Home / Courses / CSIR NET Physics / Practice / The Dirac delta and Green's functions

The Dirac delta and Green's functions

Mathematical methods · 73 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. NATWarm-up

    The value of ∫−∞∞(x2+3)δ(x−2)dx is ______.
    Answer and solution

    Answer: 7

    The delta picks out the value of the rest of the integrand where its argument vanishes: x=2, giving 4+3=7.

    Why people get this wrong: Evaluating at x=0 out of habit gives 3.

  2. MCQWarm-up

    The distribution δ(3x) equals
    1. 3δ(x)
    2. δ(x)
    3. δ(x)/3
    4. δ(x/3)
    Answer and solution

    Answer: C

    Substitute u=3x: ∫f(x)δ(3x)dx=∫f(u/3)δ(u)du3=f(0)3. So δ(ax)=δ(x)/|a|. The spike is three times narrower, so it holds a third of the area.

    Why people get this wrong: Pulling the 3 out as a factor: the delta is not linear in its argument.

  3. MCQWarm-up

    The derivative of the Heaviside step function θ(x) (0 for x<0, 1 for x>0) is
    1. 0 everywhere
    2. δ(x)
    3. δ′(x)
    4. 1/x
    Answer and solution

    Answer: B

    For any smooth f that vanishes at infinity: ∫fθ′dx=−∫f′θdx=−∫0∞f′dx=f(0). That is the action of δ(x). A unit jump differentiates into a unit spike.

    Why people get this wrong: The derivative is 0 everywhere except at one point, but that point carries all the area.

  4. MCQWarm-up

    If x is a length measured in metres, the unit of δ(x) is
    1. m
    2. dimensionless
    3. m²
    4. m⁻¹
    Answer and solution

    Answer: D

    ∫δ(x)dx=1 is a pure number and dx is in metres, so δ(x) is in m⁻¹. Likewise δ3(𝐫) is in m⁻³, which is why a point charge density is qδ3(𝐫) in C/m³.

    Why people get this wrong: Treating the delta as a pure number like 0 or 1.

  5. MCQWarm-up

    The value of ∫−∞∞cos⁡xδ(x−π)dx is
    1. 1
    2. 0
    3. −1
    4. π
    Answer and solution

    Answer: C

    Sifting: cos⁡π=−1.

    Why people get this wrong: Answering π, the location of the spike, instead of the value of the function there.

  6. MSQWarm-up

    Which of these hold (as distributions)?
    1. xδ(x)=0
    2. δ(−x)=δ(x)
    3. ∫12δ(x)dx=1
    4. f(x)δ(x−a)=f(a)δ(x−a)
    Answer and solution

    Answer: A, B, D

    ∫g(x)xδ(x)dx=0⋅g(0)=0. δ(−x)=δ(x)/|−1|. The product rule fδ(x−a)=f(a)δ(x−a) is sifting. But [1,2] does not contain x=0, so that integral is 0.

    Why people get this wrong: The delta contributes only if its spike lies inside the range of integration.

  7. NATStandard

    The value of ∫−∞∞x2δ(x2−4)dx is ______.
    Answer and solution

    Answer: 2

    For δ(g(x)) use ∑iδ(x−xi)|g′(xi)| over the simple roots. g=x2−4 has roots ±2 with |g′|=4. So the integral is 44+44=2.

    Why people get this wrong: Keeping only the root x=2, or forgetting to divide by |g′|.

  8. MCQStandard

    For a smooth f, ∫−∞∞f(x)δ′(x)dx equals
    1. f′(0)
    2. f(0)
    3. 0
    4. −f′(0)
    Answer and solution

    Answer: D

    Integrate by parts; the boundary term vanishes: ∫fδ′dx=[fδ]−∫f′δdx=−f′(0). In general ∫fδ(n)dx=(−1)nf(n)(0).

    Why people get this wrong: The sign: each derivative moved across costs a factor −1.

  9. NATStandard

    The value of ∫−∞∞e2xδ′(x)dx is ______.
    Answer and solution

    Answer: -2

    ∫fδ′=−f′(0) with f=e2x, f′=2e2x: −2.

    Why people get this wrong: Answering +2 by forgetting the minus sign from integration by parts.

  10. MCQStandard

    Which integral representation of δ(x) is correct?
    1. δ(x)=12π∫−∞∞eikxdk
    2. δ(x)=∫−∞∞eikxdk
    3. δ(x)=12π∫−∞∞e−k2xdk
    4. δ(x)=12π∫0∞cos⁡kxdk
    Answer and solution

    Answer: A

    The Fourier transform of δ is ∫δ(x)e−ikxdx=1; inverting, δ(x)=12π∫eikxdk. (Using cosines, 1π∫0∞cos⁡kxdk; the last option has the wrong factor.)

    Why people get this wrong: The 2π: check by integrating over x, which gives 2πδ(k) and then 1.

  11. MSQStandard

    As ε→0+, which of these tend to δ(x)?
    1. 1επe−x2/ε2
    2. 1πεx2+ε2
    3. 12ε for |x|<ε, and 0 otherwise
    4. 1εe−|x|/ε
    Answer and solution

    Answer: A, B, C

    A delta sequence needs area 1 and a width shrinking to zero. The Gaussian, the Lorentzian and the box all have area 1. ∫1εe−|x|/εdx=2, so that sequence tends to 2δ(x); it needs 12ε in front.

    Why people get this wrong: Check the area before trusting the shape.

  12. MCQStandard

    The value of ∫−∞∞x3δ(2x−6)dx is
    1. 27
    2. 13.5
    3. 54
    4. 108
    Answer and solution

    Answer: B

    δ(2x−6)=δ(2(x−3))=12δ(x−3). So the integral is 12⋅27=13.5.

    Why people get this wrong: Evaluating x3 at the root but forgetting the 1/|a| factor.

  13. MCQStandard

    For a linear differential operator Lˆ, the Green's function G(x,x′) is defined by
    1. LˆG(x,x′)=0
    2. LˆG(x,x′)=1
    3. LˆG(x,x′)=δ(x−x′), with the boundary conditions of the problem
    4. G(x,x′)=Lˆδ(x−x′)
    Answer and solution

    Answer: C

    G is the response at x to a unit point source at x′, satisfying the same homogeneous boundary conditions as the problem. Then u(x)=∫G(x,x′)f(x′)dx′ solves Lˆu=f, because Lˆ acts on x and ∫δ(x−x′)f(x′)dx′=f(x).

    Why people get this wrong: Forgetting that G also carries the boundary conditions.

  14. MCQStandard

    The Laplacian of 1/r (with r=|𝐫|) is
    1. 0 everywhere
    2. 4πδ3(𝐫)
    3. −δ3(𝐫)
    4. −4πδ3(𝐫)
    Answer and solution

    Answer: D

    For r≠0, ∇2(1/r)=0. But the flux of ∇(1/r)=−rˆ/r2 through any sphere is −4π, so by the divergence theorem the volume integral of ∇2(1/r) is −4π: all concentrated at the origin. Hence ∇21r=−4πδ3(𝐫); Coulomb's law and Gauss's law are this identity.

    Why people get this wrong: "Zero everywhere" misses the point source.

  15. MCQStandard

    If |𝐫0|=2, then ∫all spacee−rδ3(𝐫−𝐫0)d3r equals
    1. e−2/(16π)
    2. 4πe−2
    3. e−2
    4. 1
    Answer and solution

    Answer: C

    The three-dimensional delta sifts just like the one-dimensional one: the integral is e−|𝐫0|=e−2≈0.135. No 4πr2 factor appears, because δ3 already absorbs the volume element.

    Why people get this wrong: Adding a 4πr2 Jacobian by hand.

  16. MCQStandard

    G(x,x′) solves d2Gdx2=δ(x−x′) on [0,L] with G=0 at both ends. Across x=x′, the slope ∂G/∂x
    1. is continuous
    2. jumps by +1
    3. jumps by −1
    4. becomes infinite
    Answer and solution

    Answer: B

    Integrate the equation from x′−ε to x′+ε: [Gx]x′−x′+=1. G itself is continuous (a jump in G would make G′′ contain δ′).

    Why people get this wrong: Thinking G jumps: it is the derivative that jumps, giving a kink (a "tent" shape).

  17. NATExam level

    Solve y′′=1 on [0,1] with y(0)=y(1)=0 (for example with the Green's function G=x<(x>−1)). The value y(0.3) is ______ (to three decimal places).
    Answer and solution

    Answer: -0.105 (±0.001)

    y(x)=∫01G(x,x′)⋅1dx′=(x−1)∫0xx′dx′+x∫x1(x′−1)dx′=(x−1)x22−x(1−x)22=x(x−1)2. Check: y′′=1 and both ends are 0. At x=0.3: 0.3×(−0.7)2=−0.105.

    Why people get this wrong: Sign: G≤0 here, because y′′>0 with zero ends means the curve sags below the axis.

  18. MCQExam level

    The Green's function for d2dx2 on [0,1] with G(0,x′)=G(1,x′)=0 is (x< and x> are the smaller and larger of x and x′)
    1. x>(x<−1)
    2. x<(x>−1)
    3. |x−x′|
    4. 12|x−x′|
    Answer and solution

    Answer: B

    Left of x′ the solution of G′′=0 vanishing at 0 is Ax; right of it, vanishing at 1, is B(x−1). Continuity at x′ and a slope jump of +1 give A=x′−1, B=x′. So G=x<(x>−1). 12|x−x′| is the whole-line Green's function, which ignores the end conditions.

    Why people get this wrong: Using the free-space Green's function in a bounded problem.

  19. NATExam level

    The value of ∫04x2δ(sin⁡x)dx is ______ (to two decimal places).
    Answer and solution

    Answer: 9.87 (±0.01)

    sin⁡x=0 at x=0,π,2π,…; inside (0,4) only x=π (as 2π≈6.28>4, and the endpoint 0 has x2=0 anyway). There |cos⁡π|=1. Result: π2≈9.87.

    Why people get this wrong: Including 2π, which lies outside the range.

  20. MSQExam level

    For the underdamped oscillator mx¨+mγx˙+mω02x=F(t), let G(t) be the causal (retarded) Green's function, so that x(t)=∫G(t−t′)F(t′)dt′. Which statements are true?
    1. G(t)=0 for t<0
    2. G contains a delta function at t=0
    3. G(t)=θ(t)e−γt/2sin⁡ω1tmω1 with ω1=ω02−γ2/4
    4. G is continuous at t=0 while G˙ jumps by 1/m
    Answer and solution

    Answer: A, C, D

    Causality: no response before the kick. Integrating the equation across t=0: m[G˙]=1, so G˙ jumps from 0 to 1/m while G stays 0: a kicked mass starts at rest position with velocity 1/m. For t>0 it is a free damped oscillation with those initial values, which gives the formula. A delta in G itself would make G¨ contain δ′′, which nothing on the right-hand side balances.

    Why people get this wrong: The impulse changes velocity, not position.

  21. NATExam level

    A 2 kg mass on a spring of stiffness 8 N/m rests at equilibrium. At t=0 it receives an impulse of 4 N s. Its displacement at t=0.5 s is ______ m (to three decimal places).
    Answer and solution

    Answer: 0.841 m (±0.002)

    ω0=8/2=2 rad/s. The response is PG(t)=Pmω0sin⁡ω0t=42×2sin⁡2t=sin⁡2t. At t=0.5: sin⁡1=0.841 m.

    Why people get this wrong: Dividing by k instead of mω0.

  22. MCQExam level

    The Green's function satisfying ∇2G(𝐫,𝐫′)=δ3(𝐫−𝐫′) in all space, vanishing at infinity, is
    1. 14π|𝐫−𝐫′|
    2. −14π|𝐫−𝐫′|
    3. −1|𝐫−𝐫′|
    4. 12πln⁡|𝐫−𝐫′|
    Answer and solution

    Answer: B

    From ∇21r=−4πδ3: ∇2(−14πr)=δ3. Applied to ∇2V=−ρ/ε0 this gives Coulomb's potential V=14πε0∫ρ(𝐫′)|𝐫−𝐫′|d3r′. The logarithm is the two-dimensional Green's function.

    Why people get this wrong: The sign depends on whether the equation is ∇2G=δ or −∇2G=δ; read it.

  23. MCQExam level

    In two dimensions, a solution of ∇2G=δ2(𝐫−𝐫′) is
    1. 12πln⁡|𝐫−𝐫′|
    2. −14π|𝐫−𝐫′|
    3. 12π|𝐫−𝐫′|
    4. |𝐫−𝐫′|2/4
    Answer and solution

    Answer: A

    With ρ=|𝐫−𝐫′|: ∇(12πln⁡ρ)=ρˆ2πρ, whose flux through any circle is 12πρ⋅2πρ=1. So its Laplacian is a unit delta. This is the potential of a line charge (per unit length).

    Why people get this wrong: Using the 3D form 1/r, whose flux through a circle is not constant.

  24. MCQExam level

    Using 12π∫eik(x−a)dk=δ(x−a), the double integral 12π∫−∞∞dx∫−∞∞dkx2eik(x−3) equals
    1. 3
    2. 6
    3. 9
    4. 2π⋅9
    Answer and solution

    Answer: C

    Do the k integral first: it gives δ(x−3) (the 12π is already there). Then ∫x2δ(x−3)dx=9.

    Why people get this wrong: Leaving an extra 2π.

  25. MSQExam level

    Which statements about Green's functions are true?
    1. G depends on both the operator and the boundary conditions
    2. Once G is known, the solution for any source follows by superposition
    3. The method works just as well for non-linear equations
    4. For a self-adjoint operator with homogeneous boundary conditions, G(x,x′)=G(x′,x)
    Answer and solution

    Answer: A, B, D

    The whole method rests on linearity: a source is a sum of point sources, so the response is the same sum of point responses. It fails for non-linear equations. Reciprocity G(x,x′)=G(x′,x) holds for self-adjoint (Sturm–Liouville) problems: the response at A to a source at B equals the response at B to a source at A. Causal Green's functions in time are not symmetric.

    Why people get this wrong: Non-linear equations break superposition.

  26. NATExam level

    The value of ∫−∞∞x3δ(x2−3x+2)dx is ______.
    Answer and solution

    Answer: 9

    g=(x−1)(x−2), g′=2x−3: |g′(1)|=1, |g′(2)|=1. The integral is 11+81=9.

    Why people get this wrong: Forgetting the absolute value: g′(1)=−1, and a negative weight would give 7.

  27. MCQExam level

    The causal Green's function of dydt+ay=f(t) (with a>0) is
    1. θ(t)e−at
    2. e−a|t|/2a
    3. θ(t)eat
    4. θ(−t)e−at
    Answer and solution

    Answer: A

    G′+aG=δ(t) with G=0 for t<0. Integrating across 0: G(0+)=1. For t>0, G=e−at. So G=θ(t)e−at: an RC circuit or a cooling body responding to a sudden kick.

    Why people get this wrong: e−a|t|/2a is the symmetric Green's function of a second-order operator, not a causal first-order one.

  28. NATExam level

    For dydt+2y=θ(t) with y=0 for t<0, the value y(1) is ______ (to three decimal places).
    Answer and solution

    Answer: 0.432 (±0.002)

    y(t)=∫0tG(t−t′)⋅1dt′=∫0te−2(t−t′)dt′=1−e−2t2. At t=1: 1−0.13532=0.432.

    Why people get this wrong: Using the steady value 12 before the transient has died away.

  29. MCQExam level

    The value of ∫−∞∞exδ(x2−3x+2)dx is
    1. (e+e2)/2
    2. e3/2
    3. e2−e
    4. e+e2
    Answer and solution

    Answer: D

    δ(g(x))=∑iδ(x−xi)/|g′(xi)| over the simple roots of g. Here g=(x−1)(x−2), roots 1 and 2, g′=2x−3 with |g′|=1 at both. So the integral is e1+e2≈10.11.

    Why people get this wrong: Using g′ without the absolute value gives e2−e (the root at 1 has g′=−1).

  30. MCQExam level

    As distributions, xδ′(x) equals
    1. δ(x)
    2. 0
    3. −δ(x)
    4. δ′(x)
    Answer and solution

    Answer: C

    Test against a smooth f: ∫f(x)xδ′(x)dx=−ddx(xf)|0=−[f(0)+0⋅f′(0)]=−f(0). That is the action of −δ. So xδ′(x)=−δ(x).

    Why people get this wrong: "xδ=0, so xδ′=0" is wrong: the derivative falls on x as well.

← All topics · Timed mock test →