What Is a Triangular Matrix?
A triangular matrix is a square matrix in which all the entries either above or below the main diagonal are zero. The main diagonal runs from the top-left to the bottom-right — the entries $a_{11}, a_{22}, a_{33}, \dots$ where the row and column numbers match.
There are two kinds, named for where the non-zero entries live:
An upper triangular matrix has all zeros below the main diagonal: $a_{ij} = 0$ whenever $i > j$.
A lower triangular matrix has all zeros above the main diagonal: $a_{ij} = 0$ whenever $i < j$.
A triangular matrix must be square, and the main-diagonal entries themselves can be anything — zero or non-zero. The name comes from the triangle of non-zero values that remains once the zeros fill one corner. A matrix that is both upper and lower triangular at once is a diagonal matrix — every off-diagonal entry is zero.
What Are the Types of Triangular Matrices?
Beyond upper and lower, two refinements describe the diagonal itself:
Strictly triangular. A strictly upper (or lower) triangular matrix has zeros on the main diagonal too — so an upper triangular matrix is strictly upper triangular when $a_{ii} = 0$ for every $i$.
Unit triangular. A unit triangular matrix has ones all along the main diagonal ($a_{ii} = 1$), with the usual triangle of zeros on one side. Unit triangular matrices appear constantly in LU decomposition, where a matrix is split into a lower-unit-triangular factor and an upper-triangular factor.
So a single triangular matrix can be upper or lower, and within that, strict, unit, or neither. Here is an upper triangular matrix beside a strictly upper one:
$$\begin{bmatrix} 3 & 1 & 4 \ 0 & 5 & 9 \ 0 & 0 & 2 \end{bmatrix} \quad\text{and}\quad \begin{bmatrix} 0 & 1 & 4 \ 0 & 0 & 9 \ 0 & 0 & 0 \end{bmatrix}.$$
What Are the Properties of a Triangular Matrix?
Triangular matrices behave predictably, which is what makes them useful:
Determinant = product of the diagonal. $\det A = a_{11},a_{22},\cdots,a_{nn}$ — no expansion needed.
Eigenvalues = the diagonal entries. For a triangular matrix, the eigenvalues are read straight off the main diagonal.
Transpose flips the type. The transpose of an upper triangular matrix is lower triangular, and vice versa.
Closed under multiplication and addition. The product (and sum) of two upper triangular matrices is again upper triangular; the same holds for lower.
Invertible only if no diagonal entry is zero. A triangular matrix is invertible exactly when every main-diagonal entry is non-zero (because that is when the determinant is non-zero).
Inverse stays triangular. The inverse of an invertible upper triangular matrix is itself upper triangular.
The determinant and eigenvalue shortcuts are the two properties worth remembering above all — they turn calculations that are hard for a general matrix into a single multiplication.
What Is the Determinant of a Triangular Matrix?
For any triangular matrix — upper or lower — the determinant is simply the product of the main-diagonal entries:
$$\det A = a_{11} \cdot a_{22} \cdot a_{33} \cdots a_{nn}.$$
There is no cofactor expansion, no row reduction — just multiply down the diagonal. This is one of the most useful facts in all of determinant computation, and it is the reason numerical methods reduce a matrix to triangular form before computing its determinant: once triangular, the determinant is one line of arithmetic.
Examples of Triangular Matrix
The set runs from identifying a type, through the most common diagonal-product mistake, to computing a determinant, checking invertibility, multiplying two triangular matrices, and solving a system by back-substitution.
Example 1
Classify $A = \begin{bmatrix} 2 & 0 & 0 \ 5 & 3 & 0 \ 1 & 4 & 6 \end{bmatrix}$.
Look above the main diagonal: every entry there ($a_{12}, a_{13}, a_{23}$) is zero, while entries below the diagonal are non-zero.
Final answer: $A$ is a lower triangular matrix — the zeros sit above the diagonal, and the non-zero values form the lower triangle.
Example 2
Find the determinant of $A = \begin{bmatrix} 4 & 7 & 2 \ 0 & 5 & 9 \ 0 & 0 & 3 \end{bmatrix}$.
Wrong attempt. A student starts a full cofactor expansion along the first row, computing minors and signs, and several lines of arithmetic later arrives at $60$ — correct, but the long way round, and a single slip anywhere would have broken it.
There is no need for any of that. The matrix is upper triangular, so the determinant is just the product of the diagonal — the off-diagonal entries above never contribute.
Correct. Multiply the diagonal:
$$\det A = 4 \cdot 5 \cdot 3 = 60.$$
Final answer: $\det A = 60$. For a triangular matrix, reaching for cofactor expansion is wasted effort — the diagonal product gives it instantly.
Example 3
Is $A = \begin{bmatrix} 1 & 8 \ 0 & 0 \end{bmatrix}$ invertible?
A triangular matrix is invertible only when no diagonal entry is zero. Here $a_{22} = 0$, so the determinant is $1 \cdot 0 = 0$.
Final answer: not invertible. A single zero on the diagonal of a triangular matrix is enough to make it singular.
Example 4
State the eigenvalues of $A = \begin{bmatrix} 7 & 2 & 1 \ 0 & -3 & 5 \ 0 & 0 & 4 \end{bmatrix}$.
For a triangular matrix, the eigenvalues are exactly the main-diagonal entries — no characteristic polynomial needed.
Final answer: the eigenvalues are $7,\ -3,\ 4$. This is why triangularising a matrix is the goal of so many eigenvalue algorithms.
Example 5
Multiply the upper triangular matrices $A = \begin{bmatrix} 2 & 1 \ 0 & 3 \end{bmatrix}$ and $B = \begin{bmatrix} 1 & 4 \ 0 & 5 \end{bmatrix}$.
Carry out the multiplication entry by entry:
$$AB = \begin{bmatrix} 2 & 1 \ 0 & 3 \end{bmatrix}\begin{bmatrix} 1 & 4 \ 0 & 5 \end{bmatrix} = \begin{bmatrix} 2 & 13 \ 0 & 15 \end{bmatrix}.$$
Final answer: $AB = \begin{bmatrix} 2 & 13 \ 0 & 15 \end{bmatrix}$ — still upper triangular, and its diagonal $(2, 15)$ is the product of the two original diagonals. Triangular times triangular stays triangular.
Example 6
Solve $;2x + y - z = 3,; 3y + 2z = 11,; 4z = 8;$ by back-substitution.
The coefficient matrix is upper triangular, so solve from the bottom up. From the last equation, $z = 2$. Substitute into the second: $3y + 4 = 11$, so $y = \tfrac{7}{3}$. Substitute both into the first: $2x + \tfrac{7}{3} - 2 = 3$, giving $2x = 3 - \tfrac{1}{3} = \tfrac{8}{3}$, so $x = \tfrac{4}{3}$.
Final answer: $x = \tfrac{4}{3},\ y = \tfrac{7}{3},\ z = 2$. Each step solved for one unknown — no simultaneous juggling — which is exactly the advantage a triangular system gives.
Why the Triangular Matrix Earns Its Place
"How do you solve a thousand equations without solving them all at once?"
Triangular form is the engine behind Gaussian elimination, the method named for Carl Friedrich Gauss (1777–1855, Germany), who used systematic elimination to reduce systems to a solvable staircase. The triangular shape is the destination of that elimination — once a matrix is triangular, the hard part is over.
Where triangular matrices do real work:
Solving large linear systems. LU decomposition factors a matrix into lower- and upper-triangular pieces, then solves by two rounds of substitution — the workhorse of scientific computing, from weather models to circuit simulation.
Computing determinants fast. Reduce to triangular form, multiply the diagonal — far cheaper than cofactor expansion for large matrices.
Eigenvalue algorithms. Methods like the QR algorithm iteratively push a matrix toward triangular form so the eigenvalues appear on the diagonal.
Computer graphics and statistics. The Cholesky decomposition (a triangular factorisation of symmetric matrices) underlies fast simulation and the sampling of correlated random variables.
Where Students Trip Up on the Triangular Matrix
Mistake 1: Expanding the determinant the long way
Where it slips in: Computing the determinant of a triangular matrix.
Don't do this: Launch into full cofactor expansion when the matrix is already triangular.
The correct way: Multiply the main-diagonal entries — that is the determinant. The off-diagonal values do not matter.
Mistake 2: Mixing up upper and lower
Where it slips in: Naming the type, especially under the "zeros where?" question.
Don't do this: Call a matrix "upper triangular" because its non-zero entries are in the upper-left — the name is about where the zeros are.
The correct way: Upper triangular means zeros below the diagonal ($i > j$); lower triangular means zeros above ($i < j$). The second-guesser who reasons from the non-zero entries gets it backwards — anchor on the zeros, not the values.
Mistake 3: Forgetting the diagonal can still make it singular
Where it slips in: Assuming any triangular matrix is invertible.
Don't do this: Treat the clean triangular shape as a guarantee of invertibility.
The correct way: A triangular matrix is invertible only if every diagonal entry is non-zero. One zero on the diagonal makes the determinant zero and the matrix singular. The memorizer who learned "triangular matrices are easy to invert" misses the precondition.
Key Takeaways
A triangular matrix is a square matrix with all zeros on one side of the main diagonal.
Upper triangular has zeros below the diagonal; lower triangular has zeros above it; the two are transposes.
Strictly triangular puts zeros on the diagonal too; unit triangular puts ones on the diagonal.
The determinant is the product of the diagonal entries, and the eigenvalues are the diagonal entries.
A triangular matrix is invertible only when no diagonal entry is zero.
Triangular form powers Gaussian elimination, LU and Cholesky decomposition, and fast eigenvalue algorithms.
Practice These Before Moving On
Classify $\begin{bmatrix} 1 & 2 & 3 \ 0 & 4 & 5 \ 0 & 0 & 6 \end{bmatrix}$ and find its determinant.
State the eigenvalues of $\begin{bmatrix} 5 & 0 \ 8 & -2 \end{bmatrix}$.
Solve $;3x = 9,; 2x + y = 7;$ by forward substitution (lower triangular system).
Answer to Question 1: upper triangular; $\det = 1 \cdot 4 \cdot 6 = 24$. Answer to Question 2: eigenvalues $5$ and $-2$ (the diagonal entries). Answer to Question 3: $x = 3$, then $6 + y = 7$ so $y = 1$. If Question 1 tempted you toward cofactor expansion, return to Mistake 1.
Want a live Bhanzu trainer to walk through more triangular matrix problems? Book a free demo class — online globally.
Was this article helpful?
Your feedback helps us write better content