What a determinant summarizes

The determinant assigns one number to a square matrix. A nonzero determinant means the associated linear transformation is invertible and its rows and columns are linearly independent. A zero determinant means the transformation collapses dimension: area becomes a line in 2D, or volume becomes a plane or lower-dimensional set in 3D.

The absolute determinant is an area or volume scale factor. A 3×3 determinant of 19 means a unit cube becomes a parallelepiped with volume 19. The sign records orientation: positive preserves handedness, while negative reverses it. Sign is not a statement that geometric volume is negative.

Worked 3×3 expansion

For rows [2,1,3], [0,−1,4], and [5,2,0], the first minor removes row 1 and column 1: (−1 × 0) − (4 × 2) = −8. Multiplying by entry 2 gives −16.

The second minor is (0 × 0) − (4 × 5) = −20, but its cofactor sign is negative. Thus the cofactor is +20, and multiplying by entry 1 contributes +20. The third minor is (0 × 2) − (−1 × 5) = 5, contributing 3 × 5 = 15. Total determinant is −16 + 20 + 15 = 19.

Sarrus’ rule gives a separate 3×3 audit. The three forward diagonal products sum to 20. The three backward diagonal products sum to 1. Their difference is again 19. Sarrus applies only to 3×3 matrices; it is not a general determinant algorithm.

Cofactor signs and minors

A minor Mᵢⱼ is the determinant left after deleting row i and column j. A cofactor Cᵢⱼ equals (−1)ⁱ⁺ʲMᵢⱼ. The checkerboard signs begin plus, minus, plus across the first row and alternate in every direction.

A common mistake is to write a minus before the second term and also use an already signed cofactor, subtracting twice. Choose one notation: expand with raw minors and explicit + − + signs, or add entries times signed cofactors. The result board labels each displayed factor as a cofactor term.

Expansion can use any row or column. A row with zeros usually saves work. All valid expansions produce the same determinant, making a second row or column a useful manual check.

Row-operation rules

These rules make elimination an efficient determinant method. Reduce to upper triangular form while tracking swaps and row scaling, then multiply diagonal entries. Adding row multiples during elimination needs no determinant adjustment.

Invertibility and solving systems

For a square matrix, det(A) ≠ 0 is equivalent to invertibility, full rank, unique solvability of Ax = b for every b, and zero not being an eigenvalue. These statements link the determinant to several areas of linear algebra.

Cramer’s rule expresses solution coordinates as determinant ratios, but it is inefficient and numerically unattractive for large systems. LU or QR methods are preferred computationally.

A determinant close to zero can indicate near singularity, but determinant magnitude alone is scale dependent. Multiplying a row by a large unit conversion changes the determinant even when the underlying information is unchanged. Use a condition number for sensitivity assessment.

Properties for checking work

Transposing leaves the determinant unchanged: det(Aᵀ) = det(A). Determinants multiply: det(AB) = det(A)det(B). If A is invertible, det(A⁻¹) = 1/det(A). A triangular matrix’s determinant is the product of its diagonal.

Repeated or proportional rows make the determinant zero. The same is true for repeated or proportional columns. A zero row or column is an immediate zero. These structural observations can avoid unnecessary expansion.

Adding matrices has no simple determinant addition rule. In general det(A+B) is not det(A)+det(B). Likewise, multiplying every matrix entry by k is not merely k det(A) unless the matrix is 1×1; for n×n it is kⁿdet(A).

Numerical tolerance and rounding

The exact determinant of an integer matrix can often be verified with exact arithmetic. This browser calculator uses IEEE floating-point numbers. Decimal entries, subtractive cancellation, and large magnitude differences can leave a tiny residual rather than exact zero.

The tolerance controls the status label only; it does not alter the displayed arithmetic. A determinant of 10⁻¹³ with tolerance 10⁻¹² is labeled effectively singular. Change tolerance with regard to matrix scale and the problem’s accuracy requirements.

Do not round cofactor minors before summing. The displayed terms are rounded for reading, while the final determinant uses full internal precision. Multiplying rounded outputs may differ slightly.

Scale boundary: A small determinant does not automatically mean ill conditioning, and a large determinant does not guarantee stability. Units and singular values matter.

Triangular matrices and elimination

For an upper- or lower-triangular matrix, every nonidentity permutation product contains a zero, so the determinant is simply the product of diagonal entries. This rule makes elimination efficient: use row additions to clear entries below the diagonal, record each row swap, and avoid scaling rows unless its multiplier is recorded.

Suppose elimination ends with diagonal 4, 4, and 1 after one row swap. The triangular determinant is 16, and reversing the swap sign gives −16 for the original matrix. If a diagonal entry becomes zero and no lower pivot exists, the determinant is zero.

Fraction-preserving elimination can retain exact integers, while numerical libraries use pivoting and optimized floating-point kernels. Cofactor expansion is educational for 3×3 but grows too quickly in direct recursive form for large matrices.

Products, inverses, and basis changes

The identity det(AB) = det(A)det(B) reflects successive volume scaling: B acts first, then A. Setting B = A⁻¹ gives det(A⁻¹) = 1/det(A). A valid change-of-basis matrix must have nonzero determinant because its basis vectors are independent.

Similarity transformations B = S⁻¹AS preserve determinant because det(S⁻¹) and det(S) cancel. They also preserve the characteristic polynomial and eigenvalues. This shows that determinant is not an artifact of one coordinate representation.

Multiplicativity is a useful audit but not an addition rule. There is generally no simplification from det(A+B) to the separate determinants.

Area from coordinate differences

For planar points A, B, and C, form edge vectors B−A and C−A as columns of a 2×2 matrix. The absolute determinant is parallelogram area; half is triangle area. Zero means the points are collinear. The sign indicates ordered orientation.

Edges [3,1] and [1,4] give determinant 3×4 − 1×1 = 11, so triangle area is 5.5 square units. Swapping the edges reverses sign but not physical area. Translating every point by the same vector leaves the edge differences unchanged.

Computational-geometry orientation tests use this signed value, but near-collinear floating-point coordinates may require robust predicates rather than one fixed tolerance.

Jacobian determinants and units

A Jacobian matrix contains partial derivatives of transformed coordinates with respect to original coordinates. Its determinant gives local oriented scale. A change-of-variables integral uses its absolute value because area and volume measures are nonnegative.

Units provide a check: derivative units combine multiplicatively across the determinant. Treating a Jacobian determinant as unitless without inspecting the transformation can conceal a conversion error. When it is zero, the mapping locally collapses dimension and may lack a smooth local inverse.

A sign change can mark orientation reversal or a fold. These interpretations require derivatives at a point; this calculator evaluates only the numeric matrix entered.

Applications

Determinants appear in coordinate changes, Jacobian volume adjustments, cross products, orientation tests, eigenvalue characteristic polynomials, and formulas for inverses. In geometry, a 2×2 determinant gives signed parallelogram area; half its absolute value gives triangle area from two edge vectors.

In multivariable calculus, the absolute Jacobian determinant adjusts area or volume elements under a transformation. In statistics, covariance-matrix determinants relate to generalized variance, though interpretation requires positive semidefinite structure and scale awareness. In each use, the determinant’s context supplies units and meaning.

Frequently asked questions

References

NIST JAMA — LU determinant method

NIST Digital Library of Mathematical Functions — determinants and matrices

Wolfram MathWorld — determinant identities