Home / Courses / GATE Physics / Practice / Fluctuations, diffusion and Brownian motion (NET)

Fluctuations, diffusion and Brownian motion (NET)

Statistical physics · 55 questions and templates for GATE 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 the canonical ensemble the variance of the energy is
    1. ⟨(ΔE)2⟩=kT2CV
    2. ⟨(ΔE)2⟩=kTCV
    3. ⟨(ΔE)2⟩=CV/kT2
    4. ⟨(ΔE)2⟩=k2TCV
    Answer and solution

    Answer: A

    ⟨E⟩=−∂ln⁡Z/∂β and ⟨E2⟩−⟨E⟩2=∂2ln⁡Z/∂β2=−∂⟨E⟩/∂β=kT2∂⟨E⟩/∂T=kT2CV. A response function (heat capacity) measures the size of equilibrium fluctuations.
  2. MCQWarm-up

    A particle takes N independent steps of fixed length l in random directions. Its root-mean-square displacement from the start is
    1. Nl
    2. lN
    3. l/N
    4. lN2
    Answer and solution

    Answer: B

    R→=∑l→i, so ⟨R2⟩=∑i⟨li2⟩+∑i≠j⟨l→i⋅l→j⟩=Nl2+0, since different steps are uncorrelated. Hence Rrms=lN: distance grows as the square root of time, the hallmark of diffusion.
  3. MCQWarm-up

    Particles released at the origin diffuse in one dimension with diffusion constant D. Their mean square displacement after time t is
    1. Dt
    2. 2Dt
    3. 6Dt
    4. (Dt)2
    Answer and solution

    Answer: B

    The diffusion equation ∂tn=D∂x2n with a point start gives the Gaussian n∝e−x2/4Dt/4πDt, with variance 2Dt. In d dimensions ⟨r2⟩=2dDt: 4Dt in 2D and 6Dt in 3D.

    Why people get this wrong: 6Dt is the three-dimensional result.

  4. MCQWarm-up

    A small sphere of radius a in a fluid of viscosity η at temperature T has diffusion constant (Stokes-Einstein)
    1. D=6πηakT
    2. D=kT4πηa
    3. D=kT6πηa
    4. D=6πηakT
    Answer and solution

    Answer: C

    Einstein: D=μkT with mobility μ=1/γ (velocity per unit force). Stokes drag gives γ=6πηa, so D=kT/6πηa. Perrin used this, with measured D, to find Avogadro's number.
  5. NATWarm-up

    A polymer is modelled as a freely jointed chain of 100 segments, each 1.0 nm long. Its rms end-to-end distance is ______ nm.
    Answer and solution

    Answer: 10 (±0.05)

    A freely jointed chain is a random walk: Rrms=bN=1.0×100=10 nm, compared with a contour length of 100 nm.
  6. MCQWarm-up

    For a classical ideal gas in the grand canonical ensemble, the variance of the particle number in a volume is
    1. zero
    2. ⟨N⟩2
    3. ⟨N⟩
    4. ⟨N⟩ (a Poisson distribution)
    Answer and solution

    Answer: D

    In general ⟨(ΔN)2⟩=kT(∂⟨N⟩/∂μ)T,V. For a classical ideal gas ⟨N⟩∝eμ/kT, so the derivative is ⟨N⟩/kT and the variance is ⟨N⟩: the counts are Poisson, with relative spread 1/⟨N⟩.
  7. MCQWarm-up

    For a system of N weakly interacting particles in contact with a heat bath, the relative energy fluctuation ΔE/⟨E⟩
    1. grows like N
    2. falls like 1/N
    3. falls like 1/N
    4. does not depend on N
    Answer and solution

    Answer: C

    ⟨(ΔE)2⟩=kT2CV and both CV and ⟨E⟩ are proportional to N. So ΔE∝N and ΔE/⟨E⟩∝1/N. For N∼1023 this is about 10−12, which is why canonical and microcanonical ensembles agree for large systems.

    Why people get this wrong: The absolute spread grows with N; only the relative spread shrinks.

  8. NATStandard

    A classical monatomic ideal gas of N=1.0×104 atoms is in contact with a heat bath. The relative energy fluctuation ΔErms/⟨E⟩ is ______ ×10−3 (to two decimal places).
    Answer and solution

    Answer: 8.16 (±0.02)

    ⟨E⟩=32NkT and CV=32Nk, so ⟨(ΔE)2⟩=kT2⋅32Nk and ΔErms⟨E⟩=3N/2kT32NkT=23N=6.67×10−5=8.16×10−3. For a mole it is ∼10−12: ensembles agree.
  9. NATStandard

    A sphere of radius 0.50 μm diffuses in water (η=1.0×10−3 Pa s) at 300 K. Its diffusion constant is ______ ×10−13 m2/s (to two decimal places).
    Answer and solution

    Answer: 4.39 (±0.02)

    D=kT6πηa=4.142×10−216π×10−3×5.0×10−7=4.142×10−219.425×10−9=4.39×10−13 m2/s.
  10. NATStandard

    The sphere of the previous question (D=4.39×10−13 m2/s) is watched for 60 s. Its rms displacement along one axis is ______ μm (to one decimal place).
    Answer and solution

    Answer: 7.3 (±0.05)

    xrms=2Dt=2×4.39×10−13×60=5.27×10−11=7.26×10−6 m ≈7.3 μm, easily visible in a microscope: this is what Perrin measured.
  11. NATStandard

    A 1.0 MΩ resistor at 300 K is measured with a bandwidth of 10 kHz. The rms Johnson-Nyquist noise voltage is ______ μV (to one decimal place).
    Answer and solution

    Answer: 12.9 (±0.1)

    Vrms=4kTRΔf=4×1.381×10−23×300×106×104=1.657×10−10=1.29×10−5 V. It is a fluctuation-dissipation result: the same resistance that dissipates also sets the noise.

    Why people get this wrong: Forgetting the 4 halves the answer.

  12. NATStandard

    An AFM cantilever of spring constant 0.10 N/m is at 300 K. Its rms thermal deflection is ______ nm (to two decimal places).
    Answer and solution

    Answer: 0.2 (±0.005)

    Equipartition: 12ks⟨x2⟩=12kT, so xrms=kT/ks=4.142×10−21/0.10=2.04×10−10 m =0.20 nm. AFMs use this to calibrate ks.
  13. NATStandard

    A one-dimensional random walker takes 10 steps, each +1 or −1 with equal probability. The probability that it ends exactly where it started is ______ (to three decimal places).
    Answer and solution

    Answer: 0.246 (±0.001)

    It must take 5 steps right and 5 left: P=(105)/210=252/1024=0.246. The Gaussian approximation P(m)=22πNe−m2/2N (the factor 2 because m has the parity of N) gives 2/20π=0.252.
  14. MCQStandard

    In the Langevin equation mv˙=−γv+F(t), the random force must satisfy ⟨F(t)F(t′)⟩=Γδ(t−t′) with
    1. Γ=2γkT
    2. Γ=γkT/m
    3. Γ=kT/γ
    4. Γ=0
    Answer and solution

    Answer: A

    Solving for v and demanding 12m⟨v2⟩=12kT at long times gives ⟨v2⟩=Γ/2γm=kT/m, so Γ=2γkT. This is the fluctuation-dissipation theorem: friction and noise come from the same molecular collisions.
  15. MSQStandard

    Which of these equilibrium relations between fluctuations and response functions are correct?
    1. ⟨(ΔE)2⟩=kT2CV (canonical)
    2. ⟨(ΔN)2⟩/⟨N⟩=nkTκT (grand canonical)
    3. ⟨(ΔM)2⟩=kTχ (magnet in a field, with χ=∂M/∂H for total moment M)
    4. ⟨(ΔT)2⟩=kT2CV
    Answer and solution

    Answer: A, B, C

    Each conjugate pair gives variance =kT×(susceptibility): energy with β, N with μ (which can be rewritten with the compressibility κT), M with H. Temperature fluctuations (for a small system in contact with a bath, in Einstein's sense) have ⟨(ΔT)2⟩=kT2/CV: inverse heat capacity.
  16. NATStandard

    An ideal gas at 273 K and 1 atm contains 2.69×107 molecules in a cube of side 1.0 μm. The relative rms fluctuation of the number in that cube is ______ ×10−4 (to two decimal places).
    Answer and solution

    Answer: 1.93 (±0.01)

    Poisson statistics: ΔN/N=1/N=1/2.69×107=1.93×10−4. Density fluctuations on the scale of a wavelength scatter light (Rayleigh); near a critical point they grow huge (critical opalescence).
  17. MCQStandard

    Why does the end-point distribution of a long random walk become Gaussian, whatever the distribution of single steps (with finite variance)?
    1. Because each step is Gaussian
    2. By the central limit theorem: the sum of many independent variables with finite variance tends to a Gaussian
    3. Because of energy conservation
    4. It does not: it stays binomial
    Answer and solution

    Answer: B

    The characteristic function of the sum is the product of identical factors, [ϕ(k)]N≈[1−k2σ2/2]N→e−Nk2σ2/2, a Gaussian of variance Nσ2. Steps with infinite variance (Lévy flights) escape the theorem.
  18. MCQStandard

    The erratic motion of pollen grains (Brownian motion) is caused by
    1. convection currents in the water
    2. the grains being alive
    3. light pressure from the microscope lamp
    4. unbalanced impacts of the surrounding molecules, which do not average out for a small particle over short times
    Answer and solution

    Answer: D

    Each grain is hit about 1020 times per second; the net force averages to zero but fluctuates. Einstein (1905) showed the resulting ⟨x2⟩=2Dt with D=kT/6πηa, and Perrin's measurements of it convinced physicists that atoms exist. Dead or inorganic particles move the same way.
  19. NATStandard

    A sphere of radius 1.0 μm and density 1000 kg/m3 is in water at 300 K. Its rms velocity along one axis is ______ mm/s (two decimal places). Use k=1.381×10−23 J/K.
    Answer and solution

    Answer: 0.99 mm/s (±0.01)

    Equipartition: 12m⟨vx2⟩=12kT. m=43π(10−6)3×1000=4.19×10−15 kg, so vrms=kT/m=4.14×10−21/4.19×10−15=9.9×10−4 m/s =0.99 mm/s. This speed lasts only about m/6πηa≈0.2 μs before friction randomises it, so the observed motion is the slow diffusive drift, not this speed.

    Why people get this wrong: Using 3kT (all three components) or forgetting the cube of the radius.

  20. NATExam level

    The noise power a resistor at 300 K can deliver to a matched load is kTΔf. For a 1 Hz bandwidth, expressed in dBm (decibels relative to 1 mW), this is ______ dBm (one decimal place).
    Answer and solution

    Answer: -173.8 dBm (±0.2)

    kT=1.381×10−23×300=4.14×10−21 W per Hz. In dBm: 10log10⁡(4.14×10−21/10−3)=10log10⁡(4.14×10−18)=−173.8 dBm. Radio engineers use "−174 dBm/Hz" as the thermal noise floor.

    Why people get this wrong: 4kTΔf is the open-circuit voltage-noise factor (⟨V2⟩=4kTRΔf); the power into a matched load is a quarter of ⟨V2⟩/R, i.e. kTΔf.

  21. MSQExam level

    Which statements about Brownian diffusion are true?
    1. ⟨x2⟩=2Dt along one axis and ⟨r2⟩=6Dt in three dimensions
    2. By Stokes-Einstein, a sphere of half the radius diffuses twice as fast
    3. Raising the viscosity of the fluid raises D
    4. Brownian paths are continuous but nowhere differentiable
    Answer and solution

    Answer: A, B, D

    Each Cartesian direction contributes 2Dt independently. D=kT/6πηa∝1/a and ∝1/η. Since displacement grows as t, the ratio Δx/Δt∼Δt−1/2 diverges as Δt→0: the path has no velocity at any scale (in the idealised, overdamped limit).

    Why people get this wrong: More viscous fluids slow diffusion; D falls as 1/η.

  22. MCQExam level

    For a current of independent, randomly arriving electrons measured with a fixed bandwidth, the rms shot-noise current grows with the mean current I as
    1. I
    2. I
    3. I2
    4. it does not depend on I
    Answer and solution

    Answer: A

    Poisson arrivals: the variance of the number counted in a time window equals the mean number, so the current variance is 2eIΔf and the rms current is 2eIΔf∝I. The relative noise ∝1/I falls as the current grows.

    Why people get this wrong: The noise power (variance) is linear in I; the rms value is its square root.

  23. MCQExam level

    The Johnson-Nyquist voltage noise of a resistor at 300 K has the flat spectrum 4kTR only up to a frequency where quantum effects cut it off. That frequency is about
    1. 6 THz
    2. 6 MHz
    3. 6 GHz
    4. 60 Hz
    Answer and solution

    Answer: A

    Each mode carries the Planck energy hfehf/kT−1 instead of kT; the two agree only while hf≪kT. The cut-off is at f∼kT/h=1.38×10−23×3006.63×10−34=6×1012 Hz. Electronics never reaches it at room temperature, but at millikelvin temperatures it falls into the GHz range used by quantum circuits.

    Why people get this wrong: The noise is white for all practical electronics, but not to infinite frequency (that would give infinite power).

  24. NATExam level

    A single-particle state of an ideal Bose gas has mean occupation ⟨n⟩=2. The relative fluctuation Δn/⟨n⟩ of its occupation is ______ (three decimal places).
    Answer and solution

    Answer: 1.225 (±0.003)

    For one bosonic mode, P(n)∝e−nx (geometric), which gives ⟨(Δn)2⟩=⟨n⟩(1+⟨n⟩)=2×3=6. So Δn/⟨n⟩=6/2=1.225: larger than 1, the "bunching" of bosons. A fermion mode has ⟨n⟩(1−⟨n⟩) and a classical (Poisson) one ⟨n⟩.

    Why people get this wrong: Using the Poisson value ⟨n⟩, which holds only when ⟨n⟩≪1.

  25. MSQExam level

    The velocity of a free Brownian particle obeying mv˙=−γv+F(t) with white-noise F (an Ornstein-Uhlenbeck process). In equilibrium, which statements are correct?
    1. It is a stationary Gaussian process
    2. It is Markovian: its future depends only on the present velocity
    3. Its correlation time is m/γ
    4. Its variance ⟨v2⟩ grows linearly with time forever
    Answer and solution

    Answer: A, B, C

    A linear equation driven by Gaussian white noise gives a Gaussian process; in equilibrium its statistics do not change in time, with ⟨v2⟩=kT/m fixed. White noise has no memory, so the first-order equation makes v Markovian. ⟨v(0)v(t)⟩=kTme−γt/m, so the correlation time is m/γ. It is the position, not the velocity, whose variance grows linearly.

    Why people get this wrong: Confusing the velocity (bounded variance) with the position (diffusing).

  26. MCQExam level

    An inductor L is part of a circuit in equilibrium at temperature T. The mean square thermal current through it is
    1. kT/2L
    2. kT/L
    3. kTL
    4. kTL
    Answer and solution

    Answer: B

    The magnetic energy 12LI2 is quadratic in one variable, so equipartition gives 12L⟨I2⟩=12kT and ⟨I2⟩=kT/L, whatever the resistors. This is the twin of ⟨V2⟩=kT/C for a capacitor.

    Why people get this wrong: Dropping the 12 on only one side gives kT/2L.

  27. MCQExam level

    A Brownian particle has ⟨vx2⟩1/2=1.0 mm/s and velocity autocorrelation ⟨vx(0)vx(t)⟩=⟨vx2⟩e−t/τ with τ=0.10 μs. Using D=∫0∞⟨vx(0)vx(t)⟩dt (Green-Kubo), its diffusion constant is
    1. 1.0×10−13 m2/s
    2. 2.0×10−13 m2/s
    3. 1.0×10−7 m2/s
    4. 5.0×10−14 m2/s
    Answer and solution

    Answer: A

    D=⟨vx2⟩τ=(10−3)2×10−7=1.0×10−13 m2/s. With ⟨v2⟩=kT/m and τ=m/γ this is the Einstein relation D=kT/γ. The mass cancels: diffusion does not depend on inertia.

    Why people get this wrong: Doubling the integral (as if from −∞ to ∞) gives 2×10−13; that is ⟨x2⟩/t, not D.

  28. MSQExam level

    Which noise spectra are flat (white) at low frequencies?
    1. Johnson noise of a resistor
    2. Shot noise of a current of independent electrons
    3. Flicker (1/f) noise
    4. The thermal voltage across a capacitor C fed by a resistor R, at frequencies far above 1/RC
    Answer and solution

    Answer: A, B

    Johnson noise (4kTR per hertz) and shot noise (2eI per hertz) come from uncorrelated events on short time scales, so they are white up to very high frequencies. Flicker noise grows at low frequency as 1/f. The capacitor voltage is Johnson noise filtered by the RC low-pass: flat below 1/RC, falling as 1/f2 above.

    Why people get this wrong: Total capacitor noise kT/C is finite precisely because the spectrum is cut off above 1/RC.

  29. NATHard

    A sphere of radius 0.50 μm turns randomly in water (η=1.0×10−3 Pa s) at 300 K. Its rotational diffusion constant Dr=kT/8πηa3 is ______ s−1 (two decimal places).
    Answer and solution

    Answer: 1.32 s⁻¹ (±0.02)

    The rotational friction of a sphere is 8πηa3, and Einstein's relation gives Dr=kT/(8πηa3)=4.14×10−218π×10−3×1.25×10−19=4.14×10−213.14×10−21=1.32 s−1. An axis fixed in the sphere loses memory of its direction in about 1/2Dr≈0.4 s.

    Why people get this wrong: Rotational friction scales as a3, not a like translational friction; halving the radius speeds rotation up eightfold.

  30. MCQHard

    A tunnel junction of resistance R at 300 K carries a current I of independent electrons. Its current shot noise (2eI per unit bandwidth) equals the Johnson current noise of the same resistance (4kT/R) when the voltage IR across it is about
    1. 26 mV
    2. 52 mV
    3. 13 mV
    4. 104 mV
    Answer and solution

    Answer: B

    2eI=4kT/R gives IR=2kT/e=2×0.02585 V =51.7 mV. Below this the noise is mainly thermal, above it mainly shot noise. (The full result for a tunnel junction interpolates as 2eIcoth⁡(eV/2kT).)

    Why people get this wrong: kT/e=26 mV is the familiar thermal voltage; the factor 2 comes from 4kT against 2eI.

← All topics · Timed mock test →