Home / Courses / IIT JAM Physics / Practice / Partial derivatives and Jacobians

Partial derivatives and Jacobians

Mathematical methods · 82 questions and templates for IIT JAM Physics, in order from warm-up to top rank. Check each answer, then read the full solution.

New to this topic? Read the lesson first: Partial derivatives, Jacobians and exact differentials.

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

    For f(x,y)=x2y+y3, the partial derivative ∂f/∂x is
    1. 2xy+3y2
    2. 2xy
    3. x2+3y2
    4. 2x
    Answer and solution

    Answer: B

    Differentiate with respect to x while holding y fixed, so y behaves like a constant: ∂(x2y)/∂x=2xy and ∂(y3)/∂x=0. So fx=2xy.

    Why people get this wrong: Differentiating y3 as if y depended on x. In a partial derivative the other variables are frozen.

  2. NATWarm-up

    For f(x,y)=x2y3, the value of ∂2f∂x∂y at (1,1) is ______.
    Answer and solution

    Answer: 6

    fy=3x2y2, then ∂/∂x of that is 6xy2. At (1,1): 6. (The other order, fx=2xy3 then ∂/∂y, gives 6xy2 too.)

    Why people get this wrong: Taking a second derivative in the same variable (fxx=2y3) instead of a mixed one.

  3. MCQWarm-up

    The total differential of z=xy is
    1. xdx+ydy
    2. dxdy
    3. ydx+xdy
    4. xy(dx+dy)
    Answer and solution

    Answer: C

    dz=∂z∂xdx+∂z∂ydy=ydx+xdy. It is the product rule written with differentials.

    Why people get this wrong: Pairing each variable with its own differential (xdx) instead of with the other one.

  4. MSQWarm-up

    Which of these differentials Mdx+Ndy are exact?
    1. ydx+xdy
    2. xdy−ydx
    3. 2xydx+x2dy
    4. ycos⁡xdx+sin⁡xdy
    Answer and solution

    Answer: A, C, D

    Test ∂M/∂y=∂N/∂x. First: 1=1, exact (d(xy)). Second: M=−y, N=x: −1≠1, not exact. Third: 2x=2x, exact (d(x2y)). Fourth: cos⁡x=cos⁡x, exact (d(ysin⁡x)).

    Why people get this wrong: Comparing ∂M/∂x with ∂N/∂y: the test crosses over, M with y and N with x.

  5. NATWarm-up

    The sides of a rectangle are measured as 10 cm and 5 cm, each with a maximum error of 1%. The maximum percentage error in its area is ______ %.
    Answer and solution

    Answer: 2 %

    A=ab, so dA/A=da/a+db/b. In the worst case the errors add: 1%+1%=2%.

    Why people get this wrong: Multiplying the percentages (1%×1%) or averaging them. For products, relative errors add.

  6. MCQWarm-up

    For an ideal gas P=nRT/V, the derivative (∂P/∂T)V is
    1. −nRT/V2
    2. nRT
    3. R/V
    4. nR/V
    Answer and solution

    Answer: D

    The subscript V means volume is held constant. Then P is just (nR/V)×T, and its derivative with respect to T is nR/V.

    Why people get this wrong: −nRT/V2 is (∂P/∂V)T, the derivative with respect to the other variable.

  7. MCQWarm-up

    For f(x,y)=xy, the partial derivative ∂f∂y is
    1. 1y
    2. xy2
    3. −xy2
    4. 1−xy2
    Answer and solution

    Answer: C

    Hold x fixed and differentiate xy−1 with respect to y: x⋅(−1)y−2=−x/y2.

    Why people get this wrong: 1/y is ∂f/∂x; read which variable is changing.

  8. NATStandard

    Let z=x2+y2 with x=cos⁡t and y=2sin⁡t. The value of dz/dt at t=π/4 is ______.
    Answer and solution

    Answer: 3 (±0.001)

    Chain rule: dzdt=2xdxdt+2ydydt=2cos⁡t(−sin⁡t)+2(2sin⁡t)(2cos⁡t)=6sin⁡tcos⁡t=3sin⁡2t. At t=π/4: 3.

    Why people get this wrong: Using only one branch of the chain rule; z depends on t through both x and y.

  9. MCQStandard

    On the circle x2+y2=25, the slope dy/dx at the point (3,4) is
    1. 3/4
    2. −3/4
    3. −4/3
    4. 4/3
    Answer and solution

    Answer: B

    With F=x2+y2−25=0: dydx=−FxFy=−2x2y=−xy=−34. The tangent is perpendicular to the radius, whose slope is 4/3.

    Why people get this wrong: Forgetting the minus sign in dy/dx=−Fx/Fy, or inverting the ratio.

  10. MCQStandard

    Three variables are linked by one equation F(x,y,z)=0. The product (∂x∂y)z(∂y∂z)x(∂z∂x)y equals
    1. 1
    2. 0
    3. it depends on F
    4. −1
    Answer and solution

    Answer: D

    Each factor is minus a ratio of partial derivatives of F: (∂x/∂y)z=−Fy/Fx, (∂y/∂z)x=−Fz/Fy, (∂z/∂x)y=−Fx/Fz. The product is (−1)3=−1 for every F. Check with PV=nRT: (∂P/∂V)T(∂V/∂T)P(∂T/∂P)V=(−P/V)(nR/P)(V/nR)=−1.

    Why people get this wrong: Cancelling the "differentials" like fractions gives +1. Partial derivatives held at different constants do not cancel like that.

  11. NATStandard

    The isothermal compressibility is κT=−1V(∂V∂P)T. For an ideal gas at 2 atm, κT is ______ atm−1.
    Answer and solution

    Answer: 0.5 atm⁻¹

    V=nRT/P, so (∂V/∂P)T=−nRT/P2=−V/P. Then κT=−1V(−VP)=1P=0.5 atm−1.

    Why people get this wrong: Dropping the minus sign in the definition gives a negative compressibility, which would mean the gas expands when squeezed.

  12. MCQStandard

    For f(x,y)=(x2+y2)3/2, the combination x∂f∂x+y∂f∂y equals
    1. 3f
    2. f
    3. 2f
    4. 32f
    Answer and solution

    Answer: A

    Scaling x→λx, y→λy multiplies f by λ3: f is homogeneous of degree 3. Euler's theorem gives xfx+yfy=3f. Directly: fx=3x(x2+y2)1/2, so xfx+yfy=3(x2+y2)3/2.

    Why people get this wrong: Reading the degree off the outer power 3/2; the inside is already degree 2, so the total degree is 2×3/2=3.

  13. MSQStandard

    For f(x,y)=x2−y2 at the origin, which statements are true?
    1. The gradient of f is zero there.
    2. The origin is a saddle point.
    3. The origin is a local minimum.
    4. fxxfyy−fxy2<0 there.
    Answer and solution

    Answer: A, B, D

    ∇f=(2x,−2y)=0 at the origin. fxx=2, fyy=−2, fxy=0, so D=−4<0: a saddle. Along the x-axis f rises, along the y-axis it falls, so it is not a minimum.

    Why people get this wrong: A zero gradient does not mean a maximum or minimum; check the sign of D.

  14. NATStandard

    For f(x,y)=x2y near (2,3), use the total differential to estimate the change in f when x increases by 0.01 and y decreases by 0.02. The change is ______ (to two decimal places).
    Answer and solution

    Answer: 0.04 (±0.001)

    df=2xydx+x2dy=12(0.01)+4(−0.02)=0.12−0.08=0.04. (The exact change is 2.012×2.98−12=0.0395.)

    Why people get this wrong: Using dy=+0.02; a decrease means dy is negative.

  15. MCQStandard

    For plane polar coordinates x=rcos⁡θ, y=rsin⁡θ, the Jacobian ∂(x,y)∂(r,θ) is
    1. 1
    2. r2
    3. r
    4. 1/r
    Answer and solution

    Answer: C

    |cos⁡θ−rsin⁡θsin⁡θrcos⁡θ|=rcos2⁡θ+rsin2⁡θ=r. That is why dxdy=rdrdθ.

    Why people get this wrong: 1/r is the inverse Jacobian ∂(r,θ)/∂(x,y).

  16. NATStandard

    In a pendulum experiment g=4π2L/T2. The length is known to 0.5% and the period to 0.2%. The maximum percentage error in g is ______ %.
    Answer and solution

    Answer: 0.9 %

    Take logs and differentiate: dgg=dLL−2dTT. In the worst case the two parts add in size: 0.5%+2×0.2%=0.9%.

    Why people get this wrong: Subtracting because of the minus sign. A maximum error assumes the worst case, so sizes add.

  17. MCQStandard

    For f(x,y)=exy, the mixed derivative ∂2f/∂x∂y is
    1. xyexy
    2. exy
    3. (x+y)exy
    4. (1+xy)exy
    Answer and solution

    Answer: D

    fy=xexy. Now differentiate with respect to x by the product rule: exy+x⋅yexy=(1+xy)exy.

    Why people get this wrong: Forgetting the product rule (the factor x in fy also depends on x).

  18. MSQStandard

    For n moles of an ideal gas (with constant CV), which of these are exact differentials?
    1. CVdT
    2. PdV
    3. CVTdT+nRVdV
    4. VdP+PdV
    Answer and solution

    Answer: A, C, D

    Use T and V as variables. CVdT=dU: exact. PdV=(nRT/V)dV: ∂(0)/∂V=0 but ∂(nRT/V)/∂T=nR/V≠0, so not exact (work depends on the path). The third is dS=d(CVln⁡T+nRln⁡V): exact. The fourth is d(PV): exact.

    Why people get this wrong: Thinking PdV is exact because P and V are state variables. A differential is exact only if it is the change of some function.

  19. MCQStandard

    A function f(x,y) is rewritten in polar coordinates, x=rcos⁡θ, y=rsin⁡θ. Then ∂f∂θ (at fixed r) equals
    1. −yfx+xfy
    2. xfx+yfy
    3. yfx−xfy
    4. −yrfx+xrfy
    Answer and solution

    Answer: A

    Chain rule: ∂f∂θ=fx∂x∂θ+fy∂y∂θ=fx(−rsin⁡θ)+fy(rcos⁡θ)=−yfx+xfy. This is the operator behind the angular momentum Lz∝x∂y−y∂x. The combination xfx+yfy is r∂f/∂r.

    Why people get this wrong: ∂x/∂θ=−rsin⁡θ already contains r; dividing by r again is wrong.

  20. NATStandard

    On the sphere x2+y2+z2=9, z is a function of x and y near the point (1,2,2). The value of ∂z/∂x there is ______.
    Answer and solution

    Answer: -0.5 (±0.001)

    Differentiate with respect to x at fixed y: 2x+2z∂z∂x=0, so ∂z∂x=−xz=−12.

    Why people get this wrong: Holding y fixed means its derivative is zero; it does not enter.

  21. MCQExam level

    From dU=TdS−PdV, the equality of mixed partial derivatives gives
    1. (∂T∂V)S=(∂P∂S)V
    2. (∂T∂V)S=−(∂P∂S)V
    3. (∂T∂V)S=(∂S∂P)V
    4. (∂T∂V)S=−(∂V∂T)S
    Answer and solution

    Answer: B

    U(S,V) has (∂U/∂S)V=T and (∂U/∂V)S=−P. Since ∂2U/∂V∂S=∂2U/∂S∂V: (∂T/∂V)S=(∂(−P)/∂S)V=−(∂P/∂S)V.

    Why people get this wrong: Losing the minus sign that comes with −PdV.

  22. NATExam level

    The differential (3x2y+ay2)dx+(x3+4xy)dy is exact. The constant a is ______.
    Answer and solution

    Answer: 2

    ∂M/∂y=3x2+2ay and ∂N/∂x=3x2+4y. Equal for all x,y only if 2a=4, so a=2. (It is then d(x3y+2xy2).)

    Why people get this wrong: Matching ∂M/∂x with ∂N/∂y instead.

  23. NATExam level

    The exact differential dF=(2xy+3)dx+(x2+4y)dy has F(0,0)=0. The value of F(1,2) is ______.
    Answer and solution

    Answer: 13

    Integrate M in x: F=x2y+3x+g(y). Then Fy=x2+g′(y) must equal x2+4y, so g=2y2 (plus a constant, zero here). F=x2y+3x+2y2, and F(1,2)=2+3+8=13. Any path from (0,0) to (1,2) gives the same answer.

    Why people get this wrong: Adding the x-integral and the y-integral separately counts the shared term x2y twice.

  24. MCQExam level

    For spherical coordinates (r,θ,ϕ) with θ measured from the z-axis, the Jacobian ∂(x,y,z)∂(r,θ,ϕ) is
    1. r2sin⁡θ
    2. r2cos⁡θ
    3. rsin⁡θ
    4. r2
    Answer and solution

    Answer: A

    The three edges of the small volume cell are dr, rdθ and rsin⁡θdϕ, at right angles. Their product is r2sin⁡θdrdθdϕ, so the Jacobian is r2sin⁡θ (the determinant gives the same).

    Why people get this wrong: r2cos⁡θ appears if θ is measured from the xy-plane (latitude) instead of from the z-axis.

  25. NATExam level

    The minimum value of f(x,y)=x2+xy+y2−3x is ______.
    Answer and solution

    Answer: -3

    Set fx=2x+y−3=0 and fy=x+2y=0. So x=−2y, then −3y=3: y=−1, x=2. Hessian: fxx=2>0 and D=2⋅2−1=3>0, a minimum. f(2,−1)=4−2+1−6=−3.

    Why people get this wrong: Reporting the location (2,−1) or forgetting the cross term xy when evaluating.

  26. MSQExam level

    For f(x,y)=x3−3x+y2, which statements are true?
    1. (−1,0) is a local maximum.
    2. (1,0) is a local minimum.
    3. (−1,0) is a saddle point.
    4. f has exactly two stationary points.
    Answer and solution

    Answer: B, C, D

    fx=3x2−3=0 gives x=±1; fy=2y=0 gives y=0. Two stationary points. fxx=6x, fyy=2, fxy=0. At (1,0): D=12>0 and fxx>0: minimum. At (−1,0): D=−12<0: saddle, not a maximum (along y the function rises).

    Why people get this wrong: Looking only along x: in one variable x3−3x has a maximum at −1, but y2 makes it rise the other way.

  27. NATExam level

    Positive numbers x, y, z satisfy x+y+z=12. The largest possible value of xyz is ______.
    Answer and solution

    Answer: 64

    Lagrange: ∇(xyz)=λ∇(x+y+z) gives yz=xz=xy=λ, so x=y=z=4 and xyz=64. (AM-GM confirms it: xyz3≤(x+y+z)/3=4.)

    Why people get this wrong: Choosing unequal numbers such as 3,4,5 (product 60) by guesswork.

  28. MCQExam level

    With new variables u=x+y and v=x−y, the operator ∂2f∂x2−∂2f∂y2 becomes
    1. 2∂2f∂u∂v
    2. ∂2f∂u2−∂2f∂v2
    3. 0
    4. 4∂2f∂u∂v
    Answer and solution

    Answer: D

    Chain rule: ∂x=∂u+∂v and ∂y=∂u−∂v. So fxx=fuu+2fuv+fvv and fyy=fuu−2fuv+fvv. The difference is 4fuv. This is how the wave equation is solved by d'Alembert: fuv=0 means f=F(u)+G(v).

    Why people get this wrong: Forgetting the cross term 2fuv in each square, or getting a factor 2.

  29. NATExam level

    For u=x2−y2 and v=2xy, the Jacobian ∂(u,v)∂(x,y) at (1,2) is ______.
    Answer and solution

    Answer: 20

    |2x−2y2y2x|=4x2+4y2=4(x2+y2). At (1,2): 4×5=20. (This is the map z→z2, which stretches areas by |2z|2.)

    Why people get this wrong: A sign slip in the determinant gives 4(x2−y2)=−12.

  30. MCQExam level

    Using (∂U/∂V)T=T(∂P/∂T)V−P, find (∂U/∂V)T for one mole of a van der Waals gas, (P+a/V2)(V−b)=RT.
    1. 0
    2. −a/V2
    3. RT/(V−b)
    4. a/V2
    Answer and solution

    Answer: D

    P=RTV−b−aV2, so (∂P/∂T)V=RV−b. Then TRV−b−P=RTV−b−RTV−b+aV2=aV2. Positive: pulling the molecules apart against their attraction raises the internal energy.

    Why people get this wrong: 0 is the ideal-gas answer; the b term drops out but the a term does not.

← All topics · Timed mock test →