Registration Dossier

Physical & Chemical properties

Partition coefficient

Currently viewing:

Administrative data

Link to relevant study record(s)

partition coefficient
Type of information:
experimental study
Adequacy of study:
key study
Study period:
Not specified.
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
data from handbook or collection of data
Taken from publically available data, and is considered accurate based on the registrants experience of the substance.
no guideline followed
Principles of method if other than guideline:
Partition coefficients of 32 gaseous anesthetics in the octanol water system have been determined. It is shown that relative anesthetic potency depends on hydrophobicity of the anesthetic (as defined by log P) and on a polar factor. The presence of a polar hydrogen in the anesthetic greatly increases potency. A quantitative structure-activity relationship is formulated based on these two factors.
GLP compliance:
not specified
Type of method:
other: Gas Chromatography
Partition coefficient type:
Analytical method:
gas chromatography
Key result
log Pow
Partition coefficient:
24 °C
Remarks on result:
other: pH is not specified.
Details on results:
See below

Results and Discussion


Equations 1-3 have been formulated from the data in Table I. In these equations p in log1/p is the effective anaesthetic pressure (ATA), n is the number of data points used in deriving the equations, r is the correlation coefficient, and s is the standard deviation.


Equation 1


Log 1/P = 1.193 (±0.59) Log P – 1.327 (±0.87)






Equation 2


Log 1/P = 1.913 (±0.69) I –0.596 (±0.45)






Equation 3


Log 1/P = 1.166 (±0.25) Log P + 1.881 (±0.33) I –2.106 (±0.39)






The figures in parentheses are the 95% confidence intervals.I is an indicator variable assigned a value of 1 to all compounds containing a "polar hydrogen atom". A11 other compounds are assigned value of zero for I.


The definition of a polar hydrogen atom is simply a phenomenological one. Those compounds which contain an electronegative element (O, halogen, etc.) attached directly to a carbon atom holding a hydrogen atom are observed to be more potent anaesthetics than compounds lacking such a hydrogen atom. This given a value of 1 for such compounds and zero for all others except HCCH and N2O. The report assumes that the CC function is electronegative enough to confer polar character on its H atoms. In the case of N2O, the hydrated form may be the active species.


Equation 1 accounts for only 38% (r2) of the variance in log 1/p while eq 3 accounts for 90%. The coefficient of 1.9 with I in eq 3 indicates that on the average, other factors being equal, molecules having a “polar” hydrogen (CH3C1, HClCH, HCC13, EtOEt, etc.) are about 80 times as potent anaesthetics as are compounds lacking such a function (CH4, CF4, CFC13, CF2C12, etc.). Carbon tetrachloride (logP2.83) with logPnot far from halothane (2.30) has very poor anaesthetic activity. The coefficients with the log P terms in eq 1 and 3 are characteristic of those we have found for correlation equations for the narcotic action of organic compounds on membranes and nerve processes.


Adding a term in (logP)2 to eq 3 makes a slight improvement in the correlation (r= 0.956). However, confidence limits cannot be placed on logPo; hence, this higher order equation is of little value. Although the log Po of 2.7 is not very reliable, it does suggest a parabolic dependence of anaesthetic potency on logPwhich is in accord with the earlier findings with a series of 28 anaesthetic ethers. This could be interpreted to mean that highly lipophilic gaseous anaesthetics do not reach equilibrium between gas phase and receptor sites.


One approach to interpreting eq 1 and 3 is that anaesthetic action is brought about by two properties of organic compounds: (1) hydrophobic and (2) polar. This conclusion was reached some time ago for the hypnotic action of alcohols, amides, and barbiturates. Thus, in eq 3 the results observe the normal increase in narcotic activity as logP increases. In effect, there are two such curves; one for the polar and one for the nonpolar anesthetics, separated by about 1.9 log units. The inclination is to rationalize theIterm by postulating that perturbation of the lipid space is also brought about by a polar interaction of the anesthetic. This might be related to the dipole moment of the anesthetic; however, using µ in eq 3 instead ofIdoes not yield as good a correlation(r= 0.858 for 28 data points; see Table I). The addition of µ and µ2 to eq 3 in place ofIalso falls far short of the correlation of eq 3( r= 0.893). Thus it appears that is not nearly as effective a parameter asIto model polar character. This does not rule out a role for the dipole moment, however, since there is considerable collinearity between I and µ as shown in the following correlation matrix:



Log P



Log P













The values in the above matrix are for r2; hence, the correlation between I and µ is high although logP is orthogonal to these two vectors. The model definitely brings out a polar component in anesthetic action, the nature of which is somewhat ambivalent because of the collinearity betweenIandp.It is interesting to note that the most potent anesthetics, chloroform, halothane, methoxyflurane, and

ethrane, all contain a “polar” hydrogen atom.


A fact which must be considered is that Miller et al found an excellent correlation between anesthetic pressure and olive oil-gas partition coefficients. This partition coefficient, obtained from a nonaqueoussystem, correlates anesthetic potency without the additional termIneeded with octanol-water partition coefficients (r= 0.994 for 16 molecules from Table I for which olive oil-gas log P values are available). Assuming an approximate equilibrium between anesthetic in the gas phase and anesthetic on receptor sites causing anesthesia, it would seem that olive oil-gas partitioning models the lipid receptor site quite well. This single partition coefficient contains both the hydrophobic and polar information contained in the two terms of eq 3.

There is a very high correlation between oil-water partition coefficients and octanol-water partition coefficients. This is illustrated by eq 4 which correlates solutes that do not contain a strong hydrogen bond donor such as OH or NH2 and is therefore applicable to the compounds of Table I.


Equation 4


Log P (oil-water) – 1.1Log P (octanol-water) – 1.30






Since it is the oil-gas partition coefficient which rationalizes anesthetic potency and not the oil-water constant, the conclusion is that solubility of the gases in olive oil must be determined by dispersion and polar forces (including hydrogen bonding) in such a way that olive oil models the effects of these forces in the critical lipophilic sites of action.


Equation 3 factors these two properties modelled by olive oil. The role of polar forces and especially hydrogen bonding is deemphasized in the octanol-water or olive oil-water partition coefficient. The polar interactions of the solute in water are still operative in the octanol, especially in the octanol-water system where about 4% water is present in the octanol phase, although to a lower degree. In effect, octanol- water or olive oil-water constants do not contain much polar information and this must be introduced using the variable I.


A number of comments about specific cases in Table I should be made. The congener C2F6 is very poorly fit and has not even been used to derive eq 1 and 3. It is about 25 times less active than expected. Perfluoromethane is reasonably well fit; however, SF6 is off by 1.5 standard deviations. All of these fluoro compounds are less active than one would expect from their high partition coefficients.

The fluoro compounds are better fit by olive oil-gas partition coefficients. The results suggest that they partition into octanol more readily than into nerve membrane. Although our log P values for the rare gases are, to our knowledge, the first to be published, partition coefficients for these gases have been calculated by taking the ratio of their oil and water solubilities .The values obtained by this method are: He, 0.23; H2,0.37; O2 0.56; N2, 0.55; Ar, 0.60; Kr, 0.88; Xe, 1 16; Rn, 2.04. The values were found by direct measurement of both phases of these gases partitioned between octanol-water are very close indeed to the calculated values obtained using different oils by different investigators.


Equation 5


Log P (octanol) – 1.018 (±0.17) Log P (oil) + 0.076 (±0.17)






Equation 5 expresses the correlation between the two sets of values. This equation is quite significant (F1,5=247) even though there is only a small difference between the two sets of values. We did not measure Rn in octanol- water; however, a value can be calculated for it from eq5.


In conclusion, the report agrees with Meyer and Overton, Mullins, and Miller, Paton, and Smith that the results of this study, taken with earlier results,strongly suggest that the critical phase in which anesthetic action occurs is lipophilic in character. However, our results clearly establish that there is an important polar component which plays a major role in the disruption of nerve function.

The data set is well justified, and the partition coefficient value derived in an appropriate manner with suitable statistical analysis. The substance is not considered to be bioaccumulative on the basis of this data.
Executive summary:

The data set is well justified, and the partition coefficient value derived in an appropriate manner with suitable statistical analysis. The substance is not considered to be bioaccumulative on the basis of this data.

Description of key information

Partition Coefficient data.

Key value for chemical safety assessment

Log Kow (Log Pow):
at the temperature of:
24 °C

Additional information

The value from Hansch et al (1975) is utilised as the appropriate source of the data. The substance is not considered to be bioaccumulative on the basis of this data.