Neil, can I ask that we get this discussion back on the topic of Professor Roy, and forget about what other people have or have not said on other forums at some time in the distant past?
I was musing about his slide show last night, and I'd like to critique slide 28 more cogently. I had to study UV-vis spectroscopy quite intensively 30 years ago as part of my biochemistry degree. Since then I have been
using such spectroscopy (well, spectrometry) every day. Now that doesn't qualify me as a great expert any more than saying that I took a course in motor mechanics 30 years ago and have driven a car every day since makes me an expert on the internal combustion engine, but it does mean I have some idea of the basics.
Transmission/absorbance UV-vis spectroscopy (Beer-Lambert Law) refers to the study of the specific wavelengths of light which are absorbed by particular substances in solution. Where the absorbance is in the visible range of the spectrum (400 to 800nm) it is this property that gives a solution its colour. (Reflectance spectroscopy (Kubelka-Munk function), the study of the wavelengths which are reflected, is a different subject and governed by different rules, but it's clear from the term "Abs" on the y-axis of the figure on slide 28 that Roy is using bog-standard transmission/absorbance UV-vis spectroscopy here.)
It's a simple technique. You put a sample of the solution under test in a 1cm cuvette and shine a beam of full-spectrum light through it. You have a detector on the other side which measures what comes through the solution. This detector is a "grating spectrometer" which can look at very precise small windows of wavelengths. You simply scan up (or down) the wavelength spectrum and record what percentage of the incident beam is present in the transmitted beam. Absorbance is the inverse of transmission, and the difference between the incident beam and the transmitted beam at a particular wavelength is the absorbance of the solution at that wavelength (also known as optical density, referred to as A
l or OD
l). With me so far?
Roy's slide 28 is a simple example. The x-axis is the wavelength, running from 200 to 500nm. The y-axis records the standard absorbance units. Although it's been linearised here (a common convention) this scale is logarithmic, with sharply decreasing precision in the higher reaches. Anything with an absorbance of more than 2 units is essentially opaque, and in practice one pays no attention to any reading over about 1 unit. If you get a reading over 1 unit and want more detail then you have to dilute the sample to get the reading lower, then multiply your result by the dilution factor.
In any spectroscopic analysis there are four things between the incident beam and the detector - air, the solid material of the cuvette which contains the test solution, the solvent and the solute. Air has negligible absorbance at the relevant wavelengths. The cuvette is important. Glass is OK if you are only working at visible wavelengths, but it absorbs significantly in the UV range (below 400nm). Therefore quartz cuvettes must be used if any data are required for the UV wavelengths. The solvent is also important, as again one needs something that absorbs minimally at the interesting wavelengths. I've done all my biochemistry work in aqueous solutions (water doesn't absorb much in the relevant wavelengths), but hey, guess what, ethanol is also a useful solvent for organic compounds in this respect, as it "
absorbs very weakly at most wavelengths"! (Sorry for citing Wikipedia, but it was the most comprehensible of the links stating this - it's a common consensus though.)
What does Roy's slide 28 show? Let's look at the black line first. This is variously described as "plain EtOH", "unsuccussed ethanol" and "the original solvent".
Does this look to you like something that is "absorbing very weakly at most wavelengths"? Obviously not! It's absorbing very weakly above 400nm (in the visible range), but in the UV range it's absorbance increases steeply with decreasing wavelength until it's right "off the scale" (that is above the upper limit of sensitivity of UV spectroscopy of 1 to 1.2 absorbance units) by about 250nm.
This page (table about half way down) gives the absorbance of ethanol as 0.05 absorbance units at 240nm (considered to be the lower limit of its usefulness as a solvent for quantitative work). Roy's "ethanol" line has an absorbance of >1.2 absorbance units at 240nm!
Whatever this black line is, it ain't pure ethanol, being measured in the standard way (quartz cuvette with 1cm path length). JJM surmised it might be glass, due to the use of a glass cuvette which absorbs below 400nm. However, I have two problems with this. First, it's second nature to use quartz cuvettes for work below 400nm, and I have difficulty in believing that even Roy would make such an elementary mistake. Second, if we were just looking at the cuvette's absorbance, all the solutions should be the same, and they're not.
Now I would point out that JJM got there before me (I'm just elaborating on what he said), and I'd expect any competent specialist in the area to get there the minute he looked at that graph. What is labelled as an absorbance spectrum of ethanol just isn't so, quite blatantly and grossly. This blows the entire UV spectroscopy right out of the water in one go.
However, it gets curiouser. The three coloured lines on the slide 28 graph are supposed to be three different potencies of Nux Vomica. In fact the three absorbance lines for these are essentially identical. No surprise there. But also, they are absorbing significantly
less than the alleged solvent line.
If Roy had thought at all about what he was actually saying here, he'd realise this is very odd. He's declaring ethanol to have significant UV absorbance in the 200-400nm range (false, as we have seen), and then he's declaring that using the ethanol to prepare a homoeopathic potency
removes a significant amount of this absorbance.
Now absorbance spectra of compounds aren't just arbitrary. The absorbance behaviour is closely related to the chemical composition of the compound. Look at the two articles linked to above. In organic compounds it's all about the possession of conjugated bonds. Expert spectroscopists can deduce a lot about a compound from its UV spectrum, particluarly about what conjugated bonds are present. Any spectroscopist finding an unexpected or anomalous UV spectrum would naturally tend to start speculating about what was causing the readings - what sort of bonds were changing and where. Roy hasn't got to that bit yet though!
Of course at that point we come to a grinding halt, because we'd have to find out what bonds in ethanol were causing this absorbance between 200 and 400nm, and especially the apparent peak with a
lmax of about 320nm, before we could speculate as to what process might be removing this absorbance. And this is where we hit the brick wall, because we discover that ethanol actually
has no such absorbance, except in Professor Roy's graph.
I submit we have no idea what Roy is measuring in that graph, and we can't even absolve him for sure of the elementary sin of having used a glass or plastic cuvette. However, what he labels as "ethanol" simply is nothing like the absorbance spectrum of real ethanol, which leaves the entire thing in a heap on the floor.
Rolfe.
, sure right. I guess I'm a bit naive. What's the point of science if you have no taste for the truth, I aks myself ? The thing is that it would be exotic and interesting if there was something in it. I guess that's why real scientists, even some working in medicinal research (I've met such people), get hooked to it.