Phase behaviour of premixed 0.25% isoflurane in 50% nitrous oxide and 50% oxygen{dagger}

M. E. Tunstall and J. A. Ross*

Department of Environmental and Occupational Medicine, University of Aberdeen, University Medical School, Foresterhill, Aberdeen AB25 2ZD, UK j.a.ross@abdn.ac.uk


{dagger}Declaration of interest. Patented by Aberdeen University. If commercial production were to occur, the university and the authors would benefit.

Accepted for publication: July 18, 2002


    Abstract
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Background. Isoflurane (0.25%) in premixed nitrous oxide and oxygen 50/50, v/v, (IN2O), has been suggested for pain relief in labour.

Methods. Possible phase separation of this mixture was studied by analysis of samples from pre-filled cylinders as they were cooled.

Results. Condensation of isoflurane was found at 3.1°C in a cylinder, which held 8.7 MPa at 15°C. In a cylinder holding 13.8 MPa, which is the standard filling pressure stipulated by the National Health Service, the condensation temperature was –2.3°C. At the highest cylinder filling pressure investigated (14.15 MPa) the separation temperature was even less, –3.0°C. After exposure of cylinders to –40°C and complete phase separation of the mixture, complete mixing was achieved by 24 h storage in the horizontal position at room temperature and, either three complete inversions of the cylinder or mechanical rolling at 30 r.p.m.

Conclusions. These findings should assist the use and storage of IN2O.

Br J Anaesth 2002; 89: 814–19

Keywords: anaesthetics volatile, isoflurane; anaesthetic gases, nitrous oxide; anaesthetics, gases, premixed gases


    Introduction
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The addition of 0.2–0.25% isoflurane to premixed nitrous oxide and oxygen, 50/50, v/v, is helpful for the relief of pain in labour.1 2 Isoflurane and a 50% mixture of nitrous oxide in oxygen can be premixed together in the same storage cylinder at economical filling pressures and that such a mixture can be self-administered safely during labour3 and for pain relief during the removal of chest drains.4 Premixing anaesthetic agents with oxygen, stored under pressure in a single cylinder, allows a simple and safe method of administration but phase separation caused by cooling means that precautions are required for storage and handling. The characteristics of phase separation in a multi-component gas mixture are complex and alter with pressure. For example, if a cylinder of 50% nitrous oxide and oxygen (EntonoxTM), which is filled to 13.8 MPa at 20°C, is then cooled to –7°C, separation of nitrous oxide begins and a gas/liquid interface is formed. If the cylinder is partly emptied before cooling, for example to 11.7 MPa at 20°C, separation begins at the higher temperature of –5.5°C. Further emptying reduces separation temperature again.5 This behaviour of individual gas mixtures is unpredictable. Experiments to determine phase separation at different temperatures and pressures are required.6 7 We developed a technique of continuous sampling during externally applied cooling to measure separation of 0.25% isoflurane in equal parts of oxygen and nitrous oxide (IN2O) and derived a partial phase envelope curve.


    Methods
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Phase separation experiments
A chrome molybdenum steel pressure cylinder of a nominal 10-litre water capacity (Faber, British Standard 5045), 535 mm long, was fitted with a brass head drilled through with four ducts. Dip tubes reaching within the cylinder were connected to three of the ducts, only two of which were used in this project (Fig. 1). With the cylinder standing upright, gas could be sampled through the ducts from two different levels within. Gas sampling from the top was via a duct opening directly from the head into the cylinder. Gas sampling from the bottom was through a 1.5-mm internal diameter dip tube with a notched tip touching the inner surface of the cylinder. Both ducts led to their own stop valves attached to the brass head, and then on to tapered needle metering valves (Nupro ‘S’ Series Fine Metering Valve with a 0.79 mm orifice). A 4.72 mm internal diameter dip tube with a blind end reached 6.5 mm clear of the cylinder bottom. It held a thermistor probe (Yellow Springs series 400) calibrated against a mercury thermometer (National Physical Laboratory) for measuring the gas temperature internally in the most dependent part of the cylinder.



View larger version (17K):
[in this window]
[in a new window]
 
Fig 1 Variable scale broken diagram of test cylinder showing dip-tubes. A and B, gas sample pathway from bottom of cylinder. C and D, gas sampling and filling pathway from the top of the cylinder. E and F, insertion pathway for thermistor probe into closed end dip-tube. G, shrouded thermistor probe on surface of cylinder.

 
The cylinder was attached to a gas filling manifold assembly and evacuated (Edwards Speedivac 2 vacuum pump). Forty-eight grams of isoflurane were weighed out (Sartorius 1773) at room temperature and injected into the cylinder from a glass syringe through a septum, which was then isolated from the filling system by a quarter-turn brass ball valve (Whitey). The cylinder was then set on a weighing scale (Sartorius IB 31000P) and filled with 3529 g of EntonoxTM from a bank of cylinders to give a cylinder pressure of 14.35 MPa at 15°C. These quantities were calculated to give a final concentration of 0.25% isoflurane at standard atmospheric pressure and temperature.

Sampling tubes were attached to the cylinder head needle valves and a thermostatically controlled water circulator (Grant FH15) circulated warmed anti-freeze solution through plastic tubing coiled around the valves and attached tubing in order to prevent condensation of the test mixture after it had been sampled from the cylinder. After the warming coils were applied, the sampling arms were insulated with several layers of reflective plastic foil.

The test cylinder was cooled in a deep freeze chest (Derby), rated to operate at –40°C, within which an electric fan was used to circulate the air and a heat gun was used to slow cooling if it became too rapid. Three thermistor probes (Yellow Springs Instruments series 400) were calibrated against a mercury thermometer (National Physical Laboratory) and used for temperature recording. One was placed on the external surface of the cylinder half way along its length. It was shielded with a block of polystyrene foam backed with a 9 cm diameter pad of closed cell 4-mm-thick neoprene foam. Two more temperature probes were placed within the freezer cabinet for observing the circulating air temperature both close to floor and at close to lid levels.

Three series of experiments were performed. Before each, the test cylinder was filled as described above and then lowered into the deep freeze. The lid was closed on top of the wires and tubes trailing over the edge, and continuous gas sampling was commenced in anticipation of the changes in isoflurane concentration. Gas flow from the sampling arms was checked at intervals and the metering valves were adjusted to allow a flow of approximately 1 g of IN2O per minute using a specially marked flow meter. The gas sampled was analysed with an infra-red analyser (Datex Normac AA-102) for isoflurane and a paramagnetic analyser (Servomex type 570A/580A) for oxygen. Output signals from the analysers and thermistors were recorded on a flat bed recorder (Rikandenki RO3).

The temperature of the cylinder was allowed to decrease until a change in the concentration of isoflurane measured indicated the start of separation of the gas mixture. The first formation of isoflurane dew was taken to occur as soon as the concentration of isoflurane in the bottom sample began to decrease (Fig. 2). It was assumed that this was caused by isoflurane being lost as dew from the gas flow entering the dip tube becoming attached to immediately adjacent cold surfaces. The next stage is the formation of heavy vapour in the bottom part of the cylinder. As this happens, the concentration of isoflurane in the bottom sample increases and the concentration in the top sample decreases. The first formation of heavy vapour was noted when the decrease in isoflurane concentration stopped and a sustained increase had started (Fig. 2). The final stage is the formation of a pool of isoflurane liquid in the bottom of the cylinder. This contaminates the sampling system with large quantities of liquid isoflurane, which are then difficult to get rid of, so cooling was not allowed to progress to this stage. The valve on the sample path was closed and cooling stopped when the isoflurane concentration had increased to more than 0.3% and before liquid isoflurane had formed in the bottom of the cylinder.



View larger version (15K):
[in this window]
[in a new window]
 
Fig 2 Trace of part of a recording of a cooling experiment. The upper trace is internal cylinder temperature and is offset to the left by the chart recorder pen position. First dew formation is indicated by the decrease in isoflurane concentration at A. Heavy vapour formation is indicated by the increase in concentration at B, which is then stopped by closing the valve at END. The internal cylinder temperatures at which A and B occur are at X and Y, respectively; offset in position by the chart recorder. The weight of gas in the cylinder at point A was 1361 g.

 
After each experiment within a series, a measured quantity of gas was released to reduce the contents before the next experiment. The cylinder was then warmed, rolled at 30 r.p.m. for 30 min and allowed to stand overnight so that it would achieve a uniform temperature. The next morning the cylinder was weighed, its internal temperature and pressure recorded and another cooling study was done after a sample of the mixture had been analysed.

Remixing cylinder contents after their separation as a result of cooling
Two 12.2-litre capacity cylinders were used. They were filled with 54 g isoflurane and 4001 g EntonoxTM to give a homogenous concentration of 0.25% isoflurane and then cooled to –40°C in a deep freeze chest for at least 24 h. This causes separation of isoflurane and nitrous oxide to form a pool of liquid in the dependent portion of the cylinder. The cylinders were then removed and stored either in the vertical or horizontal position for some time at room temperature before being checked. Mixing of cylinder contents was then attempted either by three complete inversions of the cylinder or by rolling the cylinder at 30 r.p.m. for 5 min. The contents of the cylinder were regarded as homogenous when the concentration delivered was unchanged by inverting the cylinder. Thermistor probes were attached to the cylinder surface at its mid-point in order to detect the temperature fall during inversion, which would be caused by the partial re-vaporization of agitated liquid contents. Gas analysis with the cylinder inverted was not carried out if there was free liquid in the cylinder. The cylinder contents were then mixed by either rolling at 30 r.p.m. or by repeated inversions. Each cylinder was used in each of the four test conditions.


    Results
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Phase separation experiments
The temperatures at which dew and then heavy vapour started to form during cooling, related to the weight of the IN2O contained at the onset of dew formation, are shown in Table 1. Related to the weight of the gas mixture in the cylinder, they have a typical dome shape for the phase interfaces (Fig. 3). When the contents of the cylinder correspond to a filling pressure of 13.8 MPa absolute at 15°C (i.e. commercial filling pressure), the first formation of dew and then of heavy vapour occurred at –2.3 and –2.7°C, respectively. A further reduction in the cylinder contents gave a maximum temperature for first dew and for heavy vapour formation of at +3.1 and +1.3°C when contents were equivalent to filling pressures of 8.7 and 9.5 MPa, respectively (Fig. 3). The concentration of isoflurane in the test cylinder was measured as 0.25% both at the beginning and end of each series of experiments for each of the three cylinder fillings used.


View this table:
[in this window]
[in a new window]
 
Table 1 Temperatures at which isoflurane dew and heavy vapour first formed during cooling, related to the amount of IN2O gas in the cylinder at first dew formation
 


View larger version (22K):
[in this window]
[in a new window]
 
Fig 3 The effect of change in weight (filling pressure) of the contents of a cylinder of IN2O on the temperatures at which dew formation begins and heavy vapour is formed.

 
Remixing cylinder contents after their separation because of cooling
There was detectable non-homogeneity of cylinder contents after both 24 and 48 h at room temperature in the horizontal position but no indication of the presence of a liquid phase. After the 24 h test, complete mixing was achieved either by rolling for 5 min at 30 r.p.m. (Table 2), or by three complete inversions of the cylinder (Table 3). After 24 h storage in the vertical position, a decrease in surface temperatures of the cylinders on the first inversion indicated the presence of liquid. The concentration of isoflurane was 0.52% and the concentration of oxygen was 27% in the sample taken on the first inversion. An even mixture was not achieved either by three complete inversions (Table 2) or by 5 min rolling (Table 3) but was restored after 10 min rolling (Tables 2 and 3).


View this table:
[in this window]
[in a new window]
 
Table 2 Data from IN2O (0.25% isoflurane and 51% oxygen) freezing and recovery experiments
 

View this table:
[in this window]
[in a new window]
 
Table 3 Data from IN2O (0.25% isoflurane and 50% oxygen) freezing and recovery experiments. Oxygen was not measured in the first part of this experiment
 

    Discussion
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Condensation occurs in a gas when the saturated vapour pressure of the compound in question is exceeded. The saturated vapour pressure of nitrous oxide is 5.5 MPa at 20°C at atmospheric pressure8 but in a mixture of 50% nitrous oxide in oxygen at 13.7 MPa the partial pressure of nitrous oxide is 6.85 MPa without condensation. Further, the mixture is stable down to –7°C. We have identified a similar increase in the saturated vapour pressure of isoflurane when mixed with 50% nitrous oxide in oxygen. The saturated vapour pressure of isoflurane at atmospheric pressure and 20.91°C is 33.41 kPa.9 Yet at a pressure of 14.15 MPa a partial pressure of 35.38 kPa can exsist in the gaseous phase and the mixture is stable down to a temperature of –3.1°C.

Two effects explain this behaviour. The application of pressure to the condensed phase of a substance causes its vapour pressure to rise. This effect was first described by Poynting and is named after him.10 In addition, gas phase molecules attract molecules out of the condensed phase by the process of gas solvation.11 In high-pressure gaseous mixtures it is correct to think of the components as dissolved in each other. We studied the relative contribution of solvation and the Poynting effect in mixtures of isoflurane and 50% nitrous oxide in oxygen. At 10 MPa and 20°C the saturated vapour pressure of isoflurane was 95 kPa, an increase of 63 kPa from the saturated vapour pressure at atmospheric pressure.6 The Poynting effect predicted an increase in vapour pressure of 22 kPa at this pressure and so an increase of 41 kPa could be attributed to gas solvation. The solvent property of a gas depends on its density, which increases with pressure and so higher cylinder pressures are associated with greater solvent power of the gas. However, the increase in solvent power of the gas phase may not be enough to prevent condensation if a fixed concentration is studied and at pressures up to 9.35 MPa 0.25% isoflurane becomes progressively less thermally stable. At higher pressures, however, there is a progressive improvement in stability indicating that the solvent capacity of the gas mixture increases.

Although the phase behaviour of multiple component gas mixtures can be estimated by modelling techniques we have found it easier and more direct to adopt an experimental approach.6 7 This study has partially defined the phase diagram for the behaviour of 0.25% isoflurane in a mixture with equal parts of nitrous oxide and oxygen. A phase diagram is a map of the ranges of pressure and temperature at which each phase of a substance is the most stable (that is, has the lowest Gibbs energy).11 The cricondentherm is the maximum temperature of a multi-component system at which liquid and vapour can exist in equilibrium, and above which liquid cannot be formed regardless of pressure. The cricondenbar is the maximum pressure at which liquid and vapour can co-exist.12 At a temperature above its cricondentherm or at a pressure above its cricondenbar a mixture must be in a homogeneous phase.

We have defined the cricondentherm of IN2O as 3.1°C. Importantly, we have shown that this occurs at a pressure typical of a cylinder that is between one-half and two-thirds full (assuming a usual cylinder pressure for medical gas of 13.8 MPa). The temperature at which separation will start to occur in a full cylinder is –2.3°C. In other words the contents of a full cylinder are more stable with respect to temperature than of one that is partially used. This has important implications for the transport, storage and use of IN2O. An ambient outside temperature of –2.3°C is not unusual in the UK and so storage or transport in unheated areas would have to be avoided. An alternative approach might be to increase the cylinder filling pressure for the mixture, as this would make it more stable. Once the cylinder is in use, the thermal stability decreases until the cylinder pressure reaches 8.7 MPa. Further use improves stability again. Once a cylinder is in use it should be kept where ambient temperature cannot fall below 3.1°C. The freezing and recovery experiments show that the separated premixed gases are stable and that mechanical mixing after re-warming is necessary. It is better to store cylinders in the horizontal position.


    References
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Arora S, Tunstall M, Ross J. Self-administered mixture of Entonox and isoflurane in labour. Int J Obst Anesth 1992; 1: 199–202

2 Wee MYK, Hasan MA, Thomas TA. Isoflurane in labour. Anaesthesia 1993; 48: 369–72[ISI][Medline]

3 Ross JAS, Tunstall ME, Campbell DM, Lemon JS. The use of 0.25% isoflurane premixed in 50% nitrous oxide and oxygen for pain relief in labour. Anaesthesia 1999; 54: 1166–72[ISI][Medline]

4 Bryden FMM, McFarlane H, Tunstall ME, Ross JAS. Isoflurane for removal of chest drains after cardiac surgery. Anaesthesia 1997; 52: 173–5[ISI][Medline]

5 Bracken AB, Broughton GB, Hill DW. Equilibria for mixtures of oxygen with nitrous oxide and carbon dioxide and their relevance to the storage of N2O/O2 cylinders for use in analgesia. J Phys D Appl Phys 1970; 3: 1747–58[ISI]

6 Uyanik A, Marr IL, Ross JAS, Tunstall ME. Preparation and high-pressure behaviour of pre-mixed volatile liquid anaesthetics in Entonox. Br J Anaesth 1994; 73: 712P

7 Uyanik A. An investigation into the preparation, high-pressure behaviour and stability of premixed volatile liquid anaesthetics in Entonox. Aberdeen University Thesis, PhD, 1994

8 Grant WJ. Medical Gases: Their Properties and Uses, 1st Edn. Aylesbury, Buckinghamshire: HM+M Publishers, 1978

9 Nahrwold NL, Archer PG, Cohen PJ. Estimation of an equation relating saturated vapor pressure to temperature. Anesthesiology 1973; 39: 444–6[ISI][Medline]

10 Poynting JH. Change of state: solid:liquid. Philosophical Magazine 1881; 12: 32–48

11 Atkins PW. Physical transformations of pure substances. In: Atkins PW, ed. Physical Chemistry, 5th Edn. Oxford: Oxford University Press, 1994; 183–91/240–53

12 Ahmed TH. Hydrocarbon phase behavior. In: Chilingar GV, ed. Contributions in Petroleum Geology and Engineering. Houston, Texas: Gulf Publishing Company, 1989; 7: 1–35