Pulse Picker: How to Select Laser Pulses Precisely

A mode-locked laser emits a train of ultrashort pulses at a fixed, very high repetition rate—often tens of megahertz or more—but many experiments need only one pulse out of every hundred, thousand, or million. How do you keep the pulses you want and discard the rest without distorting the ones you keep? A pulse picker does exactly that: it uses a fast electro-optic switch, synchronized to the laser, to transmit selected pulses and block all others. This article explains how pulse picking works, what the Pockels effect and the driver electronics contribute, and how to specify a pulse picker for your laser wavelength and repetition rate.

Key Takeaways

• A pulse picker reduces the repetition rate of a pulsed laser by transmitting every Nth pulse (or a programmed pattern of pulses) and blocking the rest, using a Pockels cell between polarizers as a fast optical gate.
• The gate must open and close within the time between pulses. For an 80 MHz oscillator that interval is 12.5 ns, so the switching edges must be considerably faster than that while the transmitted pulse sits cleanly inside the open window.
• Conoptics pulse pickers cover 350 to 1600 nm and support pulse selection at repetition rates up to 30 MHz, so the same operating principle applies from ultraviolet-visible Ti:sapphire and frequency-doubled sources through to telecom-band fiber lasers.
• Selecting pulses raises pulse energy at a given average power, lowers average power at a given pulse energy, and leaves a long dark interval for slow processes such as fluorescence decay to finish.
• Key specification parameters are wavelength range, maximum selected-pulse rate, contrast (the suppression of unwanted pulses), and the dispersion the crystal adds to the pulses you keep.

What Is a Pulse Picker?

A pulse picker is an electro-optic device that selects individual pulses, or groups of pulses, from a continuous train and rejects the others. It sits in the beam path of a mode-locked laser, or between an oscillator and an amplifier, and it is synchronized to the laser through a trigger or a clock signal derived from the pulse train itself.

The need for pulse selection comes from a mismatch between how mode-locked lasers are built and how experiments use them. An oscillator’s repetition rate is set by its cavity length: a cavity roughly 1.9 m long round trip produces pulses about 80 million times per second. That is ideal for stable mode-locking and for averaging a weak signal, but it is a poor match for several common tasks:

• Seeding a regenerative or chirped-pulse amplifier, which can only amplify pulses at a rate (typically kilohertz) limited by the gain medium’s recovery and its pump source.
• Time-resolved measurements, such as fluorescence lifetime or pump-probe experiments, where the sample needs a long dark interval to relax before the next excitation. A fluorophore with a lifetime of several nanoseconds does not fully decay in the 12.5 ns between pulses of an 80 MHz source.
• Limiting thermal or photochemical load, where reducing the number of pulses at fixed pulse energy lowers average power on the sample without reducing the peak intensity that drives the nonlinear process.

How a Pulse Picker Works

Pulse picking combines three ideas: electro-optic phase retardation, polarization-based switching, and precise timing. The physics of the first two is the same as in any Pockels cell, and it is worth stating directly before turning to the timing.

The Pockels effect as a fast switch

In certain non-centrosymmetric crystals, an applied electric field changes the refractive index in proportion to the field. This is the linear electro-optic, or Pockels, effect. Because the crystal is birefringent, the field changes the relative phase delay between two orthogonal polarization components of the light. For a crystal of length L and electrode spacing d, the induced phase retardation scales as

Γ = π · V / V_π

where V is the applied voltage and V_π is the half-wave voltage, the voltage that produces a retardation of π (half a wavelength). When the input light is linearly polarized at 45° to the crystal axes and followed by a polarizer crossed with the input polarization, the transmitted fraction is

T = sin²(π V / 2V_π)

At V = 0 the transmission is zero, and at V = V_π it reaches unity. A pulse picker holds the crystal at zero volts (or at a bias that keeps the output blocked) for every pulse that should be rejected, and applies V_π for a brief window that surrounds the pulse that should pass.

Because V_π grows with wavelength and depends on the crystal material and geometry, the crystal and the driver must be matched to the laser. This is why a pulse picker is specified for a wavelength band: the 350 to 1600 nm coverage of Conoptics pulse pickers is achieved by choosing the appropriate crystal and coatings for each region, not by one design that is optimal everywhere.

Timing: fitting a gate between two pulses

The more demanding constraint is timing. The laser’s pulse period is

T_rep = 1 / f_rep

so an 80 MHz oscillator provides a 12.5 ns period, and a 30 MHz source provides about 33 ns. The Pockels cell driver has to do four things within that period:

• Rise from the blocking state to V_π before the selected pulse arrives.
• Hold V_π for the duration of the selected pulse (picoseconds to a few nanoseconds including any broadening).
• Fall back to the blocking state before the next pulse arrives, so the neighbor is rejected.
• Do all of this at the programmed selection rate, with jitter small compared with the gate width so every selected pulse sits in the same part of the window.

If the rising or falling edge is slow, the pulse adjacent to the selected one sees partial transmission, and the contrast between selected and rejected pulses degrades. This is why driver bandwidth, rise and fall time, and voltage stability matter as much as the crystal. A driver that supports pulse selection at up to 30 MHz has to produce a repeatable, clean waveform for every selected pulse at that rate, which sets how much bandwidth the output stage needs.

Selection patterns

The simplest mode divides the repetition rate by an integer: with N = 1000 and f_rep = 80 MHz, the output rate is 80 kHz. Programmable drivers also support burst modes, in which a group of consecutive pulses passes before the gate closes, and gated modes, in which pulses pass only while an external signal is active. In every mode the output pulse energy equals the input pulse energy times the insertion transmission of the cell and polarizers, since the picker does not add energy; it only removes pulses.

What Changes in the Pulses You Keep

A pulse picker is not perfectly transparent to the beam, and a technical reader should account for three effects.

• Insertion loss and polarization. The pulses that pass go through a crystal, polarizers, and antireflection coatings. Transmission is below 100%, and the input polarization must match the picker’s design axis.
• Dispersion. A crystal of several millimeters adds group delay dispersion, which broadens a femtosecond pulse. For very short pulses, the dispersion is usually pre-compensated or accounted for in the compressor downstream.
• Contrast. Residual leakage of rejected pulses, expressed as the ratio of selected to rejected pulse energy, determines how clean the selected train is. Leakage comes from imperfect polarization extinction, birefringence of the crystal at zero voltage, beam divergence through the crystal, and timing edges that overlap neighboring pulses.

Applications

• Amplifier seeding. A regenerative amplifier needs seed pulses at a rate matched to the pump and cavity timing. A pulse picker between the oscillator and the amplifier delivers one pulse per amplification cycle, so the amplifier does not waste gain on pulses it cannot store.
• Fluorescence lifetime imaging and time-correlated single-photon counting. Lowering the excitation rate to a few megahertz gives the fluorescence decay time to finish before the next pulse and prevents pile-up in the timing electronics.
• Nonlinear and multiphoton microscopy. Because two-photon signal depends on peak intensity squared, a lower repetition rate with the same pulse energy preserves signal per pulse while reducing the average power delivered to tissue.
• Pump-probe and ultrafast spectroscopy. Selecting pulses at a lower rate allows lock-in detection at a defined modulation frequency and gives the sample time to return to its ground state between excitations.
• Telecom-band and fiber-laser sources. With coverage out to 1600 nm, pulse selection applies to fiber-laser and OPO-based systems used in sensing, LIDAR, and communications research.

Specifying a Pulse Picker

Start from the laser and the experiment, and work toward the device. Four parameters decide most of the design.

• Wavelength and bandwidth. The crystal, coatings, and polarizers must support the central wavelength and, for femtosecond pulses, the full spectral bandwidth.
• Repetition rate and selection rate. The input repetition rate sets the gate window you have to fit between pulses, and the highest selected-pulse rate sets the driver’s required bandwidth and thermal load. Conoptics pulse pickers support pulse selection at up to 30 MHz.
• Contrast and edge speed. A higher contrast requirement calls for tighter beam alignment, a smaller beam divergence in the crystal, and faster driver edges.
• Beam size and peak intensity. The crystal aperture has to be large enough that the beam does not clip, and the peak intensity must stay below the damage threshold of the crystal and coatings, which matters especially after amplification.

Conclusion

A pulse picker turns a fixed-rate pulse train into the pulse pattern an experiment actually needs. The Pockels effect supplies a voltage-controlled polarization switch, the crossed polarizers convert that switch into intensity gating, and a fast driver synchronized to the laser fits the gate between neighboring pulses. Matching wavelength range, repetition rate, contrast, and dispersion to the application is what separates a picker that simply passes pulses from one that delivers a clean, repeatable output train.

Frequently Asked Questions

What is the difference between a pulse picker and a pulse selector?

The terms are used interchangeably. Both describe an electro-optic gate that transmits chosen pulses from a laser pulse train and rejects the rest.

How fast does the Pockels cell have to switch?

Fast enough to rise, hold, and fall within one pulse period. For an 80 MHz oscillator, that period is 12.5 ns, so the edges must be a fraction of that interval. Slower edges transmit part of the neighboring pulses and reduce contrast.

Can a pulse picker select pulses from any mode-locked laser?

It can when the wavelength, polarization, beam size, and repetition rate match the picker’s design. Conoptics pulse pickers cover 350 to 1600 nm and pulse selection at repetition rates up to 30 MHz. A trigger or clock synchronized to the pulse train is required.

Does pulse picking change the pulse energy or duration?

It does not add energy, and it transmits selected pulses with the insertion transmission of the crystal and polarizers. The crystal adds some dispersion, which broadens very short pulses slightly unless it is pre-compensated.

What limits the contrast between selected and rejected pulses?

Polarizer extinction, residual birefringence at zero voltage, beam divergence in the crystal, and driver edge speed all contribute. Higher contrast generally requires better alignment, a collimated beam, and faster, cleaner switching waveforms.

When should I use a pulse picker rather than reducing the laser’s repetition rate directly?

A mode-locked oscillator’s repetition rate is fixed by its cavity length, so reducing it means rebuilding the cavity and usually compromises stability. A pulse picker leaves the oscillator unchanged and provides a lower, programmable output rate.

Is pulse picking useful at high repetition rates, or only for amplifier seeding?

Both. At several megahertz, pulse picking supports fluorescence lifetime imaging and pump-probe work. At kilohertz rates, it seeds regenerative amplifiers. The principle is the same; the driver and timing requirements differ.

You Might Also Like

• Pockels Cell Driver: Essential for Laser Precision
• Pockels Cells vs. AOMs: Choosing the Right Switch
• Electro-Optic Camera: Capturing Light at Picosecond Speed
• Optical Pulse Modulation Systems: Fundamentals and Applications