Hermitian Matrix And Skew Hermitian Matrix

Hermitian and Skew‑Hermitian Matrices: Definitions, Properties, and Applications

A Hermitian matrix is a complex square matrix that equals its own conjugate transpose, meaning each entry (a_{ij}) satisfies (a_{ij} = \overline{a_{ji}}). This property makes Hermitian matrices the complex analogue of symmetric real matrices and guarantees that all eigenvalues are real numbers. And in contrast, a skew‑Hermitian matrix (or anti‑Hermitian matrix) meets the condition (A = -A^{}), where (A^{}) denotes the conjugate transpose. So consequently, the diagonal entries of a skew‑Hermitian matrix are purely imaginary or zero, and its eigenvalues are either zero or pure imaginary. Understanding these two families of matrices is essential for fields ranging from quantum mechanics to signal processing, because they provide natural frameworks for representing observables and transformations that preserve norm Simple, but easy to overlook..

Introduction

The study of linear algebra over the complex numbers introduces several special classes of matrices that simplify analysis and computation. Among them, Hermitian matrices and skew‑Hermitian matrices stand out for their elegant algebraic structure and powerful theoretical consequences. In practice, both types are defined by simple relationships with the conjugate transpose operation, yet they lead to dramatically different spectral behaviors. This article explores the definitions, key properties, eigenvalue characteristics, diagonalization processes, and practical applications of these matrices, offering a clear and thorough guide for students and professionals alike.

Definition and Basic Characteristics

Hermitian Matrix

A matrix (H) is Hermitian if it satisfies

[ H = H^{*} ]

where (H^{*}) is obtained by taking the transpose of (H) and then complex‑conjugating each entry. In real terms, in component form, this means (h_{ij} = \overline{h_{ji}}) for all (i, j). Real symmetric matrices are a special case because the complex conjugate of a real number is the number itself, so a real symmetric matrix is automatically Hermitian Most people skip this — try not to..

Counterintuitive, but true Most people skip this — try not to..

Key points

• The diagonal entries of a Hermitian matrix are real, because (h_{ii} = \overline{h_{ii}}) forces the entry to equal its own conjugate.
• The off‑diagonal entries occur in conjugate pairs: if (h_{ij} = a + bi), then (h_{ji} = a - bi).
• Hermitian matrices are self‑adjoint with respect to the standard inner product on (\mathbb{C}^n).

Skew‑Hermitian Matrix

A matrix (S) is skew‑Hermitian if

[ S = -S^{*} ]

Equivalently, (s_{ij} = -\overline{s_{ji}}). This condition implies that the diagonal entries satisfy (s_{ii} = -\overline{s_{ii}}), which forces each diagonal element to be either zero or a purely imaginary number. Off‑diagonal entries are also related by a negative conjugate relationship The details matter here..

Key points

• Multiplying a skew‑Hermitian matrix by the imaginary unit (i) yields a Hermitian matrix: (iS) is Hermitian because ((iS)^{} = -iS^{} = iS).
• The exponential of a skew‑Hermitian matrix, (e^{S}), is a unitary matrix, a fact exploited in the representation of rotations and phase transformations.

Fundamental Properties

Hermitian Matrices

1. Real Eigenvalues – Every eigenvalue (\lambda) of a Hermitian matrix satisfies (\lambda = \overline{\lambda}), which forces (\lambda) to be real.
2. Orthogonal Eigenvectors – Eigenvectors belonging to distinct eigenvalues are orthogonal under the standard inner product.
3. Diagonalizability – Hermitian matrices are always diagonalizable; there exists a unitary matrix (U) such勧め such that (U^{*}HU = D), where (D) is a real diagonal matrix.
4. Positive‑Definiteness – If all eigenvalues are positive, the matrix is called positive‑definite Hermitian and is crucial in optimization and statistics.
5. Trace and Determinant – The trace (sum of diagonal entries) equals the sum of eigenvalues, and the determinant equals the product of eigenvalues, both of which are real.

Skew‑Hermitian Matrices

1. Imaginary Eigenvalues – Eigenvalues of a skew‑Hermitian matrix are either zero or pure imaginary numbers.
2. Nilpotent when Real – If all entries are real, the matrix reduces to a real skew‑symmetric matrix, whose eigenvalues are zero or pure imaginary pairs (\pm i\beta).
3. Unitary Exponential – For any skew‑Hermitian (S), the matrix (U = e^{S}) satisfies (U^{*}U = I), making (U) unitary.
4. Lie Algebra Structure – The set of skew‑Hermitian matrices forms the Lie algebra (\mathfrak{su}(n)) when traceless, playing a central role in the theory of compact Lie groups.
5. Norm Preservation – Multiplying a vector by a skew‑Hermitian matrix does not change its Euclidean norm, because (|e^{S}x| = |x|) for any vector (x).

Spectral Theorem and Diagonalization

The Spectral Theorem for Hermitian matrices states that any Hermitian matrix (H) can be written as

[ H = U \Lambda U^{*}, ]

where (U) is a unitary matrix whose columns are orthonormal eigenvectors of (H), and (\Lambda) is a diagonal matrix containing the real eigenvalues of (H). This decomposition is unique up to ordering of eigenvalues and choice of eigenvectors.

For skew‑Hermitian matrices, the spectral theorem also applies but with a twist: if (S) is skew‑Hermitian, then (iS) is Hermitian, so we can write

[ iS = V \Lambda V^{*}, ]

where (\Lambda) is real diagonal. This means

[ S = -i V \Lambda V^{*}. ]

Thus the eigenvalues of (S) are (-i) times the real eigenvalues of (\Lambda), confirming the pure‑imaginary nature of the spectrum Turns out it matters..

Relationship Between the Two Classes

The two families are closely linked:

• Linear Combination – Any complex matrix (A) can be uniquely decomposed into a sum of a Hermitian part and a skew‑Hermitian part: [ A = H + S,\quad H = \frac{A + A^{}}{2},\quad S = \frac{A - A^{}}{2}. ] Here (H) is Hermitian and (S) is skew‑Hermitian.
• Commutator – The commutator of two Hermitian matrices is skew‑Hermitian: ([H_{1}, H_{2}] = H_{1}H_{2} - H_{2}H_{1}) satisfies ([H_{1}, H_{2}]^{*} = -[H_{1}, H_{2}]).
• Unitary Generation – If (U) is unitary, then (i(U - U^{*})) is skew‑Hermitian, showing how skew‑Hermitian matrices naturally arise from unitary transformations.

Examples

Example 1: Hermitian Matrix

[ H = \begin{pmatrix} 3 &

The user wants me to continue the article naturally. The article is about Hermitian and skew-Hermitian matrices, eigenvalues, determinants, spectral theorem, etc. I need to read the provided text, understand what's there, and then write the next part of the article without repeating any previous text. It ends abruptly at "3 &" in an example. I need to continue from there, providing a proper conclusion.

Not obvious, but once you see it — you'll see it everywhere The details matter here..

Let me look at the structure:

1. Spectral Theorem and Diagonalization section
2. Intro with properties of Hermitian and skew-Hermitian matrices (sum of eigenvalues, determinant)
3. Skew-Hermitian Matrices section with 5 points
4. Relationship Between the Two Classes section with 3 points

I should continue the examples section. But the user says "Finish with a proper conclusion.On the flip side, since it starts with a Hermitian matrix example, I could complete that example or add more examples. " So I need to provide content that leads to a good conclusion.

Easier said than done, but still worth knowing.

Let me think about what would be a natural continuation. The article has covered:

• Properties of Hermitian matrices (real eigenvalues, orthogonal eigenvectors, spectral theorem)
• Properties of skew-Hermitian matrices (pure imaginary eigenvalues, unitary exponential, Lie algebra)
• Relationship between them
• Examples section begins but is cut off

I should probably complete the first example, maybe add another example or two, and then write a conclusion summarizing the key points And that's really what it comes down to..

Wait, looking more carefully: the example is cut off at "H = \begin{pmatrix} 3 &". This looks like it might be starting a 2x2 matrix. Let me think... That's why actually, it could be a 3x3 matrix starting with 3 on the diagonal. But the "3 &" suggests it's continuing a matrix entry.

Not obvious, but once you see it — you'll see it everywhere.

Actually, looking at the LaTeX: \begin{pmatrix} 3 & - this is likely the start of a matrix like: [ H = \begin{pmatrix} 3 & ... \ ... & .. Simple, but easy to overlook..

I don't know exactly what the full matrix was supposed to be. But since the user said "Continue the article smoothly," I should probably either:

1. Complete the example in a reasonable way

Actually, I think the best approach is to complete the example naturally, perhaps adding a second example for skew-Hermitian, and then write a conclusion.

Let me draft:

Completing Example 1: [ H = \begin{pmatrix} 3 & 1 \ 1 & 3 \end{pmatrix} ] This is Hermitian since it equals its conjugate transpose. In practice, its eigenvalues are found by solving det(H - λI) = 0, giving λ = 3 ± 1, so λ₁ = 4, λ₂ = 2. Both are real, as expected Worth keeping that in mind..

Then maybe Example 2 for skew-Hermitian: [ S = \begin{pmatrix} 0 & -i \ i & 0 \end{pmatrix} ] This is skew-Hermitian. Its eigenvalues are ±i, pure imaginary as expected Simple as that..

Then a conclusion summarizing:

• Hermitian matrices have real eigenvalues and orthogonal eigenvectors
• Skew-Hermitian matrices have pure imaginary eigenvalues
• They form complementary decompositions
• The spectral theorem provides diagonalization for both
• These structures underpin quantum mechanics (observables vs. generators)

Let me write this out properly.

Actually, I should be careful. The user said "Do not repeat previous text.Also, " So I shouldn't quote or restate anything from before. Which means i should just continue forward. And "Finish with a proper conclusion.

So my response should be just the continuation and conclusion, no preamble.

Let me write:

Continuing the example:

[ H = \begin{pmatrix} 3 & 1 \ 1 & 3 \end{pmatrix} ]

Since (H^* = H), this matrix is Hermitian. Computing its characteristic polynomial:

[ \det(H - \lambda I) = \det\begin{pmatrix} 3-\lambda & 1 \ 1 & 3-\lambda \end{pmatrix} = (3-\lambda)^2 - 1 = \lambda^2 - 6\lambda + 8 ]

Setting this equal to zero gives (\lambda = 2) and (\lambda = 4), both real. Even so, the corresponding normalized eigenvectors are (v_1 = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 \ -1 \end{pmatrix}) for (\lambda = 2) and (v_2 = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 \ 1 \end{pmatrix}) for (\lambda = 4). Thus (H = V \Lambda V^*) with (V) unitary, confirming the spectral theorem.

For the skew‑Hermitian counterpart:

[ S = \begin{pmatrix} 0 & -i \ i & 0 \end{pmatrix} ]

One verifies (S^* = -S), so (S) is skew‑Hermitian. Its characteristic polynomial yields eigenvalues (\pm i), consistent with the theory. On top of that, (e^S = \begin{

[ H = \begin{pmatrix} 3 & 1 \ 1 & 3 \end{pmatrix} ]

Since (H^* = H), this matrix is Hermitian. Its characteristic polynomial is

[ \det(H - \lambda I) = (3-\lambda)^2 - 1 = \lambda^2 - 6\lambda + 8, ]

yielding eigenvalues (\lambda_1 = 2) and (\lambda_2 = 4), both real as guaranteed by the spectral theorem. The corresponding orthonormal eigenvectors are

[ v_1 = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 \ -1 \end{pmatrix}, \qquad v_2 = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 \ 1 \end{pmatrix}, ]

so (H = V\Lambda V^*) with (V = [v_1 ; v_2]) unitary and (\Lambda = \operatorname{diag}(2,4)) No workaround needed..

For a skew‑Hermitian example, consider

[ S = \begin{pmatrix} 0 & -i \ i & 0 \end{pmatrix}. ]

One checks (S^* = -S). Its eigenvalues are (\pm i), purely imaginary as expected. The matrix exponential (e^S) is unitary:

[ e^S = \begin{pmatrix} \cos 1 & -\sin 1 \ \sin 1 & \cos 1 \end{pmatrix}, ]

illustrating the fundamental correspondence between skew‑Hermitian matrices and the Lie algebra of the unitary group.

These two classes are not merely analogous; they are complementary pieces of a single algebraic structure. Every square matrix (A) decomposes uniquely as (A = H + S) with (H) Hermitian and (S) skew‑Hermitian:

[ H = \frac{A + A^}{2}, \qquad S = \frac{A - A^}{2}. ]

This decomposition underlies the polar decomposition (A = UP) (with (U) unitary, (P) positive semidefinite) and the Cartan decomposition of Lie algebras. On top of that, in quantum mechanics, Hermitian operators represent observables—quantities that can be measured—while skew‑Hermitian operators generate continuous symmetries via Stone’s theorem. The spectral theorem ensures that both classes are unitarily diagonalizable, making them amenable to analysis and computation Nothing fancy..

Thus, Hermitian and skew‑Hermitian matrices form the backbone of linear algebra over (\mathbb{C}): one captures the “real” geometry of quadratic forms and measurement, the other the “imaginary” geometry of rotations and dynamics. Their interplay, mediated by the adjoint operation, is the algebraic signature of complex Hilbert space.