NCLA - Ultrafast Laser Materials Processing

Laser Model	Clark MXR CPA2001
Pulse Width (FWHM)	150fs
Maximum Pulse Energy	800 uJ
Maximum Repetition Rate	1 kHz
Output stability	1% rms
Average Power	0.8 W
Polarisation state	Linear Horizontal
M²	1.4
Beam profile	Gaussian, round

The Clark system is based on the Chirped Pulse Amplification technique. The laser machining centre in the NCLA (refer figure 1.) was built around this laser source and it consists of laser beam monitoring instrumentation, optical attenuation, polarisation rotation, steering optics, high damage threshold magnification and focussing optics, a high precision CNC system, a vacuum chuck and a light vacuum chamber. All of the above components are housed in a climate controlled laboratory on a vibration isolating, air driven optical bench. A GUI (graphical user interface) was programmed to unify the various components into one console. The set-up is subject to ongoing development to incorporate increased capability including vision and in-process tuning of the laser pulse energy and the laser beam polarisation state.

System Performance � M²

Measuring the M² parameter for pulses of such high intensity is not straightforward and much debate surrounds the applicability of the ISO standard for laser beam quality in relation to the measurement of femtosecond pulses. The ISO standard was used in the measurements as there was no better method available at the time of the study. Two sets of measurements were performed, one set on the standard �raw� beam output, and one on the laser beam when passed through a pinhole aperture. The basic arrangement is shown in figure 1. The output of the laser was sampled by a CCD camera (�DataRay Wincam�) and attenuated with optical filters such that the peak response of the camera was the same at each position along the beam path. The laser beam was passed through a lens to the detector which was moved successively along the beam path at known distances. At each position along the beam path the diameter of the beam was measured in the x and y directions (D_x and D_y ) using the camera and the second moment method, and the average of three readings was taken for each data point presented. The data is graphed in figure 2. What is immediately evident from the graph is that the beam appears more radially symmetric after the beam waist position, which is at approximately 525 mm from the lens in the x direction, Z₀(x), and 550mm in the y direction, Z₀(y). It also indicates that the beam waists appear to be offset somewhat from each other in the two orthogonal directions, x and y


Figure 1: Experimental arrangement used for measuring M² (� NCLA)

Figure 2: Laser beam diameter in the x and y directions with propagation distance. (�NCLA)

It must be noted however that there is a great possibility for error in the measurement of the beam diameters at the waist of the beam. There are a number of reasons for this, chiefly the intensity is greatest there and non-linear effects are significant, the beam diameter is smallest there and hence the resolution of the camera image is least, also the non-linear effects of using filters to attenuate the laser beam energy so that the camera does not saturate contributes to the errors as well. The greater roundness after the focus suggests that there is also a non linear process happening in the region of the focus which is improving the beam quality somewhat in the far field. It would appear that the laser beam is somewhat astigmatic as well. The M² values in the two directions were then calculated from a re-arrangement of the ISO standard formula:

(1)

where: M² is the beam quality factor, D₀ is the diameter of the laser beam at the
waist, D_z is the diameter of the laser beam at a distance Z from the waist and l is the wavelength of the laser light. These calculations were performed for positions in x and y outside the beam waist only, and they are shown in figure 3.

Figure 3: M² of the �raw� laser beam with distance from the beam waist � NCLA

It is evident that the values closest to the beam waist show most discrepancy from the average values and as the x and y beam waists have an apparent offset of 25mm from each other when using a long focal length lens an error of �12.5mm can be substituted into the equation for M² for each value along the beam propagation to examine the relative errors with distance from the beam waist. When this calculation was carried out it was evident that the largest errors were indeed associated with the value nearest the beam waist, refer to table 1. The other experimental points were found to be reliable. Examining the table only values of M² relatively far away from the beam waist were used for the average calculations.

Table 1: M² errors with propagation distance from beam waist (� NCLA)

Distance From beam waists z_x ,z_y (mm)	M²_x	M²_x error	M²_y	M²_y error
25	0.950	0.475	2.857	1.429
100	1.355	0.169	2.181	0.272
125	1.318	0.132	2.395	0.239
150	1.379	0.115	2.252	0.188
175	1.317	0.094	2.205	0.157
200	1.302	0.081	2.194	0.137
225	1.301	0.072	2.265	0.126

Within the range of experimental error the values of M² were found to be: M²_x = 1.318 � 0.131 and M²_y = 2.276 � 0.218 for the raw beam. The identical procedure was carried out with a pinhole with a 3mm diameter in place before the lens, and this had the effect of transforming the Gaussian profile of the beam into a more �top-hat� beam and of transforming the elliptical beam into a more regular circular laser beam. The �raw� beam profile measured by the camera is shown in the left of figure 4, and the �pinhole� beam profile is shown in the right hand frame of figure 4. At the low peak responses used on the camera (33.7% �raw� case, 50.2% �pinhole� case) the �raw� beam exhibits reasonable goodness to fit for the Gaussian case and the �pinhole� case exhibits excellent roundness albeit with some diffraction patterns. As the degree of ellipticity was less then 15% at all distances along the beam path when the pinhole was in place the beam was deemed radially symmetric according to the ISO standard and the calculations that followed assumed one value of M² for the laser beam accordingly. Using the same arrangement as in figure 1, except with the pinhole in place the values of M² were obtained and are shown in figure 5. Ignoring the first data point which is again very close to the beam waist and has a high error associated with it, the average value for M² = 1.325 � 0.098 for the laser beam propagating through the pinhole.

Figure 4: Raw beam profile exhibiting some ellipticity (l.h.s) and �Pinhole� beam profile (r.h.s) exhibiting a more top hat profile and greater circularity (� NCLA)

The importance of these results is significant when using the system for different types of machining. For percussion machining of micron range diameter round holes, laser beam circularity and intensity distribution are important parameters to be controlled in order to achieve good machining results.

Pulse duration and Pulse energy

Very high intensities are available with femtosecond pulses but in many practical applications only a fraction of the available pulse energy of the laser beam when focused is required to operate around the ablation threshold. This is especially the case when employing short focal length objectives for producing the tiniest of features in the few micron diameter range. In the current set-up a 1/2 waveplate and a beam spitting polarisation cube are used to attenuate the beam. With this set-up and with an appropriate pulse energy meter it is possible to attenuate the beam down to the �J and nJ pulse energy levels. The method relies on the stability of the output pulse energy from the laser and the accuracy of the measurement method, and it was found that when the laser cavity was properly tuned the stability was good. Figure 1 exhibits the pulse energy stability when the mean pulse energy was 7.62�J and the pulse repetition rate was 1kHz. The minimum reading was 6.36�J and the maximum reading was 9.17�J. The standard deviation was found to be 4.67e-7J. The rms value for pulse to pulse stability was estimated at 6% for the sub 10�J pulse energy values. At even lower pulse energies it became difficult to separate laser pulse energy stability from noise in the detector signal as the nJ pulse energy probe was extremely sensitive. However, within the measurable range in the set-up used here - 0.5�J to 800�J, the laser pulse energy stability was found to vary by up to 6% rms.

Figure 1: Pulse energy distribution at a mean pulse energy of 7.6�J (� NCLA)

In terms of pulse width the compressor grating setting can be adjusted until the desired pulse width is obtained with the Clark system, and the prime driver in the experiments reported here was to have the shortest pulses available and pulse width stability. The following autocorrelator trace shows a FWHM of approximately 150fs (when transformed) for the laser when operating at 100Hz and the device was sampling for one second, at the minimum gain available to ensure the device was not saturated. This was the shortest pulse width achieved.

Figure 2: � Autocorrelator trace FWHM 210fs which for a Gaussian beam corresponds to 150fs actual pulse length (Transfer constant is 1.4 for a Gaussian beam, vertical scale arbitrary units) ( �NCLA)