From the 24^th Febraury to the 7^th March, Richard Scrivens and Hartmut Kugler visited TRINITI, Troitsk, in order undertake the pre-commissioning of the CO₂ Master Oscillator (MO) Laser, which has been ordered for the CERN Laser Ion Source (CLIS).

The aim of the MO is to provide a smooth, single-mode (in both the longitudinal and transverse planes) pulse up to 200 mJ for later amplification. The system should work at a rep. rate of 3 Hz and be able to complete 10⁷ shots without intervention.

During the two weeks, some assembly of the MO was seen, with tests of the power supplies and final specifications testing.

As an appendix, an "Agreement of Status" is attached, giving the situation at present and the final steps to be taken. This was finalized before the last days statistics were compiled.

During the visit, the MO continued to be assembled. A schematic diagram of the unit seen for the final testing session is shown in Figure 1.

Figure 1. Scheme of the Master Oscillator installation at TRINITI.

The basis for the production of single frequency, single spatial-mode, smooth tail less 10.6 mm radiation is the following. The MO uses an atmospheric pressure transverse discharge chamber for the main active medium (TED module). However, the cavity and active medium scheme mean that the gain profile only just reaches the threshold for generation. The low pressure tube (LPT - typically with a CO₂/N₂/He mixture at a few mbar with a longitudinal sub-normal discharge), provides a narrow gain profile of typically 100 - 150  MHz, similar to the inter-mode spacing of the cavity (1.8 m long). These two tubes normally provide preferential lasing on one longitudinal mode.

The single transverse mode is provided using a diaphragm with a 6.7 mm aperture in front of the partial reflector (ZnSe plate, AR coated on one side).

The possibility of tailless pulse generation appears only at extremely high laser active medium pumping levels as is the case in the Master Oscillator.

By the end of the visit, both main power supplies were installed. The TED module uses a HV current supply to charge a capacitor bank, which is discharged through a double trigger scheme using a thyratron and spark-gap. The LPT uses a pulsed HV current supply which provides around 300 mA.

The MO resonator was mounted on a optical rail, fixed to an optical table. This did not provide ideal stability, being susceptible to both temperature drifts and vibration.

The MO did not have any automated alignment system, replying on hand alignment while viewing the output beam. A gas flow system (<20 litres/hour) from a reservoir was used in place of a regenerator system. The regeneration system was tested with palladium and Hopaclite mixtures, but sparking in the TED module was seen after a few hours of operation.

The final LPT is already at CERN, a second device was used for these experiments.

After the oscillator, the beam was directed through a spatial filter, consisting of two spherical focusing mirrors to provide a focal spot through a small 1 mm aperture. This guaranteed that the final output beam was Gaussian.

The beam was measured in the time domain using a photon-drag detector and a TDS360 200 MHz bandwidth digital scope. Approximately 20% of the traces were read into a computer (fastest data rate possible with the set-up available) where statistics were performed. The average and standard deviation of the amplitude and pulse width were recorded. In addition statistics were compiled on the number of smooth and two-mode shots (a two-mode shot was defined as one for which the trace deviated by more than 10% from an 11 point smoothed version of the pulse, relative the maximum of the smoothed pulse).

Two major sets of statistics were compiled on the final two days of the visit.

The first session on the 5^th March, was completed without the final LPT pulsed power supply (which caused an unstable timing of the discharge), instead a capacitor bank and thyratron were used. No adjustment of the cavity alignment or length were performed (until the last hour).

The second session (6^th March) included the final LPT supply, re-adjustment of the cavity length when many two-mode shots were seen and a generator in the lab which caused vibration on the MO support was switched off.

The final statistics for the 5/3 session are shown in Table 1.

The tests were completed at 2.5 Hz rep. rate for a total of 9 hours (~80,000 shots). A graph showing the number of smooth and noisy shots (see Figure 2) recorded through the first session is shown in Figure 3, where the cyclic nature of the two-mode regime is clearly seen. After a slight adjustment of the cavity length for the final hours running, the percentage of two-mode shots fell to 13%, in comparison with 53% averaged over the entire day.

Figure 2. An example of a smooth and two-mode shot.

Figure 3. Plot of the number of smooth and two-mode (noisy) shots record, through the first statistics session.

During the second session, three adjustments of the cavity length were made during approximately two hours testing. The statistic for this run can be seen in Figure 4. The percentage of two-mode shots recorded was 6% over this period, but during good runs of typically half an hour (4,500 shots) no two-mode shots were seen. It is therefore foreseen to add a motor drive to adjust the position of the outcoupling partial reflector, via a simple push button facility, which may be located remotely from the MO.

Figure 4. Plot of the number of smooth and two-mode (noisy) shots recorded through the second session (with cavity length adjustment).

The final statistics for the standard deviation of the ampliture variation and the pulse widths are shown in Table 2.

Table 2. Statistics of the first MO testing session (6/3).

	Average		Standard dev (%).
Amplitude	80.8		6.7%
Width	74.8 ns		3.8%
Smooth shots: 2677		Two-mode shots: 179

Discussion and Forward Planning

It is proposed that the MO now be made ready for delivery at CERN (foreseen for the 7^th April 1997). The initially instabilities have been controlled (pending the motorized mount installation) and the initial discharge instabilities between the LPT and its final power supply have been cured using a low level (~5 mA) DC discharge, with high current pulsing at the required time. The sparking during gas mixture regeneration may be studied before delivery, but in principle a gas flow scheme may work for the required number of shots.

At CERN, the preparation for delivery must begin, including the support system (with optical tables being the most likely method in view of the instabilities and vibrations seen at TRINITI). At this time the decision should be taken as to the best height for the ensemble (MO and Lumonics). At present the Lumonics beam output is at eye-level, which is undesirable from a safety point of view.

Finally the interlocking scheme for both beam and HV safety will be considered (the MO already has a complete interlock scheme) with its integration into the existing experimental system.

Conclusion

The Master Oscillator has been seen in full working order at TRINITI, Troitsk, with the exceptions of the motorized alignment scheme and the gas regeneration system. It completed 8.10⁴ shots for statistical measurements, except without the final LPT power supply. This final power supply was tested the following day for ~10⁴ shots.

The motorized alignment system is an interim solution before the full automatic system (with measurement and feed-back) is made.

The statistics show a standard deviation in the maximum amplitude of 6.7 - 7.2 %, and for the FWHM of 71.3 ns with a standard deviation of 3.8 - 4.2 %.

The MO will be delivered to CERN in April for commissioning and acceptance tests.