Solid-State and Rare-Earth Lasers

A solid-state laser is what you get when the general machinery of laser action — population inversion, the gain coefficient, the threshold condition, the rate equations — is loaded into a very particular amplifier: a transparent crystal or glass sprinkled with optically active ions. From the prerequisite articles it inherits everything abstract: inversion gives a small-signal gain $g_0=\sigma_{21}\,\Delta N$ ; a resonator turns that gain into oscillation once it beats the round-trip loss; a four-level terminal state makes threshold cheap and a three-level one dear. What the solid host adds is the concrete quantum mechanics of the ion, whose active electrons live in an electrostatic cage — the crystal field of the lattice — and the single most important fact about a doped-solid laser is how strongly those electrons feel that cage. Two ion families answer oppositely. Rare-earth ions such as Nd³⁺, Yb³⁺, and Er³⁺ hide their 4f electrons beneath filled 5s²5p⁶ shells, so the lattice barely perturbs them: their f–f transitions stay sharp and atomic-like, giving narrow lines, large peak cross-sections, and the long metastable lifetimes that store energy for Q-switching. Transition-metal ions such as Cr³⁺ and Ti³⁺ wear their 3d electrons on the outside, where the crystal field and phonons smear every transition into a broad vibronic band — the origin of tunability and femtosecond pulses. Ruby (Cr³⁺:Al₂O₃), the first laser of all, is three-level and thirsty; Nd:YAG is four-level and efficient; Ti:sapphire is vibronic and broadly tunable — one sheet of physics splitting into very different machines.

Key takeaways

A solid-state laser is a transparent host (sapphire, YAG, YVO₄, glass) doped with optically active ions; the ions supply the levels, the host supplies transparency, strength, and heat conduction, and pumping is always optical (today mostly diodes).
Rare-earth ions have shielded 4f electrons, so their transitions are sharp and atomic-like (narrow lines, large σ, long lifetimes) — ideal for efficient CW and Q-switched sources like Nd:YAG at 1064 nm.
Transition-metal ions expose their 3d electrons to the crystal field and phonons, giving broad vibronic bands — the tunability and sub-30-fs pulses of Ti:sapphire (≈650–1100 nm).
The three- versus four-level structure of the terminal state, not the fact of being solid, sets the threshold: ruby must bleach its ground state, Nd:YAG need not.
The quantum defect $1-\lambda_p/\lambda_L$ becomes heat no matter how perfect the laser; minimising it (Yb:YAG ≈9 percent versus Nd:YAG ≈24 percent) is why high-power lasers migrate to Yb.
Key idea: shielded rare-earth 4f transitions are sharp and store energy while exposed transition-metal 3d transitions are broad and tunable, and the terminal-level structure (three- versus four-level) fixes the pump threshold — two independent choices that between them define every solid-state laser.

The core insight

Shielded 4f (rare-earth) transitions are sharp, strong, and long-lived while exposed 3d (transition-metal) transitions are broad and tunable; independently, a four-level terminal state makes threshold tiny while a three-level one forces you to bleach the ground state — and those two choices classify every solid-state laser.

Where this lives today

The doped-solid physics of this article runs in almost any modern optics facility. The industrial floor belongs to Nd:YAG and Yb fibre: diode-pumped Nd:YAG at 1064 nm — frequency-doubled to green (532 nm) or tripled to ultraviolet (355 nm) — cuts, welds, and marks metal, while kilowatt-class Yb fibre lasers, riding their tiny quantum defect, have taken over high-power cutting because their huge surface-to-volume ratio sheds heat that would fracture a rod. At the opposite extreme of duration sits Ti:sapphire, whose vibronic band feeds mode-locked oscillators and chirped-pulse amplifiers (the 2018 Nobel technique) producing few-cycle pulses and terawatt peaks — underpinning high-harmonic generation, attosecond science, and multiphoton microscopy. And the same shielding that gives Nd³⁺ its sharp line gives Er³⁺ its 1.5 μm transition, the backbone of the erbium-doped fibre amplifier that keeps the world's optical-fibre internet lit.

The physical picture

Picture the doping ion as a nearly free atom clamped inside an electrostatic cage built from the neighbouring host ions. For a rare-earth ion the optically active 4f electrons sit deep inside the filled 5s²5p⁶ shells, so they hardly notice the cage: the free-ion Russell–Saunders terms ²ˢ⁺¹L_J survive almost intact, split only slightly by the Stark effect, and the transition stays sharp — a caged bird singing its own undisturbed tune. For a transition-metal ion the active 3d electrons are the outermost ones; they grip the cage bars directly and rattle with every lattice vibration, so each transition dresses itself with phonons and smears into a broad vibronic band. That single difference — shielded versus exposed — is the fork that separates narrow-line, energy-storing lasers from broadband, tunable, ultrafast ones. A second, independent fork decides where the electron lands after emitting its laser photon, controlling the threshold — the first figure sets the two archetypes side by side.

Figure

Energy levels of a three-level laser like ruby Cr³⁺ in Al₂O₃ versus a four-level laser like Nd:YAG. In ruby the lower laser level is the crowded ground state, so more than half the ions must be excited before the 694.3 nm line shows gain; in Nd:YAG the 1064 nm transition ends on a level about 2000 cm⁻¹ above ground that empties in nanoseconds, so N₁ ≈ 0 and threshold is tiny

The energy a rare-earth ion holds while it waits is the second asset of the shielded 4f states. The metastable upper level lives extraordinarily long — Nd³⁺ about 230 μs, Yb³⁺ about 1 ms, Er³⁺ about 10 ms — because the shielding also suppresses the pathways that would drain it. That lifetime is a battery: pour pump energy in slowly, the ions hold it far longer than a cavity photon survives, then a Q-switch opens the cavity to dump the whole store in a nanosecond giant pulse. The second figure shows the spectral consequence of the shielded-versus-exposed fork: the same integrated cross-section as a narrow spike or a broad hill.

Figure

Emission cross-section versus wavelength: the shielded 4f transition of Nd³⁺ gives a sharp 1064 nm line with a large peak σ ≈ 2.8×10⁻¹⁹ cm², while the exposed 3d electron of Ti³⁺ spreads its vibronic gain across ≈650–1100 nm with a peak σ roughly 100 times smaller — narrow-and-tall versus broad-and-short at nearly fixed integrated area

The core equations

Laser action begins with gain: in any inverted medium the small-signal (unsaturated) gain coefficient is the product of the stimulated-emission cross-section and the inversion density. Every subsequent number funnels through it, converting an atomic quantity σ and a pumping quantity ΔN into a measurable amplification per metre.

g_0 = \sigma_{21}\,\Delta N = \sigma_{21}\,(N_2 - N_1) \qquad (1)

where: $g_0$ — small-signal gain coefficient per unit length (m⁻¹); $\sigma_{21}$ — stimulated-emission cross-section at the laser wavelength (m²); $\Delta N=N_2-N_1$ — population-inversion density (m⁻³) between the upper (2) and lower (1) laser levels. Gain is inversion measured in units of $1/\sigma_{21}$ , so a larger cross-section means fewer excited ions for the same gain. Because the integrated cross-section of a homogeneous line is fixed, a narrow rare-earth line concentrates σ into a tall peak while a broad vibronic band spreads it thin — Nd³⁺ reaches σ ≈ 2.8×10⁻¹⁹ cm², Ti³⁺ roughly a hundred times less. Oscillation starts only when a wave returns to its starting point at least as strong as it left, so the gain must cancel the resonator's round-trip losses. Equating round-trip gain and loss and taking a logarithm gives the threshold gain.

\gamma_{\text{th}}\,\ell = \tfrac{1}{2}\left[\alpha\ell - \tfrac{1}{2}\ln(R_1 R_2)\right] \qquad (2)

where: $\gamma_{\text{th}}$ — threshold gain coefficient (m⁻¹); $\ell$ — length of the gain medium (m); $\alpha$ — distributed internal (scattering and absorption) loss coefficient (m⁻¹); $R_1,R_2$ — mirror power reflectivities. The bracket is the round-trip loss — parasitic $\alpha\ell$ plus output coupling $-\tfrac12\ln(R_1R_2)$ — and the outer ½ converts it to the single-pass gain that must supply it. The same threshold gain demands a definite inversion, and here the four- and three-level worlds part company. Combining $\gamma_{\text{th}}=\sigma_{21}\Delta N_{\text{th}}$ with the terminal-level bookkeeping gives two very different requirements.

\Delta N_{\text{th}}^{(4)} \approx \frac{\gamma_{\text{th}}}{\sigma_{21}}\,,\qquad \Delta N_{\text{th}}^{(3)} \gtrsim \tfrac{1}{2}\,N_{\text{tot}} \qquad (3)

where: $\Delta N_{\text{th}}$ — inversion density needed at threshold; $N_{\text{tot}}$ — total active-ion density; the superscript labels the level scheme. In a four-level laser $N_1\approx0$ , so the required inversion is just the tiny $\gamma_{\text{th}}/\sigma_{21}$ — a small excited population is already net gain. In a three-level laser the lower level is the ground state, so half of all ions must be excited merely to reach transparency and still more for gain — a macroscopic fraction of the whole doping. For a four-level laser the required inversion becomes an absorbed pump power, scaling inversely with the upper-state lifetime and cross-section, both large for rare-earths.

P_{\text{th}} = \frac{h\nu_p\,V}{\eta_p\,\tau_2}\,\frac{\gamma_{\text{th}}}{\sigma_{21}} \qquad (4)

where: $P_{\text{th}}$ — absorbed pump power at threshold (W); $h\nu_p$ — pump photon energy (J); $\eta_p$ — pump quantum efficiency (fraction of absorbed pump photons reaching level 2); $\tau_2$ — upper-state (fluorescence) lifetime (s); $V$ — pumped mode volume (m³). A long lifetime accumulates inversion between decays and a large cross-section lets a little inversion do a lot — so Nd³⁺ thresholds at milliwatts. Above threshold, extra pump converts to output at a fixed slope efficiency, a chain of efficiencies capped by the quantum defect.

P_{\text{out}} = \eta_s\,(P_{\text{abs}} - P_{\text{th}}),\qquad \eta_s = \eta_p\,\eta_c\,\eta_q\,\frac{T}{T+L} \qquad (5)

where: $P_{\text{out}}$ — output power (W); $\eta_s$ — slope efficiency; $\eta_c$ — transverse mode-overlap efficiency; $\eta_q=\nu_L/\nu_p$ — quantum efficiency, the laser-to-pump photon-energy ratio; $T,L$ — output-coupler transmission and round-trip parasitic loss. No factor exceeds one, so $\eta_q$ sets a hard ceiling on efficiency — fixed by the two wavelengths alone. That ceiling, and the heat that is its shadow, is the quantum defect: the fraction of each pump photon's energy that cannot emerge as laser light and is dumped into the lattice as phonons.

\eta_q = \frac{h\nu_L}{h\nu_p} = \frac{\lambda_p}{\lambda_L},\qquad \text{quantum defect} = 1-\eta_q = 1-\frac{\lambda_p}{\lambda_L} \qquad (6)

where: $\lambda_p,\lambda_L$ — pump and laser wavelengths; $\eta_q$ — maximum fractional energy extractable per photon. Pump Nd:YAG at 808 nm to lase at 1064 nm and 24 percent is lost as heat before any other loss; pump Yb:YAG at 940 nm to lase at 1030 nm and only 9 percent is. That single ratio is why the multi-kilowatt world belongs to ytterbium.

Derivation AThreshold pump power of a four-level solid-state laser

The goal is to compute how much absorbed pump power a four-level laser such as Nd:YAG needs to reach threshold, exposing why the upper-state lifetime and the emission cross-section are the two knobs that matter most. We write the rate equation for the upper level below threshold, solve for the inversion the pump sustains, connect that rate to a real absorbed power, and demand the inversion equal the value the cavity loss dictates.

Step 1. Below threshold, stimulated emission is negligible, so the upper level fills at the volumetric pump rate $R_p$ and empties with lifetime $\tau_2$ : $\dfrac{\mathrm{d}N_2}{\mathrm{d}t}=R_p-\dfrac{N_2}{\tau_2}$ . Why: with no laser field yet, the only two things happening to level 2 are pumping in and spontaneous decay out.

Step 2. In steady state the derivative vanishes, giving $N_2^{ss}=R_p\,\tau_2$ , and because a four-level system keeps $N_1\approx0$ the inversion is simply $\Delta N\approx R_p\tau_2$ . Why: a long lifetime lets each pump event accumulate before it decays, so the same pump rate builds a larger inversion — τ₂ as a lever.

Step 3. Connect the pump rate to measurable power: each absorbed pump photon of energy $h\nu_p$ deposits, with probability $\eta_p$ , one ion in level 2, so $R_p\,V=\eta_p\,P_{\text{abs}}/h\nu_p$ . Why: the bridge from photons per second per cubic metre to watts, where the pump wavelength and pumped volume enter.

Step 4. The cavity fixes the target inversion: from $g_0=\sigma_{21}\Delta N$ at the threshold gain, $\Delta N_{\text{th}}=\gamma_{\text{th}}/\sigma_{21}$ . Why: gain climbs only until it cancels loss, so equation (2) hands us the target — larger σ means a smaller target.

Step 5. Combine Steps 2 through 4 — set $R_p\tau_2=\Delta N_{\text{th}}$ , substitute $R_p$ from Step 3, and solve for the absorbed power. Why: long $\tau_2$ and large $\sigma_{21}$ , both hallmarks of rare-earth ions, drive $P_{\text{th}}$ down — why Nd³⁺ four-level lasers are so easy to run.

P_{\text{th}}=\frac{h\nu_p\,V}{\eta_p\,\tau_2}\,\frac{\gamma_{\text{th}}}{\sigma_{21}}

Verified: consistent with Svelto Ch. 7; Nd:YAG values (σ ≈ 2.8×10⁻¹⁹ cm², τ₂ ≈ 230 μs) give thresholds of tens of milliwatts to a few watts, matching Worked example 1.

Go deeper · Derivation B · 1 minWhy a three-level laser like ruby demands more than half its ions excited

Derivation BWhy a three-level laser like ruby demands more than half its ions excited

The goal is to prove that a three-level scheme, whose lower laser level is the ground state itself, cannot even become transparent until half of all active ions are excited, making quantitative why ruby has an enormous threshold and runs pulsed. We write the population balance for the two levels holding essentially all the ions, express the inversion through the excited fraction, and show that gain needs still more — the sharp factor $(2f-1)$ .

Step 1. With fast relaxation out of the pump band, essentially every ion is in either the ground/lower level (1) or the metastable upper laser level (2), so $N_1+N_2\approx N_{\text{tot}}$ . Why: the intermediate pump levels drain so quickly they hold negligible population, collapsing the system to two reservoirs.

Step 2. Define the excited fraction $f=N_2/N_{\text{tot}}$ and eliminate $N_1$ : $\Delta N=N_2-(N_{\text{tot}}-N_2)=(2f-1)\,N_{\text{tot}}$ . Why: one variable exposes the tell-tale $-1$ that makes the ground state a liability, not an empty basement.

Step 3. The medium is transparent when $\Delta N=0$ , which requires $f=\tfrac12$ ; below that it absorbs at the laser wavelength. Why: half the ions must be excited just to stop the medium eating its own light — the entire penalty of a three-level scheme.

Step 4. Real oscillation needs $\Delta N_{\text{th}}=\gamma_{\text{th}}/\sigma_{21}>0$ , so the excited fraction must exceed one half by $\gamma_{\text{th}}/(2\sigma_{21}N_{\text{tot}})$ . Bleaching more than half of ruby's ≈1.6×10¹⁹ cm⁻³ Cr³⁺ ions takes enormous energy density — exactly why Maiman's 1960 ruby laser was flashlamp-pumped and pulsed. Why: the threshold is now a macroscopic fraction of the doping, not a whisper above zero, quantifying the gulf between ruby and Nd:YAG. *Verified: standard three-level analysis (Foot Ch. 7; Svelto Ch. 6); the factor $(2f-1)$ is textbook, and ruby's Cr³⁺ ≈1.6×10¹⁹ cm⁻³ at λ = 694.3 nm are established.*

Limiting cases

Case	Result	Meaning
Rare-earth, shielded 4f	$\Delta\nu_{\text{line}}\sim$ GHz–THz (sharp)	Weak crystal-field coupling; atomic-like ²ˢ⁺¹L_J terms, small Stark splitting. Narrow, strong, temperature-stable, long-lived: Nd³⁺, Er³⁺, Yb³⁺. Ideal for CW and Q-switched sources.
Transition-metal, exposed 3d	$\Delta\nu_{\text{band}}\sim100$ THz (broad)	Strong crystal-field and electron–phonon coupling. Broad vibronic gain: Ti³⁺:sapphire ≈650–1100 nm, Cr³⁺:LiSAF, alexandrite. Tunable, femtosecond-capable.
Three-level, terminal = ground	$P_{\text{th}}$ high; need $N_2\gtrsim\tfrac12 N_{\text{tot}}$	Ruby (694.3 nm), Er:glass. Must bleach the ground state; large threshold, usually pulsed.
Four-level, terminal above ground	$P_{\text{th}}$ low; $N_1\approx0$	Nd:YAG (1064 nm), Nd:YVO₄, Ti:sapphire. Lower level empties fast; minimal pump. Efficient CW.
Quasi-three-level, small defect	$h\nu_p\gtrsim h\nu_L,\ \eta_q\to1$	Yb:YAG (1030 nm, pumped 940 nm). Lower level thermally populated but tiny defect (≈9 percent): low heat, multi-kW scaling.

Common trapSolid-state does not mean four-level, narrow-band, or a fixed pump-to-laser offset

Three traps hide in the phrase "solid-state laser". First, solid-state is not the same as four-level: ruby is solid-state yet three-level, with a threshold thousands of times higher than Nd:YAG's. Second, solid-state is not the same as narrow-band: Ti:sapphire is a crystal like the rest, yet its vibronic gain spans hundreds of nanometres and supports sub-10-fs pulses. Third, the laser wavelength is not the pump wavelength minus a universal constant: the quantum defect varies enormously — about 9 percent for Yb³⁺ pumped near its emission, roughly 24 percent for Nd:YAG — and this ratio, not the level scheme, governs the heat load and thermal lensing that cap average power. Keep the forks straight: shielded-or-exposed sets linewidth and storage, three-or-four level sets threshold, the pump-to-laser ratio sets the heat.

Worked examples

1 · Threshold pump power of a diode-pumped Nd:YAG rod (confidence builder)

A 5 mm Nd:YAG rod is diode-pumped at 808 nm inside a cavity with a high reflector (R₁ = 1.00) and a 10 percent output coupler (R₂ = 0.90); internal loss is negligible. The pumped mode has radius w_p = 100 μm. Using σ₂₁ = 2.8×10⁻¹⁹ cm², τ₂ = 230 μs, and η_p ≈ 1, find the absorbed pump power at threshold.

Step 1. Get the round-trip loss and the threshold gain from equation (2). With α ≈ 0, the only loss is the output coupler: $-\tfrac12\ln(R_1R_2)=-\tfrac12\ln(0.90)=0.0527$ , so $\gamma_{\text{th}}\ell=\tfrac12(0.0527)=0.0263$ and $\gamma_{\text{th}}=0.0263/(5\times10^{-3}\,\text{m})=\mathbf{5.27\ m^{-1}}$ .

Step 2. Convert to the threshold inversion via equation (3): with σ₂₁ = 2.8×10⁻²³ m², $\Delta N_{\text{th}}=\gamma_{\text{th}}/\sigma_{21}=5.27/2.8\times10^{-23}=\mathbf{1.88\times10^{23}\ m^{-3}}$ (1.9×10¹⁷ cm⁻³ — only one in 10⁵ of the ≈1.4×10²⁰ cm⁻³ Nd ions, the signature of a four-level system).

Step 3. Assemble the pumped volume and the pump-photon energy. $V=\pi w_p^2\ell=\pi(10^{-4})^2(5\times10^{-3})=1.57\times10^{-10}\ \text{m}^3$ ; $h\nu_p=hc/\lambda_p=(6.626\times10^{-34})(2.998\times10^{8})/(808\times10^{-9})=2.46\times10^{-19}\ \text{J}$ .

Step 4. Insert into equation (4): $P_{\text{th}}=\dfrac{h\nu_p\,V}{\eta_p\,\tau_2}\,\Delta N_{\text{th}}$ . The numerator group $h\nu_p V\Delta N_{\text{th}}=7.3\times10^{-6}$ J, divided by τ₂ = 2.30×10⁻⁴ s, gives P_th ≈ 0.032 W, about 32 mW. Only about 3×10¹³ ions are inverted at threshold — a whisper of the total — which is why a small diode brings this laser to life while ruby would demand bleaching more than 10¹⁹ cm⁻³.

2 · Quantum defect, heat load, and thermal lens for Nd:YAG versus Yb:YAG

Both a Nd:YAG and a Yb:YAG rod absorb 100 W of pump. Using the quantum defects of equation (6) and YAG's κ = 14 W m⁻¹ K⁻¹ and dn/dT = 7.3×10⁻⁶ K⁻¹, compare the minimum heat deposited and the resulting thermal-lens focal length for a pumped radius w_p = 1 mm.

Step 1. Minimum heat is the defect times the absorbed power. Nd:YAG: $Q_h\ge0.24\times100=\mathbf{24\ W}$ ; Yb:YAG: $Q_h\ge0.09\times100=\mathbf{9\ W}$ — less than half the heat for the same pump.

Step 2. The thermal-lens power follows the rod formula $\dfrac{1}{f_{\text{th}}}\approx\dfrac{P_h}{\pi w_p^2}\,\dfrac{1}{\kappa}\left(\tfrac12\dfrac{\mathrm{d}n}{\mathrm{d}T}\right)$ . For Nd:YAG: $\dfrac{1}{f}=\dfrac{24}{\pi(10^{-3})^2}\cdot\dfrac{1}{14}\cdot\tfrac12(7.3\times10^{-6})=1.99\ \text{m}^{-1}$ , so f_th ≈ 0.50 m.

Step 3. For Yb:YAG only the heat changes, scaling by 9/24: $\dfrac{1}{f}=1.99\times\dfrac{9}{24}=0.75\ \text{m}^{-1}$ , so f_th ≈ 1.34 m — a far weaker lens. Less distortion and a wider stability window is precisely why Yb media, in fibre and thin-disk geometries that spread heat over a large surface, own the multi-kilowatt regime.

Physics of the portrait

The solid-state laser began with an object one person switched on first — a ruby rod flashing red in a Malibu lab in 1960.

Origins · 1 minTheodore Maiman and how this was discovered

Theodore Maiman

1927–2007 · built the first working laser, a pulsed ruby (Cr³⁺:Al₂O₃) device emitting at 694.3 nm, in May 1960

Theodore Harold Maiman was an engineer-physicist at Hughes Research Laboratories who, against the prevailing view that ruby was a hopeless candidate, built the first laser of any kind on 16 May 1960. The reigning wisdom held that ruby's three-level scheme, with the ground state as the lower laser level, would need impossibly hard pumping. Maiman's insight was that he did not need continuous operation: a helical flashlamp firing a millisecond pulse could momentarily bleach more than half the Cr³⁺ ions and cross threshold, exactly the $f\gtrsim\tfrac12$ condition of the three-level derivation above. His device is the archetypal three-level, transition-metal, solid-state laser, and every distinction this article draws — sharp versus broad, three- versus four-level, heat set by the quantum defect — can be read off from why his ruby was hard to pump and the Nd:YAG that followed was easy.

History

1917

— Einstein derives the A and B coefficients that predict stimulated emission

1958

— Schawlow and Townes lay out the principle of the optical maser (laser)

1960

— Maiman fires the first laser, a pulsed ruby rod at 694.3 nm

1964

— Geusic, Marcos, and Van Uitert demonstrate continuous-wave Nd:YAG at 1064 nm

1986

— Moulton reports the tunable Ti:sapphire vibronic laser

Going further

Texts. Svelto, Principles of Lasers, 5th ed., Ch. 9 (Solid-State, Dye, and Semiconductor Lasers) and Ch. 6–7 · Koechner, Solid-State Laser Engineering, 6th ed. (the standard engineering reference) · Saleh and Teich, Fundamentals of Photonics, 2nd ed., Ch. 15–16 · Foot, Atomic Physics, Ch. 7 (Lasers) · Hecht, Optics, 5th ed., §13.1 · Demtröder, Laser Spectroscopy, Vol. 1, Ch. 5.

Provenance. A verifier should confirm: (1) the gain $g_0=\sigma_{21}\Delta N$ and threshold $R_1R_2e^{2(\gamma-\alpha)\ell}=1$ (Saleh and Teich Ch. 15–16; Svelto Ch. 7); (2) the three-level inversion $\Delta N=(2f-1)N_{\text{tot}}$ with transparency at $f=\tfrac12$ (Foot Ch. 7); (3) the four-level $P_{\text{th}}=(h\nu_p V/\eta_p\tau_2)(\gamma_{\text{th}}/\sigma_{21})$ (Svelto); (4) the material numbers — ruby 694.3 nm (Maiman 1960, Cr³⁺ ≈1.6×10¹⁹ cm⁻³); Nd:YAG 1064 nm, σ ≈ 2.8×10⁻¹⁹ cm², τ₂ ≈ 230 μs, pump 808 nm; Yb:YAG 1030 nm, pump ≈940 nm; Ti:sapphire ≈650–1100 nm, pump ≈532 nm; Er 1.5 μm, Er:YAG 2.94 μm; defects 0.24 and 0.09 — against Koechner and Svelto Ch. 9. The worked results (threshold ≈32 mW; thermal-lens focal lengths ≈0.50 m and ≈1.34 m at 100 W absorbed) follow from those standard inputs.

Builds on: Population Inversion and Optical Gain · Four-Level Laser Systems · Three-Level Laser Systems · Laser Rate Equations · Stimulated Emission and the Gain Coefficient. Leads to: Threshold Condition · Cavity Q Factor and Photon Lifetime · Optical Resonators and Cavity Modes · Gaussian Beams.

Ask the tutor: Want me to walk through why ruby needs to bleach half its ions while Nd:YAG threshold-lases with barely one in a hundred thousand — or size the diode power for a laser you have in mind?

What actually makes a laser "solid-state" rather than a gas or dye laser?

The gain medium is a solid: a transparent crystal or glass host with optically active ions substituted into it, replacing host cations. The host — sapphire, YAG, YVO₄, or a phosphate glass — must be clear at both wavelengths, strong, and thermally conductive, but it supplies no levels; all the laser physics comes from the dopant ions. Because those ions sit in an insulating solid they cannot be excited by a current, so solid-state lasers are always optically pumped — historically by lamps, today overwhelmingly by laser diodes.

Why are rare-earth lines so sharp when the ion is buried inside a crystal?

Because the electrons that make the light are hidden. The optically active 4f electrons lie inside the already-filled 5s²5p⁶ shells, an electrostatic umbrella that screens them from the crystal field and lattice vibrations. The free-ion energy terms survive almost unchanged, perturbed only into small Stark sublevels, so the f–f transitions stay narrow and nearly atomic. That shielding does double duty: it keeps the lines sharp and temperature-stable, and it suppresses the channels that would shorten the upper-state lifetime — giving the long-lived metastable states behind energy storage and Q-switching.

If Ti:sapphire is four-level and efficient, why is its threshold not as low as Nd:YAG's?

Two quantities are at play. The four-level scheme keeps the lower level empty, so Ti:sapphire avoids ruby's ground-state-bleaching penalty. But threshold also scales as $P_{\text{th}}\propto1/\sigma_{21}$ , and here the vibronic breadth hurts: spreading a fixed integrated cross-section across ≈650–1100 nm makes the peak σ roughly a hundred times smaller than Nd³⁺'s, and the upper-state lifetime is short (a few μs versus 230 μs). Both raise the threshold, which is why Ti:sapphire is pumped hard by a several-watt green laser rather than a modest diode — trading low threshold for enormous bandwidth.

Why can't I just pump at the laser wavelength and get zero heat?

You can approach it — that is the quasi-three-level trick — but never quite reach it. Pumping needs a genuine energy gap between absorbed and emitted photons so ions relax into the upper laser level and away from the lower one; if the two photons were identical, light would be reabsorbed as fast as emitted. Yb:YAG pushes the defect to about 9 percent by pumping at 940 nm and lasing at 1030 nm, but its lower level sits only a few hundred wavenumbers above ground and is thermally populated, so it behaves partly like a three-level system and must be kept cool and pumped hard. The defect is a floor on heat, not the only constraint.

What sets the choice between a rod, a thin disk, and a fibre?

Heat removal, almost entirely. The quantum defect fixes how much heat you must extract per watt; the geometry fixes how easily. A cylindrical rod has a poor surface-to-volume ratio, so at high power its core runs hot, builds the thermal lens of Worked example 2, and can fracture. A thin disk is pumped through a large face and cooled through the opposite one, so heat flows along the beam axis and the transverse gradients that cause lensing nearly vanish. A fibre stretches the gain over metres of thread with enormous surface area. This is why kilowatt-class systems are almost always Yb in disk or fibre form — a small defect in a heat-friendly shape.

Is the erbium-doped fibre amplifier really the same physics as these lasers?

Yes, minus the mirrors. An erbium-doped fibre amplifier is a rare-earth solid-state gain medium — Er³⁺ ions in a glass fibre — pumped by a diode (typically 980 nm) to invert the 1.5 μm transition. Signal light passing through is amplified by stimulated emission exactly as in a laser rod; there is simply no resonator, so it amplifies rather than oscillates. The same shielded-4f physics that makes Nd³⁺ sharp gives Er³⁺ its telecom-band transition and a millisecond-class lifetime that stores energy and smooths fluctuations. That single device, repeated along undersea cables, is why global fibre-optic communication works without converting light to electronics at every span.