Sciencemadness Discussion Board

homemade mass spectrometer

neptunium - 28-12-2014 at 13:56

i know a glass blower who is currently building a mass spectrometer for me. i had various vacuum glass tubes with electrodes built from him before and i had a lot of fun with it. my ex roomate`s girlfriend thought it was appropriate to throw my stuff at him while in a rage so i lost the last one .

However, i am wondering about the detection system and after some research on how to detect heavy ion i am at an impass..
scintillators must be very thin and degrade rapidly, single farady cups requires scanning and arrays are impossible to find (cheap enough) same with microchannel detector...
I will be posting pictures of the thing when it is done (maybe early January)
there is an positive ion repulsor, an ion gun, a beam focusor and negative pole have 2 vacuum pumps with gauge that measures vacuum from atm to about 1 torr and from 10-1 to 10-10 torr. i am still looking for a turbo pump.
the mass spectrometer is about 3 feet in lenght so 10-4 Torr for a mean free path of about 1 meter is required wich my pumps cannot reach.

anyways, i am looking for a way to detect, qualify and quantify the ion hitting the back of the tube. Software instead of the oscilloscope would be the best!

any ideas?

diddi - 28-12-2014 at 15:11

sounds like a great project. you are lucky to have access to that kind of technical expertise.

neptunium - 28-12-2014 at 17:44

thanks , its not much but time and a few bucks.... anybody with enough dedication should be able to do what i do ...
also i try those alpha spectrometer. . . most of them have NO detector (on ebay) those that do are way too expansive and i talk to the tech support guy at ortec who said they ar light sensitive and cost from 700 to 4000 dollars (US)...
so ... next!

phlogiston - 29-12-2014 at 03:27

Very cool. I have been hoping someone would try building a masspec of some kind someday (lacking time and equipment to do so myself), looking forward to your progress.
Incidentally, Ben Krasnov also mentioned in one of his recentish videos he was going to work on one.

Regarding detector suggestions, a non-electronic, simple (but inconvenient) method is to use photographic paper.

How would a CCD behave in this environment? As I am sure you are well aware, charged particles of sufficient energy show up on CCD cameras. Linear CCD chips can be harvested from discarded scanners, and they are often relatively easy to interface with.

Using the search term "ionccd" you can find papers describing a modified CCD where the photosensitive array has been replaced with an array of capacitors which is even commercially used but it is probably pretty costly if at all available in single chips.

What energy do your ions have when they reach the detector?
Also, how are you deflecting them, is scanning an option you have permanently dismissed?

edit: I removed a non-electonic detector suggestion involving a dark basement, your ex-roommates girlfriend and atropine but on second thought perhaps it is a poor joke. I can hardly imagine how angry you must have been though.

[Edited on 29-12-2014 by phlogiston]

neptunium - 29-12-2014 at 06:44

a CCD seems like a great idea! ... no i did not fully rejected the scanning method yet.
the design will be such that several method of detection could be used thanks to interchangeable glass element.
i plan to use it with voltage from 1000 to 40Kv . but probably closer to 10Kv and play arround to see what is the best voltage.
the deflection will be either a permanent magnet (N55) or a strong ectro magnet on a variac.
I want to start with LiI because Iodine and Lithium are far appart in mass with 2 important natural isotopes. (3 actually but Li6 less than 8%.)
so if i put a CCD camera i should be able to see the impact of each group of ion ... im going to work on this idea... seems brilliant thanks Phlogiston!

phlogiston - 29-12-2014 at 07:56

Interesting, that brings up another thing I was wandering about.
I was assuming you would initially try to resolve isotopes of residual gas in the tube, but detecting isotopes of LiL requires that you ionize the sample...

neptunium - 29-12-2014 at 12:04

yes i have a modest ion gun (homemade) i still need to work out the tray where the sample sits

diddi - 28-1-2015 at 04:07

any update on this project? I am waiting for you to make me one as well :)

neptunium - 12-2-2015 at 18:20

i just received a turbomolecular vacuum pump and i am suppose to meet with the glass blower this weekend! (if not too much snow) i also need a few more clamp and connectors .... money is always an issue as always for us home scientist! so stay tuned i`ll have pictures soon!

[Edited on 13-2-2015 by neptunium]

diddi - 13-2-2015 at 16:16

mass spec is performed under complete vacuum? that must be a fancy pump...

neptunium - 13-2-2015 at 16:40

the mean free path is really what you are looking at 10E-4 torr to 10E-5 torr gives a MFP of about 1 meter that is plenty for the size of the tube . but to deeper the better! this pump is therotically capable of getting down to 10E-9 torr ! fancy indeed! but i had a good price for it !:cool:

WeaponsRx - 14-2-2015 at 03:08

Can you perform mass spec yet?!

Fulmen - 14-2-2015 at 04:52

Neptunium: Sir, you are my kind of crazy.

neptunium - 14-2-2015 at 05:34

Quote: Originally posted by WeaponsRx  
Can you perform mass spec yet?!

no not yet i will post pictures and i have to tweek with it see what i can get....i mam not even pretanding it will work but the fun is in the journey right?

neptunium - 14-2-2015 at 05:36

Quote: Originally posted by Fulmen  
Neptunium: Sir, you are my kind of crazy.

i am suprised nobody here tried it before actually

morganbw - 15-2-2015 at 05:11

Quote: Originally posted by neptunium  
Quote: Originally posted by WeaponsRx  
Can you perform mass spec yet?!

no not yet i will post pictures and i have to tweek with it see what i can get....i mam not even pretanding it will work but the fun is in the journey right?

The journey is all we have:)

smaerd - 15-2-2015 at 06:24

Now this is an exciting thread. I'd love to see some pictures and stuff if you'd be willing to share.

neptunium - 24-2-2015 at 18:11


here is my friend making a bromine ampoule for my periodic table ... he is the same one working on the mass spectrometer as i am writting these lines... you guys should see his UNBELIEVABLE shop! lab glassware everywhere ! and he can make you anything outta glass !

glass blower.jpg - 564kB

i will post more picture when he is done with it! i cant wait !
also my turbo vacuum pump has arrived and its a deep vacuum blast!!!

vacuum system.jpg - 622kB

i dont know if the picture is clear enough but the vacuum gauge reads 2.2E-7 Torr !!!! after about 20 minutes!

[Edited on 25-2-2015 by neptunium]

diddi - 25-2-2015 at 16:17

great to see some news about this project. that is a really good vacuum result. have you trialled your ionising mechanism at all?

neptunium - 27-2-2015 at 12:18

no that will be the suprise when its completed along with the rest of the tube..
Also the 10-7 i thought i reached was a transicent pressure and i had a short in the gauge ! :( !
now i am confident i reached 2.0x10-5 Torr which gives a mean free path somewhere arround 10 meters !
that is way overkill for a tube only about 1 meter long! those ion should be flying !!

0227151437.jpg - 575kB

mayko - 27-2-2015 at 13:31

This looks great!

One of my too-mad-to-work schemes has been to convert a CRT television into a mass spectrometer. I don't know if it's designed to withstand the vacuum involved, or if it could withstand being retrofitted (no matter how gently I depressurize, the phosphor screen seems to take damage.) But perhaps the electromagnets which direct the electron gun could be repurposed for a tunable deflector, and some sort of phosphor used as an inexpensive detector?

diddi - 27-2-2015 at 13:44

interesting ideas mayko. I don't know enough, but I do know that the deflectors are tunable. I cant see how you can mod the gun without causing irreparable damage to the crt?

id like to see a 10m long mass spec. ultra hi res!

neptunium - 28-2-2015 at 17:04

the cathode ray tubes are(were) built to accelerate electron not heavy ion so not only you`d have to find a way to introduce them in there but also reverse the accelerating voltage to negative.

a 10 meter long spectrometer huh? i guess if you were stuck with the same magnet it would get better resolution... but why make it longer if you could increase the seperating field ?
I am working on an idea that i need to test for the detector its very simple and may not require a magnetic field , the resolution might suffer a bit or it may not work at all but hey thats why we call it research right? :)

IrC - 7-3-2015 at 15:06

Method and apparatus for increasing the ionization of impurity ions in a mass spectrometer

A patent from 1966 but I wonder if it has useful information to apply to this project.

Attachment: US3254209A.pdf (444kB)
This file has been downloaded 345 times

Analysis by imparting unequal energies to ions

Attachment: US2764691A.pdf (308kB)
This file has been downloaded 339 times

Attachment: US2587555A.pdf (333kB)
This file has been downloaded 340 times

Attachment: US2642535A.pdf (578kB)
This file has been downloaded 335 times

One could redesign the older tube circuitry with modern semiconductors.

neptunium - 1-4-2015 at 09:07

interesting how they are using a photomultiplier for detector , i thought about it but much of a PMT has to be removed and the glass tube may not weld on my type of glass (glue would compromise the high vacuum )
but it definitely deserve some investigating further ... thanks IrC you always dig up suprising finds!

phlogiston - 2-4-2015 at 02:44

What if you were to include only a scintillating material in the vacuum, and a PMT outside (viewing the scintillator through a window)?

neptunium - 15-5-2015 at 12:00

finally ! i had the money to go get my mass spectrometer !
here is a few shots and a first vacuum test...
0515151532.jpg - 576kB
this is the coil and ion gun area followed by the ion beam focusing rings
0515151533.jpg - 636kB
at the other end the detector will take its place between the two electrodes (one for the signal and one for the ground)
and a general view of the pumping system by my washing machine!
I am working on upgrading my lab to have a better work area the laundry room is not equiped for fragil glass expansive vacuum pumps and high voltage wires...
so i will have a bench in a few weeks i hope in a different part of the house! this is NOT a cheap hobby guys!

0515151533a.jpg - 627kB

0515151533b.jpg - 587kB

smaerd - 19-5-2015 at 16:42

It's beautiful :)

neptunium - 23-2-2017 at 10:29

latest development

major vacuum problems, i have to re engineer the glass/electrodes connections
i can get it to about 5e-5 Torr with this pump without any problems, which would give me a mean free path above one meter (longer than the spectrometer)
still working on it ( when my job doesnt get in the way...)

[Edited on 23-2-2017 by neptunium]

wg48 - 23-2-2017 at 12:49

Hi Neptuneium. Wow a serious project. I noticed in your youtube video you had the focusing electrodes floating. I would expect that to stop any electrons getting down the tube as they will build up a negative charge..

Is the long valve in the video part of the mass spectrometer? How is the sorting of the different masses done? I don’t see any big magnets or quadruples. Is it a time of flight mass spectrometer?

Sulaiman - 23-2-2017 at 13:29

if your vacuum is only sufficient for a mean free path just a little longer than the physical path length
I would expect your spectrum to have very poor resolution;
i.e. if the mean of ions just about make it to the detector without collision, there must be a statistical spread of collisions.
I do not know, but you should check,
by how many orders of magnitude should the mean free path be longer than the physical path to achieve your desired resolution ?

I do not see why you can not just use the detected beam current directly into an electronic amplifier.
In the '70's I was working in the fA range, technology must have moved on since then, if you like I could have a look for a suitable amplifier ?

Regarding magnetic deflection with permanent magnets;
N55 sounds sexy,
but I think that you would do better with a greater volume of a lower grade magnetic material,
use a calculator like this
to try different configurations.

also, I suspect that for good resolution you may need to shield your spectrometer tube from electric and magnetic fields, especially 50/60 Hz magnetic fields.

If you do go for spectral scanning by varying the field of an electromagnet, I2.R heating is always a problem
if you have permanent magnets creating a magnetic field of half your peak desired field
then varying an electromagnet in series with the magnets only requires +/- 0.5 times the current .... 0.25 times the heating.

neptunium - 23-2-2017 at 14:48

Wg48, no the focusing copper ring are not floating they are pressed between the glass and the electrode with a small spring.
I don't know what valve you a reefering to but I was planning on using the electromagnet as a deflection way . based on preliminary test I might need a better magnet even if my tube seems long enough...

neptunium - 23-2-2017 at 14:52

Thanks for your help . this was just a preliminary test . I did plan on using an electromagnet for scanning the mass spectrum but upon review I might use one of my N55 ... Or a much much stronger electromagnet... This is really a mass spectroscope . if I can just get a visible trace of deflected ion on the phospor I'll be a happy camper....but any Improvement is indeed very welcome !!

wg48 - 23-2-2017 at 16:49


The glass tube in your youtube video I was referring to as a valve.

The two rings (I think you called them focusing rings) to the right of the cathode area on the left end of the valve glass tube thing. The two rings are not connected to anything, they are electrically floating. You should at least connect them to the filament or they will charge up and electrons or for that matter ions will be stopped from going down the tube.

So if this tube is the mass spectrometer/spectroscope tube then the deflection can be no more than the diameter of the long tube. I have got that correct?

Do you have the one side of the filament supply connected to the negative of the HV supply? It should be connected.

No glow discharge is visible in the tube. When you apply the 1000v you should get a glow discharge if your vacuum is not low enough but still less than about 20torr.

If you run the tube with filament off and apply the HV between two electrodes then start evacuating the tube you should get a glow discharge as the pressure drops eventually as the vacuum get very good the glow will stop. I suggest you try that but limit the HV current to say a few mA to avoid damage to your electrodes. See image below.

What was the pressure in the video.

You should be able to creat a glow like this as the vaccum increases

Sorry the image would not display for some reason

[Edited on 24-2-2017 by wg48]

[Edited on 24-2-2017 by wg48]

neptunium - 23-2-2017 at 18:57

The pressure was around 2e-3 Torr as shown .
I see what you mean with the focusing ring.. This was just a proof of concept and you are correct I did not connect them to any thing yet. I have to fix my vacuum leakage problem first. I do get a glow discharge from 10 Torr to about 1e-3 or so then the filament supply the electron by thermo electro generation like in a X ray tube. It does not have to be supply with a negative HV

wg48 - 24-2-2017 at 02:22

Quote: Originally posted by neptunium  
The pressure was around 2e-3 Torr as shown .
I see what you mean with the focusing ring.. This was just a proof of concept and you are correct I did not connect them to any thing yet. I have to fix my vacuum leakage problem first. I do get a glow discharge from 10 Torr to about 1e-3 or so then the filament supply the electron by thermo electro generation like in a X ray tube. It does not have to be supply with a negative HV

Your HV supply should have two output connections. One commonly designated positive and connected to the anode at the phosphor end of the tube. The other connection of your HV supply commonly called negative which I assume is connected to the electrode at the far end of the tube near the filament. That must be connected to one side of the filament so that there is a current path from the filament back to the HV supply.

neptunium - 24-2-2017 at 14:48

still work in progress but i got a trace of electrons on the phosphor! and the vacuum leaks are fixed!
i need a much better filament ( i messed it up while fixing the vacuum leak on the electrodes)but i do commonly get down to 1e-4 Torr and below easily now!
unless i manufacture an ion gun and perfectly line it up to the detector (which has to be made as well)
this phosphor is just an alpha scintillator (ZnS )

[Edited on 24-2-2017 by neptunium]