Methanol spectral line at 12.2 GHz

[Ed astro]

Hi all, 

In recent years the hydrogen line has become a popular target for amateur (radio)astronomers. However, there are many more spectral lines in the radio spectrum, and I am very interested in detecting some of these lesser-known lines. Here I will describe my attempts at detecting the methanol (CH3OH) line at 12178.593 MHz. 

My homemade 3 metre dish was unfortunately not very useful for this project, because its surface is not accurate enough for such short wavelengths (the wavelength of the methanol line is only 2.5cm, much shorter than the 21 cm hydrogen line which the dish was intended for.) Therefore I bought a 1.1 metre solid offset satellite dish. This smaller dish is also light enough for my HEQ5 mount. 

The frequency of the spectral line is way above the maximum frequency of most SDR receivers. However, it nicely falls within the frequency range of those cheap Ku band (PLL-based) LNBs intended for satellite TV reception between 10.7 and 12.7 GHz. I use an inverto single Ku band PLL LNB to convert the methanol line frequency down to L-band. The local oscillator (LO) frequency of the LNB is 10.6 GHz, so the methanol line is converted down to 12178-10600=1578 MHz, which is well within the frequency range of my airspy mini SDR. One caveat with these LNBs is that they often have two different LO's; a low-band LO at 9.75 GHz and a high band LO at 10.6 GHz. The high band LO can be turned on with a 22 KHz tone. My father built a special power supply for me, which puts out the 12V DC needed to power the LNB with the 22 KHz sinewave for band switching superimposed on it. (I am very grateful to him for building this because I am not very good at electronics myself!) The 12v DC with the 22KHz tone is fed into the LNB via a bias tee. 

At the moment the little radio telescope is outside in the cold, collecting spectra of a star forming region. I use SDR# with the IF average plugin to average the spectrum, this is the same software which I also use for hydrogen line observations. Every 3 minutes a .txt file with the averaged spectrum is saved for later processing.

When I am done with processing and plotting the spectra I will report on the results.

Best regards,

Eduard.

 

 

[andrew s]

First rate. Looking forward to the results.  Regards Andrew 

[Ed astro]

Hi all

After a week of doing observations and another afternoon processing the data I finally have some nice results.

The target I observed is the star forming region W3(OH). It is also a strong methanol maser: the 12.178 GHz methanol line, as well as other radio spectral lines, is amplified by stimulated emission. This is why we can detect the methanol  line of W3(OH) with such a small aperture. It still is an incredibly weak signal though, compared to terrestrial signals from cellphones and other electronics... Luckily, the 12 GHz band is still fairly "radio-quiet", the only major interference I have seen was from geostationary satellites, and these were far enough away from the position of W3(OH) and were not much of a problem for my observations. This is all going to change when starlink starts operating at this same band in the near future, I doubt whether it will be possible to repeat these observations with a lot more RFI.

I did 6 different observations of W3(OH), during each observation session the dish was pointed at W3(OH) for about about 2-3 hours. I also spent an equal amount of time collecting spectra while the dish was pointing away from the source, these spectra are used to subtract the artifacts of the SDR receiver from the W3(OH) spectra. this "dark subtraction" is also done with hydrogen line observations. In fact: processing methanol line spectra is not much different from hydrogen line, it is just that the signal is about 100- 1000 times weaker! One thing that had to be accounted for was the frequency offset of the LNB. After each observing session I measured the frequency of the Astra 3B satellite beacon at 11446.8 MHz to determine the frequency offset. I also corrected the radial velocity for the doppler shift caused by Earths movement around the Sun using the ATNF online VLSR calculator. All six results showed a bump in the spectrum at around -46 km/s.

Finally, I averaged all spectra from the 6 different observing sessions together to get a spectrum with much better signal to noise ratio. The total integration time of this spectrum is over 15 hours! The spectral line does not appear to be a single peak: about four different features can be distinguished in the averaged spectrum. These features originate from different maser spots inside the star forming region, these are clumps of gas where conditions are favourable for stimulated emission to occur. Because these spots move around the centre of mass in the star forming region, we see their methanol lines red- and blueshifted to different velocities in our spectrum.

You can compare my results to the spectra published on Michiel Klaassens website (under project MK23): http://parac.eu/projectmk23.htm . Michiel has done several observations of W3(OH) and other masers with his 9.3 metre dish in Portugal.

Best regards, 

Eduard

 

 

Edited March 1, 2021 by Ed astro

[Ed astro]

Hi all,

Here an update on this project. I have fixed some problems and mistakes in the frequency correction. In my previous post I mentioned that I measured the frequency of the Astra 3B satellite beacon after each observing session, but I did not really explain why this was important. The local oscillator (LO) frequency should be 10600 MHz, but in reality it can deviate tens of KHz from that frequency. It is possible to modify the LNB for better frequency stability, but I did not want to open up the LNB and mess around with the electronics. Instead, I decided to use a satellite beacon. The idea is that if we know the exact frequency of the beacon, then we can calculate the frequency deviation of the LO by measuring the (apparent) frequency of the beacon and subtracting the true frequency.

However, things were not as simple as I had imagined... While my results of W3(OH) were fairly consistent, the radial velocity of the methanol line deviated about 2km/s from the spectra made by Michiel Klaassen. It soon became clear that there was some uncertainty about the frequency of the Astra 3B beacon. The source I used (UHF- satcom's list of Ku-band beacons) listed 11446.8 MHz, while another source I later found listed 11446.75 MHz. I decided to try measuring the exact frequency of the Astra beacon by using another beacon with an exactly known frequency for the LO frequency correction. I used the two beacons which mark the downlink band of the amateur radio transponder on board of the Eshail2 satellite, at 10489.5 and 10489.795 MHz. I found that the Astra 3B beacon has a frequency of 11446.7266 MHz. I also found a small error in the spreadsheet I used for the frequency correction. Now after I fixed these issues, the velocity deviation is much smaller and close to the spectral resolution, only a few hundred m/s at most. 

Finally, I did an extra observation of W3(OH) on the evening of March 19, and the signal was still there at the right velocity. I am now very confident that I did indeed detect the methanol line.

Best regards,

Eduard

 

[athornett]

This is spectacular! Thanks for sharing. Amazing stuff. I am so excited to read about your observations. One question though you said the HEQ5 could handle it bit surely the dish is a bit heavy for this mount?

[Ed astro]

Hi,

Surprisingly, the dish is not that much heavier than the 6" scope which I normally have set up on this mount. It does catch a lot of wind, so I can only use it when the weather is calm.

 

[Ed astro]

Hi all,

Last week I tried to detect another 12.2 GHz methanol maser source known as G188.94+0.89. This object much weaker than W3(OH), so detecting it with a small dish is even more challenging. I observed this source on the evenings of March 23, 25, 27 and 29. The graph below shows the result after averaging all spectra from March 23, 25 and 29 (the spectra of March 27 were not used because these were more noisy and the signal of the maser was buried in the noise). There is a small peak at 10.9 km/s, the width of the peak is about 1.6 km/s. 

This result can be compared to the spectra taken by Michiel Klaassen in 2018 and 2019 with the 9.3 metre Sao Giao radio telescope (see http://parac.eu/projectmk22.htm). Both of Michiels spectra show that the velocity of the peak is about 10.5- 10.9 km/s, but the width is a bit different (about 1.3 km/s in 2018 and 2.5 km/s in 2019). Furthermore, the flux density (brightness) of this source also varies between 2018 and 2019. It seems that the G188.94+0.89 methanol maser is quite variable.

Best regards,

Eduard. 

[Corncrake]

Gosh!

Extraordinary, fascinating, well done and thanks for showing.

[Ouroboros]

Yes, very interesting. Maybe you explain this somewhere above but why is the horizontal axis velocity (km/s)?  I might have  expected it to be frequency (Hz).  

[Ed astro]

8 hours ago, Ouroboros said:

Yes, very interesting. Maybe you explain this somewhere above but why is the horizontal axis velocity (km/s)?  I might have  expected it to be frequency (Hz).  

Hi,

I did indeed not discuss why I used velocity on the horizontal axis instead of frequency. Professional astronomers often use the LSR velocity for spectra of objects within our own galaxy, because they are often studying the motion of objects in our galaxy or in a star forming region. LSR (Local Strandard of Rest) velocity is the velocity with respect to our local area of the Milky Way.

I could of course have just used the frequency, but the frequency gradually changes due to the doppler shift caused by the Earths rotation and movement around the Sun. In that case it would not be that easy to add spectra from different days together. The easiest way to deal with this issue is to calculate radial velocity from the frequency values using the doppler formula, and then correcting for the velocity components of Earths orbit and rotation using an online LSR velocity calculator. I used Steve Olney's calculator (https://sites.google.com/view/hawkrao/hawkrao-calculators/vlsr-calculator) but there are others as well. Using LSR velocities also makes it easier to compare my own observations to those of others or to spectra published in the scientific literature.

 

[Ouroboros]

I see, I think. Instead of plotting the methanol spectral line as intensity versus frequency you’re converting frequency to velocity relative to the sun’s. Is that right? 

[Ed astro]

1 minute ago, Ouroboros said:

I see, I think. Instead of plotting the methanol spectral line as intensity versus frequency you’re converting frequency to velocity relative to the sun’s. Is that right? 

Yes, that is indeed basically what I did. I converted the frecuency to the velocity relative to the average of the Sun and nearby stars (local standard of rest) but that is a detail.

[Ouroboros]

OK. Thanks.  I’ve learned something here. 

Can you obtain any spatial information on the source?  Or is the object effectively a point source. 

[Ed astro]

The resolution of the 1 metre dish at 12GHz is about 2 degrees but the source is much smaller than that. So it is indeed effectively a point source.

[Ed astro]

Hi all,

It has been a while since I posted in this topic. I was hoping that I could detect more methanol maser sources, but unfortunately most of the sources are much weaker than W3(OH) and G188. There was an unsuccessful attempt to detect the variable source G9.62+0.19 last November, it was rather low above the horizon and my dish was picking up too much thermal noise from the ground and from nearby trees. 

I did make a new observation of W3(OH) in September, this time by pointing the dish at the declination of W3(OH) and letting the source drift through the beam as the Earth rotates. Because the beam of my 1 metre dish is only two degrees wide at 12 GHz, the duration of the transit was less than 15 minutes. This is not much time to observe such a weak signal, so a good signal-to-noise ratio was not expected. However, the methanol maser line was still detected. This demonstrates that W3(OH) could be observed without a tracking mount. By repeating the transit scan multiple times and averaging the results it should be possible to build up more integration time and achieve a higher signal-to-noise ratio.