Tensors

Mathematical methods · 76 questions and templates for CSIR NET Physics, in order from warm-up to top rank. Check each answer, then read the full solution.

Loading questions… (they need JavaScript; nothing you type is sent anywhere)

Written questions with solutions

The same questions as a list, so you can print them or read without the practice box. Templates (fresh numbers) appear only in the practice box above.

  1. MCQWarm-up

    In three dimensions, with the summation convention, δii equals
    1. 1
    2. 0
    3. 3
    4. 9
    Answer and solution

    Answer: C

    A repeated index is summed: δii=δ11+δ22+δ33=3. In n dimensions it is n (the trace of the identity).

    Why people get this wrong: Reading δii as a single component equal to 1.

  2. MCQWarm-up

    A general tensor of rank 3 in four dimensions has how many components?
    1. 12
    2. 64
    3. 81
    4. 24
    Answer and solution

    Answer: B

    Each of the 3 indices runs over 4 values: 43=64. In general nr for rank r in n dimensions.

    Why people get this wrong: 34=81 swaps the rank and the dimension.

  3. MCQWarm-up

    Given ε123=1, the value of ε213 is
    1. 1
    2. −1
    3. 0
    4. 2
    Answer and solution

    Answer: B

    213 comes from 123 by one swap (of the first two indices), an odd permutation, so ε213=−1. It is 0 whenever two indices are equal.

    Why people get this wrong: Thinking any rearrangement gives +1: only even (cyclic) ones do.

  4. MSQWarm-up

    With the summation convention, which of these are scalars (rank 0)?
    1. AiBi
    2. Tii
    3. AiBj
    4. εijkAiBjCk
    Answer and solution

    Answer: A, B, D

    Count the free (unrepeated) indices. AiBi (dot product), Tii (trace) and εijkAiBjCk (triple product) have none. AiBj has two free indices: it is the rank-2 outer product.

    Why people get this wrong: The rank is the number of free indices, not the number of letters.

  5. MCQWarm-up

    Contracting one pair of indices of a rank-4 tensor gives a tensor of rank
    1. 3
    2. 1
    3. 0
    4. 2
    Answer and solution

    Answer: D

    Contraction sets two indices equal and sums, removing both: 4−2=2. For example Rijklδik=Rjl-type objects.

    Why people get this wrong: One contraction removes two indices, not one.

  6. NATWarm-up

    A symmetric rank-2 tensor in three dimensions has ______ independent components.
    Answer and solution

    Answer: 6

    Tij=Tji: the 3 diagonal components and the 3 above the diagonal are free, n(n+1)2=6. The inertia and stress tensors are examples.

    Why people get this wrong: Counting all 9.

  7. NATStandard

    In three dimensions, εijkεijk (summed over all three indices) equals ______.
    Answer and solution

    Answer: 6

    Each non-zero εijk is ±1 and squares to 1; there are 3!=6 permutations. Or from εijkεijm=2δkm: 2δkk=6.

    Why people get this wrong: Answering 27 (all index combinations) or 1.

  8. MCQStandard

    The contraction εijkεimn equals
    1. δjmδkn+δjnδkm
    2. δjnδkm−δjmδkn
    3. δjmδkn−δjnδkm
    4. 2δjmδkn
    Answer and solution

    Answer: C

    Check a case: j=m=2, k=n=3: the left side is ε1232=1, and δ22δ33−δ23δ32=1 ✓. With j=n=2, k=m=3: ε123ε132=−1, and the formula gives 0−1=−1 ✓. This identity proves 𝐀×(𝐁×𝐂)=𝐁(𝐀⋅𝐂)−𝐂(𝐀⋅𝐁).

    Why people get this wrong: The sign: matching positions (j with m, k with n) come with +.

  9. MCQStandard

    An antisymmetric rank-2 tensor in four dimensions (such as the electromagnetic field tensor) has how many independent components?
    1. 4
    2. 6
    3. 10
    4. 16
    Answer and solution

    Answer: B

    The diagonal vanishes and the lower triangle is minus the upper, leaving n(n−1)2=6: exactly the three components of 𝐄 and three of 𝐁.

    Why people get this wrong: 10 is the count for a symmetric rank-2 tensor in 4D.

  10. MCQStandard

    Under a rotation xi′=Rijxj, the components of a Cartesian rank-2 tensor transform as
    1. Tij′=RikTkj
    2. Tij′=RkiRljTkl only for antisymmetric T
    3. Tij′=RikRjlTkl
    4. Tij′=Tij, since tensors do not depend on axes
    Answer and solution

    Answer: C

    Each index transforms like a vector component: one factor of R per index. In matrix form T′=RTR𝖳. The tensor (the physical object) does not change; its components do.

    Why people get this wrong: A rank-2 tensor needs two factors of R.

  11. MCQStandard

    In three dimensions, δijδjkδki equals
    1. 1
    2. 3
    3. 9
    4. 27
    Answer and solution

    Answer: B

    δijδjk=δik (the delta replaces j by i). Then δikδki=δii=3.

    Why people get this wrong: Multiplying three traces: 33.

  12. MSQStandard

    A tensor is isotropic if its components are the same in every rotated Cartesian frame. Which are isotropic in three dimensions?
    1. δij
    2. εijk (under proper rotations)
    3. δijδkl
    4. any constant vector ai≠0
    Answer and solution

    Answer: A, B, C

    RikRjlδkl=(RR𝖳)ij=δij. RilRjmRknεlmn=(det⁡R)εijk=εijk for proper rotations. Products of isotropic tensors are isotropic. A non-zero vector points somewhere, so rotating changes its components: there is no isotropic rank-1 tensor.

    Why people get this wrong: A constant vector is constant in space, not invariant under rotation.

  13. MCQStandard

    In index notation, the ith component of 𝐀×𝐁 is
    1. εijkAjBk
    2. δijAjBi
    3. εijkAiBj
    4. AjBk−AkBj summed over j and k
    Answer and solution

    Answer: A

    Check i=1: ε123A2B3+ε132A3B2=A2B3−A3B2 ✓. The third option has free index k, not i. The last, summed over both, gives 0.

    Why people get this wrong: The free index must be the one in the result.

  14. MCQStandard

    For the Minkowski metric ημν=diag(1,−1,−1,−1), the contraction ημνημν equals
    1. −2
    2. 0
    3. 2
    4. 4
    Answer and solution

    Answer: D

    ημνηνλ=δμλ, so ημνημν=δμμ=4, the dimension. Component by component: 1+1+1+1, since each diagonal entry squares to 1.

    Why people get this wrong: Adding the diagonal entries 1−1−1−1=−2: that sum is not a contraction, and it is not invariant.

  15. MCQStandard

    In plane polar coordinates (r,θ), the metric tensor gij is
    1. diag(1,r)
    2. diag(1,r2)
    3. diag(r,r)
    4. diag(1,1)
    Answer and solution

    Answer: B

    The line element is ds2=dr2+r2dθ2, read off as grr=1, gθθ=r2. The scale factor hθ=r is the square root of gθθ.

    Why people get this wrong: Confusing the scale factor r with the metric component r2.

  16. MCQStandard

    For a rank-2 tensor in three dimensions, which quantity is unchanged by every rotation of the axes?
    1. T11
    2. T12
    3. T11+T22
    4. det⁡T
    Answer and solution

    Answer: D

    T′=RTR𝖳 is a similarity transformation, so the eigenvalues, the trace Tii and the determinant are invariant. Single components and partial sums like T11+T22 change.

    Why people get this wrong: Only the full trace is invariant, not part of it.

  17. NATExam level

    A 2D tensor is T=diag(3,1) in some axes. In axes rotated by 30°, the component T11′ is ______ (to two decimal places).
    Answer and solution

    Answer: 2.5 (±0.01)

    T11′=R1kR1lTkl=cos2⁡θT11+sin2⁡θT22 (the off-diagonal terms are 0). =3×34+1×14=2.5. The trace stays 4=2.5+T22′, so T22′=1.5.

    Why people get this wrong: Rotating with one factor of cos⁡θ instead of cos2⁡θ.

  18. NATExam level

    The 2D symmetric tensor T=(3111) becomes diagonal when the axes are rotated by an angle θ between −45° and 45°. Find θ in degrees: ______
    Answer and solution

    Answer: 22.5 (±0.1)

    Setting T12′=0 gives tan⁡2θ=2T12T11−T22=22=1, so 2θ=45° and θ=22.5°. The principal values are 2±2.

    Why people get this wrong: Forgetting the factor 2 in tan⁡2θ.

  19. MCQExam level

    The contraction εijkδij equals
    1. δkk
    2. 1
    3. εkkk, which is 1
    4. 0
    Answer and solution

    Answer: D

    Contracting an antisymmetric pair with a symmetric one always gives zero: εijkδij=εiik=0. In general AijSij=0 for A antisymmetric and S symmetric.

    Why people get this wrong: Any ε with two equal indices is 0.

  20. MSQExam level

    Write Tij=Sij+Aij with S symmetric and A antisymmetric (3D). Which statements are true?
    1. Sij=12(Tij+Tji)
    2. Aij has 3 independent components and is equivalent to the axial vector vk=12εkijAij
    3. SijAij=0
    4. Part of the trace of T sits in A
    Answer and solution

    Answer: A, B, C

    S=12(T+T𝖳), A=12(T−T𝖳). In 3D an antisymmetric matrix has 3 entries above the diagonal and acts like a cross product with a vector (Aijwj=εijkwjvk up to sign convention): the rotation part of a velocity gradient, or 𝐁 inside the field tensor. SijAij=SjiAji=−SijAij, so it is 0. A has zero diagonal, so the whole trace is in S.

    Why people get this wrong: The antisymmetric part is traceless.

  21. NATExam level

    With r2=xjxj, compute ∂i(xir2) at a point where r=2: ______
    Answer and solution

    Answer: 20

    Product rule: ∂i(xir2)=(∂ixi)r2+xi∂i(xjxj)=3r2+xi(2xi)=5r2. At r=2: 20. (This is ∇⋅(r2𝐫).)

    Why people get this wrong: ∂ixi=δii=3, not 1.

  22. MCQExam level

    For every vector Bj, the quantity Ci=TijBj is a vector. Then Tij is
    1. a rank-2 tensor
    2. a scalar
    3. a vector
    4. not necessarily a tensor
    Answer and solution

    Answer: A

    The quotient law: Ci′=RikCk=RikTklBl=RikTklRjlBj′. Since also Ci′=Tij′Bj′ for arbitrary B′, Tij′=RikRjlTkl: the rank-2 law. Example: Li=Iijωj shows that inertia is a tensor.

    Why people get this wrong: It must hold for arbitrary B; for one special B it proves nothing.

  23. NATExam level

    On a unit sphere, ds2=dθ2+sin2⁡θdϕ2. The length of the circle θ=60° (going once round in ϕ) is ______ (to three decimal places).
    Answer and solution

    Answer: 5.441 (±0.003)

    Along the circle dθ=0, so ds=sin⁡θdϕ and the length is 2πsin⁡60°=2π×0.866=5.441. Not 2π×60° in radians: the metric decides lengths.

    Why people get this wrong: Using cos⁡θ: θ is measured from the pole, so the radius of the circle is sin⁡θ.

  24. MCQExam level

    For plane polar coordinates (grr=1, gθθ=r2), the Christoffel symbol Γθθr is
    1. −r
    2. r
    3. 1/r
    4. 0
    Answer and solution

    Answer: A

    Γθθr=12grr(2∂θgrθ−∂rgθθ)=−12∂r(r2)=−r. In the geodesic equation it gives r¨−rθ˙2=0: the centripetal term of polar kinematics. The other non-zero one is Γrθθ=1/r.

    Why people get this wrong: 1/r is Γrθθ.

  25. NATExam level

    A point mass of 2 kg sits at (1,1,0) m. Using Iij=∑m(r2δij−xixj), the product of inertia Ixy is ______ kg m².
    Answer and solution

    Answer: -2 kg m²

    Off the diagonal δxy=0, so Ixy=−mxy=−2×1×1=−2 kg m². A non-zero product of inertia means the axes are not principal: spinning about z alone needs a torque.

    Why people get this wrong: Dropping the minus sign in the definition.

  26. MCQExam level

    Under a general orthogonal transformation R (rotations and reflections), εijk behaves as
    1. a true rank-3 tensor
    2. a pseudotensor: it picks up an extra factor det⁡R
    3. a scalar
    4. nothing definite, since it is not a tensor at all
    Answer and solution

    Answer: B

    If we insist that its components stay ±1,0 in every frame, then the transformation law must include det⁡R: εijk′=(det⁡R)RilRjmRknεlmn. That is why 𝐀×𝐁 of two polar vectors is an axial vector (angular momentum, magnetic field) that does not flip under parity.

    Why people get this wrong: Under proper rotations alone it does behave as a true isotropic tensor.

  27. MSQExam level

    Which statements about covariant and contravariant components are true?
    1. Aμ=gμνAν
    2. AμBμ is invariant
    3. In Cartesian coordinates with the Euclidean metric, covariant and contravariant components coincide
    4. The coordinate differentials dxμ transform covariantly
    Answer and solution

    Answer: A, B, C

    The metric lowers indices and its inverse raises them. A contraction of an upper with a lower index is invariant because the transformation matrices cancel. With gij=δij lowering changes nothing. dx′μ=∂x′μ∂xνdxν is the contravariant law; gradients ∂μϕ are covariant.

    Why people get this wrong: Differentials are the model contravariant vector; gradients the model covariant one.

  28. NATExam level

    In units with c=1, the invariant FμνFμν=2(B2−E2). For |𝐄|=3 and |𝐁|=5 it equals ______.
    Answer and solution

    Answer: 32

    2(25−9)=32. Because it is a full contraction it is the same in every inertial frame: if B>E in one frame, no frame can make the field purely electric.

    Why people get this wrong: Using B2+E2, which is the energy density (not invariant).

  29. MCQExam level

    In three dimensions, with the summation convention, εijkεijlAkBl equals
    1. 2𝐀⋅𝐁
    2. 𝐀⋅𝐁
    3. 6𝐀⋅𝐁
    4. 0
    Answer and solution

    Answer: A

    Contracting two indices: εijkεijl=2δkl (from εijkεimn=δjmδkn−δjnδkm with m=j: 3δkn−δkn). Then 2δklAkBl=2𝐀⋅𝐁.

    Why people get this wrong: Using δjj=1 instead of 3 in the contraction gives 0.

  30. MCQExam level

    Under spatial inversion 𝐫→−𝐫 (with 𝐩→−𝐩), the angular momentum Li=εijkxjpk
    1. is unchanged
    2. changes sign
    3. becomes zero
    4. changes in a way that depends on the axes chosen
    Answer and solution

    Answer: A

    xj→−xj and pk→−pk, while the symbol εijk is a fixed set of numbers. So Li→εijk(−xj)(−pk)=Li. A true (polar) vector would flip; 𝐋 does not, so it is an axial vector (a pseudovector), equivalently an antisymmetric rank-2 tensor xjpk−xkpj in disguise.

    Why people get this wrong: Treating 𝐋 as an ordinary vector that must flip under inversion.

← All topics · Timed mock test →