Threshold Condition

A laser is an optical oscillator, and the threshold condition is the single line that decides whether it oscillates at all. It is a meeting point of two physics chapters. From the quantum theory of light–matter interaction it inherits the gain: stimulated emission from an inverted medium amplifies a passing wave by $e^{g\ell}$ , where $g=\sigma(\nu)\,\Delta N$ is the emission cross section times the population inversion. From the electromagnetics of resonators it inherits the loss: a real cavity leaks light through its partially transmitting mirrors and scatters and absorbs a little each pass, a drain summarized by the cavity photon lifetime $\tau_c$ . The threshold condition is simply the balance sheet of these two — amplification against leakage — over one complete round trip. Below threshold the losses win: a fluctuation seeded by spontaneous emission decays before it can build, and the device merely glows with weak, incoherent fluorescence. At threshold the two exactly cancel, and a wave reproduces itself after one round trip in both amplitude and phase — a marginally stable oscillator, like a pendulum whose drive just cancels its friction. Above threshold gain beats loss and the field grows out of noise until saturation halts the runaway. Because gain rises with pump, this balance translates into a sharply defined minimum pump power, and the input–output curve shows a distinct kink there. That kink, and the gain clamping that produces the laser's linear power characteristic above it, are the practical face of one self-consistency equation, $R_1R_2\exp[2(g-\alpha)\ell]=1$ : everything about starting a laser, and the power it delivers, is read off from it.

Key takeaways

Laser oscillation requires round-trip gain to exactly balance round-trip loss (mirror transmission, absorption, scattering) — a wave that reproduces itself after one lap.
The amplitude condition $R_1R_2\exp[2(g_\mathrm{th}-\alpha)\ell]=1$ fixes the minimum gain $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ and hence the minimum inversion $\Delta N_\mathrm{th}=g_\mathrm{th}/\sigma$ .
Above threshold the gain clamps: the inversion is pinned at $\Delta N_\mathrm{th}$ however hard you pump, so every extra pump quantum leaves as an output photon and $P_\mathrm{out}$ rises linearly.
The same threshold is a rate, $v_g\,g_\mathrm{th}=1/\tau_c$ , so a high-Q cavity with a long photon lifetime lowers the required gain — good mirrors recycle each photon many times.
Four-level media reach threshold at a tiny inversion; three-level media (ruby) must invert more than half of all active ions first — why one runs CW on a diode and the other needs a flashlamp.
Key idea: laser oscillation begins when the single-pass gain, doubled over a round trip, just compensates every round-trip loss; the self-consistency condition $R_1R_2\exp[2(g-\alpha)\ell]=1$ fixes the threshold gain and clamps the inversion the instant oscillation starts.

The core insight

Laser oscillation begins when the single-pass gain, doubled over a round trip, just compensates all round-trip losses: the self-consistency condition $R_1R_2\exp[2(g-\alpha)\ell]=1$ fixes the threshold gain $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ and clamps the inversion at $\Delta N_\mathrm{th}=g_\mathrm{th}/\sigma$ once oscillation starts.

Where this lives today

The threshold condition is a daily engineering tool. When a company specifies a diode-pumped module, the numbers that dominate the datasheet — turn-on pump power, slope efficiency, the choice of a 2 percent versus a 10 percent output coupler — are direct outputs of $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ and its input–output line. Measuring internal loss uses the condition in reverse: the Findlay–Clunie method plots threshold pump against $\ln(1/R_2)$ for a family of couplers, and the intercept yields the intrinsic loss $\alpha$ no single measurement can isolate. It also governs pulsed operation by design — a Q-switch spoils the cavity to raise $g_\mathrm{th}$ above the stored gain, then restores high $Q$ so the collapsed threshold unleashes a giant pulse. At the frontier the same balance sets the ultralow thresholds of microcavity and photonic-crystal lasers, where extreme- $Q$ resonators push $\Delta N_\mathrm{th}$ so low that a handful of emitters suffice — the "thresholdless" limit nanolaser and cavity-QED research chase.

The physical picture

Follow a single photon making laps of the cavity. On each lap it stimulates the inverted medium to emit copies of itself — the gain — while losing siblings to the leaky mirror, to absorption, and to scattering. Threshold is the break-even lap: the medium replaces exactly as many photons as are lost, so a steady field persists indefinitely. Push the pump higher and replacement outpaces loss and the field grows; ease it lower and it decays into darkness. The first figure makes the knife-edge literal, tracking the intensity around one round trip and showing it return, digit for digit, to where it began.

Figure

Threshold is the break-even round trip: with mirrors R₁ = 1.00 and R₂ = 0.98, each pass through the gain medium multiplies the intensity by e^(gℓ) ≈ 1.010, the 2 percent output coupler removes just as much, so the intensity returns exactly to 1.000 — a self-reproducing wave

The amplitude condition sets how much gain is needed; a companion phase condition sets at what frequency. Only frequencies whose round-trip path is a whole number of wavelengths survive — the longitudinal cavity modes — and among those, only ones under the part of the gain profile that clears the loss line reach threshold. What keeps the arrangement stable rather than hair-trigger is gain clamping: if the field overshoots, the stronger field saturates the gain back down; if it dips, the gain recovers. The system self-regulates to gain equals loss, and the second figure shows the payoff — the laser's signature input–output curve, dead below threshold and ruler-straight above it.

Figure

Output power versus pump power for a diode-pumped Nd:YAG laser: below the threshold P_th ≈ 0.77 W the light is faint fluorescence, above it the output rises linearly because the inversion ΔN clamps at ΔN_th and every extra pump quantum becomes one output photon

The core equations

Everything begins by asking a wave to survive one lap. Track the complex field from a reference plane, through the medium to the far mirror and back, collecting a gain factor from each traversal and a reflection factor from each mirror; self-consistency demands it return to itself.

E_\mathrm{round}=E_0\,r_1 r_2\,\exp[(g-\alpha)\ell]\,\exp[(g-\alpha)\ell]\,e^{-i\,2k\ell} \qquad (1)

where: $E_0$ — complex field amplitude at a reference plane; $r_1,r_2$ — amplitude reflection coefficients, with $R_i=|r_i|^2$ ; $g$ — gain coefficient per unit length; $\alpha$ — distributed internal loss per unit length; $\ell$ — gain-medium length (taken to fill the cavity here); $k=n\omega/c$ — propagation constant. Setting $E_\mathrm{round}=E_0$ and splitting into modulus and argument gives the amplitude and phase conditions. The magnitude part is the threshold gain requirement: the round-trip intensity multiplier — two passes of net gain times the two reflectivities — must equal exactly one. More, and the field grows without bound; less, and it dies.

R_1R_2\,\exp[2(g_\mathrm{th}-\alpha)\ell]=1 \qquad (2)

where: $g_\mathrm{th}$ — threshold gain coefficient, the minimum gain for oscillation; $R_1R_2$ — product of the mirror intensity reflectivities. The factor of $2$ is the whole point: a round trip crosses the medium twice. Taking the logarithm of the amplitude condition isolates the threshold gain as a sum of loss rates — the workhorse formula of the topic.

g_\mathrm{th}=\alpha+\frac{1}{2\ell}\ln\!\frac{1}{R_1R_2} \qquad (3)

where: $\alpha$ — distributed internal loss per unit length; $\tfrac{1}{2\ell}\ln(1/R_1R_2)$ — effective mirror (output-coupling) loss per unit length. Threshold gain equals total loss: the medium makes good exactly what the cavity spills. Because gain is $g=\sigma(\nu)\,\Delta N$ , the threshold gain fixes a minimum population inversion — the bridge from resonator engineering back to the atomic physics of the gain medium.

\Delta N_\mathrm{th}=\frac{g_\mathrm{th}}{\sigma(\nu)} \qquad (4)

where: $\Delta N_\mathrm{th}$ — threshold inversion density (upper minus statistically weighted lower level), in $\mathrm{m^{-3}}$ ; $\sigma(\nu)$ — stimulated-emission cross section, in $\mathrm{m^2}$ . A large cross section (a "strong" transition) buys a low threshold inversion. The same round-trip loss reads as a decay rate rather than a length: divide the fractional loss by the round-trip time and you have the cavity photon lifetime $\tau_c$ , and threshold says the gain rate equals the loss rate.

\frac{1}{\tau_c}=v_g\left(\alpha+\frac{1}{2\ell}\ln\frac{1}{R_1R_2}\right)=v_g\,g_\mathrm{th} \qquad (5)

where: $\tau_c$ — cavity photon lifetime, the mean dwell time before a photon is lost; $v_g$ — group velocity in the medium. A high- $Q$ resonator (long $\tau_c$ ) demands only a small $g_\mathrm{th}$ — the cavity does much of the work. Above threshold the inversion cannot rise, so extra pump has nowhere to go but into the output beam — the laser's defining straight line, with intercept $P_\mathrm{th}$ and slope the slope efficiency.

P_\mathrm{out}=\eta_s\,(P_\mathrm{pump}-P_\mathrm{th}),\qquad P_\mathrm{pump}>P_\mathrm{th} \qquad (6)

where: $P_\mathrm{out}$ — output power; $\eta_s$ — slope efficiency (extractable photons per extra pump quantum, times the output-coupling fraction); $P_\mathrm{th}$ — threshold pump power. Measuring this line is the standard way to read $P_\mathrm{th}$ and $\eta_s$ off a device. Finally, how the pump translates into threshold depends sharply on the level scheme: a four-level medium need only sustain the small inversion $\Delta N_\mathrm{th}$ , while a three-level medium must first lift more than half of every active ion out of the ground state.

R_\mathrm{p,th}^{(3)}=\frac{N_\mathrm{tot}}{2\,\tau_2}\quad\text{versus}\quad R_\mathrm{p,th}^{(4)}=\frac{\Delta N_\mathrm{th}}{\tau_2} \qquad (7)

where: $N_\mathrm{tot}$ — total active-ion density; $\tau_2$ — upper-level lifetime; $R_\mathrm{p}$ — pump rate. Since $\Delta N_\mathrm{th}\ll N_\mathrm{tot}$ , the four-level pump rate is smaller by orders of magnitude.

Derivation AThreshold gain from round-trip self-consistency

The goal is to obtain the threshold gain $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ from the barest premise — that a steady field must reproduce its own intensity after one complete lap. We follow a plane wave of intensity $I_0$ from just inside mirror 1 to mirror 2 and back, and demand that what returns equals what left. Nothing quantum enters; it is pure bookkeeping of a wave that grows in the gain and shrinks at the mirrors, and the threshold formula falls out of insisting the books balance.

Step 1. On the forward pass the intensity obeys Beer–Lambert with a gain term: $I_1=I_0\,e^{(g-\alpha)\ell}$ , growing by $g$ and shrinking by $\alpha$ over $\ell$ . Why: gain and distributed loss both act exponentially, so their net effect is the single factor $e^{(g-\alpha)\ell}$ .

Step 2. Reflection off the far mirror keeps a fraction $R_2$ : $I_2=R_2\,I_1$ ; the missing $1-R_2$ is output plus mirror loss. Why: the output coupler is deliberately partially transmitting — at once a feedback element and the dominant engineered loss.

Step 3. The return pass crosses the medium again, $I_3=I_2\,e^{(g-\alpha)\ell}$ , and mirror 1 gives $I_4=R_1R_2\,I_0\,e^{2(g-\alpha)\ell}$ . Why: the wave sees the gain on both legs — this is where the factor of $2$ is born; forgetting it doubles the apparent gain requirement.

Step 4. Impose steady state $I_4=I_0$ , leaving $R_1R_2\,e^{2(g_\mathrm{th}-\alpha)\ell}=1$ . Below this the mode dies each lap; above it, it swells until saturation. Why: a self-sustained oscillator neither grows nor decays, so equality defines threshold.

Step 5. Take the logarithm and solve: $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ . Both terms are loss rates, so it reads "threshold gain equals total loss." Why: this matches Saleh and Teich Ch. 15 and Svelto Ch. 7, and letting $R\to1$ recovers the low-loss form $g_\mathrm{th}\approx\alpha+(1-R_1R_2)/2\ell$ .

Go deeper · Derivation B · 1 minThreshold as the pump where photons first appear, and why the gain clamps

Derivation BThreshold as the pump where photons first appear, and why the gain clamps

The goal is to locate threshold from the dynamics rather than the geometry — to find, from the coupled photon and inversion rate equations, the pump rate at which the photon number first becomes nonzero, and to show why the inversion freezes there afterward. This complements the round-trip picture: self-consistency told us the gain a mode needs; the rate equations give the pump that supplies it and reveal the negative feedback, gain clamping, that pins a running laser at gain equals loss however hard it is driven.

Step 1. Write the four-level rate equations for inversion density $N$ and photon number $\phi$ : $\dot N=R_\mathrm{p}-N/\tau_2-v_g\sigma N\phi$ and $\dot\phi=v_g\sigma N\phi-\phi/\tau_c$ . Why: $R_\mathrm{p}$ builds inversion, $\tau_2$ leaks it, the shared $v_g\sigma N\phi$ term moves quanta into the field, and $\tau_c$ drains it.

Step 2. Set the photon steady state $\dot\phi=0$ , which factors as $\phi\,(v_g\sigma N-1/\tau_c)=0$ : either $\phi=0$ (non-lasing) or the bracket vanishes. Why: a product equal to zero forces one factor to zero, and which branch the system takes is the threshold question.

Step 3. For any nonzero field the bracket must vanish, so $v_g\sigma N_\mathrm{th}=1/\tau_c$ , i.e. $N_\mathrm{th}=1/(v_g\sigma\tau_c)$ ; with $g=\sigma N$ this is $v_g g_\mathrm{th}=1/\tau_c$ — identically the round-trip threshold gain. Why: two independent routes landing on the same condition confirms gain equals loss is the one true threshold.

Step 4. Set $\dot N=0$ with $\phi\to0^+$ at onset: the pump need only replace spontaneous decay of the clamped inversion, so $R_\mathrm{p,th}=N_\mathrm{th}/\tau_2$ . Why: at onset there are almost no stimulated events, so the whole pump budget holds $N$ against its lifetime.

Step 5. Above threshold, pin $N$ at $N_\mathrm{th}$ to find $\phi=\tau_c\,(R_\mathrm{p}-R_\mathrm{p,th})$ : the field grows linearly with excess pump because the inversion cannot. Why: gain clamping made quantitative — the surplus has nowhere to go but photons, the origin of the straight output line.

Limiting cases

Case	Result	Meaning
Low-loss (high-R) cavity	$g_\mathrm{th}\approx\alpha+\dfrac{1-R_1R_2}{2\ell}$	For $R_1R_2\to1$ , $\ln(1/R_1R_2)\approx1-R_1R_2$ ; threshold gain is tiny, set mainly by mirror transmission.
Dominant output coupling	$g_\mathrm{th}\approx\dfrac{1}{2\ell}\ln\dfrac{1}{R_1R_2}$	With $\alpha$ negligible, threshold is fixed entirely by the mirrors — the regime for designing the output coupler.
Photon-lifetime form	$v_g\,g_\mathrm{th}=1/\tau_c$	Gain rate equals photon loss rate; a high- $Q$ cavity lowers $g_\mathrm{th}$ .
Ideal four-level medium	$R_\mathrm{p,th}=\Delta N_\mathrm{th}/\tau_2$	The lower level empties instantly, so the pump need only sustain the small inversion against decay $\tau_2$ .
Three-level medium	$R_\mathrm{p,th}\approx N_\mathrm{tot}/2\tau_2$	The ground state is the lower laser level, so more than half of all ions must be excited before net gain appears.

Worked examples

1 · A diode-pumped Nd:YAG laser (the concrete case)

A 1 at.% Nd:YAG rod fills a cavity of length $\ell=0.10\ \mathrm{m}$ , with mirrors $R_1=1.00$ and $R_2=0.98$ and internal loss $\alpha=0.10\ \mathrm{m^{-1}}$ . Using $\sigma=2.8\times10^{-23}\ \mathrm{m^2}$ , $\tau_2=230\ \mu\mathrm{s}$ , and $n=1.82$ , find the threshold gain, the threshold inversion (and the fraction of Nd ions it represents), the photon lifetime, and the absorbed pump power at threshold for a rod of $1\ \mathrm{mm^2}$ cross section pumped at 808 nm.

Step 1. Threshold gain from equation (3): $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)=0.10+5\times0.02020=\mathbf{0.201\ m^{-1}}$ . The mirror term ( $0.101\ \mathrm{m^{-1}}$ ) and the internal loss ( $0.100\ \mathrm{m^{-1}}$ ) contribute almost equally here.

Step 2. Threshold inversion from equation (4): $\Delta N_\mathrm{th}=g_\mathrm{th}/\sigma=0.201/(2.8\times10^{-23})=\mathbf{7.2\times10^{21}\ m^{-3}}$ . With 1 at.% doping the total Nd density is $N_\mathrm{tot}\approx1.38\times10^{26}\ \mathrm{m^{-3}}$ , so the inversion needed is a fraction $\Delta N_\mathrm{th}/N_\mathrm{tot}\approx\mathbf{5\times10^{-5}}$ — five hundredths of a percent of the ions, the four-level advantage in one number.

Step 3. Photon lifetime from equation (5), with $v_g=c/n=1.65\times10^8\ \mathrm{m/s}$ : $1/\tau_c=v_g\,g_\mathrm{th}=3.3\times10^7\ \mathrm{s^{-1}}$ , so $\tau_c=\mathbf{30\ ns}$ — about 25 laps before a photon escapes.

Step 4. Threshold pump: it need only replace spontaneous decay of the clamped inversion, $R_\mathrm{p,th}=\Delta N_\mathrm{th}/\tau_2=7.2\times10^{21}/(230\times10^{-6})=3.1\times10^{25}\ \mathrm{m^{-3}s^{-1}}$ . Over the rod volume $V=\ell A=10^{-7}\ \mathrm{m^3}$ that is $3.1\times10^{18}$ excitations per second, each needing one 808 nm photon of energy $hc/\lambda=2.46\times10^{-19}\ \mathrm{J}$ , so $P_\mathrm{th}=\mathbf{0.77\ W}$ — a sub-watt threshold, exactly the regime in which small diode-pumped Nd:YAG lasers run continuously.

2 · Choosing the output coupler

Keep the same rod ( $\ell=0.10\ \mathrm{m}$ , $\alpha=0.10\ \mathrm{m^{-1}}$ , $R_1=1.00$ ) but compare two output couplers, $R_2=0.98$ and $R_2=0.90$ . How much more pump does the leakier coupler demand?

Step 1. For $R_2=0.98$ : $g_\mathrm{th}=0.10+5\ln(1/0.98)=\mathbf{0.201\ m^{-1}}$ (as in Example 1).

Step 2. For $R_2=0.90$ : $g_\mathrm{th}=0.10+5\ln(1/0.90)=0.10+0.527=\mathbf{0.627\ m^{-1}}$ . The mirror loss now dwarfs the internal loss.

Step 3. Because threshold pump scales with $g_\mathrm{th}$ (through $\Delta N_\mathrm{th}=g_\mathrm{th}/\sigma$ and $R_\mathrm{p,th}=\Delta N_\mathrm{th}/\tau_2$ ), the leakier coupler raises threshold by $0.627/0.201=\mathbf{3.1\times}$ . That is the central trade: the 10 percent coupler extracts far more light per pass once running, but more than triples the pump needed to start — and if the gain cannot reach $0.627\ \mathrm{m^{-1}}$ , that laser never turns on. The optimum coupler maximizes output for a given pump by balancing useful loss against the parasitic $\alpha$ .

3 · Why ruby needs a flashlamp but Nd:YAG does not

Ruby is three-level, chromium density $N_\mathrm{tot}\approx1.6\times10^{25}\ \mathrm{m^{-3}}$ with the ground state as the lower laser level. Nd:YAG is four-level, $\Delta N_\mathrm{th}\approx7.2\times10^{21}\ \mathrm{m^{-3}}$ from Example 1. Compare the inversions each must reach.

Step 1. Ruby must lift more than half of all ions out of the ground state before net gain appears, so its threshold inversion is of order $N_\mathrm{tot}/2\approx\mathbf{8\times10^{24}\ m^{-3}}$ .

Step 2. The ratio is $8\times10^{24}/7.2\times10^{21}\approx\mathbf{1100}$ : the three-level system must invert a thousand times more ion density for the same gain. Since threshold pump scales with required inversion, ruby is flash-pumped and pulsed while Nd:YAG's tiny four-level inversion is met by a modest CW diode. The level scheme, through the threshold condition, decides flashlamp versus diode.

Physics of the portrait

The threshold condition turned a laboratory curiosity into an oscillator — where the atomic physics of stimulated emission had to meet the engineering of a resonant cavity before anyone could build a device.

Origins · 1 minthreshold condition and how this was discovered

threshold condition

round-trip gain equals round-trip loss · laser physics

The condition that an optical oscillator's gain must overcome its cavity losses was first set down quantitatively by Arthur Schawlow and Charles Townes in their 1958 paper "Infrared and Optical Masers," the blueprint that extended the microwave maser — which Townes, with Gordon and Zeiger, had built in 1954 — into the visible. Their crucial move was to choose an open Fabry–Pérot resonator and work out the pump needed for gain to balance its losses. Two years later Theodore Maiman made the first laser, a ruby three-level system that, precisely because ruby is three-level, demanded the fierce inversion only a flashlamp could deliver; within months Javan, Bennett and Herriott ran the first continuous-wave gas laser, a helium–neon device whose low, nearly four-level threshold let it run steadily. The physics underneath was old in parts — Einstein's 1917 coefficients gave the gain, Fabry and Pérot's interferometer the cavity — but it was the threshold condition that welded gain physics to resonator physics into one design rule. Townes, Basov and Prokhorov shared the 1964 Nobel Prize for the principle; Schawlow's 1981 Nobel came for the laser spectroscopy it made possible. Every line of a modern datasheet, from turn-on pump to slope efficiency, is still read off the balance those papers struck.

History

1917

— Einstein introduces the A and B coefficients, giving stimulated emission and the microscopic basis of optical gain

1954

— Gordon, Zeiger and Townes operate the first ammonia maser, demonstrating amplification by stimulated emission

1958

— Schawlow and Townes derive the threshold condition for an optical maser and propose the Fabry–Pérot laser cavity

1960

— Maiman demonstrates the first laser, a flash-pumped ruby three-level system

1960

— Javan, Bennett and Herriott achieve the first continuous-wave gas laser (helium–neon)

1964

— Townes, Basov and Prokhorov share the Nobel Prize for the maser–laser principle

Going further

Texts. Saleh and Teich, Fundamentals of Photonics, 3rd ed., Ch. 15 (Lasers) · Svelto, Principles of Lasers, 5th ed., Ch. 7 (threshold and CW operation) · Hecht, Optics, 5th ed., §13.1 · Foot, Atomic Physics, Ch. 1 and 7 (laser action) · Loudon, The Quantum Theory of Light, 3rd ed., Ch. 1 and 7 · Demtröder, Laser Spectroscopy, Vol. 1, Ch. 5 (lasers as spectroscopic sources) · Koechner, Solid-State Laser Engineering (Nd:YAG parameters).

Provenance. The equations are standard laser physics. A verifier should reproduce the round-trip derivation and $g_\mathrm{th}=\alpha+\tfrac{1}{2\ell}\ln(1/R_1R_2)$ against Saleh and Teich Ch. 15 and Svelto Ch. 7, the rate-equation threshold and gain-clamping condition $v_g\sigma N_\mathrm{th}=1/\tau_c$ against Foot Ch. 7, and the three-versus-four-level comparison against Demtröder Ch. 5. The worked Nd:YAG numbers use standard tabulated parameters ( $\sigma\approx2.8\times10^{-23}\ \mathrm{m^2}$ , $\tau_2\approx230\ \mu\mathrm{s}$ , $n=1.82$ , 1 at.% doping $N_\mathrm{tot}\approx1.38\times10^{26}\ \mathrm{m^{-3}}$ , 808 nm pump) from Koechner; reflectivity and loss values are illustrative but realistic. Symbol conventions follow Saleh and Teich.

Prerequisites. Population Inversion and Optical Gain · Stimulated Emission and the Gain Coefficient · Optical Resonators and Cavity Modes · Cavity Q Factor and Photon Lifetime. Leads to. Laser Rate Equations · Four-Level Laser Systems · Three-Level Laser Systems · Longitudinal Modes and Free Spectral Range.

Ask the tutor: Want to see how the threshold pump shifts when you change the output coupler, rod length, or doping — or how gain clamping fixes the slope efficiency? Ask and we will run the numbers together.

Is threshold the same thing as having population inversion?

No, and conflating them is the most common error. Population inversion gives net gain, $g>0$ , but that only says the wave grows on a single pass. Threshold asks the harder question: does it still grow after a complete round trip, once you subtract everything the cavity spilled. A weakly inverted medium in a leaky cavity can have real gain and never lase. The condition $R_1R_2e^{2(g_\mathrm{th}-\alpha)\ell}=1$ is that round-trip test, always stricter than "gain is positive" because it must also pay for the mirror transmission that makes the laser useful.

Why does the inversion stop rising once the laser turns on?

This is gain clamping, and it follows from the photon rate equation. Above threshold the field is nonzero, which forces the bracket $v_g\sigma N-1/\tau_c$ to vanish, pinning $N$ at $N_\mathrm{th}=1/(v_g\sigma\tau_c)$ . If the inversion tried to climb higher the gain would exceed the loss, the field would surge, and the surge would stimulate the extra inversion away — a fast negative feedback. So the pump can be doubled and the inversion does not budge; the surplus becomes output photon-for-photon. That is why the output-versus-pump curve is a straight line with a well-defined slope efficiency.

How can very high-reflectivity mirrors make the threshold gain so tiny?

Because the mirror loss enters as $\tfrac{1}{2\ell}\ln(1/R_1R_2)$ , which collapses toward zero as $R_1R_2\to1$ . A high-reflectivity cavity recycles each photon through the gain medium many times — of order $1/(1-R)$ passes — so even a feeble per-pass gain accumulates enough amplification to overcome the trickle of loss. In Example 1 the photon made about 25 laps, so a gain of only $0.2\ \mathrm{m^{-1}}$ sufficed. This is the same physics as a high- $Q$ resonator storing energy for a long photon lifetime: $v_g g_\mathrm{th}=1/\tau_c$ , so long $\tau_c$ means small $g_\mathrm{th}$ .

What does the phase part of the round-trip condition do?

Equation (1) is complex, so setting the field equal to itself splits into two statements. The magnitude part is the gain condition used throughout. The argument part, $2k\ell=2\pi q$ for integer $q$ , requires the round-trip path to be a whole number of wavelengths, selecting a comb of longitudinal modes spaced by the free spectral range. The amplitude condition then decides which comb lines have enough gain to oscillate. So phase chooses the possible frequencies and amplitude the achievable ones; the spectrum is the overlap of comb and gain profile.

Why do four-level lasers reach threshold so much more easily than three-level ones?

Because of where the lower laser level sits. In a four-level scheme it lies above the ground state and empties almost instantly, so it stays essentially empty and any excited population is net inversion — you need only the small $\Delta N_\mathrm{th}$ set by the cavity. In a three-level scheme the lower level is the ground state, which starts full; you must pump more than half of every ion up just to break even, and only the excess is gain. Example 3 put numbers on it: ruby needs about a thousand times more inversion than Nd:YAG for the same gain, which is why one is flash-pumped and pulsed and the other runs continuously on a diode.

How is the threshold condition used to run a Q-switch?

A Q-switch makes the threshold a variable. During pumping the switch holds the cavity in a high-loss state, raising $g_\mathrm{th}$ above the gain the medium can supply, so it cannot lase and instead banks a large inversion well beyond the normal threshold. When the switch flips to low loss, $g_\mathrm{th}$ crashes below the stored gain, and the surplus dumps the whole banked inversion into the field in a few cavity lifetimes — a giant nanosecond pulse. Nothing about the medium changed; only the loss term in $g_\mathrm{th}$ was switched, the cleanest demonstration that threshold is a design variable, not a fixed property.