2.2. Detection System
The optical emissions from the plasma plume were collected by a spectrograph (SpectraPro 2300i) and recorded with a Hamamatsu streak camera (model C4334) with an integrated video streak camera (
Figure 1). The streak images were time-resolved, thus enabling the monitoring of temporal evolution of the ionic and atomic emission lines [
7,
8,
9], or spatial development of the plasma. The fundamental advantage of the streak scope was its two-dimensional nature, which was especially important for measuring time-resolved LIBS spectra. The camera had a spectral range from 200 to 850 nm. The CCD chip had a resolution of 640 × 480 pixels. The data were acquired and analyzed using High Performance Digital Temporal Analyzer (HPD-TA) software provided by Hamamatsu.
The spectrograph contained a triple grating turret. Diffraction gratings of 50, 150, and 300 grooves/mm were installed. In the place of the 150 g/mm grating, we mounted the plain mirror (see
Figure 2). Thus, when grating of 150 grooves/mm was selected by HPD-TA software, the streak camera, instead of the image of the optical spectrum, took the image of the spatial distribution of the optical emissions of the laser-induced breakdown. To utilize as much of the CCD camera active area as possible, the maximal size of the spectrograph entrance slit was used. The diffraction grating of 50 grooves/mm was used when we required a wide observing wavelength window, and the grating of 300 grooves/mm was used when a better optical resolution was needed. Other optical parts of the acquisition system were chosen to obtain an overall optical magnification of 0.6. In this case, the calibration procedure showed that 1 mm on the target position corresponded to 72 pixels of the CCD camera.
To take the streak image in the time frame of interest, the proper delay time must be set on the digital delay generator (
Stanford DG 535), which triggers the streak camera (see
Figure 1). In our setup, for camera time scales up to 200 ns, the laser Q-switch trigger out-signal was used to trigger the streak camera. We used a fast 1-GHz photodiode and digital oscilloscope (Tektronix TDS 5032) to determine the time interval between the Q-switch trigger and the laser pulse (see
Figure 3). The laser excitation pulse was partially reflected by the beam splitter, and acquired in an attenuated form by the fast photodiode. The photodiode was chosen to be sensitive both to the fundamental (1064 nm) and second harmonic output (532 nm) of the laser. There was a significant delay between the Q-switch, the top trace (shown in
Figure 3), and the laser firing recorded with fast photodiode, as shown in the bottom trace in
Figure 3. A similar problem concerning the acquiring of the streak image in the time frame of interest was solved by Mohamed and Kadowaki [
11] by using an image light scope.
To determine the time that a streak image begins relative to the Q-switch trigger signal and the laser pulse, we had to tabulate the important time parameters of the detection system, provided in
Table 1. The output energy of our laser was varied by a laser controller, by setting the different timings of the Q-switch. Thus, the time interval between the Q-switch signal and the laser pulse was a consequence of the “percent of laser energy” parameter set by the operator on the laser control unit. Moreover, the same “percent of laser energy” corresponded to the same timings and different energy levels of the laser’s fundamental output and the laser’s second harmonic output. In the measurements presented in this paper, for example, 100% of the laser energy meant 270 mJ on the fundamental harmonic or 68 mJ on the second harmonic. To measure the values presented in
Table 1, the second harmonic of the laser was used as an excitation source.
Looking at
Figure 3 and
Table 1, it is easy to see that the time interval between the Q-switch trigger and the laser pulse was more or less longer than the delay set on the delay generator. So, when we say the “camera dead time”, we mean the time difference between the camera trigger signal and the moment when the camera is capable of acquiring the streak image. Because of this dead time, the streak camera needed to be triggered in advance of the laser pulse using the delay generator triggered by the Q-switch.
Since the spectral range of our streak camera was in the interval from 200 to 850 nm, the fundamental harmonic of our excitation pulse (at 1064 nm) could not be acquired by the streak camera. We recorded waveforms of the laser’s fundamental harmonic (1064 nm) and second harmonic (532 nm) using a fast photodiode and digital oscilloscope. The oscilloscope was triggered by the Q-switch, using an internal trigger delay. The delay between the fundamental and the second harmonic generator (SHG) was determined to be 2 ns, as shown in
Figure 4. The delay came from a longer optical path passing through the SHG. Measuring the length of the optical path and calculating the time by using the known value of velocity of light gave the same result as that obtained from
Figure 4.
To analyze plasma development, it is necessary to determine the initiation and duration of the excitation-laser pulse on streak images. Excitation-laser pulse is visible in
Figure 5 (at 532 nm), where the streak image of the optical spectrum of laser-induced air plasma is presented. The streak images are usually presented in pseudo-color, where different intensities are coded as different colors. However, for laser excitation at 1064 nm, or for spatial streak images, laser excitation is not necessarily visible.
2.3. Determination of Synchronization Timing Using the Fast Photodiode
Two problems can be solved by acquiring the laser excitation signal using the fast photodiode. First, the proper setting of the delay time generator used for triggering the streak camera can be calculated. If this time is not properly set, the time window of interest will not be acquired by the streak camera; usually, just a blank screen with some noise will be recorded. The proper setting of the delay time generator can be read from
Table 1, based on selected experimental parameters, and acquiring the time interval between the Q-switch and laser firing by using the fast photodiode and digital oscilloscope, as depicted in
Figure 3.
Moreover, if the start of the laser pulse is not visible on the streak image, the timing of the plasma development (recorded on the streak image) regarding the laser excitation could be calculated using the fast photodiode signal. The calculation was performed with the data provided in
Table 1. The laser pulse position from the top of the streak image equaled the difference in the delay between the Q-switch trigger signal (recorded by oscilloscope) and the laser pulse, the sum of the time set on delay generator, and the camera’s dead time. If the calculated time is negative, the laser pulse has begun before the time frame visible on the streak camera screen.
2.4. Determination of Synchronization Timing by Recording the Laser Pulse on the Streak Image
There is no doubt that the determination of synchronization timing can be best achieved by recording the laser pulse on the same streak image as the plasma optical emission.
When the second harmonic of the laser at 532 nm was used as an excitation source, the laser signal was made visible and recorded on spatial streak image, as follows. By the appropriate placement of the neutral optical attenuator on the optical axis of the camera detection system, the plasma optical emissions and the elastic scattering of the laser beam from air molecules were recorded simultaneously on the same streak image (see
Figure 6). The light attenuator enabled the camera to “see” the plasma breakdown (attenuated by attenuator) on the left-hand side and the laser excitation scattering (not attenuated) on the right-hand side of the focal point. The use of the attenuator was necessary to allow the optical signals of the plasma and scattered laser to have similar values; otherwise, after adjusting the gain of the detection system to match the intensity of the plasma optical emission, the laser scattering signal would not have sufficient intensity to be recorded. We discarded the original idea of transmitting the part of laser beam to the streak camera by a beam splitter as it was too risky.
Before performing any timing calculations, the streak image was corrected for possible geometric distortion. Looking at
Figure 5, the horizontal tilting of the streak image can be easily seen. The tilting of the streak images stems from the fact that the deflection of streak sweep is not completely straight, but rather, elliptic. This results in a geometric distortion of the streak image in the sweep direction. The distortion, which is always present on original streak images, is not obvious when looking at
Figure 6. To make time calculations based on streak images, this distortion should be corrected using the curvature correction tool provided by the camera software.
When the more powerful first harmonic of our laser was used as the excitation source, the situation was more complicated. To make laser signal at 1064 nm visible, we used a very low concentration of Rhodamine B dye embedded in PMMA thin film. On the time scales of interest in the present study, the time delay of the Rhodamine B up-conversion fluorescence response was negligible, as proven by our fluorescence measurements. Almost all of the laser beam energy was transmitted through the thin PMMA film, placed about one centimeter from the focal point at an acute (sharp) angle, but not equal to 45 degrees relative to the beam, to avoid even the partial reflection of the laser beam to the camera. The Rhodamine B fluorescence, now visible by the camera, was recorded similarly to the laser scattering shown in
Figure 6. It should be noted that, in this case, only the position of the raised edge of the laser excitation was correctly acquired. Again, the use of an optical attenuator for plasma emissions was mandatory.