|
To demonstrate we mean business, here is Ed III of Vol II. Production time has been reduced which means that some of the submitted papers have been refereed, altered, edited and published in about 6 weeks. We plan to reduce our lead times still further! The edition is a bit slim (only about 55 pages) because several authors who have promised pieces forgot about their Summer Holidays. So, your Editor had to produce some material – quickly! I hope you find my tutorial on fundamental modes useful. It’s not aimed at the expert but rather at the army of occasional users. In the piece I try to show you what fundamental modes are and how to assign the spectrum you record to specific modes. There are a wide variety of techniques for doing this job that I have not yet discussed. Perhaps the most useful is to use polarised light either in the infrared or the Raman. [An example of its infrared use came up last time]. So I will scribble up another tutorial for you next time. The contributed section has only 3 papers this time, but all are really worth your attention. Peter Chen has an exotic new method of recording spectra of gases based on Raman spectroscopy. It’s a hard technique to master, but do read the paper. I’m sure we will hear a lot more about the technique in the future. NOW – many of you will need more info than Peter provides. This is a WEB JOURNAL, so send your questions by e-mail direct to Peter [BUT PLEASE COPY TO ME AS WELL]. We will then feature any questions and answers, which will then help everyone else. One of the foremost rubber chemistry laboratories used to be called the "Malaysian Rubber Producers Research Association Laboratories" and is situated near Hertford, just north of London. Now it is known as the "Tun Abdul Rajak Laboratories". Kevin Jackson, their resident expert on the use of vibrational spectroscopy has written an excellent paper for us on elastomers and I urge you to have a good look at it. Cees Otto and his colleague Dr Greve have produced for us a superb piece on micro-Raman spectroscopy with particluar application in biology and biochemistry. Just to finish – a few days ago, we received an e-mail telling me that one of my papers had been cited in a recent review and our enquirer was interested in it. Problem was that the citation printed in the review was wrong and he asked for our help. Set me thinking – when that happens in IJVS [would it ever? – Assistant Editor] we can simply amend the error on the server, so that future readers will get the correct citation. Absolutely impossible to do anything about a hardcopy journal. Hmmm! So Good Reading – but PLEASE tell us what you think of
the Journal – be rude or even kind – lets have your ideas and suggestions. We
are getting almost no reaction from you the readers. Remember we cannot improve the
Journal without your input. This Journal covers near infrared – after all n.i.r is a vibrational technique even if it is based on overtones. We URGENTLY need articles on the instrumentation, applications, value, analysis of the data and ideas about future applications. What about on-line applications? Come on you lot – please tell us what you do, have done and aim for in the future. All contributions gratefully received by the whole readership. And remember your
paper will be published in ~ 6 weeks and will be seen by 600+
subscribers guaranteed.
2. The
Interpretation of Vibrational Spectra The Eighth Residential School in the brand new Glamorgan Business Centre at the University of Glamorgan in Wales, will be held on 13-14 April 1999 to be followed by a meeting of the "Infrared and Raman Discussion Group" on Thursday 15 April 1999. The School follows previous residential meetings and includes as main contributors Prof Pat Hendra, Southampton, and Editor of IJVS, Prof. John Van der Maas, Utrecht, Dr Clara Craver, Craver Consultants, and Prof Bill Fateley, Kansas. These leading experts have been involved in research and tuition in the field covering many successful similar courses. The Programme contains lectures and tutorials in interpretation of spectra using the latest accessories with problem solving on an individual basis. Practical work will include hands on use of spectrometers and latest accessories involving equipment and staff from leading instrument firms including Bio-Rad Laboratories Ltd , Bruker UK Limited, Nicolet Analytical and Perkin Elmer Ltd. These Companies are all providing generous sponsorship. These meetings have a tradition of great work and inspiring social activities with visits to areas of outstanding natural beauty. Well worth attending for experts and newcomers alike! Fees for the meeting are £275 for two days 13-14 April (£150 for one day) and £25 for the IRDG meeting on 15 April with 25% reduction for accredited students. Booking form and further details
from wogeorge@glam.ac.uk We are so pleased that the number of subscribers for IJVS is increasing so fast (600+ to date), but are keen to increase it even more. So some new cards have been printed with our new URL promoting IJVS and we are looking to distribute these cards in as many places as possible to spread the word. If any of you want some cards to give to friends or colleagues, then please e-mail me (mailto:louise@ijvs.demon.co.uk )and I'll send them out (a maximum of 20 cards). It will be gratefully appreciated. Thank you to all our contributors for Edition III. Some years ago, I used to teach vibrational spectroscopy to our undergraduates in Chemistry. I found most of them were happy with the idea that molecular vibrations have certain characteristic frequencies when the sample contains contaminated specific chemical groups (so-called group frequencies) but they were very wary of 'fundamental modes'. The group frequency approach is very popular amongst organic chemists and it is based on a good (if somewhat limited) theoretical basis. The problem is that users tend to stop at this level of sophistication and go no further. Eventually they get themselves into terrible trouble. The group frequency approach and some beautiful animated examples are given in IR Tutor - available on CD-ROM from Perkin Elmer. It includes the out-of-phase deformation of toluene, which I always call the " demented spider motion". Get a glimpse of IR Tutor and you'll see why! My solution was to take our students through the vibrations of chloroform and then to assign the Raman spectrum. The project was written up and published in Spectrochimica Acta at the end of one of the series of special editions devoted to F-T Raman spectroscopy and it has been suggested a re-run brought more-up-to-date might be of value to the non-specialist readers of IJVS. [PJ Hendra Spectrochemica Acta A51 (1995) 2205-2208]
Fundamental Modes
If you wished to exactly define the shape and size
of a molecule, where it is placed and its orientation, you could do it by stating a set of
cartesian co-ordinates for each atom’s nucleus referenced to axes in your
instruments' sample area. Let's draw up HCl (I'm too lazy to trot out all this rubbish for
chloroform).
Put it this way - in chloroform, we have 5 atoms so the number of XYZ co-ordinates we have is - 3 x 5 = 15 ---- but 3 simply tell us where the wretched molecule is! Now 3 are 'burned up' telling us its orientation. So we have 15-6 possible ways of defining the vibrations at most. The important things to remember about vibrations is that they have properties which are quite fundamental. They are 1. All the atoms move periodically and in phase 2. the centre of mass remains fixed at all times and 3. the molecule retains its orientation and position throughout the motion. So - returning to dear old HCl we get
Note that the big fat chlorine hardly moves. The sprightly hydrogen belts back and forward like a table tennis ball. This is because the centre of mass must remain fixed. Now Chloroform - the complex vibrational motion is made of a series of simpler fundamental vibrational motions - and there can't be more than nine of them. Draw on a piece of paper five or six structures of
chloroform. Remembering the rules above try to draw the vibrations using vectors - you
won't get them all - try to draw three. To give you the idea I will do the job for you for
H2O
(This is shown in action in IR tutor) Note the direction and the size of the vectors. So have a go at chloroform and see how well you can do. STOP HERE Introduction Samples for quantitative infrared, near infrared or Raman analysis are frequently very inhomogeneous and/or of asymmetric shape. A typical example is the pharmaceutical tablet - frequently inhomogeneous and even of layered construction, of irregular shape and frequently embossed with figures, letters and a logo. A well established procedure is to rotate the sample in the spectrometer sample area and even when appropriate to translate and rotate so that the point of analysis is scrolled over the whole surface of the specimen. However, recently people have been claimimg that sample rotation is not acceptable in F-T instruments, but is this really true? Traditionally, Raman spectroscopists have rotated samples very rapidly (typically 3000 rpm) but the reason for doing so is not averaging but rather the elimination of sample heating. Many years ago, when the laser was introduced as a source for Raman spectroscopy, sample burning appeared as a real problem. If the sample absorbs the laser radiation, the very high brightness of illumination and the typically low thermal conductivity of the sample results in an extremely rapid rise in temperature and in many cases the sample burns! Even when sample absorption is only small (the extinction coefficient is small) some heating always occurs and this has been discussed in IJVS [1]. Avoiding sample burning A problem notoriously associated with FT Raman spectrometers is that the laser heats the sample to an unknown and uncontrollable degree. Hitting the sample with 0.5 W at 1.064 m m for a poor scatterer is not at all unusual. Assuming the illuminated patch at the sample is 300m m in diameter, a fairly typical value, the brightness of illumination exceeds 700 W/cm2. When samples do not classically absorb the radiation, the temperature rises 30 or 40 ° C in typical cases. This usually doesn’t cause any problems unless of course the sample has a low melting point. An example is shown in Figure.1. When pharmaceutical A powder is illuminated with 0.5 W of laser power, the powder melts (TM = 37 ° C) as can be seen in the bottom spectrum.
Figure 1 Spectra of
pharmaceutical A. In the top spectrum, the same powder was hit with 0.5W but now the powder, held in an NMR tube, was rotated. In Figure 2, the same two spectra are displayed only on a different scale in order to show the effect of melting in the stationary spectrum.
Figure 2 Spectra of pharmaceutical A. As can be seen, rotating the sample dissipates the heat and allows one to record spectra of low temperature melting compounds with a considerable amount of laser power. There are many other examples e.g. in catalytic or inorganic chemistry where some absorption of the radiation occurs. In biomedical samples the problem is particularly severe. A good example in inorganic chemistry is NiSO4.xH2O. Figure 3 shows the effect of sample rotation to avoid laser heating. In the bottom spectrum, the sample is stationary and the heating is clearly present. In the midddle and top spectra the heating effect has vanished, the residual broad peaks near 5900 cm-1 absolute wavenumber, are not due to heating but arise from OH vibrations in the water of crystallisation.
Figure 3 Spectra of NiSO4.xH2O.
To avoid these OH vibrations, we tried the same experiment using anhydrous FeCl3 and the effect is shown in Figure 4. Oddly enough the band is more intense when the sample is rotated. This must be due to the bleaching of the fluorescence when the sample is stationary. When the sample is rotated, there is no bleaching, hence the fluorescence increases again (bleaching, or the disappearence of the fluorescence, occurs when a sample is illuminated by the laser for a certain time period).
Figure 4 Spectra of FeCl3.
Fourier Transform Raman But there can be a problem; fast rotation, combined with Fourier transform spectroscopy, whether it be infrared Raman or near infrared, leads to the appearance of spurious lines due to so-called ‘modulation’ and ‘double modulation’ of the scattering (or non-absorbed light in the case of the IR or NIR) of the sample. An example is shown in Figure 5. Where a maleic acid tablet is rotated at ~1600 rpm. In the top spectrum, the tablet is stationary, the bottom spectrum is that of the rotating sample.
Figure 5 Spectra of a tablet, consisting of maleic
acid. Note that the data is not presented in the familiar wavenumber shift, but in absolute wavenumbers. The Nd-YAG laser (n 0 = 9394 cm-1) excites a Raman spectrum between ~10,000 and ~6000 cm-1. Beyond this frequency, the detector ceases to function hence no real bands appear from 6000 to 400 cm-1. Thus, the spurious lines that appear in the MIR are due to modulation, caused by rotation of the sample. The periodic variations in scattering (or absorption) are Fourier transformed by the instrument processor and appear to the instrument like emission bands in the IR. In fact we are using the Fourier process to frequency analyse the signal from the detector. However, when rotational frequencies are kept low, these spurious bands fall well outside the relatively restricted range typical of the Raman spectrum and hence no harm will have been done in rotating the sample [2]. It must be stressed that there was no mechanical coupling between the rotor and spectrometer. Now, there is another problem associated with sample rotation. Assume we are looking at a sample with a single strong Raman or infrared band in its spectrum. This single frequency will generate a cosine function in the interferogram which the processor will then convert into a Raman band. Let’s work through an example: the laser wavelength is ~1 micrometer = 10,000 cm-1 = n 0 assume we have a Raman frequency at D n = 2000 cm-1 , resulting in Raman band (Stokes) at 8000 cm-1 or 1.25 m m wavelength. This Raman band generates a cosine function as the interferometer is scanned, as can be seen in Figure 6.
Figure 6 Interferogram after transformation of a single frequency at ~ 8000 cm-1. Assume, that the instrument scans with 0.1 cm/s, then the audio frequency of this signal will be 0.1cm/s / 1.25*10-4cm or 800 Hz. To put it another way - if you put on headphones and connected the leads across the output leads of the detector preamplifier, you would hear a high pitched whistle as the spectrum is scanned. Now let us rotate the sample at 100Hz or 6000 rpm. Rotation of the specimen will generate a cosine at 100Hz, which in our example will transform to a band in the MIR at 1000 cm-1. In addition, the new frequency will interact with the Raman one, generating weaker cosine functions at 800 ± n*100 Hz. The primary additional bands will be the most important ones, producing an
interferogram like in Figure 7. Figure 7. Effect of sample rotation near the actual Raman frequency. These cosine functions will transform to (see Figure 8):
Figure 8 Effect of sample
rotation near the actual Raman frequency Thus, rotation will inevitably introduce spurious peaks in the Raman
spectrum and hence is not acceptable. Recently, Salzer et al [3] suggested the use of step
scan measurements in order to eliminate the appearance of these spurious bands. But is
this incredibly expensive solution really needed? A couple of years ago, Ventacon Ltd
devised a simple rotator for Nicolet instruments (at the suggestion of Dr. Chris Petty),
able to hold an NMR tube and designed to rotate slowly, with the intention that by so
doing the problems described above would be avoided. The accessory in its Perkin-Elmer
guise is shown in Figure 9.
Figure 9. NMR tube rotator and battery supply Rotation is adjustable in the range 20-50 rpm hence any periodicities in Raman intensities will produce spurious line at less than 1 Hz, which will be very close to the Raman bands themselves and probably cause little or no problem. We show this in Figure10 where the top spectrum is that of crystalline KMnO4, stationary. In the bottom spectrum KMnO4 powder is rotating. As you can see the problem is not severe. Samples like KMnO4 are particularly sensitive to the double modulation problem because they show very intense and sharp bands.
Figure 10 Spectra of a KMnO4 powder. In addition to periodic variations caused by rotation, the crystals themselves will be illuminated by the laser for short periods of varying lengths. Let us look at this a little closer. The laser beam is about 300 m m in diameter. The permanganate powder particles average about 100m m in size. Figure 11 shows the NMR tube, filled with powder, rotating at 40 rpm.
Figure 11 Top view of a rotating NMR tube, Linear speed at radius 2 mm = 2*p *2mm * 40/60s = 8.4 mm/s Thus, the total Raman scattering will rise and fall rather like that
displayed in Figure 12. Figure 12. Periodic variations in the Raman scattering due to particle size. Hence, the crystallites will produce random spurious bands near the Raman frequencies at ± ~28 Hz. As expected, when the sample is both rotated and translated, the actual
spectrum displays more noise than when only rotating (since the overall noise is
proportional to the area under the total spectrum). A picture of the commercially
available NMR tube rotator and translator is shown in Figure 13.
Figure 13. NMR tube rotator and translator. When speeds of rotation and translation are varied continuously the noise level due to the appearance of spurious side bands is similar to that when the sample is just rotating, but we have the advantage of a bigger sample coverage. It must be said that this noise level is still very low, signal to noise being : 280/0.5 ~ 560, fairly typical of the performance of our instrument. Averaging An example of a machine capable of averaging the Raman spectra over a
relatively large sample area is shown in Figure 14.
Figure 14. Tablet analyser Using this device the Raman spectrum of an active in a complete pharmaceutical tablet (complete with logo and identification letters) can be recorded reliably and used to measure the quantity of active present. Conclusions Spinning samples provides an excellent means of avoiding problems of sample overheating. Secondly, it overcomes problems associated with inhomogeneous samples. When rotational speeds are kept low (below ~ 1500 rpm) the rotational peaks stay well away from the actual spectrum the increase in noise in the spectrum is negligible. Both rotating and translating samples in order to increase the sampled area does increase the noise in the actual spectrum a little, but it provides the advantage of a bigger sample coverage. References [1] Y.D. West, IJVS, Vol. 1, Ed. 1 [2] A.T.G. De Paepe, J.M. Dyke, P.J. Hendra and F.W. Langkilde, Spectrochim. Acta A, 53, (1997), 2261 [3] R. Salzer, U. Roland, R. Born and J. Sawatzki, Applied
Spectroscopy, 51, 10, (1997), 1471
|
