^{1,a)}, S. Bouillet

^{1}, R. Courchinoux

^{1}, T. Donval

^{1}, M. Josse

^{1}, J.-C. Poncetta

^{1}and H. Bercegol

^{1}

### Abstract

Known for more than , laser damage phenomena have not been measured reproducibly up to now. Laser resistance of optical components is decreased by the presence of material defects, the distribution of which can initiate a distribution of damage sites. A raster scan test procedure has been used for several years in order to determine laser damage density of large aperture UV fused silica optics. This procedure was improved in terms of accuracy and repeatability. We describe the equipment, test procedure, and data analysis to perform this damage test of large aperture optics with small beams. The originality of the refined procedure is that a shot to shot correlation is performed between the damage occurrence and the corresponding fluence by recording beam parameters of hundreds of thousands of shots during the test at . We characterize the distribution of damaging defects by the fluence at which they cause damage. Because tests are realized with small Gaussian beams (about at ), beam overlap and beam shape are two key parameters which have to be taken into account in order to determine damage density. After complete data analysis and treatment, we reached a repeatable metrology of laser damage performance. The measurement is destructive for the sample. However, the consideration of error bars on defect distributions in a series of parts allows us to compare data with other installations. This will permit to look for reproducibility, a necessary condition in order to test theoretical predictions.

We would like to gratefully acknowledge Pierre Grua, Jean-Pierre Morreeuw, and Jérôme Néauport for numerous discussions during the long gestation of this work.

I. INTRODUCTION

II. PHYSICAL BASES FOR THE MEASUREMENT OF DAMAGE DENSITY

III. RASTER SCAN PROCEDURE

IV. DATA TREATMENT TO TAKE INTO ACCOUNT BEAM SHAPE AND EXPERIMENTAL MARGINS

A. Derivation of damage density, function of local fluence

1. At high fluences, above (see Fig. 7)

2. At low damage fluences, below

B. Calculation of error bars

V. ANALYSIS AND DISCUSSION

VI. CONCLUSION

### Key Topics

- Laser damage
- 15.0
- Density measurement
- 11.0
- Crystal defects
- 10.0
- Laser theory
- 7.0
- Metrology
- 6.0

## Figures

(Color online) Damage density vs fluence measured with previous procedure (Ref. 6). Six different areas have been tested at six different predetermined control fluences. Each value of damage density is the ratio (number of damage)/(scanned area) for one area, i.e., one control fluence. This measurement is now replaced by that reported in this paper.

(Color online) Damage density vs fluence measured with previous procedure (Ref. 6). Six different areas have been tested at six different predetermined control fluences. Each value of damage density is the ratio (number of damage)/(scanned area) for one area, i.e., one control fluence. This measurement is now replaced by that reported in this paper.

(Color online) Synoptic of the whole procedure, with reference to the sections of the text. The purpose is to replace the raw data of Fig. 1 by a more accurate set of data.

(Color online) Synoptic of the whole procedure, with reference to the sections of the text. The purpose is to replace the raw data of Fig. 1 by a more accurate set of data.

(Color online) Setup for laser damage testing.

(Color online) Setup for laser damage testing.

(Color online) On the left, the successive laser pulses overlap spatially to achieve an almost uniform scanning over the area dedicated to the test. Several areas are fired at several predetermined control fluences. On the right, the beam overlap is reported.

(Color online) On the left, the successive laser pulses overlap spatially to achieve an almost uniform scanning over the area dedicated to the test. Several areas are fired at several predetermined control fluences. On the right, the beam overlap is reported.

(Color online) Peak fluence distribution over thousands of shots at a “control” fluence of . In this plot, a set of measurements is represented, ranked from the lowest fluence to the highest, as a function of probability expressed in units of standard deviation. A normal distribution is a straight line in this kind of map, with mean value at and standard deviation equal to the slope. One can see that fluences can be approximately described by a normal law with a mean value of and a standard deviation of .

(Color online) Peak fluence distribution over thousands of shots at a “control” fluence of . In this plot, a set of measurements is represented, ranked from the lowest fluence to the highest, as a function of probability expressed in units of standard deviation. A normal distribution is a straight line in this kind of map, with mean value at and standard deviation equal to the slope. One can see that fluences can be approximately described by a normal law with a mean value of and a standard deviation of .

(Color online) Setup for the postmortem observation of samples.

(Color online) Setup for the postmortem observation of samples.

(Color online) Example of result expressed as density vs peak fluence (after Sec. III) of raster scanning a fused silica window at by means of tripled Nd:YAG laser. In this case, the mean peak fluence is . Here, each shot is taken to irradiate a surface area equal to the equivalent area of the Gaussian part of the beam. From this typical curve, fluence axis may be divided into two parts, above and below .

(Color online) Example of result expressed as density vs peak fluence (after Sec. III) of raster scanning a fused silica window at by means of tripled Nd:YAG laser. In this case, the mean peak fluence is . Here, each shot is taken to irradiate a surface area equal to the equivalent area of the Gaussian part of the beam. From this typical curve, fluence axis may be divided into two parts, above and below .

(Color online) Damage density vs fluence after treatment taking care of beam shape and to derive experimental uncertainty. Circles are the same as in Fig. 7; for those, fluence is peak fluence of the beam. Squares represent treated data; in this case fluence is the local fluence. Data on plateau are issue from relation (8) and those in the high fluence range from relation (7). Error bar calculations are explained in Sec. IV. Let us note that the error bars on log density appear to be roughly the same whatever fluences. This is due to the fact that during the test procedure, the number of shots decreases with increasing fluence (see Fig. 5). Damage probability increases with fluence, then the number of damage sites is almost constant.

(Color online) Damage density vs fluence after treatment taking care of beam shape and to derive experimental uncertainty. Circles are the same as in Fig. 7; for those, fluence is peak fluence of the beam. Squares represent treated data; in this case fluence is the local fluence. Data on plateau are issue from relation (8) and those in the high fluence range from relation (7). Error bar calculations are explained in Sec. IV. Let us note that the error bars on log density appear to be roughly the same whatever fluences. This is due to the fact that during the test procedure, the number of shots decreases with increasing fluence (see Fig. 5). Damage probability increases with fluence, then the number of damage sites is almost constant.

(Color online) Damage density vs true local fluence. (a) Two scans at nearly the same fluence , one being that already shown in Figs. 7 and 8. (b) Two scans in two different fluence ranges (8.3 and , respectively).

(Color online) Damage density vs true local fluence. (a) Two scans at nearly the same fluence , one being that already shown in Figs. 7 and 8. (b) Two scans in two different fluence ranges (8.3 and , respectively).

## Tables

Interval of confidence of for a given measured value of .

Interval of confidence of for a given measured value of .

Article metrics loading...

Full text loading...

Commenting has been disabled for this content