Sciencemadness Discussion Board - analytical machinery

analytical machinery

chemrox - 18-2-2007 at 21:56

As a student, like most students of the day, I had GC, IR and NMR available at almost all times. I have an old two column GC that can be resurrected. I want to read up on detectors and would appreciate help with the refs. As a next device, what do you think? GC with the right detector will tell me how pure. IR will tell me about functiuonal groups knowing mass would be really nice but for now I have to get there through colligative properties .... sometimes used machines are fairly cheap b ut not so cheap I can go and buy whatever I like at the same time. If you could get and IR and an NMR for about the same price and you could afford one, which? Why?

Ozone - 18-2-2007 at 22:07

The initial cost is not the question. The question is what does upkeep cost? In terms of overall use The best deal you will get is a good old scanning UV-VIS spectrophotometer (crappy characterization, but used for so many other things that you be glad to have it). Following this, a simple HPLC with a UV-VIS detector (which can be hacked off of auction sites for @$500US (GC is nice, but you need a constant supply of high quality gas a,dn expensive scrubbers). NMR is the best thing since cheese--but, it's expensive both to purchase and keep running (liquid N2 or you magnet quenches--and is probably very expensively dead). So, for the home lab I would recommend the IR (the only thing that ever usually dies is the globar), but, buy a decent FTIR instrument.

Cheers,

O3

pantone159 - 18-2-2007 at 22:07

Quote:

Originally posted by chemrox
If you could get and IR and an NMR for about the same price and you could afford one, which? Why?

Assuming I already had a GC (which would be my first choice), I'd definitely go for the NMR. It gives you a lot more info to interpret to figure out your structure. IR spectra, if you don't have a library to compare to anyway, don't always give much useful information. Some, but not as much as NMR.

Personally, I think having an NMR in my Mad Scientist Lair would just be too cool. :cool:

EDIT: I just read Ozone's post, and I'm pretty sure he has a lot more practical advice than me. Do you really have to keep the magnets cold 100% of the time? That would really be a problem indeed.

[Edited on 19-2-2007 by pantone159]

Ozone - 18-2-2007 at 22:36

Well, services for dealing with magnet quenches are available for ~$18,000US a pop:

http://www.mrr.com/service/quench.shtml

They do look cool, don't they? Oh, and high field magnets are very large and heavy. The higher the MHz, the bigger the magnet. Our new one was delivered and moved in with a crane.

Cheers,

O3

[edit] the final cooling is with He (l), N2 (l) is used to pre-cool the magnets prior to bringing them up to field. The older, lower field instruments did use N2 (l), I think for primary cooling. He is much more expensive and may not be available in some countries without enormous shipping costs.

[Edited on 19-2-2007 by Ozone]

Polverone - 18-2-2007 at 22:59

Are there actually any NMR machines that get away with using high temperature superconductors and liquid nitrogen? I thought they all used niobium alloys at liquid helium temperatures -- and helium's a lot more expensive, and has much lower cooling capacity to boot. I seem to recall seeing plans for a (low frequency) DIY NMR machine that didn't use superconductors at all somewhere.

NERV - 18-2-2007 at 23:22

Quote:

Originally posted by Polverone
Are there actually any NMR machines that get away with using high temperature superconductors and liquid nitrogen? I thought they all used niobium alloys at liquid helium temperatures -- and helium's a lot more expensive, and has much lower cooling capacity to boot. I seem to recall seeing plans for a (low frequency) DIY NMR machine that didn't use superconductors at all somewhere.

You wouldn't happen to know where to find these plans? It seems like an interesting, and challenging project for an amateur scientist. Plus who wouldn't want to proudly display there own home built NMR system. Now that's serious bragging rights!

Sauron - 19-2-2007 at 02:42

I have a Sequoia Turner 340 UV-VIS, a large pile of Waters HPLC analytical (mostly Mod 600) and preparative (Mod 4000) sets (controllers and FHUs) and detectors two 486 and four 490 and a couple older PDAs 990, 991, and 994. The 990/991 are good for nothing as I don't have the dedicated PCs and interfaces that ran them though I do have the software. Never mind, I paid little for them.

So my shopping list, desultorily, is:

FTIR not too expensive to get into

NMR: Long range project. I am told you can get a second hand PNMR for about $25,000.

GC. Lots of 5890s around since Agilent dropped all support. They run $1000-$2600 without frills (autoinjector etc.) or up to double that with. Typically they will come with one or two FIDs, sometimes a TCD and a FID. Skip ones with radioactive detectors as you will have regulatory problems. Most desirable are the 5890 Series II+ which have advanced circuitry and better upgrade paths. You probably will want split/splitless injector(s).

Two things to hassle over with GCs: you will need a decent chromatography manager software package and unless you are lucky and get a machine with one installed, a HPIB (GPIB) card, that is an IEEE-488 interface card. Forget Agilent's software as they will rape you. But the most common third party GC software is a piece of shit. There are alternatives but you can expect to pay as much for the software as you do for the instrument.

Another is carrier gas. Americans use H2, Europeans prefer He, you can avoid having a big old tank of either around by investing in a H2 generator to feed your hungry lil ole GC that way. Again expect to spend as much as you did for the instrument.

At that point you are up to $7K-10K

And that is why I went with HPLC.

My ex partner does GC-MS and loves it but, I never got into mass spec. While I never say never I probably have enough on my plate for the balance of my lifetime, lord willing and the creek don't rise.

Ozone - 19-2-2007 at 08:01

@Sauron: That sounds pretty much like my rig. The ST 340's are robust and programmable (a good, relatively inexpensive choice).

I want a scanning spec, though (I have a Hach DR 2000 with data collection via chart-out to ADC with recording using Peaknet v5), A waters 510 pump, ABI absorbance detector (I have 2 differential refractive index detectors, a Waters and an Alltech) with a variety of column phases and sizes; my biggest HPLC (rather than using the pump on an LPC glass column) column is an original Dupont zorbax CN-Pr 250mm X 44mm (ID).

The 5890 series II are better GCs than the 6890, but they don't have automated flow control (which makes the 6890 win the bet). If you get one, make sure to also acquire a digital flow meter (if youve ever used a bubble flow meter, you know why!). Also keep in mind that GPIB (they called them HPIB boards) were ~$1000US *when they were supported*.

It's obscenely expensive, but Chemstation is, by far, my favorite (and most intuitive) chromatography interface. Peaksimple is crap. Peaknet (Dionex) 5 and 5.02 is good, intuitive, flexible, and can be had if you ask nicely--but--for you will need a Dionex ACI (the most flexible ADC w/ttl support I know of) or a Dionex UI-20 (straight ADC/ttl to TCP-IP or BNC) to interface it.

We use He most of the time for GC carrier gas, H2 occasionally or maybe P5 with an ECD. I was under the impression that in most places besides the US, He was prohibitively expensive (as you need 2 things for minable He, old U deposits and geology suitable for pocket capture)? I know, for example, that in South Africa they tend to adapt plumbing, mass flow, and method development around N2 carrier (once I found this out, it was obvious why my chromatograms looked so different

).

FTIR is definitely doable, and to NMR, godspeed. I'm trying to find out this rt NMR that Polverone seems to remember. It would have to very low field (even 50MHz were cryo), but might be fun to construct nonetheless. Following this an ESR seems doable...

Cheers,

O3

For fun, a picture of Florida State's new 900 MHz, 110mm ultra wide bore, 21.1T SS NMR! Keep in mind how large the Dewar to the lower left actually is

.

[Edited on 19-2-2007 by Ozone]

[Edited on 19-2-2007 by Ozone]

FSUs 900MHz 110mm 21.1T NMR.jpg - 20kB

Sauron - 19-2-2007 at 09:28

The 5890 Series II+ had AFC which is the "advanced circuitry" I referred to in earlier post, the Series II didn't but could be upgraded to it, the earlier versions could not be upgraded.

It's going to take me years to get the HPLC stuff all plumbed and settled in. Another reason I went with Waters HPLC vs HP GC: All the Waters stuff is 110/220 V 50/60 Hz. All you need to do is change out the fuses (sometimes internally too) and away you go. A 5890 is 110 V, they made 220V models but I have never seen one on used market. So a major hassle to use one here, I can step down the local voltage but still have to deal with frequency difference. An inverter will do the trick but that's still more $$$.

An NMR is pretty much just a dream.

Yeah HPIB mean Hewlett Packard Interface Board, but what the hell - they invented them so why not? GP just means General Purpose.

The Waters software (now yesterday's falvor since they switched to Empower) is Millenium32, I use v3.2 w/PDA option under Win98. It works.

The lousy 3rd party GC/LC software is Peak Simple which will run you $1500+ but itn't worth it. I will double check on the superior alternative, it is out of Czech Republic. It runs just a little more than Peak Simple.

unionised - 19-2-2007 at 12:42

Just for the record, the first NMR I used had a permanent magnet (30MHz IIRC). The only problem with NMR is the photons are too low energy so it's a bit of a pig trying to get decent sensitivity. On the other hand if what you are trying to do is characterise a pure product they are great.
By an odd coincidence I use the same HPLC software as Sauron, but I'm using it at work. We really need to upgrade our LC kit.
The toy I would personally like to have is an FTIR microscope. Is there anyone out there with much more money than sense?

Ozone - 19-2-2007 at 13:18

Not me! The nearest FTIR uscope I can get to is at CAMD. Obviously, the synchrotron facility is not worried overmuch about instrument costs.

But I agree, the FTIR 'scope is about as bad-ass as you get (have not played with a Raman yet, though).

Awesome photos too,

O3

leu - 19-2-2007 at 17:59

Quote:

I seem to recall seeing plans for a (low frequency) DIY NMR machine that didn't use superconductors at all somewhere.

This might be the page you saw:

http://www.geocities.com/CapeCanaveral/2404/

The kind people at Exstrom will provide plans upon request:

http://www.exstrom.com/magnum.html

An online source for other NMR equipment:

http://www.sdsnmr.com/larmor.htm

From the SA article

In the apparatus designed at the Aero Medical Laboratory the magnetic biasing field is supplied by a Type 220A 150 surplus magnetron magnet. The pole faces of the magnet were replaced by soft iron disks 3 1/2 inches in diameter and 7/8 inch thick to provide a field over a large area. For maximum response all protons must precess at the same rate, which means that all must be acted upon uniformly by the modulated biasing field. The intensity of the field will vary with the distance between the pole faces. Hence these must be made parallel and free from surface irregularities. Surplus magnets from magnetrons of the radial-cathode type usually bear a small white dot on the base which gives an approximate figure in gauss for the field strength that may be expected in the air gap. The magnet used in the instrument constructed at the Aero :Medical Laboratory is rated at 1,450 gauss. It was modulated by a coil consisting of 20 turns of No. 30 cotton-covered magnet wire wound on a Bakelite tube 1 5/8 inches in outside diameter and 7/8 inch long. Ten turns of the coil are wound at one end of the tube and 10 turns are wound in the same direction at the other end. A hole 5/8 inch in diameter is cut in the center of the coil form to admit the test tube. A second hole 3/8 inch in diameter is made at right angles to the first to admit a length of coaxial cable for linking the oscillator coil to the source of high-frequency current. The modulating coil is energized by the transformer which supplies the tube heaters, and it sweeps the strength of the biasing field 50 gauss above and below its mean value.
The test tube is 12 millimeters in diameter and 75 millimeters long. A two layer coil of No. 22 enameled magnet wire, consisting of 16 turns per layer, is wound on the straight portion of the tube as close as possible to the closed end. The tube and coil are mounted vertically in the Bakelite form on which the modulating coil is wound.
The circuit construction is conventional. The oscillator is designed around a 6AK5 pentode tube. When used with an oscilloscope of high sensitivity, output from the oscillator may be taken at the junction between the 22,000-ohm resistor and the 200,000-ohm resistor in the plate circuit. With 'scopes of lower sensitivity, such as the Heathkit Model 0-10, a single-stage amplifier using a 6AU6 pentode is added as shown in the circuit diagram. A variable capacitor, such as the Hammarlund Type MG 140-MI, is used for adjusting the frequency of the oscillator. These components are assembled on an aluminum chassis three inches high, five inches wide and six inches long. Input and output connections are made through RC 58/U coaxial cable equipped with UC 290/U and UG 88/U terminals. Power may be taken from any supply capable of delivering 100 milliamperes of direct current at 150 volts to the tube heaters and 60-cycle alternating current at 6.3 volts to the modulating coil.

The test solution is prepared by dissolving .4 gram of ferric nitrate in 100 cubic centimeters of distilled water. Two cubic centimeters of this solution are added to the test tube and placed in the biasing field. Power is applied. After the horizontal-sweep circuit of the oscilloscope has been made synchronous with the 60-cycle modulating voltage, a pattern should appear on the screen. The pattern may resemble a horizontal figure eight, as shown at left in the illustration. This indicates that the frequency of the oscillator coil lies outside the limits within which the particles are precessing and that resonance is not established. To search for resonance, set the oscillator capacitor for minimum frequency (the plates of the capacitor meshed fully) and adjust the intensity (feedback) control to the point where the oscillator is on the verge of going out of operation. Then increase the frequency slowly while observing the scope. It may be necessary to trim the feedback control occasionally to maintain the marginal oscillating condition. The procedure can be simplified with the aid of a short-wave radio receiver. If the receiver is equipped for continuous wave-reception, the oscillator signal will be heard as a shrill whistle. If not, it will make a rushing sound, perhaps accompanied by a 60-cycle hum. The receiver is particularly useful in checking the point at which the oscillator goes out of operation when adjusting the feedback control. If the receiver is calibrated, it may be used to calibrate the oscillator. If not, the receiver can be calibrated easily by tuning in on the time signal of Station WWV.
When resonance is established, the display will resemble the center figure. Usually two peaks appear which are joined at the bottom by loops. This indicates a displacement (phase difference) in the time at which signal arrives, at the vertical and horizontal plate of the 'scope. The Heathkit Model 'scope is equipped with a line sweep switch and a phase control for manipulating the display. When these are properly adjusted, the peaks coincide, shown in the figure at right.
What does the display mean? The height of the figure is proportional to the number of protons resonating with the oscillator; the width of the figure, to the range through which the particles precess. Accordingly if all of the particles were precessing at precisely the same rate and all flipped over precisely in resonance with the oscillator, the pattern would resemble an inverted T. The spectrometer could then be said to have perfect resolution. Evidently in the instrument all the particles do not precess at the same frequency. Part of the explanation lies in the interaction of magnetic forces within the test sample, The fields of neighboring protons merge in such a way that some particles are partially shielded from the influence of the outside field. But in this instrument the breadth of the peaks is large, explained by cross-sectional variations in the strength of the biasing field. Particles in regions of high-field intensity precess at higher rates than those in regions where the field is relatively weak. These differences are preserved when the field is modulated. Some particles are swept into resonance with the oscillator earlier or later than others, and the displayed peak is broadened accordingly. The width of the peak illustrated is 26 gauss, which means a difference of 85,000 revolutions per second in the of precession of the slowest and fastest particles.
With an instrument of high resolution many substances show fine multiple peaks. This is due to the complex magnetic interaction between systems of particles and the consequent shielding of the biasing field. Many substances are not sensitive to an external magnetic field because the magnetism of the spinning particles cancels out. But those substances that do respond can be identified by the characteristic pattern that shows up on the 'scope. The resolution of the apparatus described here is high enough for fine spectroscopic work As indicated earlier, it is intended to serve as a simple demonstration of the magnetic-resonance effect.
Modifications to adapt the apparatus for limited applications would include the provision of larger pole faces on the magnetron magnet to provide a more uniform biasing field. In contrast the 20-gauss peak-width displayed by the apparatus, the best instruments made. today resolve to a few ten thousands of a gauss; this means that irregularities in the biasing field must be kept below this figure. High resolution also requires precise and calibrated control of the intensity, frequency and amplitude of the biasing field. in this demonstration the. high sweep-rate of 60 cycles per second,: is made possible by limiting the experiment to a test solution of ferric nitrate. Few substances are so responsive.

Some images are attached:

Attachment: nmr.zip (171kB)
This file has been downloaded 738 times

Ozone - 19-2-2007 at 19:19

Holy shitballs, Leu!

That actually looks doable. No doubt the bulk of the cost lies in the power supplies and DAQ board. I think I'll try to contact them for some schematics or equivalence RE. the amps.

And:

------------------------------------------------------------------------------

Economical 60/90MHz FTNMR Instruments
Anasazi Instruments, Inc. provides complete FT-NMR systems and CW-NMR console upgrades for educational and industrial applications. Using an existing 60 or 90MHz permanent magnet (Varian EM360 or EM390), Anasazi provides all the necessary components to perform most modern measurements.

Complete NMR systems use refurbished EM360 magnets. These systems are ideal for sites which cannot support a superconducting magnet because of budget or availability of cryogens.

The Eft spectrometer can be operated by personnel with little or no FT-NMR experience. All instrument functions are computer controlled including magnet shimming.

The Eft spectrometer is an inexpensive, easy to use and robust FT-NMR spectrometer that obtains H1 spectra in less than three minutes and C13 spectra in less than ten minutes. Currently the Eft is available in three configurations: H1, C13/H1, and Wideband/H1 which covers the ranges from H1 to F19 on the high frequency channel and P31 to Si29 on the low frequency channel.

The Eft spectrometer is covered by a five-year warranty and exceptional service.

For more information on Eft NMR spectrometers please contact:

Anasazi Instruments, Inc.
4101 Cashard Ave., #103
Indianapolis, IN 46203
Phone: 866 494-9369 toll free
Email: sales@aiinmr.com
Web: www.aiinmr.com

I emailed asking for a quote on their HETCOR capable model. I'll let you all know what happens, viz. how much a 50MHz NMR goes for these days (heck, if it's reasonable I'm thinking about trying to get one for work that the students can use).

Cool!

O3

[Edited on 20-2-2007 by Ozone]

12AX7 - 19-2-2007 at 21:08

Hmm little more than a regenerative radio, huh? That is pretty easy.

"Power may be taken from any supply capable of delivering 100 milliamperes of direct current at 150 volts to the tube heaters and 60-cycle alternating current at 6.3 volts to the modulating coil."

Heh....heh....um....yeah, you'll need a few orders of magnitude more than 100mA to run the heaters at 150V! Fortunately, transistor regenerative recievers are also easy to build (though you only get that mystique when you use tubes

).

Tim

Sauron - 20-2-2007 at 19:05

@O3, I for one will be most interested to hear how they quote for their baseline PNMR setup w/magnet.

Ozone - 20-2-2007 at 19:21

Righto!

They got back to me almost immediately! The cost for a HETCOR capable instrument with computer, software and everything (installation, training, etc.) was $74,000; this is about the cost of a GC/MS with an autosampler. I know it sounds crazy, but that's pretty cheap (and they have *free* support with a 5 year warranty).

The best part: Permanent magnets. 50-90MHz (which, face it, is good enough for normal everyday structure elucidation of "small" molecules) without He or N2 (l); granted, the 90MHz is *very heavy*, but I think that delivery is included.

I think that this is the ceapest, easiest to keep-up, and comparably functional instrument I have ever seen for the money (hey, Chemistry *is* and expensive hobby

).

Cheers,

O3

I'd expect the straight 1H rig to be *significantly* cheaper.

[Edited on 21-2-2007 by Ozone]

Attachment: Brochure.pdf (1.8MB)
This file has been downloaded 935 times

pantone159 - 20-2-2007 at 21:27

Still, that's out of reach of most home users

My Mad Scientist vision is an NMR, that politely sits dormant when needed, but when the magnets are immersed in vats of LN2, just picked up from a welding supplier or something, condensation fog billowing, it all cools down to high-temp superconducting, and the low-budget home-brew NMR is ready to go.

Maybe someday.

Polverone - 20-2-2007 at 21:33

That is one heavy magnet on the 90 MHz version. Is that the major cost of a permanent-magnet NMR? I imagine that there's quite a markup on instruments like these because they're produced on a limited scale, but even if you were making a million a year of them, I doubt you can really make a "cheap" 2200 pound magnet.

I could imagine maybe dropping $35,000 on a machine like this -- I'd rather have it than a new car -- but $74,000 is a bit much. That's almost half the cost of a house around here.

Ozone - 20-2-2007 at 21:39

That's what I meant by heavy!

I'm pretty sure that the 1H machine will be quite a bit cheaper, maybe in the 35K range. Remember, I requested the HETCOR (COSY, 1H, 13C). I'll see what the 1H costs.

He aint heavy,

O3

Sauron - 20-2-2007 at 23:38

That is actually, for what it is, very cheap.

Please do get a quote on the basic version.

I was told to expect a used instrument to run $25K so c.$35K for a new one is not intimidating. It would take me 18-24 months of scrimping and savings to put that together (or longer) but I have done it before.

@Polverone, if I read the original remarks about Anasazi cirrectly, they are using refurbished magnets which would certainly reduce the cost as long as the supply holds out. Also whenever possible they utilize a customer's existing magnet.

@O3 was spot on, these things are priced even at the top end, like a GC/MS or a HPLC/MS.

I am going to point a couple of my Thai analytical-instrument dealer friends at this company just in case they do not have a Thai agent yet. If they already have a dealer that may be bad news for me, as the local dealer likely will mark the thing way up. In a kickback-driven economy it is easy to see why.

[Edited on 21-2-2007 by Sauron]

[Edited on 21-2-2007 by Sauron]

Sauron - 21-2-2007 at 06:10

@leu, I believe the article you quoted from was not in SA (Scientific American) but rather in the April 1959 issue of The Amateur Scientist, which is (or maybe was) published by the Society for Amateur Science. Anyway the website with the Amateur Scientist Collection (1920s to 1990s) is indeed sponsored by SA and The Tinker's Toolkit, so that's probably what you meant, and I didn't get it.

However I already spotted at least one VERY interesting article in the January 59 issue which will be subject of its own thread as it definitely deserves the descriptor Mad Science, nothing this mad since Ghostbusters.

Anyway the primitive NMR described is not intended for analytical use at all but merely technology demonstration and was estimated to cost in 1959 $ - $100.

It is however interesting.

Here's the attachment for the PDF of the NMR article.

[Edited on 21-2-2007 by Sauron]

Attachment: DIY NMR.pdf (125kB)
This file has been downloaded 2256 times

Maya - 21-2-2007 at 08:55

Stong definitely was the father of amatuer science!

Ozone - 21-2-2007 at 17:12

@Sauron,

Done. I should get a reply in a day or two.

Cheers,

O3

Sauron - 23-2-2007 at 22:09

I've been watching labx and surpluslab for years without seeing a non-FT (grating) IR come up for sale that I can recall.

If one did, I'd be happy to buy one if price was right. Pass the KBr. Nujol anyone?

I have seen a few FT-IR machines for sale.

leu - 24-2-2007 at 16:46

If you want a device that's not sold on LabX then perhaps you should look elsewhere since there are many other sources of used equipment

The fact is that the Amateur Scientist was a column conducted by C L Stong in Scientific American for many years and was very well known to any American, Canadian or other English speaking scientist who read that popular publication

Attachment: stong.zip (336kB)
This file has been downloaded 695 times

Sauron - 24-2-2007 at 19:02

I do look in other places but many of them I find disappointing; many I have done business with and been disappointed by. Some like eBay I refuse to deal with any longer. I could list the companies and individuals I have bought from but it would be a LONG list.

At the time Stong wrote that particular article I was 8 years old. I am most definitely an English speaking American scientist, but I was not a regular reader of SA, so being such is not a sine qua non. I have no recollection at all of The Amatuer Scientist so if I ever saw it in the pages of SA it made very little impression on me as a child, an adolescent or as an adult. I was busy reading JACS and JOC, science fiction, other literature, my uncle's back file of ARRL magazines, etc etc and somehow SA just was not high on my agenda. I have not even touched on the gun journals, and later the defense journals, I ended up as an editor for one of those and a contrinutor for many others.

In the immortal words of @Orgie: "Sorry, my bad."

[Edited on 25-2-2007 by Sauron]

hinz - 13-3-2007 at 13:42

I hope anyone can help me with the next problem with the FTIR.
Beginning is here:
http://www.sciencemadness.org/talk/viewthread.php?tid=7770
(WIN-IR is running properly, the problem was caused by a dongle in the printer port as copy protection for WIN-IR and I thought the copy protection would be read from the FTIR's interface so I overlooked the dongle which looked rather like a printer adapter than a protection dongle, but after I opened it and saw a chip everything was clear.)

Now I tried to get a proper spectra of a polystyrene film which I got as testing sample for the FTIR, but the spectra starts suddenly to oscillate wildly from 1/2500cm to the lower wavenumbers/higher wavelenght

(photos I uploaded), but between 1/4000cm and 1/2500cm are exactly the peaks you would expect of a polystyrene transmission sample like this one:
http://animl.sourceforge.net/Minutes/2005/20050617%20Virtual...

There's one piezoelectrically cooled DTGS detector inside, but DTGS detectors are normally rated for the entire mid-IR range, so this couldn't be the the reason.

It might be caused by the beamsplitter, I don't know from what it's made of (transparent in normal light,so maybe KBr or KRS-5) but can a opaque beamsplitter (if not made from something mentioned above) cause such a sudden oszillating?
When I block the beam by placing a piece of wood into the sample cell, the oszillating is in the whole area from 4000-400cm^-1
Could the reference laser for the infernometer cause this, if you look close to the photo with the detector you can see the laser beam on the detector, it's lasering through the sample cell, but I think it shouldn't do this.

I hope anyone with experience with FTIR's could help here, probably I've overlooked some settings in the software.

EDIT: Uploaded picture instead of .rar file

[Edited on 14-3-2007 by hinz]

spectra.jpg - 44kB

unionised - 14-3-2007 at 13:14

It's probably just my computer but I can't download that file. For the benfit of stupid people like me could you screen dump the image and copy it into a post here please? I'm not saying I will be able to help but I'm curious about this problem.

hinz - 14-3-2007 at 13:56

Second picture, the detector with the laser beam on it.
Today I took a spectra without background and with the intensity of the beam at the Y-axis instead of absorbance, and the intensity gets down from a high value at 1/4000cm to zero at 1/2500cm and if you look close at the spectra above, you can see an increase of noise between 1/2700 and 1/2500cm (the wavy line). It's because the signal to noise ratio gets worse till the computer doesn't know how to handle the signal and thus makes an oscillating line.

So either the beamsplitter or the detector isn't made for those wavelengths, but if you look at the detector you can clearly read "cooles TGS" which stands for Triglycine sulfate and should handle at least IR light between 1/4000 and 1/400cm.

Do you think I could place something like a second IR source (a piece of electrically heated NiCr wire) in the sample compartment and look whether the detector responds on it. Thereby I could bypass the infernometer and beamsplitters. Are you sure that this won't harm the detector if I only heat it slightly and don't place it in front of the detector.

PS: Now I know why the nice analytic equipment is that cheap at E-bay, because you have a lot of trouble till it works how it should if it ever do so

[Edited on 14-3-2007 by hinz]

DTGS detector.jpg - 61kB

unionised - 16-3-2007 at 11:15

Thanks for the picture. I guess you are right and the beamsplitter or detector doesn't respond properly. That seems odd because a TGS detector should be OK. Is there any way to check if the cooling is working?
Is there a setting in the software that controls the path length difference when the machine scans?
Unfortunately, I think you might be right about buying cheap stuff on ebay.

S.C. Wack - 16-3-2007 at 15:32

That's your fault for using IE or some other crap browser. Looks fine on my monitor.

Stuff that appears on ebay often got there for good reason, or the people who put it there often put one part in one lot, and another essential part in another lot, because they have no clue and often got it at the local auction the same way.

12AX7 - 16-3-2007 at 18:00

Quote:

Originally posted by S.C. Wack
That's your fault for using IE or some other crap browser. Looks fine on my monitor.

Well, no. IE works just fine, and besides that, I'm using Firefox and the previous page is stretched (slightly). I'm running 1280x1024 (this window is not maximized). Do you also imply, then, that FF is a "crap browser"? I'm curious if any browser would count as anything other than "crap" based on your reasoning.

Yes, that's right, I AM going to singe your asshairs, right here and right now, for generalizing about a Microsoft product. You have an IRRATIONAL FEAR of IE (and likely, OE and other products) that reflects very poorly on you in deed!

Tim

not_important - 16-3-2007 at 22:16

hinz

Your plotting the intensity was what I would have done. It sure looks like something is either absorbing the longer wavelengths, or the detector is foobar.

Most pyroelectric detectors are very broadband, while not DC to x-ray them will work at least from in the visible on out to 50 microns.

Your detector is working in that it responds to the short wave stuff, so you know that it and the electronics work. This doesn't tell you if the detector is damaged in some wave that affects its working band.

First thing I'd look for is any sign of a filter being added in. Also try to get the part number for the detector, see if it is stock or special.

It's also try your idea of another IR source, but put it so the light goes through the full path so you can get intensity readings vs wavelength off it. Unless you've a way to measure the detector output, having it see the naked IR source won't tell you much, you don't know if it is responding to the longer wavelengths or not.

Some examples of commercial detectors here
http://www.oss.goodrich.com/ThermalIRDetectors.shtml

S.C. Wack - 17-3-2007 at 04:19

12AX7 - Yes! By all means, let's derail yet another thread with nonsense that could be PM'd if mentioned at all. Unfortunately I must participate in this because I won't let that go. Sorry, hinz. As you can see here,
http://www.box.net/shared/3u5hqqfyh1
not only am I running Firefox, the only web browser that I use, but the page is not stretched at all.
So as someone running XP, I'm not qualified to diss IE? Even though you're right, it does not stretch the page. Opinion is now irrational fear? Again I concede the last word, because I'm not going to shit here more than necessary. The point was - in response to Sauron's now deleted post, previously immediately above mine; saying that the pictures should be resized because they stretch the page - there was not a size problem.

unionised - 21-3-2007 at 11:16

Just to contribute more pointless junk to this "debate" I couldn't load it as a .rar file but it works fine since it got changed.
It works on this machine in firefox and IE.

SCW, did you actually try to load the original file, it seems unlikely since you only commented on it after it was changed to one that works fine?

Ozone - 21-3-2007 at 12:02

I have seen this sort of noise before (albeit on UV-VIS scanning instruments). The result was a faulty (out-of-calibration) monochromator which had the instrument thinking it was seeing one wavelength when it wasn't. This resulted in "insufficient signal" and noise was returned. Running a wavelength calibration solved the problem. is there an autocal or wavelength calibration option in your software (it might be under something like "user maintenance" or somesuch.

Cheers,

O3

len1 - 16-4-2007 at 00:38

I got an FTIR off eBay to work recently so Ill hazard a guess at your problem. Mine took two months work of my spare time and I had to reverse engineer just about every part of the instrument since PE said they 'dont sell service manuals'. However I expected quite a bit of work, since there had to be a reason why the instrument was cheap. Your instrument looks newer, and its problem not as severe as mine, so I suspect you got a better deal.

First, I presume your instrument is running in ratio mode, as otherwise its most likely an electronics fault.

If so, it could be that a deterioration in either the detector or IR source has meant that your signal, which is not too good at short wavelengths either, has dipped below the noise level at long wavelengths. However the most frequent cause of FTIR failure is the beamsplitter. (This would
be easiest to diagnose if you ran the instrument in the inteferrogram mode, ie gathered the raw data. We need to see the reason for the oscillations (if there is one) at the FFT level) In that case all you have to do is up the overall throughput, so the signal does not fall below the noise at large wavelengths as well as the short. To up the power you can

1) Align the adjustable fixed mirror azimuthally and elevation wise - these alignments normally require removal of purge cover, so have a dry environment before you do this. This would attone for beamsplitter deterioration in angular performance

2) Up the electronic gain. My FTIR had a varibale pot for this on the analog board.

Heres a spectrum of a thin ethanol film I took yesterday - as good as on a new FTIR, (compare: http://www.800mainstreet.com/irsp/eir.html) (shows that a dumped FTIR with enough attention can clearly still perform)

As a question: does anyone have any detailed reference to 'catastrophic beamsplitter failure'? This is the warning on the purge cover of my instrument should you fail to change the dessicant at regular intervals. Yet my instrument looked like it was kept inside a shower (90% humidity outside cover, 50% inside), but was able to be coaxed to work. Has anyone seen what a beamsplitter that has undergone catastrophic failure looks like?? I notice that a replacement beamsplitter from PE costs 1/2 the price of the original instrument. Maybe this has something to do with it, ie PE want you to think that the beamsplitter has 'catastrophically failed' so you would not be tempted to adjust the instrument but instead buy a new beamsplitter (or even better a new FTIR).

Another question: is there any sense in buying an FTIR instrument without the computer and software? Is PE FTIR computer control generic? If so can the software be readily obtained (I presume not from PE seeing they dont even want to sell the service manual)

[Edited on 19-4-2007 by len1]

C6H5OH_IR_3.bmp - 235kB

chemrox - 16-4-2007 at 21:15

This is all very nice .. I love the idea of building a magnetic resonance spectrometer. I like the idea of building a polarimeter better. The latter would be useful the homemade nmr is just a curiosity. I mean how could you possibly use it?

hinz - 18-4-2007 at 15:32

Hello len1, you thin film spectra looks nice, I hope I'll get mine to work as well as your's. I've found the error in my FTIR quite some time ago, it was the beam splitter and not the detector as I supposed. The beamsplitter which was mounted was made of quartz and it is only translucent up to 4um / 1/2500cm so the oscillations is noise which is amplified till it fills the whole screen.

Here is a absorption spectra of quartz:
http://www.korth.de/transmiss/sio2.htm

Inside the beamsplitter compartment was another beamsplitter, I realised afterwards that this one was made form KBr, because I wondered for what the hole in the middle of the beamsplitter was and why the FTIR didn't found the reference laser beam with it. At first I thought it would be some kind of aperture due the hole in the middle (the quartz beamplitter is half transparent mirrored entirely).
As I opened the cover I saw that the laser beam passes through the hole in the middle of the beamsplitter on 3 photodiodes. Now it was clear to me that the hole is for the laser beam, but it doesn't still find the reference laser. I think it's because the KBr disks aren't really translucent by visible light and the mirror in the middle for the laser has drawn some water and isn't 50% reflective any more.

At the photodiodes are 3 variable resistors, probably to adjust the gain of each photodiode, but I'm not sure about this. It's strange, but wouldn't 1 photodiode be enough to measure the interferences?

I was at a local universitylast week and there I asked someone if he could get me a new beamsplitter or if he would know from where I could get a new beamsplitter and he replid that it would be difficult, because chemical IR strucure analysis of chemical compounds is dying out and thus the university doesn't have a lot of FTIR's any more.

How much costs a beamsplitter for your instrument ecactly len1, I suppose from the screen on the picture it's a PE 1600. Probably I have to buy a new one from Biorad.

BTW: The beamsplitter construction is bad. Why couldn't they place the half translucent laser mirror coaxial to the KBr mirror, So they could make the laser mirror from glas and you could just order a half germanium mirrored KBr disk from one of those various supppliers which also sell output mirrors made of KBr for CO2 lasers.

Picture of the beamsplitter:

[Edited on 19-4-2007 by hinz]

beamsplitter.jpg - 42kB

hinz - 18-4-2007 at 15:36

The 3 potentiometers and photodiodes for detection of the laser beam:

EDIT: better photo

[Edited on 19-4-2007 by hinz]

HPIM3140.bmp - 1.4MB

len1 - 18-4-2007 at 20:45

Hi Hinz

Of course, the solution to the problem was right on the screen, the fact that the transition to noise occured at 2500cm-1 just where quartz stops transmitting. I should have thought of that. Mind you I have not encountered quartz beamsplitters before, Im mainly interested in organic analysis, where quartz blocks half the region of interest.

I did not quite understand your second paragraph (I presume you are German speaking right?) so please excuse if my response is not to what you were asking.

KBr transmits visible perfectly, it looks just like glass and has a similar index. The KBr is normally coated with a film to protect the KBr from mjoisture and reduce reflection at the first surface, and a half mirror at the main interface. This coating while good for IR scatters visible light, and hence the coating is absent where the laser beam is to pass through. My PE 1600 (your guess was spot on, obviously are familiar with IR's) has a small notch in the coating on the side of the beamsplitter (not in the middle like yours) for the laser.

I got well below 50% laser reflection at the half mirror - its only about 16%, yet the FTIR can still be got to work in those circumstances.

I dont know if I understood you correctly: you do get laser signal detected properly on the quartz (else you could not have produced the absorbance diagram above) but it is failing with the second splitter you have shown? If it is failing what is the error message?

Ill wait for a reply before saying any more, but I do note that the half-mirror surface appears much darker than that on my FTIR. Mine you can actually see through visibly. I dont know if this appearance is due to oxidation, or whether indeed some metal coating like that are used. If so, I have not previoysly seen them. Maybe someone else can correct me. In any case the laser beam should have no problem passing the KBr - that is not a coating problem.

To answer your question, a new beamsplitter from PE costs $US4400. Far too much for me. As I said I think its a deliberate rip-off by PE, and most FTIR could be got to work as is. Len

PS Can you take a better picture, I have problems in seeing the path from beamsplitter to laser detector (and its very dark and low resolution)

[Edited on 19-4-2007 by len1]

hinz - 19-4-2007 at 13:31

The quartz beamsplitter is working properly, but I'm also interested in organic structure analysis thus I want to get the FTIR work with the KBr beamsplitter. But when I mount the KBr beamsp. in the FTIR, I always get a "No laser detected" error.
I think this is because the mirror in the middle for the laser is damaged by moisture. The outer surface (Ge-mirror is between the two disks) is also slightly etched by water so if I try to look trough it, I see everything blurred, like if you look trough a glass plate you once drpped in HF. I wonder if there was once a moisture protection on it.

The mirror of the beamsp. is completly intransparent in visible light, you can't look trough it, but I think this is normal, because germanium becomes transparent in the IR-region. Here is a spectra of germanium: http://korth.de/transmiss/ge.htm

I think I'll contact Biorad sometimes in the near future and ask them how much a new KBr beamsplitter costs, I hope they don't rip you off like PE does it. Then I could also ask them for a service manual, maybe there is written for what the three potentiometers at the laser interference detection circuit board are.
Thanks for your help with my FTIR and you guessed right, I'm German speaking

len1 - 19-4-2007 at 17:31

Hi Hinz,

The new picture is much better I can see whats going on now. After the beamsplitter there is a germanium mirror with a hole in it to allow the laser beam to pass through a filter and then to photodiodes behind it, while the IR ir reflected at 90 degrees to a detector. The PE 1600 is arranged differently, the laser passes through a visible transparent region in a notch on the side of the mirror, and is stopped by a single photodetector a few centimeters behind the notch.

You've also explained better what your problem is. Next question: do you see the laser beam strike any of the photodiodes? As I said the KBr is transparent to visible, you can see right through it, so the laser must pass through as well. You said the surface at the beamsplitter center appears fogged, but in the picture you posted you still can see the pattern in the paper underneath through it, so its not bad enough to completely scatter the laser signal.

Lets assume the laser beam (in actual fact two strong beams should exit the interferrometer) passes thru but doesnt strike any of the photodiodes. That was one of the problems in my spectrometer. Slight alterations in the refractive index of coatings of the beamsplitter with time are then at fault. The laser beam is refleted onto the beamsplitter from the laser via an adjustable mirror, adjust the two degrees of freedom of this mirror so the beam strikes the diodes.

Second case: the beam does strike one of the photodiodes. That could still give you a laser fault message. Thats because the electronics looks for beats between the laser beams traversing the two paths of the interferrometer. In this case you must adjust the fixed mirror at the back of the beamsplitter to make sure both beams strike the photodiodes head on. They must do this to interferre. If the computer sees no interference it still reckons theres no laser signal.

Now as to why there are 3 photodiodes. Without having the spectrometer in front of me I cant tell you 100%, but there are really only these possibilities I believe

1) The spectrometer changes the detector diode in response to changed settings. For instance when you change beamsplitters, the slightly different refraction index of the KBr channels the signal to a different diode than for the quartz.

2) Your spectrometer has attachments for an external detector, which would result in the laser beam following a slightly different course

3) There are facilities for different lasers/positionings.

If (1) is the case, then thats a possible source of the problem responsible for your fault. Unless you have switched the software from quartz to KBr, the electronics will still be trying to read the quartz diode, and not see a signal.

I did not understand what you mean by "mirror in the middle for the laser is damaged by moisture", which mirror? The half mirror in the beamsplitter? All other Ge coated mirrors are not KBr and so should not be attacked by moisture. The laser mirror should be just glass. If you mean the half-mirror in the beamsplitter, then that is not so important as the laser signal is very strong, I found even a KBr interface with no coating still has enough reflective properties (in addition to refractive) to allow lock-on to the laser beat signal. In theory the reflective property of uncoated KBr is (1.52-1)/(1.52+1) or about 20% at normal incidence. Please show me what exactly you are referring to.

Ge is opaque to visible as you say. But

1) Ge is silver white, not brown as it appears in the photo. My beamsplitter and mirror surface all appear this way.

2) The coating is only a 50% mirror, and some visible light still gets through, at least in my case.

I had enquired the cost of beamsplitters from various manufacturers, when I was trying to substitute mine for anything available (this is before I knew that it could be restored) They all charge in the order US$3000-US$5000 for a new splitter. This is why FTIR's come cheap once the original (clueless) owners find their splitter has deteriorated. Because the equipment manufacturer has arranged the economics so as to entice the to acquire a new one.

I was interested to hear that FTIR's are dying out in Germany. Not here as far as Im aware. What is being used instead? (surely MS and NMR are too bulky, and the latter also expensive).

Another question: can you help with IR libraries, does your software come with them? Len

[Edited on 20-4-2007 by len1]

Matchheads - 19-4-2007 at 19:02

We had a FTIR from the other company besides Nicolette, based in Madison, WI, and they flew us out there, but this was 1988. I was being paid $8 per hour, and they wanted formulas broken, but we should have had other lab eqpt besides that. You can run a spectrum easily, but I never trusted the library to accurately tell us what it was. It seemed like you had to superimpose the two spectra and slide one into position along the ordinate to get a match. The commands were in DOS. If they aren't lined up in the computer's view, the "closest" thing to your spectrum would be a false positive. So, did they cure that? Second, when they gave us a sample, it would often be a product which was mostly water. You couldn't dry out samples using the same method on each one. It seemed like there was no telling how dry the library sample would have been. Some ingredients are hydrates. What do you do about that (drying samples). Somebody is into instrument building, not analytical chem. That is hinz, and the guy chatting with him is len1. You must work in a lab with a budget.

len1 - 19-4-2007 at 19:26

To answer your question, the FTIR is my own not the labs. I prefer to do something else with my $20K. In the lab it doesnt matter since its the customers money. Its interesting to hear about your experience. What was in the samples? Some things FTIRs are just not designed for, and MS would have been better (such as the VS plasma) Len

len1 - 22-4-2007 at 18:15

Hinz, if you can see the laser beam hitting one of the photodiodes, and the beam is single - that is its not split into two spots, then its a matter of uping the gain using the potentiometer corresponding to that particular diode. You can find out which it is by tracing the electronics on the back of the card.

I have found another interesting effect with the 1600. Any shock to the instrument produces a pronounced change in the background. Not just in the amplitude, but in the shape as well, meaning the instrument has to be realighned. The design of the 1600 is quite shoddy, by my expectations anyway. The beamsplitter and all interferrometer mirrors are held in place by two metal prongs, rather than a solid all-round fit. The moving mirrors are freely pivoted in a v-groove. I noticed that misalignment of the fixed mirror manifests itself in an unexpected way. As I said above any misalignment reduces background transmission, that is expected, BUT it also leads to a characteristic modulation of the interferrometer throughput. The background, rather than being humpback-shaped develops deep undulations. These are periodic, so if theres a minimum transmission at 800cm-1, there will also be at 1600cm-1 and 2400cm-1. These are quite broad (sinusoidal in fact) and so there is little chance of mistaking them for absorption peaks (except if they occur in the hydrogen-bonding region). Nonetheless the background needs be as flat as possible for maximum sensitivity. I had not heard of the undulation effect of misalignment before, but have now convinced myself that theoretically that is what you would expect.

The interferrometer works by inteferring a multitude of light beams, travelling along two paths. In the simple model one normally uses, one interferres just two beams whose path difference is dL giving sin(om dL) as the interference phase. Integrating over all emited wavelengths gives the usual Fourier transform of the incident spectrum. However if one approximates the adjustable beam by two light rays each travelling length dL1 and dL2 with respect to the fixed beam then we have

sin(om dL1) + sin(om dL2) = 2 sin(om (dL1+dL2)/2) sin(om (dL1-dL2)/2)

The second term modulates the output and is independent of position since dL1-dL2 is fixed. This then forms the sinusoidal filter I noted above. The filter in reality is not as sharp as above (it doesnt have zero transmission troughs) because a multitude of rays (instead of two) need to be integrated, rather than summed, above. This gives a filter of finite depth.

Matchheads - 11-5-2007 at 11:54

I have worked for a few employers, but the timeframe was 1978-1989. As I examine these companies in the intervening years, I try to take note of which ones survived and prospered and/or went out of business or changed their names, or, worst of all, had their name splashed all over the newspapers due to an industrial accident. Because, when you work there, you can see them headed for these various ends. I am not really serious about chemistry since my boss's idea of hard work may include me chipping out the last bit of CrO3 from a drum with rubber gloves and a hammer and screwdriver, but no respirator as that would cause alarm among the rest of the employees.

This particular company was named "Custom Chemical Formulators" of Cudahy, CA. It is where the 5 and 605 freeways cross in LA. Other towns around there include Bell and there is a lot you can do there you can't do elsewhere. The cities of Southgate, Carson, or Santa Fe Springs are not garden spots.

In chemistry, there are often whole yards where you can just have the chemicals. An industry exists in which you can charge $600 per drum to take drums away, but you better document what happens to them, and they are manifested with the word WASTE before the chemical name. You need to know what is in them. Many are intact. Out of a hundred, there may be only 20 with no label and being penetrated by rainwater, a classic chemical problem.

They hired me and my boss, a marijuana grower who had his 4-year degree from Cal State Fullerton, name of Bob Stone, to work in the lab together in 1988. I was just wondering why it went the way it did, since I had my UCSD degree by then, but, it went that way because our contact downstairs hired me to work with him. If he had been given the power to hire and fire, he would have gotten a girl without a degree to spend the days with.

As it was, they depended on me to do things like ride a forklift to the ceiling with a portable pH meter and adjust a batch, in a shop where the workers could not be taught to do that. They did not speak English. Thus, Bob was able to get me fired when I took time to attend a lab auction. It was worth it. I got two 12s and a 22-L boiling flask, a mantle, a Hot Shot, from Tauber Auctioneers. I could smell the phenylacetic acid in the equipment. Some other cooks beat me out of some good stuff, too. I registered as a Mr. Davidson, as in Harley-Davidson. This auction was in North Hollywood. It was a Goldilocks auction. One that was too big was Occidental Petroleum in Irvine. They always put things in lots, an auctioneer's trick; you either buy 15 Variacs or gas regulators (just to get Hydrogen), or you budget and get very little. An old glassblower is a good source. There was even a guy in Fullerton with a daughter who specialized in IR, but she's old now. Much of it used to be recorders that would stop working or you'd "just" have to find the chart paper for it, like computer printers are now.

People were interested in gold recovery, Kyocera has the best solution. It is not cyanide. By now they must have a good way to recycle catalytic converters from cars.

Bob told me I could "have" whatever I wanted from a lot of glassware that came in; it was from before standard taper. I should have believed him, but I just picked up a crucible with the bottom coming out and told the big boss, "Dude, this is platinum. I shouldn't take it. I guess it could be nickel, but I doubt it." He comes back and says somebody gave him $600 for it. I was getting $8 per hour. He was from the salesman end of it. He was buying iodine for making teat dip. I never even took the samples of HI they had. It wasn't "stabilized". But, we went through nothing but iodine. The lab was stained with iodine. We washed it off with hypo. There's only a limited amount of iodine they can buy from South America. It's "wet process". Another location which deals in drums of 57%HI and makes it is Deepwater Chemical across from Cal State Northridge. They are wise.

To answer your question Len I know. There is even a little drying tube for GCMS samples. These guys did not have what you would call professional courtesy with say, an environmental lab. What they mainly do is whenever a big company (Micrographics, I think, needs potassium dichromate to be opened up in someone else's facility), farms out hazardous operations, they do it. So, there are lots of one-time substances. i learned a lot. What you want if you have to do it all by hand is a GOOD wet chemistry book. They had to have that to begin making soaps, during the phosphate stage and before the ionic surfactant time. Rohm and Haas is good in anionic surfactants; they also have an ephedrine synthesis which goes through pyruvate. what confused me about old books was the use of something you titrated to and calked out to a "NaO" equivalent, or some phosphate number where I really wanted it to register as greater than 100%. It's hard to make soap go liquid so it will pump out of the mixer, or to get a high cleaning power, if somebody else's product has it. What is "polyphosphoric acid"? I never found out.

What was in the samples I remember was ethyl acetate. I solved that one. We had ODCB in that shop, which is "orthodichlorobenzene". If they come in the lab with an open sample of that you just scream at them what it is, and tell them to get it the hell out. We liked to say the initials, though. Another good one is bromoacetone. You could say it made P2P in a one-pot, no catalyst, quantitative yield.

Matchheads - 11-5-2007 at 12:03

Yeah, Len1, before I leave, I see your sample is swamped out with water. Also, you ran an ethanol film, so, you meant "ATTENUATION" when you used the word 'attention', huh?

len1 - 14-6-2007 at 23:03

The alcohol I used is not really that, its just some meths I had at hand, which contains 5% water.

I dont know why Im continuing this, but for completness sakes, Ive now found out what is meant by catastrophic beamsplitter failure, a question I posed before. That actually is the reason for the profusion of cheap second hand FTIR on the market. A replacement beamsplitter costs about half the price of the instrument for most models which is why owners (and of course manufacturers) prefer to upgrade.

Heres a picture of the failure:

KBr base is stable to humidity, it can not handle condensation. when that happens a thin surface layer dissolves, and the IR reflective coating on it crumples as seen in the pic. Whats more it cant be restored.

My point is that the spectrometer can still be made to work. Clean of the coating repolish the beamsplitter. At a 1.52 to 1.0 index difference the beamsplitter surface still refelects 1.51-1/1.52+1^2 of the incident radiation. So a retruned FTIR will still function though with reduced sensitivity.

My first FTIR the owner simply gave up on thinking the beamsplitter was fut, whereas in fact he couldnt tune it. I found that after desiccation for a few weeks, not only did it function well as per diagram, but the throughput is 100% restored Len

cata_beamsplitter.bmp - 460kB

benzylchloride1 - 17-8-2009 at 12:56

Does anyone know where to obtain a pyroelectric detector for a Perkin Elmer 1310 infrared spectrophotometer? I recently bought one of these and the detector failed.

JohnWW - 17-8-2009 at 15:57

Contact Perkin-Elmer directly; find them using Google. They would surely have spare parts for their products.

densest - 14-10-2009 at 12:52

I looked at this thread because I have 3 old UVIS-205 absorbance detectors and some flow cells for them. It accepts a cuvette holder (which I don't have, but could fabricate) and can be persuaded via RS-232 to scan a range of wavelengths and output the absorbance values.

I'm very much more knowledgeable at electronics, computers, and mechanics than chemistry, which I'm learning. It would be fairly easy to come up with some control software which would accept commands and plot output from instruments like this, as long as the control codes were known.

I was intending to use the UVIS-205(s) as old fashioned scanning spectrophotometers to do quantitative analysis of things like gold and (with luck) platinum and palladium in solution. It would be truly wonderful if I could do those in the presence of much larger quantities of copper, nickel, and iron, any of which make visual color inspection impossible.

Does it make sense to try to make one of these units run for that sort of inorganic analysis?

As a tangent, how much interest would there be in trading expertise writing software for analytical instruments for chemical expertise? And is it worth my while to put one or two units up for auction (EBay or LabX, maybe) or are they so old that nobody would be interested?

benzylchloride1 - 23-6-2012 at 18:00

I just bought a Hewlett Packard 5989 quadrapole mass spectrometer. I plan on interfacing it to a HP 5890 series II gas chromatograph that I bought last year. The HP 5989 is a fully functional mass spectrometer that can interface with a GC, HPLC or direct sample injection. At the moment, the system is capable of both electron impact and chemical ionization, I hope to add APCI once I get the instrument up and running. I will have the instrument in a few weeks and will begin to gradually sort out the details. It looks like this is going to take a few years to get going, due to my limited budget being on a graduate student stipend and the expense of mass spectrometers in general. I plan on having the 60 MHz NMR spectrometer I bought last year interfaced with with a computer within a month or two. Does anyone have a source of HP Chemstation that would run on MS DOS or Windows 95 that they could sell to me reasonably cheaply? I cant get away with strip chart recorders for mass spectroscopy.

[Edited on 24-6-2012 by benzylchloride1]