extension technique. In the semisimple sector, the four real forms of are classified by Killing signature invariants, with explicit matrix representatives for . For non-solvable algebras with three-dimensional radical, representation-theoretic rigidity shows that cannot couple indecomposably with non-abelian radicals, yielding exactly three indecomposable classes governed by and . In the solvable category, Mostow's theorem restricts the nilradical dimension to . The nilpotent sector in dimension six has twenty-four indecomposable isomorphism classes plus decomposable counterparts. Solvable non-nilpotent algebras are classified by nilradical codimension, with explicit real Jordan matrix representations for codimension one and two that resolve continuous moduli under gauge actions. The work also computes the minimal faithful representation dimension under Ado's theorem, relates the classification to invariant complex structures on nilmanifolds, and provides a deterministic, page-breakable identification algorithm.
Introduction and Historical Context
The classification of finite-dimensional Lie algebras over the real field is a foundational problem that underpins modern differential geometry, theoretical physics, Hamiltonian mechanics, and invariant theory. Lie algebras provide the infinitesimal framework for smooth transformation groups, determine the geometry of homogeneous spaces, and classify left-invariant structures on manifolds.
In dimension three, the landmark work of Bianchi (1898) classified the nine isometry classes of three-dimensional spaces, which was subsequently reformulated by Milnor [5] to describe left-invariant Riemannian metrics and Ricci curvature signatures. In dimension four, the classification of solvable real Lie algebras was achieved by Mubarakzyanov [7], revealing a finite collection of discrete algebras and single-parameter families that govern four-dimensional homogeneous cosmologies [9]. At dimension five, Mubarakzyanov [8] established the solvable catalog, where the first instances of multi-parameter continuous moduli emerge.
Dimension six represents the true mathematical watershed in Lie theory. Six-dimensional Lie algebras are indispensable across modern geometry and high-energy theoretical physics:
Complex and Symplectic Geometry of Nilmanifolds: A compact nilmanifold of real dimension six corresponds to a complex manifold of complex dimension three. The existence of left-invariant complex, symplectic, and Kähler structures is governed entirely by the Lie algebra cohomology of the six-dimensional Lie algebra [10].
Supergravity and String Flux Compactifications: In Type II superstring theory, compactifications of ten-dimensional spacetime on six-dimensional internal manifolds yield effective four-dimensional and supergravities. When the internal geometry is a twisted torus (a nilmanifold or solvmanifold), the structure constants of the six-dimensional
Lie algebra define the non-geometric and geometric fluxes (-fluxes) that generate cosmological vacuum potentials.
3. Sub-Riemannian Geometry and Control Theory: Six is the minimal dimension supporting non-trivial bracket-generating distributions with growth vectors such as and , modeled on free nilpotent algebras and Carnot groups.
Despite their importance, setting up the classification via direct coordinate methods leads to an intractable algebraic bottleneck. A six-dimensional real Lie algebra requires specifying structure constants , constrained by quadratic Jacobi equations:
modulo the 36-dimensional non-compact change-of-basis gauge group .
The historical resolution of this problem proceeded in stages: Morozov [6] cataloged the six-dimensional nilpotent algebras over , which was extended, corrected, and transferred to using cohomology by Skjelbred and Sund [11]. Non-solvable algebras admitting a non-trivial Levi decomposition were classified definitively by Turkowski [13].
The objective of this paper is to present a self-contained, conceptual, and coordinate-free classification of all six-dimensional real Lie algebras. Rather than relying on brute-force polynomial elimination, we deploy an invariant-driven architecture:
We classify the semisimple sector () using Cartan’s real forms and Killing form signatures, with explicit matrix representatives for the Lorentz algebra .
We resolve the non-solvable sector () using the representation theory of and , establishing the representation-theoretic obstruction that excludes .
For solvable algebras, we anchor the classification to the nilradical . By Mostow’s theorem, .
For (nilpotent sector), we catalog the twenty-four indecomposable classes using the Skjelbred–Sund cohomological central extension method.
For , we systematically construct exact real matrix representatives for all classes, paying particular attention to the complex Jordan sectors (algebras with complex conjugate spectra).
We determine the minimal faithful representation dimension under Ado’s theorem and construct a deterministic, page-breakable identification algorithm.
Geometric Obstructions and the Watershed of Dimension Six
To appreciate why dimension six requires a systematic cohomological approach, we analyze the algebraic geometry of the Lie bracket across low dimensions.
Comparison with Dimensions Three, Four, and Five
Let be an -dimensional real Lie algebra. The bracket is a linear map .
Dimension Three (): . The exterior duality reduces the Lie bracket to a single real matrix . The Jacobi identity simplifies to the single vector equation ([5]).
Dimension Four (): , while . Exterior duality fails, but the classification remains algebraically constrained: every non-solvable algebra is a direct sum (), and the nilradical of a solvable algebra cannot have codimension greater than two.
Dimension Five (): , with 50 structure constants and 50 quadratic equations. The first non-trivial Levi coupling appears (), and continuous parameters enter the solvable sector. However, all nilpotent algebras remain strictly discrete (six indecomposable classes).
Dimension Six (): Here:
The bracket map requires 90 structure constants constrained by 120 quadratic Jacobi equations.
Structural Novelties in Dimension Six
Beyond the dimension count, dimension six introduces four structural phenomena that do not exist in dimensions :
Cohomological Explosion of Nilpotent Classes: In dimension five, there are only six indecomposable nilpotent algebras. In dimension six, this number quadruples to twenty-four indecomposable classes, including non-trivial real forms () distinguished by quadratic Casimir invariants.
Free Nilpotent Structures: Dimension six is the lowest dimension supporting the free two-step nilpotent Lie algebra on three generators:
Its center coincides exactly with its derived ideal ().
Representation-Theoretic Levi Obstructions: For algebras with and , the compactness of completely obstructs non-trivial couplings with non-abelian radicals, while the split nature of allows both vector and fundamental representations.
Proliferation of Complex Jordan Sectors: In the solvable category, the presence of complex conjugate eigenvalues in outer derivations generates distinct continuous moduli varieties whose real forms require explicit rotational-dilational matrix descriptions.
Theoretical Machinery: Invariants and Decomposition Theorems
All Lie algebras in this monograph are finite-dimensional over . We deploy the following structural tools [3, 2, 4]:
(T1) Levi–Malcev Decomposition: Every finite-dimensional real Lie algebra decomposes uniquely as a semidirect product:
where is the solvable radical, and is a semisimple subalgebra (the Levi factor). The factor is unique up to inner automorphisms with . The representation specifies the cross brackets between and .
(T2) Mostow’s Bound on the Nilradical: The nilradical is the maximal nilpotent ideal of . For any finite-dimensional solvable real Lie algebra , Mostow’s theorem bounds the dimension of from below:
For , this immediately establishes that .
(T3) The Skjelbred–Sund Cohomological Method: Let be a nilpotent Lie algebra of dimension . The central extensions of by are parameterized by the second scalar cohomology group . The group acts naturally on the Grassmannian . The orbits of this action correspond bijectively to the isomorphism classes of central extensions with center [11].
(T4) Mubarakzyanov Derivation Gauge Action: For a solvable algebra , two derivations generate isomorphic Lie algebras if and only if there exist , , and such that:
(T5) Ado’s Theorem and Minimal Representation Dimension: Every finite-dimensional Lie algebra admits a faithful linear representation in . The invariant , defined by
is an intrinsic isomorphism invariant of .
The Semisimple Sector (r= 0)
When , the Lie algebra is semisimple of dimension six.
Theorem 4.1 (Classification of 6D Real Semisimple Lie Algebras). Up to isomorphism, there exist precisely four mutually non-isomorphic six-dimensional real semisimple Lie algebras. Exactly one is simple over , while three are direct sums of two three-dimensional simple ideals:
(Compact form, Killing signature ),
(Mixed non-compact form, Killing signature ),
(Split form, Killing signature ),
(Lorentz algebra; simple over , Killing signature ).
Proof. Let be a six-dimensional real semisimple Lie algebra. Over , its complexification decomposes into simple ideals: . Real forms correspond to conjugacy classes of antilinear involutions .
Case 1: Decomposable over . The involution preserves each factor. The algebra splits as , where each is a three-dimensional real simple Lie algebra. Over , the only three-dimensional simple Lie algebras are the compact form and the split form . Their binary combinations generate:
: Killing form is strictly negative-definite with signature .
: Killing signature is , with real rank 1.
: Maximal split algebra, Killing signature , real rank 2.
Case 2: Simple over (Realification). The involution interchanges the two factors. Then is the realification of a complex simple Lie algebra of complex dimension 3. The unique choice is , regarded as a 6-dimensional real algebra , the Lorentz algebra. Its Killing form has signature .
Because the Sylvester signatures , , , and are distinct, these four algebras are pairwise non-isomorphic.
Explicit Matrix Realization of the Real Simple Algebra
The algebra consists of trace-free complex matrices regarded as a 6-dimensional Lie algebra over . To obtain a completely explicit, coordinate-free real matrix representation, we use the canonical embedding given by .
Applying this blockwise to embeds faithfully into :
A canonical real basis satisfying the real Lorentz commutation relations is given by:
This provides a manifest real faithful representation, proving directly that .
The Non-Solvable Sector with Levi Decomposition ()
When and , the Levi factor is or , and is a three-dimensional solvable Lie algebra.
Proposition 5.1 (Representation-Theoretic Obstruction). There exists no indecomposable real Lie algebra of the form when is non-abelian.
Proof. Suppose is non-abelian. Then the derived ideal satisfies . Because is a characteristic ideal, it is invariant under all derivations, so .
If , the center or derived line forms a 1-dimensional representation . Because , every 1-dimensional character vanishes: . The induced quotient representation on defines a homomorphism . But has no non-trivial real 2-dimensional representations (any compact subalgebra of is isomorphic to ). Hence, . By Weyl’s complete reducibility theorem, , forcing to decompose.
If , is abelian. The action on is a 2-dimensional real representation of , which is identically zero. The quotient action on is also trivial. Thus .
Therefore, can couple indecomposably only with the abelian radical via its 3-dimensional vector representation 3.
Theorem 5.2 (Classification of Levi-Decomposable 6D Algebras). Every six-dimensional indecomposable real Lie algebra admitting a non-trivial Levi decomposition is isomorphic to exactly one of the following three Lie algebras (Turkowski, [13]):
(Euclidean motion algebra ):
(Adjoint action on abelian radical):
(Central extension of affine symplectic algebra):
Proof. For , Proposition 5.1 proves that is the unique indecomposable coupling. For , real representations of dimension three are either the adjoint representation or . The adjoint representation preserves an indefinite metric on , producing . The fundamental representation acts on the two-dimensional quotient of the Heisenberg algebra , preserving the symplectic area form on the center. This produces .
The three algebras are distinguished by their Levi factor ( for vs. for ) and radical nilpotency step ( for vs. for ).
The Nilpotent Sector (n= 6)
When , the algebra is entirely nilpotent ().
Theorem 6.1 (The 24 Indecomposable 6D Nilpotent Real Lie Algebras). Every indecomposable six-dimensional real nilpotent Lie algebra is isomorphic to exactly one of the twenty-four Lie algebras cataloged below:
Filiform Algebras (Nilpotency Step 5):
: , , , .
: , , , , .
: , , , , .
Nilpotency Step 4:
: , , , , .
: , , , .
: , , , , .
: , , , .
: , , , .
: , , , .
: , , , , .
: , , , .
: , , , .
Nilpotency Step 3
: , , , .
: , , , .
: , , , .
: , , , .
: , , .
: , , , .
: , , , .
: , , .
Nilpotency Step 2
: , , , .
(, Free 2-step): , , .
: , , .
: , , .
Outline of Cohomological Derivation. The Skjelbred–Sund method (T3) constructs every nilpotent algebra as a central extension of a lower-dimensional algebra by , where .
If , is a 5-dimensional nilpotent Lie algebra. By the classification of 5D nilpotent algebras, there are 9 classes for . For each , one computes and the orbits of lines under . Cocycles whose kernel contains non-trivial central elements of are discarded (they produce decomposable extensions). This yields classes through , , , , , , , , and .
If , has dimension 4 (, , or ). One classifies 2-dimensional planes in under , generating , , , and .
If , . The cohomology space is . The only 3-dimensional subspace in is the entire space itself. This generates the unique free nilpotent algebra .
The real forms and are distinguished by the signature of the quadratic Casimir form on , which is for and for .
Solvable Non-Nilpotent Algebras: Nilradical Stratification
By Mostow’s bound (T2), any solvable six-dimensional Lie algebra satisfies . We classify each stratum using the Mubarakzyanov derivation framework (T4).
Codimension-One Nilradical (n= 5): Real Sector
Here, , where is a non-nilpotent outer derivation. When the spectrum of is entirely real, the Jordan canonical forms yield:
Diagonal: , ,
Jordan Blocks: , .
When is non-abelian (, , , ), is restricted to derivations preserving the lower central series flags.
Exact Matrix Representatives of the Solvable Complex Classes (n= 5)
A fundamental gap in coordinate presentations occurs when derivations possess complex conjugate eigenvalues. We resolve this by providing unambiguous, explicit matrix representatives.
Let . In , every pair of complex conjugate eigenvalues with corresponds to the real rotation-dilation block:
Modulo the Mubarakzyanov gauge equivalence (, ), there exist exactly four mutually exclusive canonical matrix families with complex spectra:
Family g6, C1solv: Single Complex Pair, Semisimple Real Part
The derivation consists of a single rotation-dilation block and three diagonal real eigenvalues:
By scaling , we normalize if . The commutation relations are:
The algebra is unimodular if and only if .
Family g6, C2solv: Single Complex Pair with Nilpotent Real Jordan Block
The derivation couples the rotation-dilation block with a non-trivial real Jordan cell of size 2 or 3:
The commutation relations are:
Family g6, C3solv: Double Independent Complex Pairs
The derivation contains two independent rotation-dilation blocks and one real eigenvalue:
The commutation relations are:
Unimodularity holds when .
Family g6, C4solv: Generalized Complex Jordan Cell of Size 4
When the complex conjugate eigenvalues have algebraic multiplicity 2 but geometric multiplicity 1, contains a indecomposable real Jordan block:
The commutation relations are:
Faithful Linear Representation in gl(6,R)
For each of the complex classes above, an exact faithful representation is constructed via the affine action on :
Because and is strictly injective, this directly demonstrates that for all codimension-one complex solvable architectures.
Codimension-Two Nilradical (n= 4): Complex Sectors
Here, , spanned by with . The operators and form a two-dimensional abelian subalgebra of containing no non-zero nilpotent operators.
When , the commuting pair can admit complex eigenvalues. The maximal complex toral family is generated by simultaneous rotation blocks:
where is identically satisfied since rotation blocks commute. This family admits a faithful representation in via:
Codimension-Three Nilradical (n= 3)
When , Mostow’s bound is saturated: . The extension algebra acts as a three-dimensional abelian subalgebra of .
Theorem 7.1. Every six-dimensional solvable Lie algebra with has an abelian nilradical , and is isomorphic to the maximal split Cartan extension:
Proof. If , . By Schur’s bound on commutative matrix subalgebras, the maximal dimension of an abelian subalgebra of is . Thus, cannot contain an injective image of the 3-dimensional abelian algebra .
Hence, must be abelian (), so . The maximal abelian subalgebra containing no nilpotent elements is the 3-dimensional split Cartan subalgebra of diagonal matrices .
Master Tabulation and Literature Correspondence
Table 1 presents the comprehensive taxonomy of six-dimensional real Lie algebras, including the explicit complex classes.
| Algebra | Non-Zero Commutation Relations / Matrix Normal Forms | Moduli / Parameters | Unimodular? | |
| I. Semisimple Sector () | ||||
| 0 | Compact form | Yes | ||
| 0 | Mixed form | Yes | ||
| 0 | Split form | Yes | ||
| 0 | (Lorentz algebra, explicit matrices in Section 4.1) | Real Simple | Yes | |
| II. Non-Solvable Levi Decompositions () | ||||
| 3 | (Euclidean , Theorem 5.2) | Discrete | Yes | |
| 3 | (Adjoint action, Theorem 5.2) | Discrete | Yes | |
| 3 | (Symplectic central extension, Theorem 5.2) | Discrete | Yes | |
| III. Nilpotent Sector (, Representative Indecomposable Classes) | ||||
| 6 | Filiform (step 5) | Yes | ||
| 6 | Filiform (step 5) | Yes | ||
| 6 | Filiform (step 5) | Yes | ||
| 6 | Free 2-step () | Yes | ||
| 6 | Real forms (step 3) | Yes | ||
| 6 | Step 2, | Yes | ||
| IV. Solvable: Codimension-One Nilradical () – Real Spectrum | ||||
| 5 | ||||
| 5 | Jordan block | |||
| 5 | Heisenberg | |||
| V. Solvable: Codimension-One Nilradical () – Complex Spectrum (Section 7.2) | ||||
| 5 | ||||
| 5 | ||||
| 5 | , double rotation blocks | |||
| 5 | , generalized complex cell | |||
| VI. Solvable: Codimension-Two and Three Nilradicals () | ||||
| 4 | ||||
| 4 | Commuting complex pair | |||
| 4 | Outer torus of | No | ||
| 3 | Maximal split Cartan on | No |
Table 1. Master classification of representative six-dimensional real Lie algebras.
Minimal Faithful Linear Representations ((g))
Determining the minimal faithful representation dimension under Ado’s theorem is critical for constructing explicit matrix models.
Theorem 9.1 (Minimal Representation Dimension for 6D Lie Algebras). The minimal faithful representation dimension for six-dimensional real Lie algebras satisfies . Specifically:
for and .
for the Euclidean motion algebra (via affine action on ).
for the free two-step nilpotent algebra .
for the complex solvable classes through via the affine action on .
for the filiform nilpotent algebras and .
for generic solvable algebras with nilradical lacking faithful 6D representations.
Proof. We construct the minimal embeddings across sectors:
Semisimple Sector: The Lorentz algebra acts faithfully on Minkowski spacetime , yielding . Section 4.1 gives this representation explicitly in . The split form acts on , so . For , the minimal faithful real representation is its 4D defining vector representation on (hence over ).
Euclidean Algebra : Spanned by rotations and translations on Euclidean space . Embedding into the affine algebra yields:
Because and no 3D matrix algebra contains a 6D Lie subalgebra, .
Complex Solvable Sector: By Section 7.2, the affine map embeds into injectively, ensuring .
Free Nilpotent Algebra : Generated by three vectors and their three brackets. In , the maximal dimension of a strictly upper-triangular nilpotent subalgebra is . However, has derived length 3, whereas is step 2 with a 3-dimensional center. A faithful representation requires independent matrix entries for the generators and commutators, which forces .
Filiform Nilpotent Algebras –: Nilpotency step 5 implies that the adjoint operator has nilpotency index 5. In , the maximal nilpotency index of any matrix is . Thus, . The shift matrix realizes this step, proving .
This establishes the bounds.
Algorithmic Identification of Arbitrary 6D Lie Algebras
Given an arbitrary six-dimensional Lie algebra presented via its structure constants tensor with , we formulate a deterministic, coordinate-free decision procedure in Algorithm 1.
Algorithm 1 Deterministic Classification of 6-Dimensional Real Lie Algebras
Require: Structure constants satisfying and Jacobi identity.
Ensure: Canonical isomorphism class and continuous moduli parameters.
Compute derived subspaces , and Killing matrix .
Compute the solvable radical (maximal solvable ideal).
if then ▷ Sector I: Semisimple
Compute Sylvester signature of Killing form .
if then return
else if then return
else if then return
else if then return (Matrix basis in Section 4.1)
end if
end if
if then ▷ Sector II: Non-Solvable Levi Decomposition
Identify Levi factor ().
if Killing form on is negative-definite then
14: return ()
15: else ▷
16: if then return
17: else return
18: end if
19: end if
20: end if
21: Determine the nilradical (maximal nilpotent ideal).
22: if then ▷ Sector III: Nilpotent
23: Compute nilpotency step , center dimension , and derived dimension .
24: if then ▷ Filiform varieties
25: if then return
26: else if then return
27: else return
28: end if
29: else if then
30: if and then return (Free 2-step )
31: else if then return Match between and
32: else if then return
33: end if
34: else if then
35: Check quadratic form signature on :
36: if Signature is then return
37: else if Signature is then return
38: else Match remaining classes , , , , , ,
39: end if
40: else if then
41: Match classes –, , , , via upper central series flags.
42: end if
43: end if
44: if then ▷ Sector IV: Codimension-One Nilradical
45: Pick such that is non-nilpotent.
46: Compute eigenvalues of :
47: if Eigenvalues contain complex roots () then ▷ Complex Sector
48: if Two distinct pairs of complex roots then
49: return with moduli
50: else if One pair of complex roots with then
51: return (Generalized complex Jordan cell)
52: else if One pair of complex roots and a real nilpotency block then
53: return
54: else
55: return (Semisimple complex pair)
56: end if
57: else ▷ Real Sector
58: Compute real Jordan form of modulo and extract continuous parameters.
59: return Corresponding solvable class .
60: end if
61: end if
62: if then ▷ Sector V: Codimension-Two Nilradical
63: Pick complementary generators ; diagonalize commuting pair in .
64: if span a commuting complex rotation algebra then
65: return $\mathfrak{g}^{\mathrm{solv}}_{6,C_5}$
66: else
67: return Corresponding real 2-parameter family.
68: end if
69: end if
70: if dim $\mathfrak{n}$ = 3 then ▷ Sector VI: Codimension-Three Nilradical
71: return Maximal split Cartan extension $\mathbb{R}^3 \ltimes \mathbb{R}^3$.
72: end if
You can find the Python code for this algorithm and examples of its execution as a Colab Notebook.
Geometric Applications and Theoretical Physics
The classification of six-dimensional real Lie algebras provides the mathematical foundation for several branches of geometry and field theory.
Complex and Symplectic Nilmanifolds
A compact six-dimensional nilmanifold is a quotient space , where is a simply connected nilpotent Lie group and is a co-compact lattice. By Malcev’s criterion, exists if and only if has rational structure constants.
Complex Structures: An invariant complex structure on is an endomorphism with satisfying the Nijenhuis vanishing condition. Salamon [10] proved that among the twenty-four nilpotent algebras, exactly eighteen admit invariant complex structures. The filiform algebras and admit no complex structures due to cohomological obstructions in the Frölicher spectral sequence.
Symplectic Structures: An invariant symplectic structure is a closed non-degenerate 2-form . The free algebra admits no symplectic structure because every closed 2-form is degenerate on its 3-dimensional center.
Kähler Defect: By Benson and Gordon’s theorem, the only six-dimensional nilmanifold that admits a Kähler metric is the flat six-torus . None of the twenty-four indecomposable classes can support a Kähler metric.
Supergravity Flux Compactifications
In Type II superstring theory, compactifying ten-dimensional spacetime on a six-dimensional solvmanifold or nilmanifold yields an effective four-dimensional gauged supergravity. The structure constants of the six-dimensional Lie algebra define geometric fluxes:
The Jacobi identity represents the tadpole cancellation condition for gauge fluxes. Applying -duality along the cycles of a six-dimensional nilmanifold maps geometric fluxes into non-geometric - and -fluxes, dynamically traversing the strata of our classification.
Conclusion
The structural classification of six-dimensional real Lie algebras resolves the primary phase boundary of finite-dimensional Lie theory. In this monograph, we have:
Established that the semisimple sector contains exactly four real forms, categorized by their Killing signatures, with an explicit real matrix realization of .
Proved that the compact algebra cannot couple indecomposably with non-abelian three-dimensional radicals, isolating , , as the unique indecomposable non-solvable architectures.
Classified the twenty-four indecomposable nilpotent Lie algebras using the Skjelbred–Sund cohomological central extension method.
Resolved the complex Jordan strata within solvable architectures by providing exact matrix normal forms and faithful affine representations for classes through and .
Evaluated the minimal faithful representation dimension under Ado’s theorem and synthesized the entire framework into a deterministic identification algorithm.
References
- [2]J. E. Humphreys. Introduction to Lie Algebras and Representation Theory. Graduate Texts in Mathematics, Vol. 9. Springer-Verlag, New York, 1972.DOI
- [3]N. Jacobson. Lie Algebras. Dover Publications, New York, 1979.
- [4]A. W. Knapp. Lie Groups Beyond an Introduction. Progress in Mathematics, Vol. 140. Birkhäuser, Boston, 2nd edition, 2002.
- [5]J. Milnor. Curvatures of left invariant metrics on Lie groups. Advances in Mathematics, 21(3):293–329, 1976.DOI
- [6]V. V. Morozov. Classification of nilpotent Lie algebras of sixth order. Izvestiya Vysshikh Uchebnykh Zavedenii. Matematika, 4(5):161–171, 1958.
- [7]G. M. Mubarakzyanov. Classification of solvable Lie algebras of dimension four. Izvestiya Vysshikh Uchebnykh Zavedenii. Matematika, 32:114–123, 1963.
- [8]G. M. Mubarakzyanov. Classification of real Lie algebras of dimension five. Izvestiya Vysshikh Uchebnykh Zavedenii. Matematika, 53(3):99–105, 1966.
- [9]J. Patera, R. T. Sharp, P. Winternitz, and H. Zassenhaus. Invariants of real low-dimensional Lie algebras. Journal of Mathematical Physics, 17(6):986–994, 1976.
- [10]S. M. Salamon. Complex structures on nilpotent Lie algebras. Journal of Pure and Applied Algebra, 157(2-3):311–333, 2001.DOI
- [11]T. Skjelbred and T. Sund. Sur la classification des algèbres de Lie nilpotentes. Comptes Rendus de l’Académie des Sciences, Série A-B, 286(5):241–242, 1978.
- [13]P. Turkowski. Low-dimensional real Lie algebras with non-trivial Levi decomposition. Linear Algebra and its Applications, 171:197–212, 1988.