Water activity in aqueous solutions of sucrose: an improved temperature dependence
Maciej STARZAK School of Chemical Engineering University of Kw aZulu-Natal Durban, South Africa
W ater activity in aqueous W ater activity in aqueous
solutions of sucrose: an improved
solutions of sucrose: an improved temperature dependence temperature dependence Mohamed MATHLOUTHILaboratoire de Chimie Phy sique Industrielle Université de Reims, Champagne-Ardenne
Outline Outline
1. Introduction
2. Previous studies
3. Experimental database
4. Selection of the water activity model
5. Relationships between experimental
variables and activity coefficients
6. Data regression
7. Results
8. Recommended equation for water activity coefficient
Water activity formulae popular amongst food technologists
2 ln x
Norrish, 1966 γ = α
A B
1000 A
- M m
γ = A
Chen, 1989 n
M m A Bm 1000 ( )
A
3 ln x x
Miyawaki, 1997 γ = α β
- 2
A B B
Models used to predict
water activity coefficient
nRedlich-Kister expansion - empirical
(includes Margules equation)
nUNIQUAC - phenomenological (two-fluid theory) n group contribution methods (UNIFAC, ASOG)
Starzak & Peacock, 1997 Q
2
2 ln x 1 bx cx
γ =
A B B B ( )
RT
Based on 1197 experimental points (56 data sets),
mainly VLE data (BPE, vapour pressure, ERH, isopiestic solutions)1.4 Water activity coefficient
0.5
1.4 Water activity coefficient Starzak & Peacock, 1997 UNIQUAC Margules eq.
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.4
0.5
Catté et al., 1994
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.4
Water activity coefficient (boiling curve)
1.8
1.6
1193 points Starzak & Peacock, 1997
1.4
1.2
1
0.8
0.6 Water activity coefficient on the boiling curve
Theoretical models of activity
for highly concentrated sucrose solutions
Van Hook, 1987:§
- - sucrose hydration - sucrose association (clustering) Starzak & Mathlouthi, 2002:
§
- - water association - sucrose hydration - sucrose association (clustering)
Previous studies (small exptl. databases) Ø Le Maguer, 1992 (UNIQUAC) Ø Caté et al., 1994 (UNIQUAC) Ø Peres & Macedo, 1996 (UNIQUAC) Ø Peres & Macedo, 1997 (UNIFAC) Ø Spiliotis & Tassios, 2000 (UNIFAC) Objective of this study: q an empirical activity equation
q wide range of temp. & concentrations
q large database
Thermodynamic data typically used
to determine the activity coefficientVLE data (BPE, VPL, ERH, q osmotic coeff. by isopiestic method) q SLE data (FPD, sucrose solubility)
q Termochemical data (heat of dilution,
excess heat capacity)EXPERIMENTAL DATABASE
Experimental points Literature sources Data type
VLE 1507
64 FPD
213
13 Solubility 265
34 Heat of dilution 283
10 Heat capacity
70
4
Distribution of experimental data
Distribution of experimental data
- -suffix Margules equation Advantages: linear with respect to its coefficients
2
ln n kA k B k x
γ α =
= ∑
Selection of water activity model n
Drawbacks: lack of sound theoretical foundation
2
3
1
2
3
4
5 ( )
ln
k k k k k k kα α θ α α θ α θ α θ α θ θ = + + + + +
General temperature dependence
Disadvantage: large number of parameters where T T
θ =
Simplified temperature dependence
Advantage: reduced number of parameters
2
2 ln ( ) n k
A k B k a b x
γ θ − =
= ∑
2
3
1
2
3
4
5 ( ) ln a a a a a a a
θ θ θ θ θ θ
= + + + + + Disadvantage: temperature effect pattern independent of composition
Sucrose activity coefficient
n k ln xγ = β B k A
∑ k
2 =
Gibbs-Duhem equation
d d ln lnγ γ A B x x
=
A B d x d xA B
Expansion coefficients of sucrose activity
( 1) ,2 n k k l l k l A k k
β = −
∑
³ ³
1
1 , 2,3,...,
1 l l l n n l A l n l
- = − = − =
A α α α
where
2
3
2
2
2
2
2
8
35
5
8
3
15
5
35
9
4
2
24
14
5
35
6 β α α α α α α β α α α α α β α α α α β α α α β α α = + + + + +
= − − − − − = + + + = − − − = +
7
3
2
7
3
4
5
6
7
3
3
4
5
6
4
7
4
5
6
7
5
5
6
7
6
6
Expansion coefficients of sucrose activity, n = 7
Expansion coefficients of sucrose activity
Matrix representation
=ß M a
1
1
1
- =
1 , 1, 2,...,
1 n k k l l l
M k n β α
−
= −
∑ =
Processing experimental data
VLE data ln
⇒ γ
AFPD data ln
⇒ γ
A Solubility ln
⇒ γ
B E
Heat of dilution H RT
/⇒ asym E
Specific heat C / R
⇒ pasym ln
VLE data A
γ ⇒
Modified Raoult’s law
(1 ) ( ) A B A P x P T
γ =
−
ERH 100
1
A B xγ = −
Equilibrium relative humidity (ERH):
BPE, vapour pressure, osmotic coeff. :
Assumed
H T C T B T T RT R T C T T T B T B T x R T T
C T C T B T T R R ∆ ∆ = + ∆ −
( ) ( ) ( )
pA pA m
A m ∆ ∆
= − − + −
∆
γ ∆ ∆
A m A m B m m
Freezing point data ln A
A m f pA m f f
2 fA m pA m A m m
1 ln(1 )
2 ( ) ln
1
( ) ( ) ln
γ ⇒ Solid-liquid equilibrium
- − ∆ + − − −
Sucrose solubility data ln ⇒ γ
% B
Solid-liquid equilibrium (asym. convention)
∞C T ( )
∆ ( T ) H ( ) T B T T T
pB
γ ∆ ∆
1 γ = − −
- B m dB
− B
∞
B m ln
RT R T T
2 γ
B m
C ( ) T ∆ T B T T
pB ∆
B T ln 1 ln( x ) − ∆ − − −
- B m
B B
R T T
2 m m
C T C T
( ) ( ) Assumed ∆ ∆
pB pB
B T T ( )
= + ∆ − B for sucrose:
R R
E H asym
Heat of dilution data ⇒
RT
Two-step procedure
q fitting each set of isothermal data toMcMillan-Mayer expansion (evaluating coefficients of expansion) q data generated from the expansion used as input data for regression
Types of experimental heats of dilution
I. Amount of heat liberated per one mole of sucrose in solution after addition of a known amount of pure water
H k k
∆
1
1 I − −
(J/mol solute) h m (f ) m (i) = − k
∑
n k2 B
=
Types of experimental heats of dilution
II. amount of heat liberated per one mole of water in original solution after addition of a known amount of dilute sucrose solution H
∆
II (J/mol solvent)
= n (i)
A
n (i) n (a) m (f ) n (a) m (a) n (i) m (i) − −
- k k k
A A A A [ ] h k
∑ n (i) k
2 =
A
Types of experimental heats of dilution
III. amount of heat liberated per one gram of water added after addition of a small amount of pure water (differential heat of dilution) H
1 ∆ k
III (J/g solvent added) h m (i)
= − k
∑ w
1000 k
2 A = m (f ) @ m (i)
E H asym
Heat of dilution data ⇒
RT Excess enthalpy of solution from McMillan-Mayer expansion k h m k
∑ E k
2 =
H (J/mol) =
1000 m M
- asym
A
/ Specific heat data E p asym
C R ⇒
Apparent molar heat capacity of sucrose: 1 ( ) 1000
M m c m c m m
+ −
Φ =
( ) B p pA c
Excess heat capacity of solution [ ]
( ) (0) (J/mol K) 1000
E c c p asym
- ×
A m m C m M
Φ − Φ =
Experimental variables & Margules expansion coefficients SLE (asym. convention) - solubility ∞
( T ) γ
B m ln
γ = B
∞ γ
B n n n
1 − b
k
2 − a T ( ) M b x a T ( ) a T ( )
k
− k l l A m
1 − k
[ ] ∑∑
∑
1 k l k
2
1 2 −
= = =
Experimental variables
& Margules expansion coefficientsExcess enthalpy of solution
Derived from free Gibbs energy: ( ) ln ln
E A A B B G RT x x
γ γ = + ( / )
E E H G RT T RT T ∂ = −
∂
2 ln
E E B asym B
H H RT x γ
∞ ∂ = +
Experimental variables & Margules expansion coefficients Excess enthalpy of solution E n
H asym b k k
2 −
( ) T a T ′ x
= B
∑
1 RT k k
2 −
= a2
3 T a T ( ) a a 2 a 3 a ′ = − + + θ θ θ
2
3
4
5 θ where T T
θ =
12 T a T Ta T a a a a θ θ θ ′ ′′
− ∑
6
2
5 2 ( ) ( )
4
3
2
3
2
[ ]
− = ′ ′′ = +
Experimental variables & Margules expansion coefficients Excess heat capacity of solution By definition: ( )
R k
B k C b T a T Ta T x
1 E n p asym k k
2 2 ( ) ( )
2
[ ]
∂ Hence:
C H R R T ∂ =
E E p asym asym
- = + + +
DATA REGRESSION Performance index exp
2
2
2 I w w ( , ) a b (ln ln )
= γ − γ
i A i A i
VLE
VLE, , ,
VLE ∑ i exp exp
2
2
2
2 w (ln ln ) w (ln ln )
γ − γ γ % − γ % + + FPD A i , A i , FPD SOL B i , B i , SOL
∑ ∑ i i
2
2 E E exp , E E exp
,
C C H H p p asym i , asym i , asym i , asym i ,
2
2 w w
− + + −
HE CE
exp ∑
∑ RT RT R R i i i i
Results of data regression: correlation diagrams
Boiling point elevation – predicted vs experimental
70 1507 points
60
50
40
30 BPE predicted, ºC
20
10
Results of FPD data regression
1.02
40
1
213 points
35
0.98 A
30
γ
0.96
25
0.94
0.92
20
0.9
15
0.88
10 Water activity coefficient,
Freezing point depression [°C]
0.86
5
213 points
0.84
Results of solubility data regression
- 0.5
100
265 points
- 1
95 Solubility curve
90
- 1.5
] ∞ B
85
γ
- 2
) / m
80
( T ∞ B
γ
- 2.5
B γ
% wt. sucrose
75
ln [
- 3
70
- 3.5
65
265 points
0.1
20 E
2
2
1.5
1
20 30°C
10
2
1.5
1
20 25°C
10
1.5
20 33°C
1
20 20°C
10
2
1.5
1
20 15°C
10
2
1.5
10
1
20 10°C
5
25°C ____ ____ experimental predicted
10
2
1
10 25°C
5
2
1
50 80°C
10
50 100 60°C
2
10
5
50 100 40°C
10
5
50 100 20°C
10
5
20 25°C
10
1
10
0.2
20
4
2
40 22°C
20
4
2
50 18°C
4
2
40 16°C
4
2
2
40 12°C
20
4
2
0.2 30°C
0.1
0.2
0.1
0.2 20°C
0.1
50 100 26°C
4
2
1.5
1.5
1
/RT [-] 5°C
10
2
1.5
1
20 0°C
10
2
1
50 30°C
50 20°C
4
2
2 18°C
1
1
0.5
0.2 18°C
0.1
0.2
0.1
20 Molality
Prediction of enthalpy of dilution 140
100
6 mol/kg (67%)
90
40°C
120
20°C
80 /RT [-] /RT [-]
100 E E
5 mol/kg (63%) 10°C
70
60°C 5°C
60
80
0°C
50
4 mol/kg (58%)
60
40
30
3 mol/kg (51%)
40
80°C
20 1000 x Enthalpy of dilution, H
1000 x Enthalpy of dilution, H
2 mol/kg (41%)
20
10
1 mol/kg (26%)
Regression of 5 heat capacity data sets
Gucker & Ayres, 1937 Gucker & Ayres, 1937 Banipal et al., 1997
0.4
0.3
0.04 T = 20°C T = 25°C
T = 25°C
0.25
0.3
0.03
0.2
0.2
0.15
0.02
0.1
/R [-]
0.1
0.01 E
0.05
2
4
6
2
4
6
0.5
1
1.5 DiPaola & Belleau, 1977 Vivien, 1906
0.05
0.4 T = 25°C
T = 15°C
0.04
0.3
0.03
0.2
0.02 Excess heat capacity, C
0.1
0.01
0.5
1 0°C
5°C 10°C
20°C 40°C 60°C 80°C
Excess heat capacity, C E /R [-]
0.2
0.4
0.6 1 mol/kg 2 mol/kg 3 mol/kg 4 mol/kg 5 mol/kg 6 mol/kg
Excess heat capacity, C
E
/R [-]
Prediction of excess heat capacity of solution- 1
- 0.5
- 1
- 0.8
- 0.6
- 0.4
- 0.2
Prediction of water activity coefficient
1.6 ISOTHERMS BETWEEN 0°C and 150°C A
γ
1.4
1.2
1
0.8 Water activity coefficient,
0.6
Prediction of water activity coefficient
1 A
0.9 γ
0.8 80°C 100°C
0.7 150°C
60°C
0°C 20°C 40°C
Water activity coefficient,0.6
0.5
Final water activity coefficient equation
γ θ θ θ θ θ θ
1 = -1.9831 b
4 = -277.88 b
3 = 1399.9 a
2 = -1102.3 a
1 = -861.42 a
= -268.59 a
= + + = + + + + a
A B B B a b x b x x a a a a a a
2
4 ln ( ) (1 ) ( ) ln
3
2
1
2
2
1
2
2 = 0.92730
Acknowledgement Acknowledgement Association Andrew van Hook for the Advancement of the Knowledge on Sugar, Reims, France