Introduction: Why Grating Size Matters in CPA Stretchers
In Chirped Pulse Amplification (CPA) systems, the pulse stretcher is the first critical component. It introduces Group Delay Dispersion (GDD) to stretch femtosecond pulses to nanosecond durations, reducing peak power to protect amplifier gain media from optical damage. The grating pair inside the stretcher must be large enough to accommodate the full spectral spread of the beam — if not, beam clipping degrades pulse quality and can permanently damage expensive optics.
This guide walks through the complete 5-step calculation chain from initial pulse parameters to final grating dimensions, with an interactive calculator you can use right now.
The 5-Step Calculation Chain
Designing a grating stretcher requires solving a chain of coupled equations. Each step feeds into the next:
- FT-Limited Pulse Width τ_in — From spectral bandwidth via the time-bandwidth product
- Required GDD — From the desired stretch ratio τ_out/τ_in
- Grating Angles — Littrow configuration and diffraction angles across the full spectrum
- Grating Separation L — Physical distance between gratings in each stretching stage
- Grating Physical Dimensions — Minimum width (dispersion axis) and height (non-dispersion axis)
Step 1: Fourier Transform-Limited Pulse Width
For a transform-limited pulse, the time-bandwidth product (TBP) relates spectral bandwidth to the shortest achievable pulse duration:
τ_in = C_B × λ₀² / (c × Δλ)
where C_B is the TBP constant: 0.441 for Gaussian, 0.315 for sech², and 0.142 for Lorentzian pulse shapes. For a typical Ti:Sapphire system (λ₀ = 800 nm, Δλ = 25 nm), this gives τ_in ≈ 37.6 fs.
Step 2: Required Group Delay Dispersion (GDD)
The GDD needed to stretch a pulse from τ_in to τ_out is:
GDD_total = τ_in / (4 ln 2) × √(τ_out² − τ_in²)
For multi-stage stretching (N stages), the total GDD is split equally: GDD_stage = GDD_total / N. Multi-stage designs dramatically reduce the required grating separation L, but each additional stage adds diffraction losses (4 grating hits per double-pass stage).
Step 3: Grating Angles
The Littrow angle (where incidence angle equals diffraction angle) is the optimal operating point:
sin(θ_Littrow) = m N λ₀ / 2
The diffraction angles for the spectral extremes (λ_min and λ_max) determine the spatial spread of the beam on the second grating. Critical note: Always use the full spectral width (typically 2× FWHM), not just the FWHM. Using only FWHM underestimates grating size by over 50%.
Step 4: Grating Separation
The perpendicular distance between gratings for each stage:
L = |GDD_stage| × π c² d² cos³θ / λ₀³
The cos³θ term means high line-density gratings (large incidence angle) dramatically shorten L — but at the cost of wider spectral spread on the second grating, demanding larger optics.
Step 5: Physical Grating Dimensions
Dispersion Axis (Width)
The second grating (G2) is always the bottleneck. Its minimum width must accommodate:
- Spectral dispersion spread across the full bandwidth
- Beam footprint at each spectral component
- Edge safety margins (typically 3 mm per side)
- Substrate walk-off for transmission gratings
Non-Dispersion Axis (Height)
A double-pass stretcher produces 4 symmetric beam spots on each grating. The minimum height must cover all 4 spots plus safety margins:
H_min = (N_beams − 1) × spacing + D₀ + 2A
Interactive Calculator
Use this calculator to design your stretcher. Enter your pulse parameters, grating specs, and beam size — it computes all 5 steps including a visual beam clipping diagram and inventory grating compatibility check:
Practical Example: Gitterwerk Transmission Grating
Consider designing a stretcher for a Ti:Sapphire CPA system using Gitterwerk 664287 transmission gratings (133 × 32 mm, 909.1 l/mm, efficiency ≥97.5%):
- Spectral FWHM: 25 nm at 800 nm center wavelength
- Target stretched pulse: 1 ns
- 2-stage stretching design
With 2-stage stretching, each stage requires only ~47 mm grating separation — an extremely compact design. The dispersion axis margin is ~10.5 mm, requiring careful alignment but entirely feasible in practice. Total throughput is 81.7% (each stage: 97.5%⁴ = 90.4%).
Key Design Considerations
- Full width vs. FWHM: Always calculate grating sizes using full spectral width (2× FWHM recommended). Using only FWHM underestimates D₂ by over 50%.
- Multi-stage trade-off: Each additional stage halves L and D₂, but adds 4 more grating diffraction events. At 97.5% efficiency per hit: 2-stage = 81.7% total, 3-stage = 73.8%.
- Transmission grating walk-off: At steep incidence angles, the substrate causes significant lateral beam displacement. At 68.7° with 3 mm fused silica: ~5.2 mm walk-off — half the margin budget.
- Non-dispersion height: Don't forget the 4-beam footprint in double-pass configuration. Many engineers only check the dispersion axis and discover height clipping too late.
- Incidence angle vs. efficiency: High line-density gratings give compact L but demand high polarization-dependent diffraction efficiency at steep angles. Always verify the manufacturer's efficiency curve at your operating angle.