THE CHEMISTRY OF THE HIGHER OXIDES OF NITROGEN AS RELATED TO THE MANUFACTURE, STORAGE AND ADMINISTRATION OF NITROUS OXIDE

Brit. J. Anaesth. (1967), 39, 345

THE CHEMISTRY OF THE HIGHER OXIDES OF NITROGEN AS RELATED TO THE MANUFACTURE, STORAGE AND ADMINISTRATION OF NITROUS OXIDE BY

A. T. AUSTIN

School of Chemistry, University of. Leeds, England SUMMARY

PRODUCTION OF NITROUS OXTDE

Nitrous oxide for anaesthetic purposes is commercially produced by heating ammonium nitrate (m.p. 169.6°C) to 245-270°. The reaction is usually expressed by the following equation:

(O2N.NH2) which is assumed to be formed by interaction of the nitronium cation (NO3+) with ammonia (Barclay, Crewe and Smith, 1963; Friedman and Bigeleisen, 1950; Wood and Wise, 1955). H+ HO.NO3 >H 3 O.N0 s (nitric acid) (nitricacidium ion)

However, the reaction is not just a simple decomposition. In the highly ionized liquid melt, reversible proton transfer between ammonium cations and nitrate anions occurs. ,+ + ONO2 - ;

3

NH, >O3N.NH3 (nitramide)

+ HO.NO 2

Part of the ammonia and a lesser equivalent of nitric acid formed in this way escape into the gas phase where recombination occurs and ammonium nitrate condenses as a sublimate (or a concentrated solution) in the cooler parts of the system. The ammonia and nitric acid remaining in the now slightly acidic melt undergo oxidativereductive interaction to form chemical species which initiate a series of reactions leading mainly to nitrous oxide with relatively small amounts of nitric oxide (NO), nitrogen dioxide (NO2) and nitrogen (N5). No systematic mechanistic study of the formation of nitrous oxide in molten ammonium nitrate has been made. However, it is widely held that the gas arises from decomposition of nitramide

2

(nitronium ion)

The evidence, however, is inconclusive. An alternative or concurrent mechanism postulates oxidation of ammonia to hydroxylamine (HO.NH,) by the nitric acid with reduction to nitrous acid (HO.NO). This acid gives rise to dinitrogen trioxide (ON.NOa) which reacts, via established routes, (i) with hydroxylamine to give nitrous oxide, (ii) with ammonia to give nitrogen, or (iii) simply decomposes into nitric oxide and nitrogen dioxide. 2HO.NO;=iON.NO 1 + H3O (i) NO (nitrosohydroxylamine)

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

The chemical basis for commercial production and purification of nitrous oxide is outlined and discussed. Nitric oxide is the most likely toxic contaminant but the concentration in the effluent gas from a cylinder diminishes rapidly due to fractionation. Nitric oxide reacts with oxygen but the velocity of the reaction is influenced by concentration and is very slow at concentrations below about 0.1 per cent. Most methods of detection and estimation of nitric oxide depend upon prior oxidation. The reaction between nitrogen dioxide and water is complex and toxic effects result from the formation of hydrogen ions, nitric oxide, nitrate and nitrite ions.

346 (ii)

BRITISH JOURNAL OF ANAESTHESIA HJ

(nitrosoamine) (iii)

REMOVAL OF CONTAMINANTS FROM NITROUS OXIDE

The scrubbed gas is compressed, dried and liquefied. Crude fractionation then allows nitrogen and NO to "blow off". The gas is further compressed, again passed through a drier, after which it is again liquefied and run into evacuated cylinders. In another purification process (see Appendix II) the gas is scrubbed with alkali and then passed through beds of finely divided moist iron, which "removes" NO by converting it into nitrous oxide. An acid wash is then followed by a water wash after which the gas is dried using lump caustic soda. The gas is compressed and further "scavenged" by passing successively through another tower packed with finely divided iron and two towers packed with lump caustic soda. The compressed gas is then liquefied by cooling. Although thermal decomposition of pure molten ammonium nitrate occurs smoothly up to 260-270 °C, the decomposition becomes explosively vigorous above 290 °C when greatly increased proportions of NO, NO 3 and N 3 are formed with the N 3 O. Strict precautions are, therefore, normally taken during manufacture to guard against overheating. The purification procedure should, however, be capable of dealing with occasional surges of contaminated gas. Impurities other than those cited above are unlikely to be present in the gas. Chlorine would

COMPOSITION OF EFFLUENT GAS FROM CYLINDERS CONTAINING LIQUID NITROUS OXIDE CONTAMINATED WITH NO AND NO 3

Any NO that passes through the purification train will, on reaching the cylinder, be dissolved under pressure in liquid nitrous oxide. The use for anaesthetic purposes of a cylinder contaminated in this way can be extremely dangerous. Owing to the greater volatility of NO compared with N 3 O, the first gas to leave a cylinder contaminated with NO will be much richer in NO than the liquid in the cylinder. Based on extrapolated values of the vapour pressures of NO (b.p. -151.6°C) and N 3 O (b.p. -88.5°C) which show that NO is approximately 30 times more volatile than N 3 O at 20°, calculations indicate that a 0.1 per cent (1 g/kg) solution of NO in liquid NSO will give off initially a gas containing 3 per cent NO. The composition of the evolved gas and the composition of the liquid in the cylinder will vary continuously during use of the cylinder and the changes can be calculated. The escaping tendency of a substance from a liquid is directly proportional to the partial vapour pressure of that substance. Now, if A and B be the number of moles of NO and N.O in the liquid phase in the cylinder and a differential amount of the liquid is vaporized, then the

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

Current practice for removing NO, NO 3 and N 2 from nitrous oxide varies. One process (see Appendix I) combines a series of gas washings with a crude fractionation of the liquefied gas. The raw gas is successively scrubbed, (i) with water, which removes residual ammonia and nitric acid, and most of the NO,; (ii) with alkaline permanganate, which removes the remainder of the NO a and NO; (iii) with acid which completes the removal of ammonia and any entrained alkali.

be removed by the alkaline wash but is not likely to be present given the ready availability of high grade synthetic ammonium nitrate and the recognition that chlorine-free water should be used for making the hot concentrated aqueous solution if this is the form in which the ammonium nitrate is handled. Carbon monoxide (CO) should be absent. But its presence has caused concern and recently it was incriminated as the likely cause of a fatality following nitrous oxide anaesthesia (Simona, 1959). Its presence could arise by foreign matter of an organic origin being introduced into the reactor where it would be oxidized by the free nitric acid in the hot melt to carbon monoxide and carbon dioxide. Such contamination could originate, for example, in cellulose fibres from paper sacks, where this mode of handling ammonium nitrate is employed.

THE CHEMISTRY OF THE HIGHER OXIDES OF NITROGEN decrease in the amount of A, ( - dA), and the decrease in the amount of B, ( - dB), are related: -dA p£o

347

mole. These calculations are summarized in table I. TABLE I

Where p. v0 is the partial pressure of NO and Psio that of N , 0 in the liquid.

Residual NO in "400 gallon" (1800 /.) cylinder of N,O originally contaminated with 0.1 per cent (1 g/kg) NO. Original state A

Assuming Raoult's Law holds, A Puo

— *si

(A+B)

and

B =

*NiO

(A+B)

-dA P*o A -dB PB_ B where P N0 and P y i 0 are the saturated vapour pressures of NO and N,O at the same temperatures. From extrapolated data, = 30 at 20°C Therefore, -dA

-dA

-dB = 30 B"

On integration and rearrangement, log A , - l o g A, log BL - log B3

=3Q

where A^ is the number of moles of NO in the liquid phase at the beginning of a period of gas take-off, A, the number of moles at the end; B t is the number of moles of N , 0 at the beginning and B. that at the end. From this expression it can be calculated that a 0.1 per cent (1 g/kg) contamination by NO in a "400 gallon" cylinder of N , 0 is reduced to "half value" after the evolution of 40 litres of gas, representing 2 per cent use. The concentration of NO in the effluent gas would have fallen from 3 to 1.5 per cent and would reach 0.75 per cent after the evolution of another 40 litres. Similar calculations show that 90 per cent of the NO is removed after 130 litres of gas have been taken from the cylinder, representing only 7 per cent use and 99 per cent will have been removed after 250 litres of gas have been taken off. When the cylinder is "half empty" the residual NO amounts to no more than 7 x l O - 1 0

D

Vol. of gas taken from cylinder (1.)

0

40

130

250

900

Percentage of cylinder contents used

0

2

7

14

50

Percentage of original contamination still present 100

50

10

1

Percentage of NO in effluent gas

3

1.5

0.3

0.03

10-10

<10-1J

It is seen how practically all NO contamination is released with the gas from the initial take-off. The risk of inhaling any NO contaminant is, therefore, maximal during the very early stages of a cylinder's use. A cylinder that is "quarter used" would give off gas virtually free of NO. In this context it is to be noted that some manufacturers of nitrous oxide for anaesthetic use deliberately "overfill" the cylinders and then allow 15 per cent by weight to "blow off" (Appendix II). Conversely, it can be shown how the relative involatility of NO 3 (b.p. 21 °C) would cause this compound to concentrate in the cylinder so that only gas issuing from a very nearly empty cylinder would be appreciably contaminated. It should be noted that, for both NO/N 3 O and N 0 2 / N , 0 mixtures, the fractionation is accentuated on lowering the temperature which happens during anaesthetic use. Analytical procedures employed during the manufacture of nitrous oxide should, therefore, be adapted to take gaseous as well as liquid samples and be capable of detecting NO as well as NO 2 . Any method that relied only on liquid sampling or which was unsuited for detecting NO would leave large percentages of NO undetected. Medical practice in the U.K. treats each

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

Therefore,

B

348

BRITISH JOURNAL OF ANAESTHESIA

cylinder as containing pure nitrous oxide and the cylinders are run from the initial opening until empty. It is said that in former times, anaesthetists would "blow off the brown gas" before starting to use a new cylinder of nitrous oxide. This could relate to contamination with nitric oxide which would fall to low values after venting, say, 150 litres. The brown cloud would be nitrogen dioxide formed by aerial oxidation of the vented nitric oxide. CHEMICAL PROPERTIES OF NITRIC OXIDE AND PEROXIDE"

Nitrous oxide is inert under conditions obtaining in anaesthesia, and chemical changes are confined to the contaminants. Nitric oxide (NO), b.p. -151.6°C, is a colourless gas, density 1.04 (relative to air) and only sparingly soluble in water or alkali with which there is no reaction at room temperature and pressure. It readily forms complexes with many metals and salts, the reversible reaction with ferrous sulphate to give the dark brown "Fe(N0)S0 4 " being well known as a method for detecting nitrites and nitrates. At ordinary temperature nitric oxide combines with molecular oxygen to form a red-brown gas (nitrogen dioxide —NO2) at a rate dependent on the concentration of oxygen and the square of the concentration of nitric oxide. At very low concentrations the oxidation is slow, e.g. at 10 parts per million, 7 hours are required for 50 per cent oxidation. At 10,000 parts per million (1 per cent), however, 50 per cent oxidation is achieved in approximately 24 seconds (Magill, 1956; Rabson, Quilliam and Goldblatt, 1960). Further values are listed in table II. TABLE II

Oxidation rate oj nitric oxide, NO, in air (20 per cent O3) at 2O°C. Oxidation time

Concentrations lccniTBiions (p.pjn.) 10,000 (1%) 1,000 (0.1%) 100 10 1

24 sec

1.4 min

4 min

14 min

40 min

2.3 hours

This is the basis of one method of estimation of NO (Kain et al., 1967). Nitrogen dioxide (NOt), b.p. 21°C, is a redbrown gas which exhibits great reactivity. It combines with itself to form a colourless dimer, dinitrogen tetroxide. via an easily reversible equilibrium ("nitrogen peroxide" is the name commonly used for this gaseous equilibrium mixture). In a mixture of NO 2 and N 2 O 4 at 25 °C the monomeric NO, is present to the extent of 25 per cent and at 37 °C NOj is present to the extent of 30 per cent. However, the proportion of monomer increases as the partial pressure is reduced, as for example when diluted by other gases (such as N 2 O). On cooling the gas condenses to an orange-red liquid, b.p. 21.15 °C. This colour becomes less intense as the liquid is further cooled when a colourless solid, m.p. 11.2°C, is formed. The rapidity of interconversion from monomer to dimer frequently makes it extremely difficult to ascertain which is the reacting species. Nitrogen dioxide also reacts readily with nitric oxide to give dinitrogen trioxide in an easily reversible reaction, NOs + N O ^ O N . N O ,

25 per cent 50 per cent 90 per cent 8.4 sec

5NO + 2KMnO4 + 3H,SO4 >5NO, + K,SO4 + 2MnSO, + 3H,O.

3.6 min 36 min 6 hours

7 hours 63 hours

24 hours 72 hours 648

hours

(ii)

At 27 °C dinitrogen trioxide is present to the extent of 3 per cent. It condenses to a royal blue liquid. Both dinitrogen tetroxide and dinitrogen trioxide show extraordinary diversity of chemical behaviour. The latter is fairly readily soluble in water to give nitrous acid.

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

"NITROGEN

Oxidation of nitric oxide by potassium permanganate involves the intermediate formation of NO,. This is easily retained and further oxidized in an alkaline solution and the overall oxidation by alkaline potassium permanganate can be expressed: NO + KMnO 4 >KNO3 + MnO, In acid solution the NO 2 is not so easily retained and it may be swept out of solution by a carrier gas before full oxidation to nitrate is effected.

THE CHEMISTRY OF THE HIGHER OXIDES OF NITROGEN ON.NO, + H , O ^ 2 H O . N O (iii) Dinitrogen tetroxide dissolves (slightly less readily) to give nitrous and nitric acid in equal amounts. O2N.NO2

.NO + HO.NO,.. .(iv)

3HO.NO;=i2NO + H3O + HO.NO, As these reactions are reversible, the interaction between NliO< and water may be represented by the following equation for the overall change: 3O2N.NO2

2NO + 4HO.NO,

These reactions come to equilibrium at room temperature when measurable quantities of the reactants are present (Yost and Russell, 1946). It will be seen that, in the lungs of a patient, nitric oxide may be formed from nitrogen dioxide as well as nitrogen dioxide from nitric oxide. These interconversions must complicate any attempt to study the separate pharmacological effects of the two gases. Dinitrogen tetroxide and dinitrogen trioxide are each powerful nitrosating entities, replacing amino and hydroxylic hydrogens by the nitroso group (—N = O) with alacrity. The N— and O— nitroso-compounds are themselves usually highly reactive compounds and rapidly undergo further changes (Austin, 1961). In a system containing the species NO15 N 2 O 4 , N 2 O 3 , NO, HO.NO, ONO 2 -, ONO-, H+, H2O and O 2 , the problem of isolating and defining particular reactions is prodigiously difficult. But the task may be simplified by focusing attention on patterns of behaviour that characterize such systems, viz. (i) nitrosation (by ON.NO,, O.N.NO 2 and possibly NO);

(ii) the slow "regeneration" of NO from the decomposition of nitrous acid; (iii) changes in acidity attendant on the formation of nitric acid (and to a less extent, nitrous acid); (iv) the oxidative power of O2N.NOj and ON.NO 2 ; (v) the release of the biologically active nitrite ion; (vi) reaction with haemoglobin (Toothill, 1967). ADDENDUM

The sharp fractionation recorded in table I indicates the behaviour that could be expected of a cylinder contaminated with NO only. Should NOS also be present—signifying a more extensive breakdown in the purification of the nitrous oxide —a "holding back" effect on the NO would be expected by virtue of compound formation, N 2 O, (p. 348, eqn. ii). The release of NO would then be less sharp and extend over a greater period of cylinder use. For the same concentration of NO to be attained in the effluent gas, the combined NO and NO2 contamination would have to be many times greater than the 0.1 per cent stated for NO alone in table I. Conversely, the presence of NO would "lift" NO2 which would then appear in the effluent gas earlier than in the absence of NO. Any imbalance in favour of NO of the 1:1 molar proportion necessary for formation of N 2 O 3 would result in a quick initial "blow off' of NO followed by a slower evolution of this gas. The danger inherent in setting limits for the combined concentration of NO and NO2 in any specification of purity is therefore apparent. The NO and NO2 concentrations should be stated separately and, for each gas, there should be a set limit. This applies particularly to NO should it be in molar excess of NO 2 . REFERENCES

Austin, A. T. (1961). Nitrosation in organic chemistry. Set. Progr., 47, 619. Barclay, K. S., Crewe, J. M., and Smith, E. J. (1963). Order of reaction in the thermal decomposition of ammonium nitrate. Nature, 1S8, 1054. Friedman, L., and Bigeleisen, J. (1950). Oxygen and nitrogen isotope effects in the decomposition of ammonium nitrate. J. chem. Phys., 18, 1325. Kain, M. L., Commins, B. T., Dixon-Lewis, G., and Nunn, J. F. (1967). Detection and determination of high:r oxides of nitrogen. Brit. J. Anaesth., 39, 425.

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

Nitrous acid is a fairly weak acid (KA = 6X 10 ~4) and is unstable in the un-ionized fonn. It is, however, extensively ionized in very dilute solution and together with the fully-ionized nitric acid accounts for the acidity of their aqueous solutions. Nitrous acid decomposes via the formation of dinitrogen trioxide (iii) into nitric oxide and nitrogen dioxide (ii) which dimerizes to dinitrogen tetroxide (i) which undergoes hydration (iv). The overall decomposition of nitrous acid may thus be represented:

349

350

BRITISH JOURNAL OF ANAESTHESIA

ETUDE CHIMIQUE DES OXYDES AZOTIQUES FORTS DU POINT DE VUE DE LA FABRICATION, DU STOCKAGE ET DE VADMINISTRATION DU PROTOXYDE D'AZOTE SOMMAIRE

On esquisse ct discute les bases chimiques president a la fabrication et a la purification du protoxyde d'azote. L'oxyde azotique s'avere Stre l'agent contaminant le plus vraisemblablement toxique, mais sa concentration dans le gaz provenant d'un cylindre diminue rapide-

ment en raison du fractionnement. L'oxyde azotique entre en reaction avec l'oxygene, mais la rapidite de la reaction est influencee par la concentration et se presente comme etant tres basse a des concentrations inferieures a 0,1 pour-cent environ. La plupart des precedes de detection et de dosage de l'oxyde azotique sont bases sur une oxydation prealable. La reaction entre le peroxyde d'azote et l'eau est complexe et les eflFets toxiques r&ultent de la formation d'ions hydrogene, d'oxyde nitrique et d'ions nitrate et nitrite. DIE CHEMIE DER HOHEREN STICKSTOFFOXYDE IN BEZIEHUNG ZUR HERSTELLUNG, LAGERUNG UND ANWENDUNG VON STICKOXYDUL ZUSAMMENFASSUNG

Es werden die chemischen Grundlagen der industriellen Herstellung und Reinigung von Stickoxydul (Lachgas) beschrieben und erortert. Stickoxyd ist die am ehesten in Frage kornmende toxische Verunreinigung, aber die Konzentration in dem aus einer Flasche stromenden Gas nimmt aufgrund der Fraktionierung rasch ab. Stickoxyd reagiert mit Sauerstoflf, doch die Reaktionsgeschwindigkeit wird von der Konzentration beeinflufit und ist bei Konzentrationen unter etwa 0.1 Prozent sehr gering. Die meisten Verfahren zum Nachweis und zur Messung von Stickoxyd hangen von einer vorherigen Oxydation ab. Die Reaktion zwischen Stickstoffdioxyd und Wasser ist verwickelt; toxische Wirkungen resultieren aus der Bildung von WasserstoffIonen, Stickoxyd, Nitrat- und Nitrit-Ionen.

A SYMPOSIUM on T H E ANAESTHETIC PROBLEMS FACING DEVELOPING COUNTRIES

will be held on June 10, 1967, at 10.30 ajn. in the POSTGRADUATE LECTURE THEATRE, ADDENBROOKE'S HOSPITAL, TRUMPINGTON STREET, CAMBRIDGE.

Chairman: Dr. A. H. Galley. Speakers: Professor Leslie Banks: "Health Services in developing countries" Dr. J. V. Farman: "Why should we help developing countries?" Dr. T. B. Boulton: "Technical considerations in anaesthesia overseas" Dr. I. C. Geddes, Dr. A. G. Brown: "Personnel employed in anaesthesia" Dr. Aileen K. Adams: "Anaesthesia abroad; the present position" If you wish to attend please inform the Secretary of the Medical School, Tennis Court Road, Cambridge, from whom further details may be obtained. A fee of £1 Is. will be charged. A buffet lunch will be available for 6s. to all those whose applications arrive by June 5.

Downloaded from http://bja.oxfordjournals.org/ at Carleton University on June 13, 2015

Magill, P. (1956). Air Pollution Handbook, p. 3. New York: McGraw Hill Book Co. Rabson, S. R., Quilliam, J. H., and Goldblatt, E. (1960). The elimination of nitrous fumes from blasting gases. J. S. Afr. Inst. Mm. Metall., 61, 152. Simona, E. (1959). [Concerning the degree of purity of anaesthetic gas especially with regard to carbon monoxide.] Bull, schtveiz. Akad. med. Wiss., 15, 423. Toothill, C (1967). The chemistry of the in vivo reactions between haemoglobin and the higher oxides of nitrogen. Brit. J. Anaesth., 39, 405. Wood, B. J., and Wise, H. (1955). Acid catalysis in the thermal decomposition of ammonium nitrate. J. chem. Phys., 23, 693. Yost, D. M., and Russell, H. (1946). Systematic Inorganic Chemistry of the fifth- and sixth-group nonmetaUic elements. Oxford: University Press.