2 May 00
Maxime Boccas

The main thing I tried to determine is whether the regular C02 cleaning helps and, if so, by how much? SEE the conclusion at the end if you don't want to read all the text. For further details, look at the time plots in the corresponding sections for each telescope in the previous menu (back with your browser button)

I compared 2 periods:

• Before regular use of CO2 - 24 July 96 to 7 August 98 = 24 months with monthly (I believe) CO2 cleaning for the first 9 months approximately. I have the values after the 96 aluminizing and I find the following slopes:

--Reflectivity decrease: 0.25% per month at 400nm

                                    0.14% per month at 700nm

--Scatter increase: 0.35% per month at 400nm

                            0.33% per month at 700nm

• Regular use of CO2 - 7 August 98 to 13 April 00 = 21 months with fortnightly CO2 cleaning over the entire period. Looking at the plots of scattering, one sees what looks like 2 different slopes along the year, the steeper one repeating after about 12 months (although even more pronounced)! When one looks at the corresponding months, it is even more striking to see that they actually match the dry/wet seasons (winter/summer)!

--Oct 98 to March 99: increase of 0.25% per month at 400nm

                                   increase of 0.20% per month at 700nm

--April 99 to Oct 00: increase of 0.08% per month at 400nm

--Nov 99 to April 00: increase of 0.96% per month at 400nm

                                 increase of 0.70% per month at 700nm

Scattering is a direct measure of amount of dust deposited on the mirror, which in turns is closely connected to the humidity of the ambient air (wet air = wet dust = sticky dust!). The amount of dust carried by the air could itself show seasonal variations (again probably because of humidity), as well as other reasons. So the seasonal tendency has a reasonable explanation: because the same effect is also detected in the 1.5m telescope data (see below), and with even more reproducibility, this speculation must be considered seriously.

Overall scatter increase: 0.38% per month at 400nm

                                   0.29% per month at 700nm

Conclusion: This is not more or less the same as the previous period when little cleaning was made!! The CO2 cleaning efficiency is unclear (but at least no negative action)...

Reflectivity is trickier because no significant seasonal tendency is seen where I would have expected some (because reflectivity and scatter are somehow complementary; scatter+reflectivity+"absorption"=100%)! If there is a correlation with the seasonal dampness, it is lost in the noise of the measurements.

--Reflectivity decrease: 0.18% per month at 400nm

                                   0.11% per month at 700nm

Conclusion: this is slightly better than the previous period hence the CO2 cleaning seems useful...

--Reflectivity: 0.10% at 400nm

                    0.27% at 700nm

--Scatter: 0.06% at 400nm

               -0.05% at 700nm

Conclusion: CO2 cleaning seems to improve reflectivity but is not significant for scatter, which obviously confirms the previous conclusions.

I compared 2 periods:

• Before regular use of CO2 - 9 April 97 to 1 September 98 = 17 months without CO2 cleaning (or just a few at the beginning) assuming the mirror started at typical values reached at the last aluminizing of 98 (I don't have values for the aluminizing of 97), I find the following slopes:

--Reflectivity decrease: 0.18% per month at 400nm

                                 0.07% per month at 700nm

--Scatter increase: 0.24% per month at 400nm

                  0.24% per month at 700nm

• Regular use of CO2 - 1 September 98 to 21 March 00 = 18 months with monthly cleaning (1 missed in January 00). Looking at the scatter plots, one sees again what looks like 2 different slopes along the year, the first one repeating quite precisely after 12 months. They also match the dry/wet seasons (winter/summer):

--Sept 98 to March 99: increase of 0.36% per month at 400nm

                                     increase of 0.30% per month at 700nm

--April 99 to August 00: virtually no change (slope = 0) at 400 and 700nm

--Sept 99 to March 00: increase of 0.46% per month at 400nm

                                     increase of 0.35% per month at 700nm

Overall scatter increase: 0.27% per month at 400nm

                                     0.22% per month at 700nm

Conclusion: This is not much different from the previous period when no cleaning was made!!

The CO2 cleaning seems useless (but at least no negative action)...

Now looking at the reflectivity again, there is no significant seasonal tendency. If there is a damped seasonal tendency, it is also lost in the noise of the measurements.

--Reflectivity decrease: 0.09% per month at 400nm

                                 0.03% per month at 700nm

Conclusion: this is a factor 2 better than the previous period hence the CO2 cleaning seems quite useful...

--Reflectivity: 0.33% at 400nm

                     0.15% at 700nm

--Scatter: -0.19% at 400nm

               -0.33% at 700nm

Conclusion: CO2 cleaning seems to improve reflectivity and scatter, which obviously sort of the previous conclusions.

I compared 2 periods:

• June 97 to June 99:

--Reflectivity decrease: 0.33% per month at 400nm

                                    0.18% per month at 700nm

--Scatter increase: 0.32% per month at 400nm

                              0.44% per month at 700nm

• June 99 to April 00:

--Reflectivity decrease: 0.30% per month at 400nm

                                    0.17% per month at 700nm

--Scatter increase: 0.36% per month at 400nm

                              0.52% per month at 700nm

The numbers show some coherence but the weird effect is that scattering at longer wavelength (700nm) does increase faster than at shorter wavelength (400nm), effect seen only in that telescope. This is not expected.

We also notice that the rates are higher than at the 4m and the 1.5m. This might be explained by : 1.) the fact that we don’t use CO2 cleaning (assuming that CO2 cleaning is as efficient as in the 1.5m)

and/or 2.) the fans pulling air down the tube may increase the incidence of dust onto the mirror.

Below is a table summarizing the slope in percent per month and the gains:

From the slopes, the only case where a benefit seems visible is the reflectivity at the 1.5m. From the gains, only the scatter at the 4m is not convincing. This latter form is our favorite (because slopes are delicate to determine).

One can compare the 1.5m and 4m telescope because they have similarly opened domes and exposed mirrors. (Note that the absolute numbers of scatter and reflectivity after aluminizing are different because the aluminizing chambers are different and the processes don't have the same efficiency for reasons we are investigating). Comparing with the 0.9m is more difficult because the dome is less open (no lateral louvers) and the mirror is hidden at the bottom of a long tube that is equipped with fans.

The scattered light measured is the Total Integrated Scatter (TIS) equivalent to (4.PI.d/l)^2 where d is the rms height of surface irregularities and l is the wavelength. Thus scatter scales down in the IR.

Since we implemented the regular CO2 cleaning almost 2 years ago at both telescopes (4m and 1.5m), we have used 16 cylinders (as far as I have been able to account for) of carbonic snow and spent about 16 x 41.000 pesos = $656,000 pesos = US$1,260 to refill them.