Home / Courses / CSIR NET Physics / Practice / Semiconductors and the Fermi level

Semiconductors and the Fermi level

Solid state physics and devices · 20 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

    Perfectly pure silicon at absolute zero temperature is
    1. a good conductor, because silicon has four valence electrons
    2. an insulator
    3. a superconductor
    4. n-type, because electrons are lighter than holes
    Answer and solution

    Answer: B

    At 0 K every valence-band state is full and every conduction-band state is empty; with a gap of about 1.17 eV and no thermal energy, no electron can be excited. A full band carries no current, so the crystal insulates. Conduction appears only when temperature, light or impurities supply carriers.

    Why people get this wrong: A semiconductor is not "half a conductor" at all temperatures: its conductivity comes entirely from excited or added carriers.

  2. NATStandard

    Electrons in a semiconductor have mobility 1400 cm2/V s and conductivity effective mass 0.26me. Their mean time between scattering events is ______ ps (to three decimal places).
    Answer and solution

    Answer: 0.207 ps (±0.003)

    From the Drude picture μ=eτ/m∗, so τ=μm∗e=(0.140 m2/V s)(0.26×9.11×10−31 kg)1.60×10−19 C=2.07×10−13 s =0.207 ps.

    Why people get this wrong: Leaving the mobility in cm2/V s gives an answer 104 times too big; use 1 cm2=10−4 m2.

  3. MCQStandard

    In a non-degenerate n-type semiconductor at 300 K, the donor density (all ionised, much larger than ni) is multiplied by 10. The Fermi level
    1. moves down by about 60 meV
    2. moves up by about 26 meV
    3. moves up by about 0.6 eV
    4. moves up by about 60 meV
    Answer and solution

    Answer: D

    n=Nce−(Ec−EF)/kT≈Nd, so Ec−EF=kTln⁡(Nc/Nd). Multiplying Nd by 10 lowers Ec−EF by kTln⁡10=0.02585×2.303=0.0595 eV: the Fermi level rises about 60 meV towards the conduction band.

    Why people get this wrong: Using kT itself (26 meV): a factor e, not 10, moves it by kT.

  4. NATExam level

    Electrons in silicon at 300 K have conductivity mass 0.26me and low-field mobility 1400 cm2/V s. Taking their thermal speed as 3kBT/m∗, the electric field at which the (low-field) drift speed would reach 10% of the thermal speed is ______ kV/cm (to two decimal places).
    Answer and solution

    Answer: 1.64 kV/cm (±0.02)

    vth=3(1.381×10−23)(300)0.26×9.11×10−31=2.29×105 m/s. Drift speed vd=μE, so E=0.1vthμ=2.29×1040.140=1.64×105 V/m =1.64 kV/cm. Beyond fields of this order the simple vd=μE fails and the drift speed saturates near 105 m/s.

    Why people get this wrong: Converting V/m to kV/cm: 1 kV/cm =105 V/m, not 103.

  5. MSQExam level

    Which statements about the Hall effect in semiconductors are correct?
    1. An intrinsic semiconductor whose electrons are more mobile than its holes has a negative Hall coefficient.
    2. A p-type sample can show a Hall coefficient that changes sign as it is heated into the intrinsic range.
    3. At the same current and field, a thicker sample (along B→) gives a larger Hall voltage.
    4. For a single carrier type, |RH|σ gives the carrier mobility.
    Answer and solution

    Answer: A, B, D

    With both carriers RH=pμp2−nμn2e(pμp+nμn)2. Intrinsic (n=p) with μn>μp gives RH<0. In a p-type sample RH>0 at first, but once heating makes n≈p the more mobile electrons win and RH turns negative. The Hall voltage is VH=IBnet, smaller for a thicker sample t. For one carrier, RH=1/ne and σ=neμ, so |RH|σ=μ (the Hall mobility).

    Why people get this wrong: Assuming the sign of RH is fixed by the doping: it is set by pμp2−nμn2, which temperature can flip.

  6. MCQExam level

    As the temperature of silicon rises from near 0 K to 300 K, its band gap
    1. decreases slightly, from about 1.17 eV to 1.12 eV
    2. stays exactly constant, since it is a property of the crystal
    3. increases from 1.12 eV to about 1.4 eV
    4. falls to zero, which is why the conductivity rises
    Answer and solution

    Answer: A

    Thermal expansion and electron-phonon interaction shift the band edges so that the gap shrinks a little with temperature (an empirical fit is the Varshni formula). For silicon the change from 0 to 300 K is only about 0.05 eV. The large rise in intrinsic conductivity comes from the factor e−Eg/2kT, not from the small change in Eg.

    Why people get this wrong: Blaming the conductivity rise on a closing gap: the Boltzmann factor does almost all the work.

  7. NATHard

    The effective density of states of a band with effective mass m∗=me at 300 K is Nc=2(2πm∗kTh2)3/2. Its value is ______ ×1019 cm−3 (to two decimal places).
    Answer and solution

    Answer: 2.51 ×10¹⁹ cm⁻³ (±0.02)

    2πmekTh2=2π×9.109×10−31×1.381×10−23×300(6.626×10−34)2=5.40×1016 m−2. Raised to the power 3/2: 1.255×1025 m−3; times 2 (spin): 2.51×1025 m−3 =2.51×1019 cm−3. For another mass multiply by (m∗/me)3/2.

    Why people get this wrong: Forgetting the factor 2 for the two spin states, or the 106 between m−3 and cm−3.

  8. NATHard

    n-type silicon (εr=11.7) has 1016 cm−3 ionised donors at 300 K. Its extrinsic Debye length LD=εkBTe2Nd is ______ nm (to one decimal place).
    Answer and solution

    Answer: 40.9 nm (±0.4)

    Nd=1016 cm−3=1022 m−3. εkBTe2Nd=(11.7)(8.854×10−12)(0.02585 V)(1.602×10−19)(1022)=1.67×10−15 m2 (using kBT/e=0.02585 V), so LD=4.09×10−8 m =40.9 nm. Charge imbalances smaller than this are smeared out; it also sets how sharply a depletion edge fades.

    Why people get this wrong: Forgetting the cm−3 to m−3 conversion (factor 106) changes LD by a factor 103.

  9. MCQHard

    n-type silicon with Nd=1016 cm−3, μn=1350 cm2/V s and εr=11.7 has some excess majority charge put into it. The time over which this space charge disappears (the dielectric relaxation time) is about
    1. 0.5 μs
    2. 1 ns
    3. 0.5 ps
    4. 5 fs
    Answer and solution

    Answer: C

    Gauss's law plus Ohm's law give ∂ρ/∂t=−(σ/ε)ρ, so τd=ε/σ. σ=eNdμn=(1.602×10−19)(1022)(0.135)=216 S/m, and ε=11.7×8.854×10−12=1.04×10−10 F/m. τd=4.8×10−13 s ≈0.5 ps. Majority carriers neutralise charge almost instantly; minority carriers, which must recombine, live for microseconds.

    Why people get this wrong: Confusing dielectric relaxation (majority carriers, picoseconds) with minority-carrier lifetime (recombination, often microseconds).

  10. MSQHard

    Which statements about recombination in semiconductors are correct?
    1. Shockley-Read-Hall recombination through traps is most effective for trap levels near the middle of the gap.
    2. Auger recombination grows rapidly with carrier density and dominates in very heavily doped or strongly injected material.
    3. For radiative recombination at low-level injection, the minority lifetime does not depend on the majority carrier density.
    4. In direct-gap materials such as GaAs, radiative band-to-band recombination is important.
    Answer and solution

    Answer: A, B, D

    In a direct gap an electron and hole can recombine and emit a photon without a phonon, so radiative recombination is fast (hence LEDs and lasers). SRH capture is most efficient when the trap is equally easy to fill from either band, i.e. near midgap. Auger involves three carriers, rate ∝n2p or np2, so it rises steeply at high density. Radiative rate =B(np−ni2); for holes in n-type, τp=1/(Bn0), which falls as the doping rises.

    Why people get this wrong: Treating lifetime as a fixed material constant: it depends on doping, trap density and injection level.

  11. NATTop rank

    Silicon is doped with 1015 donors per cm3. Take ni=1010 cm−3 at 300 K, Eg=1.12 eV and ni∝e−Eg/2kT (ignore the T3/2 factor). The temperature at which ni equals Nd is ______ K (to the nearest kelvin). Take k=8.617×10−5 eV/K.
    Answer and solution

    Answer: 640 K (±4)

    ni(T)ni(300)=exp⁡[Eg2k(1300−1T)]=105. So 1300−1T=ln⁡105Eg/2k=11.516499=1.772×10−3 K−1, giving 1T=3.333×10−3−1.772×10−3=1.562×10−3 and T=640 K. Above about this temperature the device loses its doping-controlled behaviour.

    Why people get this wrong: Solving e−Eg/2kT=10−5 directly, without referring to the known value at 300 K, leaves out the prefactor.

  12. NATTop rank

    At a high temperature where kT=0.050 eV, a semiconductor with Nd donors per unit volume (all ionised) has ni=Nd. The Fermi level lies above the intrinsic level by ______ meV (to one decimal place).
    Answer and solution

    Answer: 24.1 meV (±0.2)

    Neutrality gives n=Nd2+Nd24+ni2=ni(12+54)=1.618ni (the golden ratio). Then EF−Ei=kTln⁡(n/ni)=0.050×ln⁡1.618=0.050×0.4812=0.0241 eV =24.1 meV.

    Why people get this wrong: Taking n=Nd gives ln⁡1=0, as if the sample were intrinsic.

  13. MSQTop rank

    A semiconductor has μn>μp. Its conductivity is varied by doping, always with np=ni2. Which statements are correct?
    1. The lowest possible conductivity occurs in slightly p-type material, with p=niμn/μp.
    2. Intrinsic material does not have the lowest possible conductivity.
    3. The lowest possible conductivity is 2eniμnμp.
    4. The lowest conductivity occurs where n=p.
    Answer and solution

    Answer: A, B, C

    σ(n)=e(nμn+ni2nμp). Setting dσ/dn=0: μn=ni2μp/n2, so n=niμp/μn<ni and p=niμn/μp>ni: slightly p-type (true). Hence intrinsic is not the minimum (true). Substituting: σmin⁡=eni(μnμp+μnμp)=2eniμnμp (true). n=p is the intrinsic point, not the minimum (false).

    Why people get this wrong: Assuming pure material is always the most resistive. Adding a few acceptors removes fast electrons and adds only slow holes.

  14. NATTop rank

    For silicon with μn=1350 and μp=480 cm2/V s, the ratio of the intrinsic conductivity to the lowest conductivity reachable by doping is ______ (to three decimal places).
    Answer and solution

    Answer: 1.137 (±0.002)

    σi=eni(μn+μp) and σmin⁡=2eniμnμp. Ratio =μn+μp2μnμp=18302648000=18301610=1.137: the arithmetic mean over the geometric mean of the mobilities, always at least 1.

    Why people get this wrong: Answering 1, assuming intrinsic is the minimum.

  15. MCQTop rank

    Silicon is doped so heavily that the Fermi level moves into the conduction band. Which statement is correct?
    1. The electron density still equals Nce−(Ec−EF)/kT exactly.
    2. The Boltzmann approximation fails, and n must be found from the full Fermi-Dirac distribution.
    3. The mass action law np=ni2 becomes exact.
    4. The material becomes an insulator, because all states are filled.
    Answer and solution

    Answer: B

    The formula n=Nce−(Ec−EF)/kT assumes Ec−EF≫kT, so that f≈e−(E−EF)/kT. Once EF is above Ec, states near the band edge are mostly full and the Fermi-Dirac form (a Fermi-Dirac integral) is needed. The mass action law also stops being exact, and the gap itself shrinks (band-gap narrowing). Such degenerate material conducts like a poor metal.

    Why people get this wrong: Using the simple exponential formulas outside their range: they overestimate n when EF is near or inside a band.

  16. NATTop rank

    Donors in a semiconductor with εr=12 and m∗=0.20me are modelled as hydrogen atoms. Their orbits begin to overlap and the electrons delocalise (the Mott criterion) when Nd1/3a∗≈0.25, where a∗ is the donor Bohr radius. Taking a0=0.0529 nm, this critical density is ______ ×1017 cm−3 (to two decimal places).
    Answer and solution

    Answer: 4.88 ×10¹⁷ cm⁻³ (±0.06)

    Donor Bohr radius: a∗=a0εrm∗/me=0.0529×120.20=3.175 nm =3.175×10−7 cm. Then Nd=(0.25/a∗)3=(7.87×105 cm−1)3=4.88×1017 cm−3. Above this doping the semiconductor stays conducting even at T→0.

    Why people get this wrong: Using a∗=a0m∗/εr (inverted). A small mass and a large permittivity make the orbit bigger.

  17. NATTop rank

    Silicon's conduction band has 6 equivalent ellipsoidal valleys, each with longitudinal mass ml=0.98me and transverse mass mt=0.19me. Its density-of-states effective mass (which sets Nc) is ______ me (to two decimal places).
    Answer and solution

    Answer: 1.08 $m_e$ (±0.01)

    One ellipsoidal valley has the same density of states as a sphere with mass (mlmt2)1/3=(0.98×0.0361)1/3=0.328me. The density of states goes as m3/2, and six valleys multiply it by 6, so mdos3/2=6(mlmt2)1/2, giving mdos=62/3(mlmt2)1/3=3.30×0.328=1.08me.

    Why people get this wrong: Using the conductivity mass 3/(1/ml+2/mt)=0.26me, or multiplying the single-valley mass by 6 instead of 62/3.

← All topics · Timed mock test →