Searching for an Experimental Procedure for the Measurement of

Alkane-Water Partition Coefficient

Hind Gharbi

Dissertation presented to
Escola Superior de Tecnologia e Gestão
Instituto Politécnico de Bragança
In order to obtain Master’s Degree in
Chemical Engineering

This work was supervised by

Maria Olga Amorim e Sá Ferreira
Simão Pedro Almeida Pinho
Radhia El Aissi

July, 2019
 Acknowledgements
I would to thank my IPB supervisors for making this work possible.

Professor Simao Pinho for being such a good guidance, for all the discussions pushing me to
always look at things differently.

Professor Maria Olga Amorim Ferreira for showing me how work must be done in her one kind
gentle way.

I would like to think all the research team for always being disposed to help and for the good
vibes.

Most important of all I dedicate this achievement to my mom and thank here for being my rock,
for believing and supporting me, for the friendship, for the dictature, for the tenderness,
love…for every(good :p) thing I am and everything I am not.

For my dad, you’ll not see this but I miss you every day more, doing this work I thought of you
during every step of the way. The brilliant and unique mind that you are is letting a huge hollow
in this world...still feel your love and kindness.

To the sweetest and craziest human being ever, I love these big beautiful eyes that always look
at me as the most powerful super hero (will always be yours)! I thank God every day for giving
me such a wonderful big little sister.

For the most supportive family ever T.R, K.S, T.A, K.Z and all my cousins, love you guys.

I can’t miss on thanking a noble soul that I have been blessed to meet and love. Thank you for
believing in me when I made it hard for you. Thank you for the friendship, partnership and
relationship. Mehdi.

To Mariam Thank you for always dancing the robo-dance with me, where ever the place,
whenever the time and under all circumstances.

To Emna, Khouloud, Anis thank you for the dreams, laughs, support and everything.
Abstract

The amount and variety of pollutants that are detectable and toxic, even in small concentrations,
has been growing. Despite the importance of studying these chemicals and their effect on the
environment, there is a lack not only of partition coefficients data of these pollutants between
water and alkanes, but also of experimental protocols to measure that information with
accuracy. This information is essential to predict their fate once released in the environment.

In this context, the main objectives of this work were to perform a literature review about the
methodologies used to measure partition coefficients, to collect and analyze a database of
partition coefficients of compounds with diverse chemical structure between water and alkanes
and, finally, to implement an experimental procedure to measure the partition coefficients at
298.15 K.

A variant of the shake-flask method was combined with UV-Visible spectroscopy and refractive
index methods of analysis to perform the partition coefficients measurements between
isooctane and water (Pisoo/w). Two solutes for which the solubility data are well established,
such as toluene and benzoic acid, were selected, covering a wide range of numerical values of
the partition coefficients. Hopefully, after, the methodology will be applied on compounds
highly relevant from the environmental point of view, but not yet studied. Solubility
measurements were also carried out as this type of data are essential for designing the partition
coefficients experiments.

In general, the results obtained are in close agreement with the scarce information available in
literature. The results of the partition coefficient of toluene was logPisoo/w = 3.24 ± 0.07 and the
distribution ratio of benzoic acid was log Disoo/w = 0.26 ± 0.03. The errors obtained in the
corresponding material balances were 3% and 7% for toluene and benzoic acid, respectively.
Resumo

A quantidade e a variedade de poluentes detetáveis e tóxicos, mesmo em pequenas

concentrações, tem vindo a aumentar. Apesar da importância de se estudar estes compostos e
os seus efeitos sobre o meio ambiente, faltam dados de coeficientes de partição de poluentes
entre água e alcanos, mas também protocolos experimentais para medir essa informação com
precisão. Esta informação é essencial para prever o seu destino uma vez libertados no ambiente.

Neste contexto, os principais objetivos deste trabalho foram realizar uma revisão bibliográfica
sobre as metodologias utilizadas para medir coeficientes de partição, recolher e analisar uma
base de dados de coeficientes de partição entre água e alcanos de compostos com estrutura
química diversa e, finalmente, implementar um procedimento experimental para medir os
coeficientes de partição a 298.15 K.

Para realizar as medições dos coeficientes de partição entre isooctano e água (Piso/w), aplicou-
se uma variante do método do frasco agitado, combinada com métodos de análise de
espectroscopia UV-Visível e de medição do índice de refração. Selecionaram-se dois solutos
para os quais os dados de solubilidade estão bem estabelecidos, o tolueno e o ácido benzóico,
cobrindo uma ampla gama de valores de coeficientes de partição. Espera-se que, mais tarde, a
metodologia seja aplicada a outros compostos relevantes do ponto de vista ambiental, mas ainda
não estudados. Foram também realizadas medições de solubilidade, pois este tipo de dados é
essencial para projetar os ensaios dos coeficientes de partição.

Em geral, os resultados obtidos estão de acordo com a reduzida informação disponível na

literatura. Os resultados obtidos para o coeficiente de partição do tolueno foram log Piso/w = 3.24
± 0.07 e a razão de distribuição do ácido benzóico foi log Diso/w = 0.26 ± 0.03. Os erros
associados aos balanços de material foram de 3% e 7% para o tolueno e o ácido benzóico,
respetivamente.
Table of Contents

List of Symbols and Acronyms: ................................................................................................. ii

List of Figures ........................................................................................................................... iii
List of Tables ............................................................................................................................. iv
Introduction .......................................................................................................... 1
1.1. Motivation ......................................................................................................................... 1
1.2. Objectives ......................................................................................................................... 2
Literature review .................................................................................................. 3
2.1. Distribution and partition coefficients ................................................................................. 3
2.2. Experimental methods ......................................................................................................... 4
2.2.1. Slow stirring method protocol .......................................................................................... 5
2.2.2. Shake flask method protocol ............................................................................................ 6
2.3 Partition coefficients database .............................................................................................. 7
2.4. Prediction methods ............................................................................................................ 15
2.5. Selected system and solutes .............................................................................................. 18
Experimental work ............................................................................................. 20
3.1 Materials ......................................................................................................................... 20
3.2 Methodologies ................................................................................................................ 20
3.2.1 Solubility measurement ................................................................................................... 20
3.2.2 Partition coefficient measurement ................................................................................... 21
3.3 Results and discussion ........................................................................................................ 23
3.3.1. Solubility data ................................................................................................................ 23
3.3.2 Partition coefficients data ................................................................................................ 24
3.3.2.1 Partition coefficient of toluene in isooctane/water ....................................................... 24
3.3.2.2 Distribution ratio of benzoic acid in isooctane-water .................................................. 25
Conclusions and future work.............................................................................. 27
References ................................................................................................................................ 28
Appendix A .............................................................................................................................. 31
Appendix B. Calibration curves ............................................................................................... 35

i
List of Symbols and Acronyms:

List of Symbols
P Partition coefficient
Poct/w Partition coefficient of octanol-water system
P16/w Partition coefficient of hexadecane-water system
Palk/w Partition coefficient of alkane-water system
Pcyc/w Partition coefficient of cyclohexane-water system
Piso/w Partition coefficient of isooctane-water system

List of Acronyms
COSMO-RS COnductor like Screening for Real Solvents
MOSCED MOdified Separation of cohesive Energy Density
SMD Solvation Model based on Density
SF Shake Flask
SS Slow stirring
RPM Revolutions Per Minute
Abs Absorbance
BA Benzoic acid
UV-vis UV-visible
HPLC High Performance Liquid Chromatography

ii
List of Figures

Figure 2.1 Linear relationship between log Poct/w and log Pcyc/w................................................ 8
Figure 2.2 Linear relationship between log P16/w and log Pcyc/w. ............................................... 8
Figure 2.3 Linear relationship between log Palk/w and log Pcyc/w................................................ 9
Figure 2.4 Evolution of log P of toluidine according to the position of substituents.............. 10
Figure 2.5 Evolution of log P of chloroaniline according to the position of substituents....... 10
Figure 2.6 Evolution of log P of methoxyaniline according to the position of substituents. .. 10
Figure 2.7 Evolution of log P of nitroaniline according to the position of substituents. ........ 11
Figure 2.8 Evolution of log P according to number of methyl group in methylaniline. ......... 11
Figure 2.9 Evolution of log P of chlorophenol according to the positions of substituents. .... 12
Figure 2.10 Evolution of log P of bromophenol according to the positions of substituents. .. 12
Figure 2.11 Evolution of log P of methylbenzaldehyde according to the positions of the
substituent................................................................................................................................. 13
Figure 2.12 Evolution of log P according to the number of carbons in alcohols. ................... 15
Figure 2.13 Evolution of log P according to the number of carbon in carboxylic acid. ......... 15
Figure 2.14 A representation showing related variables relevant to environmental partitioning.
The central quantities in black font are the fundamental thermodynamic parameters, those in
red are the measured physico-chemical properties. Figure adapted from[15]. ........................ 16
Figure 2.15 Diagram of log Poct/w according with log Poct/w adapted from [15]. ..................... 17
Figure 2.16 Partitioning of N-propylbenzamide and urea based pesticides in water, octanol and
air. ............................................................................................................................................. 19
Figure 3.1 Experimental set up of the solubility experiments. ................................................ 21
Figure 3.2 A) Glass graduated tubes; B) Thermostatic and shaking equipment (ThermoMixer
C, Eppendorf). .......................................................................................................................... 22
Figure 3.3 A) Refractometer, B) prism surface of the refractometer. ..................................... 23
Figure 3.4:Solubility of oxalic acid in water (g/100g of water) as a function of temperature: ⚫
[29], ⚫ [26], ⚫ [27], ⚫[28], x this work. ................................................................................. 24
Figure B.1 Calibration curves of benzoic acid in water and isooctane. .................................. 35
Figure B.2 Calibration curve of toluene in water. ................................................................... 35
Figure B.3 Calibration curve of toluene in isooctane. ............................................................. 36

iii
List of Tables

Table 2.1 General characteristics of some measurement and estimation methods for P. ......... 4
Table 2.2 Comparative table for SS and SF methods. ............................................................... 5
Table 2.3 Log Pcyc/w/log Pcyc/w for position isomers of dimethylphenol. ................................. 13
Table 2.4 Log Pcyc/w c of different position isomers of trimethylphenol. ................................ 14
Table 3.1 Properties of the compound used in this work.. ...................................................... 20
Table 3.2 Solubility of oxalic acid measured in this work as a function of temperature (standard
deviation in parenthesis). .......................................................................................................... 23
Table 3.3 Experimental value of log Piso/w obtained for toluene. ............................................ 24
Table 3.4 Experimental value of log Piso/w obtained for benzoic acid. .................................... 25
Table 3.5 Estimated log (wBA, in alkane/wBA, in water) of BA in several alkanes with water. .......... 26
Table A.1: Abraham's table of observed and calculated partition coefficient in four systems.
.................................................................................................................................................. 31

iv
 Introduction

1.1. Motivation

One of the most important concerns of the scientific community nowadays is the effect of
human’s activity (industry, transport, chemical weapon…) on the environment. While our
activities are expanding and prospering, the waste released in the environment are extending
as well, containing substances that can be toxic for the wildlife (animals, vegetables, micro-
organisms), thus threatening the natural equilibrium of the whole ecosystem.

To limit and even eliminate these kind of effects, predicting the fate [1] of chemicals released
to the environment is a priority. To assess the environmental risk of a substance, the scientific
community developed a list of parameters which can show the preferential compartments
where a given species will concentrate as well as the effect of a chemical on the environment.
It has been consented that the fate of chemicals is firstly controlled by partition coefficients
[2,3]. It indicates the extent to which the chemical will concentrate in an aqueous phase, air,
soil or an organism.

Measurements for partition coefficient has long been done for n-octanol-water. To avoid
issues with water saturated octanol phases, an heterogenous structure (caused by miscibility
of n-octanol and water), researchers have been also concentrating on the measurement of
alkane-water partition coefficients [4].

In this way, an experimental protocol will be developed and tested, firstly for solutes for which
the partition coefficients are well established, covering a wide range of numerical values of
these partition coefficients. Hopefully, after, the methodology will be applied on compounds
highly relevant from the environmental point of view, but not yet studied.

1
1.2. Objectives
The main objective of this work is the compilation and analysis of the experimental data
published so far of partition coefficients of organic solutes in water-alkane biphasic systems.
In addition, a literature review about the experimental methods for the determination of the
partition coefficient will be carried out, with the perspective of optimizing and finding a
method more suitable to our laboratory conditions and needs. As a complement, a
methodology will be also applied to measure the solubility of solid solutes in liquid solvents
as this type of data is essential for designing the partition coefficients experiments.

2
 Literature review

2.1. Distribution and partition coefficients

A substance may distribute itself between two immiscible phases (solvents). This mass
transfer is accompanied by very small heat changes, and the migration of molecules from one
solvent to the other does not depend on a large positive value of the enthalpy change.
Therefore, it is assumed that this phenomena is controlled by an entropic effect [2]. Once the
system has reached a state of equilibrium (under specific conditions), the concentration ratio
of the compound in the organic phase and the aqueous phase is calculated to have a
representative value of the partition of the solute between the two phases. This ratio defines
what is called the partition coefficient [5]. Being the quotient between the equilibrium
concentration in the organic phase (Corganic phase ) and the aqueous phase (Cwater ) given by
equation (1) [6] and typically measured between 20 and 25 °C. The partition coefficient (P)
is dimensionless and is usually given in the form of its logarithm to base ten [4].

P = Corganic phase/ Cwater (1)

The distribution ratio D between an organic phase and water is a measure of P that accounts
for the pH dependency of an ionizable organic chemical, and is a measure of the distribution
of dissociated and non-dissociated species in organic and water as a function of pH. The extent
to which an ionizable compound is dissociated across environmentally relevant pH ranges
may have a marked effect on properties such as water solubility. It should be noted that neutral
form of the species is generally less water soluble and is thus more hydrophobic as compared
to the ionized or charged form. The relation between both is given by equation (2), where Ka
means the acid dissociation constant [30].

log 𝑃 = log 𝐷 + 𝑙𝑜𝑔[1 + 10(𝑝𝐻−𝑝𝐾𝑎) ] (2)

3
2.2. Experimental methods

During 1989, Sangster published an extensive review on experimental methods to measure

the partition coefficient. These are briefly summarized in Table 2.1, highlighting their main
characteristics and with a brief discussion to point out the most important specification of each
technique [8].

Going through the open literature it was concluded that nowadays the methods that were the
most effective, and so to say mostly used, are the shake flask method (SF) and the slow stirring
method (SS), which are analyzed in more detail in Table 2.2

Table 2.1 General characteristics of some measurement and estimation methods of P.

Optimal log P
Method Applicability Strengths Weaknesses
range
Time consuming Requires
Direct method attention Many details of
General for
Shake-flask -2.5 to 6 reliable when procedure
neutral species
precautions taken Less accurate at log P > 5

Column become a strim of

Direct method with
solute.
no danger of
General, except New column needed for each
emulsion
Generator 2 to 7 for very solute.
Closed system
column hydrophilic Limited to lipophilic
Efficient
compounds compounds.
temperature control
Rather elaborate analytical
Faster than SF
equipment
Reverse phase
Fast convenient,
high pressure 0 to 6 Neutral species Correlational method
Impurities does not
liquid only Single correlation inaccurate
interfere
chromatography
Neutral species
Water solubility 2 to 6 Correlational method
of limited water Convenient
correlation Single correlation inaccurate
solubility
Depends on
Ease of measurement
Henrian activity route to Thermodynamically
Neutral Species Accuracy depends on route
coefficient activity exact
to activity coefficient
coefficient
Fast convenient
Same as SF
Fragmental reliable for Requires expertise for
Useful for Same as SF
constants molecules without complicated molecules
extrapolation
many polar groups

4
 Table 2.2 Comparative table for SS and SF methods.
Slow stirring method Shake flask method

Log P estimated for the

Log P > 4 Log P < 4
chemical

No induced miscibility of the organic Take way less time than the slow
Advantages
phase into water stirring method

Suitable components With water miscibility Without water miscibility

System Octanol-water Cyclohexane-water

It is essential to indicate that in the case where the chemical species studied are not stable in
the pH of the solvent, a buffer most be used depending on the considered system. Being the
most used methods [5], the next section will give a general idea about the shake flask and the
slow stirring method.

2.2.1. Slow stirring method protocol

In this method, the following conditions should be considered [9]:

- Temperature constant during the experiment.

- Reaction vessel of larger than one liter has to be considered, so that a sufficient volume of
water can be obtained for chemical extraction and analysis.

- Purity of octanol at least +99 % (extraction with acid or distillation can be used to prepare
the octanol).

- Water should be glass or quartz distilled, or obtained from a purification system, or HPLC-
grade water may be used. Filtration through a 0.22 µm filter is required for distilled water.

- Both solvents are mutually saturated prior to the experiment by equilibrating them in a
sufficiently large vessel. This is accomplished by slow-stirring the two-phase system for two
days.

Regarding the analytical method for quantifying the amount of solute needed, the
concentration of the test substance should not exceed 70% of its solubility with a maximum
concentration of 0.1 M in either phase.

The volume of the phases should be chosen such that the 1-octanol layer is sufficiently thick
(> 0.5 cm) in order to allow the sampling. Typical phase ratios used for the determinations of

5
compounds with log P of 4.5 and higher are 20 experiments: 20 to 50 ml of 1-octanol and 950
to 980 ml of water in one liter vessel for log P of 4.5 and higher.

The stirring system

The 1-octanol-water system is stirred until equilibrium is attained. In a pilot experiment the
length of the equilibration period is assessed by performing a slow-stirring experiment and
sampling the water and organic phases. The sampling time points should be separated by a
minimum period of five hours (at least 3 samples).

The stirring rate should be increased slowly. Also it should be adjusted so that a vortex at the
interface is formed between water and 1-octanol of 0.5 to maximally 2.5 cm depth.

Equilibration time
The minimum equilibration time is one day before sampling can be started. It is assumed that
the equilibrium is achieved when a regression of the concentration ratio against time over a
time span of four time points yields a slope that is not significantly different from zero at a p-
level of 0.05.

Sampling
The stirrer should be turned off prior to sampling and the liquids should be allowed to stop
moving. The organic phase sample volume is close to 100 µL.

2.2.2. Shake flask method protocol

This protocol is a simplified version that aims to give a general idea of the shake flask method
that are currently being used [10,11].

-Organic phase:

10 μL of 10 mM compound in dimethyl sulfoxide DMSO.

5 µL of 200 μM propanolol in acetonitrile + 500 μL cyclohexane.

-Shake the mixture for 50 minutes using a plate shaker at 800 rpm.

-Separation of the two solvents by centrifugation for 5 minutes at 3700 RPM in a plate
centrifuge.

-Cyclohexane wells: 45 µl of octanol (to avoid accumulation in the HPLC column) + 5 µl

cyclohexane from the top phase.

6
-Water wells: 50 µl of water phase.

The slow stirring method is a modality of the shake flask method. Despite of the fact that they
are the most used experimental method, they remain quite complicated and needing permanent
checking.

2.3 Partition coefficients database

Abraham and coworkers made one of the most complete compilation of data published until
this day (Appendix A) that have been used in most of the subsequent works [12,13].

In what follows, we will use the data collected and provided by of Abraham’s works to try to
look for a mathematical correlation between Pcyc/w values and P of other types to help
predicting the Pcyc/w values and check for its consistency.

Some clarifications can be given about the data shown in Table A.1 (Appendix A)[12].
- For water-hexadecane, water-alkane and water-cyclohexane some values of P were obtained
directly from solubilities in water and organic phase.

- Sulfoxides, alkyl anilines, haloanilines and alkylpyridines were excluded from the water-
octanol set, because of a variable basicity problem.

- All aliphatic aldehydes were also excluded from the water-octanol set because of formation
of the hydrate, RCH(OH)2.

Correlation between partition coefficients:

In this section, log Poct/w , log P16/w or log Palk/w are shown in function of log Pcyc/w in Figures
2.1 , 2.2 and 2.3 to identify if there is a relation between the values that can help do an
approximation of log Pcyc/w for unknown components.

7
- log Poct/w = f(log Pcyc/w):

2
logPoct/w

y = 0.4357x + 1,2876
1 R² = 0.4807

0
-6 -4 -2 0 2 4 6

-1

-2
logPcyc/w

Figure 2.1 Linear relationship between log Poct/w and log Pcyc/w.

The determination coefficient is R2 = 0.4807, meaning that only 48.7% of change on log Pcyc/w
can be explained by the change on log Poct/w, which make this relation not very valuable, and
a big dispersion is observed.

- log P16/w = f(log Pcyc/w)

5
4 y = 0.9462x - 0.1203
R² = 0.9908
3
2
1
logP16/w

0
-6 -4 -2 -1 0 2 4 6

-2
-3
-4
-5
-6
logPcyc/w

Figure 2.2 Linear relationship between log P16/w and log Pcyc/w.

The determination coefficient is R2 = 0.9908, meaning that 99.08% of the change on log Pcyc/w
can be related to the change on log P16/w, which make this relation very interesting. The points
are almost aligned showing a very small dispersion.

8
- log Palk/w = f(log Pcyc/w):

5
4 y = 0.9678x - 0.0715
R² = 0.989
3
2
1
0
logPalk/w

-6 -4 -2 -1 0 2 4 6

-2
-3
-4
-5
-6
logPcyc/w

Figure 2.3 Linear relationship between log Palk/w and log Pcyc/w.

The determination coefficient is R2 = 0.989, meaning that 98.9% of the change on log Pcyc/w
can be related to the change on log Palk/w, which make this also an interesting relation.

Even though a good linear relation has been found, these data need some additional analysis
and several approaches should be followed before using them as explained above.

Aiming to predict or approximate the behavior of the value of log P, an attempt has been made
in the coming graphics to find a potential relationship between the chemical structure of the
studied elements and the value of the partition coefficient in the systems. In this way, log P
has been plotted according to different structural characteristics of a chemical group or
families.

Correlation between log P and chemical structure.

This aspect is explored from Figures 2.4 to 2.13.

9
 1,00
Toluidine-alk/w
0,90
Toluidine-cyc/w
0,80
logP 0,70
0,60
0,50
0,40
2 3 4
CH3 Position

Figure 2.4 Evolution of log P of toluidine according to the position of substituents.

1,30
1,20 Chloroaniline-16/w
1,10 Chloroaniline-alk/w
1,00 Chloroaniline-cyc/w
logP

0,90
0,80
0,70
0,60
0,50
2 3 4
Position ortho/meta/para

Figure 2.5 Evolution of log P of chloroaniline according to the position of substituents.

0,7
0,6 Methoxyaniline-16/w
0,5
Methoxyaniline-alk/w
0,4
Methoxyaniline-cyc/w
0,3
0,2
0,1
logP

0
-0,1
-0,2
-0,3
-0,4
-0,5
-0,6
2 3 4
Position ortho/meta/para

Figure2.6 Evolution of log P of methoxyaniline according to the position of substituents.

10
 3
Nitroaniline-oct/w
2,5
Nitroaniline-alk/w
2
Nitroaniline-cyc/w
1,5
Titre de l'axe
1
0,5
0
-0,5
-1
-1,5
2 3 4
Position ortho/metha/para

Figure 2.7 Evolution of log P of nitroaniline according to the position of substituents.

There is an interesting similarity between Figures 2.4 to 2.10 in which the log P of these
solutes follow the same behavior for the alkane-water and cyclohexane-water systems. The
log P is decreasing respectively from the ortho, meta to para disposition for all the
compounds.

In the following Figures we will continue to study aniline but with one and then two methyl
groups on the functional group. Log P is significantly higher for a dimethyl than for a
methylaniline.

2,5

2
logP

1,5

0,5

0
1 2
number of CH3
N-Methylaniline_alk/w N,N-Dimethylaniline_cyc/w

Figure 2.8 Evolution of log P according to number of methyl group in aniline.

11
The following two figures show the variation of log P for phenol substituted with halogens
(Cl or Br) in the ortho, meta and para positions.

2,5

1,5
logP

1 Chlorophenol-oct/w

0,5 Chlorophenol-alk/w
Chlorophenol-cyc/w
0

-0,5

-1
2 3 4
Position orhto-meta-para

Figure 2.9 Evolution of log P of chlorophenol according to the positions of substituents.

2,5

1,5
logP

0,5

-0,5
2 3 4 Bromophenol-oct/w
positon ortho-meta para Bromophenol-cyc/w

Figure 2.10 Evolution of log P of bromophenol according to the positions of substituents.

For phenols, there is a different behavior of log P in the octanol-water system on the other
hand for cyclohexane-water and alkane-water, log P is higher respectively in the ortho, meta
and para positions.

Figure 2.11 shows the variation of log Pcyc/w of benzaldehyde substituted with a methyl group:

12
 1,94
1,92
1,9
1,88
1,86
logPcyc/w

1,84
1,82
1,8
1,78
1,76
2 3 4
Ortho-meta-para position for substuent

Figure 2.11 Evolution of log Pcyc/w of methylbenzaldehyde according to the positions of the
substituent.

For methylbenzaldehyde, log P position meta < log P position para, and only complete data
is available for cyclohexane-water partition system.

Tables 2.3 and 2.4 show the collected results for tri and tetrasubstituted phenols, respectively.

Table 2.3 Log Pcyc/w and log Palk/w for position isomers of dimethylphenol.
log Pcyc/w log Palk/w

2,6-Dimethylphenol 0.99 0.81

2,3-Dimethylphenol 0.56 0.42

2,4-Dimethylphenol 0.43 0.29

2,5-Dimethylphenol 0.33 0.21

3,4-Dimethylphenol 0.20 0.08

13
 Table2.4 log Pcyc/w of different position isomers of trimethylphenol.
log Pcyc/w

2,3,5-Trimethylphenol 0.92

2,4,5-Trimethylphenol 0.86

3,4,5-Trimethylphenol 0.61

According to the data, in general, the value of log P is greater when the CH3 groups are closer
to the functional group OH, log P decreases with the increase of the distance between CH3
substituent and the main functional group.

In order to find a correlation between the number of carbons in a family of compounds and its
partition coefficient, the log P of alcohols or carboxylic acids is presented in Figures 2.12 and
2.13, respectively, according to the length of the hydrocarbon chain. It is easy to conclude that
all the partition coefficients increase with the number of carbons in the structure of alcohols
or carboxylic acids, and octanol-water partition coefficients are always larger than all the other
partition coefficients.

alcohol functional group

3
2
1
0
logP

-1
-2
-3
-4
0 1 2 3 4 5 6 7 8
number of C/chain
cyc/w alk/w 16/w oct/w

14
 Figure 2.12 Evolution of log P according to the number of carbons in the alcohols series.

carboxylic acid functional group

2
1,5
1
0,5
0
-0,5
logP

-1
-1,5
-2
-2,5
-3
-3,5
0 1 2 3 4 5 6
number of C/chain
oct/w 16/w alk/w cyc/w

Figure 2.13 Evolution of log P according to the number of carbon in the carboxylic acids series.

2.4. Prediction methods

Predicting models are ways to calculate a physical parameter by means of mathematical
equations, and correlations with other variables that represent physico-chemical properties
such as: charge densities, molecular surface [14], volume and free energy of the molecule we
are interested in. The advent, speed and low cost of computing technologies facilitated the
development of fundamental quantum mechanical methods of calculating solute–solvent free
energy interactions leading to “first principles” estimation methods that require no
experimental data, except for validation [12].

Prediction models for partition coefficient are based on an indirect path of calculation. Hence
partition coefficient values are calculated from parameters that are related to.

Figure 2.14 illustrates the relation between variables and why it is possible to calculate one
from the other. In the following illustration, Mackay and coworker consider octanol as the
organic phase and not cyclohexane but it does not affect the accuracy of the reasoning.

15
 Figure 2.14 A representation showing related variables relevant to environmental partitioning. The
central quantities in black font are the fundamental thermodynamic parameters, those in red are the
measured physico-chemical properties. Figure adapted from [15].

There are several calculation model such as COSMO-RS [16], MOSCED [16], OPLS-AA
[17], among others. Most of these models, or an improved version of them, have been used
for the SAMPL5 (Statistical Assessment of the Modeling of Proteins and Ligands) challenge
which is a blind predictive challenge and one of the most important works that have been done
on the cyclohexane-water partition coefficient until now [18]. This project gave interesting
results that helped to figure the relevance of the calculation models that have been used. A
brief presentation of two of these models will be given in the following sections.

COSMO-RS (COnductor-like Screening MOdel for Real Solvents)

Principle
Conventionally, the non-interacting gas phase state of the molecule was used as a
thermodynamic reference state. The COSMO model is based on solvation theory in which
quantum-mechanical calculations are used to predict, by energy minimization, the preferred
configuration and electron distribution of a molecule within an electrostatically responsive
solvation environment. The “COSMO” approach is much closer to a realistic solvation state,
so the error in using it as a reference for real solvents is much smaller than that associated
with the gas phase. Continuing the same path to get closer to the real state; to determine the
relative Gibbs free energies of the actual solvents one needs to regulate the solvent response
in the COSMO-RS calculation on the ideal or perfect solvent (i.e infinite dielectric constant).
COSMO-RS theory then calculates the relative energy of interaction of the molecule with its
16
surrounding “solvent” with respect to this ideal COSMO state [19]. That is where the approach
gets its name COSMO-Real Solvent [20]. To have a general definition of COSMO-RS: it is a
solvation theory based on the determination of solvation energy in a realistic solvent,
thermodynamically referenced to a perfect or ideal solvation environment.

Accuracy of the model

We can notice from Figure 2.15that the dispersion of the predicted and experimental values
of Poct/w is very small in such a way that the linear line passes through almost all the points.
The following graph is adapted from Mackay [15].

Figure 2.15 Diagram of experimental and COSMO-RS log Poct/w adapted from [15].

MOSCED model (Modified Separation of cohesive Energy Density)

Solubility is an important parameter in the study and the prediction of the partition coefficient.
MOSCED (Modified Separation of cohesive Energy Density) [16] is one of several solubility
parameter methods used nowadays. With only knowledge of solute chemical structure, the
parameters for the MOSCED method can be obtained by performing electronic structure
calculations with the SMD continuum solvent model. The specificity of this model is that it
can predict infinite dilution activity coefficients as a function of temperature which means that
it is not limited to 298 K. In addition, once parameterized, MOSCED calculates the infinite
dilution activity coefficient of the solute in the solvent, in addition to the infinite dilution
activity coefficient of the solvent in the liquid solute (hypothetical).

Despite the advantages cited above the accuracy of the model has not yet been confirmed
since it was used for the SAMPL5 challenge which gave the following data: correlation
coefficient = 0.6 ± 0.1 (error), root mean square of 2.8 ± 0.3 [21].

17
2.5. Selected system and solutes
Despite of the lack of effectiveness of the experimental procedures and the calculation models
available, partition coefficients are very important in toxicology and environmental chemistry
[22]. Figure 2.16, adapted from a work within our research group [23], explains the
importance of partitioning and the capacity of chemicals to contaminate water, soil and air.
On the other hand, it highlights the gap between the experimental and calculated values and
confirms the weakness of the available data.

The main objective in this work is to implement an experimental methodology to measure the
partition coefficients using the following model solutes: toluene and benzoic acid. The
experimental methodology developed in this work, will be applied in future work to important
compounds such as nitrophenol (and isomers), hydroquinone, p-benzoquinone, catechol,
paracetamol, acetamide, acetic acid and oxalic acid. The isooctane-water system will be used
for the partition phases. The choice of the partition system was not done randomly; knowing
that phenol, hydroquinone, p-benzoquinone, catechol and carboxylic acids are the oxidized
intermediates produced by the peroxidation of 2-NP (2-nitrophenol) in aqueous phase
considering them in a biphasic mixture of isooctane-water is in fact a simulation of
contaminated oily streams.

In addition, considering the importance of knowing the solubility data of the solutes in water
to design the partition assays, an experimental methodology to measure solubility data will be
also developed.

18
Figure 2.16 Partitioning of N-propylbenzamide and urea based pesticides in water, octanol and air.

19
 Experimental work

3.1 Materials
All the compounds described in Table 3.1 were used as received, and the solids kept in a
desiccator to avoid water contamination. Ultra-pure water (distilled twice) was used for the
solubility and the partition coefficient measurements.

Table 3.1 Properties of the compounds used in this work.

Compound CAS number Purity Supplier
(weight fraction)
Oxalic acid 2-hydrate 6153-56-6 0.995 Applichem Panreac
Isooctane 540-84-1 0.999 Acros Organics
Toluene 108-88-3 0.9985 Acros Organics
Benzoic Acid 65-85-0 0.99 Acros Organics

3.2 Methodologies
3.2.1 Solubility measurement
The solubility experiments were carried out by the analytical isothermal stirring method using
the experimental setup in Figure 3.1. A saturated solution was prepared mixing a small excess
of solid solute with about 50 ml of solvent. To reach equilibrium, the solution was
continuously stirred for at least 30 hours, and placed in a thermostatic bath at 25 °C (Lauda
Instruments, model E20, Ecoline 025). Later, the solution was allowed to settle at least 14 h
before sampling [25]. From preliminary experiments, it was found that after equilibrium at 25
ºC, and for temperatures between 30 and 40 ºC, a stirring period of 20 hours followed by a
period of 3 hours of settling time was adequate.
Samples (at least 1 ml) of the saturated liquid phase were after collected using plastic syringes
coupled with polypropylene filters (0.45 μm), previously heated, in order to avoid any
precipitation. The gravimetric method was chosen for the quantitative analysis. Therefore, the
samples were placed into pre-weighed glass vessels and immediately weighed (±0.1 mg). The

20
 Figure 3.1 Experimental set-up for the solubility experiments.
next step was to evaporate all the solvent, first under the hood until there is no more liquid
then into the oven at 70 °C. The drying procedure was followed by taking weight
measurements every 48 h until a constant value was reached.

3.2.2 Partition coefficient measurement

The partition coefficient experiments were carried out by a variant of the shake flask
procedure. The methodology developed in this work was divided in three parts as follows:

Part 1: Preparation of the tubes

- Preparation of the mother solution containing the solute (either organic or aqueous,
depending on the solubility of the solute in each phase) by weighing exactly the right amount
of solute used.

-Transfer the mother solution to the glass tube (Figure 3.2A) with a glass pipette.

- Addition of the other phase to the tubes, with the same volume of the mother solution phase.
To minimize the amount of air above the liquid, around 7 ml of each phase were prepared to
fill the 15 ml tubes.

Part 2: Establishment of the equilibrium and phase separation

-To help transfer solute from one phase to another, once closed, the tubes were placed in a
thermostatic shaker (ThermoMixer C, Eppendorf 15ml) at 25 °C (Figure 3.2B) and with
shaking rate of 300 rpm for 6 hours.

- The solutions were allowed to settle one week, at 25 ºC, before the composition analysis of
both phases.

21
 A) B)

Figure 3.2 A) Glass graduated tubes; B) Thermostatic and shaking equipment (ThermoMixer C,
Eppendorf)
Part3: Chemical analysis of the phases in equilibrium

Preferably, the concentrations of the solute in both phases should be evaluated.

- Considering that the aqueous phase (lower phase) should be sampled by a procedure that
minimizes the risk of contamination from the upper phase, first the organic phase was sampled
with a glass pipette, and after removed, letting just a thin layer (interface layer). After a syringe
with removable needle goes through this layer while expelling air gently [4] to sample the
aqueous phase.

Depending on the order of magnitude of the concentrations the analysis were carried out by
UV spectrophotometry (model Jasco V-730) or by measuring the refractive index (Abbemat
500, Anton Paar) with a reproducibility within ±1×10-5. Figure 3.3 shows the refractometer
equipment used. Adequate dilutions were made to have proportions fitting those of the
calibration curves. Once the analysis was done, the total amount of solute present in both
phases could be estimated and compared with the quantity originally introduced, assuming
both phases to be totally immiscible.

A calibration curve for toluene in isooctane was prepared in few steps (Appendix B):

- measurement of the refractive index of water (reference substance).

- measurement of the refractive index of the standards. First, the prism was washed with the
standard itself to avoid contaminations with previous substances. The filling height should be
at least 1 mm above prism surface, which means minimum of 1ml of sample.

-As a final measurement, the water reference value is again checked.

22
 A) B)

Figure 3.3 A) Refractometer, B) prism surface of the refractometer

Calibration curves (Appendix B) were also prepared for UV measurements with a maximum
wavelength of 262 nm for toluene and 273 nm for benzoic acid: preparation of six to seven
standards covering an adequate range of concentrations, using the mixed solvents water +
ethanol (50:50 in weight) or isooctane + ethanol (50:50 in weight).

3.3 Results and discussion

3.3.1 Solubility data
The solubility measured for oxalic acid at 298.2 K was 11.28 ± 0.08 g/100 g of water which
is in closer agreement with the values obtained by Hyva et al. [25], Flottman [27] and Apelblat
[26] who obtained respectively 12, 16 and 11.12 g/100 g of water.

Our data, presented in Table 3.2, was measured between 298.2 and 313.2 K. The change of
the solubility with temperature can be seen in Figure 3.4. Our data can be compared with the
data reported by Apelblat [26] who measured the solubility in a wider temperature range. The
solubility reaches 47.5 g/100g of water at 338.15 K, showing that temperature has an
important effect on the solubility of oxalic acid.

Table 3.2 Solubility of oxalic acid measured in this work as a function of temperature (standard
deviation in parenthesis).
Temperature (K) Solubility (g oxalic
acid/100 g of water)
298.2 11.28 (0.08)
303.2 14.03 (0.14)
308.2 16.89 (0.39)
313.2 21.67 (0.11)

23
 50

Solubility (g oxalic acid/100g of water)

45
40
35
30
25
20
15
10
5
0
270 280 290 300 310 320 330 340 350
Temperature in K

Figure 3.4 Solubility of oxalic acid in water (g/100g of water) as a function of temperature:
⚫ [27], ⚫ [25], ⚫ [28], ⚫[26], x this work.

3.3.2 Partition coefficients data

3.3.2.1 Partition coefficient of toluene in isooctane/water

The partition coefficient measurements where carried in at least six tubes with the same
composition, to assess the precision of the methodology. Considering the low amount of water
phase, the very low solubility of toluene in water, and the analytical methods restrictions, a
considerable concentration of toluene was dissolved initially in the organic phase (23% w/w).

Table 3.3 presents the average partition coefficient of toluene in isooctane-water system
obtained in this work and the corresponding standard deviation.

Table 3.3 Experimental values of log Pisooct/w obtained for toluene.

log Pisooct/w
Experiment Average value ± standard deviation Individual values
First round 3.08±0.16 3.13; 3.17; 3.32; 3.00; 2.87; 2.89
Second round 3.24±0,07 3.34; 3.26; 3.18; 3.22; 3.16; 3.30

In preliminary experiments, the mass fractions of toluene in the isooctane phase were
determined by UV-Vis spectroscopy which implied a long analytical procedure with several
dilutions (dilution factor of 103). The partition coefficient obtained was 3.08 with a standard

24
deviation of 0.16 and the material balance resulted in an average relative error of 24%. To
improve the consistency of the material balance results, another method of analysis was
adopted for the organic phase measurements. A calibration curve was prepared using a
refractometer in an appropriate range of concentrations as described in the previous section.
Following this approach, the average value of 3.24 for the partition coefficient was obtained,
with 0.07 standard deviation and a material balance with an average error of 3%. This
procedure confirms the importance of choosing the adequate analytical methods.

For alkane-water systems, there is a clear lack of partition coefficients data. But a comparison
can be done with an approximation of log P of toluene in other system with an organic phase
almost completely immiscible with water such as cyclohexane [13]. According to Abraham’s
data [12] log Pcyc/w is equal to 2.85, calculated using the ratio of molar concentrations, which
can be considered in the same range and in good agreement with log Pisooct/w obtained from
the experimental work. It should also be emphasized that the value obtained here can be quite
different from the value obtained under the conditions of low solute concentration. Therefore,
further studies should be carried out by increasing the volume of the water phase compared to
the volume of the organic phase, to allow the analytical quantification of the aqueous phase.

3.3.2.2 Distribution ratio of benzoic acid in isooctane/water

For this system, an initial solution of benzoic acid in isooctane (BA mass fraction equal to
0.00527) was prepared for the partition experiments. The value of distribution coefficient
obtained from this work is log Disooct/w = 0.26 with a standard deviation of 0.035 which
apparently disagrees with Abraham’s value [12] for the partition coefficient which is -0.71 in
cyclohexane-water.

Table 3.4 Experimental values of log Pisooct/w obtained for benzoic acid.
log Pisooct/w
Average value ± standard deviation Individual values
0.2607±0.03 0.2887/0.2947/0.2462/0.2615/0.2764/0.2656/0.1919

For solvents that are ‘‘almost’’ completely immiscible with water, such as alkanes,
cyclohexane, hexane, heptane, octane, nonane, decane and most aromatic solvents, as a first
approximation the ratio of the solubilities of the solute in the organic phase and in water can

25
be considered since it will be nearly identical to direct partition [13]. Table 3.5 presents the
solubility in weight fraction at 298.15 K of benzoic acid (BA) in different alkanes [29] and
their distribution ratios calculated considering the BA solubility in water which is 0.003 in
weight fraction at 298.15 K [14].

Table 3.5 Estimated log (wBA, in alkane/wBA, in water) of BA in several alkanes with water.
Solubility in weight fraction log (wBA, in
alkane/wBA, in water)*

Hexane 0.0067 0.35

Heptane 0.0093 0.49
Decane 0.0095 0.50
Cyclohexane 0.007 0.37
*obtained from the ration of weight fractions

We can see that the ratio of solubilities and the partition coefficient available in the literature
apparently are not in agreement. It should be noticed that BA is a weak acid that dissociates
in water and that is why it is more appropriate to consider a distribution coefficient (D) in this
case. Therefore, more experiments should be carried out to obtain the value of the distribution
ratio as function of pH and then, estimate the value of log P, using Equation 2.

26
 Conclusions and future work

Going through the experiments for the implementation of a procedure to measure partition
coefficients, one of the most delicate steps was the sampling of the water phase even though
we had satisfactory results. An improvement could be achieved by using tubes that can open
from below(faucet) which could be interesting an interesting option at least to confirm the
current method.
Also, it is highly recommended to avoid big factor dilutions on the volatile compounds (solute
or organic solvent), as it can reduce considerably the accuracy of the results. To that problem
the refractometer analysis can be considered as a solution since it showed its effectiveness on
toluene-isooctane measurements.
For salts and components that react in water, giving rise to different chemical species in
solution, the importance of the pH must be considered in these studies since the pH must be
known in order to establish then the concentration of each species in solution. In these cases,
measurements must be conducted under pH control.
After the partition experiments and going through literature, it is possible to postulate that for
partition in organic solvents that are almost totally immiscible with water, the effective
partition coefficient and the partition coefficient deduced from solubilities in both phases are
approximate values. In this case, parallel analysis of partition coefficient and solubility are
very interesting and could give very relevant results.
To continue the assessment of compounds that are important in environmental studies, it is
very interesting to consider first a set of experiments to evaluate the time needed to attain
equilibria and to change the volumes of each phase.

27
References

1. Drami A, Majstorovi H. Filip Andri. Linear modeling of the soil-water partition

coefficient normalized to organic carbon content by reversed-phase thin-layer
chromatography. Journal of Chromatography. 2016;1458:136–44.
2. Leo A, Hansch C, Elkins D. Partition coefficients and their uses. Chemical Reviews.
1971;71(6):525–616.
3. Jonker MITOJ. Detrmining octanol – water partition coefficient for extremely
hydrophobic chemicals by combining “slow stirring” and solid-phase micro extraction.
Environmental Toxicology Chemistry. 2016;35(6):1371–7.
4. OECD/OCDE. Partition coefficient (n-octanol/water): shake flask method. Vol. 107,
OECD Guideline for the Testing of Chemicals. 1995.
5. Berthod A, Carda-broch S. Determination of liquid – liquid partition coefficients by
separation methods. J Chromatography. 2004;1037:3–14.
6. Brooke D, Nielsen I. An interlaboratory evaluation of the stir-flask method for the
determination of octanol-wat partition coefficients. Chemosphere. 1990;21:119–33.
7. Sangster JRL. Octanol-Water partition coefficients of simple organic compounds.
Journal of Phys ChemRef Data. 1989;18(3):1111–226.
8. Morikawa G, Suzuka C, Shoji A, Shibusawa Y, Yanagida A. High-throughput
determination of octanol / water partition coefficients using a shake-flask method and novel
two-phase solvent system. Journal of Pharmaceutical Biomedical Analysis. 2016;117:338–
44.
9. OECD. OECD Guidelines for Testing of chemicals- partition coefficient (1-
Octanol/Water): Slow-stirring method. 2006;(March):1–16.
10. Rustenburg AS, Dancer J, Lin B, Feng JA, Ortwine DF, Mobley DL, et al. Measuring
experimental cyclohexane-water distribution coefficients for the SAMPL5 challenge. Journal
Computer Aided Molecular Design. 2016;30(11):945–58.
11. Finizio A, Vighi M, Sandroni D. Determination of N-Octanol/water Partition
coefficient(Kow) of Pesticide Critical Review and Comparison of Methods. Chemosphere.
1997;34(I):131–61.
12. Abraham, Michel H, Harpert. Hydrogen Bonding.32. An Analysis of water-octanol
and water-alkane partitioning and the ∆IogP parameter of seiler. J Pharmaceutical Science.
1994;83(8):1085–100.
13. Daniels CR, Charlton AK, Acree WE, Abraham MH. Thermodinamical behaviour of
dissolved carboxylic acid solutes : part 2 mathematical correlation of ketoprofen solubilities
with the Abraham general solvation model. Physics Chemistry Liquids. 2004;42(3):305–12.
14. Blokhus AM, Haavik J, Skauge A, Barth T. A0019 - Partion coefficients and
interfacial activity for Polar components in oilWater model Systems. J Colloid Interface
Science. 1999;41:33–41.

28
15. Mackay D, Celsie AKD, Parnis JM. The evolution and future of environmental
partition coefficients. Environmental Review. 2016;24(February):101–13.
16. Eckert F, Klamt A, Gmbh C, Kg C, Leverkusen D-. Fast Solvent screening via
quantum chemistry : COSMO-RS Approach. AIChE J. 2002;48(2):369–85.
17. Phifer JR, Cox CE, Ferreira L, Nogueira G, Karolyne A, Barbosa P, et al. Predicting
the equilibrium solubility of solid polycyclic aromatic hydrocarbons and dibenzothiophene
using a combination of MOSCED plus molecular simulation or electronic structure
calculations. Molecular Physics. 2017;115:1286–300.
18. Kenney IM, Beckstein O, Ligand Á. Prediction of cyclohexane-water distribution
coefficients for the SAMPL5 data set using molecular dynamics simulations with the OPLS-
AA force field. J Comput Aided Molecular Design. 2016;30:1045–58.
19. Bannan CC, Burley KH, Chiu M, Shirts MR, Gilson MK, Mobley DL. Blind
prediction of cyclohexane – water distribution coefficients from the SAMPL5 challenge. J
Comput Aided Molecular Design. 2016;30(11):927–44.
20. Klamt A, Diedenhofen M. Blind prediction test of free energies of hydration with
COSMO-RS. J Comput Aided Molecular Design. 2010;24:357–60.
21. Klamt A, Eckert F, Reinisch J, Wichmann K. Prediction of cyclohexane-water
distribution coefficients with COSMO-RS on the SAMPL5 data set. J Comput Aided
Molecular Design. 2016;30(11):959–67.
22. Samantha SD, Jeremy MB, Paluch AS. Predicting cyclohexane / water distribution
coefficients for the SAMPL5 challenge using MOSCED and the SMD solvation model. J
Comput Aided Mol. 2016;1007–17.
23. Cevc G. Partition coefficient vs . binding constant : How best to assess molecular
lipophilicity. European Journal of Pharmaceutics and Biopharmaceutics. 2015;92:204–15.
24. Schroder B, Monia M, Joao C, Simao P. Aqueous solubilities of fi ve N-(
diethylaminothiocarbonyl) benzimido derivatives at T=298.15 K. Chemosphere.
2016;160:45–53.
25. Ferreira O, Pinho S. Solubility of flavonoids in pure solvents. Industrial Engineering
Chemistry Research. 2012;51:6586–90.
26. Hyva A, Lihavainen H, Gaman A, Vairila L, Ojala H, Kulmala M, et al. Surface
tensions and densities of Oxalic, Malonic, Succinic, Maleic, Malic, and cis -Pinonic Acids.
Journal of Chemical Engineering and Data 2006;255–60.
27. Fr Flottman Liislichkeits gleichgewichte Aus dem Physikalisch-chemischen institut
der Universiti~t Marburg 1). 1927;(November).
28. Apelblat Alexender ME. Solubility of oxalic , malonic , and tartaric acids in water
from 278.15 to 338.15 K. Journal of Chemistry and Thermodynamics. 1987;19:317–20.
29. Faizal M, Smagghe FJ, Malmary GH, Lozar J, Mollnier JI. Water-Oxalic Acid-3-
Methyl-l-butanol Ternary Systems. Journal Chemical Engineering and Data.
1990;354(3):352–4.

29
30. European centre for ecotoxicology and toxicology of chemicals. Technical report
N.123 Environmental exposure assessment of ionisable organic compound. 2013

30
Appendix A

Table A.1 Abraham's table of observed and calculated partition coefficient in four systems.
LogPoct LogP16 LogPalk LogPcyc
obs calc obs calc obs calc obs calc
Helium 0.28 0.35 0.28 0.39 0.65 0.58 0.46 0.44
Neon 0.28 0.41 0.38 0.46 0.75 0.65 0.58 0.52
Argon 0.74 0.81 0.78 0.93 1.09 1.1 0.99 1.01
Krypton 0.89 1.03 1 1.18 1.28 1.34 1.24 1.27
Xenon 1.28 1.34 1.35 1.55 1.58 1.7 1.65 1.66
Hydrogen 0.52 0.57 0.78 0.75 0.69 0.63
Methane 1.09 1.04 1.14 1.2 1.37 1.36 1.33 1.29
Ethane 1.81 1.58 1.83 1.82 2.05 1.96 2.06 1.94
Propane 2.36 2.11 2.49 2.44 2.73 2.56 2.79 2.59
n-Butane 2.89 2.65 3.13 3.07 3.38 3.16 3.62 3.25
n-Pentane 1.39 3.19 3.87 3.69 4.06 3.77 4.27 3.9
Cyclohexane 3.44 3.38 3.91 3.87 3.72 3.94 4.15 4.13
Trichloromethane 1.97 2.1 1.69 1.68 1.74 1.83
Trichloroethene 2.42 2.52 2.68 2.58 2.86 2.79
Diethyl ether 0.89 1.08 0.85 0.76 0.66 0.92
Propanone -0.24 -0.16 -1.09 -1.03 -0.91 -0.92 -0.96 -0.95
Butanone 0.29 0.3 -0.43 -0.37 -0.26 -0.28 -0.25 -0.25
Pentan-2-one 0.91 0.85 0.18 0.27 0.34 0.35 0.44 0.42
Hexan-2-one 1.38 1.38 0.85 0.9 1 0.94 1.12 1.07
Hexan-3-one 1.27 1.1
Heptan-2-one 1.98 1.91 1.53 1.57 1.67 1.54 1.78 1.71
Methyl acetate 0.18 0.25 -0.39 -0.36 -0.2 -0.26 -0.19 -0.26
Ethyl acetate 0.73 0.79 0.15 0.28 0.29 0.36 0.34 0.4
n-Propyl acetate 1.24 1.34 0.77 0.92 0.9 0.99 1.1 1.08
n-Butyl acetate 1.78 1.86 1.41 1.53 1.67 1.57 1.75 1.72
n-Pentyl acetate 2.3 2.39 2 2.15 2.19 2.17 2.39 2.37
Methyl propanoate 0.82 0.82 0.28 0.32 0.5 0.41 0.57 0.46
Methyl pentanoate 1.96 1.88 1.51 1.56 1.81 1.75
Methyl hexanoate 2.42 2.4 2.04 2.16 2.39 2.38
Acetonitrile -0.34 -0.29 -1.11 -1.23 -1.46 -1.19
Ethylamine -0.13 -0.39 -1.62 -1.69 -1.78 -1.54 -1.8 -1.61
n-Propylamine 0.48 0.15 -1.08 -1.07 -0.98 -0.95 -0.98 -0.96
n-Butyl amine 0.97 0.67 -0.49 -0.47 -0.57 -0.36 -0.29 -0.33
Diethyl amine 0.58 0.42 -0.6 -0.52 -0.35 -0.41 -0.34 -0.37
Trimethylamine 0.22 0.04 -0.73 -0.61 -0.48 -0.48 -0.44 -0.46
Triethylamine 1.45 1.27 0.72 0.74 0.91 0.81 1.1 0.97
Nitromethane -0.35 -0.19 -1.06 -1.09 -0.98 -0.97 -0.93 -1.04
1-Nitropropane 0.87 0.84 0.44 0.33 0.45 0.4 0.53 0.44
Acetamide -1.26 -1.43 -4.68 -4.7 -4.88 -4.77
Propanamide -4.07 -4.17
N-butylacetamide -1.9 -2

31
N.N-Dimethylformamide -1.01 -1.18 -2.35 -2.51 -2.71 -2.47
Acetic acid -0.17 -0.18 -3.16 -3.04 -3.06 -2.89 -3.05 -3.08
Propanoic acid 0.33 0.31 -2.45 -2.45 -2.32 -2.32 -2.4 -2.47
Butanoic acid 0.79 0.86 -1.83 -1.8 -1.7 -1.68 -1.76 -1.78
Pentanoic acid 1.39 1.42 -1.14 -1.14 -0.92 -1.05 -1.1 -1.09
Methanol -0.74 -0.66 -2.77 -2.9 -2.8 -2.72 -2.49 -2.91
Ethanol -0.3 -0.15 -2.19 -2.1 -2.08 -1.94 -1.89 -2.06
Propan-1-ol 0.25 0.38 -1.53 -1.5 -1.39 -1.34 -1.49 -1.42
Butan-1-ol 0.88 0.91 -0.86 -0.86 -0.82 -0.75 -0.87 -0.77
2-Methylpropan-2-ol 0.76 0.94 -0.89 -0.82 -0.6 -0.7 -0.85 -0.73
Butan-2-ol 0.61 0.69 -1.05 -1.01 -0.8 -0.9 -0.96 -0.92
2-Methypropan-2-ol 0.35 0.6 -1.32 -1.07 -1.18 -0.94 -1.15 -0.96
Pentan-1-ol 1.56 1.45 -0.24 -0.24 -0.28 -0.15 -0.26 -0.13
Pentan-2-ol 1.19 1.22 -0.38 -0.4 -0.27 -0.31 -0.39 -0.28
Hexan-1-ol 2.03 1.98 0.38 0.38 0.48 0.45 0.45 0.53
Hexan-2-ol 1.76 1.75 0.32 0.29 0.23 0.37
Heptan-1-ol 2.72 2.52 1.03 1.01 1.04 1.06 1.12 1.18
2.2.2-Trifluoroethanol 0.41 0.53 -1.93 -1.9 -2.04 -1.95
Hexafluoroethanol 1.66 1.71 -1.37 -1.11 -1.46 -1.19
Methylthiol 0.65 0.69 0.93 0.59
n-Propylthiol 1.81 1.76 1.91 1.69 1.52 1.89
Methyl ethyl sulfide 1.54 1.59 1.73 1.64
Trimethyl phosphate -0.78 -0.77 -2.2 -2.12 -2.22 -2.08
Triethyl phosphate 0.8 0.68 -0.78 -0.51 -0.14 -0.33
Tri-n-propyl phosphate 1.87 1.95 0.91 0.83 1.18 1.15
Tri-n-butyl phosphate 2.74 2.49 2.9 2.96
Benzene 2.13 2.13 2.15 2.15 2.3 2.21 2.35 2.36
Toluene 2.73 2.66 2.68 2.76 2.85 2.81 2.99 3.01
Chlorobenzene 2.89 2.76 2.84 2.89 2.95 2.93 3.13 3.14
Methyl phenyl ether 2.11 2.19 2.09 2 2.07 2.03 2.19 2.24
Ethyl phenyl ether 2.51 2.66 2.61 2.54 2.77 2.81
Benzaldehyde 1.48 1.47 1.06 0.99 1.05 1.02 1.13 1.21
2-Methylbenzaldehyde 2.09 2.05 1.86 1.92
3-Methylbenzaldehyde 1.8 1.78
4-Methylbenzaldehyde 2.09 2.02 1.6 1.63 1.82 1.87
Acetophenone 1.58 1.69 1.14 1.16 1.11 1.17 1.27 1.4
Ethyl phenyl ketone 2.19 2.18 2.02 2
Benzophenone 3.18 3.24 3.29 3.19
Methyl benzoate 2.12 2.1 1.56 1.72 1.82 1.73 2.08 1.98
Ethylphenyl acetate 2.4 2.42
Benzonitrile 1.56 1.51 0.95 1.07 1.11 1.24
Phenylacetonitrile 1.56 1.6 1.31 1.24
Aniline -0.05 -0.08 0.05 0.04
o-Toluidine 0.36 0.4 0.52 0.51 0.61 0.69
m-Toluidine 0.45 0.45 0.58 0.62
p-Toluidine 0.39 0.43 0.56 0.61

32
2.6-Dimethylaniline 1.21 1.23 1.35 1.47
2-Chloroaniline 1.07 1.05 1.09 1.08 1.17 1.27
3-Chloroaniline 0.64 0.64 0.71 0.67 0.89 0.84
4-Chloroaniline 0.56 0.55 0.58 0.57 0.69 0.74
2-Methoxyaniline 0.33 0.38 0.39 0.4 0.52 0.6
3-Methoxyaniline -0.33 -0.27 -0.28 -0.25 -0.13 -0.07
4-Methoxyaniline -0.54 -0.47 -0.54 -0.45 -0.38 -0.26
2-Nitroaniline 1.85 1.85 0.22 0.23 0.2 0.24 0.36 0.42
3-Nitroaniline 1.37 1.54 -0.61 -0.62 -0.42 -0.48
4-Nitroaniline 1.39 1.24 -1.2 -1.15 -1 -1.03
N-Methylaniline 1.04 0.84 1.18 1.04
N.N-Dimethylaniline 2.17 2.24 2.31 2.24 2.47 2.54
1-Naphtylamine 1.15 0.91 1.26 1.26
2-Naphthylamine 1.06 0.9 1.21 1.25
Benzylamine 1.09 0.79 -0.21 -0.36 -0.12 -0.18
Nitrobenzene 1.85 1.84 1.54 1.46 1.45 1.48 1.69 1.68
Benzamide 0.64 0.47 -2.3 -2.34 -1.92 -2.28
Acetanilide 1.16 1.04 -1.85 -1.69 -1.31 -1.59
Benzoic acid 1.87 1.74 -0.71 -0.74 -0.85 -0.7
Phenol 1.46 1.54 -1.08 -0.98 -0.92 -0.9 -0.8 -0.89
o-Cresol 1.98 2.13 -0.09 0 0.08 0.06 -0.04 0.14
m-Cresol 1.98 1.96 -0.22 -0.36 -0.34 -0.3
p-Cresol 1.97 2.07 -0.19 -0.26 -0.3 -0.2 -0.35 -0.13
2.3-Dimethylphenol 0.43 0.42 0.51 0.56
2.4-Dimethylphenol 2.3 2.42 0.36 0.29 0.59 0.43
2.5-Dimethylphenol 0.43 0.37 0.57 0.5
2.6-Dimethylphenol 0.82 0.81 1 0.99
3.4-Dimethylphenol 0.21 0.08 0.25 0.2
3.5-Dimethylphenol 0.26 0.21 0.38 0.33
2-Ethylphenol 2.47 2.44 0.36 0.48
4-Ethylphenol 2.58 2.39 0.24 0.17 0.37 0.29
2.3.5-Trimethylphenol 0.97 0.92
2.4.5-Trimethylphenol 0.94 0.86
3.4.5-Trimethylphenol 0.63 0.61
2-Propylphenol 2.93 2.95 1.18 1.1
3-Propylphenol 0.83 0.89
4-Propylphenol 0.86 0.76 1.03 0.92
4-Butylphenol 3.56 2.87 0.77 0.86
4-Butylphenol 3.56 3.45 1.48 1.57
4-Phenylphenol 0.98 0.95
2-Fluorophenol 1.71 1.88 -0.43 -0.43 -0.3 -0.42
3-Fluorophenol 1.93 1.89 -0.7 -0.74
4-Fluorophenol 1.77 1.69 -0.7 -0.81 -1 -0.83
2-Chlorophenol 2.15 2 0.84 0.61 0.87 0.74
3-Chlorophenol 2.5 2.41 -0.08 -0.18 -0.12 -0.14
4-Chlorophenol 2.4 2.22 -0.76 -0.43 -0.39 -0.38 -0.35 -0.34

33
2-Bromophenol 2.35 2.28 1.05 0.81 1.16 0.98
3-Bromophenol 2.63 2.57 -0.06 -0.01
4-Bromophenol 2.59 2.42 -0.1 -0.24 -0.08 -0.2 -0.09 -0.12
2-Iodophenol 2.65 2.54 1.01 0.84 1.26 1.08
4-Iodophenol 2.91 2.85 0.35 0.23 0.57 0.39
2-Methoxyphenol 1.32 1.53 0.36 0.22 0.47 0.38
2-Cyanophenol -1.7 -1.52
4-Cyanophenol 1.6 1.47 -2.04 -2.03 -2.14 -2.01
Indole 0.79 0.82
Pyrazole -2.91 -2.85
Imidazole -3.7 -3.69
N-Methylimidazole -2.16 -1.96

34
Appendix B. Calibration curves

1,800
1,600
1,400
1,200
1,000
Abs

0,800
0,600
0,400
0,200
0,000
0,00E+00 5,00E-05 1,00E-04 1,50E-04 2,00E-04 2,50E-04
Weight fraction of Benzoic acid in solvent

Isooctane-Ethanol Water-Ethanol
Linéaire (Isooctane-Ethanol) Linéaire (Water-Ethanol)

Figure B.1 Calibration curves of benzoic acid in water and isooctane.

1,000
y = 2273,8x - 0,0089
0,900
R² = 0,9986
0,800
0,700
0,600
Abs

0,500
0,400
0,300
0,200
0,100
0,000
0,00E+00 5,00E-05 1,00E-04 1,50E-04 2,00E-04 2,50E-04 3,00E-04 3,50E-04 4,00E-04 4,50E-04
weight fraction of toluene in water

Figure B.2 Calibration curve of toluene in water.

35
 1,425
y = 0,0899x + 1,3907
R² = 0,9998
1,42

1,415

refractive index 1,41

1,405

1,4

1,395
0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4
weight fraction of toluene in isooctane

Figure B.3 Calibration curve of toluene in isooctane.

36