Back Issues

JUL/AUG 2013  

Follow us:

Find MICROmanufacturing on TwitterFind MICROmanufacturing on FacebookFind MICROmanufacturing on YouTubeMICROmanufacturing RSS feed

RPMC adds <200nm DPSS lasers to lineup

Editor's Note: RPMC Lasers Inc. has strengthened its ultraviolet-laser offerings by adding Actinix’s Model 3193 <200nm DPSS laser to its product line. Following is a white paper on the 3193.

3193 image

The Model 3193 solid-state 193nm laser.

The Model 3193 is a solid-state 193 nm laser providing a high spatial coherence for demanding metrology and microscopy applications. The system operates at five-kilohertz producing 5 mW average power. The system uses a proven, efficient far-UV light generation method that provides reliable, stable, hands-free operation for extended periods of use. The 3193 produces a FWHM bandwidth of 7 pm at 193 nm.

  • System Architecture

Light at 193.4 nm is generated by frequency-mixing a 266-nm ultraviolet beam with a 708-nm infra-red beam. Each of the interacting beams is generated from a common 532 nm pump laser, a JDSU Q-201 diode-pumped, doubled Nd:YAG laser operating at 5 kHz. This laser generates 15-ns-wide (FWHM) pulses with energy up to 2.5-mJ/pulse giving an average power of ~ 12.5 W. The UV beam at 266 nm is the second-harmonic of the pump laser light and is generated by Type-1 phase matching in a CLBO crystal. The IR beam at 708.6-nm is the signal wave of a 532-nm-pumped optical parametric oscillator (OPO). The IR and UV beams, each with power levels ~1 W, are combined using beam combiner. The flux-grown BBO crystal is uncoated and contained in a dry atmosphere within a sealed housing. A TEC is used to cool the crystal. The 193-nm light is separated from the other wavelengths with a Pellin-Broca prism. The 193-nm beam generated is nearly diffraction-limited (M2 values < 1.2 in both directions) with a TEM00 spatial mode.

  • Beam Quality

A typical beam image and profile for the 193 nm output beam from the Actinix 3193 light source are shown in the figure below. Also shown are fringes from two pinholes spaced 500 microns apart showing the high spatial coherence of the source.

mode

Measurements of the beam profile were made at five positions, at distances ranging from 194 mm to 2512 mm from the output aperture of the laser. At all positions the beams have very accurately Gaussian profiles, with a slight asymmetry between the x- and y-axes. For each image, the dimensions of the beam were determined by fitting slices of pixel values along the x- and y-axes to a Gaussian form. The results of these fits are shown in Table 1. To find the location and size of the beam waist, these data can be fit to the form that was first derived by Siegman,

image 2

Here w0 and z0 denote the beam waist and its location, and the parameter M2 is measure of “beam quality.”

Images of beam profile acquired with a Startech UV down-converter. Image (a) shows the beam at a distance of 194 mm from the output aperture of the laser; image (b) shows the beam 2512 mm from the output aperture. The images have the same scale, corresponding to 2.71 mm on a side.

Z [mm] Wx(z) [mm] Wy(z) [mm]
194 0.319 0.281
370 0.365 0.328
668 0.445 0.424
1557 0.711 0.729
2512 1.067 1.046

The results of such a fitting procedure are:

x direction y direction
w0,x = 0.297 mm w0,y = 0.254 mm
z0,x = -115 mm z0,y = -102 mm
M2= 1.12 M2 = 1.00

Note: this is not the ISO approved method for determining beam parameters; our intent here is to show that the beam is a very clean Gaussian with low M2.

  • Spectral Bandwidth

A Michelson interferometer is the appropriate instrument to measure the spectral properties of a moderate bandwidth light source such as the 3193. A series of interferograms was recorded with the laser running under normal operating conditions. The optical path difference of the interferometer arms was varied between -5.5 mm and +5.5 mm in steps of 0.50 mm. The figure below shows two examples of the fringe patterns observed at the output port of the interferometer. In one case the arms of the interferometer were of equal length, to within a few microns; in the other the movable mirror was translated by 1.0 mm, corresponding to an optical path difference of 2.0 mm and, for this laser, a fringe visibility of about 50%.

The 3193 laser has an exceptionally clean, Gaussian transverse profile, which is essential for the quantitative analysis of the images from the Michelson interferometer. Plots of intensity vs. position in a slice across the CCD sensor are shown below. The intensity variation can be described as a function of position x in the image plane by a Gaussian profile with sinusoidal modulation,

Here I0 is a constant background; the unmodulated beam profile is parameterized by the amplitude A, centroid coordinate x0, and width sbeam; and the interference fringes are parameterized by the visibility V , spatial frequency kx, and phase constant d. For each fringe profile, a nonlinear least-squares fitting routine adjusted these seven parameters for best agreement with the measured fringe pattern.

For the case of optimal fringe contrast, fringes (a) above, the fitting procedure gives V = 0.951; with the movable mirror translated by 1.0 mm, fringes (b) , the fitted visibility was V = 0.541.

The fitted fringe visibility V is plotted as a function of the optical path difference, s, of the interferometer arm (relative to an arbitrary origin).

The visibility has a Gaussian dependence on the length of the interferometer arm, as shown by the solid curve. The standard deviation of this curve is 0.991 mm; since the optical path length difference is twice the mirror displacement, the visibility expressed as a function of optical path difference has a standard deviation ss = 1.982 mm.

The measurements of V(s) for the 3193 laser produce a Gaussian line shape,

which corresponds to a Gaussian visibility function,

where

The FWHM bandwidth expressed in wavenumbers is

Expressed in frequency units with 1 cm-1 = 30 GHz,

In terms of wavelength, the spectral line shape is also Gaussian with a standard deviation of

The full width at half maximum (FWHM) as a measure of line width is given by

Using these relations with our measured data implies

The coherence length is then given as:

 

[The 3193 engineering development was funded in part by NSF grant 0450620 and by a joint development contract with IBM. In particular we'd like to thank Dr. John Hoffnagle for the spatial mode and spectral analysis work described here.]