19.1 Semiconductors and the p–n Junction
Energy bands explain why copper conducts, glass does not and silicon sits in between. A few impurity atoms per million turn silicon into n-type or p-type material obeying $n_en_h=n_i^2$, and joining the two makes a p–n junction whose barrier lets current pass one way only.
Every phone, laptop, LED bulb and solar panel is built from one surprising material: silicon, the main element of sand, made extremely pure and then deliberately "spoiled" with a few atoms of phosphorus or boron per million. Pure silicon at room temperature conducts electricity about a million million times worse than copper, yet a trace of the right impurity makes it conduct a million times better than pure silicon — and the impurity decides whether the current is carried by electrons or by "holes". Put a piece of each kind side by side and the boundary between them behaves like a one-way valve for current: the semiconductor diode.
In this section we build that story from the ground up. Energy bands explain why some solids are conductors, some insulators and some semiconductors; the idea of a hole explains how a semiconductor conducts; doping and the law $n_en_h=n_i^2$ give n-type and p-type material; and the p–n junction, with its depletion layer and barrier potential, explains the diode's I–V characteristic in forward and reverse bias. Section 19.2 then puts the diode to work in circuits.
Energy bands: conductors, insulators and semiconductors
An isolated atom has sharp energy levels. Bring $N$ identical atoms close together to form a crystal and the outer electrons of neighbouring atoms begin to feel each other's nuclei. Each atomic level then splits into $N$ levels, so close together (a crystal has some $10^{23}$ atoms per cubic centimetre) that they form an almost continuous band of allowed energies. Between the bands lie energy ranges that no electron in the crystal can have. For conduction, only the two highest bands matter.
In a solid, the valence band is the band of energies occupied by the valence (bonding) electrons; at absolute zero it is the highest band that is completely or partly filled. The conduction band is the next band above it; electrons in it are free to move through the crystal. The energy difference between the top of the valence band $E_v$ and the bottom of the conduction band $E_c$ is the band gap (forbidden gap)
$$E_g=E_c-E_v .$$No electron of the crystal can have an energy inside the gap. $E_g$ is usually quoted in electronvolts.
An electron can carry current only if it can pick up a little energy from the applied field, which means it needs empty levels just above it. A completely full band therefore carries no current, however strong the field — every electron's neighbour level is already taken. This one idea sorts all solids into three groups.
- Conductors (metals): the conduction band is partly filled, or the valence and conduction bands overlap ($E_g=0$). Plenty of electrons have empty levels next to them, so conduction is easy: resistivity $\rho\sim10^{-8}$–$10^{-6}\ \Omega\,\text{m}$.
- Insulators: the valence band is full, the conduction band is empty and the gap is large, $E_g\gt3\ \text{eV}$ (diamond $5.4\ \text{eV}$, glass and quartz about $9\ \text{eV}$). Thermal energy cannot lift electrons across: $\rho\sim10^{11}$–$10^{19}\ \Omega\,\text{m}$.
- Semiconductors: the same arrangement but with a small gap, $E_g\lt3\ \text{eV}$ (silicon $1.1\ \text{eV}$, germanium $0.7\ \text{eV}$, gallium arsenide $1.4\ \text{eV}$). At $0\ \text{K}$ a semiconductor is an insulator; at room temperature a few electrons are thermally excited across the gap: $\rho\sim10^{-5}$–$10^{6}\ \Omega\,\text{m}$.
Temperature: in a metal the number of free electrons is fixed and heating only increases collisions, so $\rho$ rises with temperature (positive temperature coefficient). In a semiconductor heating excites many more electrons across the gap, which far outweighs the extra collisions, so $\rho$ falls steeply with temperature (negative temperature coefficient).
Why does a gap of only $1.1\ \text{eV}$ make such a difference? At room temperature ($T=300\ \text{K}$) the typical thermal energy is $kT=(8.62\times10^{-5}\ \text{eV/K})(300\ \text{K})=0.026\ \text{eV}$, about $\tfrac1{40}\ \text{eV}$. To cross the gap of silicon an electron needs some $40$ times this — rare, but with $5\times10^{28}$ atoms per cubic metre "rare" still means about $10^{16}$ free electrons per cubic metre. For diamond the gap is $200$ times $kT$ and essentially no electron makes it.
A photon can also lift an electron across the gap, but only if its energy is at least $E_g$. Using $hc=1240\ \text{eV nm}$, the longest wavelength a semiconductor can absorb in this way is
$$\lambda_{\max}=\frac{hc}{E_g},\qquad \lambda_{\max}\,(\text{nm})=\frac{1240}{E_g\,(\text{eV})}.$$This threshold, the same idea as the threshold wavelength of the photoelectric effect in Section 17.1, decides which light a photodiode or solar cell can use and what colour an LED emits (Section 19.2).
The resistance of a pure silicon rod is measured at $20^\circ\text{C}$ and again at $80^\circ\text{C}$. Compared with a copper wire heated through the same range, the silicon rod's resistance
Intrinsic and extrinsic semiconductors
Silicon and germanium have four valence electrons. In the crystal each atom shares one electron with each of its four nearest neighbours, forming four covalent bonds; at $0\ \text{K}$ every valence electron is locked in a bond (the valence band is full) and nothing can conduct. At room temperature thermal vibration breaks a few bonds. The freed electron wanders through the crystal as a conduction electron; the broken bond it leaves behind is a vacancy that a neighbouring bonded electron can jump into, which moves the vacancy to the neighbour's bond. The vacancy behaves exactly like a particle of charge $+e$ moving through the crystal.
A hole is a vacancy in the valence band (a missing bonding electron). It acts as a mobile carrier of charge $+e$: when valence electrons move one way to fill vacancies, the holes move the other way.
An intrinsic semiconductor is a perfectly pure semiconductor. Every conduction electron in it comes from a broken bond and leaves exactly one hole, so
$$n_e=n_h=n_i ,$$where $n_e$ and $n_h$ are the numbers of free electrons and holes per unit volume and $n_i$ is the intrinsic carrier concentration. $n_i$ depends on the material and rises steeply with temperature; at $300\ \text{K}$, $n_i\approx1.5\times10^{16}\ \text{m}^{-3}$ for silicon and $2.4\times10^{19}\ \text{m}^{-3}$ for germanium.
Electron–hole pairs are created all the time (generation) and disappear when a free electron falls back into a hole (recombination). In equilibrium the two rates are equal and $n_i$ stays constant. In an electric field the electrons drift against the field and the holes along it, so both currents are in the direction of the field and add: $I=I_e+I_h$. Using the drift-velocity picture of Section 12.1 for each carrier separately gives the conductivity.
If electrons and holes have mobilities $\mu_e$ and $\mu_h$ (drift speed per unit field), the current density in a field $E$ is $j=e(n_e\mu_e+n_h\mu_h)E$, so the conductivity and resistivity are
$$\sigma=\frac1\rho=e\,(n_e\mu_e+n_h\mu_h).$$Electrons are usually more mobile than holes (silicon at $300\ \text{K}$: $\mu_e\approx0.135$, $\mu_h\approx0.048\ \text{m}^2\text{V}^{-1}\text{s}^{-1}$). In a doped sample the majority-carrier term dominates.
An intrinsic semiconductor conducts far too poorly to be useful, and its conductivity is at the mercy of temperature. The cure is doping: adding a tiny, controlled amount of a suitable impurity — typically one atom in $10^6$ to $10^8$ — which changes the number of carriers enormously without disturbing the crystal.
An extrinsic semiconductor is a semiconductor doped with impurity atoms.
- n-type: doped with a pentavalent impurity (phosphorus P, arsenic As, antimony Sb). Four of its five valence electrons form bonds; the fifth is so loosely held (about $0.01$–$0.05\ \text{eV}$) that it is freed at room temperature, leaving a fixed positive ion. The impurity is a donor. Electrons are the majority carriers, holes the minority carriers.
- p-type: doped with a trivalent impurity (boron B, aluminium Al, gallium Ga, indium In). It has only three electrons for four bonds, so it readily takes an electron from a neighbouring bond, creating a hole and becoming a fixed negative ion. The impurity is an acceptor. Holes are the majority carriers, electrons the minority carriers.
In the band picture, donors add a level just below the conduction band (the donor level) and acceptors add a level just above the valence band (the acceptor level). A doped crystal is still electrically neutral: each mobile electron or hole is balanced by a fixed ion.
Doping adds majority carriers, but what happens to the minority carriers? More free electrons mean more chances for a hole to meet an electron and recombine, so the hole population falls. The two effects are tied together by a remarkably simple law.
In a semiconductor in thermal equilibrium at temperature $T$ (no light, no current being forced through), whether pure or doped,
$$n_e\,n_h=n_i^2 ,$$where $n_i$ is the intrinsic concentration of the same material at the same temperature.
Why: the recombination rate is proportional to $n_en_h$ (an electron must meet a hole), while the generation rate depends only on the material and $T$. In equilibrium they are equal, so $n_en_h$ is fixed by $T$ alone — and in the pure crystal it equals $n_i\cdot n_i$.
Doped crystals: with donor concentration $N_D\gg n_i$, every donor is ionised and $n_e\approx N_D$, so $n_h=n_i^2/N_D$. With acceptor concentration $N_A\gg n_i$: $n_h\approx N_A$, $n_e=n_i^2/N_A$. If both are present the larger one wins and only the difference counts ($n_e\approx N_D-N_A$ when $N_D\gt N_A$).
Silicon at $300\ \text{K}$ ($n_i=1.5\times10^{16}\ \text{m}^{-3}$) is made n-type with $N_D=10^{22}\ \text{m}^{-3}$. How many holes per cubic metre are left?
A sample of silicon is doped with boron. Which statement about the sample is correct?
The p–n junction: depletion layer and barrier potential
A p–n junction is a single crystal that is p-type on one side and n-type on the other, made by diffusing or implanting acceptors into part of an n-type wafer (or the reverse). It is not two pieces pressed together: the crystal lattice runs straight through the boundary.
Consider what happens the moment the junction is formed. The n side is crowded with free electrons and the p side with holes, so — exactly like ink spreading in water — electrons diffuse from n to p and holes from p to n. Each electron that crosses soon recombines with a hole on the p side, and each hole that crosses recombines with an electron on the n side. The diffusion therefore strips a thin layer on each side of its mobile carriers and uncovers the fixed ions: positive donor ions on the n side and negative acceptor ions on the p side. These two layers of charge create an electric field pointing from n to p, which pushes back against further diffusion. It also sweeps any minority carriers that wander into it across the junction — electrons from p to n and holes from n to p — the drift current.
The depletion region (depletion layer) of a p–n junction is the thin region on either side of the junction that contains no mobile charge carriers, only the immobile ions — negative acceptor ions on the p side and positive donor ions on the n side. Its width is typically $0.1$–$1\ \mu\text{m}$.
The ions set up an electric field from the n side to the p side, and the n side is at a higher potential than the p side by the barrier potential $V_0$ (about $0.7\ \text{V}$ for silicon and $0.3\ \text{V}$ for germanium at room temperature). An electron going from n to p, or a hole going from p to n, must climb a potential-energy hill of height $eV_0$.
The depletion layer grows until the field it creates is strong enough that the diffusion current of majority carriers (holes p→n, electrons n→p, both a current from p to n) is exactly balanced by the drift current of minority carriers (a current from n to p). With no external voltage the net current is zero.
- The barrier potential exists only inside the crystal. A voltmeter connected across an unbiased diode reads zero: the contact potentials where the metal leads meet the semiconductor cancel it, and an unbiased diode is not a source of energy.
- A heavily doped junction has a narrower depletion layer (the needed ion charge is found in a thinner slice) and therefore a much stronger field — the basis of the Zener diode in Section 19.2.
Even though $V_0$ is less than a volt, the depletion layer is so thin that the field in it is enormous: $0.7\ \text{V}$ across $0.5\ \mu\text{m}$ is on average $1.4\times10^6\ \text{V/m}$, close to the field that makes air break down in a spark.
Not named in the JEE Main 2026 list or the JEE Advanced 2026 syllabus. For interest only; nothing here is used in the exercises, the Exam corner or the quiz.
The depletion layer as a whole is neutral: the negative charge uncovered on the p side equals the positive charge on the n side. If it reaches a distance $x_p$ into the p side and $x_n$ into the n side, then $eN_Ax_p=eN_Dx_n$, i.e. $\dfrac{x_n}{x_p}=\dfrac{N_A}{N_D}$. A junction made of heavily doped p-material ($\text{p}^+$) on lightly doped n-material is depleted almost entirely on the n side. Solving Gauss's law across the layer also shows that the width grows as $\sqrt{V_0-V}$ with the applied voltage $V$ — the rule used to draw the depletion layer in Interactive 2.
In the depletion region of an unbiased p–n junction there are
Forward and reverse bias: the diode's I–V characteristic
A p–n junction with a metal contact on each side is a semiconductor diode. Its circuit symbol is an arrowhead pointing from the p side (the anode) to the n side (the cathode, marked by the bar): the arrow shows the direction in which conventional current flows easily. What happens depends on which way round the external voltage $V$ is applied.
A diode is forward biased when its p side is connected to the higher potential (positive terminal of the battery) and its n side to the lower. It is reverse biased when the p side is at the lower potential.
- Forward bias: the applied voltage opposes the barrier, which falls to $V_0-V$; the depletion layer becomes narrower. Majority carriers diffuse across in large numbers, and the current rises steeply (roughly exponentially) once $V$ approaches $V_0$.
- Reverse bias: the applied voltage adds to the barrier, which rises to $V_0+V$; the depletion layer becomes wider. Diffusion stops; only the drift of minority carriers remains, giving a tiny reverse saturation current $I_0$ (nanoamperes for silicon, microamperes for germanium) that hardly changes with $V$.
- Forward region: almost no current until the knee (cut-in, threshold) voltage — about $0.7\ \text{V}$ for silicon and $0.3\ \text{V}$ for germanium — after which the current rises very steeply (milliamperes). The curve is non-linear: the diode is a non-ohmic device.
- Reverse region: a small, nearly constant current $-I_0$ carried by minority carriers. $I_0$ depends on how many minority carriers are generated, so it increases with temperature (and with light falling on the junction), not with the reverse voltage.
- Breakdown: at a large enough reverse voltage $V_{br}$ (tens to hundreds of volts for an ordinary diode) the reverse current increases suddenly. An ordinary diode is destroyed by the heating unless the current is limited; the Zener diode (Section 19.2) is designed to work there.
- Model of the curve: below breakdown the characteristic is well described by $I=I_0\left(e^{eV/\eta kT}-1\right)$, with $\eta$ between 1 and 2. At room temperature $kT/e\approx25\ \text{mV}$, so in forward bias the current is multiplied by $e\approx2.7$ for every extra $25$–$50\ \text{mV}$.
Because the curve is not a straight line, "the resistance of a diode" needs care. Two different quantities are used.
At an operating point $(V,I)$ on the characteristic, the static (d.c.) resistance is $R=\dfrac{V}{I}$, and the dynamic (a.c.) resistance is the reciprocal of the slope of the curve,
$$r_d=\frac{\Delta V}{\Delta I}.$$In forward bias above the knee $r_d$ is a few ohms and falls as the current rises; in reverse bias it is many megohms.
- Check the scales: forward current is usually in mA and reverse current in μA or nA, and the two voltage axes are often drawn to different scales.
- For dynamic resistance take two nearby points on the same straight part of the curve and divide $\Delta V$ by $\Delta I$ (convert mA to A). Do not use $V/I$.
- For "which diode / which bias" questions, first decide which terminal is at the higher potential, then whether that is the p side.
- If the model $I=I_0(e^{V/V_T}-1)$ is given, use ratios: well above the knee $\dfrac{I_2}{I_1}=e^{(V_2-V_1)/V_T}$, so $\Delta V=V_T\ln\dfrac{I_2}{I_1}$; in reverse bias $I\approx-I_0$.
The diode of Example 6 carries $1.63\ \text{mA}$ at $+0.60\ \text{V}$. Raise the bias to $+0.70\ \text{V}$ — one-sixth more voltage. The current becomes about
1. "n-type is negative, p-type is positive". Both are electrically neutral. "n" and "p" name the sign of the majority carriers; every carrier is balanced by a fixed ion.
2. Getting the minority concentration wrong. Majority $\approx$ doping concentration; minority $=n_i^2/\text{majority}$ — always less than $n_i$. Do not add $n_i$ to the minority carriers, and do not set them to zero.
3. Width and bias. Forward bias narrows the depletion layer and lowers the barrier to $V_0-V$; reverse bias widens it and raises the barrier to $V_0+V$. Swapping these is the most common slip.
4. What carries the reverse current. It is the drift of minority carriers; it is almost independent of the reverse voltage but rises with temperature. The forward current is due to majority carriers diffusing.
5. Static versus dynamic resistance. $r_d=\Delta V/\Delta I$ from the slope; $V/I$ is the static value. Convert mA to A before dividing.
The reverse bias across a silicon diode is increased from $2\ \text{V}$ to $10\ \text{V}$ (well below breakdown), at constant temperature. The reverse current
Exercises
Take $e=1.6\times10^{-19}\ \text{C}$, $m_e=9.1\times10^{-31}\ \text{kg}$, $k=8.62\times10^{-5}\ \text{eV/K}$, $hc=1240\ \text{eV nm}$. For silicon at $300\ \text{K}$: $n_i=1.5\times10^{16}\ \text{m}^{-3}$ and $5.0\times10^{28}$ atoms per $\text{m}^3$ unless a part says otherwise. Assume every impurity atom is ionised.
- 1 Find $kT$ in electronvolts at $300\ \text{K}$. How many times larger is the band gap of silicon ($1.1\ \text{eV}$)?
- 2 Germanium has $E_g=0.67\ \text{eV}$. What is the longest wavelength of light that can create electron–hole pairs in it? In which part of the spectrum is it?
- 3 A semiconductor absorbs light strongly only for wavelengths shorter than $500\ \text{nm}$. Estimate its band gap.
- 4 Classify each solid as a conductor, semiconductor or insulator: (a) bands overlap; (b) $E_g=0.3\ \text{eV}$; (c) $E_g=1.4\ \text{eV}$; (d) $E_g=5.4\ \text{eV}$; (e) $E_g=9\ \text{eV}$.
- 5 Explain, using bands, why the resistance of a copper wire increases when it is heated while that of a germanium crystal decreases.
- 6 Why is a pure semiconductor a perfect insulator at $0\ \text{K}$?
- 7 Find the photon energy of red light of wavelength $700\ \text{nm}$ and use it to explain why silicon absorbs all visible light while diamond ($5.4\ \text{eV}$) absorbs none.
- 1 Pure germanium at $300\ \text{K}$ has $n_i=2.4\times10^{19}\ \text{m}^{-3}$. What are $n_e$ and $n_h$?
- 2 Silicon is doped with phosphorus at one atom per $10^7$ silicon atoms. Find $N_D$ and the hole concentration.
- 3 State the type of semiconductor and its majority carriers when silicon is doped with (a) boron, (b) antimony, (c) gallium, (d) arsenic.
- 4 A p-type silicon sample has $n_h=9.0\times10^{21}\ \text{m}^{-3}$. Find its electron concentration.
- 5 A student says: "An n-type crystal has extra electrons, so it must be negatively charged." Correct the statement.
- 6 An n-type silicon sample has $n_e=5.0\times10^{21}\ \text{m}^{-3}$ and $\mu_e=0.135\ \text{m}^2\text{V}^{-1}\text{s}^{-1}$. Neglecting holes, find its resistivity.
- 7 A germanium sample ($n_i=2.4\times10^{19}\ \text{m}^{-3}$) has $n_e=7.2\times10^{19}\ \text{m}^{-3}$. Find $n_h$, and the conductivity if $\mu_e=0.39$ and $\mu_h=0.19\ \text{m}^2\text{V}^{-1}\text{s}^{-1}$.
- 8 Why does the band diagram of an n-type semiconductor show a donor level just below the conduction band, and what does its small distance from the band ($\sim0.05\ \text{eV}$) tell you?
- 9 A silicon crystal contains $N_D=1.0\times10^{21}\ \text{m}^{-3}$ donors and $N_A=4.0\times10^{21}\ \text{m}^{-3}$ acceptors. State its type and find $n_h$ and $n_e$.
- 1 When a p–n junction is formed, in which direction do (a) electrons diffuse, (b) holes diffuse, (c) the drift current flow? What finally stops the growth of the depletion layer?
- 2 A silicon junction has a barrier potential of $0.70\ \text{V}$ and a depletion layer $0.50\ \mu\text{m}$ wide. Find the average electric field in the depletion layer.
- 3 Which side of the depletion layer is positively charged, and which side is at the higher potential?
- 4 How much energy, in eV and in joules, must a conduction electron from the n side have to cross into the p side of the junction in part 2?
- 5 An electron on the n side moves straight towards a junction of barrier potential $0.60\ \text{V}$ with speed $8.0\times10^5\ \text{m/s}$. Does it cross? Find its speed on the p side.
- 6 How does the width of the depletion layer of a junction made from very heavily doped material compare with that of a lightly doped one? What does this do to the electric field in it?
- 7 A student connects a sensitive voltmeter across an unbiased silicon diode, hoping to read $0.7\ \text{V}$. What does the meter read, and why?
- 1 For forward bias, to which terminal of the battery is the p side connected? What happens to the barrier height and the width of the depletion layer?
- 2 A silicon junction ($V_0=0.70\ \text{V}$) has $0.30\ \text{V}$ forward bias applied. Find the barrier height. What is it with $4.0\ \text{V}$ reverse bias?
- 3 A diode carries $5.0\ \text{mA}$ at $0.65\ \text{V}$ and $25\ \text{mA}$ at $0.75\ \text{V}$. Find its dynamic resistance in this range.
- 4 For the diode of part 3, find the static resistance at $0.75\ \text{V}$.
- 5 In reverse bias the same diode carries $2.0\ \mu\text{A}$ at $5.0\ \text{V}$ and $2.1\ \mu\text{A}$ at $15\ \text{V}$. Find its dynamic resistance in reverse bias, in megohms.
- 6 A diode obeys $I=I_0(e^{V/V_T}-1)$ with $V_T=0.050\ \text{V}$. Well above the knee, by how much must the forward voltage rise to multiply the current by 100?
- 7 Why does the reverse saturation current of a diode roughly double for every $10^\circ\text{C}$ rise in temperature, while the forward current at a fixed large current is much less affected?
- 8 A germanium and a silicon diode are compared. Which has the lower knee voltage and which the larger reverse current? Explain both using band gaps.
E1. JEE Main · MCQ A silicon crystal with intrinsic carrier concentration $1.5\times10^{16}\ \text{m}^{-3}$ is doped with $3.0\times10^{22}\ \text{m}^{-3}$ phosphorus atoms. The concentration of holes in the doped crystal is (A) $1.5\times10^{16}\ \text{m}^{-3}$ (B) $7.5\times10^{9}\ \text{m}^{-3}$ (C) $3.0\times10^{22}\ \text{m}^{-3}$ (D) $2.0\times10^{6}\ \text{m}^{-3}$
E2. JEE Main · Numerical A semiconductor has a band gap of $2.48\ \text{eV}$. Taking $hc=1240\ \text{eV nm}$, find the longest wavelength, in nanometres, of light that can produce electron–hole pairs in it. Give the answer to the nearest integer.
E3. JEE Main · MCQ When a p–n junction diode is reverse biased, (A) the depletion layer becomes narrower and a large current flows due to majority carriers (B) the depletion layer becomes wider and a large current flows due to majority carriers (C) the depletion layer becomes wider and a small current flows due to minority carriers (D) the barrier potential falls and no current flows
E4. JEE Main · Numerical Pure silicon at $300\ \text{K}$ has equal electron and hole concentrations of $1.5\times10^{16}\ \text{m}^{-3}$. Doping with indium raises the hole concentration to $4.5\times10^{22}\ \text{m}^{-3}$. Find the electron concentration in the doped silicon, in units of $10^{9}\ \text{m}^{-3}$. Give the answer to the nearest integer.
E5. JEE Main · MCQ The current through a diode is $I=I_0\left(e^{eV/kT}-1\right)$ with $kT/e=25\ \text{mV}$. At a forward voltage well above the knee, the voltage is increased by $0.10\ \text{V}$. The current is multiplied by about (A) $55$ (B) $4$ (C) $2.7$ (D) $400$
E6. JEE Main · Numerical In a pure semiconductor at room temperature, $n_e=n_h=2.5\times10^{19}\ \text{m}^{-3}$. The electron and hole mobilities are $0.40$ and $0.20\ \text{m}^2\text{V}^{-1}\text{s}^{-1}$. Taking $e=1.6\times10^{-19}\ \text{C}$, find its resistivity in $\Omega\,\text{cm}$. Give the answer to the nearest integer.
Research question. Can an ordinary silicon diode be used as a thermometer — how does its forward voltage at a fixed current change with temperature, and how does its reverse current change?
Why it matters. The temperature sensors inside phone processors, laptop chargers, car engines and medical thermometers are mostly p–n junctions: at a constant current the forward voltage falls by about $2\ \text{mV}$ for every kelvin. Engineers designing solar panels and power electronics in hot climates must also allow for the rapid rise of reverse (leakage) current with temperature. Measuring both effects shows the band picture at work.
What to do.
- Connect a small silicon diode (such as a 1N4148) in forward bias in series with a $4.7\ \text{k}\Omega$ resistor and a $9\ \text{V}$ battery, so the current is about $2\ \text{mA}$. Measure the diode voltage with a multimeter.
- Seal the diode in a thin plastic bag and put it, with a thermometer, in water at about $20$, $35$, $50$, $65$ and $80^\circ\text{C}$ (cool water added to a kettle's output). Record the diode voltage at each temperature.
- Plot $V$ against $T$. Is it a straight line? Find the slope in $\text{mV/K}$.
- If you have a meter with a microampere range, reverse-bias the diode at $9\ \text{V}$ and record the reverse current at the same temperatures. Plot $\ln I_0$ against $T$ and estimate how many kelvin it takes for $I_0$ to double.
- Use your calibration line as a thermometer: measure the voltage with the diode held in your hand and read off the temperature.
Tools: Silicon diode, resistor, $9\ \text{V}$ battery, digital multimeter, thermometer, kettle, beaker, plastic bag, spreadsheet.
Hand in: The $V$–$T$ graph with its slope, the reverse-current data (if measured), your hand-temperature reading, and a paragraph explaining both trends with energy bands and carrier numbers.
Going further. Repeat at a current ten times larger. Does the slope change? Compare your diode with a germanium diode or an LED and relate the differences to their band gaps.
🧪 Self-check quiz
Q1. In a semiconductor at room temperature, the band gap is
Q2. Germanium is doped with arsenic. The majority carriers and the charge of the fixed impurity ions are
Q3. In a doped semiconductor at a fixed temperature, the electron concentration is made 100 times larger by extra donors. The hole concentration becomes
Q4. When a p–n junction is forward biased,
Q5. A diode carries $10\ \text{mA}$ at $0.70\ \text{V}$ and $20\ \text{mA}$ at $0.72\ \text{V}$. Its dynamic resistance in this range is