Nobel Lecture, December 11, 1923

Quantitative Micro-Analysis of Organic Substances

In accepting the honour of addressing this illustrious assembly on this proud occasion today, I am complying with § 9 of the Statutes of the Nobel Foundation, and I would like to take the opportunity at the same time of expressing my humble thanks to the Swedish Academy of Sciences for awarding me the Nobel Prize for Chemistry for the year 1923.

Already two years ago I had the honour of delivering two lectures to experts in the Kemiska Sällskapet (Chemical Society) on the subject of the quantitative organic micro-analysis developed by me. The nature of my task today does not permit me to present all the details described and demonstrated at that time, neither does the time allocated allow me to execute two important experiments; I can therefore only give a survey of the origin, the development and the final results up to date of quantitative organic microanalysis.

I had felt an inner compulsion to carry out the analysis of organic substances with much smaller quantities than had hitherto been used at a time when I was working on an investigation of a substance which could be obtained in extraordinarily small quantities; I was therefore forced to decide whether to process tons of material or whether to search for new methods, which would enable me to arrive at correct analytical results while using much smaller quantities then previously. I decided on the latter course of action. The most obvious plan was to apply reductions to all and everything, namely to move the decimal point back one or two places, not only in relation to the quantity of the substance, but also in relation to all the apparatus, reagent quantities, etc., involved.

This resulted not only in the necessity of carrying out all weighings with far greater accuracy than the customary analytical balances permitted, but the balance also had to be constructed in such a manner that the diversely shaped devices that were to be weighed could be accommodated on them. The assay balance for precious metals by Wilhelm Kuhlmann in Hamburg, which at that time was already being used in other laboratories for weighing with accuracy of 0.01-0.02 mg, fulfilled these requirements to a high degree. The designer, most fortunately, carried out great improvements to it at my suggestion, so that it is now possible to weigh with an accuracy of ±0.001 mg over a range of 20 grams using the micro-chemical balance designed by him and now in common use.

The assumption that our aim could be achieved by shifting the decimal point for all factors proved to be entirely misleading; for, following the reduction of the quantity of the substance, effects appeared which decisively influenced the result of the micro-analysis, whereas they had been insignificant in macro-analysis and had therefore often not even been noticed.

This can already be very simply illustrated in relation to nitrogen determination according to Dumas. The volumetric determination of the nitrogen set free during this process is influenced to such an extent by the 50% potassium hydroxide solution adhering to the interior of the measuring tube of the micro-nitrometer, that 2% must be deducted from the volume reading. Requirements regarding the carbon dioxide used for deaeration and for expelling the nitrogen are very much more exacting than one had hitherto been accustomed, and it was found necessary to develop a carbon dioxide supply which was completely free from air. Up to now there has been no lack of work in this direction and it is peculiar that nearly always the massive, compact marble was suspect as carrier of air; it was even found necessary to boil it off in a vacuum, and the fact was overlooked that aqueous solutions like hydrochloric acid are capable of absorbing considerable quantities of air, and also of nitrogen. Further, it was found to be imperative that the alkali sealing off the volume of nitrogen that had been set free, should not foam; treatment with barium hydroxide was found to be the means by which the substances that caused the foaming could be extracted from the alkaline solution. Finally, the position of the reduced copper mass in the tube lining was found to be of importance and the most correct position proved to be in the middle of the copper oxide lining in the hottest part of the tube.

The development of the determination of carbon and hydrogen was considerably more difficult, because the problem here was firstly to absorb the resulting combustion products free from foreign admixtures and, secondly, in such a manner that the increase in weight could only be related to the resulting water and the carbon dioxide which had been produced. Thus, for instance, even the use of clean air and pure oxygen can result in incorrect weight increase, if the gases are conducted through new rubber tubes, because these release noticeable amounts of vapours containing carbon and hydrogen into the stream of gas. Artificial ageing, for instance by passing steam through these tubes for one hour, can eliminate this undesirable property.

If the substance which is to be analysed contains nitrogen, halogen, or sulphur, gaseous products may result during combustion, which might erroneously be weighed and recorded as carbon dioxide, because they are also retained by the absorbing agent for carbon dioxide, i.e. soda lime. In such cases, therefore, entry of such products into the absorption apparatus must be prevented at all costs. Whereas for macro-analysis different tube linings are recommended for the various substances in accordance with their composition, my main aim was to develop a tube lining which would always retain everything that was not carbon dioxide or water. I therefore call it the universal lining. It consists of a mixture of copper oxide and lead chromate between two layers of silver and finally a layer of lead peroxide on asbestos heated to 180°C.

Since the invention of Liebig’s potassium apparatus and the calcium chloride tube, many variations have been made in the course of nearly a century to the usual absorption apparatus for elemental analysis. I chose for this purpose the simplest form, a tube. It is quite obvious that the slightest impurities must cause special errors. Warming caused by wiping, wrongly makes these small devices appear to be lighter in weight; then, when they have not been filled, they attain a constant weight within 10-15 minutes. However, if they contain an absorbing agent and if their connections are open at both ends and are not tapered, they will increase in weight at the same rate after that period has elapsed, that is to say, they absorb from the air water vapour and therefore become heavier. I have finally succeeded in completely preventing this diffusing-in of water vapour by fitting capillary tapering pieces between larger and smaller expansions of the flow line; this arrangement represents such a gradual diffusion gradient that even for the sensitive weighing instrument of the micro-analyst unnoticeable weight increases during the periods under practical consideration will result.

On the other hand, these capillary taperings present considerable resistance to the gas stream resulting in an increase of pressure in the interior of the combustion tube. New difficulties arose hereby; the tube connections between the apparatuses, particularly between the warm outlet of the combustion tube and the calcium chloride tube – the most critical spot – could cause the loss of combustion products in consequence of the increased pressure, if the tube connection is only slightly damaged, and especially in view of the fact that rubber has the peculiarity of allowing carbon dioxide to diffuse easily. I eliminated these hazards by equalizing the pressure at the critical place to that outside, i.e. the atmospheric pressure. This effect can easily be achieved by means of a Mariotte’s flask.

Right from the start I had rejected the idea of forming the connections between the apparatuses and the combustion tube by ground glass joints and of using cocks for closing the absorption apparatus, because these cannot safely and easily be wiped and would also be in danger of being broken on account of their fragility. On the contrary, I am convinced that microanalysis, its development and wider use will be best served by avoiding cocks and ground joints.

When all these conditions are observed, it is quite easy to obtain accurate analysis results for 2-4 mg of the substance. The smallest quantity ever used was 1 mg and the divergence was still within the usual permissible margin of error. In doing this we have quite exceptionally reduced the amount of substance required for the analysis. Bearing in mind that Liebig used 0.5 g of substance and originally even 1 g, and that towards the end of the last century 0.15-0.2 g was usually considered to be the quantity required, we may well say that the development of micro-analysis has reduced the quantities required for the determination to about one hundredth part of what ten years ago was considered to be the amount required. This reduction does not impair the accuracy in any way and I can even say that my students demand far greater accuracy from micro-analysis than they previously expected from macro-analysis. This probably follows from the fact that the basic conditions are far better and can be more completely allowed for than was possible with macro-analysis. There are two further advantages that should be borne in mind. Firstly, there are great savings in gas and reagents, and secondly, and this is even more valuable, there is a saving in time. Thus it is possible to attain the same result in at least one third of the time required for macro-analysis; a decision then to repeat micro-analyses is much easier made, as the work is much shorter and more pleasant.

For the determination of halogens and sulphur Carius’ method was originally generally used. The precipitates obtained then used to be sucked out in a small Gooch crucible, I particularly would like to mention here briefly that I was the first to recognize that the presence of barium chloride was essential for the decomposition of sulphur-containing substances with nitric acid; without this, losses were certain to result. Later a completely new analytical process for the decomposition of organic substances was discovered, which consisted in burning them in an oxygen stream and then leading the combustion products over red-hot platinum in order to complete the decomposition. With halogen determinations, these combustion products are collected in an alkaline sulphite solution, whereas the products from sulphur determinations are collected in perhydrol, with which porcelain beads are wetted. All this is done in one single hard-glass tube and this may be used many hundreds of times for the same purpose. All the halogen is obtained with the washing liquid in the form of sodium halide, all the sulphur in the form of sulphuric acid. Here also there were initial difficulties with the determination of the halogen, because the values obtained after acidifying the alkaline sulphite solution with nitric acid were too high. The explanation was that the oxidation of sulphite to sulphuric acid does not proceed directly, but by the precipitation of elementary sulphur, which would indicate that this reaction does not proceed monomolecularly, but probably by way of a polythionic acid. For, if we first use perhydrol as oxidizing agent in alkaline solution and then acidify with nitric acid, sulphur is not precipitated and fully correct results are obtained. The resulting silver halides are automatically sucked on to a filter tube and weighed after drying. Similar analytical instructions resulted with sulphur determinations; if these instructions are observed, correctness of the analytical results can be guaranteed.

My assistant, lecturer Dr. Hans Lieb has developed methods for the determination of phosphorus and arsenic in organic substances, by which satisfactory results are achieved.

Apart from the quantitative determination of the elements making up the structure of organic substances, the quantitative content of certain groups of atoms is also of the utmost importance. I have therefore developed the determination of the carboxyl group by acidimetry by using phenolphthalein as indicator; I have further elaborated the determination of the methoxy and methylimide groups. I do not want to go into details, but would nevertheless stress the enormous simplification of the micro-apparatus as compared with the macro-procedure in connection with the methoxy determination; the weighing of the substance in tin-foil caps is new in principle and of great advantage, as the hydriodic acid does not bump because of its tin iodide content. For the methylimide determination, the fact that the ratio of the hydriodic acid to the substance quantity used (approx. 3 mg) would be far more favourable than for the macro-analytical process, had not been foreseen. This was for the first time noticeable in connection with physostigmine, as only by micro-analytical methods a third methyl or ethyl group was detected, where previously only two had always been assumed, so that the structure of this body had to be revised for this reason.

Determination of the molecular weight by the boiling-point method has also become practicable through micro-analysis, as we are able to obtain reliable values for a quantity of 7 mg.

It is hard to believe that the simplest of all determinations, i.e. the determination of the residue through fuming with sulphuric acid, as is necessary with the determination of Na, K, Ca, etc. in organic substances, could only be adapted in such a way as not to make too great demands on the skill of the analyst and yet give satisfactory results that beginners always think they are mere accidents, long after the previously described difficult methods had been developed.

Finally it might not be superfluous to raise the question what positive results micro-analysis has achieved so far.

In this connection we may mention that one of the first results was the discovery that the yellow colouring matter in corpora lutea of cattle ovaries was by micro-analytical carbon hydrogen determination found to be carotin, the substance which is in carrots widely found in the vegetable kingdom. This was extracted from approximately 10,000 ovaries in the Willstätter Laboratory in Zurich and would have been sufficient for only one macro-analysis. Thus, in this case already micro-analysis has made possible important progress in our knowledge, by proving that the same substance appears as colouring matter in both natural kingdoms, the vegetable as well as the animal kingdom.

Also, in view of the minute quantity of material available, I want to draw your attention to the valuable information on molecular structure obtained by means of micro-analysis, through the work of Adolf Windaus on cymarin and cholesterol, and Wieland‘s work on the decomposition of bile acids, where micro-analysis proved its value.

Last but not least I want to mention that quite recently a Swede, C.A. Yllner applied micro-analytical methods successfully at my Institute in Graz for the identification of alkaloids in forensic cases, that is, when these had to be extracted from dissected corpses. During the last few years my methods have also been applied in technology for electrolytic metal analysis; and for the analysis of wine very clever micro-analytical methods have been developed, apart from those discussed here.

I therefore am very hopeful that in the future quantitative organic microanalysis will find many more fields of application and expansion, and that it will make possible much insight and discernment which otherwise might not have been possible, especially following the magnificent recognition that has been awarded to it by the Swedish Academy of Sciences.