Eigenvalue and Eigenvector Calculator – 2×2 Matrix
Enter the four elements of a 2×2 matrix to get its eigenvalues and eigenvectors, plus the determinant, trace, and characteristic polynomial with a full worked solution.
2×2 Matrix Eigenvalues and Eigenvectors Solver
Eigenvalues and Eigenvectors Formula for 2×2 Matrix
Where:
- \( A = \begin{bmatrix} a & b \\ c & d \end{bmatrix} \) — the 2×2 matrix
- \( \lambda \) — the eigenvalues (the roots of the characteristic equation)
- \( \vec{v} \) — the corresponding eigenvectors, found by solving \( (A - \lambda I)\vec{v} = \vec{0} \)
- \( \operatorname{tr}(A) = a + d \) — the trace, and \( \det(A) = ad - bc \) — the determinant
The first equation defines the characteristic polynomial \( \lambda^2 - \operatorname{tr}(A)\lambda + \det(A) = 0 \). Solving it (a simple quadratic) gives both eigenvalues, and each eigenvalue then yields an eigenvector from the homogeneous system \( (A - \lambda I)\vec{v} = \vec{0} \).
Eigenvalues and Eigenvectors 2×2 – Calculation Example
Example: find the eigenvalues and eigenvectors of the matrix
\[ A = \begin{bmatrix} 4 & 2 \\ 1 & 3 \end{bmatrix} \]
Step 1 — trace and determinant: \( \operatorname{tr}(A) = 4 + 3 = 7 \), \( \det(A) = 4 \cdot 3 - 2 \cdot 1 = 10 \).
Step 2 — characteristic equation: \( \lambda^2 - 7\lambda + 10 = 0 \).
\[ \lambda_{1,2} = \frac{7 \pm \sqrt{49 - 40}}{2} = \frac{7 \pm 3}{2} \]
Step 3 — eigenvalues: \( \lambda_1 = 5 \), \( \lambda_2 = 2 \).
Step 4 — eigenvectors from \( (A - \lambda I)\vec{v} = \vec{0} \). For \( \lambda_1 = 5 \): \( (4-5)x_1 + 2x_2 = 0 \), so \( x_2 = 0.5x_1 \) and \( \vec{v}_1 = (1,\ 0.5) \). For \( \lambda_2 = 2 \): \( (4-2)x_1 + 2x_2 = 0 \), so \( x_2 = -x_1 \) and \( \vec{v}_2 = (1,\ -1) \).
Result: \( \lambda_1 = 5 \) with eigenvector \( (1,\ 0.5) \); \( \lambda_2 = 2 \) with eigenvector \( (1,\ -1) \). Verification: \( A\vec{v}_1 = (5,\ 2.5) = 5\vec{v}_1 \) and \( A\vec{v}_2 = (2,\ -2) = 2\vec{v}_2 \).
How to Use the Eigenvalue Calculator
- Enter the matrix element a₁₁ (row 1, column 1). The default value is 4.
- Enter a₁₂ (row 1, column 2). The default value is 2.
- Enter a₂₁ (row 2, column 1). The default value is 1.
- Enter a₂₂ (row 2, column 2). The default value is 3.
- Read the results: the determinant |A|, the trace, the matrix \( A - cI \), the characteristic polynomial, both eigenvalues, and the eigenvector for each eigenvalue.
What Are Eigenvalues and Eigenvectors?
An eigenvector of a matrix \( A \) is a nonzero vector that does not change direction when the matrix is applied to it — it is only stretched or compressed. The scaling factor of that stretch is the eigenvalue \( \lambda \). Formally, \( \vec{v} \) is an eigenvector if it satisfies
\[ A\vec{v} = \lambda\vec{v} \]
Geometrically, a 2×2 matrix transforms the plane: it rotates, stretches, and shears vectors. Most vectors end up pointing somewhere new, but up to two special directions survive the transformation unchanged (apart from scaling). Those directions are the eigenvectors, and their scaling factors are the eigenvalues. This is why eigen decomposition reveals the "essential" behavior of a transformation, stripping away everything else.
Eigenvalues encode a great deal of information in just two numbers. Their product equals the determinant (how the matrix scales areas), and their sum equals the trace. If both eigenvalues are real and positive, the transformation stretches space in both eigen-directions; if one is negative, one direction is flipped; if both have absolute value less than 1, the transformation shrinks space. When both eigenvalues are equal, the matrix may or may not have two independent eigenvectors — this is the distinction between diagonalizable and defective matrices.
Eigenvalues and eigenvectors appear everywhere in applied mathematics: stability analysis of differential equations and dynamical systems, principal component analysis in statistics, vibration modes in mechanical engineering, quantum mechanics (where eigenvalues are the measurable values of observables), and page-rank style algorithms. Where a matrix repeats its action many times — \( A^n \) — the largest eigenvalue dominates the long-term behavior.
Assumptions and Limitations
- Real eigenvalues only. The calculator solves the characteristic quadratic over the real numbers. If the discriminant \( \operatorname{tr}(A)^2 - 4\det(A) < 0 \), the eigenvalues are complex and the tool reports "No real roots". This happens, for example, for rotation matrices.
- Eigenvectors are not unique. Any nonzero multiple of an eigenvector is also an eigenvector. The calculator fixes \( x_1 = 1 \) and reports \( (1,\ x_2) \); the same direction could equally be written as \( (2,\ 2x_2) \) or with opposite sign. Only the direction matters.
- Repeated eigenvalues. When the discriminant is zero, \( \lambda_1 = \lambda_2 \) and there may be only one independent eigenvector (a defective matrix). The calculator will return the same eigenvector twice in that case.
- Zero rows. If both \( a_{12} \) and \( a_{21} \) are zero (a diagonal matrix), the eigenvalues are simply the diagonal entries \( a_{11} \) and \( a_{22} \), and the eigenvectors fall back to the standard basis directions.
- Rounding. Results are rounded to 4 decimal places, which is sufficient for most textbook problems but may hide small values in ill-conditioned matrices.
To work with larger matrices, try the Eigenvalues and Eigenvectors 3×3 Calculator. If you need other matrix operations first, the Matrix Multiplication Calculator and the Determinant Multiplication Calculator are useful companions, since eigenvalues are often computed from matrix powers and products of determinants.
Frequently Asked Questions
How do I find eigenvalues and eigenvectors of a 2×2 matrix by hand?
Compute the trace \( a + d \) and determinant \( ad - bc \), then solve the quadratic \( \lambda^2 - (a+d)\lambda + (ad-bc) = 0 \) with the quadratic formula. For each eigenvalue, plug \( \lambda \) back into \( (A - \lambda I)\vec{v} = \vec{0} \) and solve the resulting linear system for the vector. One free variable always remains, so choose a convenient value like \( x_1 = 1 \) and solve for \( x_2 \).
What does it mean if the calculator says "No real roots"?
The discriminant \( \operatorname{tr}(A)^2 - 4\det(A) \) is negative, so the eigenvalues are a complex conjugate pair. This is typical for rotation matrices and matrices that rotate the plane more than they stretch it — no real direction survives the transformation unchanged. To work with such matrices you need complex arithmetic: the eigenvalues are \( \frac{\operatorname{tr}(A) \pm i\sqrt{4\det(A) - \operatorname{tr}(A)^2}}{2} \).
Why is my eigenvector different from the textbook answer?
Eigenvectors are not unique. Any nonzero scalar multiple of an eigenvector is also a valid eigenvector, so \( (1, 0.5) \), \( (2, 1) \), and \( (-4, -2) \) all represent the same direction. Textbooks often normalize eigenvectors to unit length or clear fractions, which changes the numbers but not the answer. Only the direction (the ratio of the components) matters, so compare the ratios rather than the raw values.
What is a defective matrix, and how do I know if my matrix is one?
A matrix is defective when it has a repeated eigenvalue \( (\operatorname{tr}(A)^2 = 4\det(A)) \) but fewer than two independent eigenvectors. Check whether \( A - \lambda I \) is the zero matrix: if it is, every vector is an eigenvector and the matrix is fine (diagonalizable); if not, only one independent eigenvector exists and the matrix cannot be diagonalized. This matters for computing powers \( A^n \) and solving systems of differential equations.
What do the eigenvalues tell me about the matrix?
The product of the eigenvalues equals the determinant, and their sum equals the trace — a quick way to verify your answer. If both eigenvalues are positive, the transformation preserves orientation and stretches space; if exactly one is negative, one direction is flipped. Eigenvalues with absolute value below 1 indicate shrinking along that direction. In dynamical systems, both eigenvalues inside the unit circle mean the system is stable; the eigenvalue with the largest absolute value dominates long-term behavior.
Can an eigenvector be the zero vector?
No. By definition, \( \vec{v} \) must be nonzero, because \( A\vec{0} = \lambda\vec{0} \) holds for every \( \lambda \), making the definition meaningless. If your calculation produces only the zero vector as a solution, you likely made an arithmetic error — substituting an eigenvalue into \( (A - \lambda I) \) must make the system singular (determinant zero), which guarantees nonzero solutions exist.
How can I verify my eigenvalue and eigenvector results?
Multiply the original matrix by each eigenvector and check that the result equals the eigenvalue times that vector: \( A\vec{v} = \lambda\vec{v} \). Also confirm that the two eigenvalues add up to the trace and multiply to the determinant. A common mistake is a sign error in \( \det(A) = ad - bc \) or in solving the linear system — these checks catch both immediately.