A phenomenological theory of polycomponent interactions

in non-ideal mixtures. Application to NH

3
/H
2
O and
other working pairs
M.D. Staicovici
Thermopower Equipment Research and Design Institute (SC ICPET- CERCETARE SA) 236 Vitan Road, 74369, Bucharest, Romania
Received 23 October 1998; received in revised form 23 June 1999; accepted 24 June 1999
Abstract
The paper proposes an original linear phenomenological theory (Ph T) of evolution physical mono-, bi- and
particular polycomponent gasliquid interactions with non-ideal mixture. The expressions of the phenomenological
factors (entropy source, force, coecient and coupled heat and mass transfer currents) are deduced. The theory is
particularized to the NH
3
/H
2
O and other gasliquid systems used in the thermal absorption technology. The work's
conclusions are listed next. The paper raises the problem of ammonia bubble absorption which is dicult to answer with
current theory of interface mass transfer and absorption as a surface phenomenon. The heat and mass transfer at the
gasliquid interface is governed by the thermodynamic force, which applies also to solidliquid, solidgas, or liquid
liquid, gasgas type interactions and continuous or discontinuous media. The paper mentions a postulate referring to
the force behavior approaching an ideal point, previously formulated by the author. According to its consequence, the
mass and heat currents suer an ideal point approaching (i.p.a.) eect, not mentioned so far in the specialized literature,
consisting in a continuous increase of their absolute value by several percent (for a pure component), to several hun-
dred times (for a binary system) when the interacting system approaches an ideal state, as compared to the values of
states which are far from the same ideal point. In this way, ``far from equilibrium'' becomes synonymous to ``low
interaction''. The classic assessment of the interface mass transfer by analogy with heat transfer lacks basic physics.
The (Ph T) satisfactorily explains the problem of ammonia bubble absorption. Absorption is a mass phenomenon, not
a surface one. An intensive way of improving absorption is emphasized, which seeks to promote the i.p.a. eect
appearance rather than the extensive way currently used, based on increasing gasliquid interaction area. To this
extent, the bubble absorber is hereby proposed for ecient absorption. The i.p.a. eect existence oers an additional
chance for a satisfactory explanation of the Marangoni eect. #2000 Elsevier Science Ltd and IIR. All rights reserved.
Keywords: Refrigerating system; Absorption system; Ammonia/water; Absorption; Eciency
Interactions entre les composants de me langes non-ide aux :
the orie et application au uide actif NH
3
/ H
2
O et a d'autres
uides actifs
Re sume
Cet article propose une approche lineaire selon l'evolution des interactions gazliquide dans un melange non-ideal ou ce
melange comporte 2, 3 ou davantage de composants, selon les phenomenes presents. L'auteur donne les expressions
couvrant facteurs impliques (source d'entropie, force, coecient de transferts de chaleur et de masse). Le uide actif NH
3
/
H
2
O est utilise comme exemple principal et d'autres systemes gazliquide sont utilises pour illustrer la technologie
0140-7007/00/$20.00 # 2000 Elsevier Science Ltd and IIR. All rights reserved.
PI I : S0140- 7007( 99) 00034- 1
International Journal of Refrigeration 23 (2000) 153167
www.elsevier.com/locate/ijrefrig
d'absorption thermique. Les conclusions sont fournies. Un probleme souleve dans cette communication est celui de l'ab-
sorption de la bulle d'ammoniac : il est dicile a resoudre ce probleme dans l'etat actuel de la theorie sur l'interface entre le
transfert de masse et l'absorption. Le transfert de chaleur et de massea l'interface gaz-liquide est gouverne par la force
thermodynamique qui s'applique egalement aux interactions solidesolide, solidegaz, liquideliquide ou gazgaz en mili-
eux continu ou discontinu. La communication mentionne egalement un postulat sur le comportement des forces a proximite
d'un point ideal, deja expose par l'auteur. Selon ce postulat, les ux de chaleur et de masse subissent un certain eet en
s'approchant du point ideal (i.p.a.) ; ce point n'a pas ete evoque dans la litterature. Ce point implique une augmentation
constante des valeurs absolues, variant d'un faible pourcentage (pour un composant unique)a plusieurs centaines de fois
pour un systeme binaire dans lequel le systeme est proche d'un etat ideal, ce qui n'est pas le cas des systemes loin de ce
point ideal. De cette maniere, l'expression loin de l'equilibre devient synonyme de faible interaction . Les principes
de physique de base ne tiennent pas compte de l'evaluation classique du transfert de masse a l'interface par rapport au
transfert de chaleur. L'approche phenomenologique fournit une explication pour l'absorption de la bulle d'ammoniac.
L'absorption est un phenomene de masse, pas un phenomene de surface. Une fac on d'ameliorer l'absorption de fac on
marquee est soulignee ; cette approche est destinee a promouvoir l'eet i.p.a. et non d'utiliser la methode actuellement
utilisee qui est fondee sur l'augmentation de la supercie d'interaction entre le gaz et le liquide. Un absorbeur de bulle est
propose comme methode d'absorption ecace. L'eet i.p.a. donne une explication supplementaire pour l'eet de Mar-
angoni. # 2000 Elsevier Science Ltd and IIR. All rights reserved.
Mots cles: Syste me frigorique ; Syste me a absorption ; Ammoniac / eau ; Absorption ; Ecacite
Nomenclature
A reduced excess heat
c
p
specic heat at constant pressure (kJ/kmol K)
C local thermal capacity (kJ/kmol)
f gure of merit quantifying the increase of a
phenomenological function approaching an
ideal state, as compared to its far equilibrium
states
h specic enthalpy (kJ/kmol)
j interface current (mass, heat, etc.) (mol/s, kg/
s, W)
k global mass transfer coecient (kg/m
2
s)
m bubble feeding mass ow rate (kg/s)
M match function intervening in the generalized
thermodynamic force equation
L
i,j
phenomenological coecient associated with
the coupled currents i and j (mol
2
K/Js, kgKs/
m
2
)
N mass ux (kg/m
2
s)
Nu Nusselt number
p partial pressure (kPa, bar, ata)
Pr Prandtl number
R gas constant (kJ/kmolK)
Re Reynolds number
s specic entropy (kJ/kmolK)
Sc Schmidt number
Sh Sherwood number
S
.
entropy source (W/K)
t time (s)
T,t temperature (K,

Celsius)
U total internal energy (J)
V volume, molar volume (m
3
, m
3
/kmol)
x liquid phase molar fraction (mol/mol)
x
+
liquid phase molar fraction in equilibrium
with the gas phase molar fraction y (mol/mol)
X generalized thermodynamic force (kJ/kmolK)
y gas phase molar fraction (mol/mol)
y
+
gas phase molar fraction in equilibrium with
the liquid phase molar fraction x (mol/mol)
Greek symbols
d nite dierence (variation)
L driving force (according to classical inter-
pretation)
A
X
exergy destruction velocity (W)
" chemical potential (kJ/kmol)
' supercial tension (N/m)
Subscripts
abs absorption
c condensation
e equilibrium
g gas, gaseous phase
gen generation
i mass
q heat
l liquid, liquid phase
max maximum
p constant pressure, pure
ph phenomenological
r reduced
T constant temperature
v vaporization
/
before interaction
//
after interaction
0 environmental
1,2 mixture species
154 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
1. Introduction
The NH
3
/H
2
O system is the most common working
combination in the thermal absorption technology.
Although there is quite a comprehensive theoretical and
practical experience in the use of such combination, the
gasliquid interaction in this medium is, most surpris-
ingly, still dealt with empirically. The analogy with heat
transfer is commonly used for the calculation of dis-
continuous media (interface) mass transfer,
N = kL (1)
In Eq.(1), N is the mass ow transferred as a result of
the action of the motive force L. The motive force is
equal to the deviation from the equilibrium conditions.
Proportionality factor k is the mass transfer coecient.
According to the known concept ``of the two lms'' [1],
N is proportional to the dierence of the concentrations
of the gas or liquid phases and to the global mass
transfer coecients of the phases
N = K
g
y y
+
( ) = K
l
x x
+
( ) (2)
In Eq.(2), x* and y* are the concentrations of the liquid
and gas phases in equilibrium with the gas phase of
concentration y and the liquid phase of concentration x,
respectively. Coecients K
g
and K
l
are expressed as
functions of the global resistance to mass transfer,
which have additive properties to the partial resistance
of the phases, similarly to the heat transfer. More often
than not it is considered that mass transfer is controlled
by the phase with the greatest partial resistance. In
order to decrease global resistance, the dispersion of the
process controlling phase is usually used. That is why,
according to the classic interpretation, it is implicitly
admitted that the NH
3
/H
2
O gasliquid interaction is a
surface phenomenon.
The literature referring to mass transfer has mainly
dealt with the dependence of transfer coecients on
molecular diusivity. This approach includes ``the two
lms'' concept referred to above, as well as the theory of
penetration [2], and renewal [3]. Specialized literature
has paid less attention to the relation between absorp-
tion rate and hydrodynamic conditions. Kropholler and
Carr [4], Chilton and Colburn [4] and Ciborowski and
Richlicki [5], have proved experimentally that in certain
given conditions there is a Reynolds analogy in the case
of mass transfer
Sh = a
m
Re
b
m
Sc
c
m
(3)
and
Nu = a
h
Re
b
h
Pr
c
h
(4)
in the case of heat transfer, in two-phase gas, liquid ow
in pipes. Several authors have studied heat and mass
transfer in the NH
3
/H
2
O system, [511]. Most of them
express their results by means of correlations similar to
Eqs. (3) and (4). A detailed study was conducted by
Keizer [12], for two-phase NH
3
/H
2
O ow of the slug
type in vertical pipe absorbers. For the purpose of eval-
uating mass transfer, he made use of prediction correla-
tions of partial coecient k
l
, which are due to Banerjee
et al. [13], Gregory and Scott [14], Jepsen [15], Jagota
[16], Kulic and Rhodes [17], Kasturi and Stepanek [18]
and others. Comparing his experimental data with the
ones predicted and nding that they sensibly dier,
sometimes even by an order of magnitude, Keizer places
under a question mark the applicability of the correla-
tions of the authors mentioned above to the prediction
of mass transfer in the case of the NH
3
/H
2
O system and
comes to the conclusion that it cannot be precisely
determined which partial resistance, of the liquid phase
or of the gas phase, respectively, controls the NH
3
/H
2
O
absorption process. Finally his correlations are satisfac-
tory by adapting Eq. (3).
In an attempt to nd a more accurate method of pre-
diction of NH
3
/H
2
O gasliquid interactions, this paper
will take into consideration the non-equilibrium ther-
modynamics [19]. Although it seems to be a natural
direction of investigation, it has not been used so far to
analyse neither the NH
3
/H
2
O system nor other working
combinations.
Herein below a simple experiment made by the
author, which strengthens the conviction that the
method of approaching the interactions which form the
object of this study should be reviewed, will be descri-
bed. The experiment consists of injecting a single gas-
eous ammonia bubble into the NH
3
/H
2
O solution at
normal pressure, ambient temperature and small molar
fractions. Based upon the numerous observations
recorded, it has been found out that subject to the gas
phase feeding rate the evolution of the bubble may be
characterized as follows: (i) at low and moderate feeding
rates (10
8
10
7
kg/s), the dynamics of the bubble takes
place in two distinct stages: (a) growth, when the bubble
progressively expands its volume up to a maximum
value, concurrently with gas absorption, followed by (b)
collapse, when the bubble is rapidly absorbed, reducing
its volume up to zero, although it continues to be fed
with gas through the injection nozzle; throughout its
evolution the bubble does not detach itself from the
nozzle and has an almost spherical shape; in the case of
a continuous gas feeding, the phenomenon becomes
quasi-periodical; the frequency of complete evolution
increases with the increase of the ow rate; the duration
of the increase stage is much longer than the one of the
collapse stage; (ii) at high ow rates, collapse no longer
occurs, but a high frequency phenomenon of oscillation
of the bubble volume between two values, a maximum
and minimum one, is noticed; the bubble is no longer
spherical but elongates in the direction of the nozzle axis
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 155
and has an irregular contour. The study of the bubble in
its complete growthcollapse evolution gives rise to two
questions which are dicult to answer by means of the
current theory of interface mass transfer and absorption
as a surface phenomenon, synthesized by Eqs. (1) and
(2) (the problem of the ammonia bubble absorption, [20]):
(i) if during the growth phase, as the bubble increases its
volume, the molar fraction x of the interface approaches
the saturation point so that the dierence in the con-
centration of phases (x
+
x) and implicitly, the current
absorbed decrease, then how could the occurrence of the
collapse phase, which, quite to the contrary, assumes a
continuous increase of the current absorbed in order to
exceed the relatively constant value of the feeding ow
rate, be explained?; (ii) if absorption is considered to be a
surface phenomenon, then, yet again, how could collapse,
during which the current absorbed, proportional to the
gasliquid contact area, decreases by the radius second
power, instead of increasing as noted in the rst ques-
tion above, be explained? As will be demonstrated fur-
ther, these questions may be accurately answered by
applying the non-equilibrium thermodynamics (i.e.
phenomenological equations) to the interface of the
subject interactions. Its application brings to light an
important feature of the mass and heat currents beha-
vior near an ideal (equilibrium) point, which the specia-
lized literature does not point out, namely the ideal
point approaching eect. Previous author's works use
the phenomenological theory to model absorption of
the ammonia bubble absorption, having as a tool a non-
empirical linear Phenomenological Hydro-GasoDyna-
mical (PhHGD) approach. The nal part of the paper
presents some results of this modeling, in order to elu-
cidate from both qualitative and quantitative points of
view the problem of the ammonia bubble absorption,
and shows the conclusions deriving from it.
2. Phenomenological equations of evolution physical
mono-, bi- and particular polycomponent interactions
with non-ideal mixture
In the general case, interacting binary two-phase evo-
lution phenomena, of gasliquid, solidliquid, solidgas
or gasgas, liquidliquid type, may be considered. For
the sake of clarity, here we shall use the notation for the
physical interactions of gasliquid type. The liquid
phase may additionally contain other dissolved non-
volatile components. The state parameters characteriz-
ing the interface, i.e. temperature, pressure and molar
fraction, marked as T
g
Y p
g
and y for gas and as T
l
Y p
l
and
x for liquid, are time functions. At interface, surface
tension forces, given by Laplace equation, written in a
new wording [21], by equality
Lp = div '

(5)
where '

is a normal positive vector to the surface and

equal to the surface tension of the gasliquid, are gen-
erated. The theoretical approach holds for both open or
closed systems. It evolves towards the state of equili-
brium through a nite sequence of stationary states. It is
equally assumed that the total energy of the system is
approximately equal to its internal energy, that is its
macroscopic kinetic energy, the external elds of any
nature which act on it as well as its internal tensions are
negligible. Under these circumstances, when the
mechanical work exchanged by the system with the
exterior is negligible, the heat exchanged between the
system and its environment and between its various
areas is sensibly equal to the total enthalpy exchange.
The gasliquid interaction implies the simultaneous
existence of coupled mass and heat currents ``j'', noted
by indices ``i'' and ``q'', respectively, of gas and liquid,
marked by indices ``g'' and, ``l'', respectively, which get
in and out of the system, marked as prime (`) and sec-
ondary (``), respectively.
The entropy source, S
X
, shows the contribution of the
currents to the global entropy yield. In a stationary
regime this is written [22]
S
X
=
A
T
o
=

j
i
/
s
i
/

j
i
//
s
i
//
_ _

j
q
/
T
/

j
//
q
T
//
_ _
(6)
where A
X
is the exergy dissipation velocity and T
o
is the
temperature of the innite reservoir. Dimensionally j
i
/
and j
q
/
are expressed in mol/s or kg/s and in W, respec-
tively. Eq.(6) is particularized to two opposed elemen-
tary processes which are permanently encountered in the
heat absorption technology: (i) the gas interacts with the
liquid, resulting in a liquid with new properties
(absorption); (ii) the liquid decomposes in liquid and gas
with new properties (generation). Taking the above into
account as well as the common conventional signs in
thermodynamics, Eq. (6) can be rewritten for case (i) as
S
X
= j
iYg
/
s
g
/
j
iYl
/
s
l
/
j
iYl
//
s
l
//
_ _
j
qYg
/

1
T
/
g
j
qYl
/

1
T
l
/
j
qYl
//
j
q
_ _

1
T
l
//
_ _
(7)
where j
q
is a heat current, considered as an excess ele-
ment resulting from the deviation from the ideal state of
the mixture. The following currents balances are con-
sidered in the system
process heat
j
qYg
/
j
qYl
/
= j
qYl
//
j
q
(8)
total mass
j
iYg
/
j
iYl
/
= j
iYl
//
(9)
156 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
mass for one of the species
yj
iYg
/
xj
iYl
/
= x x ( )j
iYl
//
(10)
In Eq. (10) x is the nite variation of molar fraction,
suered by the liquid as a result of its interaction with
the gas phase. The heat ows are extensive elements.
Taking into account the previous observation concern-
ing the calculation of the heat exchange by means of
total enthalpy, they can be written as
j
qYg
/
= j
iYg
/
h
/
g
j
qYl
/
= j
iYl
/
h
/
l
j
qYl
//
= j
iYl
//
h
//
l
(11)
Taking into consideration Eqs (9) and (10), the rst
bracket of Eq. (7) can be written as
j
iYg
s
/
g
j
iYl
/
s
/
l
j
iYl
//
s
//
l
_ _
=
j
iYg
/
s
/
g
s
/
l
_ _
y x ( )
s
//
l
s
/
l
x
_ _
(12)
It is noted that the rst bracket of Eq. (7) represents the
contribution of mass currents to the entropy source
only, therefore the ratio in the right side of Eq. (12)
becomes
s
//
l
s
/
l
x

ds
l
dx
_ _
pYT
(13)
Temperature T
l
//
is related to the nite variation of
liquid temperature due to the deviation from the ideal
state T = T
l
//
T
l
/
by
1
T
l
//
=
1
T
l
1
T
T
l
/
_ _ =
1
T
l
/
1
T
T
l
/

T
T
l
/
_ _
2
. . .
_ _
(14)
Following simple calculations by means of Eqs. (14) and
(8)(11) the second bracket of Eq.(7) may be written as
j
qYg
/
1
T
g
/
j
qYl
//
1
T
l
/
j
qYl
//
j
q
_ _
1
T
l
//
_ _
=
j
iYg
/
1
T
g
/

1
T
l
/
_ _
h
/
g

1
T
l
/
h
/
g
h
/
l

h
/
l
j
r
_ _
T
T
l
/
1
T
T
l
/
_ _
_

_
_

_
(15)
where
j
r
=
j
iYg
/
j
iYl
//
(16)
is the reduced mass current of gas (mol mol
1
), involved
in the interaction. The ratio
T
T
l
/
is calculated by means of
the liquid specic heat equation at constant pressure
c
pYl
=
dh
l
dT
_ _
pYx
(17)
resulting in
T
T
l
/
=
h
/
l
T
l
/
c
pYl
= Aj
r
(18)
where the non-dimensional factor A in Eq. (18), herein
called ``reduced excess heat'' is obtained with Eqs. (9),
(10) and (16) and has the expression
A = y x ( )
dh
/
l
dx
_ _
pYT

1
T
l
/
c
pYl
(19)
Eqs. (12)(19) are introduced in Eq. (7) and the entropy
source is written as
S
X
= j
iYg
/
X ( ) (20)
where
X =
_
s
/
g
s
/
l
_ _
y x ( )
ds
/
l
dx
_ _
pYT
_ _
h
/
g
1
T
g
/

1
T
l
/
_ _

A
T
l
/
1 Aj
r
( )
h
/
g
h
/
l
_ _
j
r
h
/
l
_ _
_
(21)
Eq. (20), written as above, points out the proportion-
ality of the entropy source with the coupling gas mass
current j
iYg
/
and with factor X ( ), which is the thermo-
dynamic (generalized) force of the irreversible process at
hand [22,23]. The force diers of forces governing the
coupled heat and mass transfer in continuous media and
originating from Fourier's and Fick's laws, respectively.
It only depends on the entry values of the system, for
which reason the marking primary (
/
) is hereinafter
given up to. Because it depends only of the interface p,
T, x and y state parameters, force is a way independent
function, therefore a state function. According to Eqs.
(8)(11) the coupled heat current j
q
has the expression
j
q
= j
iYg
/
C (22)
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 157
where C is the local thermal capacity (availability) of the
binary mixture [24]
C = h
/
g
h
/
l
_ _
y x ( )
dh
/
l
dx
_ _
pYT
(23)
This paper is conned to considering that the currents
involved are linear functions of the thermodynamic for-
ces. Later on, this consideration will be justied. Hence,
the phenomenological equations are reduced to the
simplest linear expression
j
iYg
/
= L
11
X ( ) (24)
where L
11
is the strictly positive phenomenological
coecient of proportionality, having the dimension of
mol
2
KJ
1
s
1
. In a rst approximation L
11
is con-
sidered to be constant, independent of X. With Eq. (24),
the entropy source, [Eq. (20)], acquires the denite
positive form
S
X
= L
11
X
2
(25)
for all interface p, T, x and y values, except the ideal
points, given by
p
l
= p
g
= p
e
T
l
= T
g
= T
e
"
1 2 ( )Yl
= "
1 2 ( )Yg
where, according to current opinion it vanishes [23].
Later, the paper will refute this last statement. It is use-
ful to further proceed to non-dimensioning the equa-
tions referred to above. In order to do this, the
following reduced elements are considered
L
11Yr
=
L
11
j
iYl
//
R
_ _; X
r
=
X
R
; S
X
r
=
S
X
Rj
iYl
//
h
r
=
h
T
b
R
; c
pYr
= c
p
aR; s
r
= saR
p
r
=
p
p
b
; T
r
=
T
T
b
(26)
In Eq. (26) R, p
b
and T
b
are the gas constant, a reduc-
tion pressure and, respectively, an absolute reduction
temperature. In view of the above, the previously
deduced equations can be re-written in a reduced form
(index ``r'') as follows
S
X
r
= j
r
X ( ) = L
11Yr
X
2
r
(27)
j
r
= L
11Yr
X
r
( ) (28)
X
r
= s
gYr
s
lYr
_ _
y x ( )
ds
lYr
dx
_ _
p
r
T
r
_ _
h
gYr
1
T
gYr

1
T
lYr
_ _

A
r
T
lYr
1 A
r
j
r
( )
h
gYr
h
lYr
_ _
j
r
h
lYr
_ _
(29)
A
r
=
dh
lYr
dx
_ _
p
r
T
r
y x
T
lYr
c
pYlYr
(30)
j
qYr
= j
r
C
r
(31)
where
C
r
= h
gYr
h
lYr
_ _
y x ( )
dh
lYr
dx
_ _
p
r
T
r
(32)
The partial reduced thermodynamic force of a purely
mass nature, given by the rst square bracket of the
right side of Eq. (29) is marked as X
iYr
X
iYr
= s
gYr
s
lYr
_ _
y x ( )
ds
lYr
dx
_ _
p
r
T
r
(33)
We have to remark here the similitude of Eqs. (32) and
(33).
Bearing in mind that L
11Yr
must be in a rst approx-
imation independent of X
r
, from Eqs. (28) and (29) the
reduced thermodynamic force is expressed in an implicit
form as [20,25,26]
X
r
=
X
iYr
h
gYr
1
T
gYr

1
T
lYr
_ _

A
r
h
lYr
T
lYr
1 A
r
L
11Yr
X
r
( )
_ _
1
A
r
h
gYr
h
lYr
_ _
T
lYr
1 A
r
L
11Yr
X
r
( )
_ _ M p
rYe
Y x
e
_ _
(34)
where M p
rYe
Y x
e
( ) is a new function discussed later in
Section 3. It may be derived rather solving Eq. (34) with
respect to X
r
( ) by means of a simple iterative method
[27] (which the author preferred), but also as a second
degree algebraic equation [26]. Force is useful in
approaching interaction mass and heat currents of
mono-, bi- and particular poly-component gasliquid
interactions with non-ideal mixture, coupling it with
Eqs. (28) and (31) and the classic hydro-gasodynamics.
For instance, the two last types of interactions are
encountered in the binary systems NH
3
/H
2
O, H
2
O/
LiBr, NH
3
/LiNO
3
, NH
3
/NaSCN. CH
3
NH
2
/H
2
O etc.,
or in the ternary systems such as NH
3
/H
2
O-LiBr,
CH
3
NH
2
/H
2
O-LiBr etc. The resulting Phenomen-
ological Hydro-GasoDynamical (PhHGD) method was
already successfully used to model a few ammonia/
water absorption/generation applications [20,25,28,29].
158 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
The thermodynamic force enjoys several important
properties [26]. First, as already mentioned, force
applies also to interactions of solidliquid, solidgas or
gasgas, liquidliquid type, provided that appropriate
indexes in Eq. (34) be used (e.g. for a liquidsolid inter-
action with a solid nal phase, the indices g and l change
in l and s, respectively). Force is valid for both con-
tinuous or discontinuous media. Obviously, the currents
assessment needs to analytically determine the right
phenomenological coecients, which often is a dicult
task. Second, force must be not dened in the ideal
points, because otherwise it should take at the same time
and with the same probability two opposite values, cor-
responding to the elementary processes at hand, with
opposite signs of the homologue currents, which is
impossible. This property extends to the currents and to
the entropy source, not dened in an ideal point as well.
Concerning the entropy source, our conclusion contra-
dicts the current opinion, previously mentioned in this
paragraph. The set of ideal points separates the strictly
positive thermodynamic forces ( X
r
b 0), character-
ized by absorption currents, j
r
b 0, from the strictly
negative forces, ( X
r
` 0), which fact gives rise to
generating currents, j
r
` 0. The third property is related
to the force behavior near an ideal point. So far, the
specialized literature does not point out this aspect and
refers only to the total entropy increase to maximum
values when the system approaches such a point, caus-
ing the entropy source strictly positive denition. In this
work it is considered that real systems naturally making
for equilibrium obey the following postulate [26]: The
generalized thermodynamic forces of closed and open
systems continuously increase in absolute value to max-
imal nite values approaching an ideal point. Here are
included the systems with physical or chemical interac-
tions in the rst place. The phenomenological factors,
proportional to the force, must have the same force
feature.
The case of pure component interactions is important
and worth to paying a little bit more attention to. The
Eq. (34) holds here too, simply considering that x and y
take close values to 0, or 1. However, the reduced excess
heat A
r
and the term y x ( )

ds
lYr
dx

p
r
YT
r
in Eq. (33) are
vanishing when x and y approach 0 or 1, so that Eq.
(34) simplies in case of pure components, namely
X
rYp
= s
gYr
s
lYr
_ _
h
gYr
1
T
gYr

1
T
lYr
_ _
(35)
In Eq. (35), the two terms of the right member have
dierent signs, but X
rYp
is strictly negative and holds
for vaporization
X
rYpY)
= X
rYp
j
rY)
= L
11Yr
X
rYpY)
_ _
T
gYr
b T
lYr
(36)
In case of condensation, the currents are changing the
sign, therefore
X
rYpYc
= X
rYp
_ _
j
rYc
= L
11Yr
X
rYpYc
_ _
T
gYr
` T
lYr
(37)
When the system approaches an ideal point, the second
term in the right side of Eq. (35) vanishes and force
takes maximum absolute values, in accordance with the
postulate. The next paragraph includes case studies for
binary and pure component interactions.
The phenomenological coecient may be theoreti-
cally estimated making a simple imaginary experiment.
Let us consider an isolated gaseous mixture bubble, of
volume V
g
t ( ), at pressure p
g
t ( ) and temperature T
g
,
which interacts with the liquid phase surrounding it. For
convenience, it is assumed that the bubble has an iso-
therm adiabatic evolution in the liquid. In accordance
with the rst principle of thermodynamics, the variation
of its internal energy within the time unit is equal to the
power exchanged between the gas and its environment,
namely
U
X
g
= p
g
V
X
g
or,
j
r
&
l
[V
X
g
[u
g
= p
g
V
X
g
(38)
where u
g
= c
vYg
T
g
is the gas specic internal energy.
Upon absorption V
X
g
` 0, and upon generation V
X
g
b 0
and Eq. (38) is simplied to the expression
j
r
=
p
g
&
l
c
)Yg
T
g
(39)
Eq. (39) shows that positive reduced currents are
obtained upon absorption, and negative reduced cur-
rents are obtained upon generation, as previously men-
tioned herein. This equation is valid when the
homobaricity condition in the bubble is met [30]
R
X
aC
_ _
2
`` 1
where R
X
and C are the interface velocity and the velocity
of sound in gas, respectively. Physically, the above con-
dition is met in the case of mass currents with low
values, that is, anticipating the conclusions hereof, mass
currents which are far from the state of equilibrium,
when the pressure from the wall is sensibly equal to the
pressure in the rest of the gas volume and has a known
value. The determination of the phenomenological
coecient is not aected by the position of the system
relative to the point of equilibrium in respect of which it
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 159
has been assumed that L
11Yr
is independent of X
r
. L
11Yr
is
calculated by means of Eq. (28)
L
11Yr
=
j
r
X
r
(40)
where j
r
and X
r
are determined through an iterative
process from Eqs. (39) and respectively (34) [20]. A
similar result is obtained if, instead of the energy equa-
tion, the mass balance in the bubble is written
j
r
&
l
[V
X
[ = &
g
V
X
(41)
resulting in
j
r
=
&
g
&
l
(42)
For an ideal gas
&
g
=
p
g
RT
g
(43)
and the reduced mass ow acquires the form
j
r
=
p
g
&
l
RT
g
(44)
which is likewise valid and useful in homobaricity con-
ditions.
3. Application of phenomenological equations to the
gasliquid systems. The ideal point approaching
(I.P.A.) eect
The thermal factors involved in the calculation of the
thermodynamic force have been rst analytically
expressed for the NH
3
/H
2
O system [20], and its pure
components [26,29], by means of Ziegler and Trepp
state equation [31]. In a second step, the applications
were extended also to other known working combina-
tions, namely NH
3
/LiNO
3
and NH
3
/NaSCN, using
thermodynamic and physical property data already
published for them [32]. From the experience, the gas
liquid interactions of these systems are nonlinear and
the assumption of a linear set of phenomenological
equations needs to be justied. Indeed, the currents are
linear with respect to the thermodynamic force, only. In
fact, they are strong nonlinear with respect to all inter-
face variables through the thermal factors s
r
Y h
r
Y A
r
and
C
r
intervening in the calculation of the thermodynamic
force [31,32].
For the beginning we shall present the results of the
ammonia/water system. The phenomenological coe-
cient has been estimated by means of Eqs. (40), (39) and
(34), resulting in L
11Yr
= 10
3
[R[
1
= 1a8314. Prior to
starting the application, we paid special attention to the
consistency of the thermodynamic force equation with
respect to the equilibrium thermodynamics, regarding
the ideal points calculation of a gasliquid system and
particularly of the ammoniawater mixtures. As we
already mentioned, the force is not dened in an ideal
point and never cancels. Practically, such a point is
determined by progressively restricting the interval of
denition throughout which the function has nite
values, of equal absolute value and contrary signs, up to
the desired level of accuracy (here usually obtained up
to the sixth decimal). In principle, the two approaches,
equilibrium and phenomenological, are equivalent, as
far as the determination of the ideal point is concerned.
However, although in both cases the same state equa-
tion for the binary mixture under study has been used,
the phenomenological calculus have led to sensibly dif-
ferent results in some parts of the solubility eld [20].
First attempts to approach the results of the two meth-
ods failed when a nonlinear dependence of the mass
current with respect to the thermodynamic force was
considered. This explains why we used only a linear
dependence of the mass current with respect to the
thermodynamic force [Eq. (24)]. Best results were
obtained amplifying the second term of the denomi-
nator in Eq. (34) by a match function M p
rYe
Y x
e
( ),
already mentioned in Section 2. Several values of the
match function are given in Fig. 1 for ve reduced ideal
pressures and ve ideal solution mass fractions. These
values were good correlated by a four-order double
polynomial expansion:
Fig. 1. Several values of the match function in Eq. (34) for
ammonia/water gasliquid mixtures (p
b
= 10Lai; T
b
= 100 K).
Fig. 1. Plusieurs valeurs correspondent a Eq. (34) pour les mel-
anges ammoniaceau (gazliquide) (p
b
=10 bar ; T
b
= 100 K).
160 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
M p
rYe
Y x
e
_ _
= ex

4
iYj=0
A
iYj
p
i
rYe
x
j
e
_ _
(45)
The coecients A
iYj
are given below:
A
00
= 1X55107 10
0
; A
01
= 4X30114 10
0
;
A
02
= 2X50709 10
1
; A
03
= 5X14562 10
1
;
A
04
= 3X53064 10
1
; A
10
= 2X20458 10
1
;
A
11
= 3X43605 10
2
; A
12
= 2X15534 10
3
;
A
13
= 5X56704 10
3
; A
14
= 4X87992 10
3
;
A
20
= 5X72617 10
1
; A
21
= 9X54441 10
2
;
A
22
= 6X00841 10
3
; A
23
= 1X54652 10
4
;
A
24
= 1X35339 10
4
; A
30
= 5X53949 10
1
;
A
31
= 9X34452 10
2
; A
32
= 5X88969 10
3
;
A
33
= 1X51448 10
4
; A
34
= 1X32402 10
4
;
A
40
= 1X74164 10
1
; A
41
= 2X94977 10
2
;
A
42
= 1X86140 10
3
; A
43
= 4X78901 10
3
;
A
44
= 4X18852 10
3
The relative dierences (%) in the approximation of
equilibrium mass fraction of the mixture solution with
Eq. (34), are plotted in Fig. 2. The dierences could
come also from the dierent units used for x
e
, namely
mass fraction in hx diagram, respectively molar frac-
tion in the equation of state of Ziegler and Trepp, and
the use of a unique computing value for the phenomen-
ological coecient. Eq. (34) covers an usual domain for
refrigeration purposes with the NH
3
/H
2
O system,
namely of 1.04p415.0 bar, 0.14x40.45 and
0.54y40.998.
The presence of the match function in force equation
must be explained, because a less experienced reader
may think that its use was necessary to compensate for
the much simplify assumptions considered in our theo-
retical approach. In fact, M p
rYe
Y x
e
( ) compensates the
arbitrarily chosen zero reference state of a system gas/
liquid enthalpy and entropy, which Eq. (34) is quite
sensitive to. To this extent, Eq. (45) is valid for Ziegler
and Trepp work, only.
Prior to using Eq. (34) in an evolution process
assessment, we need the ``nearest'' ideal point which the
system evolves to, in order to calculate the match func-
tion. Normally, this should result from an iterative cal-
culus, searching, according to a Prigogine theorem, the
way of minimum entropy production. However, not all
but in many cases a simpler practical method can be
considered, remarking that most systems have state
parameters with quasiconstant values all along the evo-
lution period, provening from their connection with the
innite reservoirs or other engineering purposes etc.
Depending on the system variance, these parameters
partially or completely dene the nearest ideal point. In
our case, the ammonia/water system is biphase and
bicomponent, so it suces to have two such parameters,
but the type of the parameters may change upon the
specic application. For instance, for the absorption
process of a closed system absorption refrigeration
plant, these parameters are the rich solution nal tem-
perature, close to the sink source temperature and the
absorbed gas molar fraction, but for the generation
process, they change to the nal poor solution tem-
perature, close to the warm source temperature and the
condensing pressure. In case of the bubble absorption
experience, mentioned in the introduction, the nearest
ideal point is dened by the quasiconstant absorption
pressure (p
l
= p
g
) and gas molar fraction y.
The phenomenological Eqs. (27), (28), (31) and (34)
have been represented for exemplication as partial
functions of the reduced state parameters xY t
lYr
Y y and
t
gYr
, respectively, in Figs. 37 for the NH
3
/H
2
O system
(at interface, gas and liquid are in mechanical equili-
brium, p
l
p
g
, and no pressure partial functions exist in
fact). The thermodynamic force of pure components
water and ammonia is given in two study cases in Fig. 8,
and that of NH
3
/LiNO
3
and NH
3
/NaSCN systems in
four study cases in Figs 9 and 10, respectively. The
curves are dierent in shape, depending upon the vari-
able used in the abscissa and the nature of the system.
Partial functions which depend on the parameters of the
same phase have the same absorption or generation
monotonous variation. According to the postulate and
its consequence, the diagrams show the continuously
increase to nite values of force, entropy source, and of
the coupled mass and heat currents, approaching an
ideal state. In case of currents, an ideal point approach-
ing (i.p.a) eect is emphasized: depending on how strong
Fig. 2. The relative dierences % in the approximation of
equilibrium mass fraction of ammonia/water solutions with Eq.
(34).
Fig. 2. Dierences entre la fraction massique des solutions
ammoniac / eau et l'Eq. (34), exprime en pourcentages.
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 161
the currents coupling is, the approach to an ideal state
determines an increase eect of the function absolute
value by several percent (in case of pure components), to
several hundred times (for the binary systems), as com-
pared to the states which are far from the same ideal
point. In this way, the term of ``far from equilibrium
state'' changes its classical meaning and corresponds to
its opposite, namely of ``low interaction''. The i.p.a.
eect may be quantied by a gure of merit, ``f '', equal
to the ratio of two same type (absorption/condensation
or generation/vaporization) function values, corre-
sponding to the states closer, respectively, further to the
ideal state. Figs. 310 include also the maximal ``f ''
values for the chosen examples. As a direct consequence
from the above it results that estimation of the interface
mass transfer by analogy with heat transfer lacks basic
physics. Force in Eq. (2) tends to zero approaching an
ideal point, which is not true, therefore is erroneous.
The specialized literature does not contain systematic
theoretical or experimental accounts of the non-equili-
brium interaction at hand which the results obtained
herein might be compared to. The only conrmation of
such results is given by their compliance with a few
experimental works which are in agreement with the
graphs shown in Figs. 310, proving the increase of the
gaseous mass ow driven as a two-phase binary system
Fig. 4. The i.p.a. eect for the reduced mass (a) and heat cur-
rent (b) as functions of solution mass fraction.
Fig. 4. L'eet i.p.a de la masse reduite (a) et le ux de chaleur
(b) en fonction des fractions de masse de la solution.
Fig. 5. The i.p.a. eect behavior for the reduced entropy source (a)
and thermodynamic force (b) as functions of solution temperature.
Fig. 5. Comportement de l'eet i.p.a. pour la source d'entropie
diminuee (a) et la force thermodynamique (b) en fonction de la
temperature de la solution.
Fig. 3. The i.p.a. eect behavior for the reduced entropy source (a)
and thermodynamic force (b) as functions of solution mass fraction.
Fig. 3. Comportement de l'eet i.p.a. pour la source d'entropie
diminuee (a) et la force thermodynamique (b) en fonction des
fractions de masse de la solution.
162 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
is approaching a state of saturation. To this eect,
mention should be made of the experimental work of
Hui and Thome [33], who investigated the boiling of
binary azeotropic ethanolwater and ethanolbenzene
mixtures and ascertained the increase of the number of
vaporization points, until it became impossible for such
points to be individually distinguished, as the molar
fraction of the mixtures approached the saturation
value.
4. The PhHGD approach of the single bubble absorption
The author used the PhHGD approach to model local
absorption/generation processes in the ammonia/water
system [20]. Because generation in an absorption plant
passes more or less through a sequence of ideal points,
but absorption, on the contrary, working usually with
rectied vapor is far from the ideal point and takes place
therefore at low gasliquid interaction, the author paid
more attention to the absorption process. However,
taking into account the ndings of Section 3, he had
additional good reasons to model the bubble absorp-
tion: (i) it was used to satisfactorily explain the ammonia
bubble absorption problem, raised in the introduction;
(ii) it contributed to a simpler experimental validation of
the PhHGD method [20,28], and (iii) it served to perform
an analytical study of absorption for the system at hand,
based on the single bubble dynamics. Here we present
only a few results referring to the rst point and some
conclusions of the analytical study.
Fig. 7. The i.p.a. eect behavior for the reduced entropy source
(a) and thermodynamic force (b) as functions of gas reduced
temperature.
Fig. 7. Comportement de l'eet i.p.a. pour la source d'entropie
diminuee (a) et la force thermodynamique (b) en fonction de la
temperature reduite des gaz.
Fig. 6. The i.p.a. eect behavior for the reduced entropy source (a)
and thermodynamic force (b) as functions of gas mass fraction.
Fig. 6. Comportement de l'eet i.p.a. pour la source d'entropie
diminuee (a) et la force thermodynamique (b) en fonction des
fractions de masse des gaz.
Fig. 8. The water and ammonia i.p.a. eect behavior for the
reduced thermodynamic force as function of liquidgas reduced
temperature.
Fig. 8. Impact du comportement l'eet i.p.a. de l'ammoniac / eau
pour le force thermodynamique reduite en fonction de la tem-
perature reduite du melange liquide / gaz.
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 163
Absorption takes place between gas and the sub-
cooled absorbent, when forces and currents are positive.
The set of the PhHGD code entry data corresponds to
the absorption of the ammonia bubble (y=0.992) in
water (x

= 10
4
), at normal pressure (p

= 1X001 bar)
and environment temperature (T

= 293 K). The solu-

tion of the motion equation is shown in Fig. 11. Fig. 12
elucidates the problem of ammonia bubble absorption.
Indeed, the reduced absorbed mass current j
r
and the
actual one
j
/
iYg
= j
r
&
l
[V
X
[ (46)
continously increase during bubble growth time, causing
the bubble pressure decrease (Fig. 13). At a certain
moment, the actual current equals the bubble mass
feeding rate m and the collapse starts. Depending on the
interface parameters, it continues to increase also after
the starting of the collapse at an even higher speed, up
to the total absorption of the bubble. This further
explains why the collapse duration is much shorter than
the duration of the growth. The cumulated absorbed
mass current
j
/
iYgYa
t ( )
_
t
0
j
/
iYg
z ( )oz (47)
and the corresponding coupled heat current j
//
qYlYa
are
plotted against the bubble growth time in Fig. 14.
Referring to the ammonia/water analytical study, the
PhHGD code proved to be a rened tool of investiga-
tion of the intrinsic absorption properties of this com-
bination. According to it, a few rst conclusions have
Fig. 9. The i.p.a. eect behavior for partial thermodynamic
forces of liquid mass fraction (a) and liquid temperature (b) in
case of the NH
3
/LiNO
3
system.
Fig. 9. Comportement de l'eet i.p.a. pour les forces thermo-
dynamiques partielles pour la fraction liquidemasse (a) et le
temperature du liquide (b) du systeme NH
3
/LiNO
3
.
Fig. 10. The i.p.a. eect behavior for partial thermodynamic
forces of liquid mass fraction (a) and liquid temperature (b) in
case of the NH
3
/NaSCN system.
Fig. 10. Comportement de l'eet i.p.a. pour les forces thermo-
dynamiques partielles pour la fraction liquidemasse (a) et le
temperature du liquide (b) du systeme NH
3
/NaSCN.
Fig. 11. Dynamics of ammonia bubble absorption.
Fig. 11. Dynamique de l'absorption de la bulle d'ammoniac.
164 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
been drawn concerning best mass/heat transfer: (i) gas
and initial liquid temperatures should have close values
to those of the ideal point; (ii) working with as much as
low absorption pressures is desirable (absorption e-
ciency increases about 3 times when pressure decreases
from 2.9 to 0.2 bar), but this must be obviously corre-
lated with the other design parameters of a plant, and
(iii) absorption eciency considerably decreases when
the gas mass fraction approaches the unity; bearing in
mind its high energy consumption, vapor rectication,
commonly used in absorption plants, has this time an
additional reason to be avoided, that of decreasing
mass/heat transfer in absorption processes; to this
extent, on the opposite side, resorption plants gain twice
eliminating rectication. More details about this analy-
tical study can be found in author's past works,
[20,28,29].
The phenomenological modelling of the bubble
absorption may be a valuable lesson for a future
advanced thermal absorption technology. Next, we
synthesize what is to be learned from it. First, the
absorption process of gasliquid working combinations
used in the thermal absorption technology is a mass phe-
nomenon and not a surface one, as pointed out also by
Eq. (46). From here, corroborated also by conclusion
relative to the motrice force, given in Section 3, it results
also that the use of the mass transfer coecients K
g
and
K
l
as surface factors is improper. Second, absorption
can be improved in an intensive way, seeking not remo-
teness from the ideal point (see the classical point of
view), but nearness to it, when the currents naturally
increase up to very high values without any additional
technical improvements (see gure of merit f=740.7 in
Fig. 12). The tendency manifested in the construction of
NH
3
/H
2
O absorbers of increasing the gasliquid contact
area through the dispersion of one of the phases in order
to increase the absorption eciency may not result in
the anticipated eect. Quite the contrary, by extend-
ing the contact area the more rapid evolution of the
interface towards the equilibrium parameters is delayed
(the mass fraction x of the liquid at interface decreases
with the increase of the contact area, at the same gas
feeding ow rate) and absorption decreases, as com-
pared to the one achieved in a less dispersed phase.
Moreover, the auxiliary pumping energy consumption is
also increased. Thus, in the light of the above, the
increase of the contact area is merely an extensive tech-
nical solution which helps to compensate the absorption
decrease caused by an articial remoteness from the
ideal point, oering in exchange a more stable func-
tioning of the absorbing device. In addition to the
above, it is noted that quasispherical bubbles with a
minimum contact area for the same volume stand the
greatest chances to achieve the best mass and heat
transfer through i.p.a. eect appearance. This gives rise
to the idea of constructing an NH
3
/H
2
O bubble ow
absorber [20]. It is at plate and mounted horizontally.
In comparison with a present NH
3
/H
2
O absorber
with vertical pipes or with liquid phase dispersion, the
bubble absorber could oer the following advantages:
Fig. 12. Phenomenological elucidation of the ammonia bubble
absorption problem.
Fig. 12. Elucidation de l'absorption de la bulle d'ammoniac.
Fig. 13. Tube and bubble pressure variation during ammonia
bubble absorption.
Fig. 13. Variation de la pression a l'interieur du tube et de la bulle
lors de l'absorption de la bulle d'ammoniac.
Fig. 14. Cumulated absorbed mass and coupled heat currents
during ammonia bubble absorption.
Fig. 14. Flux de masse et de chaleur absorbes cumules lors de
l'absorption de la bulle d'ammoniac.
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 165
(i) maximum absorption eciency; (ii) minimum pres-
sure loss on gas side; (iii) it is suited to a modern com-
pact plate construction; (iv) minimum auxiliary energy
consumption. Third, intensive theoretical and experi-
mental research work has been carried out during the
last decade aiming at improving absorption, by means
of additives (surfactants) which stimulate the Mar-
angoni convection [3436]. Up to the present date there
is no satisfactory explanation of the mechanism which
generates this eect [37]. Most of the unsuccessful
attempts have been based on the eect of supercial
tension reduction in the liquid, obtained by means of
additive agents. The non-equilibrium thermodynamics
oers yet another chance to the clarication of this
problem. Without having the intention to give a quali-
tative and so much more quantitative complete expla-
nation, we will however make a primary analysis of the
cause which can lead to the occurrence of the Mar-
angoni eect from the phenomenological point of view
[20]. Here, contrary to the eect obtained by the dis-
persion of phases, by means of additive agents the
interface area decreases considerably, and its parameters
approach the equilibrium values. Interface becomes
non-homogeneous, in the sense that violent absorption
centers appear at its level which alternate with weak
interaction zones, due to the presence of the surfactant
(which can be dissolved in the absorber or not). This
non-homogeneity, inuenced to a small extent by the
supercial tension, causes an accentuated perturbation
of the liquid surface, which is characteristic of Mar-
angoni instability.
5. Conclusions
The above leads to the conclusions presented herein
below. These conclusions refer to all mono-, bi- and
particular polycomponet gasliquid working combina-
tions used in the thermal absorption technology,
including in the rst place the ammoniawater system:
1. The paper raises the problem of ammonia bubble
absorption, which is dicult to answer by means
of current theory of interface mass transfer and
absorption as a surface phenomenon.
2. According to the linear phenomenological
approach, the heat and mass transfer at the gas
liquid interface is governed by the thermodynamic
force, dierent to that of forces governing the heat
and mass transfer in continuous media and origi-
nating from Fourier's and Fick's laws, respectively.
The force equation is extendable to other interac-
tions occuring in continuous or discontinuous
media, like solidliquid, solidgas, or liquid
liquid, gasgas type. The paper mentions a postu-
late referring to the force behavior approaching
an ideal point, introduced by the author in a pre-
vious work. According to its consequence, the
mass and heat currents suer an ideal point
approaching (i.p.a.) eect, not mentioned so far in
the specialized literature, consisting of a con-
tinuous increase of their absolute value by several
percent (for a pure component), to several hun-
dred times (for a binary system) when the inter-
acting system approaches an ideal state, as
compared to the values of states which are far
from the same ideal point. In this way, ``far from
equilibrium'' becomes synonymous with ``low
interaction (currents)''. The entropy source, force
and coupled currents are not dened in an ideal
point.
3. Estimation of the interface mass transfer by ana-
logy with heat transfer, expressed by Eqs. (1) and
(2), lacks basic physics.
4. The phenomenological approach elucidates the
problem of ammonia bubble absorption.
5. Absorption process is a mass phenomenon and
not a surface one. The use of the mass transfer
coecients as surface factors is improper.
6. An intensive way of improving absorption is
emphasized, which seeks to promote the i.p.a.
eect appearance. This would replace the exten-
sive way currently used, based on increasing
gasliquid interaction area. To this extent, the
bubble absorber is hereby proposed for ecient
absorption.
7. The i.p.a. eect existence oers an additional chance
for a satifactory explanation of the Marangoni
eect.
Acknowledgements
The author thanks Professor M.D. Cazacu at the
Polytechnical University of Bucharest for the valuable
assistance in performing the crude material of this paper
and math. D-tru Mihai at SC ICPET-Cercetare SA for
the help in numerical solving.
References
[1] Lewis WK, Whitman WG. Ind Eng Chem 1924;1:1215.
[2] Higbie R. Trans AIChE 1935;31:36589.
[3] Danckwerts PV. Industr Engng Chem 1954;43:14607.
[4] Hobler T. Mass transfer and absorbers. London: Pergamon
Press, 1966.
[5] Ciborowski JW, Richlicki RM. Mass transfer in two-
phase gas-liquid wavy ow. Int J Heat Mass Transfer
1971;14:12616.
[6] Anderson JD, Bollinger RE, Lamb DE. Gas-phase con-
trolled mass transfer in two-phase annular horizontal
ow. AIChEJ 1964;10:6405.
166 M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167
[7] Heuss JM, King CJ, Wilke CR. Gasliquid mass transfer
in cocurrent froth ow. AIChEJ 1965;11:86672.
[8] Chadzidakis S. Wa rmeu bertragung am waagerechten mit
Ammoniak-Wasser-Lo sungen berieselten Rohr. Ka lte-
techhnik-Klimatisierung 1971;23(7):2028.
[9] Hera D. Contributions to the study of heat and mass
transfer in absorption processes in ammoniawater solutions
in lm ow in horizontal pipes. Ph.D. thesis, Bucharest
Technical Civil Engineering University, 1987 (in Romanian).
[10] Chiriac F., Mihaila C., Tarca C. Baltaretu F. Heat and
mass transfer in the boiling of ammonia/water solution.
Proc. 19th Int. Congr. Refr., vol. IIIa, The Hague, 2025
August 1995. p. 6771.
[11] Mihaila C. C., Chiriac F. N., Sota I.D. Baltaretu F. The
modeling of a lm absorber with ammonia/water solution.
Proc. 19th Int. Congr. Refr., vol. IIIa, The Hague, 2025
August 1995. p. 15460.
[12] Keizer C. Absorption refrigeration machines. Ph.D. thesis,
Dept. of Mechanical Engineering, Delft University of
Technology, Delft, 1982.
[13] Banerjee S, Scott DS, Rhodes E. Studies on cocurrent gas-
liquid ow in helically coiled tubes II. Can J Chem Engng
1970;48:54251.
[14] Gregory G.A. Scott D.S. Physical and chemical mass
transfer in horizontal cocurrent gasliquid slug ow. In:
Rhodes E., Scott D.S. editors. Plenum Press, 1969. p. 633
65.
[15] Jepsen JC. Mass transfer in two-phase ow in horizontal
pipelines. AIChE Journal 1970;16:70511.
[16] Jagota AK, Rhodes E, Scott DS. Mass transfer in
upwards cocurrent gasliquid annular ow. The Chem
Engng Journal 1973;5:2331.
[17] Kulic E, Rhodes E. Chemical mass transfer in concurrent
gas-liquid slug ow in helical coils. Can J Chem Engng
1974;52:11423.
[18] Kasturi G, Stepanek JB. Two-phase ow IV: gas and
liquid side mass transfer coecients. Chem Engng Sci
1974;29:184956.
[19] de Groot C. Mazur P. Nonequilibrium thermodynamics.
Mir, 1964 (in Russian).
[20] Staicovici M.D. Contributions to two-phase ows and heat
transfer in solar cooling installations. Ph.D. thesis, Buchar-
est Polytechnics University, July 1998 (in Romanian).
[21] Staicovici MD. Variational numerical and analytical
approach of the liquid capillary rise eect. Revue Rou-
maine des Sciences Techniques, se rie Me canique Aplique e
1998;43(1):8599.
[22] Chartier P, Gross M, Spiegler KS. Applications de la
thermodynamique du non-e quilibre. Paris: Hermann,
1975.
[23] Vilcu R, Dobrescu A. Thermodynamics of irreversible
processes. Bucharest: Technical Publishing House, 1982.
[24] Staicovici MD. Poybranched regenerative GAX cooling
cycles. Int Journal of Refrigeration 1995;18:31829.
[25] Staicovici M.D. Numerical phenomenological approach of
transport provided biphase ow for NH
3
/H
2
O absorption/
generation elementary processes. Proc.VIIIth National
Conference of Thermopower, vol. II, Pitesti, 2930 May,
1998. p. 26980.
[26] Staicovici M.D. A phenomenological approach of evolu-
tion physical mono-, bi-and particular polycomponent
gasliquid interactions with non-ideal mixture. Part I:
Theory and its application to the NH
3
/H
2
O gasliquid
interactions. Int J Heat Mass Transfer, submitted for
publication.
[27] Dodescu G. Numerical methods in algebra. Bucharest:
Technical Publishing House, 1979.
[28] Staicovici M.D. A phenomenological approach of evolu-
tion physical mono-, bi-and particular polycomponent
gasliquid interactions with nonideal mixture. Part II:
modeling of single bubble absorption/generation pro-
cesses, analytical study of absorption and experiments. Int
J Heat Mass Transfer, in press.
[29] Staicovici M.D. A phenomenological hydro-gazodynami-
cal approach of the NH
3
/H
2
O gasliquid interactions. I.
C. Refr., Sydney, submitted for publication.
[30] Nigmatulin RI, Khabeev NS, Nagiev FB. Dynamics, heat
and mass transfer of vapourgas bubbles in a liquid. Int J
Heat Mass Transfer 1981;24(6):103344.
[31] Ziegler B, Trepp C. Equation of state for ammoniawater
mixtures. Int Journal of Refrigeration 1984;7:1016.
[32] Infante Ferreira CA. Thermodynamic and physical prop-
erty data equations for ammonialithium nitrate and
ammoniasodium thiocyanate solutions. Solar Energy
1984;32(2):2316.
[33] Hui TO, Thome JR. Study of binary mixture boiling:
boiling site density and subcooled heat transfer. Int J Heat
Mass Transfer 1985;28(5):91928.
[34] Rie DH, Kashiwagi T. Computer simulation of vapor-
absorption enhancement into H
2
O/LiBr absorbed by
Marangoni convection. JSME Int Journal, II 1991;
34:35561.
[35] Homan L, Greiter I, Wagner A, Weiss V, Alefeld G.
Experimental investigation of heat transfer in a horizontal
tube falling lm absorber with aqueous solutions of LiBr
with and without surfactants. Int J Refrig 1996;19(5):33141.
[36] Mo ller R, Lnoche KF. Surfactants with NH
3
H
2
O. Int J
Refrig 1996;19(5):31721.
[37] Ziegler F, Grossman G. Heat-transfer enhancement by
additives. Int J Refrig 1996;19(5):3019.
M.D. Staicovici / International Journal of Refrigeration 23 (2000) 153167 167