[Instrument usage skills] Working principle and adjustment method of autocorrelator (1)

[Instrument usage skills] Working principle and adjustment method of autocorrelator (1)

Laser pulse width generally refers to the full width at half maximum of the pulse. As the name suggests, it is the full pulse width (FWHM) where the pulse height is 50% of the maximum value.

With the advent of lasers in 1960, pulse lasers have been widely used due to their high peak power. Currently, pulse lasers are used as light sources in chemical reaction dynamics, nonlinear optics, laser processing, laser ranging and other scientific and technological fields. With the development of mode-locked lasers, the output pulse width of pulse lasers has been shortened from the ns level to the ps and fs levels.

For general ns pulse lasers, a fast-response photodiode (such as EOT series, rise and fall time is less than 1ns) can be used to convert the optical signal into an electrical signal, and display it on an oscilloscope with a large bandwidth (for example, a bandwidth greater than 350MHz), so that the pulse width to be measured can be directly read. However, for such short pulses as ps or fs, it has exceeded the response time of the photodiode, so it can only be measured with a stripe camera and a scanning autocorrelator. Of course, Frog and Spider can also measure the pulse width and phase of ultrashort pulses, but their optical paths and post-data processing are more complicated than those of autocorrelators, so I won’t go into details here.

The single-shot autocorrelator is a new instrument developed over the past ten years specifically for measuring pulse width. It has the advantages of high resolution, high sensitivity and ease of use. There are currently many models of autocorrelators that can be used to detect the instantaneous width of ultrashort optical pulses, providing optimal sensitivity and braiding rate, and are suitable for measuring fs pulses of mode-locked dye or sapphire lasers and ps pulses of pulsed semiconductor lasers or Nd-YAG/YLF lasers.

Coherent Inc. Autocorrelator

Based on Coherent Inc.'s autocorrelator for 800nm ​​wavelength, here is an explanation of its working principle and adjustment method. Figure 1 below is the optical path diagram of the autocorrelator:

Figure 1. (a) Optical path diagram and (b) actual structure diagram of Coherent Inc.'s autocorrelator SSA

From Figure 1 above, the structure of the autocorrelator is similar to the Michelson interferometer. According to its nonlinear matching form, it can be divided into collinear structure or non-collinear structure. Because the non-collinear form can naturally separate the measured sum-frequency signal and the original fundamental frequency signal, thereby eliminating the background fundamental frequency light and achieving higher measurement accuracy. The autocorrelator SSA in this example has a non-collinear structure.

Note that the beam splitter BS1 is used here to split the input light, that is, the transmitted light a and the reflected light b. The transmitted light a is incident on the delay device DL. After passing through the reflector M2, the reflected light b must pass through a beam splitter BS2 that is the same as the previous BS1 to compensate for the dispersion and chirp of the reflected light b relative to the transmitted light a. The half-wave plate here is used to change the polarization of the incident light, thereby adjusting the intensity of the sum-frequency signal generated after the light to be measured passes through the nonlinear crystal BBO.

The initial incident light I(t) (here, the center wavelength of 800nm is taken as an example) passes through BS1 in Figure 1(a) and is divided into two beams I(t) and I(t-τ). After passing through the reflector M5, it is injected into the non-linear crystal BBO in a non-collinear form. By adjusting the angle of the delayer DL and the BBO crystal, the transmitted light a and the reflected light b pass through the crystal to generate the sum frequency signal S(τ), that is, a blue bright line, as shown in Figure 2. (b). When each beam of light is injected into the BBO crystal alone, no second harmonic is generated. The sum-frequency optical signal S(τ) is only related to the intensity of the two incident lights:

The generated sum-frequency signal S(τ) is received and recorded by the photomultiplier tube, and can be displayed on an oscilloscope, as shown in Figure 2. (a).

Figure 2. (a) The oscilloscope displays the electrical pulse; (b) The sum-frequency signal after the nonlinear crystal

In this way, the pulse width (FWHM) displayed on the oscilloscope has a certain linear relationship with the actual light pulse width (FWHM). That is:

It can be seen that using the light pulse intensity correlation method to measure the pulse width is essentially to convert the measurement of the invisible time into the measurement of the visible length, and to convert the measurement of the light pulse shape into the measurement of the correlation function S(τ) waveform. The time interval of its half-width (FWHM) is the pulse width (FWHM).

The proportional coefficient a is different for different types of autocorrelators. For details, please refer to the instruction manual of the instrument. However, even if it is the same instrument, its proportional coefficient will be slightly adjusted depending on the adjustment of the optical path or the frequency multiplication state, and it needs to be calibrated in time.

The proportional coefficient a can be determined by the following method:

As shown in Figure 1. (a), when moving the delay line DL, you can see the blue line after the sum frequency crystal moving up and down, and at the same time, the autocorrelation signal displayed on the oscilloscope moves accordingly on the time axis. For example, when the delay line moves 0.24mm (the optical path of the light moves Δ1 = 0.48mm), the pulse displayed on the oscilloscope moves 5.406ms on the time line,

That is, the actual time change caused by moving the delayer is:

The time displacement of the pulse displayed on the oscilloscope is:

The actual pulse width also needs to consider the waveform of the laser pulse. Assuming that the optical pulse we input here is Gaussian, it needs to be multiplied by a transformation coefficient k=0.707

Then the proportional coefficient a

If the half-height width of the pulse measured on the oscilloscope is τ_scope=152.4μs, that is, the corresponding pulse width is

Attached, if the laser pulse waveform is Gaussian, hyperbolic secant, or unilateral exponential, the transformation coefficients are 0.707, 0.648, and 0.5 respectively.

Based on the basic principle of the autocorrelator, its structure can also be adopted in different ways. In addition to collinear structures and non-collinear structures, delayers can also adopt different structures, such as a movable pair of mirrors at right angles to each other, a movable right-angle prism, a rotating glass plate, etc. Its light splitting structure can also be divided into splitting light with a beam splitter and splitting light in space according to the energy and pulse width of the light to be measured. Detectors can also be of different types, such as photoelectric probes, CCD detectors, etc.

As for the optical path adjustment of the autocorrelator, we will introduce it in detail in the next issue.