Worked Interview Design Case

This is a whiteboard-level first pass, not a production prescription. Its purpose is to demonstrate traceable reasoning, estimates, and the tests needed to retire uncertainty.

Prompt

Outline a compact diode-pumped Nd:YAG laser delivering at least 5 W continuous wave at 1064 nm with \(M^2\le1.2\) and stable linear polarization. Explain the resonator, pump, thermal, coating, measurement, and tolerance decisions.

1. Restate requirements and assumptions

State unknowns rather than hiding them. For a first calculation assume:

Item	First-pass value	Must later come from
Pump wavelength and incident power	808 nm, 20 W	Qualified diode spectrum over current, temperature, and life
Pump transport efficiency	0.90	Coating, fiber, lens, and alignment budget
Absorbed fraction in crystal	0.90	Doping, length, spectrum, temperature, and double-pass design
Gain length	10 mm	Selected crystal and absorption/thermal model
Cavity length	100 mm optical first pass	Package, mode size, FSR, and tolerance trade
Mirrors	Plane high reflector and \(R=200\,\mathrm{mm}\) output coupler	Coating availability and thermal-lens sweep
Output-coupler transmission	5%	Saturated-gain/output-coupling optimization
Other round-trip power loss	2% represented exponentially	Loss measurement and coating/scatter budget

2. Pump and efficiency estimate

From (1),

\[P_{\rm abs}=(0.90)(20\,\mathrm W)(0.90)=16.2\,\mathrm W.\]

The ideal Stokes efficiency is

\[\eta_{\rm Stokes}=\frac{808}{1064}=0.759,\]

so at least \(24.1\%\) of absorbed pump becomes quantum-defect heat before other losses. If a provisional absorbed-power threshold is 3 W and absorbed slope efficiency is 45%,

\[P_{\rm out}\simeq0.45(16.2-3.0)=5.94\,\mathrm W.\]

That passes the paper requirement but has little system margin. The next model must split transport, absorption, overlap, internal loss, output coupling, and thermal roll-over rather than tune the single 45% number.

If the practical heat fraction is provisionally 35%,

\[P_{\rm heat}\simeq0.35(16.2)=5.67\,\mathrm W.\]

This number drives the first thermal and mount model.

3. Cold-cavity eigenmode

For the plane-concave cavity,

\[g_1=1,\qquad g_2=1-\frac{100}{200}=0.5,\qquad g_1g_2=0.5.\]

The cold cavity is comfortably inside the ideal stability interval. With the waist at the plane mirror,

\[z_R=\sqrt{L(R-L)}=100\,\mathrm{mm},\]

\[w_0=\sqrt{\frac{\lambda z_R}{\pi}} \simeq184\,\mu\mathrm m,\]

and at the curved mirror

\[w(L)=w_0\sqrt{1+(L/z_R)^2}\simeq260\,\mu\mathrm m.\]

These numbers are not final because the crystal’s refractive surfaces and pump-dependent thermal lens belong in the round-trip matrix. They do establish the approximate pump-waist scale and optic clear-aperture requirement.

4. Threshold and circulating-power checks

Using (2) with \(R_1=0.999\), \(R_2=0.95\), \(L_g=0.010\,\mathrm m\), and \(\mathcal L_i=0.02\),

\[g_{\rm th}\simeq \frac{\ln[1/(0.999\times0.95)]+0.02}{0.020} \simeq3.6\,\mathrm{m^{-1}}.\]

The spectroscopy and inversion model must demonstrate this gain over the actual pumped volume.

A 5-W output through \(T=0.05\) implies roughly

\[P_{\rm circ}\simeq\frac{P_{\rm out}}{T}=100\,\mathrm W\]

incident on the output coupler in this simplified travelling-wave power accounting. For a Gaussian radius of 260 µm, its on-axis irradiance is

\[I_{\rm pk}=\frac{2P_{\rm circ}}{\pi w^2} \simeq9.4\times10^8\,\mathrm{W/m^2} =94\,\mathrm{kW/cm^2}.\]

Repeat this calculation at every optic and include standing-wave, defect, and transient enhancement before selecting coatings.

5. Pump overlap and thermal sweep

Start with an approximately 220-µm pump radius near the gain region so the pump slightly exceeds the cold TEM00 radius. Then calculate the full longitudinal absorbed-pump distribution and overlap integral in (2).

Insert the gain medium and a variable thermal lens into the cavity matrix. Sweep from cold through the worst credible hot dioptric power, including tolerances. For every point record:

stability trace and distance to both stability boundaries;
beam radius through the pumped volume and on every coating;
pump/mode overlap and expected higher-order-mode discrimination;
clear-aperture clipping loss;
waist location and external mode-matching change; and
mirror-tilt sensitivity.

If the thermal sweep crosses a stability boundary, change resonator geometry; do not merely plan to align more carefully.

6. Polarization, feedback, and mechanics

Use a polarization-selective element only if the gain/crystal geometry does not provide sufficient stable polarization. Budget the insertion loss and thermal depolarization. Tilt or wedge transmissive intracavity optics so ghosts cannot form a parasitic cavity, while tracking the astigmatism they introduce.

Mount the crystal with a modeled and repeatable thermal interface, allow differential expansion, and avoid stress concentrations. Choose mirror mounts whose angular drift and resonances fit the alignment-sensitivity budget. Add isolation or slight angle to prevent output-path feedback into the resonator and pump diode where permitted by system requirements.

7. Verification plan

Requirement or risk	Test	Pass evidence
Output and efficiency	Incident and absorbed pump; output-versus-pump at stabilized temperatures	At least 5 W with declared efficiency basis and roll-over margin
TEM00 beam quality	Two-axis second-moment caustic fit	\(M_x^2,M_y^2\le1.2\) with residuals and uncertainty
Polarization	Analyzer sweep over power and temperature	Required extinction ratio without mode or power instability
Thermal robustness	Pump steps with waist, mode, power, and coolant logging	No stability crossing; model updated from measured thermal lens
Alignment tolerance	Controlled mirror perturbation and environmental test	Requirement retained within allocated angular/positional range
Coating/damage margin	Irradiance budget plus qualified optic data and inspection	Required statistical margin at wavelength and exposure condition

The beam-quality row must be converted into the full acceptance statement described in Beam Quality: Specify, Measure, and Interpret; the coating row must use the local-exposure and qualification framework in Laser-Induced Damage Engineering.

8. Optional Q-switched extension

If the same 5-W average output is instead delivered at 20 kHz in 10-ns pulses,

\[E_p=250\,\mu\mathrm J, \qquad P_{\rm peak}\simeq25\,\mathrm{kW}.\]

At a 300-µm \(1/e^2\) radius, the Gaussian on-axis fluence is

\[F_{\rm pk}\simeq\frac{2E_p}{\pi w^2} \simeq0.177\,\mathrm{J/cm^2},\]

and peak irradiance is about \(17.7\,\mathrm{MW/cm^2}\). This immediately changes the Q-switch, coating, bulk-damage, nonlinear, detector, and beam-dump requirements even though average power is unchanged.

What makes this a strong interview answer

It produces numbers quickly but labels them as assumptions, connects optical and thermal models, treats the thermal lens as a range, checks intracavity rather than output power alone, and ends with measurements capable of disproving the model.