Experimental study of water evaporation rate, at the

surface of aqueous solution, under the effect of a

discontinuity of chemical potential – Effect of water
activity and air pressure
Jean-Claude Benet, Samuel Ouoba, Francois Ouedraogo, Fabien Cherblanc

To cite this version:

Jean-Claude Benet, Samuel Ouoba, Francois Ouedraogo, Fabien Cherblanc. Experimental study of
water evaporation rate, at the surface of aqueous solution, under the effect of a discontinuity of
chemical potential – Effect of water activity and air pressure. Experimental Thermal and Fluid
Science, Elsevier, 2021, 121, pp.110233. �10.1016/j.expthermflusci.2020.110233�. �hal-03015745�

HAL Id: hal-03015745

https://hal.archives-ouvertes.fr/hal-03015745
Submitted on 20 Nov 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Experimental study of water evaporation rate, at the surface of
aqueous solution, under the effect of a discontinuity of chemical
potential - effect of water activity and air pressure

J.-C. Bénet1, S. Ouoba2, F. Ouedraogo3, F. Cherblanc1

Corresponding author: fabien.cherblanc@umontpellier.fr

1
LMGC, Université de Montpellier, CNRS, Montpellier, France
2
LPCE, Université Joseph Ki-Zerbo, Ouagadougou, Burkina-Faso
3
GERME & TI, Université Nazi Boni, Bobo Dioulasso, Burkina Faso.

Abstract

The aim of this communication is to analyse the influence of water activity and total pressure
on water evaporation. The system is composed of liquid and gas phases, separated by a plane
surface, contained in a cylinder whose volume is regulated by a piston. Water activity is
regulated by saturated salt solutions and pressure by the piston. The experimental device and
procedures were defined to limit the temperature variation at the interface. A transient method
is used. From a steady state, a volume increment is imposed; the resulting non-equilibrium
leads to an increase in the partial pressure of water vapour to the equilibrium pressure imposed
by the solution. Numerical calculation shows little variation in temperature in the gas-liquid
interface under the experimental conditions. An evaporation model is adopted taking into
account chemical potential discontinuity at the interface. The surface flux evaporation and
chemical potential jump at the interface are deducted from the total pressure recording. In the
neighbourhood of equilibrium, the surface flux of phase change is shown to be proportional to
the chemical potential jump. The surface coefficient of evaporation increases with the total
pressure of the gas phase and water activity.

Key words: surface evaporation rate; chemical potential; water activity; gas pressure;
transient method.

1
Nomenclature

𝑎𝑤 : Activity of water (/)

𝐶𝛼 : Specific heat capacity of the phases; 𝛼 = 𝑔 for the gas phase, 𝛼 = 𝑙 for the liquid
phase (J.kg-1.K-1)

𝐷𝑣 : Diffusion coefficient of water vapour in air (m².s-1)

ℎ𝑔 : Thickness of the gas phase (m)

ℎ𝑙 : Thickness of the liquid phase (m)

𝐽𝑣 : Evaporation flux by surface unit and unit time (kg.m-2.s-1)

𝐽𝑑𝑖 : Diffusive flux of the vapour (i=v) and air (i=a) (kg.m-2.s-1)

𝐿𝑆 : Isothermal phase change coefficient by unit surface (kg.K.s.m-4)

𝑀 : Molar mass of water (0.018 kg.mol-1)

𝑝𝑔 : Pressure of the gas phase (Pa)

𝑝𝑖 : Partial pressure of air, (i=a) and vapour (i=v) (Pa)

𝑝𝑣𝑒𝑞 : Partial pressure of water vapour at the equilibrium with water in a solution (Pa).

𝑝𝑣𝑠 (𝑇) : Saturated vapour pressure at temperature (T)

𝑝+ : Reference pressure (Pa)

𝑝0 : Standard pressure (101325 Pa)

Q : Energy flux (J.m-2.s-1)

𝑅 : Perfect gas constant (8.314 J.K-1mol-1)

𝑇 : Temperature (K)

𝑇𝑙 : Temperature of the liquid phase (K)

𝑇𝑣 : Temperature of water vapour (K)

𝑇0 : Standard temperature (273.15 K)

𝑣𝛼𝑘 : Gas phase velocity (𝛼 = 𝑔), of air (𝛼 = 𝑎) and water vapour(𝛼 = 𝑣) (m.s-1)

𝑧 : Elevation from the liquid-gas interface (m)

2
Greek symbols

𝜆𝛼 : Thermal conductivity of the phases; (𝛼 = 𝑔) for the gas phase, (𝛼 = 𝑙) for the liquid
phase (W.m-1.K-1)

𝜇𝑣 : Mass chemical potential of water vapour (J.kg-1)

𝜇𝑤 : Mass chemical potential of liquid water (J.kg-1)

𝜇+ : Reference mass chemical potential reference, function of the temperature 𝑇, at

reference pressure 𝑃+ (J.kg-1)

𝜌𝛼 : Mass density of gas phase (𝛼 = 𝑔) of air (𝛼 = 𝑎) and water vapour (𝛼 = 𝑣) (kg.m-3)

𝜀 : Correcting coefficient for evaporation rate (/)

1. Introduction

Liquid-gas phase change is involved in many mass and energy transfer. Examples include
evaporation over large bodies of water (lakes and seas) and their impact on climate, evaporation
from the soil surface associated with desertification, industrial distillation and refining
processes, high temperature geothermal processes, drying of food and industrial products.
Surface evaporation involves mass and energy transfer phenomena related to:
(i) temperature discontinuity at the interface and its variations on both sides of the interface,
(ii) convective and diffusive transport in both phases,
(iii) chemical potential discontinuity at the interface, generally associated with a difference
between the vapour partial pressure and the equilibrium vapour pressure of liquid water.

The simultaneous consideration of all these mechanisms and their coupling makes theoretical
and experimental analysis difficult. Concerning phenomena (i) and (ii), Fang and Ward [1]
performed a phase change experiment for a spherical interface separating pure water and water
vapour. Partial pressure was kept constant at values between 194.7 and 596 Pa by a flux of
vapour. The tests were carried out for temperatures of 15, 26 and 35°C. The average surface
evaporation flux varied between 70 μl/h and 160 μl/h. The temperature was measured in the
fluid and gas phases by thermocouples. The measurements showed an interfacial temperature
discontinuity up to 8°C for evaporation rates of 160 μl/h. The temperature on the vapour side
was greater than the temperature on the liquid side.
Concerning phenomena (i) and (ii), Gatapova et al. [2] brought heat locally to a thin layer of
water at atmospheric pressure. Temperature and heat flux are recorded during experiments.
Various thermocouples (4 μm diameter) allow measuring the temperature jump at the water-
air interface. A temperature jump of about 0.2°C is detected for the liquid surface temperature
of 35.2°C and the heater temperature of 43.6°C. They conclude that the value of temperature
jump depends not only on evaporation intensity but also on the temperature gradients in liquid
and gas phases near the interface.

3
Kazemi et al. [3] conducted an experimental and theoretical study on the non-stationary
evaporation of water at low pressure. They proposed a mathematical model of the system which
takes into account mass, momentum and heat transport in liquid and gas phases, as well as the
phase change at the interface. The study allowed the description of the mass transfer through
the interface. The experiments consisted in measuring the velocity field in the central plane of
the evaporating water. In addition, the temperature distribution at the liquid-vapour interface
was measured with a fine thermocouple. The results showed good correlation between
simulation and experiment. So, this model was used to reveal certain behaviours on the
phenomenon of evaporation difficult to access by experiments alone.

In the experiments proposed by Badam et al. [4], concerning phenomena (i), a cylindrical water
vapour interface is subjected to heat input. Heaters provided energy to the interface. A moving
thermocouple was used to measure the temperature profile in the liquid and vapour phases. The
water flow at the interface was determined from the quantity of water injected into the liquid
phase. These measurements were used to deduce the flow of evaporated water and the
temperature discontinuity at the interface, as well as heat flows in the liquid and gas phases
near the interface. Temperature discontinuity depended on the imposed energy; it was about
2°C without input of heat energy, and reached values up to 10°C when the temperature was
regulated at 80°C. These values are approximately 10 to 20 times higher than those given by
the kinetic theory of gases [4].

The experiments mentioned that the addition or removal of mass or heat at the interface
increases phase change. The release of latent heat of cools the two phases and leads to
temperature discontinuity at the interface. These experiments are performed in steady state, the
temperature jump is determined at the end of the experiment, which supposes that the interface
has reached an equilibrium, which is far from immediate.

In the environment, solvent, usually water, is generally impure. When the water contains salts
or other chemical compounds in solution, the activity of the water is modified as well as the
evaporation rate. Limited to aqueous saline saturated solutions, the objective of the
experimental study proposed here is to validate a macroscopic evaporation law triggered by a
chemical potential at the interface (phenomenon (iii) [4, 5, 6]. The influence of aqueous water
activity and total gas pressure on the evaporation rate will be analysed in this contribution.

2. Experimental studies of gas-liquid evaporation with low temperature

variation

The choice of a jump in chemical potential as the cause of evaporation requires carrying out an
experiment that minimizes the magnitude of mechanisms (i) and (ii); this is achieved if the
experiment is carried out in a closed environment in the absence of diffusion and if the
temperature variation at the interface is very low. However, the imposed conditions must be
able to create an interfacial discontinuity of the chemical potential. To avoid a prevailing
temperature jump at the interface and to restrict the movement of fluids, the experiments are
carried out under transient conditions. In a close cylinder containing liquid and vapour water,
the phase change will be triggered by a rapid supply of mechanical energy. This energy will be
provided by an increment of volume occupied by the gaseous phase, causing the vaporization
of a small quantity of water. Physical quantities will be recorded during the return to
4
equilibrium in order to analyse the phase change resulting from a discontinuity of the chemical
potential. However, this experimental procedure can still induce evaporation under the action
of mechanisms (i) and (ii), requiring to discuss the assumptions made.

Description of the experimental device

The experimental device, presented in Fig.1, was recently developed to characterize the liquid-
gas equilibrium in hygroscopic porous materials [7]. The same apparatus is used here to analyse
the kinetics between two equilibrium situations.

Water activity is controlled by means of standard saturated salt solutions (Tab.2). The solution
under investigation was placed in a cylindrical tank (a), it thickness being of the order of 5 mm.
A piston pump (c) was placed above the solution in order to apply gas pressure in (g) below
the atmospheric pressure. The piston chamber volume is modified by a screw system (d) with
a graduated ruler (e). O-ring gaskets (f) ensured perfect air tightness so that the system could
be considered closed.

Figure 1: Schema of the experimental device

Table 1: Dimensional characteristics of the experimental device

Solution height 15 mm

Piston diameter 44.5 mm

Piston displacement for one rotation 1.75 mm

Volume increment for one rotation 2.722 cm3

5
Some characteristic dimensions are given Tab.1. A pressure transducer (b) (Druck,
PMP4030AB) was used to record the total pressure of the gas phase, 𝑝𝑔 , throughout the process
with accuracy of ±10 Pa. The whole device was placed in a thermo-regulated bath at 30°C to
ensure a constant temperature at the outer surface of the cylinder. As the device was made of
stainless steel its high thermal inertia led to low temperature variations. Data acquisition was
performed through a National Instrument DAQ card and analysed using LabView software.

Operating mode

The tests were carried out for total pressures of 55, 75 and 95 kPa. In order to regulate both the
total pressure of the gas phase and the partial pressure of water vapour, the following procedure
was adopted.

1/ To adjust the final total pressure, preliminary experiments were carried out. The solutions
were introduced and the piston was placed in the low position. By using the pressure sensor,
the total pressure was brought to the desired value. The device was then disassembled and the
height of the cavity (g), noted ℎ𝑔 was measured. The gaseous phase being composed of 96% of
air, in the most unfavourable case of pure water, ℎ𝑔 depended little on the composition of the
solution. For the three pressures (55, 75 and 95 kPa), the average values of ℎ𝑔 were respectively:
14.22, 24.09 and 26.24 mm.

2/ As can be seen in Tab.2 the minimum equilibrium partial pressure was 500 Pa for LiCl. It
was decided to impose a partial pressure variation of 500 Pa with respect to the equilibrium
pressure, for all salts. To regulate the pressure, the piston was placed in ℎ𝑔0 and brought very
quickly to position ℎ𝑔 so that evaporation did not have time to develop. The law of perfect
gases applied to the vapour between ℎ𝑔0 and ℎ𝑔 gives:

𝑝𝑣𝑒𝑞 × ℎ𝑔0 = (𝑝𝑣𝑒𝑞 − 500)ℎ𝑔 (1)

This relation made it possible to calculate ℎ𝑔0 which takes the average values of 12.42 mm for
55 kPa, 20.87 mm for 75 kPa and 22.37 mm for 95 kPa. The couple (ℎ𝑔0 , ℎ𝑔 ) set the initial value
of the vapour partial pressure to 500 Pa and the total pressure to the desired pressure.

3/ With the device open, the piston was placed in position ℎ𝑔0 . After closing, the device was
immersed in a bath thermo-regulated at 30°C for 30 minutes. When equilibrium was reached,
the partial pressure of the vapour stabilized at the equilibrium partial pressure fixed by the
solution. The piston was moved up by the screw system. The value of the total pressure was
monitored precisely by means of the pressure transducer in order to reach the desired total
pressure. Raising the piston induced the expansion of the gas phase in the cavity (g) and
therefore decreased the partial pressure of the water vapour.

4/ The phase change took place in a closed space (g) of fixed volume. The total pressure was
recorded until a new equilibrium was obtained.

6
5/ At the final equilibrium, the device was disassembled and the value of ℎ𝑔 was measured
(Tab.3). This value differed little from the expected value, but it was more precise and was
adopted during exploitation of the experiments. The thickness of the solution was measured at
the beginning and at the end of the experiment; there was no variation in the position of the
surface during the experiment.

Presentation of the raw results

The typical changes in gas pressure during return to equilibrium are given in Fig.2 for the
various salt solutions. This figure concerns a temperature of 30°C and total gas pressure in the
order of 95 kPa. Similar evolutions were recorded for 𝑝𝑔 = 75 kPa and 55 kPa. At the end of
the experiment, the total pressure stabilized at 𝑝𝑔 (𝑡 = ∞), the vapour pressure was equal to its
𝑒𝑞
equilibrium vapour pressure, 𝑝𝑣 fixed by the solution. This allowed the partial air pressure
𝑝𝑎 (𝑡 = ∞) to be determined.

𝑝𝑎 (𝑡 = ∞) = 𝑝𝑔 (𝑡 = ∞) − 𝑝𝑣𝑒𝑞 (2)

Table 2: Water activity for various saturated salt solutions at 30°C

Salt Water activity Equilibrium vapour pressure

(/) (Pa)

Lithium chloride LiCl 0.12±0.002 504

Magnesium chloride MgN2O6 0.23±0.001 966

Potassium acetate KCH3CO2 0.55±0.001 2310

Sodium nitrite NaNO2 0.66±0.001 2772

Sodium chloride NaCl 0.751±0.001 3154

Potassium chloride KCl 0.85±0.004 3570

Potassium sulfate K2SO4 0.970±0.004 4074

Pure water 1.0 4200

7
 Figure 2: Evolution of the gas pressure pg as a function of time measured with different
saturated salt solutions (pg  95 kPa)

The average value of the pressure difference between the initial time and the end of the
experiment was (Tab.2): 420 Pa at 95 kPa, 462 Pa at 75 kPa and 490 Pa at 55 kPa, while the
predicted value was 500 Pa. Beyond the fluctuations of the experimental conditions, the
observed increase seems to show that the hypothesis of no vaporization when the piston goes
from ℎ𝑔0 to ℎ𝑔 is not entirely satisfied and that the higher the total pressure, the faster the speed.

3 Assumptions

Parasitic phenomena may occur during the return to equilibrium; it should be checked that their
order of magnitude remains low. Verification then focuses on five hypotheses:
H1- variation of the thickness of the liquid layer of water,
H2- temperature jump at the interface,
H3- vapour diffusion flux at the interface,
H4- temperature evolution of the liquid phase,
H5- vapour diffusion above solution.

H1- Variation of the thickness of the liquid slide

The value of the variation in partial pressure of the vapour is in order of 400 Pa for pure water
(Fig.2), with a thickness ℎ𝑔 of 30 mm and a piston area of 0.15×10-2 m2, the perfect gas law
gives an evaporated water mass of 0.13×10-6 kg and therefore a variation of the thickness of
the liquid phase of 0.86×10-4 mm. The hypothesis H1 of a stationary liquid - gas interface can
be adopted.

8
H2- Temperature jump at the interface

With a latent heat of 2400 kJ.kg-1, the heat necessary to evaporate the water mass is 0.156 J. It
is assumed that this heat is entirely taken from the liquid phase of thickness 5 mm and mass of
0.75×10-2 kg. With a specific heat of 4200 J.kg-1K-1, the variation of the average temperature
of the solution is -0.005 K. This temperature drop is due to evaporation; the conduction in the
metal walls and the radiation of the metal surfaces reduce the variation in temperature. The
phenomena of convection in liquid and gaseous phases have the effect of homogenizing the
temperature. The value calculated above is an upper limit of the fall in average temperature.

H3- Vapour diffusion flux at the interface

The total surface flux of material at the following interface z is the sum of the surface flux of
phase change of water and the surface flux of gases dissolved in water. In the case of nitrogen
(partial pressure 78 kPa), in equilibrium with dissolved nitrogen, with a Henry coefficient of
8.5×109 Pa.m3.mol-1, the mass of nitrogen dissolved in water in the initial state is 8.82×10-8 kg,
before the experiments. In the unfavourable hypothesis that this mass of air would be found in
gaseous form in part (g) of the device at the end, the perfect gas law would give a pressure
variation of 0.8×10-2 Pa. This is negligible in comparison with measured pressure variations
(Fig.2). The same order of magnitude is obtained with the oxygen of the air. This calculation
shows that the pressure variations of Fig.2 are entirely due to the evaporation of the water. With
the law of perfect gases applied to the mass of gas contained in (g), the vapour flux is given
by:

𝑀 𝑑
𝐽𝑣 = 𝑅𝑇 ℎ𝑔 𝑑𝑡 𝑝𝑔 (3)

H4- Temperature evolution at the liquid surface

With the vapour flow (3) it is possible to refine the temperature calculation at the interface
given in §3. The heat equation has been solved in the liquid phase of 5 mm tick, in the absence
of convection movements. The thermal conductivity is taken to be 0.65 W.m-1K-1. The initial
temperature of the liquid phase is 30°C. At the base of the liquid phase the temperature is
assumed constant and equal to 30°C. At the surface, the heat flow is imposed equal to the latent
heat of vaporization of water multiplied by the vapour flow (3), the derivative of the total
pressure being deduced from Fig.2. Based on numerical computation, Fig.3 shows the
evolutions of the average temperature of liquid phase and of the temperature at the surface of
the liquid phase. The maximum temperature variation at the interface is -0.007 K. The variation
of the average temperature of the liquid phase is -0.002 K. These values confirm the small
variations determined by the approximate calculation (-0.005 K) and help to justify hypothesis
H4 of a small temperature variation at the interface.

Table 3: Characteristic values of experiments

Total reference pressure (kPa) 55 75 95

Value of hg0 (mm) 22.37±0.05 20.87±0.05 12.42±0.05

Mean value of hg measured at the end of the 26.24±0.05 24.10±0.05 14.21±0.05

experiment (mm)

9
 Mean value of the variation of gas pressure (Fig 2) 420±10 462±10 490±10
(Pa)

Figure 3: Evolution of the average and surface temperature of the liquid phase from
numerical computation - Case of pure water

It remains to be verified that the phase change is not managed by a thermal equilibrium at the
surface. Between 28°C and 32°C, for pure water, the variation of saturating pressure variation
can be approximated by a line

𝛥𝑝𝑣𝑒𝑞 = 232.81 × 𝛥𝑇 (4)

With 𝛥𝑇 = -0.007°C, this relation gives a variation of the surface pressure of 1.7 Pa, which
represents only a small part of the variation of partial pressure, this being of the order of 400 Pa.

H5- Vapour diffusion above the solution

In order to evaluate the impact of diffusion on the phase change, the partial pressure profile of
the vapour in the cavity (g) (Fig.1) can be determined. With hypothesis H4, using the ideal gas
law, the vapour transfer in the gas phase is described by a classical diffusion equation:

𝜕𝑝𝑣 𝜕2 𝑝𝑣
− 𝐷𝑣 =0 (5)
𝜕𝑡 𝜕𝑧 2

Initially the partial pressure of the vapour is equal to the pressure imposed by the salt. The rapid
displacement of the piston (time less than 5 s) can be considered instantaneous with respect to

10
the duration of the experiment which is of the order of 1000 s (Fig.2). Water vapour from the
surface of the solution, progresses by diffusion into the gaseous phase.

The boundary conditions are:

• On the surface of the solution z=0: flow 𝐽𝑣 deduced from (3) and diffusion law, gives
the boundary condition in z=0 for 𝑝𝑣 :

𝜕𝑝 𝜕𝑝𝑔
𝐷𝑣 ( 𝜕𝑧𝑣) = −ℎ𝑔 (6)
𝑧=0 𝜕𝑡

• On the surface of the piston 𝑧 = ℎ𝑔 , the flux is zero, which according to Fick law results
in:

𝜕𝑝𝑣
=0 (7)
𝜕𝑧

The diffusion coefficient of water vapour in the air 𝐷𝑣 is given by the empirical relation [4]:
1.88
−4 𝑃0 𝑇
𝐷𝑣 = 0.217 × 10 (𝑃 ) (𝑇 ) in (m2s-1) (8)
𝑔 𝑔

Equations (5) to (8) are used to calculate the partial vapour pressure profile using the finite
differences method along an axis normal to the interface. These profiles are shown in Fig.4 in
the case of pure water.

Figure 4: Vapour partial pressure profiles from numerical computation - Case of pure water

11
After 40 s the difference in the partial pressure of vapour between the liquid surface and the
lower face of the piston represents 1% of the partial pressure of vapour. Hypothesis H4 of a
nearly uniform partial pressure profile was adopted. This implies that the partial pressure of
the air is constant throughout the experiment; it is deduced from the values in the final state by
(2).

3. Macroscopic model of evaporation at the interface during experiments

A macroscopic model of the interface can be deduced from the linear Thermodynamics of
Irreversible Processes [5, 6, 8, 9, 10]. Development of mass, momentum, energy and entropy
balance leads to expressing the production rate of entropy at the interface due to phase change.
In the dissipation, the driving force of phase change appears to be the water chemical potential
and temperature jumps across the interface. The linearization between flux and forces, in the
neighbourhood of equilibrium leads to expressing the evaporation rate 𝐽𝑣 and energy flux 𝑄 ,
positively counted in the direction: solution → gas phase:

𝐿𝑆𝑤𝑤 1 1
𝐽𝑣 = (𝜇𝑤,𝑇𝑙 − 𝜇𝑣,𝑇𝑙 ) − 𝐿𝑆𝑤𝑞 ( − ) (9)
𝑇𝑙 𝑇𝑙 𝑇𝑔

1 1
𝑄 = 𝐿𝑆𝑤𝑞 (𝜇𝑤,𝑇𝑙 − 𝜇𝑣,𝑇𝑙 ) − 𝐿𝑆𝑞𝑞 (𝑇 − 𝑇 ) (10)
𝑙 𝑔

By using the global heat balance in the vicinity of the interface, it is shown [9] that the chemical
potentials 𝜇𝑤,𝑇𝑙 and 𝜇𝑣,𝑇𝑙 are the chemical potentials evaluated at the temperature of the liquid
phase 𝐿𝑆𝑤𝑤 , 𝐿𝑆𝑞𝑞 , 𝐿𝑆𝑤𝑞 are the surface Onsager transfer coefficients. Since the TPI approach is
macroscopic, it gives no information on the structure of the phenomenological coefficients;
their knowledge is based on the experiments.

With the perfect gas hypothesis, the mass chemical potential of water vapour is [2]:
𝑅𝑇𝑙
𝜇𝑣,𝑇𝑙 = 𝜇 + (𝑇𝑙 , 𝑝+ ) + 𝑙𝑛 𝑝𝑣 (11)
𝑀

where 𝜇 + is a reference chemical potential at the temperature of the liquid phase: 𝑇𝑙 and at the
reference pressure 𝑝+ here taken equal to 1. M is the molar mass of water.

At vapour-liquid equilibrium, the partial pressure of vapour is 𝑝𝑣𝑒𝑞 . Equality of chemical

potential of liquid and water vapour leads from 11:
𝑅𝑇𝑙
𝜇𝑤,𝑇𝑙 = 𝜇 + (𝑇𝑙 , 𝑝+ ) + 𝑙𝑛 𝑝𝑣𝑒𝑞 (12)
𝑀

the relation (9) takes the form:

𝑅 𝑝 1 1
𝐽𝑣 = −𝐿𝑆𝑤𝑤 𝑀 𝑙𝑛 𝑝𝑒𝑞
𝑣
− 𝐿𝑆𝑤𝑞 (𝑇 − 𝑇 ) (13)
𝑣 𝑙 𝑔

12
The value 𝑝𝑣𝑒𝑞 differs from saturated vapour pressure at the same temperature because of the
presence of solute in the solution or the effects of interfaces in the case of multiphase medium.
The deviation of 𝑝𝑣𝑒𝑞 from 𝑝𝑣𝑠 can be expressed using the activity of water:
𝑒𝑞
𝑝𝑣
𝑎𝑤 = 𝑝 (14)
𝑣𝑠 (𝑇𝑙 )

With hypotheses H1 to H5 the liquid-gas system can be considered as the juxtaposition of two
homogeneous systems with partial pressure of water as the state variable. In the context of
these hypotheses, the evaporation law that will be used later is written
𝑅 𝑝
𝐽𝑣 = −𝐿𝑆 𝑀 𝑙𝑛 𝑝𝑒𝑞
𝑣
(15)
𝑣

The isothermal surface change coefficient is noted later 𝐿𝑆 .

4- Determination of the phase change coefficient in relation to water activity

and pressure gas phase

The following analysis aims to validate the relation (15) and to study the influence of water
activity in the solution and the total pressure of the gas phase on the surface coefficient 𝐿𝑆 .

The flux by unit surface as a function of time 𝐽𝑣 (𝑡) is derived from Fig.2 using (3). To prevent
𝑑𝑝𝑔
unphysical changes of the derivative of 𝑑𝑡 due to fluctuations in experimental points,
smoothing is carried out: the average of the experimental values at n contiguous experimental
points (𝑛 = 21) is calculated and assigned to the middle of the interval. By displacement of
the interval of n points, a smoothed experimental curve is generated. Fig.5 shows the variation
𝑑𝑝𝑔
of smoothed 𝑑𝑡 (𝑡) over a pure water, a solution of K2S04 and a solution of MgNO2 for various
pressures. These points are approximated by polynomial, logarithmic or exponential functions
𝑑
to give analytical expressions of 𝑑𝑡 𝑝𝑔 at the interface.

13
 Figure 5: Examples of the evolution of smoothed derivative of total gas pressure

The partial pressure of vapour 𝑝𝑣 in (15) is deduced by (2), at all time, using assumption H5 of
constant partial pressure of air. Fig.6 shows the variation of partial pressure of the vapour at
the interface for 55, 75 and 95 kPa, in the case of a solution of KCl (𝑎𝑤 = 0.85).

14
 e
v

Figure 6: Evolution of vapour partial pressure at the surface of a KCl solution for different
total gas pressure
𝑑
Knowing 𝑝 and 𝑝𝑣 at any time, it is possible to trace the flux as a function of the
𝑑𝑡 𝑔
𝑅 𝑝 𝑣
thermodynamic force at the interface 𝑙𝑛 𝑝𝑒𝑞 (15). The flux-force relationship is shown in
𝑀 𝑣
Fig.7 for pure water and Fig.8 for a solution of MgNO206 (𝑎𝑤 = 0.23) and for the three
pressures.
Generally, the flux-force relation relative to a dissipative phenomenon reveals a linear domain
near-equilibrium, and a non-linear domain far from equilibrium (Fig.7) [12, 13]. This is
particularly the case for chemical reactions [10] and phase change phenomena in soils [6, 14,
15], which are particular chemical reactions

The coefficient 𝐿𝑆 is deduced from the slopes of the linear parts for the imposed activities and
pressures. The value of 𝐿𝑆 are shown in Fig.9, in relation to the water activity, for the three
pressures. The values of the coefficients 𝐿𝑆 are also reported in Tab.4.

It is found that at constant pressure, the coefficient 𝐿𝑆 decreases with the activity and. For a
given activity, this coefficient increases with the pressure of the gas phase.

15
 v

/ ln v
/ vs

Figure 7: Variation of isothermal surface flux as a function of the thermodynamic force for
different total gas pressure - Case of pure water
v

/ ln v
/ vs

Figure 8: Variation of isothermal surface flux as a function of the thermodynamic force for
different total gas pressure - Case of MgN206 (aw=0.23)

16
Figures 7 and 8 provide information on the extent of the linearity domain. It is observed in
these figures, as is the case for the other experimental conditions, that the linearity range
decreases as the vaporization rate increases.

Table 4: Value of the isothermal phase change coefficient LS versus water activity for
different gas pressure

Water activity (/) Pg = 55 kPa Pg = 75 kPa Pg = 95 kPa

0.12 1.76×10-10 1.06×10-9 1.23×10-9

0.23 7.06×10-10 1.32×10-9 1.59×10-9

0.55 1.59×10-9 4.23×10-9 4.59×10-9

0.66 2.38×10-9 5.12×10-9 5.47×10-9

0.75 3.26×10-9 7.76×10-9

0.85 4.23×10-9 7.39×10-9 8.47×10-9

0.97 4.59×10-9 7.41×10-9

1.00 5.29×10-9 8.64×10-9 1.15×10-8

Figure 9: Variation of the isothermal surface phase change coefficient LS as a function of

water activity for different total gas pressure

17
By definition, the mass chemical potential of water represents the variation of internal energy
of the liquid phase when a unit mass of water is extracted. It reports on the binding energy of
water with its environment. When activity, of liquid water tends to zero, the chemical potential
of water (12) approaches −∞; this implies the need to apply a large amount of energy to remove
molecules of water. The high value of the water binding in the solution makes the phenomena
of evaporation very slow; the phase change coefficient approaches zero. When 𝑎𝑤 increases,
the binding of water molecules decreases and the efficiency of the phenomena of evaporation
increases. This results in an increase of the coefficient 𝐿𝑆 when water activity increases.

The effect of total pressure of the gas phase can be explained using a molecular description [8,
16]. For these authors three mechanisms are involved in the phase change process (Fig.10):
evaporation and condensation of a water molecule, expulsion of a water molecule by a
molecule of the gas phase (molecular exchange).

Figure 10: Different evaporation mechanisms of a molecule of water [16]

In the experiments, the gas phase was mainly composed of air; the ratio between vapour
pressure of water and air was of the order of a few percent. When the air pressure decreased
the number of collisions between air molecules and molecules constituting the interface also
decreased, evaporation by molecular exchange was attenuated resulting in a decrease of
coefficient 𝐿𝑆 .

6. Global validation experiment

Analysis of the experimental results is dependent on the hypotheses proposed and discussed in
section 3. Although these are well verified numerically, an experimental study is not immune
to a parasitic phenomenon. For this reason, a global experimental verification study is
proposed. His goal is to verify if the relation (15) and the coefficient 𝐿𝑆 are able to predict the
evolution of pressure during an experiment totally different from those at the origin of their
study. During the validation experiment, the dimensions of the measuring chamber, the method
for measuring the partial vapour pressure, and the method for lowering the partial vapour
pressure at the initial state were modified.

18
 Figure 11: Schema of the experimental device used for validation

Experimental verification was performed on a device shown in Fig.11. This device consisted
of a cylindrical sealed enclosure of 10 cm in diameter and 10 cm high. At its base a saturated
saline solution was introduced, at the top a Hobo sensor was placed which directly gave the
relative humidity of the air by measuring the capacity of a capacitive element absorbing
moisture. The manufacturer gave accuracy of 5% in the measuring range, between 25% and
90%. On the other hand, the balance between the sensing element and the surrounding
atmosphere was not immediate, introducing a delay in measurement.

The processing of an experiment includes the following steps:

1/ At the initial time, the saline solution is placed in the mould. The height of the part occupied
by the air is measured.

2/ Ventilation occurred on the surface, resulting in a change in the relative humidity of the gas
phase to reach ambient humidity (of the order of 50%). This technique only works for saline
solutions that fix water activity above 50%: NaCl (𝑎𝑤 = 0.75) and NaNO2 (𝑎𝑤 = 0.66).

3/ The mould is then sealed; recording of the relative humidity by the Hobo sensor begins. For
the reasons just mentioned, the experiment is reliable only for two salts. Experimental
evolutions are shown in Fig.12.

4/ The theoretical model used differs from the model presented in H5 by the boundary condition
at the interface; the flux is calculated by (15) with the coefficients of Fig.9; thus, all the results
of the study are used. The model gives the evolution of the relative humidity of the air Fig.12.
For NaCl, this figure highlights the slowness of the sensor to reach the value of 0.75 imposed
by the salt.

19
5/ The evolution of air humidity given by the sensor is compared to the theoretical evolution
(Fig.12). In spite of a divergence, largely imputed to the capacitive sensor, there is a general
similarity of the curves and a similar equilibrium return time.

Figure 12: Comparison between the evolutions of relative humidity obtained from numerical
computation and measured by a capacitive sensor for 2 different salt solutions

6. Conclusion

The experimental device isolates the phase change under the action of a chemical potential
discontinuity. It provides further insight compared to experiments introducing a heat input [4]
or a renewal of vapour at the interface [1]. Vaporization is produced by a sufficiently fast
mechanical energy supply to be assumed to be an impulse allowing to categorize this method
as a transient method. The advantages are the short duration of the measurement and the low
impact on the global physical parameters, like temperature or liquid activity. The recording of
the total gas pressure makes it possible to analyse the return to equilibrium and measure the
phase change rate. In the proposed study, the phase change generated by the discontinuity of
the water chemical potential is the predominant phenomenon. The hypotheses used to exploit
the experiment were validated numerically from the experimental results. Additionally, a
global validation experiment confirms all the hypotheses. The short duration of experiments
makes it possible to analyse the influence of various parameters, such as the temperature, the
gas pressure and the liquid activity.

20
The decrease in phase change coefficient 𝐿𝑆 with activity reveals an increase in the retention
of water molecules by the solution. The increase in 𝐿𝑆 with the pressure reveals an increase in
the molecular exchanges which is predominant. A finer analysis of the 𝐿𝑆 coefficient is
meaningless with the macroscopic approach used here, only a microscopic analysis would go
further. Given the precautions taken to minimize any phenomena other than the discontinuity
of the chemical potential, the experimental results should be used as an experimental basis for
validating microscopic models.

In complex tri-phasic media: soils, food products, biological tissues ..., the activity of water is
modified by liquid-gas and liquid-solid interfaces. For instance, the relationship between water
content and activity is accurately given by the desorption isotherm. The evaporation at the
surface of these media is translated by a relationship similar to (15) [5, 6]. The experimental
technique presented here was used in the case of trichloroethylene [17] to analyse the phase
change behaviour at the surface of a soil. An experiment different from the one presented here,
has shown that the law (15) is verified [6]. The exploitation of these experiments confirms that
the vaporization rate increases with the water activity and the total gas pressure.

Acknowledgments

This research was supported by the University of Montpellier, the Centre National de la
Recherche Scientifique (France), the Joseph Ki-Zerbo University, Ouagadougou and the Nazi
Boni University, Bobo Dioulasso (Burkina-Faso).

Pr. Jean-Claude Bénet deceased on July 27th 2019 while he was finishing this manuscript. This
publication is an homage to its extensive contribution on this subject.

References

[1] G. Fang, C.A. Ward, Temperature measured close to the interface of an evaporating liquid,
Physical Review 59 (1999) 417-428.

[2] E.Y. Gatapova, I. Graur, O.A. Kabov, V.A. Aniskin, M.A. Filipenko, F. Sharipov, The
temperature jump at water-air interface during evaporation, International Journal of Heat and
Mass Transfer 104 (2017) 800-812.

[3] M. A. Kazemi, D. S. Nobes, and J. A. W. Elliott, Experimental and Numerical Study of the
Evaporation of Water at Low Pressures, Langmuir 33 (2017) 4578-4591.

[4] V.K. Badam, V. Kumar F. Durst, K. Danov, Experimental and theoretical

investigations on interfacial temperature jumps during evaporation, Experimental Thermal
and Fluid Science 32 (2007) 276-292.

[5] J.-C. Bénet, P. Jouanna, Non equilibre thermodynamique dans les milieux poreux non
saturés avec changement de phase, International Journal of Heat and Mass Transfer 26 (1983)
1585-1595.
21
[6] J.-C. Bénet, A.-L. Lozano, F. Cherblanc, B. Cousin, Phase Change of Water in a
Hygroscopic Porous Medium - Phenomenological Relation and Experimental Analysis for Soil
Water, Journal of Non-Equilibrium Thermodynamics 34 (2009) 133-153.

[7] S. Ouoba, F. Cherblanc, B. Cousin, J.-C. Bénet, A new experimental method to determine
the sorption isotherm of a Liquid in a porous medium, Environmental Science & Technology
44 (2010) 5914-5919.

[8] D.A. De Vries, A.J. Kruger, On the value of the diffusion coefficient of water vapour in air,
Colloque international du CNRS n°160, Paris 18-20 avril 1966, Editions du CNRS (1967) 61-
72.

[9] D. Bedeaux, S. Kjelstrup, Transfer coefficients for evaporation, Physica A 270 (1999) 413-
426.

[10] D. Bedeaux, S. Kjelstrup, Irreversible thermodynamics - a tool to describe phase

transitions far from global equilibrium, Chemical Engineering Science 59 (2004) 109-118.

[11] E.A. Guggenheim, Thermodynamique, Dunod, Paris (1965).

[12] I. Prigogine, D. Kondepudi, Thermodynamique, Odile Jacob, Paris (1999).

[13] C. Vidal, G. Dewel, P. Borckmans, Au-delà de l’équilibre, Hermann Paris, 1993.

[14] A.-L. Lozano, F. Cherblanc, B. Cousin, J.-C. Bénet, Experimental study and modelling of
the water phase change kinetics in soils, European Journal of Soil Science 59 (2008) 939-949.

[15] A.-L. Lozano, F. Cherblanc, J.-C. Bénet, Water Evaporation Versus Condensation in a
hygroscopic soil, Transport in Porous Media 80 (2009) 209-222.

[16] M. Matsumoto, Molecular dynamics of fluid phase change, Fluid Phase Equilibria 144
(1998) 307-314.

[17] S. Ouoba, T. Daho, F. Cherblanc, J. Koulidiati, J-C. Bénet, A new experimental method to
determine the evaporation coefficient of trichloroethylene (TCE) in an arid soil, Transport in
Porous Media 106 (2015) 339-353.

22