• Foundation
  • About 15 min
  • 5 practice questions

Powers, the identity and zero matrices, and rules that fail for matrices

Compute A², A³ and general powers of a square matrix, use the identity and zero matrices correctly, and avoid the algebra rules that hold for numbers but fail for matrices, such as (A + B)² = A² + 2AB + B².

If something looks wrong, please tell us.

In this lesson a matrix is written in a line, with a semicolon between rows: [1 2; 0 1] means the matrix with first row 1, 2 and second row 0, 1. Once you can multiply two matrices, the next step is to multiply a matrix by itself. This gives powers, and it also shows where matrix algebra differs from ordinary algebra.

Powers of a square matrix

Only a square matrix A can be multiplied by itself. Define A² = A·A, A³ = A²·A and so on. The identity matrix I (ones on the main diagonal, zeros elsewhere) plays the role of the number 1: AI = IA = A for every square A of the same order, and we define A⁰ = I. The zero matrix O has every entry 0, with A + O = A and AO = OA = O.

Worked example 1: a pattern in the powers

Let A = [1 1; 0 1]. Then A² = [1 1; 0 1][1 1; 0 1]. Row 1 times column 1 is 1·1 + 1·0 = 1; row 1 times column 2 is 1·1 + 1·1 = 2; row 2 times the columns gives 0 and 1. So A² = [1 2; 0 1]. Next A³ = A²·A = [1 2; 0 1][1 1; 0 1] = [1 3; 0 1]. The pattern is Aⁿ = [1 n; 0 1], so A⁵ = [1 5; 0 1]. For a diagonal matrix the rule is simpler still: the n-th power raises each diagonal entry to the n-th power, so [2 0; 0 3]³ = [8 0; 0 27].

For matrices, AB is usually not equal to BA, so any rule of ordinary algebra that depends on swapping the order of a product can fail.

Rules that fail

Rule for numbersWhat is true for matrices
ab = baAB = BA only for some pairs
(a + b)² = a² + 2ab + b²(A + B)² = A² + AB + BA + B²
(a − b)(a + b) = a² − b²(A − B)(A + B) = A² + AB − BA − B²
ab = 0 means a = 0 or b = 0AB = O does not force A = O or B = O
(ab)ⁿ = aⁿbⁿ(AB)² = ABAB, which is A²B² only when AB = BA

Worked example 2: why the middle term changes

Let A = [1 2; 0 1] and B = [1 0; 3 1]. Then AB = [1·1 + 2·3, 1·0 + 2·1; 0·1 + 1·3, 0 + 1] = [7 2; 3 1], while BA = [1 2; 3 3·2 + 1] = [1 2; 3 7]. They differ, so AB ≠ BA. Now A + B = [2 2; 3 2], and (A + B)² = [2·2 + 2·3, 2·2 + 2·2; 3·2 + 2·3, 3·2 + 2·2] = [10 8; 12 10]. The expansion A² + AB + BA + B² gives [10 8; 12 10] as well, since A² = [1 4; 0 1], B² = [1 0; 6 1], and adding [1 4; 0 1] + [7 2; 3 1] + [1 2; 3 7] + [1 0; 6 1] = [10 8; 12 10]. By contrast A² + 2AB + B² = [16 8; 12 4], which is a different matrix. So the formula with 2AB is wrong here.

Worked example 3: a product that is zero

Take A = [1 0; 0 0] and B = [0 0; 0 1]. Neither is the zero matrix, but AB = [1·0 + 0·0, 1·0 + 0·1; 0, 0] = O. Therefore, from AB = O you cannot conclude that A or B is O. This is also why AB = AC does not allow cancelling A to get B = C.

When the ordinary rules do work

If AB = BA (the matrices commute), the ordinary expansions hold for that pair (but not cancellation, and AB = O still need not mean A = O or B = O): (A + B)² = A² + 2AB + B² and so on. Every matrix commutes with I and with itself, so (A + I)² = A² + 2A + I, using AI = IA = A. The same is true for powers of one matrix, for example A²A³ = A⁵.

Key takeaways: powers need a square matrix; look for a pattern by computing A², A³; I is the matrix 1 and O the matrix 0; (A + B)² has two cross terms AB and BA; AB = O does not force A or B to be O; the ordinary rules return when AB = BA.

Check your understanding

  1. 1. If A = [1 1; 0 1], what is A³?

    1. [1 2; 0 1]
    2. [1 3; 0 1]
    3. [1 3; 0 3]
    4. [3 3; 0 3]
    Show answer

    Answer: B. A² = [1 2; 0 1], and A³ = A²A = [1 3; 0 1].

  2. 2. For two square matrices A and B, (A + B)² equals:

    1. A² + 2AB + B² always
    2. A² + B² only
    3. A² + 2BA + B² always
    4. A² + AB + BA + B²
    Show answer

    Answer: D. Expanding (A + B)(A + B) gives four terms in order: AA + AB + BA + BB. The two cross terms combine to 2AB only if AB = BA.

  3. 3. What is [2 0; 0 3]³?

    1. [8 0; 0 27]
    2. [6 0; 0 9]
    3. [8 0; 0 9]
    4. [2 0; 0 27]
    Show answer

    Answer: A. A diagonal matrix is cubed entry by entry on the diagonal: 2³ = 8 and 3³ = 27.

  4. 4. If AB = O for two matrices A and B, then:

    1. A must be the zero matrix
    2. B must be the zero matrix
    3. nothing definite follows about A or B
    4. A and B must both be zero matrices
    Show answer

    Answer: C. A = [1 0; 0 0] and B = [0 0; 0 1] give AB = O although neither is O, so the conclusion does not follow.

  5. 5. For A = [1 2; 0 1] and B = [1 0; 3 1], what is AB?

    1. [1 2; 3 7]
    2. [1 0; 0 1]
    3. [7 2; 3 7]
    4. [7 2; 3 1]
    Show answer

    Answer: D. Row 1 of A times the columns of B: 1·1 + 2·3 = 7 and 1·0 + 2·1 = 2. Row 2: 0·1 + 1·3 = 3 and 0 + 1 = 1.

Privacy · Terms · Refunds · Delivery