All infrared users should be familiar with their instruments’ background. F.T.I.R.s are, in the traditional parlance, single beam instruments; the spectrum of the source unattenuated by a sample is scanned and then compared later with that of the source plus sample. The point by point ratio of the two spectra being the absorption spectrum of the sample alone.

The software we all use frequently memorises the background and then does the ratioing process automatically, so many people are almost oblivious of the quality or otherwise of the background. However, the details of the background have a profound effect on the quality of the spectra we record, particularly if we are making quantitative measurements or are trying to measure very weak absorptions. The most persistent problem is water absorption.

If a ratioing process is to be meaningful, the instrument background must be invariant, it must not change to any significant degree between its acquisition and recording the spectrum of the sample, otherwise any differences will appear as positive or negative spurious bands.

Most instruments, and particularly the lower cost routine ones, have sealed interferometers and sources and the sealed volume is desiccated. The sample area and the detector compartment are usually left open to the atmosphere. In these cases, the idea is that most of the optical path is relatively unattenuated by water vapour, whilst the bit that goes through the sample area and on to the detector is heavily absorbed but to a constant level (set by the level of water vapour in the lab). The problem with this approach is that, particularly at the higher resolutions (2 or 1cm-1 or better*), the attenuation can be very severe even in air conditioned labs and hence if the bands of interest lie within the envelope of the water absorption, we have severe problems with quantitative measurements.

*The apparent band intensities increase as the resolution is improved. For details see Edition III, where we will publish an article by W. F. Maddams.

To solve these problems, all instruments are provided with facilities to purge with dry gas (dried air or liquid nitrogen blow-off are both popular). The dry gas is normally passed through both the interferometer box and the sample and detector areas. We do this at Southampton using air dried from a commercial recycling molecular sieve system. Assuming the instrument is purged overnight, we would then expect to see an almost perfectly water-free background in the morning. Opening the sample area to introduce a specimen will, of course, contaminate the optical path with laboratory and hence wet air, but we are assured by the instrument makers that purging for a few minutes will ‘sweep away’ the water vapour and restore perfection to the background.

There are two snags with this approach:

  1. How dry is the gas supply, how effective is the purging process and how good a background should you expect after prolonged purging? How ‘good’ is a ‘good’ background?
  2. How long do you have to purge to restore the background after opening the sample area? Clearly, if the restoration is incomplete, running a spectrum MUST lead to spurious bands in the finished printout after ratioing.

We have suspected for some time that the air we use here at Southampton to purge our instruments is not very dry but we have no adequate method of measuring the water level. What WE regard as a “satisfactory” background after overnight drying may well be poor. To check, we have contacted other laboratories but found some would accept our background and others would not, hence there is no real standard to which we can all aim. I therefore asked our hard-pressed Editorial Advisory Board for their help. I asked them to dry their instruments overnight using their normal routine and then run backgrounds at 1, 2 and 4cm-1 resolution, before opening the instrument sample area.

The outcome is very interesting. We had several responses and the range of level of water vapour absorption was vast. Presumably people are happy with their arrangements, so the obvious question to ask is why? Presumably because they know no better!

The instrument at DSM Research gave the most outstanding results. If purged overnight, it shows no sign of bands on a full range display at 4cm-1 and about 2% at worst at 2cm-1 resolution. At 1cm-1 the water vapour bands can be discerned and peak at around 3%.

Our instruments at Southampton (and we have several operating off the same supply) show peak absorptions near 1600cm-1 of at least 30% at 1cm-1 resolution on a good day! Thus, our air supply must be rubbish.

If your gas supply is really dry (and DSM use nitrogen blow-off through a round-the-building manifold) you should see almost no absorption at 2cm-1. If you do, your gas supply is wet. Your water level may only be a few p.p.m. but clearly you can do better.

The folks at the Malaysian Rubber Producers Research Association produce a background inferior to the D.S.M. setup, so their gas (in this case nitrogen) is not quite as dry as it could be. At 1cm-1 resolution, purged overnight at 1.5l/min gives a strongest water vapour band of around 13%. At 4cm-1 resolution and the same flushing rate, the lid was opened and a sample introduced. The CO2 absorption near 2300cm-1 once the lid was closed was ~35% and the strongest water line around 1600cm-1 about the same. Backgrounds were then recorded after 5,10,15,20,40,50 and 60 mins flushing. The last background shows the CO2 level down to about 2%, but the water is more persistent at nearer 10%. The really important point here is – the intensity of the bands changes significantly between 20 and 40 mins purging but not thereafter.

The lesson is clear; once the sample compartment is opened you must purge for a long time if you are to regenerate your background quality. If you do not, you will contaminate your spectra.

My account above is only an opener; in the next edition, as promised, we will run an article by Bill Maddams looking at the problem in more detail. There are a number of experimental ruses for minimising the problem and we will cover these too in the next edition.

Thanks are due to Dr. Sjaak Bremmers at D.S.M. in Geleen, Holland and to Dr. Kevin Jackson of the Malaysian Rubber Producers Research Association at Breckendonbury in the U.K.

REF: Int. J. Vib. Spect., [] 1, 2, 3 (1996)