I promise, nothing was broken! And nobody was injured. We had simply quite a little fright…

Observing binary stars

We are in 2001, and I am coming to the beautiful South Africa for the second time to accumulate observations in the SAAO for my Ph.D. thesis.

The SAAO telescopes plateau close to Sutherland, where one can see the big dome of the 1.9m Radcliffe telescope I used. Credits: SAAO

It was not planned, but having chosen stars of the Magellanic Clouds for my Ph.D. was a fantastic idea. These “clouds” are two irregular galaxies, only visible from the southern hemisphere, and high in the sky only during the nights of late September to early January.

The subject of my Ph.D. was to find double stars among a list of very specific massive stars in the Magellanic Clouds. These so-called “binary” stars are small systems comprising two stars of different mass. This is quite common, in the Universe. And when their orbital period is short (say, a few days up to a few dozens of days), they influence each other a lot. And their evolution story can be dramatically different – and interesting! – compared to solitary cases.

So how do you detect binary stars? The most direct way is to use the Doppler effect, of course. The same effect that is used to detect the first, and event today, most of the exoplanets. Basically, the light of the source is dominated by one of the component (the most massive star of the two, in my case, and the only star, in the case of a system star + planet). This is the one you see in the data.

The radial velocity method to detect exoplanet is based on the detection of variations in the velocity of the central star, due to the changing direction of the gravitational pull from an (unseen) exoplanet as it orbits the star. When the star moves towards us, its spectrum is blueshifted, while it is redshifted when it moves away from us. By regularly looking at the spectrum of a star – and so, measure its velocity – one can see if it moves periodically due to the influence of a companion. Credits: Wikipedia

And measuring the Doppler effect is measuring the small shift of the spectrum over time. To detect these shifts, you must obtain a spectrum of the source. Then, wait a bit (say one or two nights, in my case). Then, obtain another spectrum. And again, wait a bit etc. And when you detect (large) shifts, you may have found one of those binary star. Hence, you continue accumulating spectra, until you see the beginning of a regular pattern.

Various cyclic patterns detected during my Ph.D., where data points were folded in time (after a little period-detection process). All these stars (their name stars with “BAT” because they belong to the BAT catalogue).

The problem of calibration

Of course, to detect real shifts from the stars, your spectra must be very well calibrated. That is, under every circumstances, the expected position of the spectrum of the star on the detector must be well measured. And to do that, we make spectra of a special lamp whose lines are verry narrow and their position extremely well known. This is a standard technique in spectroscopy.

Hence, when you make a spectrum of this lamp, you can say for every pixel of the detector to what wavelength it corresponds, and thus you have a scale. Say you measure the center of the line Helium II at the position of the pixel 345, you can build a scale across the detector that can now be used a reference, i.e. by comparison with the spectrum of the star.

So how do you do a spectrum of such a lamp with a telescope? Well, the lamp is inside the spectrograph instrument! So basically, you can decide, from within the telescope control room, if you take a spectrum of a star, or that of the lamp. Easy.

But there is another reason why the lamp is inside the instrument. It’s because you must calibrate exactly at the current position of the telescope. Especially if you are looking for precise radial velocities. For a simple reason: depending on how the telescope is placed, the distortions inside the instruments due to the simple gravity impacts the optical paths, hence the calibration, hence the scale on the detector, hence the radial velocities.

These distortions are quite small, since they are provoked only by the instrument weight. But these instruments are also very sensitive. Hence, even a tiny distortion has a measurable effect on the detector.

Thus, to make sure we get good velocities, our technique at that time was to bracket the spectrum of a star, with two spectra of the lamps: one before, and one after. This has a small drawback of course. All the precious night time we spent on calibration is not used for science observations. But that’s the small price you have to pay to get robust radial velocity measurements.

Hence, for every target of the night, after having pointed the telescope to it, we started the bracket with a calibration spectrum. And likewise, after the star’s spectrum, we were taking again another calibration.

And this is where it gets tricky…

The Radcliffe 1.9m (74′) telescope in SAAO

Here is a recent picture of the Radcliffe 1.9m telescope I was using.

Despite the large counterweight in front, and the unusual two-pier assymetrical mounting, one can see the telescope standing vertically, with an instrument attached at the bottom (in the Cassegrain focus – Read all details in this page).

This is a great telescope, and the data I obtained at SAAO was absolutely essential for my ability to have any meaningful result for my Ph.D.

At that time, the spectrograph attached at the bottom of the tube was controlled through cables attached to an electronic rack, standing against the wall of the dome (roughly where a small white-metal gear is shown in the bottom-right of the picture).

Oups!

So here is the trick. We are early late January / early February. It’s already late in the season for observing the Magellanic Clouds (due to their Right Ascension coordinates). It means that they go by the meridian quite early in the night (or even before for the SMC, see below). And at the end of the night, all my targets are quite low in altitude.

As you can see in this Arcsecond’s Night Explorer Screenshot: nights are short at that time of the year in the SAAO. And both the Small and Large Magellanic Clouds (SMC and LMC respectively) are quite low in altitude (<25º) at the end of the night!

But the telescope follows. And I am a Ph.D. students whose unique goal was to accumulate as many spectra as possible.

Hence, for the last stars of the night the telescope was following the low stars of the LMC. And we were blindly following always the same sequence: spectrum of a lamp, exposure on a star, then spectrum of a lamp again, keeping the telescope in its position (to ensure we calibrate with the right / distorted optical paths).

But suddenly, we noticed the telescope was not tracking the expected trajectory anymore…

As if something was preventing the telescope to follow the normal course, since the coordinates on the screen were not slowly slewing anymore….

Suddenly, Fred M. the night assistant jumped out of his chair, and rushed into the dome! The cables!

The cables relying the instrument to the rack were extending at their maximum, preventing the telescope to move any centimeter forward!

Oups…

Thanks to Fred’s quick reaction, we avoided to damage anything.

But we had a fright!

And we finished the last observations without the “after-science” calibration spectrum.

Conclusion and … security alarms in the Blanco?

Apart from a very few examples (such as the ESO’s Very Large Telelescopes in Chile), all these telescopes are very unique and custom systems. They don’t necessarily have attached all the appliances and securities and software to check everything. That’s the responsability of the astronomer and his/her night assistant (when one is present).

Even at the end of the night, you’d better “know your night and your telescope”!

A lesson I learned again a year later, in the Blanco telescope in the CTIO, in Chile. The Blanco telescope is a 4-meter giant…

To be continued!

No responses yet

Leave a Reply

Your email address will not be published. Required fields are marked *