An augmented matrix is a matrix that represents a system of linear equations by joining the coefficient matrix to the column of constants, separated by a vertical bar. It is written $[A \mid b]$, where $A$ holds the coefficients and $b$ holds the numbers on the right of each equals sign. The bar does the job of every equals sign at once.
What Is an Augmented Matrix?
Start with a system of linear equations, equations where every variable appears only to the first power, such as: $$2x + 3y = 8$$ $$x - y = -1$$
The coefficient matrix collects only the numbers multiplying the variables: $$A = \begin{bmatrix} 2 & 3 \ 1 & -1 \end{bmatrix}$$
The constant column holds the right-hand sides: $$b = \begin{bmatrix} 8 \ -1 \end{bmatrix}$$
The augmented matrix joins them, with a bar where the equals signs were: $$[A \mid b] = \left[\begin{array}{cc|c} 2 & 3 & 8 \ 1 & -1 & -1 \end{array}\right]$$
Each row is one equation. Each column on the left is one variable. The word "augmented" simply means the coefficient matrix has been extended by one extra column, the constants.
Why the vertical line?
It stands for the equals sign. Everything left of the bar is a coefficient; the single column right of the bar is what each equation equals. The bar keeps the two roles from blurring together.
The Structure of $[A \mid b]$
Every piece of the notation maps onto one piece of the system. This key names each part.
Symbol | Meaning |
|---|---|
$[A \mid b]$ | the augmented matrix, the whole grid |
$A$ | the coefficient matrix, the numbers multiplying the variables |
$b$ | the constant column, the right-hand side of each equation |
$\mid$ (the bar) | stands in for the equals sign |
each row | one full equation |
each left column | one variable |
Read the grid this way and it is never ambiguous: rows are equations, left columns are variables, and the single column past the bar holds what each equation equals.
How Do You Build an Augmented Matrix From a System?
Line the equations up first, then read them straight into the grid.
Align the variables. Write every equation as (x-term) + (y-term) + (z-term) = constant, in the same order in every row.
Fill missing terms with 0. If an equation has no $y$, its $y$-coefficient is 0, the slot is not left blank.
Copy signs exactly. A $-y$ becomes $-1$ in the grid, not $1$.
Drop in the constant on the far side of the bar.
This turns a system built for simultaneous equations into a compact object you can manipulate by Gaussian elimination, the row-reduction procedure that solves the system.
Row Operations and What They Reveal
Solving an augmented matrix means simplifying it with elementary row operations, moves that change the numbers but never change the system's solution. There are exactly three.
Swap two rows. Written $R_i \leftrightarrow R_j$. Reordering equations changes nothing about the answer.
Multiply a row by a nonzero number. Written $R_i \to kR_i$, with $k \neq 0$. Scaling a whole equation keeps it true.
Add a multiple of one row to another. Written $R_i \to R_i + kR_j$. This is the workhorse move that clears a variable from a row.
Each operation acts on a whole row, because a row is a whole equation. That is why column operations across the bar are never allowed.
Once the matrix is reduced, its final shape tells you how many solutions the system has.
What the reduced matrix shows | What it means |
|---|---|
Every variable pinned to one value | Unique solution |
A row like $\left[\, 0 \;\; 0 \mid 5 \,\right]$, i.e. $0 = 5$ | No solution (inconsistent) |
A full row of zeros, $\left[\, 0 \;\; 0 \mid 0 \,\right]$, i.e. $0 = 0$ | Infinitely many solutions |
The no-solution case is worked in Example 6 below. The infinite case is its mirror image: instead of an impossible statement, a whole row collapses to $0 = 0$, which is always true and therefore adds no new information, leaving one variable free to take any value.
Examples of the Augmented Matrix
The examples move from a clean two-variable setup to a system with hidden zeros and a full solve.
Example 1
Write the augmented matrix for $3x + 2y = 12$ and $x + 4y = 14$.
Row 1 reads off the first equation.
Row 2 reads off the second.
$$\left[\begin{array}{cc|c} 3 & 2 & 12 \ 1 & 4 & 14 \end{array}\right]$$
Final answer: the matrix above.
Example 2
A student writes the augmented matrix for $2x - y = 5$ and $x = 3$ as a $2 \times 3$ grid but writes the second row as $\left[, 1 ;; 3 ,\right]$.
The tempting move is to copy "$x = 3$" straight across as one coefficient and one constant.
That gives a row with only two entries, which does not line up with the first row's three slots.
The equation $x = 3$ has no $y$ term, so its $y$-coefficient is 0, not missing.
Rewrite it as $1x + 0y = 3$.
Now the matrix is: $$\left[\begin{array}{cc|c} 2 & -1 & 5 \ 1 & 0 & 3 \end{array}\right]$$
Final answer: the second row is $\left[, 1 ;; 0 \mid 3 ,\right]$. A missing variable is a zero coefficient, never a blank.
Example 3
Build the augmented matrix for the three-variable system $x + y + z = 6$, $2y + 5z = -4$, $2x + 5y - z = 27$.
The second equation has no $x$ term, so its $x$-coefficient is 0.
$$\left[\begin{array}{ccc|c} 1 & 1 & 1 & 6 \ 0 & 2 & 5 & -4 \ 2 & 5 & -1 & 27 \end{array}\right]$$
Final answer: the $3 \times 4$ matrix above.
Example 4
Solve $4x + 3y = 11$ and $5x - 3y = 7$ using the augmented matrix.
Write it: $$\left[\begin{array}{cc|c} 4 & 3 & 11 \ 5 & -3 & 7 \end{array}\right]$$
Add Row 1 to Row 2 to clear $y$ in the new row: $$R_2 \to R_1 + R_2: \quad \left[\begin{array}{cc|c} 4 & 3 & 11 \ 9 & 0 & 18 \end{array}\right]$$
Row 2 now says $9x = 18$, so $x = 2$.
Substitute into Row 1: $4(2) + 3y = 11$, so $3y = 3$ and $y = 1$.
Final answer: $x = 2$, $y = 1$.
Example 5
Convert the matrix $\left[\begin{array}{cc|c} 1 & 0 & 7 \ 0 & 1 & -2 \end{array}\right]$ back into a system.
Row 1 reads $1x + 0y = 7$.
Row 2 reads $0x + 1y = -2$.
Final answer: $x = 7$, $y = -2$, a matrix already in solved form.
Example 6
How does an augmented matrix show a system has no solution?
Suppose row-reducing lands on: $$\left[\begin{array}{cc|c} 1 & 2 & 3 \ 0 & 0 & 5 \end{array}\right]$$
Row 2 reads $0x + 0y = 5$, which claims $0 = 5$.
That is impossible.
Final answer: a row of zeros left of the bar with a nonzero constant right of the bar means the system is inconsistent, no solution.
Why Mathematicians Reached for the Grid
The augmented matrix exists because rewriting the same variables over and over is wasted effort. The insight is small but powerful: once the variables are lined up, only the numbers matter.
It strips away the clutter. The letters $x, y, z$ carry no information once positions are fixed, the column already says which variable a number belongs to.
It makes the method mechanical. Row operations are pure bookkeeping, which is exactly why they can be handed to a computer. Every linear-algebra solver on a modern machine runs on this idea.
It scales. Two equations or two hundred, the procedure is identical, the grid just gets taller.
The matrix idea was formalised by Arthur Cayley in A Memoir on the Theory of Matrices (1858), which set out matrices as objects in their own right. Treating a whole system as one grid to be reduced is the direct descendant of that shift.
Where the Setup Goes Wrong
Most augmented-matrix errors happen before any arithmetic, in the setup.
Mistake 1: Leaving a missing variable blank
Where it slips in: an equation like $x = 3$ or $2y + 5z = -4$ that skips a variable.
Don't do this: write a short row with fewer entries than the other rows.
The correct way: insert a $0$ for the missing coefficient so every row has the same width. The first instinct is to copy only the terms that appear; the discipline is to place a zero wherever a variable is silent.
Mistake 2: Dropping the sign
Where it slips in: copying a subtraction into the grid.
Don't do this: write $5$ for the coefficient in $5x - 3y = 7$'s $y$ slot.
The correct way: the coefficient is $-3$. The second-guesser who rewrites $-3y$ as $3y$ "to keep it positive" changes the equation. Copy the sign exactly.
Mistake 3: Forgetting what the bar means
Where it slips in: performing a column operation across the bar.
Don't do this: treat the constant column as just another coefficient column and mix it into a variable column.
The correct way: the bar marks the equals sign, so only row operations are valid, they act on whole equations at once.
Conclusion
An augmented matrix packs a linear system into one grid $[A \mid b]$, with a bar for the equals sign.
Each row is an equation; each left column is a variable; the right column holds the constants.
Missing variables become $0$ coefficients, and signs must be copied exactly.
Row operations solve it, and a row like $0 = 5$ signals no solution.
To practise systems with a teacher, explore Bhanzu's algebra tutor, help with algebra, or online math classes.
Read More
Was this article helpful?
Your feedback helps us write better content