Skew-Hermitian Matrix: Definition, Rules & Examples

What Is A Skew-Hermitian Matrix?

A skew-Hermitian matrix is a square matrix with complex entries whose conjugate transpose is equal to the negative of the matrix itself. In symbols, $A$ is skew-Hermitian when $A^{} = -A$, where$A^{}$ means "transpose the matrix, then replace every entry by its complex conjugate."

Written entry by entry, the rule says $a_{ij} = -\overline{a_{ji}}$ for every pair of positions. The bar denotes the complex conjugate, so $\overline{2+3i} = 2-3i$. Only a square matrix can be skew-Hermitian, because the conjugate transpose has to have the same order as the original for the comparison to make sense.

The name splits into two halves. "Skew" signals the minus sign, the same way a real skew-symmetric matrix satisfies $A^{T} = -A$. "Hermitian" signals that we use the conjugate transpose, not the plain transpose, because the entries are complex numbers.

How Do You Check If A Matrix Is Skew-Hermitian?

To test a matrix, build its conjugate transpose in two steps, then compare with $-A$. First take the transpose, swapping rows and columns. Then conjugate every entry, flipping the sign of each imaginary part.

Take the matrix

$$A = \begin{bmatrix} i & 2+i \ -2+i & 3i \end{bmatrix}$$

Step 1, transpose it (rows become columns):

$$A^{T} = \begin{bmatrix} i & -2+i \ 2+i & 3i \end{bmatrix}$$

Step 2, conjugate every entry to get $A^{*}$:

$$A^{*} = \begin{bmatrix} -i & -2-i \ 2-i & -3i \end{bmatrix}$$

Now compare with $-A$:

$$-A = \begin{bmatrix} -i & -2-i \ 2-i & -3i \end{bmatrix}$$

The two matrices match, so $A^{*} = -A$ and $A$ is skew-Hermitian.

What Does A Skew-Hermitian Matrix Look Like?

Every skew-Hermitian matrix follows one visible pattern. The diagonal entries are purely imaginary or zero, and each off-diagonal entry is the negative conjugate of the entry in its mirror position across the main diagonal.

A general $2 \times 2$ skew-Hermitian matrix has the form

$$A = \begin{bmatrix} ip & q+ri \ -q+ri & is \end{bmatrix}$$

where $p$, $q$, $r$, and $s$ are real numbers. Notice the diagonal holds $ip$ and $is$, both purely imaginary. The off-diagonal pair $q+ri$ and $-q+ri$ are negative conjugates of each other, since $-\overline{q+ri} = -(q-ri) = -q+ri$.

The diagonal rule is the fastest sanity check. If any diagonal entry has a non-zero real part, such as $2$ or $3+i$, the matrix cannot be skew-Hermitian, no matter what the off-diagonal entries do.

What Are The Properties Of A Skew-Hermitian Matrix?

Skew-Hermitian matrices carry a tidy set of properties, and most follow directly from the condition $A^{*} = -A$.

• Diagonal entries: always purely imaginary or zero, because $a_{ii} = -\overline{a_{ii}}$ forces the real part to vanish.

• Trace: purely imaginary or zero, since the trace is the sum of purely imaginary diagonal entries.

• Closed under addition: if $A$ and $B$ are skew-Hermitian, so is $A + B$.

• Real scalar multiples: if $A$ is skew-Hermitian and $k$ is a real number, then $kA$ is skew-Hermitian.

• Link to Hermitian: $A$ is skew-Hermitian exactly when $iA$ is Hermitian. Multiplying by $i$ turns one type into the other.

• Normal matrix: every skew-Hermitian matrix commutes with its own conjugate transpose, so it is a normal matrix and can be diagonalised by a unitary matrix.

One caution about scalars. Multiplying by a real number keeps a matrix skew-Hermitian, but multiplying by a general complex number does not. That is why the "real scalar" wording matters.

Why Are The Eigenvalues Of A Skew-Hermitian Matrix Purely Imaginary?

The eigenvalues of a skew-Hermitian matrix are always purely imaginary or zero. This is the single most quoted fact about them, and there is a short reason behind it.

Suppose $A\mathbf{x} = \lambda \mathbf{x}$ for a non-zero vector $\mathbf{x}$. Taking conjugate transposes and using $A^{} = -A$leads to$\overline{\lambda} = -\lambda$.

A number equal to the negative of its own conjugate must have zero real part, so $\lambda$ is purely imaginary or zero. The same algebra explains why a Hermitian matrix, where $A^{} = A$, has real eigenvalues instead.

A clean example is the real matrix

$$A = \begin{bmatrix} 0 & 1 \ -1 & 0 \end{bmatrix}$$

which is skew-symmetric, and therefore skew-Hermitian, since it has real entries. Its eigenvalues are $\lambda = i$ and $\lambda = -i$, both purely imaginary, and each has modulus $1$. That pair of eigenvalues is exactly why this matrix rotates vectors in the plane.

How Is A Skew-Hermitian Matrix Different From Hermitian And Skew-Symmetric Matrices?

Three matrix types sit close together, and mixing them up is the most common source of error. The difference is only in two choices: plain transpose or conjugate transpose, and a plus sign or a minus sign.

Table: How skew-Hermitian, Hermitian, and skew-symmetric matrices compare.

Read the second column top to bottom. Hermitian and skew-Hermitian both use the conjugate transpose $A^{*}$; skew-symmetric and symmetric use the plain transpose $A^{T}$. A real skew-symmetric matrix is a special case of skew-Hermitian, because conjugation does nothing to real numbers. The reverse is not true, since a skew-Hermitian matrix can hold genuinely complex entries.

How Do You Split Any Matrix Into Hermitian And Skew-Hermitian Parts?

Any square complex matrix can be written as a Hermitian part plus a skew-Hermitian part, in exactly one way. The formulas are

$$A = \underbrace{\tfrac{1}{2}\left(A + A^{*}\right)}*{\text{Hermitian part}} + \underbrace{\tfrac{1}{2}\left(A - A^{*}\right)}*{\text{skew-Hermitian part}}$$

The first bracket is unchanged by the conjugate transpose, so it is Hermitian. The second bracket flips sign under the conjugate transpose, so it is skew-Hermitian. This mirrors how any ordinary number splits into a real part plus an imaginary part, which is a large part of why the two matrix types matter together.

Example 1: Find the skew-Hermitian part of $A = \begin{bmatrix} 2 & 3+i \ 1 & 4 \end{bmatrix}$.

First the conjugate transpose:

$$A^{*} = \begin{bmatrix} 2 & 1 \ 3-i & 4 \end{bmatrix}$$

Then subtract and halve:

$$\tfrac{1}{2}(A - A^{*}) = \tfrac{1}{2}\begin{bmatrix} 0 & 2+i \ -2+i & 0 \end{bmatrix} = \begin{bmatrix} 0 & 1+\tfrac{1}{2}i \ -1+\tfrac{1}{2}i & 0 \end{bmatrix}$$

Final answer: the skew-Hermitian part is $\begin{bmatrix} 0 & 1+\tfrac{1}{2}i \ -1+\tfrac{1}{2}i & 0 \end{bmatrix}$, and you can confirm its diagonal is zero, as every skew-Hermitian matrix demands.

Why Do Skew-Hermitian Matrices Matter?

Skew-Hermitian matrices are not a textbook curiosity invented to add a minus sign. They exist because two ideas that look unrelated turn out to share one structure.

• They are the matrix version of imaginary numbers. A single imaginary number $bi$ satisfies $\overline{bi} = -bi$. A skew-Hermitian matrix satisfies the exact same relation, $A^{*} = -A$, one dimension up. Just as imaginary numbers describe rotation in the plane, skew-Hermitian matrices describe rotation in higher-dimensional and complex spaces.

• They generate rotations and quantum evolution. In physics, a Hermitian matrix stands for a measurable quantity, and multiplying it by $i$ produces a skew-Hermitian matrix that drives how a system changes over time. The rule $A$ is skew-Hermitian exactly when $iA$ is Hermitian is the mathematical hinge of that relationship.

That second point is the deep reason the type is worth naming. Skew-Hermitian matrices are the generators of unitary transformations, the length-preserving motions of complex space, in the same way imaginary numbers generate ordinary rotations. The minus sign is not decoration. It is what keeps those motions from stretching or shrinking the space they act on.

Who Discovered The Skew-Hermitian Matrix?

The idea is named for the French mathematician who first studied the conjugate-transpose symmetry that both Hermitian and skew-Hermitian matrices share.

Two names anchor the modern picture:

• Charles Hermite (1822 to 1901, France) studied matrices equal to their own conjugate transpose, the Hermitian matrices, and the skew-Hermitian type is the natural mirror image of his condition.

• Arthur Cayley (1821 to 1895, England) built the general algebra of matrices in the 1850s, the framework in which conditions like $A^{*} = -A$ can even be written down.

Where Are Skew-Hermitian Matrices Used In The Real World?

The same conjugate-transpose symmetry shows up wherever complex numbers meet motion or measurement.

• Quantum mechanics: the operator that evolves a quantum state in time is built from a Hermitian energy matrix multiplied by $i$, which makes it skew-Hermitian, so the state rotates without losing total probability.

• Computer graphics and robotics: the generators of 3D rotations are real skew-symmetric matrices, the real-entry case of skew-Hermitian, used to describe angular velocity and to interpolate smooth turns.

• Signal processing: covariance and correlation structures for complex-valued signals split into Hermitian and skew-Hermitian parts, separating the symmetric information from the rotational part.

• Numerical linear algebra: the skew-Hermitian part of a matrix appears in stability analysis and in iterative solvers that treat the symmetric and antisymmetric halves of a system separately.

One condition, $A^{*} = -A$, quietly runs quantum evolution, robot motion, and signal analysis. Mathematics keeps reusing the same idea across fields that seem to have nothing in common.

What Are The Most Common Skew-Hermitian Matrix Mistakes?

These three errors account for most lost marks on this topic, and each traces back to the definition itself. They were confirmed against the CK-12 and Wikipedia definitions and the GeeksforGeeks worked examples.

Using the plain transpose instead of the conjugate transpose.

Where it slips in:

A student checks whether $A^{T} = -A$ and forgets to conjugate the entries, treating the complex matrix as if it were real.

Don't do this:

Do not stop at transposing. For complex entries, transpose alone is the wrong test.

The correct way:

Transpose the matrix, then replace every entry by its complex conjugate to form $A^{*}$, and only then compare with $-A$. The condition is $a_{ij} = -\overline{a_{ji}}$, with the bar included.

Allowing any value on the diagonal.

Where it slips in:

A student writes a diagonal entry like $2$ or $3+i$ and still calls the matrix skew-Hermitian because the off-diagonal pattern looks right.

Don't do this:

Do not accept a diagonal entry with a non-zero real part. That alone breaks the condition.

The correct way:

Check that every diagonal entry is purely imaginary or zero, since $a_{ii} = -\overline{a_{ii}}$ forces the real part to be $0$. Values like $i$, $-4i$, and $0$ are fine; $2$ and $3+i$ are not.

Confusing skew-Hermitian with skew-symmetric.

Where it slips in:

A student applies the real skew-symmetric rule $A^{T} = -A$ to a matrix with complex entries, or assumes the diagonal must be all zeros.

Don't do this:

Do not force the skew-symmetric diagonal-of-zeros rule onto the complex case. A skew-Hermitian diagonal can hold purely imaginary values, not just $0$.

The correct way:

Use $A^{*} = -A$ for complex matrices and $A^{T} = -A$ for real ones. Remember that a real skew-symmetric matrix is one special case of skew-Hermitian, not the whole story.

Practice Problems On Skew-Hermitian Matrices

Try each one, then check the answer that follows.

1. Is $A = \begin{bmatrix} 0 & 2 \ -2 & 0 \end{bmatrix}$ skew-Hermitian?
   (Answer: yes. It is real and skew-symmetric, so $A^{} = A^{T} = -A$.)*

2. Is $A = \begin{bmatrix} 2i & 1 \ -1 & 5 \end{bmatrix}$ skew-Hermitian?
   (Answer: no. The entry $5$ on the diagonal has a non-zero real part.)

3. Find the value of $z$ that makes $A = \begin{bmatrix} 0 & z \ 3+i & 0 \end{bmatrix}$ skew-Hermitian.
   (Answer: $z = -\overline{3+i} = -3+i$.)

4. What kind of matrix is $iA$ when $A$ is skew-Hermitian?
   (Answer: Hermitian, since $(iA)^{} = -i,A^{} = -i(-A) = iA$.)

5. Find the skew-Hermitian part of $A = \begin{bmatrix} 1 & 2i \ 0 & 3 \end{bmatrix}$.
   (Answer: $\tfrac{1}{2}(A - A^{}) = \begin{bmatrix} 0 & i \ i & 0 \end{bmatrix}$.)*

6. What are the eigenvalues of $A = \begin{bmatrix} 0 & 2 \ -2 & 0 \end{bmatrix}$?
   (Answer: $\lambda = 2i$ and $\lambda = -2i$, both purely imaginary.)

Where Should You Go Next After Skew-Hermitian Matrices?

Skew-Hermitian matrices sit at the meeting point of complex numbers and linear algebra, and a few natural doors open from here.

1. Hermitian matrix. The mirror-image type, where $A^{*} = A$, real eigenvalues, and the workhorse of quantum measurement.

2. Eigenvalues. Go deeper on why the imaginary-axis result holds and how eigenvalues are found in general.

3. Matrices and determinants. The broader toolkit that these special matrices live inside.

If your child is building toward complex linear algebra, a live Bhanzu trainer teaches these matrix types starting from the "why" (the imaginary-number analogy and the transformations they generate) in the Bhanzu algebra program.