UNIVERSIDADE ESTADUAL DE CAMPINAS

FACULDADE DE ENGENHARIA DE ALIMENTOS
DEPARTAMENTO DE ENGENHARIA DE ALIMENTOS



EMPREGO DE MÉTODOS DE CONTRIBUIÇÃO DE GRUPOS NO CÁLCULO E
PREDIÇÃO DE PROPRIEDADES FÍSICO-QUÍMICAS


Luciana Ninni
Engenheira de Alimentos
Mestre em Engenharia de Alimentos

Prof. Dr. Antonio José de Almeida Meirelles
Orientador






Tese apresentada à Faculdade de
Engenharia de Alimentos da
Universidade Estadual de Campinas para
a obtenção do título de doutor em
Engenharia de Alimentos.




Campinas/SP
Novembro de 2003
BANCA EXAMINADORA



Prof: Dr. Antonio José de Almeida Meirelles (FEA/UNICAMP)


Prof
a
. Dr
a
. Maria Alvina Krähenbühl (FEQ/UNICAMP)


Prof. Dr. Pedro de Alcântara Pessôa Filho (POLI/USP)


Prof. Dr. Martín Aznar (FEQ/UNICAMP)


Prof. Dr. Fernando Antonio Cabral (FEA/UNICAMP)


Prof. Dr. Eduardo Augusto Caldas Batista (DZT/UEPG)


Prof
a
. Dr
a
. Maria Ângela de Almeida Meireles (FEA/UNICAMP)








































Aos meus dois amores,
Dirk e nosso filho.
AGRADECIMENTOS

Ao Prof. Dr. Antonio José de Almeida Meirelles, pela orientação, incentivo e presença
sempre.

Ao Prof. Dr. Gerd Maurer pela orientação em parte deste trabalho e auxílio para a realização
do mesmo.

Ao CNPq pela bolsa de estudos e pelo financiamento da pesquisa no Brasil e ao convênio
CNPq/DAAD pelo financiamento da bolsa de estudos na Alemanha.

Aos membros da banca examinadora pelas observações e sugestões feitas para a melhor
elaboração da versão final deste trabalho.

Aos meus pais e minhas irmãs que sempre me incentivam e apóiam.

Às minhas amigas Cíntia, Sueli, Christiane, Cristina e Roberta pelos bons momentos que
passamos juntas.

À Melissa, Wong e Helcio pela ajuda nos trabalhos experimentais.

À minha vó que me faz rir e que tanto amo.

À Silke Lammertz e Viktor Ermatchkov pela ajuda inicial nos laboratórios na Universidade
Kaiserslautern.

Aos meus amigos Raquel, Vanderlei e Hicham, que estão na Alemanha, pelas conversas
alegres e apoio quando precisei.

ix
ÍNDICE GERAL
Índice de Tabelas................................................................................................. xiii
Índice de Figuras............................................................................................... xviii
Resumo .............................................................................................................. xxiii
Abstract ............................................................................................................... xxv
CAPÍTULO 1
1. REVISÃO BIBLIOGRÁFICA....................................................................................................... 1
Introdução........................................................................................................................................... 3
Objetivos ............................................................................................................................................ 6
1.1 Fundamentos e modelos termodinâmicos .................................................................................... 7
1.1.1 Bases termodinâmicas ......................................................................................................... 7
1.1.2 Modelos para energia livre de Gibbs................................................................................... 9
1.2 Métodos de contribuição de grupos ........................................................................................... 12
1.2.1 Para cálculo de coeficientes de atividade.......................................................................... 12
1.2.2 Para cálculo de viscosidades de misturas.......................................................................... 18
1.3 Compostos orgânicos de interesse na área de alimentos e biotecnologia .................................. 22
1.3.1 Polióis................................................................................................................................ 23
1.3.2 Aminoácidos...................................................................................................................... 23
1.3.3 Polietileno glicóis.............................................................................................................. 24
1.3.4 Maltodextrinas (MD)......................................................................................................... 25
1.4 Propriedades físico-químicas estudadas..................................................................................... 26
1.4.1 Atividade de água.............................................................................................................. 26
1.4.2 Solubilidade....................................................................................................................... 27
1.4.3 Depressão do ponto de congelamento............................................................................... 28
1.4.4 Elevação do ponto de ebulição.......................................................................................... 28
1.4.5 pH...................................................................................................................................... 29
1.4.6 Espalhamento de luz.......................................................................................................... 29
1.4.7 Entalpia de diluição........................................................................................................... 30
1.5 Nomenclatura ............................................................................................................................. 30
1.6 Referências bibliográficas .......................................................................................................... 34
x
CAPÍTULO 2
2. WATER ACTIVITY IN POLYOL SYSTEMS........................................................................... 39
2.1 Abstract ...................................................................................................................................... 41
2.2 Introduction ................................................................................................................................ 41
2.3 Experimental Section ................................................................................................................. 42
2.4 Results and Discussion............................................................................................................... 43
2.4.1 Water activity.................................................................................................................... 43
2.4.2 Readjustment of group interaction parameters.................................................................. 47
2.4.3 Predictions with the new set of parameters....................................................................... 53
2.5 Literature Cited .......................................................................................................................... 54
2.6 Acknowledgment ....................................................................................................................... 57

CAPÍTULO 3
3. WATER ACTITITY, PH AND DENSITY OF AQUEOUS
AMINO ACIDS SOLUTIONS........................................................................................................ 59
3.1 Abstract ...................................................................................................................................... 61
3.2 Introduction ................................................................................................................................ 61
3.3 Materials and methods ............................................................................................................... 62
3.3.1 Materials............................................................................................................................ 62
3.3.2 Experimental Procedure .................................................................................................... 62
3.3.3 Experimental Results......................................................................................................... 63
3.4 Thermodynamic Modeling......................................................................................................... 68
3.4.1 Modified UNIFAC Model for Electrolytes....................................................................... 68
3.4.2 Solubility of Amino Acids in Aqueous Solutions............................................................. 74
3.5 Results and Discussion............................................................................................................... 75
3.5.1 Density data....................................................................................................................... 75
3.5.2 Equilibrium data................................................................................................................ 77
3.6 Conclusions ................................................................................................................................ 84
3.7 References .................................................................................................................................. 84
3.8 Acknowledgment ....................................................................................................................... 87

xi
CAPÍTULO 4
4. KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL) BLENDS............................ 89
4.1 Abstract ...................................................................................................................................... 91
4.2 Introduction ................................................................................................................................ 91
4.3 Experimental .............................................................................................................................. 92
4.3.1 Materials............................................................................................................................ 92
4.3.2 Apparatus and procedures ................................................................................................. 93
4.4 Results and Discussion............................................................................................................... 93
4.4.1 Excess molar Gibbs energy of activation for viscous flow............................................... 95
4.4.2 GC-UNIMOD model......................................................................................................... 96
4.5 Conclusions .............................................................................................................................. 101
4.6 Literature Cited ........................................................................................................................ 101
4.7 Acknowledgments.................................................................................................................... 103

CAPÍTULO 5
5. KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL)
AQUEOUS SOLUTIONS.............................................................................................................. 105
5.1 Abstract .................................................................................................................................... 107
5.2 Introduction .............................................................................................................................. 107
5.3 Experimental Section ............................................................................................................... 108
5.3.1 Materials.......................................................................................................................... 108
5.3.2 Apparatus and Procedures............................................................................................... 109
5.4 Results and Discussion............................................................................................................. 109
5.4.1 Kumar’s equation for viscosity correlation in PEG mixtures ......................................... 112
5.4.2 Viscosity prediction in multicomponent systems............................................................ 121
5.5 Literature Cited ........................................................................................................................ 122
xii
CAPÍTULO 6
6. THERMODYNAMIC PROPERTIES OF MALTODEXTRIN AQUEOUS
SOLUTIONS.................................................................................................................................. 127
6.1 Abstract .................................................................................................................................... 129
6.2 Introduction .............................................................................................................................. 129
6.3 Experimental ............................................................................................................................ 130
6.3.1 Materials.......................................................................................................................... 130
6.3.2 Methods........................................................................................................................... 131
6.4 Modeling .................................................................................................................................. 145
6.4.1 Osmotic virial equation................................................................................................... 145
6.4.2 VERS Model ......................................................................................................................... 151
6.5 Conclusions .............................................................................................................................. 159
6.6 Acknowledgement.................................................................................................................... 160
6.7 References ................................................................................................................................ 160

ANEXOS
ANEXO 1....................................................................................................................................... 165
Conclusões Gerais .......................................................................................................................... 175
xiii
Índice de Tabelas
CAPÍTULO 2
Table 2.1 Water activity in binary polyol solutions as a function of mass
fraction of polyol (w
2
) .............................................................................................. 43
Table 2.1 (cont.) ....................................................................................................................... 44
Table 2.2 Water activity in ternary polyol solutions at 25 °C.................................................. 45
Table 2.3 Group assignment for polyols .................................................................................. 46
Table 2.4 UNIFAC-Lyngby interaction parameters................................................................. 47
Table 2.5 ASOG interaction parameters .................................................................................. 47
Table 2.6 Systems used for adjusting the group interaction parameters .................................. 48
Table 2.7 Mean relative deviations between experimental and
calculated aw data points.......................................................................................... 50
Table 2.8 Mean relative deviations between experimental and
calculated solubility data.......................................................................................... 51
Table 2.9 Thermodynamic data of polyols and water .............................................................. 52
Table 2.10 Values of ∆A and ∆B for the calculations of ∆Cp with linear
temperature dependency........................................................................................... 52
Table 2.11 Water activity prediction in polyol mixtures............................................................ 53

CAPÍTULO 3
Table 3.1 Experimental water activities of glycine, dl-alanine, l-arginine
and l-proline as a function of amino acid mass fraction (w
a
)
at 25.0±0.1 °C........................................................................................................... 64

xiv
Table 3.2 Experimental pH of glycine, dl-alanine, l-arginine and l-proline
in water and in basic and acid buffers (w
a
=amino acid mass fraction)
24.7±0.3 °C .............................................................................................................. 65
Table 3.3 Densities of aqueous solutions of glycine, dl-alanine, l-arginine
and l-proline as a function of total solute mass fraction (w
t
) at
25.0±0.1 °C .............................................................................................................. 66
Table 3.4 Division of groups for the amino acids used in this work........................................ 68
Table 3.5 Coefficients for correlating pK values with temperature......................................... 70
Table 3.6 ∆s and ∆h values used in this work (in bold script) and
obtained from literature............................................................................................ 74
Table 3.7 Linear fitting for the densities of amino acids, salts, acid and
base in aqueous solutions ......................................................................................... 76
Table 3.8 Group interaction energies u
ij
(K) between ions and amino
acid characteristic groups ......................................................................................... 78
Table 3.9 Group interaction energies u
ij
between like groups.................................................. 78
Table 3.10 Average relative deviations between calculated and experimental data.................. 82

CAPÍTULO 4
Table 4.1 Poly(ethylene glycol) characterization..................................................................... 92
Table 4.2 Viscosity of binary blends of poly(ethylene glycol)s
at various temperatures............................................................................................. 93
Table 4.2 (continued) ............................................................................................................... 94
Table 4.3 Viscosity of binary blends containing poly(ethylene glycol)
400 at various concentrations................................................................................... 94

xv
Table 4.4 Viscosities of ternary mixtures of poly(ethylene glycol)s
at various temperatures............................................................................................. 95
Table 4.5 Viscosities of a multicomponent mixture of poly(ethylene glycol)s
at various temperatures............................................................................................. 95
Table 4.6 Coefficients of the Redlich-Kister equation (A
p
) and average absolute
deviations (AAD) between experimental and calculated G
*E
values....................... 97
Table 4.7 Average absolute deviations (AAD) for the viscosity estimation............................ 98
Table 4.8 Energy interaction parameters (in Kelvin) for the GC-UNIMOD model ................ 99

CAPÍTULO 5
Table 5.1 Average relative molar masses (M
n
), polydispersity index and
water content of PEGs used in this work ............................................................... 109
Table 5.2 Kinematic viscosities, ν, in aqueous solutions of PEGs a 293.15 K
and various mass fractions, w................................................................................. 110
Table 5.3 Kinematic viscosities, ν, in aqueous solutions of PEGs at various temperatures
and mass fractions, w ............................................................................................. 110
Table 5.3 (cont.) ..................................................................................................................... 111
Table 5.4 Kinematic viscosities, ν, of multicomponent poly(ethylene glycol)
aqueous solutions at various temperatures ............................................................. 112
Table 5.5 Linear fittings for the densities of poly(ethylene glycols) in aqueous solutions.... 114
Table 5.6 Adjusted coefficients of eqs 5.2, 5.4 and 5.5 for calculating kinematic
viscosities of aqueous PEG solutions..................................................................... 115
Table 5.7 Constants for the generalized equation .................................................................. 118
Table 5.8 Results of the generalized correlation for viscosity calculations ........................... 119

xvi
CAPÍTULO 6
Table 6.1 Experimental data on the number- and mass-average molar mass of
the maltodextrins studied in this work ................................................................... 131
Table 6.2 Experimental results for the water activities in aqueous solutions of
MD 13 – 17 at 298.15 K......................................................................................... 137
Table 6.3 Experimental results for the water activities in aqueous solutions
of MD 16 – 19 at 298.15 K.................................................................................... 138
Table 6.4 Experimental results for the water activities in aqueous solutions
of MD 4 – 7 at 298.15 K........................................................................................ 138
Table 6.5 Experimental results for the water activities in aqueous solutions
of MD 13 – 17 and 16 – 19 at 318.15 K................................................................. 139
Table 6.6 Experimental heats of dilution of maltodextrin 13 – 17 at 313.15 K..................... 141
Table 6.7 Experimental heats of dilution of maltodextrin 16 – 19 at 313.15 K..................... 141
Table 6.8 Specific density of aqueous MD 13 – 17 and MD 16 – 19 solutions
at 298.15 K............................................................................................................. 142
Table 6.9 Excess Rayleigh ratios of MD 13 – 17 in water at 298.15 K................................. 143
Table 6.10 Excess Rayleigh ratios of MD 16 – 19 in water at 298.15 K................................. 144
Table 6.11 Division of components in the maltodextrins samples........................................... 146
Table 6.12 Second and third virial coefficients of the saccharides and
polydisperse fraction studied in this work.............................................................. 151
Table 6.13 Interaction parameters of VERS model for both methods ..................................... 155
Table 6.14 Average relative deviations between experimental and calculated
data by VERS model .............................................................................................. 159

xvii
ANEXO 1
Tabela A.1 Viscosidades cinemáticas experimentais dos PEGs 8000 e 10 000....................... 166
Tabela A.2 Parâmetros da equação A.1 .................................................................................... 167
Tabela A.3 Parâmetros das equações A.2 – A.4....................................................................... 167
Tabela A.4 Parâmetros para o modelo UNIMOD..................................................................... 169
Tabela A.5 Parâmetros de interação binária para o modelo UNIMOD (Opção 1)................... 169
Tabela A.6 Parâmetros de interação binária para o modelo UNIMOD (Opção 2)................... 169
Tabela A.7 Parâmetros de interação binária para o modelo UNIMOD (Opção 3)................... 170
Tabela A.8 Resultados do ajuste de parâmetros de acordo com a opção 3
para o cálculo de viscosidades de PEGs em solução aquosa ................................. 172

xviii
Índice de Figuras
CAPÍTULO 1
Figura 1.1 Estrutura química da maltodextrina (x=ligação α-1,6 com possibilidade de
ramificação; y= ligação α-1,4 para a formação de cadeia linear). ........................... 26
CAPÍTULO 2
Figure 2.1 Prediction of water activities in the glycerol – water system................................... 45
Figure 2.2 Experimental and calculated water activities for xylitol solutions
at 25 °C..................................................................................................................... 49
Figure 2.3 Solid-liquid equilibria for sorbitol aqueous solutions (solubility
and freezing point depression) ................................................................................. 49
Figure 2.4 Experimental and calculated solubilities for D-mannitol
at different temperatures........................................................................................... 50
Figure 2.5 Predictions of water activities for the ternary system
water – xylitol – sorbitol at 25 °C............................................................................ 54

CAPÍTULO 3
Figure 3.1 Water activities determined in the present work and data
taken from ref 5 for the amino acid glycine............................................................. 63
Figure 3.2 Comparison between experimental densities from literature
and those determined in the present work................................................................ 67
Figure 3.3 Experimental and calculated a
w
of aqueous solutions
containing l-proline .................................................................................................. 79
Figure 3.4 Experimental and estimated pH of glycine solutions............................................... 79
Figure 3.5 Experimental and calculated solubility of some amino acids .................................. 80

xix
Figure 3.6 Experimental and estimated solubilities of glycine at various
pH values and NaCl concentrations ......................................................................... 81
Figure 3.7 Comparison between the distribution of the ionic species of glycine as
a function of real () and ideal (----) pH................................................................ 83
CAPÍTULO 4
Figure 4.1 Excess molar Gibbs energy of activation for viscous flow of PEG
binary mixtures (■ 400+600 at 293.15 K; □ 400+600 at 333.15 K;
▲ 400+1500; 400+3350; — Redlich-Kister) ..................................................... 96
Figure 4.2 Experimental and correlated viscosities of binary PEG mixtures
(experimental data of Table 4.2: ▲ 200+400; ┉ 200+600;
■ 400+600; ○ 400+1500; 400+3350; ▼ 600+1000;
+ 600+3350; 1000+3350; — GC-UNIMOD/ option 2)....................................... 99
Figure 4.3 Experimental and predicted viscosities of multicomponent PEG mixtures
(∇ 400+600+1000; 400+1000+3350; ┉ 400+600+1000+1500+3350;
600+1000+1500; ○ 1000+1500+3350; — GC-UNIMOD/ option 2) .............. 100

CAPÍTULO 5
Figure 5.1 B
ref
values as a function of PEG molecular mass................................................... 116
Figure 5.2 H
ref
values as a function of PEG molecular mass .................................................. 117
Figure 5.3 Experimental and calculated viscosities at 293.15 K and various
concentrations: ■, PEG 400; □, PEG 600; •, PEG 1000; ○, PEG 1500;
▲, PEG 3350; —, calculated values ..................................................................... 119


xx
Figure 5.4 Experimental and calculated viscosities at various temperatures
and concentrations: ▲, PEG 400, w=0.2487; ■, PEG 600, w=0.4899;
□, PEG 1000, w=0.4777; •, PEG 1500, w=0.4534; ○, PEG 8000,
w=0.2554; ∆, PEG 10 000, w=0.2441 ................................................................... 120

CAPÍTULO 6
Figure 6.1 Gel permeation profiles of maltodextrins............................................................... 132
Figure 6.2 Results of the experiment GPC/MALLS for the maltodextrins studied
in this work............................................................................................................. 134
Figure 6.3 Results of Gauss-Lorenz functions for the molar mass distrbution
of MD 16-19........................................................................................................... 135
Figure 6.4 Molar mass as a function of elution volume, using the MALLS data
and elution volumes for glucose, maltose and maltoriose from
the Gauss-Lorenz functions.................................................................................... 136
Figure 6.5 Reduced water activity in aqueous solutions of saccharides at 298.15 K.............. 148
Figure 6.6 Dependence of second and third virial coefficients on the molar mass ................. 149
Figure 6.7 Reduced water activity in aqueous solutions of saccharides at 298.15 K.............. 150
Figure 6.8 Laser-light scattering data of aqueous maltodextrin solutions............................... 150
Figure 6.9 Experimental and calculated water activities of water – maltodextrin
and some mono- and oligosaccharides at 298.15 K (method 1) ............................ 155
Figure 6.10 Experimental and predicted freezing temperatures for glucose and
maltose at various concentrations (method 1)........................................................ 156
Figure 6.11 Experimental and predicted boiling temperatures for glucose (method 1) ............ 156

xxi
Figure 6.12 Experimental and calculated heat of dilution in aqueous solutions of
MD 13 – 17 and MD 16 – 19 at 313.15 K (method 1)........................................... 157
Figure 6.13 Results of the predictions (method 1) for a
w
of maltodextrin
fractions (n is the polymerization number). Experimental data
from Radosta et al. (1989)...................................................................................... 158

ANEXO 1
Figura A.1 Viscosidades cinemáticas experimentais e calculadas para os
PEGs 8000 e 10 000 e simulação de viscosidades calculadas para
PEGs 6000, 12 000 e 15 000.................................................................................. 168
Figura A.2 Viscosidades calculadas e experimentais a 293,15 K e várias concentrações: ■,
PEG 400; □, PEG 600; •, PEG 1000; ○, PEG 1500; ▲, PEG 3350; —, valores
calculados ............................................................................................................... 173
Figura A.3 Viscosidades calculadas pelo modelo GC-UNIMOD e experimentais
a 298,15 K e várias concentrações ......................................................................... 173
Figura A.4 Predição das viscosidades de uma mistura quaternária.......................................... 174
xxiii

RESUMO

Métodos de contribuição de grupos têm sido ferramenta útil no cálculo de coeficientes de
atividade nos mais variados sistemas. Em sistemas aquosos, e mais especificamente naqueles
que possuem compostos biológicos como açúcares, aminoácidos, sais orgânicos, polímeros, etc.,
o uso de métodos de contribuição de grupos para a correlação e predição do equilíbrio de fases
tem crescido nos últimos anos. Em alimentos e em biotecnologia, os sistemas são geralmente
multicomponentes e por essa razão, a aplicação de métodos de contribuição de grupos torna-se
atraente partindo-se do princípio que, de posse dos parâmetros de interação entre grupos, é
possível calcular propriedades de sistemas mais complexos a partir de parâmetros obtidos da
correlação a dados experimentais de sistemas mais simples. Neste trabalho, os métodos de
contribuição de grupos ASOG, UNIFAC, VERS e UNIMOD foram empregados em diferentes
sistemas para a correlação e predição de propriedades físico-químicas. Os compostos estudados
foram: polióis, aminoácidos, maltodextrinas e polietileno glicóis (PEG)s, sendo os três primeiros
compostos geralmente encontrados em sistemas alimentícios, e o PEG é bastante utilizado em
sistemas aquosos bifásicos para o estudo do equilíbrio e separação de biomoléculas.
Propriedades físico-químicas como atividade de água, solubilidade, depressão do ponto de
congelamento, pH, espalhamento de luz, entalpia de diluição e viscosidade, são algumas das
propriedades determinadas e/ou retiradas da literatura usadas para a correlação e predição
empregando os modelos acima citados. A intenção deste trabalho foi então, testar os métodos de
contribuição de grupos para o cálculo de propriedades físico-químicas tratando também das
particularidades de cada sistema estudado, como a dissociação parcial dos aminoácidos, o efeito
de proximidade dos grupamentos hidroxilas nos polióis e a polidispersão das maltodextrinas.
Também são apresentados resultados da correlação e predição de viscosidades de sistemas
aquosos contendo PEGs por um modelo semi-empírico que não trata da contribuição de grupos
mas considera a hidratação das moléculas desses polímeros em meio aquoso.
xxiv


xxv

ABSTRACT

Group contribution methods have been used as a useful tool for calculating activity coefficients
in different types of systems. In aqueous systems, specifically those containing biological
compounds such as sugars, amino acids, organic salts, polymers, etc., the use of group
contribution models for correlating and predicting phase equilibrium has increased in the last
years. In the food and in the biotechnology areas, the systems are generally multicomponent, and
for this reason, the use of group-contribution methods becomes an attractive tool considering the
possibility of calculation of properties in complex systems with parameters adjusted for simple
ones. In this work, the group-contribution models ASOG, UNIFAC, VERS and UNIMOD were
utilized in different types of systems for the correlation and prediction of physical chemical
properties. The studied compounds were: polyols, amino acids, maltodextrins and polyethylene
(glycols) (PEG)s, which the first three compounds are found in food systems, and PEGs are wide
used in aqueous-two-phase systems to study equilibrium and separation of biomolecules.
Physical-chemical properties such as water activity, solubility, freezing point depression, pH,
light scattering, enthalpy of dilution and viscosity are some of the properties experimentally
determined and/or found in literature used in this work for the correlation and prediction using
the above cited models. Thus, the objective of this work was to test the group-contribution
models for calculating physical-chemical properties considering the particularities of each system
such as partial dissociation phenomena in aqueous amino acid systems, proximity effect of
hydroxyl groups in polyols, and polydispersity of maltodextrins. It is also presented in this work,
the results of viscosity correlation and prediction in aqueous PEG solutions by a semi-empirical
equation that does not consider group contributions but take into account the hydration of the
polymer molecules in aqueous media.
Capítulo 1

1

Capítulo 1

REVISÃO BIBLIOGRÁFICA




Capítulo 1

2

Introdução

3
Introdução
Métodos de contribuição de grupos são ferramentas úteis na modelagem e predição do
equilíbrio de fases em variados sistemas. Por essa razão, têm sido empregados em simulação de
processos industriais como a destilação e na modelagem de propriedades termodinâmicas de
sistemas mais complexos contendo compostos orgânicos. Dentre os modelos de contribuição de
grupos descritos na literatura, pode-se citar os modelos UNIFAC, ASOG e VERS que são
baseados em expressões para a energia livre de Gibbs. Esses modelos possuem vantagens, como
por exemplo, a de necessitar apenas de um pequeno número de parâmetros ajustáveis a serem
conhecidos, além de parâmetros estruturais que podem ser prontamente calculados ou retirados
de tabelas na literatura. Dessa forma, se existem parâmetros disponíveis para os grupos presentes
em uma mistura, é possível realizar a predição dos coeficientes de atividade dos compostos nessa
mistura ou de sistemas mais complexos. Métodos de contribuição de grupos vêm sendo bastante
empregados para descrever as propriedades termodinâmicas de misturas de compostos de
interesse da indústria alimentícia, farmacêutica e de processos biotecnológicos, como por
exemplo: açúcares, sais orgânicos, polímeros e aminoácidos.
Apesar de serem encontrados na literatura valores de parâmetros desses modelos para vários
grupos que formam as moléculas de compostos orgânicos e inorgânicos, o emprego desses
parâmetros em sistemas contendo compostos biológicos muitas vezes não gera bons resultados
por causa da complexidade dos sistemas envolvidos, da presença de compostos muito diferentes
nas misturas, e também das particularidades dos compostos biológicos, como hidratação,
dissociações, efeitos de proximidade de grupos polares, diferenças de tamanhos de moléculas
entre compostos homólogos, etc. Por isso, um ajuste dos parâmetros dos modelos torna-se muitas
vezes necessário para uma melhor descrição das propriedades físico-químicas de tais sistemas. A
proposta deste trabalho foi justamente aplicar métodos de contribuição de grupos para descrever
propriedades de sistemas aquosos contendo compostos presentes em alimentos (polióis,
aminoácidos e maltodextrinas), bem como polímeros empregados em separação de compostos
biológicos através de sistemas aquosos bifásicos (polietileno glicóis e maltodextrinas). A seguir,
são apresentados os resumos dos capítulos deste trabalho que se referem a trabalhos publicados e
a serem submetidos a revistas científicas da área. Também segue o resumo de um anexo com
alguns resultados que não foram publicados mas que são também parte deste trabalho.

Capítulo 1

4
1. Capítulo 1. Revisão Bibliográfica, apresentando os fundamentos termodinâmicos do
equilíbrio de fases e de outras propriedades físico-químicas, e os modelos de coeficiente
de atividade e de outras propriedades de interesse neste trabalho.
2. Capítulo 2. Determinação experimental e modelagem da atividade de água em sistemas
contendo quatro diferentes polióis. O trabalho apresenta resultados de correlações
empregando os modelos ASOG e UNIFAC, sendo que para o modelo UNIFAC foram
testadas duas diferentes formas de divisão de grupos para as moléculas envolvidas. A
justificativa de reajuste de parâmetros é que a presença de grupos polares muito próximos
(grupos hidroxilas consecutivos ligados à cadeia de carbonos da molécula) podem
provocar um efeito de proximidade intramolecular. Na correlação, dados de solubilidade
dos poliálcoois foram também utilizados. As atividades de água de misturas ternárias
foram empregadas em teste para avaliar a capacidade preditiva dos modelos com os
novos parâmetros ajustados.
3. Capítulo 3. Algumas propriedades de soluções aquosas contendo aminoácidos como
atividade de água, pH e densidade foram determinadas em três diferentes tipos de
solventes. O modelo UNIFAC-Lyngby de contribuição de grupos combinado com a
equação de Debye-Hückel foram empregados na correlação e modelagem das
propriedades determinadas neste trabalho, assim como outros dados de equilíbrio
apresentados na literatura. A dissociação parcial desses compostos em água e em dois
tampões foi considerada nas correlações.
4. Capítulo 4. Correlação e predição, pelo modelo GC-UNIMOD, de viscosidades
cinemáticas de misturas contendo polietileno glicóis, determinadas neste trabalho como
função da temperatura. As viscosidades de misturas binárias foram também empregadas
no cálculo da energia de ativação para o escoamento viscoso.
5. Capítulo 5. Determinação de viscosidades cinemáticas de misturas aquosas contendo
polietileno glicóis numa ampla faixa de massas molares, temperatura e concentração. Na
correlação, ao contrário dos outros capítulos, não foi empregado um modelo de
contribuição de grupos, mas sim uma correlação semi-empírica que descreveu bem as
viscosidades considerando o grau de hidratação das diferentes moléculas estudadas
Introdução

5
(correlação de Kumar). A predição das viscosidades de misturas com até cinco
componentes foi satisfatória.
6. Capítulo 6. Determinação de atividade de água, espalhamento de luz e entalpia de
diluição em sistemas contendo três diferentes maltodextrinas e correlação dos dados por
equação osmótica virial e o modelo VERS (Virial Equation with Relative Surface
Fractions). Os parâmetros do modelo VERS foram também empregados na predição da
atividade de água, depressão do ponto de congelamento e de ebulição em sistemas
contendo sacarídeos como glicose, maltose, maltotriose, etc.
7. Anexo 1. Resultados da correlação e predição das viscosidades cinemáticas de misturas
aquosas, contendo polietileno glicóis, pelo modelo GC-UNIMOD. Apesar das diversas
tentativas para melhorar os resultados de correlação e predição, o resultado geral foi
insatisfatório. Por esse motivo, optou-se por outra abordagem, apresentada no Capítulo
5, que gerou resultados consideravelmente melhores.

Capítulo 1

6
Objetivos
Este trabalho tem por objetivos determinar, modelar e predizer propriedades físico-químicas
como atividade de água (a
w
), pH, solubilidade, depressão do ponto de congelamento, entalpia de
diluição, elevação do ponto de ebulição e viscosidade de soluções contendo componentes
presentes em alimentos e empregados em processos biotecnológicos. Foi estudado o
comportamento de sistemas contendo polióis, aminoácidos e polímeros naturais em misturas com
dois ou mais componentes. Esses compostos possuem diversas aplicações industriais, incluindo
usos como modificadores de textura, umectantes, plasticizantes e outros, além de suas
propriedades em solução serem importantes no projeto de equipamentos e processos. Apesar de
importantes, poucos dados físicos e termodinâmicos para esses compostos estão disponíveis na
literatura. Com base em um banco de dados que incluiu os dados experimentais determinados
neste trabalho e obtidos na literatura, os métodos de contribuição de grupos UNIFAC, ASOG e o
modelo VERS foram utilizados na modelagem de propriedades termodinâmicas através do ajuste
de parâmetros dos modelos – quando necessário – e tratamento matemático adequado para cada
conjunto de dados.
Fundamentos e modelos termodinâmicos

7
1. Revisão Bibliográfica
1.1 Fundamentos e modelos termodinâmicos
Neste capítulo serão comentados os fundamentos termodinâmicos e modelos que neles se
baseiam para melhor compreensão da modelagem, assim como do cálculo do equilíbrio de fases
dos sistemas aquosos estudados. Como foram estudados diferentes compostos – poliálcoois,
aminoácidos e polímeros – os procedimentos no tratamento e cálculo das propriedades serão
apresentados em detalhe para cada sistema nos itens abaixo, e nos capítulos correspondentes a
cada composto.
1.1.1 Bases Termodinâmicas
Os modelos baseados na energia livre de Gibbs de excesso são normalmente empregados
para a descrição do equilíbrio de fases.
A energia livre de Gibbs G depende das variáveis temperatura, pressão e número de mol dos
componentes que formam a mistura.
) n ... n , n , P , T ( G G
N , 2 1
= (1.1)
Na forma diferencial a equação fundamental de Gibbs é dada por:

=
µ + + − =
N
i
i i
dn VdP SdT dG
1
(1.2)
em que S é a entropia, uma quantidade de estado extensiva do sistema:
j
n , P
T
G
S
|
¹
|

\
|


= −
(1.3)
e o volume, também uma quantidade extensiva é dado por:
j
n T
p
G
V
,
|
|
¹
|

\
|


=
(1.4)
e o potencial químico µ
i
é a energia livre de Gibbs parcial molar do componente i:
Capítulo 1

8
i j
n , P , T
i
i
n
G

|
|
¹
|

\
|


= µ
(1.5)
Já que a função G é homogênea de primeira ordem em massa ou número de mol, pode-se
aplicar o teorema de Euler para obter a seguinte função:
G n
n
G
n
N
i
i i
n , P , T
i
N
i
i
i j
= µ =
|
|
¹
|

\
|


∑ ∑
= =

1 1
(1.6)
e,
∑ ∑
= =
µ + µ =
N
i
i i
N
i
i i
d n dn dG
1 1
(1.7)
Das equações 1.2 e 1.7 obtém-se a equação de Gibbs-Duhem, uma relação fundamental da
termodinâmica:
0
1
= µ + −

=
N
i
i i
d n VdP SdT
(1.8)
e considerando um sistema a temperatura e pressão constantes, a equação 1.8 resume-se a:
0
1
= µ

=
N
i
i i
d n
(1.9)
Existe equílibrio termodinâmico de fases em um sistema – que não possui interações com a
vizinhança – quando não se observa macroscopicamente mudanças nas suas variáveis num
determinado período e a inexistência de fluxos líquidos. A condição para o equilíbrio de fases de
um sistema à temperatura e pressão constantes pode ser escrita com o auxílio da segunda lei da
termodinâmica como:
mínima constante = ⇒ = G P , T
(1.10)
Partindo-se da condição de equilíbrio de fases apresentada na equação 1.10 acima, tem-se que:
Fundamentos e modelos termodinâmicos

9
0 e 0
2
> = G d dG
(1.11, 1.12)
que permite concluir que o equilíbrio corresponde a:
π
π
π
µ = = µ ′ ′ ′ =

µ =

µ
= = ′ ′ ′ = ′ ′ = ′
= = ′ ′ ′ = ′ ′ = ′
i i i i
...
P ... P P P
T ... T T T
(1.13)
em que o sobrescrito refere-se a fase e o subescrito ao componente do sistema.
1.1.2 Modelos para energia livre de Gibbs
Como foi apresentado na equação 1.2, a energia de Gibbs é função do potencial químico.
Numa solução ideal, o potencial químico do componente i é dado por:
i i
x RT
i
ln
o id
+ µ = µ
(1.14)
onde
o
i
µ é o potencial químico do componente i puro, à mesma temperatura e pressão da solução
e x
i
é a fração molar do componente i.
Já o potencial químico de um componente i numa solução real é calculado como a soma de
duas contribuições, uma ideal e outra de excesso, como a seguir:
43 42 1 4 4 3 4 4 2 1
E id
ln ln
o
i i
i i i i
RT x RT
µ µ
γ + + µ = µ
(1.15)
em que γ
i
é o coeficiente de atividade do componente i e ( )
j
n , T , P
i i
n G ∂ ∂ = µ
E E
o termo de
excesso.
O termo de excesso – que inclui os coeficientes de atividade – descreve o comportamento
real do componente. Dessa forma, os coeficientes de atividade possuem um significado decisivo
na modelagem do equilíbrio de fases. Neste trabalho serão investigados alguns modelos que
relacionam a energia livre em excesso para o cálculo de propriedades físico-químicas em
sistemas aquosos contendo diferentes componentes. Por isso será apresentado a seguir, um
resumo sobre os modelos para o cálculo de coeficientes de atividade.
Capítulo 1

10
Os modelos de energia livre de Gibbs em excesso podem ser divididos entre empíricos e
teóricos, de acordo com suas abordagens fundamentais.
1. Abordagem empírica
Com relação à abordagem empírica poder-se-ia citar uma série de trabalhos que se utilizaram
de ajuste polinomial aos dados experimentais para obter uma equação simples que estime os
coeficientes de atividade. Um exemplo desse tipo de ajuste é a abordagem de Porter (1920) que
leva ao cálculo de coeficientes de atividade de compostos bem semelhantes numa mistura.
Extensões desse tipo de abordagem são as equações de Margules (Prausnitz, 1969) e de Redlich e
Kister (Redlich and Kister, 1948).
2. Abordagem teórica
Aqui pode-se citar três diferentes grupos de modelos: um que emprega a teoria da solução
regular; outro a teoria da solução atérmica e um terceiro grupo que se utiliza das duas teorias
anteriores como base para compor os modelos.
(a) Teoria da solução regular
A energia livre de Gibbs pode também ser escrita – com base nas tranformadas de Legendre –
da seguinte forma:
E E E E E E
TS PV U TS H G − + = − = (1.16)
De acordo com a experiência, o volume e a entropia de excesso podem ser considerados
desprezíveis em misturas apolares constituídas por compostos semelhantes. Essas são
considerações da teoria da solução regular.
E E
E
E
U G
S
V
=

=
=
43 42 1
0
0
(1.17)
A teoria de soluções regulares chamada também teoria de Scatchard-Hildebrand (Prausnitz,
1986) fornece coeficientes de atividade maiores ou iguais a 1, isto é, uma solução regular
pode exibir apenas desvio positivo da idealidade segundo a Lei de Raoult. Nessa abordagem
Fundamentos e modelos termodinâmicos

11
considera-se que a interação entre duas moléculas em solução depende apenas da distância
entre essas moléculas sendo aleatória a distribuição das moléculas em posição e orientação.
(b) Teoria da solução atérmica
Ao contrário da teoria de soluções regulares, a teoria da solução atérmica não considera
interações entre moléculas ou espécies químicas em soluções e sim efeitos de configuração.
É denominada atérmica, pois o excesso de entalpia é desprezado, e dessa forma, os
coeficientes de atividade são independentes da temperatura.
E E E
TS G H − = ⇒ = 0
(1.18)
2
,
/
RT
H
T
RT G
E
n p
E
j
− =
|
|
¹
|

\
|


(1.19)

=
≠ γ = ⇒ =
N
i
i i
E
E
T f n
RT
G
H
1
) ( ln 0
(1.20)
A entropia de excesso pode ser calculada com o auxílio de modelos que consideram as
diferenças de tamanho entre as moléculas num sistema. Um exemplo desse tipo de
abordagem é o modelo de Flory-Huggins (Flory, 1941) que prediz desvios negativos da
idealidade segundo a lei de Raoult.
(c) Abordagem combinada
Essa abordagem utiliza-se das teorias das soluções regulares e atérmicas, considerando
então efeitos de interação e configuração nos modelos para o cálculo dos coeficientes de
atividade. Um exemplo típico desse grupo é o modelo UNIQUAC (Universal Quasi-
Chemical Equation) proposto por Abrams and Prausnitz (1975). Alguns desses modelos
foram estendidos para compor os chamados modelos de contribuição de grupos. Tais
modelos consideram que o comportamento dos compostos em solução é o resultado das
interações e das diferenças de tamanho e forma dos grupos funcionais dessas substâncias. Os
modelos desse tipo mais conhecidos são o ASOG (Analytical Solution Of Groups) de Kojima
and Tochigi (1979) e o modelo UNIFAC (UNIQUAC Functional-Group Activity Coefficient)
de Fredenslund and Jones (1975). Pode-se citar também nesse grupo o modelo VERS (Virial
Capítulo 1

12
Equation with Relative Surface Fractions) desenvolvido por Großmann (1994) e baseado no
desenvolvimento virial da equação de Pitzer (Pitzer, 1991) para soluções de eletrólitos. A
seguir, estão apresentadas as equações dos modelos de contribuição de grupos empregados
neste trabalho: UNIFAC, ASOG, VERS e GC-UNIMOD (esse último baseado no modelo de
contribuição de grupos UNIFAC, utilizado para o cálculo de viscosidades de misturas e
desenvolvido por Cao et al., 1993).
1.2 Métodos de contribuição de grupos
1.2.1 Para cálculo de coeficientes de atividade
O cálculo de coeficientes de atividade por contribuição de grupos foi primeiramente sugerido
em 1925 por Langmuir (Prausnitz, 1986) mas o emprego de tais métodos expandiu-se com o
aumento do número de dados experimentais e o surgimento de modelos como o ASOG
(Analytical Solutions of Groups) e o UNIFAC (Universal Functional Activity Coefficient).
Abaixo, são apresentadas as equações desses modelos e também do modelo VERS (Virial
Equation with Relative Surface Fractions).
Modelo UNIFAC
Nesse modelo, o cálculo do coeficiente de atividade de um componente é dividido em duas
partes, como no modelo UNIQUAC: uma parte fornece a contribuição devido às diferenças no
tamanho e na forma molecular e a outra está relacionada às interações energéticas entre grupos
como mostrado abaixo, para o componente i de uma mistura:
residual ial combinator
ln
R
i
C
i i
ln ln γ + γ = γ
(1.21)
I. Parte combinatorial
No modelo UNIFAC, o termo combinatorial da equação UNIQUAC é usado diretamente.
Apenas propriedades dos compostos puros entram nesta parte da equação:

φ
− +
φ
θ
+
φ
= γ
j
j j
i
i
i
i
i
i
i
i C
i
l x
x
l ln q
z
x 2
ln ln
(1.22)
Métodos de contribuição de grupos

13
) r ( ) q r (
z
l
i i i i
1
2
− − − =
(1.23)
onde z é chamado de número de coordenação e representa o número de locais nos quais pode
haver interação entre moléculas. Geralmente, para líquidos em condições moderadas, o valor de
z é próximo de 10.

= θ
j
j j
i i
i
x q
x q


= φ
j
j j
i i
i
x r
x r
(1.24, 1.25)
em que θ
i
e φ
i
são frações de área e volume do componente i, respectivamente.
Os parâmetros r
i
e q
i
do componente puro são calculados como a soma dos parâmetros de volume
(R
k
) e de área (Q
k
) do grupo k na molécula i. Estes, por sua vez, são obtidos a partir do volume e
área de van der Waals V
k
e A
k
, dados por Bondi (1968):
∑ =
k
k
i
k i
R v r ∑ =
k
k
i
k i
Q v q (1.26, 1.27)
17 15, / V R
k k
=
9
10 5 2 × = , / A Q
k k
(1.28, 1.29)
em que v
k
i
é um número inteiro que representa o número de grupos do tipo k na molécula i. Os
fatores de normalização 15,17 e 2,5×10
9
são dados por Abrams and Prausnitz (1975).
II. Parte Residual
A parte residual do modelo UNIQUAC é substituída no modelo UNIFAC pelo conceito de
solução de grupos. A solução é tratada como uma mistura de grupos funcionais com interações
energéticas (Wilson and Deal, 1962):
∑ Γ − Γ = γ
k
i
k k
i
k
R
i
v ] ln [ln ln
(1.30)
Γ
k
é o coeficiente de atividade residual do grupo k e Γ
k
i
o coeficiente de atividade residual do
grupo k numa solução referência contendo apenas moléculas do tipo i. Γ
k
pode ser calculado
como:
Capítulo 1

14
(
(
¸
(

¸


|
|
¹
|

\
|
∑ ψ θ ψ θ −
|
|
¹
|

\
|
∑ ψ θ − = Γ
m n
nm m km m
m
mk m k k
ln Q ln 1
(1.31)

= θ
n
n n
m m
m
X Q
X Q

∑∑

=
i k
i
i
k
i
i
m
m
x v
i
x v
X (1.32, 1.33)
em que θ
m
é a fração de área do grupo m, X
m
é a fração do grupo m na mistura e ψ
nm
é o
parâmetro de interação entre os grupos m e n, e é dado por:
[ ] ) ( exp T a exp
RT
U U
nm
nn mn
nm
− =
(
¸
(

¸

|
¹
|

\
| −
− = ψ
(1.34)
U
mn
é uma medida da energia de interação entre os grupos m e n, e a
mn
representa os parâmetros
binários de interação de grupos com unidades em Kelvin e a
mn
≠a
nm
.
Modelo ASOG
No modelo ASOG o coeficiente de atividade de um componente i na solução é calculado pela
soma das contribuições devido às diferenças nos tamanhos moleculares (γ
i
FH
) e das interações
entre grupos (γ
i
G
) que descrevem as diferenças de forças intermoleculares:
G
i
FH
i i
ln γ + γ = γ ln ln
(1.35)
A contribuição devido às diferenças nos tamanhos das moléculas é calculada aplicando-se
equação similar à equação de Flory e Huggins citados por Prausnitz et al. (1986).


|
|
|
|
|
|
¹
|

\
|

+ = γ
=
n
j
FH
i i
FH
i
n
j
FH
j j
FH
i FH
i
v x
v
v x
ln ln
1 = 1
v
1
(1.36)
A parte residual, que também contribui para o cálculo do coeficiente de atividade, é dada por:
Métodos de contribuição de grupos

15
] [
i
k k
k
i
k
G
i
ln ln v ln Γ − Γ ∑ = γ
(1.37)
O coeficiente de atividade do grupo k (Γ
k
) é função da temperatura e da fração de cada grupo na
solução; ele é definido analiticamente pela equação de Wilson (1964), como a seguir:
∑ ∑

− + − = Γ
n n
m
nm m
nk n
kn n k
a X
a X
a X ln ln 1
(1.38)
∑ ∑

=
j k
i
k j
j
j
n j
n
v x
v x
X
(1.39)
T
n
m a ln
kn
kn kn
+ =
(1.40)
em que v
i
FH
é o número de átomos não hidrogênio na molécula i; v
k
i
o número de átomos não
hidrogênio no grupo k da molécula i; x
j
a fração molar do componente j; Γ
k
e Γ
k
i
são os
coeficientes de atividade do grupo k no sistema e no estado padrão (componente i puro),
respectivamente. O parâmetro a
kn
representa as interações energéticas entre os grupos k e n
(a
kn
≠a
nk
); m
kn
e n
kn
são parâmetros independentes da temperatura característicos das interações
entre os grupos k e n.
Modelo VERS
Neste trabalho, um modelo semi-empírico é empregado para descrever os dados de atividade
de água e entalpia de diluição.
Como de costume, o modelo começa com a energia livre de Gibbs em excesso G
E
. O
excesso é definido pela normalização assimétrica como mostrado abaixo:
1 1 → γ →
w w
x para o solvente (1.41)
1 0 → γ →
i i
x para os solutos (1.42)
Capítulo 1

16
A influência da pressão sob G
E
é desprezível. Então, quando um componente 1 não iônico é
dissolvido em água, o excesso da energia livre de Gibbs é expressa da maneira mais simples:
2
)] 1 ( [
11
. conc A
RT n
E
G
w
=
(1.43)
A
11
representa a interação entre duas moléculas do componente 1 dissolvidas em água. Esse
coeficiente é multiplicado pela probabilidade dessa interação, que é considerada como sendo o
quadrado da concentração de soluto. A concentração do soluto é expressa como:
w w
M
. conc
Θ
Θ
=
1
1000
) 1 (
(1.44)

= Θ
j
j j
i i
i
q m
q m

s componente
os todos
(1.45)
em que Θ
1
é a fração de área do componente 1, Θ
w
a fração de área da água, m
i
a molalidade do
componente i e q
i
o parâmetro de área do componente i. Das equações (1.43)-(1.44) resulta:
11
1
2
1000
A
M RT n
E
G
w w w
|
|
¹
|

\
|
Θ
Θ
=
(1.46)
A expressão (1.46) pode ser facilmente estendida para interações ternárias assim como para mais
solutos, da seguinte forma:
ijk
w
k
w
j
w i w j w k w
i
w
ij
w
j
w i w j w
i
w w
B
M
A
M RT n
E
G
Θ
Θ
Θ
Θ
∑ ∑ ∑
Θ
Θ
|
|
¹
|

\
|
+
Θ
Θ
∑ ∑
Θ
Θ
|
|
¹
|

\
|
=
≠ ≠ ≠ ≠ ≠
3
1000 1000
2
(1.47)
As somas na equação (1.47) devem ser feitas para todas as espécies de solutos presentes. Para
considerar a influência da massa molar do soluto, o parâmetro de área q
i
é calculado por método
de contribuição de grupos:
Métodos de contribuição de grupos

17
∑ =
l
l
i
l i
Q v q
grupos
os todos
) (
(1.48)
e para os parâmetros A
ij
e B
ijk
:
∑ ∑ Θ Θ =
l m
lm
j
m
i
l ij
a A

grupos todos

grupos todos
) ( ) (
(1.49)
∑ ∑ ∑ Θ Θ Θ =
l m n
lmn
k
n
j
m
i
l ijk
b B

grupos todos

grupos todos

grupos todos
) ( ) ( ) (
(1.50)
i
l i
l
i
l
q
Q
) ( ) (
ν = Θ
(1.51)
O parâmetro v
l
(i)
representa o número de grupos l na molécula de soluto i. Q
t
é o parâmetro de
área do grupo l, a
lm
e b
lmn
são parâmetros de interação de grupos. Esse modelo foi denominado
VERS (Virial Equation with Relative Surface Fractions).
Para descrever os dados calorimétricos e a influência da temperatura na atividade de água, a
seguinte expressão empírica, que apresenta a influência da temperatura nos parâmetros de
interação binários a
lm
foi empregada:
) ln( )) ( 1 )( (
0
) 2 (
0
) 1 ( ) 0 (
T / T T / T K / T a
lm lm lm lm
β + − β + β =
(1.52)
em que T
0
é a temperatura de referência igual a 25
o
C. Assume-se que todos os parâmetros são
simétricos, ou seja, que
) ( ) ( j
ml
j
lm
β = β e
nml nlm mnl ln m m ln lmn
b b b b b b = = = = = .
A partir das equações acima, pode-se então escrever equações para a atividade de água (a
w
) e
atividade do soluto (
*
i,m
a ):
Capítulo 1

18
ijk
k
j
i j k
i
j
j
i j
i
i
i
w
B
M
i
A
M
m
M
a ln
1 1 1 1 1 1 1 1 1 1 1
1
1
2
1000
2
1000
1000 Θ
Θ
Θ
Θ
∑ ∑ ∑
Θ
Θ
|
|
¹
|

\
|

Θ
Θ
∑ ∑
Θ
Θ
− − =
≠ ≠ ≠ ≠ ≠

(1.53)
ijk
j k
k
j
i
ij
j
j
i
i
m i
B
q
q
M
A
q
q
M
m a ∑ ∑
Θ
Θ
Θ
Θ
|
|
¹
|

\
|
+ ∑
Θ
Θ
|
|
¹
|

\
|
+ =
≠ ≠ ≠ 1 1 1 1 1 1 1 1 1
1
,
2
1000
3
1000
2 ln
*
ln (1.54)
assim como para a entalpia molar de excesso da água:
j
n , p
E
i E
w
T
) T / (
T h
|
|
|
¹
|

\
|

µ ∂
− =
2
(1.55)
T
A
M
T
RT
h
ij
w i w j w
j
w
i
w
E
w


∑ ∑
Θ
Θ
Θ
Θ
|
|
¹
|

\
|
=
≠ ≠
2
1000
(1.56)
e do soluto:
T
A
q
q
M
T
RT
h
ij
w j w
j
w
i
w
E
w i



Θ
Θ
|
|
¹
|

\
|
− =


2
1000
2 (1.57)
com:
|
|
|
¹
|

\
|
β
+ β ∑ ∑ Θ Θ =


T T
A
lm
lm
j
m
i
l
ij
l m
) 2 (
) 1 ( ) ( ) (

grupos todos

grupos todos
(1.58)
1.2.2 Para cálculo de viscosidade de misturas
Modelo GC-UNIMOD
O modelo GC-UNIMOD proposto por Cao et al. (1993) é baseado no método de contribuição
de grupos UNIFAC. O GC-UNIMOD é também um modelo de contribuição de grupos
funcionais empregado no cálculo da viscosidade de misturas líquidas multicomponentes. Nesse
Métodos de contribuição de grupos

19
modelo, a viscosidade de uma mistura pode ser calculada como a soma de duas partes como no
modelo UNIFAC: uma combinatorial e uma residual. Para os cálculos, são normalmente
utilizados parâmetros do UNIFAC para o equilíbrio líquido-vapor (UNIFAC-VLE). Segue
abaixo uma apresentação das equações matemáticas correspondentes ao modelo GC-UNIMOD
para o cálculo da viscosidade cinemática em que: ν corresponde à viscosidade cinemática da
mistura, ξ
i
C
à contribuição combinatorial e ξ
i
R
à residual do componente i.
) (
1
∑ ξ ξ = ν
=
+
n
i
R
i
C
i
) ln( (1.59)
I. Parte Combinatorial
|
|
¹
|

\
|
φ
φ +
|
|
¹
|

\
|
ν φ = ξ
i
i
i
i
i i
C
i
x
ln
M
M
2 ln
(1.60)
k
k
i
k i
R v r ∑ =
(1.61)

= φ
n
j
j j
i i
i
r x
r x
1 =
(1.62)
em que M
i
, M, r
i
, x
i
, R
k
, v
k
i
, v
i
e φ
i
são respectivamente, massa molar do componente i, massa
molar da mistura, número de segmentos na molécula i, fração molar do componente i na mistura,
parâmetro de volume do grupo k, número de grupos k na molécula i, viscosidade cinemática do
líquido i e fração de volume do componente i.
II. Parte Residual
] [
i
ki ki
k
i
k
R
i
v Ξ − Ξ ∑ = ξ
(1.63)
Ξ ΞΞ Ξ
ki
e Ξ ΞΞ Ξ
i
ki
correspondem respectivamente à viscosidade residual do grupo k para o componente i
quando numa mistura e à viscosidade residual do grupo k no componente i quando em uma
solução contendo apenas o componente i.
Capítulo 1

20
( ) ∑ ψ θ φ − = Ξ
k
km km i
visc
mi
m
m
mi
N
R
Q
ln (1.64)
∑ Ψ θ
Ψ θ
= θ
k
kn k
mn m
mn
(1.65)
|
¹
|

\
|
= Ψ
T
mn
mn
a
exp
(1.66)

= θ
k
k k
m m
m
Q X
Q X
(1.67)
|
|
¹
|

\
| −


=
z
r r q
Q N
i i i
k
visc
ki
1
2
(1.68)
k
k
i
k i
Q v q ∑ =
(1.69)
a
mn
, q
i
, Q
k
, X
k
ψ
mn
, θ
k
, θ
mn
correspondem a: parâmetro de interação binária entre os grupos m e n;
parâmetro de área para molécula i; parâmetro de área superficial para o grupo k; fração molar do
grupo k; parâmetro de interação binária entre os grupos m e n; fração de área do grupo k e
composição local dos grupos.
A equação abaixo representa o cálculo global da viscosidade, com as expressões matemáticas
para as contribuições combinatorial e residual.

(
(
¸
(

¸

∑ |
¹
|

\
|
Ξ − Ξ +
|
|
¹
|

\
|
φ
φ +
|
|
¹
|

\
|
ν φ ν
n
i k
i
ki ki
i
k
i
i
i
i
i
v
x
M
M
i
1 =
ln 2 ln = ) ln(
(1.70)
As viscosidades dos compostos puros requeridas pelo modelo GC-UNIMOD podem ser retiradas
da literatura (quando houver dados) ou calculadas por modelos do tipo proposto por Joback and
Reid (1987). Alguns trabalhos do grupo de pesquisa em que esta fase foi realizada estão
relacionados com a determinação experimental e modelagem de propriedades físicas, como por
exemplo: Valeri and Meirelles (1997); Chumpitaz et al. (1999) e Cruz et al. (2000). Trata-se
Métodos de contribuição de grupos

21
agora de estender tal abordagem a misturas multicomponentes que de fato são encontradas na
prática industrial.
Equação de Kumar
Apesar de não se tratar de um modelo de contribuição de grupos, as equações da correlação
de Kumar, empregadas nas investigações da viscosidade de misturas de polietileno glicóis no
Capítulo 5, são apresentadas neste item. Primeiramente a correlação foi proposta para o cálculo
de viscosidades de soluções de sais (Kumar, 1993) e posteriormente, também empregada no
cálculo de viscosidades de misturas de sais e açúcares (Pereira, 2001).
A equação de Kumar para viscosidade foi desenvolvida assumindo que um sal AB, quando
dissolvido em um solvente S, forma um complexo solvatado como, a seguir:
hS AB S h AB ⋅ ⇔ +
(1.71)
em que h é o número de moléculas de solvente.
Se N
1
, N
2
e N
c
são, respectivamente, o número de moles de sal, solvente e complexo solvatado, as
frações molares do sal (x
1
) e do complexo solvatado (x
c
) podem ser escritas como:
) (
2 1 1 1
N N N x + =
(1.72)
[ ] ) (
2 c c c c
N h N N N x − + =
(1.73)
Assumindo que N
1
=N
c
, pode-se obter N
2
da equação 1.72 como:
1 1 2
) - (1 x x N N
c
=
(1.74)
Substituindo N
2
da equação 1.74 na equação 1.73:
) 1 (
1 1
h x x x
c
− =
(1.75)
Seguindo a proporcionalidade entre viscosidade e concentração proposta por Harned and Owen
(1965):
) - (1 B 1) - (
1 1
h x x x
c rel
= ∝ η
(1.76)
Capítulo 1

22
em que, η
rel
=η/η
0
; η e η
0
são as viscosidades da solução e do solvente, respectivamente e B é
uma constante. A equação 1.76 pode também ser reescrita como:
) (B 1 B 1) - 1/(
1
x / / h
rel
+ − = η
(1.77)
A equação acima é a expressão proposta para correlacionar viscosidade com fração molar do
sal na solução. O parâmetro h indica a solvatação ou hidratação, se o solvente é água e o outro
parâmetro, B, representa as interações entre íon e solvente como discutido por Feakins and
Laurence (1966).
1.3 Compostos orgânicos de interesse na área de alimentos e biotecnologia
O crescente avanço tecnológico tem despertado interesse para o desenvolvimento de
processos eficientes na indústria, em particular na área biotecnológica e na indústria de alimentos.
Para o projeto de processos de purificação, recuperação de biomoléculas ou mesmo
desenvolvimento de produtos é geralmente necessário o conhecimento de propriedades físicas
dos mesmos para a otimização de operações na indústria.
Dentro desse contexto, o estudo do equilíbrio de fases desempenha papel importante em
processos de extração líquido-líquido como nos sistemas aquosos bifásicos, SABs (técnica
comumente empregada para separação e purificação de biomoléculas), processos de cristalização
e crioconcentração que envolvem equilíbrio sólido-líquido, estudos da atividade de água
(equilíbrio líquido-vapor) como parâmetro de controle de qualidade em alimentos, e também na
tecnologia de alimentos já que tanto os produtos naturais quanto os processados apresentam
microestrutura complexa e multifásica.
Dessa forma, deve-se enfatizar a importância do estudo do comportamento de compostos
alimentícios em soluções aquosas. O conhecimento de propriedades termodinâmicas de sistemas
aquosos simples pode levar a uma melhor compreensão de biomoléculas em seus ambientes
naturais. Além de compostos orgânicos, outro material estudado neste trabalho, também
interessante e muito utilizado nas mais diversas áreas industriais e investigações científicas –
como, por exemplo, em sistemas aquosos bifásicos – são os polietileno glicóis.
Dentre muitos materiais comumente encontrados em biotecnologia e na indústria de
alimentos estão polímeros, aminoácidos, triglicerídeos, açúcares, ácidos orgânicos, etc. O
Compostos orgânicos de interesse na área de alimentos e biotecnologia

23
conhecimento de propriedades físico-químicas tais como solubilidade, atividade de água, ponto
de congelamento, pH, massa molar e parâmetros de interação soluto-solvente de misturas que
incluem esses compostos é necessário sob o ponto de vista de projeto de equipamentos e
processos e também para o melhor entendimento da formação da estrutura no alimento. Do
ponto de vista da simulação de processos, o uso de modelos matemáticos existentes e
consolidados na literatura torna-se uma ferramenta útil para o cálculo das propriedades de
interesse. Nos próximos itens, segue-se uma descrição rápida dos compostos que são utilizados
neste trabalho – polióis, aminoácidos, polietileno glicóis e maltodextrinas – para esclarecimento
de suas estruturas químicas, propriedades físico-químicas e empregos na indústria e pesquisa. A
seguir, são também descritos os métodos de cálculo das propriedades físico-químicas que se
deseja estudar.
1.3.1 Polióis
Polióis são polihidróxi-álcoois empregados em diversas áreas, incluindo a de alimentos, a
farmacêutica e a de cosméticos. Na indústria de alimentos, por exemplo, soluções de sorbitol são
usadas como umectantes e plasticizantes em algumas formulações de alimentos e filmes
comestíveis (Moreton and Armstrong, 1998). O poder adoçante é outra importante característica
desses compostos, que são utilizados na confecção de doces dietéticos, produtos de higiene e
fármacos (Kirk and Othmer, 1983). Em geral, são produzidos industrialmente a partir da
hidrogenação catalítica de carboidratos comestíveis ou são produzidos por processos biológicos
(Billaux et al., 1991). Existe atualmente um grande interesse em se produzir xilitol (poliálcool)
por via microbiológica (Azuma et al., 2000; Silva et al., 1998). Dessa forma, o emprego de
métodos para a estimativa de dados como solubilidade apresenta-se como uma ferramenta útil no
planejamento de processos de recuperação de compostos - como xilitol – produzidos por
fermentação.
1.3.2 Aminoácidos
Dentre vários compostos bioquímicos, os aminoácidos são de grande interesse para a
pesquisa devido à sua simplicidade e importância para o entendimento do comportamento de
moléculas como peptídeos e proteínas em soluções mais complexas. Em geral, os aminoácidos
são produzidos por microorganismos em meio aquoso contendo solutos como sais, ácidos
orgânicos, etc.
Capítulo 1

24
Processos de separação baseados na precipitação e cristalização têm sido largamente
utilizados para concentração e purificação de biomoléculas (Cussler et al., 1989). De acordo com
Eyal and Bressler (1993), o custo de separação e concentração de biomoléculas a partir de meios
aquosos diluídos, nos quais eles são geralmente produzidos, pode alcançar 90% do custo total de
obtenção de bioprodutos. No projeto de equipamentos e processos é necessário o conhecimento
de algumas propriedades físico-químicas como solubilidade, atividade de água e a influência do
pH sobre essas propriedades em misturas contendo biomoléculas. Dessa forma, é interessante o
estudo de propriedades físico-químicas de sistemas contendo tais compostos a várias
concentrações e pHs.
1.3.3 Polietileno glicóis
Polietileno glicóis (PEGs) são polímeros de cadeia linear formados por unidades de
oxietileno, cujas características principais são alta solubilidade em água, baixa toxicidade e boa
estabilidade. Dependendo do tamanho da cadeia, apresentam-se sólidos (MM 3000 a 20000),
semi-sólidos (MM 1000 a 2000) ou líquidos viscosos (MM 200 a 700) à temperatura ambiente
(Davidson, 1980). Estas propriedades têm favorecido o uso dos PEGs em diversas aplicações
comerciais das indústrias farmacêuticas, cosmética, química e de alimentos. Na área
biotecnológica são empregados em sistemas aquosos bifásicos para a separação e purificação de
biomoléculas em meios biocompatíveis, já que o PEG possui baixa toxicidade (Chirife e Ferro
Fontan, 1980; Coimbra et al., 1995; Silva e Meirelles, 2000a, b; Alves et al., 2000, Sé e Aznar,
2002). Dessa forma, o conhecimento do comportamento reológico de sistemas contendo PEGs é
importante para avaliação, otimização e também projeto de equipamentos.
A estrutura química dos PEGs é representada por:
H O) CH (CH HO
n 2 2
− −
onde n representa o grau de polimerização.
O polímero de mais alta massa molar produzido comercialmente como um composto puro é o
tetraetilenoglicol com n=4. À medida que o grau de polimerização aumenta, não é mais possível
a separação de compostos puros por destilação devido à baixa volatilidade dos mesmos e, dessa
forma, são produzidas misturas de moléculas com várias massas molares. O polímero é então
caracterizado por um grau de polimerização médio. Comercialmente são identificados por
Compostos orgânicos de interesse na área de alimentos e biotecnologia

25
números que indicam suas massas molares aproximadas, como por exemplo, PEG 400, PEG
1000, etc.
1.3.4 Maltodextrinas (MD)
Maltodextrinas são polissacarídeos solúveis em água que têm sido empregados numa ampla
variedade de produtos industriais nas áreas de alimentos e farmacêutica. Atualmente, a
necessidade de proteção do meio ambiente coloca os polímeros biodegradáveis, como os
polissacarídeos, em vantagem para fins industriais (Swift, 1998). Os polissacarídeos apresentam
distribuição de massa molar, geralmente representada pelas massas molares médias: nominal
(M
n
) ou mássica (M
w
). Essas massas molares são calculadas como:


=
i
i
i
i i
n
M n
M
n
(1.78)


=
i
i
i
i i
m
M m
M
w
(1.79)
em que n
i
é número de moles e m
i
a massa da espécie i. A partir desses valores pode-se também
definir o quociente M
w
/M
n
denominado índice de polidispersão que é uma medida da amplitude
da distribuição.
As MD apresentam-se como uma mistura de sacarídeos com uma ampla distribuição de
massas molares entre polissacarídeos e oligossacarídeos. Suas massas molares variam de 900 a
18000 g/mol. A ampla distribuição de massas molares confere às maltodextrinas variadas
características relacionadas com propriedades físico-químicas como viscosidade, pressão de
vapor, capacidade de formação de gel, etc. (Kasapis et al., 1993; Mothé e Rao, 1999).
Dentre os polissacarídeos, as maltodextrinas são polímeros de grande interesse comercial
devido à sua alta solubilidade em água (Gliksmann, 1986; Marchal, 1999) e aplicação potencial
em sistemas aquosos bifásicos (Silva e Meirelles, 2000 a,b). A estrutura química da
maltodextrina é apresentada a seguir:
Capítulo 1

26

Figura 1.1 Estrutura química da maltodextrina (x=ligação α-1,6 com possibilidade de
ramificação; y= ligação α-1,4 para a formação de cadeia linear).
1.4 Propriedades físico-químicas estudadas
No presente trabalho, métodos de contribuição de grupos foram empregados para a
estimativa de propriedades como atividade de água, solubilidade, depressão do ponto de
congelamento, elevação do ponto de ebulição, pH, espalhamento de luz, entalpia de diluição e
viscosidade. As equações termodinâmicas para o cálculo dessas propriedades são apresentadas a
seguir, e o tratamento termodinâmico mais detalhado nos capítulos seguintes referentes a cada
um dos compostos estudados.
1.4.1 Atividade de água
O conceito termodinâmico de atividade de água pode ser relacionado à razão entre a
fugacidade da água numa amostra (f
w
) e a fugacidade da água pura (f
w
o
) à mesma temperatura e
pressão:
o
) , (
o
) , , (
w
P
w
P
P T
w
f
w
x P T
w
f
w w
x
w
a = = γ =
(1.80)
Em condições moderadas de temperatura e pressão, a fase vapor comporta-se como um gás
ideal e a razão de fugacidades pode então ser aproximada pela razão entre a pressão parcial da
água numa amostra (P
w
) e a pressão de vapor da água pura na mesma temperatura (P
w
o
). O
produto γ
w
x
w
é a atividade da água numa mistura, em que o coeficiente de atividade γ
w
pode ser
calculado pelos métodos de contribuição de grupos.
Propriedades físico-químicas estudadas

27
1.4.2 Solubilidade
O critério de cálculo da solubilidade de um componente sólido i em solução é a igualdade de
fugacidades de i nas fases líquida (L) e sólida (S) no equilíbrio. Considerando que não há
solubilidade do solvente na fase sólida, pode-se escrever:
) , , ( ) , (
i
x P T
L
i
f P T
S
i
f =
(1.81)
em que, x
i
é a fração molar do soluto, e T a temperatura do sistema. Assumindo que a
temperatura do ponto triplo é aproximadamente a temperatura de fusão do sólido (T
f
) e que ∆C
p

apresenta dependência linear com a temperatura numa faixa entre T e T
f
, a equação resultante
para predizer a concentração de saturação do sólido i num líquido pode ser expressa como:
|
|
¹
|

\
| ∆

|
|
¹
|

\
|


+
|
|
¹
|

\
|


− = γ
T
T
ln
R
C
T
T
R
C
T
T
RT
H
i
x
i
ln
f p f p f
f
f
1 1 (1.82)
A consideração que ∆C
p
é uma função linear da temperatura já foi utilizada em trabalhos prévios
(Catté et al., 1994; Peres and Macedo, 1996, 1997) para sistemas contendo açúcares. A
expressão representando a diferença entre os calores específicos do líquido e do sólido (∆C
p
) é
dada por:
) (
ref
T T B A
p
C − ∆ + ∆ = ∆
(1.83)
em que T
ref
é uma temperatura de referência, ∆A e ∆B são dois parâmetros ajustáveis e ∆H
f
a
entalpia de fusão do sólido. Considerando a dependência linear de ∆C
p
com a temperatura pode-
se calcular a atividade do soluto como:
)
f
T T (
R
B
f
T
T
ln
R
ref
BT A
f
T T
f
T
R
B
f
T
R
ref
BT A
R
f
H
i
x
i
ln


+
|
|
¹
|

\
| ∆ − ∆
+
|
|
¹
|

\
|

(
(
¸
(

¸


+
∆ − ∆
+

− = γ
2
1 1
2
2
(1.84)
Com a equação 1.84 calcula-se a solubilidade x
i
, sabendo-se as propriedades do componente i
puro e o coeficiente de atividade de i em solução. Esse cálculo requer um procedimento iterativo
Capítulo 1

28
para a estimativa simultânea da solubilidade e do coeficiente de atividade, que neste trabalho será
obtido dos métodos de contribuição de grupos.
1.4.3 Depressão do ponto de congelamento
Pode-se calcular a depressão do ponto de congelamento de uma mistura empregando a
seguinte expressão (Ferro Fontan and Chirife, 1981):
|
|
¹
|

\
|



|
|
¹
|

\
|


|
|
¹
|

\
|


= γ =
mist
w
w p
mist
w
w p
mist w
w f
w w w
T
T
R
C
T
T
R
C
T T R
H
x a 1 ln
1 1
) ln( ln
, , ,
(1.85)
em que a
w
é a atividade de água, R a constante dos gases, T
w
e T
mist
os pontos de congelamento da
água pura e da mistura com concentração conhecida, respectivamente, ∆H
f,w
é a entalpia de fusão
da água na temperatura T
w
, e ∆C
p,w
é a diferença entre os calores específicos da água líquida e do
gelo à temperatura T
mist
, assumindo que ∆C
p,w
é independente da temperatura entre T
w
e T
mist
.
Dessa forma, sabendo-se a concentração da mistura, o coeficiente de atividade da água
(calculado por método de contribuição de grupos) e as propriedades da água pura, pode-se
estimar a temperatura de congelamento da mistura (T
mist
).
Um parâmetro geralmente estudado em equilíbrio sólido-líquido de soluções é o ponto
eutético. Nesse ponto, existem em equilíbrio uma fase líquida com o(s) componente(s) sólido(s),
ou seja, no ponto eutético, uma solução de composição x
e
formada por exemplo por dois
componentes A e B, está simultâneamente em equilíbrio com o sólido puro A e o sólido puro B.
1.4.4 Elevação do ponto de ebulição
Em um sistema soluto-solvente em equilíbrio líquido-vapor, assumindo que a fase de vapor
tem o comportamento de um gás ideal, a relação básica para o cálculo da temperatura de ebulição
deste sistema é:
o
w w w
P x P γ =
(1.86)
em que P é a pressão total do sistema, γ
w
o coeficiente de atividade da água, x
w
a fração molar da
água e
o
w
P a pressão de vapor da água pura que pode ser calculada usando a equação de Antoine
com as constantes dadas por Gmehling (1977).
Propriedades físico-químicas estudadas

29
A temperatura de ebulição (T
eb
) pode então ser calculada iterativamente empregando a
equação 1.86 e conhecendo-se a composição da mistura (x
w
), uma estimativa inicial para a T
eb
e o
coeficiente de atividade da água que pode ser calculado por método de contribuição de grupos.
Dessa forma, um novo valor para a T
eb
pode ser obtido e comparado ao valor inicial estimado.
1.4.5 pH
O pH de uma solução contendo eletrólitos fracos pode ser estimado pela seguinte equação
que considera a não idealidade do sistema:
(
¸
(

¸

γ − =
+
+
H
c
H
c pH log
(1.87)
O procedimento numérico desenvolvido por Achard et al. (1994) que considera o fenômeno
de dissociação parcial das espécies em solução apresenta-se como uma metodologia eficiente
para o cálculo da concentração de íons H
+
(
+
H
c ) e do coeficiente de atividade (
c
H
+
γ ),
simultaneamente. Este procedimento é apresentado em detalhe no capítulo sobre soluções de
aminoácidos (Capítulo 3).
1.4.6 Espalhamento de luz
Em vários livros-texto pode-se encontrar a derivação da equação básica para a avaliação dos
dados de espalhamento de luz laser (Kurata, 1982). Segundo essa derivação pode-se encontrar
uma expressão que relaciona os dados medidos nos experimentos de espalhamento de luz laser
com a concentração do soluto (c
s
) e o potencial químico do solvente (µ
1
).
s
c RT
E
R
s
Kc

µ ∂ ρ
− =
1 1
(1.88)
Os termos da parte esquerda da equação 1.88 são experimentais. A constante K é dada por:
( )
4
2 2
1
2
4
λ
∂ ∂ π
=
A
s
N
c n n
K
(1.89)
Capítulo 1

30
em que n é o índice de refração do solvente puro, N
A
é o número de Avogrado e λ é o
comprimento de onda do laser. O incremento (∂n/∂c) é determinado por refratômetro
interferométrico para cada polímero e temperatura de trabalho. O excesso Rayleigh (R
E
)
considera a diferença entre a intensidade de luz espalhada pela solução polimérica e a intensidade
espalhada pelo solvente puro.
1.4.7 Entalpia de diluição
A entalpia parcial de excesso de um componente i está relacionada com o coeficiente de atividade
da seguinte maneira:
j
n , p
T
i
ln
RT
E
i
h
|
|
¹
|

\
|

γ ∂
= −
2
(1.90)
em que
E
i
h é a entalpia parcial molar do componente i menos a entalpia molar do componente i
puro a mesma temperatura e pressão.

1.5 Nomenclatura
Símbolos latinos
A Área de van der Waals
A Coeficiente virial
a Atividade
a Parâmetro de interação
B Coeficiente virial
b Parâmetros de interação
c concentração molar (g mol
-1
)
Cp Capacidade calorífica
Nomenclatura

31
f Fugacidade
G Energia de Gibbs
H Entalpia
h Entalpia
K Constante ótica
M Massa molar
m Parâmetros de interação independente da temperatura
n Índice de refração
N
A
Número de Avogrado
N Número de componentes
N Número de viscosidade
n Número de moles
n Parâmetros de interação independente da temperatura
P Pressão
q Parâmetro de área de componente
Q Parâmetro de área de grupo
r Parâmetro de volume de componente
R Constante universal dos gases (R=8.31441 J (mol K)
-1
)
R Parâmetro de volume
R Relação de Rayleigh
S Entropia
Capítulo 1

32
T Temperatura
U Energia de interação
U Energia interna
v Número de grupos
V Volume
V Volume de van der Waals
X Fração de grupo
x Fração molar
z Número de coordenação
Símbolos gregos
β Parâmetro de interação entre grupos no modelo VERS
γ Coeficiente de atividade
∆ Medida de variação
θ Ângulo
Θ Fração de área
λ Comprimento de onda de luz laser
µ Potencial químico
ν Viscosidade
Ξ Viscosidade residual
Ψ Parâmetro de interação binária
Г Coeficiente de atividade de grupo
Nomenclatura

33
Subscrita
0 Referência
1 Componente 1
11 Coeficiente virial binário entre duas moléculas do componente 1
f Fusão
H
+
Íon hidrogênio
i Componente i
ij Coeficiente virial binário dos componentes i e j
ijk Coeficiente virial ternário dos componentes i, j e k
j Componente j
k Componente k no modelo VERS
k Grupo k
l Grupo l no modelo VERS
m Grupo m
mist mistura
n Grupo n
s Soluto
w Água
Sobrescrita
(0), (1), (2) Índices dos parâmetros do modelo VERS
o Componente puro
C Termo combinatorial
Capítulo 1

34
E Medida de excesso
FH Termo de Flory-Huggins
id ideal
R Termo residual
visc Viscosidade
Siglas
MD Maltodextrina
MM Massa molar
PEG Polietileno glicol
1.6 Referêcias bibliográficas
Abrams, D.S.; Prausnitz, J.M. Statistical thermodynamics of liquid mixtures: a new
expression for the excess Gibbs energy of partly or completely miscible systems. AIChE J. 1975,
21, 116 – 128.
Achard, C.; Dussap, C. G. Prediction of pH in complex aqueous mixtures using a group
contribution method. AIChE J. 1994, 40, 1210 – 1222.
Alves, J. G. L. F.; Chumpitaz, L. D. A.; da Silva, L. H. M.; Franco, T. T.; Meirelles, A. J. A.
Partitioning of whey proteins, bovine serum albumin, and porcine insulin in aqueous two-phase
systems. J. Chrom. B 2000, 743, 235 – 239.
Azuma, M., Ikeuchi, t., Kiritami, R., Kato, J., Ooshima, H. Increase in xylitol production by
Candida tropicalis upon addition of salt. Biomass & Bioenergy 2000, 19, 129 – 135.
Billaux, M.S.; Flourie,B.; Jacquemin C.; Messing, B. Sugar alcohols. Em Handbook of
sweeteners. Edit. Marie, S.; Piggott, J.R., Glasgow:Blackie, N.Y. USA, 1991.
Bondi, A. Physical Properties of Molecular Crystals, Liquids and Glasses. Wiley, New
York, 1968.
Referências bibliográficas

35
Cao, W.; Knudsen, K.; Fredenslund, A.; Rasmussen, P. Group-contribution viscosity
predictions of liquid mixtures using UNIFAC-VLE parameters. Ind. Eng. Chem. Res. 1993, 32,
2088 – 2092.
Catté, M.; Dussap, C-G; Achard, C.; Gros, J-B. Excess properties and solid-liquid equilibria
for aqueous solutions of sugars using UNIQUAC model. Fluid Phase Equilibria 1994, 96, 33 –
50.
Chirife, J.; Ferro-Fontan, C.F. A Study of the water activity lowering behavior of
polyethylene glycols in the intermediate moisture range. J. Food Sci. 1980, 45, 1717 – 1719.
Chumpitaz, L. D. A.; Coutinho, L. F.; Meirelles, A. J. A. Surface tension of fatty acids and
triglycerides. J. Am. Oil Chem. Society 1999, 76 (3), 379 – 382.
Coimbra, J. S. R.; Thommes, J.; Meirelles, A. J. A.; Kula, M.-R. Performance of a Graesser
contactor in the continuous extraction of whey proteins: mixing, mass transfer and efficiency.
Bioseparation 1995, 5, 259 – 268.
Cruz, M. S.; Chumpitaz, L. D. A.; Alves, J. G. L. F.; Meirelles, A. J. A. Kinematic viscosities
of poly(ethylene glycols). J. Chem. Eng. Data 2000, 45, 61 – 63.
Cussler, E. L., Belter, P. A., Hu, W. S. Crystalization for bioseparations. Chemtech. 1989, 19,
376 – 384.
Davidson, R. L. Handbook of water-soluble gums and resins. McGraw-Hill Book Company,
New York, 1980.
Eyal, A.M.; Bressler, E. Mini review industrial separation of carboxylic and amino acids by
liquid membranes: applicability, process considerations and potential advantages. Biotech.
Bioeng. 1993, 41, 287 – 291.
Feakins, D.; Lawrence, K. G. Relative viscosities of solutions of sodium and potassium
chlorides and bromides in n-methylformamide at 25, 35 and 45 degrees. J. Chem. Soc. A 1966, 2,
212 – 216.
Ferro Fontan, C.; Chirife, J. The evaluation of water activity in aqueous solutions from
freezing point depression. J. Food Technology 1981, 16, 21 – 30.
Capítulo 1

36
Flory, P. J. Thermodynamics of high polymer solutions. J. Chem. Phys. 1941, 9, 660 – 661.
Fredenslund A.; Jones R.L.; Prausnitz, J.M. Group contribution estimation of activity
coefficients in nonideal liquid mixtures. AIChE J. 1975, 21, 1086 – 1099.
Glicksman, M. Food Hydrocolloids v III, CRC Press Inc., New York, 1986.
Gmehling, J.; Onken, U. Vapor-Liquid Equilibirum Data Collection. Frankfurt: Dechema,
1981.
Groβmann, C. Untersuchungen zur Verteilung von Aminosäuren und Peptiden auf wässrige
Zwei-Phasen-Systeme. Dissertação de Doutorado da Universidade de Kaiserslautern,
Alemanha, 1994.
Harned H. A.; Owen, B. B. The Physical Chemistry of Electrolytic Solutions. Reinhold
Publications Co., 3
rd
. Ed., New York, 1965.
Kasapis, S., Morris, E. R., Norton, I. T., Clark, A. H. Phase equilibria and gelation in
gelatin/maltodextrin systems – Part I: Gelation of individual components. Carbohydrate
Polymers 1993, 21, 243 – 248.
Kirk, R.E.; OThmer, D.F. Encyclopedia of Chemical Technology; John Wiley & Sons: New
York, 1983.
Kumar A. A simple correlation for estimating viscosities of solutions of salts in aqueous,
non-aqueous and mixed solvents applicable to high concentration, temperature and pressure. Can.
J. Chem. Eng. 1993, 71, 948 – 954.
Kurata, M. Thermodynamics of Polymer Solutions; Harwood Academic Publishers: New
York, 1982.
Marchal, L. M.; Beeftink, H. H.; Tramper, J. Towards a rational design of commercial
maltodextrins. Trends in Food Science & Technology 1999, 10, 345 – 355.
Moreton, R. C., Armstrong, N. A. The effect of film composition on the diffusion of ethanol
through soft gelatin films. Int. J. Pharm 1998, 161, 123 – 131.
Referências bibliográficas

37
Mothé, C. G., Rao, M. A. Rheological behavior of aqueous dispersions of cashew gum and
gum arabic: effect of concentration and blending. Food Hydrocolloids 1999, 13, 501 – 506.
Pereira, G.; Moreira, R.; Vázquez, M. J.; Chenlo, F. Kinematic viscosity prediction for
aqueous solutions with various solutes. Chem. Eng. J. 2001, 81, 35 – 40.
Peres, A. M.; Macedo, E.A. A modified unifac model for the calculation of thermodynamic
properties of aqueous and non-aqueous solutions containing sugars. Fluid Phase Equilibria 1997,
139, 47 – 74.
Peres, A.M.; Macedo, E.A. Thermodynamic properties of sugars in aqueous solutions:
correlation and prediction using a modified uniquac model. Fluid Phase Equilibria 1996, 123, 71
– 95.
Pitzer, K. S. Activity Coefficients in Electrolyte Solutions. CRC Press, Boca Raton, 1991.
Porter, A. W. On the vapour pressures of mixtures. Trans. Faraday Soc. 1920, 14, 335 – 345.
Prausnitz, J. M. Molecular thermodynamics of fluid phase equilibria. Englewood Cliffs, N.
J.: Prentice-Hall, 1969.
Prausnitz, J.M.; Lichtenthaler, R.N.; Azevedo, E.G. Molecular thermodynamics of fluid
phase equilibria. Prentice Hall Inc., New Jersey, 1986.
Redlich, O.; Kister, C. E. Algebraic representation of thermodynamics properties and the
classification of solutions. Ind. Eng. Chem. 1948, 40, 345 – 348.
Sé, R. A. G.; Aznar, M. Liquid-liquid equilibrium of the aqueous two-phase system water +
peg 4000 + potassium phosphate at four temperatures: experimental determination and
thermodynamic modeling. J. Chem. Eng. Data 2002, 47, 1401 – 1405.
Silva, L. H. M., Meirelles, A. J. A. Phase equilibrium and protein partitioning in aqueous
mixtures of maltodextrin with polypropylene glycol. Carbohydrate Polymers 2001, 46, 267 –
274.
Silva, L. H. M., Meirelles, A. J. A. Phase equilibrium in polyethylene glycol/maltodextrin
aqueous two-phase systems. Carbohydrate Polymers 2000, 42, 273 – 278.
Capítulo 1

38
Silva, S. S., Felipe, M. G. A. , Mancilha, I. M. Factors that affect the biosynthesis of xylitol
by xylose-fermenting yeasts. Appl. Biochem. Biotech. 1998, 70 – 72, 331 – 339.
Swift, G. Requirements for biodegradable water-soluble polymers. Polymer Degradation and
Stability, 1998, 59, 19 – 24.
Valeri, D.; Meirelles, A. J. A. Viscosities of fatty acids, triglycerides and their binary
mixtures. J. Am. Oil Chem. Society 1997, 74, 1221 – 1226.
Wilson, G. M.; Deal, C. H. Activity coefficients and molecular structure. Ind. Eng. Chem.
Fundam. 1962, 20 – 24.
Wilson, G.M. Vapor-liquid equilibrium XI. A new expression for the excess free energy of
mixing. J. Am. Chem. Soc. 1964, 86, 127 – 134.
Capítulo 2 39

Capítulo 2

WATER ACTIVITY IN POLYOL SYSTEMS
L. Ninni, M.S. Camargo, Antonio J. A. Meirelles
Department of Food Engineering (DEA/FEA), State University of Campinas – UNICAMP,
Cidade Universitária “Zeferino Vaz”, P.O. Box 6121 Zip Code 13083-970 Campinas – SP,
Brazil










Trabalho publicado na revista Journal of Chemical Engineering and Data, v. 45, p. 654 – 660,
2000.

Capítulo 2 40
Introduction 41
2. Water activity in polyol systems
2.1 Abstract
Water activities of binary and ternary mixtures containing polyols were measured using an
electronic hygrometer with temperature ranging from (10 to 35) °C. The concentrations of the
mixtures varied according to the solubility limit for each polyol (D-sorbitol, D-mannitol, xylitol,
meso-erythritol and glycerol). Results were compared with the group contribution-based models
ASOG and UNIFAC. The predictions using parameters from the literature were poor, probably
as a consequence of the strongly polar hydroxyl groups bounded to consecutive carbon atoms in
the polyol molecule. Better agreement was obtained by readjusting some of the interaction
parameters. The data bank used in this procedure included water activity data as well as polyol
solubility data taken from the literature. The best results were achieved using the UNIFAC-
Larsen model with an average relative deviation of 0.9% for water activity and solubility data.
2.2 Introduction
Polyols are polyhydroxy alcohols or sugar alcohols used in many diverse fields including
foods, pharmaceuticals, and cosmetics. The knowledge of the phase equilibria in systems
containing biological products, such as polyols, is an important support for modeling and
designing industrial processes like concentration and purification in a separation unit.
Thermodynamic models have been used for calculating some physicochemical properties – such
as water activity- in food systems containing amino acids, sugars, organic salts and other solutes
(Velezmoro and Meirelles, 1998; Ninni et al., 1999a, 1999b; Velezmoro et al., 2000). Water
activity is an important physical chemical property in food engineering, since many chemical,
enzymatic, and microbiological reactions that influence food stability are dependent on the
availability of water.
Group contribution-based models such as UNIFAC (Fredenslund et al., 1975) and ASOG
(Kogima and Tochigi, 1979) are considered to provide good results for estimating activity
coefficients in the liquid phase. Correa et al. (1994) studied the behavior of some aqueous polyol
solutions: water activities (a
w
) of binary (polyol + water) and ternary (urea + polyol + water and
urea + sugar + water) solutions were measured and results compared with those of the ASOG
group contribution model. New specific groups were defined: glucose and fructose rings, urea,
Capítulo 2 42
polyalcohol, and cyclic polyalcohol. The results showed that readjustment of binary interaction
parameters provided better agreement between the model and experimental a
w
values when
compared with the predictions using the original parameters from Kojima and Tochigi (1979).
Recently, Peres and Macedo (1997) have shown that the UNIFAC-Larsen model can be
successfully used for calculating thermodynamic properties of aqueous and nonaqueous solutions
containing sugars.
In this work, water activity has been measured for the binary (polyol + water) and ternary
(polyol + polyol + water) systems using an electronic hygrometer. Furthermore, new parameters
for UNIFAC-Larsen and ASOG models were determined, allowing the calculation of water
activity, solubility, and freezing point data of aqueous polyol solutions with low deviations
between experimental and calculated values.
2.3 Experimental Section
Water activities of binary and ternary aqueous polyol systems were determined from (10 to
35) ºC. The following polyols used were: D-sorbitol, D-mannitol, xylitol, meso-erythritol and
glycerol. They were analytical grade reagents from SIGMA with purity > 99%. The solutions
were prepared by mass percent with distilled water using an analytical balance (Sartorius,
Goettingen, Germany) with ±0.1 mg accuracy. The compositions were accurate to ±0.01%
approximately in mass fraction. Before this solution were prepared, the water content in the solid
polyols was determined by Karl Fischer titration (Metrohm, Herisau, Switzerland). The amount
of water varied from 0.08 to 0.64 mass %, and it was considered for calculating the water
concentration in solutions.
An electronic hygrometer AQUA-LAB CX-2 (DECAGON, Device Inc., Pullmann, USA)
previously calibrated with saturated salt solutions was used for measuring the a
w
. The
temperature inside the hygrometer was regulated at the desired ±0.1 °C by circulation of
thermostated water from water bath (Cole Parmer Instruments Co., Chicago, USA).
Measurements were made in triplicate with a reproducibility of ±0.001 a
w
units.

Results and Discussion 43
2.4 Results and Discussion
2.4.1 Water activity
The experimental water activity data obtained in this work for the binary and ternary
mixtures are given in Tables 2.1 and 2.2, respectively. The concentrations are in mass fraction.
For the same mass concentration, it was observed that the polyols with low molecular weight are
better water activity depressors than those with high molecular weight. This difference is well
visualized at high solute concentrations. The change of a
w
with temperature is small. As can be
seen in Figure 2.1, the experimental data for the system containing glycerol are in very good
agreement with the available literature data (Scatchard et al., 1938) which were measured using
the isopiestic method. The mean relative deviation between these two data sets is 0.1%. Such
result confirms the accuracy of experimental a
w
data measured using the AQUA-LAB CX-2
(DECAGON Devices Inc., Pullman, USA). Similar results were also reported by Velezmoro et
al. (2000) and Ninni et al. (1999a) for sugar solutions and poly(ethylene glycol) solutions,
respectively. The work of Roa and Daza (1991) also emphasizes a good performance of a prior
version of this electronic hygrometer (AQUA-LAB CX-1) for measuring water activity for
various kinds of food systems.
Table 2.1. Water activity in binary polyol solutions as a function of mass fraction of polyol (w
2
)
water (1) + sorbitol (2) water (1) + mannitol (2)
10.0°C 25.0°C 35.0°C 25.0°C 35.0°C
w
2
a
w
w
2
a
w
w
2
a
w
w
2
a
w
w
2
a
w

0.0519 0.996 0.0497 0.997 0.0495 0.996 0.0510 0.996 0.0501 0.995
0.0992 0.992 0.0993 0.992 0.0995 0.991 0.1036 0.991 0.0751 0.992
0.1986 0.979 0.1491 0.986 0.1985 0.978 0.1263 0.987 0.1035 0.989
0.2965 0.960 0.1978 0.977 0.2982 0.961 0.1501 0.985 0.1246 0.987
0.3966 0.934 0.2485 0.970 0.3953 0.937 0.1750 0.982 0.1467 0.983
0.4918 0.899 0.2980 0.960 0.4944 0.901 0.1745 0.980
0.5952 0.844 0.3478 0.949 0.5706 0.862
0.3974 0.935 0.6946 0.755
0.4468 0.919
0.4966 0.897
0.5466 0.872
0.5955 0.843
0.6457 0.803

Capítulo 2 44
Table 2.1 (cont.)
water (1) + xylitol (2) water (1) + erythritol (2)
10.0°C 25.0°C 35.0°C 25.0°C 30.0°C
w
2
a
w
w
2
a
w
w
2
a
w
w
2
a
w
w
2
a
w
0.0502 0.996 0.0501 0.993 0.0502 0.994 0.0500 0.992 0.0486 0.993
0.1020 0.989 0.0997 0.988 0.1011 0.987 0.1095 0.982 0.0966 0.985
0.2005 0.974 0.1499 0.981 0.1510 0.980 0.1547 0.973 0.1501 0.975
0.3009 0.951 0.1997 0.972 0.2033 0.971 0.2002 0.963 0.2022 0.964
0.3906 0.925 0.2496 0.963 0.2504 0.962 0.2524 0.950 0.2334 0.956
0.5021 0.876 0.2983 0.951 0.2996 0.951 0.2996 0.936 0.2914 0.940
0.5423 0.854 0.3496 0.937 0.3479 0.938 0.3470 0.922 0.3345 0.926
0.3995 0.921 0.3990 0.922 0.3690 0.912
0.4492 0.901 0.4489 0.903
0.4995 0.878 0.4969 0.880
0.5489 0.849 0.5533 0.850
0.5991 0.816 0.5670 0.840
0.6490 0.774 0.5993 0.818
water (1) + glycerol (2)
25.0°C 35.0°C
w
2
a
w
w
2
a
w

0.0506 0.991 0.0498 0.990
0.0999 0.980 0.1001 0.979
0.1498 0.967 0.1499 0.967
0.1999 0.952 0.1996 0.953
0.2502 0.936 0.2496 0.937
0.3002 0.918 0.2995 0.919
0.3495 0.896 0.3494 0.897
0.3994 0.872 0.3995 0.874
0.4497 0.844 0.4491 0.847
0.4992 0.812 0.4986 0.816
0.5491 0.775 0.5489 0.780
0.5993 0.733 0.5970 0.740
0.6495 0.683 0.6487 0.689
0.6989 0.628 0.6987 0.632
0.7489 0.557
0.7986 0.483
0.8487 0.399

Results and Discussion 45
Table 2.2. Water activity in ternary polyol solutions at 25.0ºC
water(1)+xylitol(2)+sorbitol(3) water(1)+glycerol(2)+mannitol(3) water(1)+glycerol(2)+sorbitol(3)
w
2
w
3
a
w
w
2
w
3
a
w
w
2
w
3
a
w

0.0264 0.0249 0.993 0.0254 0.0250 0.991 0.0254 0.0254 0.993
0.0494 0.0491 0.988 0.0376 0.0383 0.988 0.0511 0.0501 0.984
0.0765 0.0756 0.981 0.0509 0.0511 0.982 0.1002 0.1152 0.961
0.1005 0.0994 0.973 0.0610 0.0605 0.979 0.1491 0.1490 0.938
0.1259 0.1245 0.965 0.0770 0.0761 0.974 0.2105 0.1958 0.897
0.1493 0.1485 0.954 0.0872 0.0866 0.970 0.2626 0.2389 0.848
0.1764 0.1779 0.940 0.3039 0.2960 0.780
0.1911 0.1904 0.932 0.3501 0.3480 0.684
0.2266 0.2229 0.908
0.2494 0.2482 0.887
0.3013 0.2974 0.827
0.3262 0.3226 0.787
water(1)+glycerol(2)+xylitol(3) water(1)+xylitol(2)+mannitol(3) water(1)+erythritol(2)+mannitol(3)
w
2
w
3
a
w
w
2
w
3
a
w
w
2
w
3
a
w

0.0255 0.0251 0.992 0.0249 0.0250 0.993 0.0249 0.0250 0.994
0.0525 0.0505 0.982 0.0499 0.0500 0.987 0.0499 0.0498 0.987
0.1029 0.1023 0.960 0.0749 0.0750 0.980 0.0750 0.0749 0.979
0.1498 0.1503 0.895 0.0999 0.0999 0.972 0.0999 0.1004 0.971
0.2040 0.1977 0.871 0.1248 0.1249 0.964 0.1246 0.1245 0.959
0.2988 0.2998 0.772 0.1398 0.1399 0.957 0.1397 0.1398 0.953
0.3246 0.3249 0.726

0.0 0.2 0.4 0.6 0.8 1.0
0.4
0.5
0.6
0.7
0.8
0.9
1.0
exp. (Scatchard et al, 1938)
exp. (this work)
UNIFAC-Larsen with parameters from Larsen et al. (1987)
UNIFAC-Larsen (this work)
ASOG with parameters from Correa et al. (1994)
ASOG (this work)
a
w
w

Figure 2.1. Prediction of water activities in the glycerol – water system.
Capítulo 2 46
The prediction of water activity using different versions of the UNIFAC model (Fredenslund
et al., 1975; Larsen et al., 1987) and the ASOG model (Kojima and Tochigi, 1979) was carried
out as a first estimate. The group assignments utilized for the polyols in the various attempts are
summarized in Table 2.3. It was observed that the models were not sufficiently accurate to
predict values of a
w
at high solute concentrations. This poor estimation is attributed to the effect
of strongly polar hydroxyl groups bounded to consecutive carbon atoms in the molecule. This
suggests that there is an intramolecular proximity effect between these constituent groups, as
proposed by Wu and Sandler (1991a, 1991b) and Abildskow et al. (1996). Another reason, also
relative to chemical structure of the polyols, could be referred to the models being unable to
distinguish between the molecular structure of isomers (mannitol and sorbitol).
Table 2.3. Group assignment for polyols
UNIFAC
a
UNIFAC
b
(v
k
i
) ASOG
c

v
k
i
v
k
i
v
k
i
v
i
FH

glycerol 1CHOH, 2CH
2
OH 2CH
2
, 1CH, 3OH 2.8CH
2
, 3POH 6
meso-erythritol 2CHOH, 2CH
2
OH 2CH
2
, 2CH, 4OH 3.6CH
2
, 4POH 8
xylitol 3CHOH, 2CH
2
OH 2CH
2
, 3CH, 5OH 4.4CH
2
, 5POH 10
D-mannitol 4CHOH, 2CH
2
OH 2CH
2
, 4CH, 6OH 5.2CH
2
, 6POH 12
D-sorbitol 4CHOH, 2CH
2
OH 2CH
2
, 4CH, 6OH 5.2CH
2
, 6POH 12
a
CH
2
OH, CHOH are groups proposed by Wu and Sandler (1991a, b).
b
CH
2
, CH, OH are groups proposed by
Skjold-Jorgensen et al. (1979).
c
POH is a group proposed by Correa et al. (1994). v
k
i
is the number of groups k in
molecule i. v
i
FH
is the number of atoms (other than hydrogen atoms) in molecule i.
For the ASOG model with parameters from Correa et al. (1994), the predictions were
similarly poor. This could be a consequence of using a restricted range of solute concentrations
for the adjustment of the parameters. This was well observed in the system containing glycerol,
which presented a mean deviation of 0.8% in a restricted range of water activity (0.998 to 0.875)
but a high deviation (15,6%) for the whole concentration range studied in this work (see Figure
2.1).
An average relative deviation, including all experimental a
w
values from this work and from
the literature (Scatchard et al., 1938; Robinson and Stockes, 1961; Bower and Robinson, 1963),
was found as 2.1 % for the ASOG model (with parameters from Correa et al. (1994)); 1.2% for
the original UNIFAC model (Skjold-Jorgensen et al., 1979) and 1.8% when the UNIFAC-Larsen
model (Larsen et al., 1987) was used in the predictions.
On the basis of these results and considering the proximity effect of the hydroxyl groups,
some of the UNIFAC and ASOG interaction parameters were readjusted.
Results and Discussion 47
2.4.2 Readjustment of group interaction parameters
To readjust some of the interaction parameters of the UNIFAC and ASOG models, different
strategies were used. In the UNIFAC model, we have used the original equation and the
UNIFAC-Larsen version. For each version we have assumed two alternatives for the division of
groups: the first one proposed by Wu and Sandler (1991a,b) and the second one suggested by
Skjold-Jorgensen et al. (1979), (see Table 2.3). For the ASOG model we have used the group
assignment proposed by Correa et al. (1994). The binary interaction parameters readjusted in this
work are given in Tables 2.4 and 2.5 for the UNIFAC-Larsen and ASOG models, respectively.
Table 2.4. UNIFAC-Larsen interaction parameters
CH
2
CH OH H
2
O
CH
2
0.0 972.8
b
1857.0
b

CH 0.0 972.8
b
1857.0
b

OH 637.5
b
637.5
b
278.7
a

H
2
O 410.7
b
410.7
b
-175.9
a

a
Parameters readjusted in this work.
b
Parameters obtained from Larsen et al. (1987).

Table 2.5. ASOG interaction parameters
CH
2
POH H
2
O
CH
2
2434.7
a

-9.0831
a

-277.3
b

-0.2727
b

n
m
POH -2.3184
a

-3.2184
a

-42.76
a

-0.2868
a

n
m
H
2
O -2382.7
b

0.5045
b

257.5
a

-0.6705
a



n
m
a
Parameters readjusted in this work.
b
Parameters obtained from Kogima and Tochigi (1979).

The other group interaction parameters were set equal to the values available in the literature
(Kogima and Tochigi, 1979; Skjold-Jorgensen, 1979; Larsen et al., 1987), and they are also given
in Tables 2.4 and 2.5 for the UNIFAC-Larsen and ASOG models. To readjust some of the
parameters, the Marquardt method (Marquardt, 1963) was used for minimizing the following
objective function (OF)
∑ ∑

+

k exp
exp cal
n w
w w
sol
sol sol
a
a a
= OF
exp
exp calc
(2.1)

Capítulo 2 48
where n and k are the number of experimental water activity and solubility data, respectively;
sol is the solubility; and the subscripts calc and exp mean calculated and experimental values.
The systems used for readjusting the interaction parameters are given in Table 2.6.

Table 2.6. Systems used for readjusting the group interaction parameters
aqueous
system
temp
range/
o
C
reference data
Water Activity
D-sorbitol 10-35 this work
D-sorbitol 25 Bower and Robinson, 1963
D-mannitol 25-35 this work
D-mannitol 25 Robinson and Stokes, 1961
xylitol 10-35 this work
meso-erythritol 25-30 this work
Solubility
D-sorbitol 0-50 Mullin, 1993; Billaux et al., 1991
D-mannitol 0-100 Mullin, 1993; Billaux et al., 1991
xylitol 20-50 Billaux et al., 1991
erythritol 20-25 Röper et al., 1993

The best results were obtained using the UNIFAC-Larsen model with the group assignment
proposed by Skjold-Jorgensen et al. (1979). The results calculated by both the UNIFAC-Larsen
and ASOG models are shown in Figures 2.2 – 2.4 for the binary systems of xylitol, D-sorbitol,
and D- mannitol. The UNIFAC-Larsen model was capable of providing good results even at
high solute concentrations. Moreover, in comparison with the obtained results from Correa et al.
(1994), the readjustment of the ASOG parameters between the POH group and the other mixture
constituent groups improved the calculated a
w
values for these systems. In these cases, the use of
a wide range of solute concentrations and different types of polyols makes it possible to attain
significant interaction parameters for use in calculation of the thermodynamical properties
studied.


Results and Discussion 49
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.75
0.80
0.85
0.90
0.95
1.00
experimental (this work)
UNIFAC-Larsen with parameters from Larsen et al. (1987)
UNIFAC-Larsen (this work)
ASOG with parameters from Correa et al. (1994)
ASOG (this work)
a
w
w

Figure 2.2. Experimental and calculated water activities for xylitol solutions at 25 °C.


0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90
-20
-10
0
10
20
30
40
50
60
exp. solubility (Billaux et al., 1991)
exp. freezing point (Uraji et al., 1997)
exp.solubility (Mullin, 1993)
pred. freezing point UNIFAC-Larsen
calc. solubility UNIFAC-Larsen
pred. freezing point ASOG
calc. solubility ASOG
t
e
m
p
e
r
a
t
u
r
e

/


o

C
solubility (w/w)

Figure 2.3. Solid-liquid equilibria for sorbitol aqueous solutions (solubility and freezing point
depression).
Capítulo 2 50
-10 0 10 20 30 40 50 60 70 80 90 100 110
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
experimental (Billaux et al., 1991)
experimental (Mullin, 1993)
UNIFAC-Larsen
ASOG
s
o
l
u
b
i
l
i
t
y

(
w
/
w
)
temperature /
o
C

Figure 2.4. Experimental and calculated solubilities for D-mannitol at different temperatures.

Tables 2.7 shows a comparison between the average relative deviations obtained in this work
and those from the literature for water activities. The deviations between experimental and
calculated solubilities using the sets of parameters adjusted in this work are given in Table 2.8.

Table 2.7. Mean relative deviations between experimental and calculated a
w
data points
% deviation
UNIFAC-Larsen ASOG
aqueous system
Larsen et al.
(1987)
this work Correa et al.
(1994)
this work
reference
D-sorbitol 2.51 0.51 3.04 0.64 this work
D-sorbitol 0.68 0.23 0.35 0.12 Bower and Robinson, 1963
D-mannitol 0.08 0.17 0.30 0.16 this work
D-mannitol 0.11 0.06 0.19 0.02 Robinson and Stockes, 1961
xylitol 2.01 0.26 2.11 0.76 this work
meso-erythritol 0.69 0.19 0.42 0.60 this work
avg deviation 1.01 0.24 1.07 0.38

Results and Discussion 51
Table 2.8. Mean relative deviations between experimental and calculated solubility data
% deviation
aqueous system temperature
range/ºC
UNIFAC-
Larsen
ASOG
reference
D-sorbitol 0 - 50 0.91 1.32 Mullin,1993; Billaux et al., 1991
D-mannitol 0 - 100 1.86 20.80 Mullin, 1993; Billaux et al., 1991
xylitol 20 - 50 0.51 0.80 Billaux et al., 1991
erythritol 20 - 25 3.23 4.22 Röper et al., 1993
avg 1.63 6.79

For the calculation of polyol solubility in water the same equation adopted by Peres and
Macedo (1996, 1997) was used. The thermodynamic properties needed for the determination of
solubilities are reported in Table 2.9 as well as the data used for water in the calculations of the
freezing point depression of sorbitol and xylitol solutions. The values in bold type were used in
this work. Table 2.9 also indicates published values for
p
C ∆ from the literature, and excluding
the data for mannitol and erythritol, values were measured at only one temperature. There is only
one literature source in which experimental C
p
values were measured for a wide range of
temperatures. These are for the solid and liquid phases of erythritol and mannitol (Spaght et al.,
1932). The other cited literature (Barone et al. 1990) used a group contribution approach to
estimate the C
p
values for the liquid polyol. Moreover, in some cases, the experimental values
taken from distinct literature sources are expressively different. For these reasons we opted for
estimating
p
C ∆ as a linear function of temperature. This assumption was already used in
previews works from Catté et al. (1994) and Peres and Macedo (1996, 1997) for systems
containing sugars. The expression for representing the difference between the heat capacities of
the pure liquid and those of the pure solid polyols (
p
C ∆ ) is given below:
( ) ∆ ∆ ∆ C A + B T T
p ref
= −
(2.2)
where T
ref
is a reference temperature which was set equal to 25°C and where ∆A and ∆B are
two adjustable parameters. The parameters ∆A and ∆B were adjusted for each polyol using the
experimental solubilities at various temperatures taken from the literature (Billaux et al., 1991;
Mullin, 1993; Röper, 1993). The values for ∆A and ∆B are presented in Table 2.10. It should be
observed that the ∆A values estimated in this work are relatively close to the experimental
p
C ∆
Capítulo 2 52
values at 25 °C. For instance, in the case of the UNIFAC-Larsen model (with the groups CH
2
,
CH, OH, and H
2
O), a mean relative deviation of about 16% was found between the calculated
values and the
p
C ∆ experimental data at 25 °C. In general, these deviations were higher for the
other models used in this work. For example, in the case of the original UNIFAC model, the
deviation between the ∆A and
p
C ∆ values amounts to approximately 35%. Furthermore, the
minimum values for the deviation concerning solubility data were obtained for the UNIFAC-
Larsen model (with the groups CH
2
, CH, OH, and H
2
O), as can be seen in Table 2.8.
Table 2.9. Thermodynamic data of polyols and water
polyol melting temp,
T
m
/K
enthalpy of fusion,

fus
H/kJ⋅mol
-1

∆C
p
(25ºC)/
J⋅K
-1
⋅mol
-1

sorbitol 366.5
a
30.2 191
a

mannitol 439.1
a

433.2
b
56.1
a

53.58
b
, 52.8
c
191
a

290.36
b
xylitol 365.7
a
367.0
d
37.4
a
38.0
c

157
a

erythritol 390.9
a
391.6
b
39.4
a
42.36
b
40.3
c

122
a

155.42
b
water 273.15
e
6.002
e
38.03
e
a
Barone et al. (1990).
b
Spaght et al. (1932), (the values for ∆C
p
are results from regression of the experimental
data).
c
Raemy and Schweizer (1983).
d
Fassman (1975).
e
Daubert and Danner (1985), (the ∆C
p
for water was
considered constant with temperature).

Table 2.10. Values of ∆A and ∆B for the calculations of ∆C
p
with linear temperature dependency
UNIFAC-Larsen ASOG
∆A ∆B ∆A ∆B polyol
J⋅K
-1
⋅mol
-1
J⋅K
-2
⋅mol
-1


J⋅K
-1
⋅mol
-1
J⋅K
-2
⋅mol
-1

D-sorbitol 214.9 -2.9712 215.7 -4.1643
D-mannitol 215.0 0.0618 225.8 -0.2073
xylitol 128.0 3.3261 124.5 2.7738
meso-erythritol 117.2 1.0924 115.8 1.9598

It must be stressed that, for the UNIFAC-Larsen model, we readjusted only two temperature-
independent interaction parameters between the groups OH/H
2
O and H
2
O/OH. Otherwise, for
the ASOG model, better results were achieved only after the readjustment of four pairs of
temperature-dependent interaction parameters.
Results and Discussion 53
2.4.3 Predictions with the new set of parameters
The experimental freezing point data available in the literature as well as the a
w
values of the
ternary systems and of the binary system containing glycerol were used only for comparison with
predicted values. For freezing point calculations, an expression proposed by Ferro Fontan and
Chirife (1981) was used. Table 2.11 indicates the average mean deviations obtained for the water
activities of the ternary mixtures and the binary systems containing glycerol. It can be noted that
the deviations for glycerol were higher than those for the ternary systems but significantly lower
in comparison to the predictions with parameters from the literature (see Figure 2.1). Correa et
al. (1994) have already commented about the difficulties in obtaining good results for the systems
with glycerol. Figure 2.5 shows calculated and experimental a
w
values for the ternary mixture
xylitol-sorbitol-water at 25°C. However, the predictions of freezing point depression did not
present the same accuracy, with mean deviations of 10% for the UNIFAC-Larsen model and 47%
for the ASOG model. The experimental data are from Uraji et al. (1997), and the results
achieved for sorbitol are also represented in Figure 2.3. The prediction of the eutectic points for
xylitol and sorbitol aqueous solutions was also performed. These results are given in Table 2.12
for the ASOG and UNIFAC-Larsen models. Note that a good prediction for the eutectic point
concentration was obtained for both systems.
Table 2.11. Water activity prediction in polyol mixtures
% deviation
UNIFAC-Larsen ASOG
polyol mixture a
w
range
Larsen et al.
(1987)
this work

Correa et al.
(1994)
this work
xylitol+sorbitol 0.993 - 0.787 2.10 0.23 2.16 0.27
glycerol+mannitol 0.991 - 0.970 0.16 0.08 0.08 0.18
glycerol+sorbitol 0.993 - 0.684 3.99 0.95 2.63 2.63
glycerol+xylitol 0.992 - 0.726 4.48 1.87 3.65 3.65
xylitol+mannitol 0.993 - 0.957 0.35 0.09 0.17 0.14
erythritol+mannitol 0.994 - 0.953 0.37 0.13 0.16 0.12
glycerol 0.990 – 0.399 6.30 2.55 15.60 6.40
glycerol
a
0.998 – 0.762 2.20 0.88 2.94 2.54
avg 2.49 0.85 4.21 1.99
a
Scatchard et al. (1938).
Capítulo 2 54
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.75
0.80
0.85
0.90
0.95
1.00
experimental (this work)
UNIFAC-Larsen (with parameters from Larsen et al., 1987)
ASOG (with parameters from Correa et al., 1994)
UNIFAC-Larsen (this work)
ASOG (this work)
a
w
concentration (w/w)

Figure 2.5. Predictions of water activities for the ternary system water – xylitol – sorbitol at
25.0 °C.

Table 2.12. Experimental and calculated eutetic points of xylitol and sorbitol aqueous solutions
experimental ASOG UNIFAC-Larsen
T/°C conc/(w/w) T/°C conc/(w/w) T/°C conc/(w/w)
xylitol -12.2 0.43 -5.3 0.43 -10.8 0.44
sorbitol -15.5 0.54 -8.5 0.53 -13.9 0.53

2.5 Literature Cited
Abildskov, J.; Constantinou, L.; Gani, R. Towards the development of a second-order
approximation in activity coefficient models based on group contributions. Fluid Phase Equilib.
1996, 118, 1 – 12.
Barone, G.; Della Gatta, G.; Ferro, D.; Piacente, V. Enthalpies and entropies of sublimation,
vaporization and fusion of nine polyhydric alcohols. J. Chem. Soc. Faraday Trans. 1990, 86, 75
– 79.
Billaux, M.S.; Flourie, B., Jacquimin, C. Sugar alcohols. In: Handbook of sweeteners; Marie,
S., Piggott, J.R., Eds.; Glasgow: Blackie: New York, 1991.
Literature Cited 55
Bower, V.E.; Robinson, R.A. Isopiestic vapour pressure measurements of the ternary system:
sorbitol-sodium chloride-water at 25°C. J. Phys. Chem. 1963, 67, 1540 – 1541.
Catté, M.; Dussap, C-G; Achard, C.; Gros, J-B. Excess properties and solid-liquid equilibria
for aqueous solutions of sugars using UNIQUAC model. Fluid Phase Equilib. 1994, 96, 33 – 50.
Correa, A.; Comesaña, J.F.; Sereno, A.M. Measurement of water activity in water-urea-sugar
and water-urea-polyol systems, and its prediction by the ASOG group contribution method.
Fluid Phase Equilib. 1994, 98, 189 – 199.
Daubert, T.E.; Danner, R.P. Data Compilation Tables of Properties of Pure Compounds;
AIChE/DIPPR: New York, 1985.
Fassman, G.D. Handbook of biochemistry and molecular biology, CRC Press: Cleveland,
OH, 1975.
Ferro Fontan, C.; Chirife, J. The evaluation of water activity in aqueous solutions from
freezing point depression. J. Food Technol. 1981, 16, 21 – 30.
Fredenslund A.; Jones R.L.; Prausnitz, J.M. Group contribution estimation of activity
coefficients in nonideal liquid mixtures. AIChE J. 1975, 21, 1086 – 1099.
Kojima, K.; Tochigi, K. Prediction of Vapour-Liquid Equilibria by the ASOG Method,
Elsevier: Tokyo, 1979.
Larsen, B.L.; Rasmussen, P.; Fredenslund, A. A modified UNIFAC group contribution
model for prediction of phase equilibria and heats of mixing. Ind. Eng. Chem. Res. 1987, 26,
2274 – 2286.
Marquardt, D.W. An algorithm for least-squares estimation of nonlinear parameters. J. Soc.
Indust. Appl. Math. 1963, 11, 431 – 441.
Mullin J.W. Crystallization, Butherworth – Heinemann: New York, 1993.
Ninni, L.; Camargo, M.S.; Meirelles, A.J.A. Water activity in poly(ethylene glycol) aqueous
solutions. Thermochim. Acta 1999a, 328, 169 – 176.

Capítulo 2 56
Ninni, L.; Camargo, M.S.; Meirelles, A.J.A. Modeling and prediction of pH and water
activity in aqueous amino acid solutions. Computers and Chemical Engineering 1999b, 23
(supplement) S383 – S386.
Peres, A.M.; Macedo, E.A. Thermodynamic properties of sugars in aqueous solutions:
correlation and prediction using a modified UNIQUAC model. Fluid Phase Equilib. 1996, 123,
71 – 95.
Peres, A.M.; Macedo, E.A. A modified UNIFAC model for the calculation of
thermodynamic properties of aqueous and non-aqueous solutions containing sugars. Fluid Phase
Equilib. 1997, 139, 47 – 74.
Raemy, A.; Schweizer, T.F. Thermal behaviour of carbohydrates studied by heat flow
calorimeter. J. Therm. Anal. 1983, 28, 95 – 108.
Roa, V.; Daza, M.S.T. Evaluation of water activity measurements with a dew point electronic
humidity meter. Lebensm.-Wiss. Technol. 1991, 24, 208 –213.
Robinson, R.A.; Stockes, R.H. Activity coefficients in aqueous solutions of sucrose, mannitol
and their mixtures at 25°C. J. Phys. Chem. 1961, 65, 1954 – 1958.
Röper, H.; Goossens, J. Erythritol, a new material for food and non-food applications. Starch
1993, 11, 400 – 405.
Scatchard, G.; Hamer, W.J.; Wood, S.E. Isotonic solutions. 1. The chemical potencial of
water in aqueous solutions of sodium chloride, potassium chloride, sulphuric acid, sucrose, urea
and glycerol at 25°C. J. Am. Chem. Soc. 1938,60, 3061 – 3070.
Skjold-Jorgensen, S.; Kolbe, B; Gmehling, J.; Rasmussen, P. Vapor-liquid equilibria by
UNIFAC group contribution. Revision and extension. Ind. Eng. Chem. Process Des. Dev. 1979,
18, 714 – 722.
Spaght, M.E.; Thomas, S.B.; Parks, G.S. Some heat capacity data on organic compounds
obtained with a radiation calorimeter. J. Phys. Chem. 1932, 36, 882 – 888.

Literature Cited 57
Uraji, T.; Kohno, H.; Nakashima, K.; Shimoyamada, M.; Watanabe, K. Thermal Analyses of
polyol-aqueous solutions below 0°C by differential scanning calorimetry. J. Japanese Soc. Food
Sci. Tech. 1997, 44, 380 – 385.
Velezmoro, C.; Meirelles, A.J.A. Water activity in solutions containing organic acids. Drying
Technol. 1998, 16, 1789 –1805.
Velezmoro, C.; Meirelles, A.J.A.; Vitali, A. Engineering and Food at ICEF 7, Proceedings of
the Seventh International Congress on Engineering and Food, Part 1, Brighnton, UK, 1997, p.
A145.
Wu, H.S.; Sandler, S.I. Use of ab initio quantum mechanics calculations in group
contribution methods. 1. Theory and the basis for group identifications. Ind. Eng. Chem. Res.
1991a, 30, 881 – 889.
Wu, H.S.; Sandler, S.I. Use of ab initio quantum mechanics calculations in group
contribution methods. 2. Test of new groups in UNIFAC. Ind. Eng. Chem. Res. 1991b, 30, 881 –
897.
2.6 Acknowledgment
This work was supported financially by research grants from Fundação de Amparo à
Pesquisa do Estado de São Paulo (FAPESP - Proc. 95/02617-5) and Conselho Nacional de
Desenvolvimento Científico e Tecnológico (CNPq - Proc. 521011/95-7).
Capítulo 3 59

Capítulo 3

WATER ACTIVITY, PH AND DENSITY OF
AQUEOUS AMINO ACIDS SOLUTIONS
Luciana Ninni, Antonio J. A. Meirelles

Department of Food Engineering (DEA/FEA), Faculty of Food Engineering, State University of
Campinas (UNICAMP), Cidade Universitária "Zeferino Vaz", PO Box 6121, Zip Code 13083-
970, Campinas, SP, Brasil










Trabalho publicado na revista Biotechnology Progress, v. 17 (4), p. 703 – 711, 2001.
Capítulo 3 60



Introduction

61
3. Water Activity, pH and Density of Aqueous Amino Acids Solutions
3.1 Abstract
The water activity, pH and density of some aqueous amino acid solutions were determined at
25 °C in three different types of solvents. Previous published experimental data on water activity
and solubility of amino acids in aqueous solutions were used together with data from this work to
test the applicability of a group contribution model. The activity coefficients were estimated by
the UNIFAC-Larsen model combined with the Debye-Hückel equation, taking also into account
the partial dissociation phenomena of species in solution. Interaction energies between the
charged species Na
+
and Cl
-
and the specific groups of amino acids (COOH and NH
2
) were
adjusted using experimental solubility data.
3.2 Introduction
The crescent advances in the biotechnological industry have raised interest for the
development of efficient separation, concentration and purification processes of biomolecules. In
the design of equipments and processes it is necessary the knowledge of some physical-chemical
properties such as solubility, water activity (a
w
) and the influence of different pH values on these
properties in mixtures containing biomolecules. Among the biochemicals, amino acids are of
much research interest because of their simplicity and importance for understanding the behavior
of molecules such as peptides and proteins in more complex solutions. In general the amino acids
are produced by microorganisms in aqueous media containing solutes like salts, organic acids,
etc. Thus, it is relevant to study the physical-chemical properties of systems containing such
compounds at various concentrations and pH values.
In this work we present experimental data on water activity, pH and density of aqueous
solutions containing the amino acids glycine, dl-alanine, l-arginine and l-proline in three different
types of solvents: water, acid and basic buffers. The iterative procedure proposed by Achard et al.
(1994a) was utilized for modeling the properties (a
w
, pH, and solubility), considering the partial
dissociation of the amino acids and the other solutes that form the aqueous systems. The
UNIFAC-Larsen model and the Debye-Hückel equation were associated with this procedure to
allow for the non-ideality of the systems studied. Such approach has been successfully applied to
other aqueous solutions of weak electrolytes like citric, tartaric and malic acids for the prediction
Capítulo 3 62
of water activity and pH of those solutions (Maffia and Meirelles, 2000; Velezmoro and
Meirelles, 1998).
Moreover, the procedure mentioned above is further tested by using a comprehensive data
bank on other physical-chemical properties of amino acid solutions available in the literature.
Data on solubility of amino acids in water, buffers and salt solutions were included in this data
bank. The objective of the work reported here is to test and extend the procedure proposed by
Achard et al. (1994a,b) to the estimation of physical-chemical properties of amino acid solutions.
In contrast to prior works the present study takes into account a more comprehensive data bank
on different physical-chemical properties of amino acid solutions and model these properties
considering the partial dissociation of weak electrolytes.
3.3 Materials and Methods
3.3.1 Materials
Samples of the amino acids glycine, dl-alanine, l-arginine and l-proline were supplied by
Merck with purity greater than 99 mass %. No further purification was employed. Other
chemicals of >99 mass % purity (NaOH, citric acid and sodium citrate) were also purchased from
Merck and used to prepare the buffer systems. Solutions were prepared gravimetrically with
deionised water (Millipore, USA) using an analytical balance (Sartorius, Germany) with
precision of 1 × 10
-4
g. The compositions were accurate to ± 0.0001 approximately in mass
fraction.
3.3.2 Experimental Procedure
Water activity was measured at 25.0±0.1 °C and pH at room temperature (24.7±0.3 °C) at
various solute concentrations. The influence of pH on water activity was also verified by the use
of acidic and basic buffer solutions. The buffers were prepared with citric acid/sodium citrate and
glycine/NaOH, resulting in 3.09 and 9.22 pH values, and 0.0113 (ratio: 0.0084 citric acid/0.0029
sodium citrate) and 0.0043 (ratio: 0.0038 glycine/0.0005 NaOH) total solid mass fractions,
respectively.
The pH was measured with a Chem Cadet 5986-50 pH meter (Cole-Parmer Instrument Co.,
USA) calibrated with buffer solutions (pH=4 and 7) supplied by Merck, and the results used later
in the modeling of this property. The mean standard deviation obtained for the whole set of
experimental data was 0.02 pH units. An electric hygrometer AQUA-LAB CX-2 (Decagon,
Experimental Results

63
USA) previously calibrated with saturated salt solutions (analytical grade reagents from Merck,
>99 mass % purity: potassium sulfate, potassium chloride, sodium chloride and potassium
carbonate) was used for measuring a
w
. The temperature inside the hygrometer was regulated at
25.0 ± 0.1 °C by circulation of thermostated water from a water bath (Cole Parmer Instrument
Co., USA). Measurements were made in triplicate with a reproducibility of ± 0.001 a
w
units.
Density measurements were carried out in triplicate using a digital densimeter (DMA 58, Anton
Paar, Austria) (for the systems amino acid + buffer) and a picnometer (for the systems amino acid
+ water) at 25.0±0.1 °C. The digital densimeter was calibrated with air and water as standards
while the picnometer was calibrated with water at the corresponding working temperature. The
accuracy of the density measurements was estimated as 3 × 10
-5
and 3 × 10
-4
g cm
-3
for the digital
densimeter and picnometer, respectively. Experimental density measurements were performed in
the systems cited above for the conversion of solute mass fractions into molar concentrations.
3.3.3 Experimental Results
Experimental water activity, pH and density data for the amino acid solutions are
summarized respectively in Tables 3.1 to 3.3. Kuramochi et al. (1997b) measured the partial
pressure of water for glycine and alanine aqueous solutions. The corresponding experimental
water activity data for glycine were compared to the data determined in the present work (see
Figure 3.1). The mean relative deviations between both data sets were 0.151% for glycine and
0.051% for alanine, resulting in an average value of 0.101% for both amino acid solutions.
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22
0.95
0.96
0.97
0.98
0.99
1.00
R
2
=0.99811
this work
(Kuramochi et al., 1997b)
a
w
w
a

Figure 3.1. Water activities determined in the present work and data taken from ref 5 for the
amino acid glycine.
Capítulo 3 64
Table 3.1. Experimental water activities of glycine, dl-alanine, l-arginine and proline as a
function of amino acid mass fraction (w
a
), at 25.0 ±0.1 °C
w
a
a
w
w
a
a
w
w
a
a
w

glycine + water glycine + acid buffer glycine + basic buffer
0.0241 0.996 0.0252 0.993 0.0309 0.992
0.0505 0.990 0.0496 0.988 0.0573 0.985
0.0759 0.983 0.0714 0.982 0.0772 0.981
0.0997 0.978 0.0996 0.975 0.0997 0.975
0.1227 0.972 0.1210 0.970 0.1233 0.970
0.1500 0.966 0.1481 0.963 0.1484 0.964
0.1712 0.960 0.1713 0.957 0.1770 0.956
dl-alanine + water dl-alanine + acid buffer dl-alanine + basic buffer
0.0251 0.995 0.0247 0.994 0.0243 0.993
0.0503 0.990 0.0498 0.990 0.0530 0.988
0.0677 0.987 0.0615 0.987 0.0640 0.985
0.0766 0.985 0.0752 0.984 0.0721 0.982
0.1000 0.979 0.0889 0.981 0.0865 0.980
0.1237 0.974 0.0981 0.979 0.0945 0.978
0.1353 0.971 0.1195 0.973 0.1242 0.971
0.1315 0.971
l-proline + water l-proline + acid buffer l-proline + basic buffer
0.0504 0.992 0.0495 0.991 0.0498 0.993
0.0961 0.983 0.0918 0.983 0.0943 0.983
0.1503 0.972 0.2021 0.957 0.2003 0.958
0.2002 0.959 0.2951 0.923 0.2936 0.924
0.2501 0.944 0.4025 0.866 0.3355 0.905
0.2942 0.927 0.4900 0.799
0.3497 0.901
0.3799 0.886
l-arginine + water l-arginine + acid buffer l-arginine + basic buffer
0.0253 0.998 0.0247 0.998 0.0258 0.997
0.0376 0.997 0.0498 0.996 0.0494 0.994
0.0500 0.996 0.0616 0.995 0.0545 0.993
0.0597 0.995 0.0748 0.994 0.0744 0.991
0.0755 0.994 0.0863 0.992 0.0873 0.990
0.0876 0.992 0.0997 0.990 0.0996 0.989
0.1006 0.992 0.1110 0.991 0.1123 0.988
0.1127 0.990 0.1244 0.988 0.1233 0.986
0.1239 0.989
Experimental Results

65
Table 3.2. Experimental pH of glycine, dl-alanine, l-arginine and l-proline in water and in basic
and acid buffers (w
a
=amino acid mass fraction) 24.7 ±0.3 °C
w
a
pI w
a
acid buffer w
a
basic buffer
glycine
0.0241 5.79 0.0252 3.75 0.0309 8.31
0.0505 5.81 0.0496 3.95 0.0573 7.99
0.0759 5.86 0.0714 4.06 0.0772 7.89
0.0997 5.92 0.0996 4.19 0.0997 7.78
0.1227 5.96 0.1210 4.27 0.1233 7.69
0.1500 5.97 0.1481 4.36 0.1484 7.65
0.1712 5.99 0.1713 4.46 0.1770 7.58
dl-alanine
0.0251 5.71 0.0247 3.67 0.0243 8.58
0.0503 5.75 0.0498 3.90 0.0530 8.25
0.0677 5.78 0.0615 3.92 0.0640 8.17
0.0766 5.80 0.0752 3.99 0.0721 8.11
0.1000 5.88 0.0889 4.04 0.0865 8.04
0.1237 5.91 0.0981 4.10 0.0945 8.01
0.1353 5.95 0.1195 4.20 0.1242 7.84
0.1315 4.21
l-proline
0.0504 6.15 0.0495 3.62 0.0498 8.81
0.0961 6.21 0.0918 3.82 0.0943 8.61
0.1503 6.32 0.2021 4.22 0.2003 8.32
0.2002 6.39 0.2951 4.52 0.2936 8.21
0.2501 6.46 0.4025 4.88 0.3355 8.14
0.2942 6.52 0.4900 5.25
0.3497 6.65
0.3799 6.76
l-arginine
0.0253 11.01 0.0498 9.49 0.0258 9.96
0.0376 11.03 0.0616 9.64 0.0494 10.08
0.0500 11.09 0.0748 9.77 0.0545 10.12
0.0597 11.13 0.0863 9.84 0.0744 10.25
0.0755 11.17 0.0997 9.93 0.0873 10.38
0.0876 11.20 0.1110 10.00 0.0996 10.42
0.1006 11.23 0.1244 10.09 0.1123 10.44
0.1127 11.25 0.1233 10.52
0.1239 11.27

Capítulo 3 66
Table 3.3. Densities of aqueous solutions of glycine, dl-alanine, l-arginine and l-proline as a
function of total solute mass fraction (w
t
), 25.0 ±0.1 °C
a
w
t
ρ (g cm
-3
) w
t
ρ (g cm
-3
) w
t
ρ (g cm
-3
)
glycine + water glycine + acid buffer glycine + basic buffer
0.0241 1.0069 0.0365 1.01249 0.0352 1.01283
0.0505 1.0195 0.0609 1.02283 0.0616 1.02453
0.0759 1.0283 0.0827 1.03221 0.0815 1.03282
0.0997 1.0402 0.1109 1.04395 0.1040 1.04262
0.1227 1.0505 0.1323 1.05320 0.1276 1.05341
0.1500 1.0590 0.1594 1.06450 0.1527 1.06379
0.1712 1.0718 0.1826 1.07421 0.1813 1.07590
dl-alanine + water dl-alanine + acid buffer dl-alanine + basic buffer
0.0255 1.0053 0.0360 1.00991 0.0286 1.00578
0.0514 1.0137 0.0611 1.01783 0.0573 1.01458
0.1000 1.029 0.0728 1.02144 0.0683 1.01846
0.1157 1.0345 0.0865 1.02597 0.0764 1.02043
0.1237 1.0371 0.1002 1.03055 0.0908 1.02450
0.1352 1.0409 0.1094 1.03332 0.0988 1.02750
0.1308 1.04010 0.1285 1.03661
0.1428 1.04393
l-proline + water l-proline + acid buffer l-proline + basic buffer
0.0280 1.0048 0.0608 1.01572 0.0541 1.01281
0.0961 1.0250 0.1031 1.02796 0.0986 1.02576
0.1503 1.0398 0.2134 1.06098 0.2046 1.05680
0.2002 1.0551 0.3064 1.08851 0.2979 1.08500
0.2756 1.0766 0.4138 1.12111 0.3398 1.09632
0.2942 1.0841 0.5013 1.14700
0.3497 1.1022
0.3799 1.1097
l-arginine + water l-arginine + acid buffer l-arginine + basic buffer
0.0187 1.0022 0.0360 1.00848 0.0301 1.00626
0.0376 1.0077 0.0611 1.01641 0.0588 1.01379
0.0533 1.0118 0.0729 1.02052 0.0787 1.01896
0.0597 1.0133 0.0861 1.02416 0.1039 1.02512
0.0755 1.0178 0.0976 1.02777 0.1276 1.03288
0.0876 1.0219 0.111 1.03185
0.1008 1.0250 0.1223 1.03467
0.1127 1.0283 0.1357 1.03870
0.1239 1.0315
a
w
t
corresponds to total solute mass fraction considering the concentration of the amino acid (w
a
) and the solutes of
the buffers (w
b
), i.e., w
t
=w
a
+w
b
.
Experimental Results

67
A very slight influence of pH on a
w
values was noted for arginine, glycine, proline and
alanine. The addition of buffer solutions containing a strong electrolyte (NaOH) or an acid/salt
(citric acid/sodium citrate) solution gives rise to interactions between the amino acid groups and
the ionic species from the acid, base or salt. The buffer solutions were used in very low
concentrations (total buffer solutes weight fractions: 0.0113 and 0.0043 for acid and basic
buffers, respectively) yielding a
w
values equal to 0.999 measured by the electric hygrometer, at
25.0 °C. Therefore, there was no significant interference of the buffer solutions upon the a
w
data
measured in the present work through short-range type interactions. The presence of buffers
containing ionizable species contributes to changes in the net charges of the amino acids
according to pH value. As a consequence, it was observed that higher concentrations of the
cationic and anionic species in solution lead to a slightly decrease in the a
w
values.
Figure 3.2 shows a comparison between density data for aqueous solutions of dl-alanine and
glycine at 25.0 °C reported in literature (Soto et al., 1998; Dalton and Schmidt, 1933) and
determined in this work. As can be seen, the agreement is very satisfactory (mean relative
deviations equal to 0.07% and 0.004% for aqueous solutions of glycine and dl-alanine,
respectively).
0.00 0.05 0.10 0.15 0.20
0.98
1.00
1.02
1.04
1.06
1.08
1.10
glycine:
(Soto et al., 1998)
this work
dl-alanine:
(Dalton and Schmidt, 1933)
this work
linear regression
d
e
n
s
i
t
y

(
g

c
m
-
3
)
w
a

Figure 3.2. Comparison between experimental densities from literature and those determined in
the present work.
Capítulo 3 68
3.4 Thermodynamic Modeling
3.4.1 Modified UNIFAC Model for Electrolytes
Recently, attention has been given to the modeling of phase behavior in aqueous systems
containing biochemicals like amino acids (Khoshkbarchi and Vera, 1996, 1997; Kuramochi et al.,
1996, 1997a; Pradhan and Vera, 1998; Pinho et al., 1994). In several publications, the non-
ideality of such systems was represented either by local composition models (Nass, 1988; Chen et
al., 1989) or by group contribution models (Pinho et al., 1994; Gupta and Heidemann, 1990;
Ninni et al., 1999). To account for the electrostatic forces in the amino acid solutions, the Debye-
Hückel term (Pitzer, 1980) has been added to the activity coefficient models in some reported
works in the literature (Pinho et al., 1994; Peres and Macedo, 1994). In this work we used the
model presented by Kuramochi et al. (1996, 1997a), based on the UNIFAC-Larsen model,
combined with the Debye-Hückel term. Amino acids molecules were divided into several main
groups including new group assignments proposed by the cited authors: α-CH
2
(alkane bounded
to the α carbon atom in the amino acid molecule), sc-CH
2
(side chain alkane), NH
2
and COOH.
In Table 4, the constituent groups of the amino acids considered in this work are given.
Table 3.4. Division of groups for the amino acids used in this work
amino acid groups
a
glycine
b
α-CH
2
, NH
2
, COOH
l-arginine α-CH, 2×scCH
2
, COOH, NH
2
, C(NH
2
)NH, CH
2
NH
dl-alanine NH
2
, α-CH, COOH, scCH
3
dl-valine NH
2
, COOH, α-CH, scCH, 2×scCH
3
serine NH
2
, COOH, α-CH, scCH
2
, OH
l-proline CH
2
NH, COOH, α-CH, 2×scCH
2

a
sc=side chain, α=group bounded to the α carbon in the molecule.
b
For the zwitterion glycine the groups are α-
CH
2
,
+
3
NH , COO
-
; for the anionic: α-CH
2
, NH
2
, COO
-
; and for the cationic form: α-CH
2
,
+
3
NH , COOH.

The amino acids appear in various ionic forms - cationic, anionic and zwitterion - when
dissolved in water according to the pH of the solution. For example, the distribution of the
different ionic forms of the amino acid glycine as a function of pH was obtained by Max et al.
(1998) using a numerical treatment of the infrared spectra data. The following species were
Thermodynamic Modeling

69
observed for glycine in water: cationic (pH 0 to 5); zwitterion (pH 0 to 12.5) and anionic (pH 7 to
14).
For modeling the amino acid systems it is assumed that the constituent groups of the anionic
and cationic species are the same as those of the zwitterion form (see Table 3.4). Structural
parameters R
k
and Q
k
for the ionic groups of the amino acids were considered equal to the neutral
ones. To estimate the concentrations of the different ionic forms present in solution, the iterative
procedure proposed by Achard et al. (1994a) was utilized in the present work. The method takes
into account the partial dissociation phenomena of the amino acids and combines equilibrium
relations, mass and electroneutrality balances for calculating the true concentrations of the
species and their activity coefficients. This set of equations requires initial solute and water
concentrations, as well as equilibrium constants describing the chemical equilibrium. In aqueous
solution, the following series of reactions takes place simultaneously when an amino acid is
dissolved in an acidic buffer containing citric acid ( citric H
3
)/sodium citrate ( citric Na
3
):
− +
⇔ RCOO NH RCOOH NH
3 2
K
D
(3.1)
+ − + +
+ ⇔ H RCOO NH RCOOH NH
3 3
K
1
(3.2)
+ − − +
+ ⇔ H RCOO NH RCOO NH
2 3
K
2
(3.3)
-
2
+
3
citric H H citric H + ⇔ K
A1
(3.4)
- - + -
2
Hcitric H citric H + ⇔ K
A2
(3.5)
- - - + - -
citric H Hcitric + ⇔ K
A3
(3.6)
- - - +
3
citric Na 3 citric Na + ⇒ (3.7)
− +
+ ⇔ OH H O H
2
K
w
(3.8)
Reaction 3.1 corresponds to the formation of the zwitterion (
− +
RCOO NH
3
) with dual electric
charges, which can originate the amino acid cationic and anionic species (reactions 3.2 and 3.3,
respectively). The equilibrium constant K
D
is generally very large (10
5
- 10
6
). This means that the
uncharged amino acid is almost completely converted to the zwitterion form. The equilibrium
constants (K
i
and K
Ai
, i=1, 2...) can be obtained as:
K p
10

. Values for pK
i
and pK
w
used in this
work were found in (Greenstein and Winitz, 1961; King, 1951; Smith et al., 1937; Izatt and
Capítulo 3 70
Christensen, 1970). For the amino acids and water, the effect of the temperature on the
equilibrium constants were taken into account by adjusting eq 3.9 to the available experimental
data.
( )
p A
B
C
i
K
T
= +
+
(3.9)
where A, B and C are adjustable parameters and T the temperature in Kelvin. Values of A, B
and C are given in Table 3.5. Dissociation constants (pK
Ai
) for citric acid at 25 °C were found in
(Izatt and Christensen, 1970) and their values are following: pK
A1
=3.131; pK
A2
=4.762;
pK
A3
=6.397.
Table 3.5. Coefficients for correlating pK values with temperature
a

substance A B C deviation
b

glycine pK
1
pK
2

2.2641
3.2125
3.6254
1663.1
-256.2
-44.88
0.071
0.004
dl-alanine pK
1

pK
2

2.2839
3.3074
3.1010
1594.9
-252.4
-55.04
0.008
0.010
l-proline pK
1
pK
2

1.9489
3.1602
0.1900
2232.3
-271.1
0.2012
0.155
0.011
dl-valine pK
1
pK
2

2.2984
3.1443
-0.0430
1650.4
-276.14
-47.11
0.272
0.009
dl-serine pK
1
pK
2

1.9823
2.5850
13.944
1724.7
-229.7
-37.73
0.015
0.003
water pK
w
8.0618 1077.3 -116.5 0.011
a
According to eq 3.9 in the text, temperature in K.
b
Mean relative deviation (%) between estimated and
experimental values calculated by:
n K
K K
n n
cal
n n 100

p
p p
exp
exp
×
(
(
¸
(

¸

|
|
¹
|

\
|


,where n=number of experimental data points.

Reactions 3.4 to 3.7 correspond to citric acid and sodium citrate dissociations in the aqueous
system. Eleven species are present in the mixture mentioned above. The system of equations
(equilibrium relations, mass and electroneutrality balances) that allows to calculate
simultaneously the activity coefficients and the concentration of species is given below:
Thermodynamic Modeling

71
equilibrium relations
oL
w
L
m
v
v
c
c c
K




*
RCOOH) (NH
*
H
*
) RCOO (NH
RCOOH NH
H RCOO NH
1
3
3
3
3
+
+ − +
+
+ − +
=
γ
γ γ
(3.10)
oL
w
L
m
v
v
c
c c
K




*
) RCOO (NH
*
H
*
) RCOO (NH
RCOO NH
H RCOO NH
2
3
2
3
2
− +
+ −
− +
+ −
=
γ
γ γ
(3.11)
oL
w
L
m
v
v
c
c c
K




*
citric H
*
citric H
*
H
citric H
citric H H
A1
3
2
+
3
2
+
γ
γ γ
− −
= (3.12)
oL
w
L
m
v
v
c
c c
K



*
citric H
*
Hcitric
*
H
citric H
Hcitric H
A2
2
- - +
-
2
- - +
γ
γ γ
= (3.13)
oL
w
L
m
v
v
c
c c
K



*
Hcitric
*
citric
*
H
Hcitric
citric H
A3
- - +
- -
- - - +
γ
γ γ
= (3.14)
2
*
OH
*
H
OH H



|
|
¹
|

\
|
=
− +
− +
oL
w
L
m
w
w
v
v
a
c c K
γ γ
(3.15)
electroneutrality balance
− − − − − − − − + + +
+ + + + = + +
citric Hcitric citric H RCOO NH OH Na RCOOH NH H
3 2 c
2 2 3
c c c c c c c (3.16)
mass balances
- - - - - -
2
3 3 3 citric Hcitric citric H
citric H citric,0 Na citric,0 H
c c c c c c + + + = + (3.17)
+
=
Na
citric,0 Na
3
c c (3.18)
− + − + − +
+ = =
RCOO NH RCOOH NH RCOO NH 0 , RCOO NH
RCOOH NH
2 3 3 3
2
+c c c c c (3.19)

+ =
OH
O H O,0 H
2 2
c c c (3.20)
where c in the equations above are the molar concentrations (mol L
-1
) of the species;
L
m
v and v
w
oL

are the molar volumes of the mixture and water (L mol
-1
), respectively and
*
i
γ refers to the
activity coefficient of specie i in the asymmetric convention (infinite dilution basis). The
subscript 0 refers to the initial concentration of solute.
Capítulo 3 72
In this work, the activity coefficient (in the asymmetric convention) of a specie i is expressed
by long-range (electrostatic interactions) and short-range (physical interactions) contributions:
R L
i
R S
i i
− −
γ + γ = γ
* * *
ln ln ln (3.21)
where, the superscript S-R and L-R refer respectively to short- and long-range interactions.
For the solvent (water), the activity coefficient is calculated in the symmetric convention as
follows:
ln ln ln γ γ γ
w w
S R
w
L R
= +
− −
(3.22)
In addition, the short-range contribution combines the UNIFAC-Larsen model with solvation
equations for chemical interactions between water and ionic species. As a result, structural
parameters of solvated species, concentrations, and activity coefficients are estimated considering
the occurrence of solvation. The equations corresponding to this phenomenon are presented
below:
structural parameters of species k
R R Nh R
k
H
k k w
= + (3.23)
Q Q Nh Q
k
H
k k w
= + (3.24)
where superscript H refers to the hydrated state, Nh
k
is the hydration number of specie k, and
R
w
and Q
w
stands for volume and area parameters of water, respectively.
mole fractions of hydrated species
x
x Nh x
Nh x
w
H
w j j
j j w
j j
j j w
=






,
,
1
(3.25)
x
x
Nh x
i
H i
j j
j j w
=



1
,
(3.26)
where x
w
H
and
H
i
x are the mole fractions of water and ionic species i, respectively. The
summations in eq 3.25 and 3.26 should be performed upon all species j in solution except for
water.
Thermodynamic Modeling

73
conversion between hydrated and non-hydrated activity coefficients
γ γ
w
S R
w
S R H w
H
w
x
x
− −
=
,
(3.27)
[ ]
γ γ γ
i
S R
i
S R H i
H
i
w
S R H
w
H
Nh
i
x
x
x
− − −

=
, ,
(3.28)
γ
w
S R H − ,
and
H R S
i
, −
γ are calculated by the UNIFAC-Larsen equation using mole fractions and
structural parameters of the hydrated species.
The selected UNIFAC model proposed by Larsen et al. (1987) gives symmetric activity
coefficients in the mole fraction scale (
i
γ ) that have to be normalized to the standard state for the
ionic species (asymmetric convention):

− =
i i i
γ γ γ ln ln ln
*
(3.29)
where

i
γ is the activity coefficient at infinite dilution of all solutes in the mixture:
i
w
x
i
γ = γ


1
lim (3.30)
where x
w
is the water mole fraction.
The long-range contribution to the activity coefficients was calculated using the Debye-
Hückel equation, with the closest approach parameter equal to 17.1, as suggested by Achard et al.
(1994a).
The solution non-ideality is considered in the eqs 3.10 to 3.15 in which true concentration
ratios can be estimated using the framework developed by Achard et al. (1994a). It follows that
the pH and a
w
values can be estimated using the true concentrations of the ionic species (a
w
= γ
w

x
w
and ] log[
H
*
H
+
+
− = c pH
c
γ ). In the particular case of pH, activity of H
+
ions in molar
concentration scale can be obtained from the calculated activity in molar fraction scale by the
equation below:
(
(
¸
(

¸

γ
− =
(
¸
(

¸

γ − =
+ +
+
+ oL
w
c
v
x
c pH
H
*
H
H
*
log log
H
(3.31)
Capítulo 3 74
where
+
H
c and
+
H
x stand for molar concentration and molar fraction of ions H
+
, respectively;
and v
w
oL
is the molar volume of water.
3.4.2 Solubility of Amino Acids in Aqueous Solutions
The solid-liquid equilibrium of an aqueous amino acid solution and the pure solid amino
acid, can be written in terms of the zwitterion form:
S L
o
f f x
A
*
A A
=
± ±
γ (3.32)
L
o
f is the reference fugacity in the asymmetric convention, and
S
f
A
is the fugacity of the
pure solid amino acid.
The ratio
L
o
S
/f f
A
can be related to the temperature of the system through the following
relation:
|
¹
|

\
|
− =
RT
∆h
R
∆s
f
f
L
o
S
exp
A
(3.33)
where ∆s and ∆h are the change in molar enthalpy and entropy of the amino acid from the
reference state to the solid state (Prausnitz et al., 1996). These values are generally obtained from
solubility studies. The activity coefficient
*

γ can be estimated by the UNIFAC-Larsen model
combined with the Debye-Hückel equation, once the mole fraction of the amino acid is known.
Therefore, it is necessary to perform an iterative method to obtain the solubility of that molecule
using eqs 3.32 and 3.33. ∆s and ∆h in eq 3.33 were adjusted by regression from the solubility
data of the amino acids dl-alanine, l-proline and dl-serine. The values of ∆s and ∆h for the other
amino acids studied in this work (glycine and dl-valine) were maintained equal to the literature
values (see Table 3.6).
Table 3.6. ∆s e ∆h values used in this work (in bold script) and obtained from literature
amino acid ∆s/R ∆h/R (K)
glycine 2.4184
b
; 2.61
b
, 2.042
d
1654.55
b
; 1711.1
c
, 885.689
d

dl-alanine 1.3628
a
; 0.2127
b
, 2.060
d
1390.96
a
; 1115.07
b
;1110.97
d
, 1107.2
c

dl-valine -0.1305
b
, 2.305
d
1264.96
b
; 754.9
c
, 1217.946
d

dl-serine 2.7173
b
; 4.003
a
2265.35
b
; 2717.7
c
; 2641.65
a
l-proline 5.1835
a
; 2.6235
b
; 1.96
c
1476.6
a
; 959.77
b
; 654.2
c
a
Adjusted in this work.
b
Khoshkbarchi and Vera, 1996,
c
Fasman, 1976,
d
Gupta and Heidemann, 1990.
Thermodynamic Modeling

75
The total solubility (x
A
) can be written as:
-
A A A
A
x x x x + + =
+ ±
(3.34)
where,
-
A A A
, , x x x
+ ±
are the mole fractions of the zwitterion, cationic and anionic species.
There are several literature data reporting the influence of acids, bases and salts on the
aqueous solubility of amino acids within a certain pH range. Pradhan and Vera (1998) reported
the effect of the addition of strong acids and bases on the solubility of amino acids at various pH
values. It was observed that there was no significant influence of different ions (K
+
, Na
+
, Cl
-
, and
NO
3
-
) on the solubility of dl-alanine at 25 °C within the pH range of 2 to 10. It is also reported in
literature (Carta and Tola, 1996) experimental solubility data of some amino acids at various pH
and NaCl concentrations. The pH was varied from 0 to 13 using HCl or NaOH. The experimental
data shows the dependence of the solubility of amino acids on the addition of salts, bases and
acids. Experimental data reported by Needham et al. (1971) shows the influence of acid or base
on the solubility of amino acids in two different solvents: water and ethanol. In the present work,
the solubility of amino acids in aqueous electrolyte solutions has been estimated using the
UNIFAC-Larsen model combined with the Debye-Hückel equation.
3.5 Results and Discussion
3.5.1 Density data
In the calculations of activity coefficients and true concentration of species, it is necessary to
estimate the densities of the amino acid solutions. Linear equations were fitted to experimental
density data at 25 °C for the amino acid systems considered in the present work. The equations
are given in Table 3.7. To estimate solubilities of the amino acids in water, density data at
temperatures other than 25 °C were also required. We used the following approach to estimate
such densities: the densities of the amino acid (i) solutions at 25 °C relative to water (subscript w)
at the same reference temperature were multiplied by the density of water at the desired
temperature, as can be seen by eq 3.35:
) (
) 25 (
) 25 (
) ( t
C
C
t
w
w
i
i
ρ
ρ
ρ
= ρ
o
o
(3.35)
Capítulo 3 76
Table 3.7. Linear fitting for the densities of amino acids, salts, acid and base in aqueous
solutions
a

system equation: ρ=ρ(w
t
) g
cm
-3

reference R
2

glycine + water 0.4327w
t
+0.9967 Soto et al. (1998); this
work
3,36
glycine + acid buffer 0.4229w
t
+0.9971 this work 4,80
glycine + basic buffer 0.4349w
t
+0.9974 this work 3,88
dl-alanine + water 0.3241w
t
+0.9970 Dalton and Schmidt
(1933); this work
4,60
dl-alanine + acid
buffer
0.3258w
t
+0.9977 this work 3,52
dl-alanine + basic
buffer
0.3072w
t
+0.9970 this work 3,62
l-proline + water 0.2982w
t
+0.9962 this work 3,32
l-proline + acid buffer 0.2988w
t
+0.9972 this work 4,80
l-proline + basic
buffer
0.2935w
t
+0.9970 this work 4,20
l-arginine + water 0.2782w
t
+0.9970 this work 3,64
l-arginine + acid
buffer
0.3073w
t
+0.9975 this work 3,16
l-arginine + basic
buffer
0.2781w
t
+0.9974 this work 3,48
l-serine + water 0.4450w
t
+0.9998 Yan et al. (1999) 4,60
dl-valine + water 0.2276w
t
+0.9970 Dalton and Schmidt (1933) 4,60
NaOH + water 1.0599w
t
+1.0013 Lobo and Quaresma (1989) 4,80
HCl + water 0.5037w
t
+0.9994 Lobo and Quaresma (1989) 4,53
NaCl + water 0.7585w
t
+0.9980 Lobo and Quaresma (1989) 3,42
a
Correlation coefficients such as 0.999980 given as 4,80.

The prediction capacity of eq 3.35 was tested using experimental densities of binary aqueous
solutions containing a non-electrolyte solute (sucrose) (Norrish, 1967) or an electrolyte solute
(KCl) (Lobo and Quaresma, 1989) at various temperatures (20 - 90 °C for sucrose and 25 - 100
°C for KCl) and solute concentrations (1 - 85 mass % for sucrose and 0.7 - 20 mass % for KCl).
Results and Discussion

77
In the case of sucrose solutions the reference temperature was 20 °C and for KCl solutions it was
25 °C. Mean relative deviations between experimental density data and those calculated by eq
3.35 were equal to 0.09% and 0.1% for sucrose and KCl solutions, respectively. Such result
clearly shows the good performance of the selected approach to estimate densities at various
temperatures. Eq 3.35 is based on the assumption that dρ
i
/dw
i
(the derivative of solution density
in relation to solute concentration) does not dependent on temperature, so that the solution
density at some specified solute concentration changes with temperature only as a consequence of
the change in the solvent density.
For the ternary (amino acid + base/acid + water, or amino acid + salt + water) and the
quaternary (amino acid + base/acid + salt + water) systems it was utilized a semi-empirical
equation for calculating the corresponding densities. It requires densities of the binary solutions
(solute + solvent) as is shown below (Zafarani-Moattar et al., 1995):
( ) ρ ρ ρ ρ
mix w i w
i i w
− = −


,
(3.36)
ρ
mix
, ρ
w
, and ρ
i
represent the densities of the multicomponent mixture, density of pure water
and density of the binary aqueous mixtures containing component i, respectively. Component i
could be a salt, an acid or a base. The summation in eq 3.36 should be performed upon all the
components i in solution, except for water.
To test the accuracy of the eq 3.36, calculated densities for aqueous solutions of glycine and
salts were compared with experimental ones from literature (Lobo and Quaresma, 1989; Soto et
al., 1999). The relative mean deviation was found to be 0.3% for the ternary systems: glycine +
NaCl + water and glycine + KCl + water. This result confirms the applicability of the equation
for calculating the densities of the multicomponent systems involved in this work.
3.5.2 Equilibrium data
The UNIFAC-Larsen model was then used to calculate the short-range contribution to the
activity coefficients of the different species in solution considering the hydration of ions H
+
and
Na
+
, with hydration numbers (Nh) equal to 2.959 and 2.606, respectively (Achard et al., 1994a).
The interaction parameters used were found in Achard et al. (1994a) (for interaction parameters
involving charged species), Kuramochi et al. (1996, 1997a) (for interaction parameters involving
new groups such as sc-CH, α-CH) and Larsen et al. (1987). Interaction energies (u
ij
) between the
charged species Na
+
and Cl
-
and the specific groups of the amino acids (COOH and NH
2
) were
Capítulo 3 78
adjusted using the available solubility data. The option for adjusting the interaction energies u
ij
,
instead of the interaction parameters a
ij
, allows to reduce the number of parameters to be fitted.
Interaction parameters a
ij
are related to the interaction energies u
ij
and u
jj
according to eq 3.37,
given below:
jj ij ij
u u a − = (3.37)
Adjusted group interaction energies (u
ij
) between the amino acid groups (COOH and NH
2
)
and the ions Na
+
and Cl
-
are given in Table 3.8. The u
jj
values (u
COOH,COOH
and u
NH2,NH2
) required
for a
ij
calculations had been estimated by the following expression (Achard et al., 1994b) and are
reported in Table 3.9:
u a a u
jj jw wj ww
= − + (3.38)
where the subscripts w and j stand for water and specie j, respectively.
Values for
w
a
, COOH
,
COOH , w
a ,
w
a
,
2
NH
,
2
NH , w
a were taken from Larsen's tables (Larsen et al.,
1987), and the values for u
w w ,
,
+ +
Na , Na
u and
-
Cl ,
-
Cl
u were obtained in ref (Achard et al., 1994b).
Table 3.8. Group interaction energies u
ij
(K) between ions and amino acid characteristic groups
Na
+
Cl
-

COOH 442.785 -1459.36
NH
2
357.724 -1521.50

Table 3.9. Group interaction energies u
jj
between like groups

u
jj
(K)
COOH-COOH -622.18
NH
2
-NH
2
-878.10
In Figure 3.3 it is shown a comparison between experimental and predicted a
w
values for the
amino acid l-proline in water and in very diluted buffer solutions. Figure 3.4 depicts the estimated
pH values for mixtures containing glycine in three different solvents (water, acid buffer and basic
buffer). As can be observed, the UNIFAC-Larsen model combined with the Debye-Hückel term
yields a good fit to these data. Mean relative average deviations for the systems with glycine, l-
proline, l-arginine and dl-alanine are given in Table 3.10.
Results and Discussion

79
0.0 0.1 0.2 0.3 0.4 0.5
0.80
0.85
0.90
0.95
1.00
pI
model
acid buffer
model
basic buffer
model
a
w
w

Figure 3.3. Experimental and calculated a
w
of aqueous solutions containing l-proline.

The deviations for the solubility data are also given in Table 3.10, and some of the results are
shown in Figure 3.5. The model reproduces well the observed experimental data.
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
pI
acid buffer
basic buffer
model
p
H
w
a
(g g
-1
)

Figure 3.4. Experimental and estimated pH of glycine solutions.
Capítulo 3 80
0 20 40 60 80 100 120
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
dl-valine
dl-serine
dl-alanine
glycine
l-proline
w
a
t (
o
C)

Figure 3.5. Experimental and calculated solubility of some amino acids.

Results concerning solubility of glycine in salt solutions (glycine + NaCl + water +
HCl/NaOH) at different pH values are presented in Figure 3.6. Most experimental data are from
the work of Carta and Tola (1996). It must be stressed that studies about pH influence on
solubility of amino acids do not report the concentrations of acid and base utilized. Some works
(Pradhan and Vera, 1998) report the final pH of a mixture containing the amino acid, the acid or
the base, while others (Carta and Tola, 1996) provide the pH of the initial solutions – before the
addition of the amino acid. In both cases, the amount of acid or base in the present work was
selected – through analysis of the estimated pH by the model – to obtain the desirable pH or
when pH
calc
=pH
exp
. For example, when experimental solubility data as a function of final pH is
given for a mixture containing the amino acid + acid or base, the concentrations of acid or base
were selected when the pH estimated by the model reached the experimental pH. In the other
case, when the pH values of the initial solutions were reported, the total acid or base
concentrations were selected from the analysis of the experimental and estimated pH for aqueous
solutions containing the acid or the base only. Finally, the concentrations of acid or base were
utilized to estimate solubility of some amino acids in a wide range of pH values. Figure 3.6
shows the results of the calculated solubilities of the amino acid glycine at various pH values and
NaCl concentrations. From Figure 3.6, it can be seen that good agreement is observed between
experimental and calculated solubilities at high NaCl concentrations (5 and 15%), while the
Results and Discussion

81
calculated results for the amino acid at the isoelectric point agree well with experimental data
from (Needham et al., 1971, Dalton and Schmidt, 1933).
Table 3.10 also shows the mean relative deviations for a
w
, pH and solubility of the amino acids in
aqueous solutions considering the ideal case, that is activity coefficients for all species equal to 1
(
i i ∀ ,
γ =1). It was observed that, the a
w
values calculated for the ideal case had an expressive
difference from experimental results for the systems containing the amino acid proline (mean
relative deviations of about 1.4%). For the other systems containing the amino acids glycine,
alanine and arginine this difference was less pronounced but greater than that obtained
considering the non-ideality of the systems. Such deviations are outside the range of the
experimental error for a
w
measurements provided by the manufacturer (0.3%) and also much
higher than the standard deviation observed in the experimental measurements (0.1%).

0 2 4 6 8 10 12 14
0.16
0.18
0.20
0.22
0.24
0.26
0.28

Carta and Tola (1996) 0% NaCl
Carta and Tola (1996) 5% NaCl
Carta and Tola (1996) 15% NaCl
Needham et al. (1971)
Dalton and Schmidt (1933)
w
a
pH

Figure 3.6. Experimental and estimated solubilities of glycine at various pH values and NaCl
concentrations.
Capítulo 3 82
Table 3.10. Average relative deviations between calculated and experimental data
property % deviation reference
real ideal
a
w

pI
this work
literature
acid buffer
basic buffer


0.11
0.08
0.18
0.15

0.51
0.12
0.96
0.40

this work
Kuramochi et al., 1997b
this work
this work
pH
pI
acid buffer
basic buffer

3.63
2.93
0.94
3.61
3.12
0.91
this work
this work
this work
solubility
proline + water
alanine + water
glycine + water
valine + water
serine + water
glycine + NaCl + HCl or
NaOH
alanine + NaCl
valine + NaCl
serine + NaCl
alanine + HCl or NaOH
glycine + HCl or NaOH

4.26
4.51
4.99
3.77
2.43
5.41

0.13
0.55
6.73
6.47
7.66
59.60
11.26
14.99
7.18
6.90
14.32

8.64
12.06
17.76
12.44
28.76
Fasman, 1976
Fasman, 1976
Fasman, 1976
Fasman, 1976
Fasman, 1976
Carta and Tola, 1996

Khoshkbarchi and Vera, 1997
Khoshkbarchi and Vera, 1997
Khoshkbarchi and Vera, 1997
Pradhan and Vera, 1998
Needham et al., 1971
total deviation 3.05 17.63

Results and Discussion

83
Concerning the deviations between real and ideal pH values, it was verified slight differences
between them. This case can be further visualized in Figure 3.7 in which the distribution of the
three ionic species of glycine as a function of pH is shown. Unfortunately, numerical data for the
distribution of the species is not available in the literature, but it can be noted in Figure 3.7 that
the ranges in which such species appear in the various forms agree well with those reported by
(Max et al., 1998) (cationic pH 0 to 5; zwitterionic pH 0 to 12.5 and anionic pH 7 to 14).
Although it is not possible to verify which of the curves (real or ideal pH) represent well the ionic
distribution of glycine, those curves yield different concentrations of the species within the pH
range of 1-3.5 and 8-11.
Large discrepancies can also be noted between calculated (ideal case) and experimental data
for solubility (see Table 3.10). In the case of glycine and valine the values of ∆s and ∆h were
taken from literature and then were not adjusted in the present work even though the results of the
calculated solubility considering the ideal case presented very high deviations from experimental
data.

0 2 4 6 8 10 12 14
0.0
0.2
0.4
0.6
0.8
1.0
NH
+
3
RCOOH NH
2
RCOO
-
NH
+
3
RCOO
-
r
e
l
a
t
i
v
e

c
o
n
c
e
n
t
r
a
t
i
o
n
pH

Figure 3.7. Comparison between the distribution of the ionic species of glycine as a function of
real () and ideal (----) pH.
Capítulo 3 84
3.6 Conclusions
This work has presented experimental data on water activity, pH and density of mixtures
containing amino acids. The use of the UNIFAC group contribution model with solvation
equations as proposed by Achard et al. (1994a) combined with the Debye-Hückel term, was
extended to estimate physical-chemical properties such as water activity, pH and solubility of
amino acids. The model provided good results for such properties. The interaction parameters
between Na
+
and Cl
-
and the amino acid specific groups were adjusted, yielding reasonable
results concerning pH influence on the solubility of the amino acids in more complex mixtures.
3.7 References
Achard, C.; Dussap, C.-G.; Gros, J. B. Representation of Vapour-Liquid Equilibria in Water-
Alcohol-Electrolyte Mixtures with a Modified UNIFAC Group-Contribution Method. Fluid
Phase Equilibria 1994b, 98, 71- 89.
Achard, C.; Dussap, C.-G.; Gros, J.-B. Prediction of pH in Complex Aqueous Mixtures Using a
Group Contribution Method. AIChE J. 1994a, 40, 1210-1222.
Carta, R.; Tola, G. Solubilities of l-Cystine, l-Tyrosine, l-Leucine, and Glycine in Aqueous
Solutions at Various pHs and NaCl Concentrations. J. Chem. Eng. Data 1996, 41, 414-417.
Chen, C. C.; Zhu, Y.; Evans, L. B. Phase Partitioning of Biomolecules: Solubilities of
Aminoacids. Biotechnol. Prog. 1989, 5, 111-118.
Dalton, J. B.; Schmidt, C. L. A. The Solubilities of Certain Amino Acids in Water, the Densities
of their Solutions at Twenty-Five Degrees, and the Calculated Heats of Solution and Partial
Molal Volumes. J. Biol. Chem. 1933, 103, 549-578.
Fasman, G. D. Handbook of Biochemistry and Molecular Biology, 3rd ed., Physical and
Chemical Data; CRC Press: Cleveland, 1976; pp 115.
Greenstein, J. P.; Winitz, M. Chemistry of the Amino Acids Vol. 1; John Wiley & Sons: New
York, 1961; pp 486-491.
Gupta, R. B.; Heidemann, R. A. Solubility Models for Amino Acids and Antibiotics. AIChE J.
1990, 36, 333-341.
References

85
Izatt, R. M.; Christensen, J. J. Heats of Proton Ionization, pK and Related Thermodynamic
Quantities. In Handbook of Biochemistry: Selected Data for Molecular Biology, 2nd ed.; H.
A. Sober, Ed., The Chem. Rubber Co.: Cleveland, 1970; pp J 79.
Khoshkbarchi, M. K.; Vera, J. H. A Simplified Pertubed Hard-Sphere Model for the Activity
Coefficients of Amino Acids and Peptides in Aqueous Solutions. Ind. Eng. Chem. Res. 1996,
35, 4319-4327.
Khoshkbarchi, M. K.; Vera, J. H. Effect of NaCl and KCl on the Solubility of Amino Acids in
Aqueous Solutions at 298.2 K: Measurement and Modeling. Ind. Eng. Chem. Res. 1997, 36,
2445-2451.
King, E. The Ionization Constants of Glycine and the Effect of Sodium Chloride upon its Second
Ionization. J. Am. Chem. Soc. 1951, 73, 155-159.
Kuramochi, H., Noritomi, D. H., Hoshino, D., Nagahama, K. Measurements of Vapor Pressures
of Aqueous Amino Acids Solutions and Determination of Activity Coefficients of Amino
Acids. J. Chem. Eng. Data 1997b, 42, 470-474.
Kuramochi, H., Noritomi, D. H., Nagahama, K. Representation of Activity Coefficients of
Fundamental Biochemicals in Water by the UNIFAC Model. Fluid Phase Equilibria 1997a,
130, 117-132.
Kuramochi, H.; Noritomi, H.; Hoshino, D.; Nagahama, K. Measurement of Solubilities of Two
Amino Acids in Water and Prediction by the UNIFAC Model. Biotechnol. Prog. 1996, 12,
371-379.
Larsen, B. L.; Rasmussen, P.; Fredenslund, A. A Modified UNIFAC Group-Contribution Model
for Prediction of Phase Equilibria and Heats of Mixing. Ind. Eng, Chem. Res. 1987, 26,
2274-2286.
Lobo, V. M. M.; Quaresma, J. L. Handbook of Electrolyte Solutions; Elsevier Sci. Publishers:
New York, 1989; parts A and B, pp 1643, 1818-1819.
Maffia, M. C.; Meirelles, A. J. A. Water Activity and pH in Aqueous Polycarboxylic Acid
Systems. J. Chem. Eng. Data 2000,46, 582-587.
Max, J-J; Trudel M.; Chapados, C. Infrared Titration of Aqueous Glycine. Applied Spectroscopy
1998, 52, 226-233.
Capítulo 3 86
Nass, K. K. Representation of the Solubility Behavior of Amino Acids in Water. AIChE J. 1988,
34, 1257-1266.
Needham, T. E.; Paruta, A. N.; Gerraughty, R. J. Solubility of Amino Acids in Pure Solvent
Systems. Journal of Pharmaceutical Sciences 1971, 60, 565-567.
Ninni, L.; Camargo, M. S.; Meirelles, A. J. A. Modeling and Prediction of pH and Water Activity
in Aqueous Amino Acid Solutions. Comp. Chem. Eng. 1999, 23 (Supplement), 383-386.
Norrish, R. S. Selected Tables of Physical Properties of Sugar Solutions. Scientific and
Technical Surveys; The British Food Manufacturing Industries Research Association:
Leatherhead, 1967; pp 26.
Peres, A. M., Macedo, E. A. Representation of Solubilities of Amino Acids Using the UNIQUAC
Model for Electrolytes. Chem. Eng. Sci. 1994, 49, 3803-3812.
Pinho, S. P., Silva, C. M., Macedo, E. A. Solubility of Amino Acids: a Group Contribution
Model Involving Phase and Chemical Equilibria. Ind. Eng. Chem. Res. 1994, 33, 1341-1347.
Pitzer, K. S. Electrolytes: From Dilute Solutions to Fused Salts. J. Am. Chem. Soc. 1980, 102,
2902-2906.
Pradhan, A. A.; Vera, J. H. Effect of Acids and Bases on the Solubility of Amino Acids. Fluid
Phase Equilibria 1998, 152, 121-132.
Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. Molecular Thermodynamics of Fluid Phase
Equilibria; Englewood Cliffs: New York, 1996; pp 418-420.
Smith, P. K.; Taylor, A. C.; Smith, E. R. B. Thermodynamic Properties of Solutions of Amino
Acids and Related Substances. J. Biol. Chem. 1937, 109-123.
Soto, A.; Alberto, A.; Khoshkbarchi, M. K. Effect of Cation and Anion of an Electrolyte on
Apparent Molar Volume, Isoentropic Compressibility and Refractive Index of Glycine in
Aqueous Solutions. Biophysical Chemistry 1999, 76, 73-82.
Soto, A.; Alberto, A.; Khoshkbarchi, M. K. Experimental Data and Modelling of Apparent Molar
Volumes, Isoentropic Compressibilities and Refractive Indices in Aqueous Solutions of
Glycine + NaCl. Biophysical Chemistry 1998, 74, 165-173.
Velezmoro, C. E.; Meirelles, A. J. A. Water Activity in Solutions Containing Organic Acids.
Drying Technol. 1998, 16, 1789-1805.
References

87
Yan, Z. N.; Wang, J. J.; Liu, W. B.; Lu, J. S. Apparent Molar Volumes and Viscosity B-
Coefficients of Some Alpha-Amino Acids in Aqueous Solutions from 278.15 to 308.15 K.
Thermochimica Acta 1999, 334, 17-27.
Zafarani-Moattar, M. T.; Salabat, A.; Kabiri-Badr, M. Volumetric Properties of PEG + Salt +
Water. Journal of Chemical Engineering Data 1995, 40, 559-562.
3.8 Acknowledgment
This work was supported financially by research grants from Fundação de Amparo à
Pesquisa do Estado de São Paulo (FAPESP - Proc. 98/12302-3) and Conselho Nacional de
Desenvolvimento Científico e Tecnológico (CNPq - Proc. 145941/99, 466680/00-7 and
521011/95-7).
Introduction

89

Capítulo 4

KINEMATIC VISCOSITIES OF
POLY(ETHYLENE GLYCOL) BLENDS
Luciana Ninni, Wong H. Fung, Antonio J. A. Meirelles
LASEFI - Faculty of Food Engineering - State University of Campinas (UNICAMP)
P.O. Box 6121 Zip Code 13083-970 Campinas – SP, Brazil










Trabalho publicado na revista Journal of Chemical & Engineering Data, v. 47, p. 835 – 838,
2002.

Capítulo 4

90

Introduction

91
4. Kinematic Viscosities of Poly(ethylene glycol) Blends
4.1 Abstract
Kinematic viscosities of binary and multicomponent mixtures containing poly(ethylene
glycols) were measured as a function of temperature. Viscosities of the binary mixtures were
used to calculate the excess molar Gibbs energy of activation for viscous flow G
*E
. The G
*E

values are positive over the entire composition range and they increase as the difference in the
molecular masses of the two polymers increases. The group-contribution viscosity model GC-
UNIMOD was employed to correlate the viscosity of the binary systems and then to predict the
viscosities of the multicomponent mixtures. Average absolute deviation around 3.5% was
obtained for the viscosity calculations using the GC-UNIMOD model.
4.2 Introduction
Poly(ethylene glycol) (PEG) belongs to a class of synthetic polymers that finds several
industrial applications due to their availability in a wide range of molecular masses besides
having unusual combination of properties such as water solubility, lubricity, low toxicity, etc.
Blends of different molecular masses are desired to obtain combinations of properties. For
instance, blends of PEG 300 and 1450 are available from most suppliers for use in
pharmaceutical and cosmetic products. PEGs are also used as processing aids in making other
products. An example is the use of molten PEGs in heat transfer baths (Powell, 1980). Hence,
because of the industrial interest, experimental viscosity data and models for estimating
viscosities of PEG mixtures are important. In a previous paper (Cruz et al., 2000), a generalized
correlation based on the number of carbon atoms was developed for calculating kinematic
viscosities of pure PEGs in a molecular mass range between 200 and 3350 as a function of
temperature. In the case of mixtures, there are many empirical or semi-empirical equations in the
literature (Domínguez et al., 2000; Nhaesi and Asfour, 2000) for estimating viscosity data.
Among them there is the group-contribution viscosity model GC-UNIMOD (Cao et al., 1993),
which requires binary interaction parameters between constituent groups of molecules in the
mixture and viscosities of pure components. The GC-UNIMOD has been used to predict the
viscosity of systems containing mixtures of organic solvents (Domínguez et al., 2000; Nhaesi and
Asfour, 2000) and working fluid pairs such as methanol and some poly(ethylene glycol) dimethyl
ether (PEGDME) mixtures (Herraiz et al., 1999). In the case of the last mentioned system, the
Capítulo 4

92
authors reported an average absolute deviation around 20 % for viscosity predictions and
observed that these predictions were getting worse with the increase of the molecular size of
PEGDME. The aim of this paper is to extend our prior studies (Cruz et al., 2000) to binary and
multicomponent blends of PEGs using the GC-UNIMOD model. Therefore, the kinematic
viscosities of blends of PEGs covering a wide range of mixture molecular masses were measured
at various temperatures, and the GC-UNIMOD model was used to correlate and predict these
experimental data.
4.3 Experimental
4.3.1 Materials
The samples of poly(ethylene glycol) of different average molecular masses, ranging from
(200 to 3350) g mol
-1
, were supplied by Sigma, except PEG 1500, which was purchased from
Fluka. They were analytical grade reagents used without further purification. Their
polydispersity index was determined by gel permeation chromatography (GPC) using a
Ultrahydrogel column Waters device. The following experimental conditions were employed for
the GPC runs: water as the mobile phase at a rate of 0.8 mL min
-1
, injection temperature of
313.15 K, volume of the sample injected equal to 100 µL, and a refractive index detector.
The water content of the PEG samples was previously determined by Karl Fischer titration
using a Metrohm device (Switzerland). In Table 4.1 the characteristics of the polymers used in
this work are given together with the nominal and true carbon numbers for each PEG. The
denomination nominal carbon number refers to carbon numbers calculated from the molecular
mass given by PEG manufacturers such as PEG 200, PEG 3350, while the true carbon number
was evaluated from the results of the GPC molecular mass distribution.
Table 4.1. Poly(ethylene glycol) Characterization
PEG
avg molecular
mass
(M
n
)
polydispersity
index
water content
(mass %)
NCN
a
TCN
b

200 202 1.095 0.20 ± 0.02 8 8
400 400 1.086 0.23 ± 0.02 18 18
600 616 1.069 0.27 ± 0.02 26 28
1000 987 1.067 1.54 ± 0.02 44 44
1500 1468 1.069 1.09 ± 0.01 68 66
3350 2806 1.073 0.75 ± 0.01 152 126
a
Nominal carbon number based on the molecular mass according to the PEG denomination.
b
True carbon number
based on the average molecular mass determined by GPC.
Results and Discussion

93
4.3.2 Apparatus and procedures
The solutions were prepared by mass on an analytical balance (Sartorius Analytic - GmbH)
with ± 0.1 mg accuracy. The estimated error in mass fraction was 1 in 10 000. Calibrated
Cannon-Fenske type viscometers were used to measure the kinematic viscosities at different
temperatures (Cannon Instrument Co.). The viscometer sizes were 150, 200, 300 and 350,
appropriate for the range of viscosity values measured in the present work. The viscometers were
placed in a constant temperature water bath controlled to ± 0.1 °C. An electronic timer with 0.01
s accuracy was used for measuring the efflux time. The experiments were replicated at least five
times for each PEG mixture and the results reported are the average values.
4.4 Results and Discussion
The viscosity data are presented in Tables 4.2 to 4.5 as a function of mass fraction (w) and
temperature T. The kinematic viscosities ν were calculated from the efflux time and the
instrument constant provided by the manufacturer. The maximal standard deviation of the
measured viscosities was 1.3 × 10
-7
m
2
s
-1
and the minimal, 1.9 × 10
-9
m
2
s
-1
, resulting in an
average deviation of 2.6 × 10
-8
m
2
s
-1
. The average uncertainty of the viscosity values was
estimated to be equal to 0.3 %.
Table 4.2. Viscosity of Binary Blends of Poly(ethylene glycol)s at Various Temperatures
T/K w
1
ν/10
-6
m
2
s
-1
w
1
ν/10
-6
m
2
s
-1
w
1
ν/10
-6
m
2
s
-1

200(1)+400(2) 200(1)+600(2) 400(1)+600(2)
293.15 0.4999 77.53 0.5031 93.89 0.5043 127.44
298.15 0.4999 59.73 0.5031 72.95 0.4999 99.61
303.15 0.4999 46.49 0.5031 58.41 0.5007 77.66
308.15 0.4999 37.18 0.5031 46.52 0.5007 60.86
313.15 0.4999 30.13 0.5031 37.90 0.5007 49.25
318.15 0.4999 24.73 0.5031 30.94 0.5007 40.49
323.15 0.4999 20.62 0.5031 25.80 0.5007 33.58
600(1)+1000(2) 400(1)+1500(2) 600(1)+3350(2)
333.15 0.5001 36.19 0.4918 40.99 0.4999 97.32
343.15 0.5001 26.44 0.4918 29.81 0.4999 70.85
353.15 0.5001 21.27 0.4836 23.11 0.4999 54.27
363.15 0.5001 15.87 0.4836 18.20 0.4999 41.97
Capítulo 4

94
Table 4.2. (continued)
T/K w
1
ν/10
-6
m
2
s
-1
w
1
ν/10
-6
m
2
s
-1


1000(1)+3350(
2)
400(1)+3350(2)
333.15 0.5001 114.79 0.5000 88.09
343.15 0.5001 84.10 0.5000 63.76
353.15 0.5001 62.07 0.5000 48.92
363.15 0.5001 49.94 0.5000 38.66

Table 4.3. Viscosity of Binary Blends Containing Poly(ethylene glycol) 400 at Various
Concentrations
T/K w
1
ν/10
-6
m
2
s
-1
w
1
ν/10
-6
m
2
s
-1
w
1
ν/10
-6
m
2
s
-1

400(1)+600(2) 400(1)+1500(2) 400(1)+3350(2)
293.15 0.1010 154.96
293.15 0.2000 146.25
293.15 0.2999 139.42
293.15 0.3992 133.61
293.15 0.7000 121.82
293.15 0.8998 110.98
293.15 0.9482 105.77
333.15 0.0996 28.04 0.0511 70.91 0.0273 224.82
333.15 0.2001 26.87 0.0997 68.69 0.0513 219.00
333.15 0.2999 25.63 0.1513 63.49 0.1029 196.70
333.15 0.3992 24.84 0.2500 56.83 0.2000 162.98
333.15 0.5000 23.76 0.3496 50.02 0.2500 147.14
333.15 0.7000 21.80 0.7006 31.17 0.6998 52.48
333.15 0.8998 20.04 0.9003 23.27 0.8002 39.62
333.15 0.9482 19.78

In this work, the GC-UNIMOD model was selected for viscosity calculations. Moreover, the
viscosity data for the systems containing PEG 400 have been used to calculate the excess molar
Gibbs energy of activation for viscous flow.
Results and Discussion

95
Table 4.4. Viscosities of Ternary Mixtures of Poly(ethylene glycol)s at Various Temperatures
T/K w
1
w
2
ν/10
-6
m
2
s
-1
w
1
w
2
ν/10
-6
m
2
s
-1

400(1)+600(2)+1000(3) 600(1)+1000(2)+1500(3)
333.15 0.3339 0.3334 30.26 0.3332 0.3335 47.10
343.15 0.3339 0.3334 22.01 0.3332 0.3335 34.20
353.15 0.3339 0.3334 16.77 0.3332 0.3335 26.67
363.15 0.3339 0.3334 13.21 0.3332 0.3335 21.00
400(1)+1000(2)+3350(3) 1000(1)+1500(2)+3350(3)
333.15 0.3334 0.3331 71.25 0.3330 0.3334 109.27
343.15 0.3334 0.3331 52.29 0.3330 0.3334 79.28
353.15 0.3334 0.3331 38.85 0.3330 0.3334 60.59
363.15 0.3334 0.3331 31.23 0.3330 0.3334 47.59

Table 4.5. Viscosities of a Multicomponent Mixture of Poly(ethylene glycol)s at Various
Temperatures
T/K w
1
w
2
w
3
w
4
ν/10
-6
m
2
s
-1

400(1)+600(2)+1000(3)+1500(4)+3350(5)
333.15 0.2000 0.1997 0.1999 0.2002 61.40
343.15 0.2000 0.1997 0.1999 0.2002 44.84
353.15 0.2000 0.1997 0.1999 0.2002 34.13
363.15 0.2000 0.1997 0.1999 0.2002 26.77
4.4.1 Excess molar Gibbs energy of activation for viscous flow
Viscosities of binary mixtures (PEGs 400+600, 400+1500 and 400+3350) have been used to
calculate the excess molar Gibbs energy of activation for viscous flow, G
*E
, through the
following equation:
|
¹
|

\
|
− =

=
n
i
i i i
M x M RT G
1
E *
) ln( ) ln( ν ν (4.1)
where M is the mixture molar mass, R is the gas constant, T is the absolute temperature, ν
and ν
i
are the kinematic viscosities of the mixture and of the pure components, respectively, x
i
is
component i mole fraction, and n is the number of components in the mixture. The pure
component viscosities were calculated using the equation proposed by Cruz et al. (2000) and the
true carbon numbers given in Table 4.1.
Capítulo 4

96
The G
*E
(in J mol
-1
) for the binary mixtures was fitted to a Redlich-Kister’s equation:

=
− =
3
0
E *
) (
p
p
j i p j i
x x A x x G
(4.2)
where x
i
and x
j
denotes the mole fractions of components i and j, respectively, and A
p
are
adjustable parameters. The G
*E
experimental values and the Redlich-Kister’s fitted polynomials
are plotted in Figure 4.1. The coefficients A
p
and the average absolute deviations (AAD) between
experimental and calculated G
*E
values are given in Table 4.6. The G
*E
values are positive over
the entire composition range and they increase as the difference in the molecular masses of the
two polymers increases. The increase of the temperature has a similar effect (Fig. 4.1).
0.0 0.2 0.4 0.6 0.8 1.0
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
G
*
E
/

1
0
3

J

m
o
l
-
1

x
1

Figure 4.1. Excess molar Gibbs energy of activation for viscous flow of PEG binary mixtures
(■ 400+600 at 293.15 K; □ 400+600 at 333.15 K; ▲ 400+1500; 400+3350; —
Redlich-Kister).

4.4.2 GC-UNIMOD model
The GC-UNIMOD viscosity equation, like the UNIFAC (Fredenslund et al., 1975) model, is
expressed as the sum of combinatorial and residual terms. The combinatorial term of both the
viscosity equation and the UNIFAC model is dependent on the molecular size and shape, while
Results and Discussion

97
the residual term takes into account the interaction energy between different groups present in the
mixture. In order to estimate the kinematic viscosities of the PEG mixtures, the viscosities of the
pure components must be known. The equation for pure PEGs (Cruz et al., 2000) was used for
calculating the polymer viscosities required in the GC-UNIMOD model.
Table 4.6. Coefficients of the Redlich-Kister Equation (A
p
) and Average Absolute Deviations
(AAD) between Experimental and Calculated G
*E
values
T/K A
0
A
1
A
2
A
3
AAD
400 (1)+600 (2)
293.15 495.3292 70.6134 409.3651 5.30
333.15 1172.855 73.1809 1249.925 5.20
400 (1)+1500 (2)
333.15 6549.193 3957.076 2049.027 5960.119 4.01
400 (1)+3350(2)
333.15 15771.32 8805.127 16811.02 22896.8 4.63

As a first attempt, the GC-UNIMOD model, with energy interaction parameters obtained
from UNIFAC-VLE tables (Skjold-Jorgensen, 1979), was tested for estimating the viscosities of
the binary mixtures measured in this work. For these predictions the PEG molecule was divided
in three functional groups: CH
2
, OH and CH
2
O. An average absolute deviation of 15.9 % was
obtained. It was also verified that specially for the binary mixtures containing polymers with
very different molecular masses (for example, the mixture of PEGs 600 and 3350) the model
predictions were poor. A readjustment of the energy interaction parameters becomes necessary
to improve the predictive capability of the model. The experimental data of the binary mixtures
(Tables 4.2 and 4.3) were used for this readjustment.
As suggested by Herraiz et al. (1999), different possibilities of dividing the polymer
molecules in functional groups were considered. PEG molecules can be divided in the following
groups: OH, CH
2
O, CH
2
(option 1) (Skjold-Jorgensen, 1979); OH, CH
2
CH
2
O, CH
2
(option 2)
(Ninni et al., 1999; Rasmussen and Rasmussen, 1989); or CH
2
OH, CH
2
O, CH
2
(option 3)
(Herskowitz and Gottlieb, 1984). As can be seen, three pairs of interaction parameters have to be
adjusted in each one of the above options. The new parameters were obtained using a nonlinear
estimation method (Marquardt, 1963). The average absolute deviations (AAD) between
experimental and estimated values were calculated according to the following equation:
Capítulo 4

98
m
m
i i
i i 100
AAD
1 , exp
, calc , exp
×
(
(
¸
(

¸

|
|
¹
|

\
|

=

=
ν
ν ν
(4.3)
where ν
exp,i
and ν
calc,i
are the experimental and calculated kinematic viscosities, respectively,
and m is the number of experimental points.
Better agreement between experimental and calculated data for the binary mixtures was
obtained using options 1 and 2, with AAD global values of 3.62 % and 3.76 %, respectively (see
Table 4.7). The GC-UNIMOD model was also capable of predicting rather well the viscosities of
the multicomponent mixtures, but in this case options 2 and 3 result in slightly better AAD global
values, around 2.9 % (Table 4.7). The energy interaction parameters for option 2 are given in
Table 4.8. Figure 4.2 shows the viscosity values calculated by the GC-UNIMOD model
compared with the experimental binary data given in Table 4.2. The comparison for the ternary
and quinary mixtures is shown in Figure 4.3.
Table 4.7. Average Absolute Deviations (AAD) for the Viscosity Estimation
AAD AAD
Binary
Mixtures Option 1

Option 2


Option 3

Multicomponent
Mixtures Option 1 Option 2 Option 3
200+400 4.16 4.20 4.74 400+600+1000 5.53 5.50 5.77
200+600 7.45 7.53 8.54 600+1000+1500 1.09 1.12 1.00
400+600 3.20 3.43 3.27 400+1000+3350 1.46 1.04 0.94
600+1000 2.20 2.18 2.28 1000+1500+3350 4.86 4.59 4.74
400+1500

2.89 2.75 3.07
400+600+1000+
1500+3350
2.52 2.03 2.03
600+3350 1.90 1.10 1.09
1000+3350 0.97 0.82 0.91
400+3350

4.54 5.36 6.75

global 3.62 3.76 4.14 3.09 2.86 2.90

Results and Discussion

99
Table 4.8. Energy Interaction Parameters (in Kelvin) for the GC-UNIMOD Model
CH
2
CH
2
O CH
2
OH
CH
2
CH
2
O - -93.8155 -250.3
CH
2
52.7182 - -101.3
OH 41.4333 85.4514 -

290 300 310 320 330 340 350 360 370 380 390
20
40
60
80
100
120
140
ν
/

1
0
-
6

m
2

s
-
1
T/ K

Figure 4.2. Experimental and correlated viscosities of binary PEG mixtures (experimental data
of Table 4.2: ▲ 200+400; ┉ 200+600; ■ 400+600; ○ 400+1500; 400+3350; ▼
600+1000; + 600+3350; 1000+3350; — GC-UNIMOD/ option 2).

As can be observed in Table 4.7 the highest AAD values were obtained for those binary
mixtures containing PEGs 200 and/or 400. In the case of the multicomponent systems no
experimental values were measured for mixtures containing PEG 200. Probably this is the reason
why the AAD values for the predictions (multicomponent mixtures) were slightly lower than
those obtained for the correlation of the binary mixtures (Table 4.7).
Capítulo 4

100
330 335 340 345 350 355 360 365
0
20
40
60
80
100
120
ν
/

1
0
-
6

m
2

s
-
1
T/ K

Figure 4.3. Experimental and predicted viscosities of multicomponent PEG mixtures (∇
400+600+1000; 400+1000+3350; ┉ 400+600+1000+1500+3350;
600+1000+1500; ○ 1000+1500+3350; — GC-UNIMOD/ option 2).

The GC-UNIMOD model, with new set of interaction parameters, works well for estimating
the viscosity data measured in the present work. For comparison purposes, some correlation and
prediction results reported in the literature should be mentioned. The authors of the GC-
UNIMOD reported deviations in the range of 2.7 to 5.3 % for predictions of mixture viscosities
using energy interaction parameters from UNIFAC-VLE tables. It is also depicted in the
literature (Herraiz et al., 1999) predictions of the GC-UNIMOD model for mixtures of some
polyethylene glycol dimethyl ethers and methanol over a wide range of temperatures. The results
showed that the predictions were getting worse with the increase of the polymer molecular mass.
The reported average absolute deviations were around 20 %. Rabelo et al. (2000) employed the
GC-UNIMOD model for predicting the viscosity of fatty systems. In this case the AAD values
varied in the range of 0.8 to 14.0 %. The comparison can also include some empirical equations
used for correlating experimental viscosities of aqueous systems containing PEGs. Some of these
works (Gündüz, 1996; Mei et al., 1995; González-Tello et al., 1994; Gündüz, 2000) reported
results for concentrated and/or diluted solutions with AAD values varying in the range of 1.2 to
6.8 %.
Conclusions

101
The unique experimental data available in the literature (Powell, 1980) for PEG mixtures is
the viscosity of a blend of PEGs 300 and 1500 at equal mass proportion of both compounds. This
data was compared to the value predicted by the GC-UNIMOD model (option 2), yielding an
AAD value of 10.6 %. Although such deviation is comparatively greater than those obtained for
the correlation and prediction of our experimental data, it should be noted that for this prediction
we had to use the nominal carbon number, not the true carbon number, to estimate the pure PEG
viscosities.
4.5 Conclusions
In this work the kinematic viscosities of blends of poly(ethylene glycol)s with nominal
molecular masses ranging from (200 to 3350) g mol
-1
were determined at various temperatures.
The GC-UNIMOD model was tested for correlating and predicting viscosity data. In most cases
a good agreement between experimental and calculated values was obtained.
4.6 Literature Cited
Cao, W.; Knudsen, K.; Fredenslund, A.; Rasmussen, P. Group-Contribution Viscosity
Predictions of Liquid Mixtures Using UNIFAC-VLE Parameters. Ind. Eng. Chem. Res. 1993, 32,
2088-2092.
Cruz, M. S.; Chumpitaz, L. D. A.; Alves, J. G. L. F.; Meirelles, A. J. A. Kinematic
Viscosities of Poly(Ethylene Glycols). J. Chem. Eng. Data 2000, 45, 61-63.
Domínguez, M.; Pardo, J. I.; Gascón, I.; Royo, F. M.; Urieta, J. S. Viscosities of the Ternary
Mixture (2-butanol + n-hexane + 1-butylamine) at 298.15 and 313.15 K. Fluid Phase Equilib.
2000, 169, 277-292.
Fredenslund, A.; Jones, R. L.; Prausnitz, J. M. Group Contribution Estimation of Activity
Coefficients in Nonideal Liquid Mixtures. AIChE J. 1975, 21, 1086-1099.
González-Tello, P.; Camacho, F.; Blázquez, G. Density and Viscosity of Concentrated
Aqueous Solutions of Polyethylene Glycol. J. Chem. Eng. Data 1994, 39, 611-614.
Gündüz, U. Evaluation of Viscosities of Polymer-Water Solutions used in Aqueous Two-
Phase Systems. J. Chromatogr. B 1996, 680, 263-266.
Capítulo 4

102
Gündüz, U. Viscosity Prediction of Polyethylene Glycol - Dextran - Water Solutions used in
Aqueous Two-Phase Systems. J. Chromatogr. B 2000, 743, 181-185.
Herraiz, J.; Shen, S.; Fernández, J.; Coronas, A. Thermophysical Properties of Methanol +
Some Polyethylene Glycol Dimethyl Ether by UNIFAC and DISQUAC Group-Contribution
Models for Absorption Heat Pumps. Fluid Phase Equilib. 1999, 155, 327-337.
Herskowitz, M.; Gottlieb, M. Vapor-Liquid Equilibrium in Aqueous Solutions of Various
Glycols and Poly(ethylene glycols). 2. Tetraethylene Glycol and Estimation of UNIFAC
Parameters. J. Chem. Eng. Data 1984, 29, 450-452.
Marquardt, D. W. An Algorithm for Least-Square Estimation of Nonlinear Parameters. J.
Soc. Indust. Appl. Math. 1963, 11, 431-441.
Mei, L.-H.; Lin, D.-Q.; Zhu, Z.-Q.; Han, Z.-X. Densities and Viscosities of Polyethylene
Glycol + Salt + Water Systems at 20°C. J. Chem. Eng. Data 1995, 40, 1168-1171.
Nhaesi, A. H.; Asfour, A-F A. Prediction of the Viscosity of Multi-Component Liquid
Mixtures: a Generalized McAllister Three-Body Interaction Model. Chem. Eng. Sci. 2000, 55,
2861-2873.
Ninni, L.; Camargo, M. S.; Meirelles, A. J. A. Water Activity in Poly(Ethylene Glycol)
Aqueous Solutions. Thermochim. Acta 1999, 328, 169-176.
Powell, G. M. Polyethylene Glycol. In Handbook of Water Soluble Gums and Resins;
Davidson, R. L., Ed.; McGraw-Hill Book Company: New York, 1980; Chapter 18.
Rabelo, J.; Batista, E.; Cavaleri, F. W.; Meirelles, A. J. A. Viscosity Prediction for Fatty
Systems. J. Am. Oil Chem. Soc. 2000, 77, 1255-1261.
Rasmussen, D.; Rasmussen, P. Phase Equilibria in Aqueous Polymer Solutions. Chem. Eng.
Progress 1989, 85, 50-56.
Skjold-Jorgensen, S.; Kolbe, B.; Gmehling, J.; Rasmussen, P. Vapor-Liquid Equilibria by
UNIFAC Group-Contribution. Revision and extension. Ind. Eng. Chem. Process Des. Dev. 1979,
18, 714-722.
Literature Cited

103
4.7 Acknowledgments
The authors gratefully acknowledge the financial support of FAPESP (Fundação de Amparo
à Pesquisa do Estado de São Paulo – 01/10137-6) and CNPq (Conselho Nacional de
Desenvolvimento Científico e Tecnológico - 521011/95-7, 466680/00-7, 145941/99).
Capítulo 5

105

Capítulo 5

KINEMATIC VISCOSITIES OF
POLY(ETHYLENE GLYCOL) AQUEOUS
SOLUTIONS
Luciana Ninni, Helcio Burd, Wong H. Fung, Antonio J. A. Meirelles
Department of Food Engineering (DEA/FEA), State University of Campinas – UNICAMP,
Cidade Universitária “Zeferino Vaz”, PO Box 6121, Zip Code 13083-970, Campinas, SP, Brazil










Trabalho publicado na revista Journal of Chemical & Engineering Data, v. 48, p. 324 – 329,
2003.
Capítulo 5

106


Introduction

107
5. Kinematic Viscosities of Poly(ethylene glycol) Aqueous Solutions
5.1 Abstract
Kinematic viscosities of aqueous mixtures containing poly(ethylene glycol)s (PEG) with
nominal molecular masses ranging from (200 to 10,000) g⋅mol
-1
have been determined at various
concentrations and temperatures. The binary experimental data were used for adjusting the
parameters of a Kumar-like equation. Relative errors around 5.5% were observed between
calculated and experimental results. The adjusted parameters also allowed the viscosity
prediction for multicomponent mixtures, with an overall deviation of 9.9%.
5.2 Introduction
Poly(ethylene glycol) (PEG) is a linear, neutral polyether, available in a variety of molecular
masses, soluble in water and most organic solvents (Powell, 1980). PEG has a variety of
properties that make it suitable for applications in the biomedical and biotechnological areas.
Among them one can cite the formation of two-phase systems with aqueous solutions of other
polymers or salts, its non-toxicity, and the formation of complex with metal cations, etc. (Harris,
1992) Aqueous two-phase systems are increasingly used for separation of biomolecules, cells
and cell particles (Bradoo et al., 1999; Gündüz, 2000). Such systems are composed of two
incompatible polymers, e.g. dextran and poly(ethylene glycol) (PEG), or of one polymer (PEG)
and a salt (usually a phosphate or sulfate salt). The use of PEGs in aqueous two-phase systems
(ATPS) has been reported by several authors (Zaslavsky, 1995; Silva et al., 2001; Grossmann et
al., 1998), since such systems allow the separation and purification of biomolecules in a gentle
environment. Extraction using ATPS can be performed continuously in available commercial
separators (Coimbra et al., 1995; Porto et al., 2000; Gongxi et al., 1999). The advantages of
using aqueous two-phase extraction lie in volume reduction, high capacity and fast separation;
moreover, it is relatively straightforward to scale up a separation process (Persson et al., 1999).
In addition to the knowledge of phase equilibrium, data on the properties of the phases are
necessary for the design of ATPS extraction processes in large-scale applications. The knowledge
of physical properties, such as the viscosity of aqueous mixtures containing PEGs, is important
because these solutions are used in many industrial and biotechnological applications (Albertson,
1971). However, it is not viable to measure viscosities at all conditions of interest, and
consequently, methods for the estimation of viscosities at various temperatures, concentrations
Capítulo 5

108
and solute types are of great practical interest. Numerous empirical and semi-empirical methods
have been proposed to calculate liquid viscosity of mixtures (Mehrotra et al., 1996; Gündüz,
1996; Pereira et al., 2001; Cao et al., 1993; Gündüz, 2000). Nevertheless, some equations are
inappropriate for concentrated and/or multicomponent mixtures.
In the present work, the reformulated Kumar’s equation (Pereira et al., 2001) has been
employed to correlate and predict kinematic viscosities of aqueous solutions containing PEGs in
a wide range of polymer concentrations and at different temperatures. The difference between
the data treatment from this work and other related works (Gündüz, 1996; Pereira et al., 2001;
Cao et al., 1993; Gündüz, 2000; Mei et al., 1995; Gonzaléz-Tello et al., 1994) is that viscosity of
concentrated PEG solutions were employed together with literature viscosity data at various
temperatures, concentrations and PEG molecular masses. Moreover, for each polymer employed,
hydration numbers were estimated and compared with available values in the literature (Antonsen
and Hoffman, 1992; Kirinčič and Klofutar, 1999; Moulik and Gupta, 1989, Bieze et al., 1994;
Bisal et al., 1990). A generalized correlation for estimating kinematic viscosities of PEG
aqueous solutions as a function of polymer molecular mass, concentration and temperature was
also developed.
5.3 Experimental Section
5.3.1 Materials
Analytical grade PEGs with nominal molecular masses ranging from (200 to 10,000) g⋅mol
-1

were purchased from SIGMA and used without further purification. Water was distilled,
deionized with a Mili-Q water system (Millipore, USA) and used for preparing the PEG
solutions. The solutions were prepared on a mass basis using an analytical balance (Sartorius
Analytic, GmbH) accurate to ±0.0001 g. PEG concentrations in the solutions varied up to 50
mass % for systems containing PEGs 200, 400, 600, 1000 and 1500, up to 30 mass% for PEG
3350, and up to 25 mass % for PEGs 8000 and 10 000. The estimated error in the mass fractions
was 2 in 10 000. PEGs were characterized according to their molecular masses and water content
using the procedures described in a previous article (Cruz et al., 2000). The characteristics of the
polymers used in the present work are given in Table 5.1. The water content in the polymer
samples was taken into account for calculating the solution concentrations.
Experimental Section

109
Table 5.1. Average relative molar masses (M
n
), polydispersity index and water content of PEGs
used in this work
PEG
avg rel molar mass
(M
n
)
polydispersity
index
water content (mass %)
200 202 1.095 0.20 ± 0.02
400 400 1.086 0.23 ± 0.02
600 616 1.069 0.27 ± 0.02
1000 987 1.067 1.54 ± 0.02
1500 1529 1.069 1.09 ± 0.01
3350 2806 1.073 0.75 ± 0.01
8000 7975 1.099 0.36 ± 0.08
10 000 10 475 1.106 0.66 ± 0.05

5.3.2 Apparatus and Procedures
Calibrated Cannon-Fenske glass capillary viscometers (sizes 50, 75, 100, 150 and 200) were
used to measure the kinematic viscosities (Cannon Instrument Co., USA). The viscometers were
placed in a water bath (Anton Paar, Austria) for holding the temperature constant within ±0.1 °C.
The measurements were performed at temperatures varying between (293.15 and 323.15) K. An
electronic timer with 0.01 s accuracy was used for measuring the efflux time. The experiments
were replicated at least three times for each PEG mixture and the results given below are the
average values. The standard deviations of the viscosity determinations varied within the range
1.9×10
-7
m
2
⋅s
-1
to 1.1×10
-9
m
2
⋅s
-1
, being the lowest figures obtained for the lowest viscosity
values. The variation coefficient varies within the range 0.1% to 0.7%, so that the uncertainty of
the experimental measurements can be estimated as not higher than 0.7%.
5.4 Results and Discussion
The kinematic viscosities for the binary and multicomponent PEG solutions determined in this
work are given in Tables 5.2 – 5.4.
Capítulo 5

110
Table 5.2. Kinematic viscosities, ν, in aqueous solutions of PEGs at 293.15 K and various mass
fractions, w
w ν / 10
-6
m
2
s
-1
w ν / 10
-6
m
2
s
-1
w ν / 10
-6
m
2
s
-1

PEG 400 PEG 600 PEG 1000
0.0248 1.121 0.0249 1.143 0.0243 1.169
0.0497 1.239 0.0501 1.311 0.0492 1.356
0.0748 1.379 0.0749 1.470 0.0737 1.588
0.1002 1.543 0.0996 1.656 0.0985 1.854
0.1529 1.941 0.1491 2.143 0.1477 2.510
0.1998 2.415 0.1994 2.793 0.1960 3.406
0.2998 3.956 0.2993 4.900 0.2953 6.304
PEG 1500 PEG 3350
0.0246 1.216 0.0248 1.345
0.0498 1.456 0.0492 1.747
0.0741 1.751 0.0744 2.267
0.0992 2.094 0.0995 2.936
0.1482 2.994 0.1489 4.669
0.1980 4.252 0.1983 7.215
0.2967 8.443 0.2987 16.693

Table 5.3. Kinematic viscosities, ν, in aqueous solutions of PEGs at various temperatures and
mass fractions, w
T/K w ν / 10
-6
m
2
s
-1
w ν / 10
-6
m
2
s
-1

PEG 200 PEG 400
293.15 0.2487 3.167
303.15 0.2399 1.829 0.2487 2.365
313.15 0.2399 1.416 0.2487 1.828
323.15 0.2399 1.148
293.15 0.4902 7.231 0.4958 10.917
303.15 0.4902 4.971 0.4958 7.425
313.15 0.4902 3.649
323.15 0.4902 2.835 0.4958 4.180
PEG 600 PEG 1000
293.15 0.2490 3.684 0.2275 4.425
303.15 0.2490 2.745 0.2275 3.237
313.15 0.2490 2.089 0.2275 2.539
323.15 0.2490 1.696 0.2275 2.028
293.15 0.4899 15.534 0.4777 21.823
303.15 0.4899 10.389 0.4777 14.524
313.15 0.4899 7.361 0.4777 10.256
323.15 0.4899 5.501 0.4777 7.502

Results and Discussion

111
Table 5.3. (cont.)
PEG 1500 PEG 3350
293.15 0.1223 2.387 0.0969 2.965
303.15 0.1223 1.833 0.0969 2.329
313.15 0.1223 1.457 0.0969 1.838
323.15 0.1223 1.191 0.0969 1.496
293.15 0.2349 6.046 0.2492 12.125
303.15 0.2349 4.473 0.2492 8.863
313.15 0.2349 3.444 0.2492 6.757
323.15 0.2349 2.716 0.2492 5.312
293.15 0.4534 28.784 0.3389 22.824
303.15 0.4534 19.228 0.3389 16.250
313.15 0.3389 12.086
323.15 0.4534 9.967 0.3389 9.281
PEG 8000 PEG 10 000
293.15 0.0530 2.748 0.0499 3.119
303.15 0.0530 2.128 0.0499 2.376
313.15 0.0530 1.679 0.0499 1.922
323.15 0.0530 1.364 0.0499 1.519
293.15 0.1000 5.623
303.15 0.1000 4.396 0.0977 5.196
313.15 0.1000 3.338 0.0977 4.037
323.15 0.1000 2.664 0.0977 2.980
293.15 0.2554 33.655
303.15 0.2554 24.476 0.2441 31.977
313.15 0.2554 18.904 0.2441 24.496
323.15 0.2554 14.715 0.2441 19.314

Capítulo 5

112
Table 5.4. Kinematic viscosities, ν, of multicomponent poly(ethylene glycol) aqueous solutions
at various temperatures
T/K w
1
w
2
w
3
w
4
ν / 10
-6
m
2
s
-1

PEG 400 (1) + PEG 600(2) + PEG 1000(3) + PEG 1500(4)
293.15 0.1005 0.1006 0.0999 0.1032 11.430
303.15 0.1005 0.1006 0.0999 0.1032 8.166
313.15 0.1005 0.1006 0.0999 0.1032 5.936
323.15 0.1005 0.1006 0.0999 0.1032 4.657

PEG 3350 (1) + PEG 8000(2) + PEG 10 000(3)
293.15 0.0997 0.1012 0.0495 23.468
303.15 0.0997 0.1012 0.0495 17.170
313.15 0.0997 0.1012 0.0495 13.051
323.15 0.0997 0.1012 0.0495 10.248

PEG 400 (1) + PEG 8000(2)
293.15 0.1003 0.1000 8.221
303.15 0.1003 0.1000 6.078
313.15 0.1003 0.1000 4.675
323.15 0.1003 0.1000 3.702

5.4.1 Kumar’s equation for viscosity correlation in PEG mixtures
The Kumar expression (Kumar, 1993) is one of the equations presented in the literature for
calculating dynamic viscosities of solutions as a function of solute concentration. Initially,
Kumar’s expression was formulated to estimate viscosities in aqueous and non-aqueous salt
solutions. Such correlation involves two parameters, the hydration number and the ion-solvent
interaction coefficient, as given below:
hx
bx

+ =
1
1
0
η
η
(5.1)
where η is the solution dynamic viscosity, η
0
is the solvent dynamic viscosity, x is the solute
mole fraction, b and h are the solute-solvent interaction coefficient and the hydration number
(number of molecules of bound solvent per molecule of solute), respectively.
The Kumar expression was later reformulated (Pereira et al., 2001) to allow the calculation
of kinematic viscosities of aqueous solutions containing salts or organic solutes. The resulting
equation is a function of relative density between solution and solvent, mass fraction of solute
Results and Discussion

113
and temperature. Pereira et al. (2001) investigated the predictive capacity of the equation for
ternary, quaternary and quinary solutions.
The reformulated Kumar equation was used in this work to correlate and predict the
kinematic viscosities of the aqueous PEG mixtures. The equations for binary and
multicomponent systems are given as follows:
(
¸
(

¸


+
ρ
= ν
Hw
Bw
1
1
1
rel
rel
(5.2)
(
(
(
¸
(

¸


+ =


i
i
i
i
Hw
Bw
1
1
1
rel
rel
ρ
ν
(5.3)
where the parameters B and H are interpreted, respectively, as a coefficient of solute-solvent
interaction and as the number of solvent molecules solvating the solute, both functions of
temperature; w is the mass fraction of solute; ν
rel
and ρ
rel
are the ratios between the kinematic
viscosities and densities of the solution and the corresponding values for the pure solvent,
respectively and i represents the solutes in the multicomponent mixture. Dependence of the
parameters B and H on temperature is given by polynomials in (T - T
ref
), as follows:
n
ref
n
n ref
) ( T T B B B − + =

(5.4)
n
ref
n
n ref
) ( T T H H H − + =

(5.5)
where T
ref
is the reference temperature equal to 298.15 K.
As can be seen, eqs 5.2 and 5.3 require the information of densities of pure water and of
aqueous PEG solutions. For this reason, linear equations were fitted to experimental density data
of PEGs in water (González-Tello et al., 1994; Kirinčič and Klofutar, 1998) at 298.15 K. The
coefficients of the linear equations are given in Table 5.5.


Capítulo 5

114
Table 5.5. Linear fittings for the densities of poly(ethylene glycol)s in aqueous solutions
equation: ρ /(10
3
kg m
-3
) = a + b w
PEG a b R
2
PEG a b R
2

300 0.996 04 0.166 97 3,66
a
4000 0.996 01 0.17836 3,71
400 0.996 11 0.169 85 3,74 6000 0.995 75 0.18059 3,72
600 0.996 23 0.172 33 3,68 8000
b
0.996 67 0.17701 3,49
900 0.995 68 0.176 62 3,76 10 000 0.995 42 0.18295 3,75
1000 0.995 70 0.176 30 3,76 12 000 0.995 86 0.18234 3,68
1500 0.995 63 0.175 87 3,72 15 000 0.995 86 0.18140 3,76
2000 0.995 77 0.178 62 3,75 20 000 0.996 04 0.18068 3,63
3000 0.995 74 0.179 43 3,71 35 000 0.996 17 0.18117 3,80
3350
b
0.996 48 0.175 64 3,67
a
Correlation coefficients such as 0.999 80 given as 3,80.
b
Experimental density data from González-Tello et al. (1994). All other sets of density data were taken from
Kirinčič and Klofutar, 1998.

In addition, to estimate the densities of the PEGs in water at temperatures other than 298.15
K the following approach was used, since these data are not available in the literature: the
densities of the PEG solutions at 298.15 K relative to water (subscript w) at the same reference
temperature were multiplied by the density of water at the desired temperature:
) (
) 15 . 298 (
) 15 . 298 (
) ( T
K
K
T
w
w
i
i
ρ
ρ
ρ
ρ =
(5.6)
In dealing with multicomponent systems, a semi-empirical equation was used for calculating
the corresponding densities. It requires densities of the binary solutions (solute + solvent), and is
given by:
( )


− = −
w i i
w i w mix
,
ρ ρ ρ ρ
(5.7)
where ρ
mix
, ρ
w
, and ρ
i
represent the densities of the multicomponent mixture, of pure water
and of the binary aqueous mixtures containing component i, respectively. The summation in eq
5.7 should be performed upon all the components i in solution, except for water. Equations 5.6
and 5.7 were used in a prior work (Ninni and Meirelles, 2001) for estimating the densities of
aqueous solutions containing different solutes with very good agreement between experimental
and calculated values.
Results and Discussion

115
By nonlinear regression (Marquardt, 1963) of eq 5.2, the values of B
ref
, B
1
, B
2
and H
ref
, H
1
,
H
2
were obtained for each PEG + water system. A data bank comprising experimental viscosities
of binary systems reported in this work and those from literature (Mei et al., 1995; Kirinčič and
Klofutar, 1999), in a wide range of PEG molecular masses, concentrations and temperatures was
employed in the correlations. The data bank comprises a total of 259 experimental points. The
adjusted parameters as well as the average absolute deviations (AAD) between experimental and
calculated viscosities are presented in Table 5.6.

Table 5.6. Adjusted coefficients of eqs 5.2, 5.4 and 5.5 for calculating kinematic viscosities of
aqueous PEG solutions
a
PEG w T/K B
ref
B
1
/10
-3
B
2
/10
-3
H
ref
H
1
/10
-3
H
2
/10
-3
AAD
b
200 ≤0.49 303 – 323 3.4444 55.3 -2.38 1.3862 -5.44 0.106 5.9
300 ≤0.29 298 4.5680 0.0 0.0 1.4256 0.0 0.0 2.4
400 ≤0.50 293 – 323 5.4186 37.8 -2.38 1.4487 -12.7 0.29 3.4
600 ≤0.49 293 – 323 6.7086 -1.05 -0.002 1.5693 -9.62 0.068 4.9
900 ≤0.21 298 7.3506 0.0 0.0 1.5891 0.0 0.0 3.7
1000 ≤0.50 293 – 323 7.6304 2.0 1.03 1.6762 0.385 -0.36 5.7
1500 ≤0.45 293 – 323 10.3957 -1.89 1.10 1.7942 -7.76 0.0 7.9
2000 ≤0.40 298 11.3580 0.0 0.0 1.9760 0.0 0.0 4.7
3000 ≤0.12 298 13.1893 0.0 0.0 2.2100 0.0 0.0 2.7
3350 ≤0.34 293 – 323 16.2610 -0.02 1.27 2.2050 1.94 0.039 10.8
4000 ≤0.10 298 16.6385 0.0 0.0 2.3424 0.0 0.0 2.6
6000 ≤0.25 298 24.3935 0.0 0.0 2.7669 0.0 0.0 7.8
8000 ≤0.26 293 – 323 29.8474 -4.56 -4.60 2.9778 -8.34 0.253 4.3
10 000 ≤0.24 293 – 323 34.1380 5.36 -10.3 3.2098 3.66 0.0 7.0
12 000 ≤0.04 298 41.0170 0.0 0.0 3.2500 0.0 0.0 2.5
15 000 ≤0.04 298 45.3602 0.0 0.0 3.4566 0.0 0.0 3.4
20 000 ≤0.03 298 48.9736 0.0 0.0 3.6566 0.0 0.0 2.3
35 000 ≤0.03 298 70.2625 0.0 0.0 4.0611 0.0 0.0 3.8
global 5.4
a
Experimental viscosity data from the present work and from the literature (Mei et al., 1995; Kirinčič and Klofutar,
1999).
b

n
n
i i
i i 100
AAD
1 , exp
, calc , exp
×
(
(
¸
(

¸

|
|
¹
|

\
|

=

=
ν
ν ν


Capítulo 5

116
It was observed that H
ref
and B
ref
exhibit a dependence on the polymer molecular mass as
depicted in Figures 5.1 and 5.2. For this reason, it is interesting to compare H
ref
values obtained
in the present work with the hydration numbers reported in the literature, since there are reports
about the variation of hydration numbers with molecular mass of different PEG molecules.
0 5000 10000 15000 20000 25000 30000 35000
0
10
20
30
40
50
60
70
B
ref
PEG molecular mass

Figure 5.1. B
ref
values as a function of PEG molecular mass.

The number of water molecules, which are thought to be bounded in the PEG molecules,
varies widely according to the experimental measurement techniques, which include
conductometry (Bisal et al., 1990), differential scanning calorimetry (DSC) (Antonsen and
Hoffman, 1992), I. R. spectrophotometry (Moulik and Gupta, 1989), and viscometry (Kirinčič
and Klofutar, 1999). Values between 0.9 and 6 water molecules per monomer have been quoted
(Bieze et al., 1994). It is also reported that PEGs of higher molecular masses show increased
hydration (Antonsen and Hoffman, 1992).
According to Moulik and Gupta (1989), 4 to 6 water molecules per oxygen for a PEG of
molecular mass 20 000 have been obtained, and a minor temperature dependence of hydration
has been observed. The same authors obtained hydration numbers (molecules of bound water per
monomer unit) of 1.53 for PEG 400 and 1.93 for PEG 600. Bahri and Guveli (1988) have
reported that 0.16 to 3.20 water molecules may bind a single oxygen center for PEGs with
molecular masses in the range of (200 and 1000) g⋅mol
-1
. Hydration numbers obtained from
Results and Discussion

117
viscometry were presented by Bisal et al. (1990) for PEGs 200, 300, 400, 600 and 1000 with
values equal to 0.79, 0.95, 1.07, 1.21 and 2.45, respectively. Hydration numbers determined by
viscometry are also reported by Kirinčič and Klofutar (1999). The values varied from 2.06 to
28.44 for PEGs with molecular masses between (300 and 35 000) g⋅mol
-1
. The explanation given
for such high values was that, as the molecular mass increases, the polymer chain begins to fold
in on itself, forming segment-segment interactions as it traps additional, more loosely bound
water between the segments. A different result was obtained by DSC determinations of bound
water (Antonsen and Hoffman, 1992). In the work of Antonsen and Hoffman (1992), the amount
of water bound per polymer repeated unit varied from 2.3 to 3.8 for PEG molecular masses
between 200 and 2.3 × 10
5
g⋅mol
-1
. They also observed that the amount bound at higher
molecular masses is greater than at low molecular masses.
0 5000 10000 15000 20000 25000 30000 35000
1.0
1.5
2.0
2.5
3.0
3.5
4.0
H
r
e
f
PEG molecular mass

Figure 5.2. H
ref
values as a function of PEG molecular mass.

Hence, comparing the H
ref
values from Table 5.6 and the hydration numbers reported in the
literature, it seems that in the present work the H
ref
values might be interpreted as the number of
molecules of bound water per oxyethylene unit. In general, based on the figures presented
previously, it can be observed that the hydration numbers at 298.15 K (H
ref
) obtained in this work
are in agreement with part of the data reported in the literature. Therefore, the physical meaning
of eq 5.2 for calculating viscosities of aqueous PEG solutions becomes more consistent despite
Capítulo 5

118
some discrepancies, as described above. In any case, the data presented here are closely related
to the general conclusion that between one and four water molecules are bound per repeated
oxyethylene unit of the polymer, and that hydration numbers rise with the PEG molecular mass.
Considering that the values of B
ref
and H
ref
are functions of the polymer molecular masses (as
shown in Figures 5.1 and 5.2), such values were regressed to an exponential type equation as
given below:
) / exp(
3 w 2 1 ref
C M C C Y − + = (5.8)
where Y
ref
is B
ref
or H
ref
, M
w
is the molecular mass of the PEG molecule and C
i
's are adjusted
constants. Correlation coefficients of 0.9974 and 0.9977 were obtained in the regression for
estimating B
ref
and H
ref
values, respectively. The molecular masses used in this correlation were
the values given by the PEG denomination, such as PEG 1000 or PEG 3350. Although Table 5.1
gives the average molecular masses for the PEGs used in the present work, such information is
not available for the viscosity data taken from the literature.
Taking B
ref
and H
ref
as functions of PEG molecular mass, a generalized correlation could be
obtained using eq 5.2 for calculating the viscosities of the systems studied. Substituting eq 5.8 in
eqs 5.4 and 5.5, a generalized expression was obtained for the H and B parameters of eq 5.2. The
parameters of this generalized equation were estimated employing the entire set of experimental
data. Results of the regression are given in Tables 5.7 and 5.8 and Figures 5.3 and 5.4.

Table 5.7. Constants for the generalized equation
H
a
B
a
C
1
= 3.99096 C
1
= 83.9795
C
2
= -2.59861 C
2
= -80.2152
C
3
= 7891.5 C
3
= 22077.1
H
1
= 1.08 × 10
-3
B
1
= -0.1424
H
2
= -0.32 × 10
-3
B
2
= 6.32 × 10
-3

a
H
ref
and B
ref
are calculated by eq 5.8 using the parameters C
i
given above.
Results and Discussion

119
Table 5.8. Results of the generalized correlation for viscosity calculations
PEG AAD% PEG AAD%
200 16.7 3350 7.6
300 4.4 4000 2.9
400 5.6 6000 6.8
600 5.2 8000 5.0
900 2.8 10 000 7.0
1000 6.4 12 000 2.8
1500 6.6 15 000 3.1
2000 3.6 20 000 3.5
3000 3.1 35 000 3.5
global 5.5


0.00 0.05 0.10 0.15 0.20 0.25 0.30
0
5
10
15
20
ν

/

1
0
-
6

m
2

s
-
1
w

Figure 5.3. Experimental and calculated viscosities at 293.15 K and various concentrations: ■,
PEG 400; □, PEG 600; •, PEG 1000; ○, PEG 1500; ▲, PEG 3350; —, calculated
values.

Capítulo 5

120
290 295 300 305 310 315 320 325
0
5
10
15
20
25
30
35
40
45
50
ν

/

1
0
-
6

m
2

s
-
1
T / K

Figure 5.4. Experimental and calculated viscosities at various temperatures and concentrations:
▲, PEG 400, w=0.2487; ■, PEG 600, w=0.4899; □, PEG 1000, w=0.4777; •, PEG
1500, w=0.4534; ○, PEG 8000, w=0.2554; ∆, PEG 10 000, w=0.2441.

The AAD-value for the whole set of experimental data was 5.5% (Table 5.8). The systems
containing PEG 200 or 3350 are those with the higher deviations between experimental and
calculated viscosities. In case of PEG 3350 the main reason for such deviation is probably the
expressive difference between the nominal and the average molecular masses (Table 5.1), a
difference not so large for the other PEG denominations. In case of PEG 200 the relatively low
experimental viscosity values might have contributed for the large deviation observed, but it also
seems that such values do not fit so well in the suggested dependence of parameters B
ref
and H
ref

on the polymer molecular mass. Some authors suggested that for poly(ethylene glycol)s of low
molecular mass, such as PEG 200, only tightly bound water is associated to the polymer chain.
For polymers of larger molecular mass the chain begins to fold in itself, trapping additional water
between the segments of the polymer chain (Antonsen and Hoffman, 1992; Kirinčič and Klofutar,
1999). This difference observed in the polymer behavior according to its molecular mass can
eventually explain the larger deviation obtained for aqueous solutions of PEG 200.
It is also important to contrast the results of the correlation obtained in the present work with
those reported in the literature for the viscosity of aqueous PEG systems. González-Tello et al.
Results and Discussion

121
(1994) reported an AAD-value of 3.5% for dynamic viscosities (η) greater than 10 mPa·s, and
17.5% for viscosities lower than 10 mPa·s. In that work, the equation proposed to calculate
viscosities of aqueous PEG solutions takes into account the influences of poly(ethylene glycol)
concentration and temperature in the range (277 to 313) K. The same model equation was tested
by Günduz (1996) for PEG 8000 and dextran 580 000 in water at 293.15 K, with a resultant AAD
of 1.9% and 1.5%, respectively. It must be stressed that in this latter case the results were
obtained by adjustment of four parameters for each polymer separately. For a mixture of both
polymers (PEG + dextran) in water, a Grunberg equation was used for correlating the viscosities,
resulting in an AAD of 6.8% (Günduz, 1996). A subsequent work of the same author (Günduz,
2000) presented AAD-values up to 9.4% for aqueous mixtures of PEG + dextran of different
molecular masses. Polymer concentrations up to 7 mass % and temperatures between (303.15
and 343.15) K were considered.
Concluding the above remarks, it can be observed that the results achieved in the present
work are in good agreement with the AAD values usually presented in the literature. The
advantage of the equation presented here, in comparison with other equations, is that it has a
generalized form in relation to PEG molecular mass, concentration of polymer and temperature.
5.4.2 Viscosity prediction in multicomponent systems.
The predictive capacity of eq 5.3 for estimating viscosities of the ternary, quaternary and
quinary aqueous PEG systems was investigated. The values of viscosity calculated by eq 5.3
were compared with the experimental data from Table 5.4. An AAD of 9.9% was obtained. It
should be observed that the predicted values are in all cases lower than the experimental
viscosities. Probably there is an additional effect upon viscosity attributable to the relatively high
concentration of polymers with different molecular masses. This effect is not properly described
by the parameters adjusted to aqueous binary mixtures containing one polymer. In spite of this, it
should be noted that such AAD values obtained in this work are similar to those reported in
literature for viscosity prediction of multicomponent mixtures. Cao et al. (1993) reported
deviations in the range of (2.7 to 5.3)% for viscosity predictions by the GC-UNIMOD model.
They used a large data bank, including mostly binary mixtures of alkanes, alcohols, ketones,
ethers, and esters. Using the same model Rabelo et al. (2000) reported deviations in the range of
(0.8 and 14.0)% for some multicomponent fatty mixtures. It is also mentioned in the literature
predictions of the GC-UNIMOD model for mixtures of some polyethylene glycol dimethyl ethers
Capítulo 5

122
and methanol over a wide range of temperatures. The reported average absolute deviations were
around 20%. On the other hand, predictions with good accuracy were obtained by Pereira et al.
(2001) employing the modified Kumar equation for estimating viscosities of multicomponent
mixtures. However, only low molecular mass solutes (salts, glycerin, monosaccharides, etc) were
employed at relatively low solute concentrations (up to 15 mass %).
In general, the results obtained in this work, either in correlation or in prediction, were
satisfactory. Moreover, the obtained generalized equation based on the polymer molecular mass
is an useful tool since only few parameters are required for viscosity estimations in systems
containing PEGs with molecular masses between 200 and 35 000 g⋅mol
-1
.
5.5 Literature Cited
Albertson P. A. Partition of Cell Particles and Macromolecules; Wiley: New York, 1971.
Antonsen, K. P.; Hoffman, A.S. In Poly(ethylene glycol) Chemistry; Harris, J. M., Ed.; Plenum
Press: New York, 1992.
Bahri, H.; Guveli, D.E. Viscosity B Coefficients of Polyethylene Glycols in Water. Colloid
Polym. Sci. 1988, 266, 141-144.
Bieze, T.W.N.; Barnes, A.C.; Huige, C.J.M.; Enderby, J.E.; Leyte, J.C. Distribution of Water
around Poly(ethylene oxide): A Neutron Diffraction Study. J. Phys. Chem. 1994, 98, 6568-
6576.
Bisal, S.; Bhattacharya, P.K.; Moulik, S.P. Conductivity Study of Microemulsions: Evaluation of
Hydration of Oil/Water Microemulsions Applying Bruggeman Equation. J. Phys. Chem.
1990, 94, 4212-4216.
Bradoo, S.; Saxena, R. K.; Gupta, R. Partitioning and Resolution of Mixture of Two Lipases from
Bacillus stearothermophilus SB-1 in Aqueous Two-Phase System. Process Biochem. 1999,
35, 57-62.
Cao, W.; Knudsen, K.; Fredenslund, A.; Rasmussen, P. Group-Contribution Viscosity Predictions
of Liquid Mixtures Using UNIFAC-VLE Parameters. Ind. Eng. Chem. Res. 1993, 32, 2088-
2092.
Literature Cited

123
Coimbra, J. S. R.; Thömmes, J.; Meirelles, A. J. A.; Kula, M. R. Performance of Graesser
Contactor in the Continuous Extraction of Whey Proteins: Mixing, Mass Transfer and
Efficiency. Biosseparation 1995, 5, 259-268.
Cruz, M.S.; Chumpitaz, L.D.A.; Alves, J.G.L.F.; Meirelles, A.J.A. Kinematic Viscosities of
Poly(Ethylene Glycols). J. Chem. Eng. Data 2000, 45, 61-63.
Gongxi, Y.; Ziqiang, Z.; Yixin, G.; Zhaoxiong, H. Measurement of Mass Transfer Coefficients
and their Modelling of Continuous Countercurrent Aqueous Two-Phase System in a Packed
Extraction Column. Chin. J. Chem. Eng. 1999, 7, 196-204.
González-Tello, P.; Camacho, F.; Blázquez, G. Density and Viscosity of Concentrated Aqueous
Solutions of Polyethylene Glycol. J. Chem. Eng. Data 1994, 39, 611-614.
Grossmann, C.; Tintinger, R.; Zhu, J. D.; Maurer, G. Partitioning of Low Molecular Combination
Peptides in Aqueous Two-Phase Systems of Poly(Ethylene Glycol) and Dextran in the
Presence of Small Amounts of K
2
HPO
4
/KH
2
PO
4
Buffer at 293 K: Experimental Results and
Predictions. Biotechnol. Bioeng. 1998, 60, 699-711.
Gündüz, U. Evaluation of Viscosities of Polymer-Water Solutions used in Aqueous Two-Phase
Systems. J. Chromatogr. B 1996, 680, 263-266.
Günduz, U. Optimization of Bovine Serum Albumin Partition Coefficient in Aqueous Two-Phase
Systems. Bioseparation 2000, 9 , 277-281.
Gündüz, U. Viscosity Prediction of Polyethylene Glycol-Dextran-Water Solutions Used in
Aqueous Two-Phase Systems. J. Chromatogr. B 2000, 743, 181-185.
Harris, J. M. In Poly(ethylene glycol) Chemistry; Harris, J. M., Ed.; Plenum Press: New York,
1992.
Kirinčič, S.; Klofutar, C. A Volumetric Study of Aqueous Solutions of Poly(etheylene glycol)s at
298.15 K. Fluid Phase Equilib. 1998, 149, 233-247.
Kirinčič, S.; Klofutar, C. Viscosity of Aqueous Solutions of Poly(Ethylene Glycol)s at 298.15 K.
Fluid Phase Equilib. 1999, 155, 311-325.
Capítulo 5

124
Kumar A. A Simple Correlation for Estimating Viscosities of Solutions of Salts in Aqueous,
Non-aqueous and Mixed Solvents Applicable to High Concentration, Temperature and
Pressure. Can. J. Chem. Eng. 1993, 71, 948-954.
Marquardt, D.W. An Algorithm for Least-Square Estimation of Nonlinear Parameters. J. Soc.
Indust. Appl. Math. 1963, 11, 431-441.
Mehrotra, A. K.; Monnery, W. D.; Svrcek, W. Y. A Review of Practical Calculation Methods for
the Viscosity of Liquid Hydrocarbons and their Mixtures. Fluid Phase Equilib. 1996, 117,
344-355.
Mei, L.-H.; Lin, D.-Q.; Zhu, Z.-Q.; Han, Z.-X. Densities and Viscosities of Polyethylene Glycol
+ Salt + Water Systems at 20°C. J. Chem. Eng. Data 1995, 40, 1168-1171.
Moulik, S. P.; Gupta, S. Hydration Studies on Some Polyhydroxy Non-Electrolytes and Non-
Ionic Surfactants. Can. J. Chem. 1989, 67, 356-363.
Ninni, L.; Meirelles, A.J.A. Water Activity, pH and Density of Aqueous Amino Acids Solutions.
Biotechnol. Progr. 2001, 17, 703-711.
Pereira, G.; Moreira, R.; Vázquez, M. J.; Chenlo, F. Kinematic Viscosity Prediction for Aqueous
Solutions with Various Solutes. Chem. Eng. J. 2001, 81, 35-40.
Persson, J.; Johansson, H-O; Tjerneld, F. Purification of Protein and Recycling of Polymers in a
New Aqueous Two-Phase System Using Two Thermoseparating Polymers. J. Chromatogr.
A 1999, 864, 31-48.
Porto, A. L. F.; Sarubbo, L. A.; Lima, J. L.; Aires-Barros, M. R.; Cabral, J. M. S.; Tambourgi, E.
B. Hydrodynamics and Mass Transfer in Aqueous Two-Phase Protein Extraction Using a
Continuous Perforated Rotating Disc Contactor. Bioprocess Eng. 2000, 22, 215-218.
Powell, G. M. In Handbook of Water Soluble Gums and Resins; Davidson, R. L., Ed.; McGraw-
Hill: New York, 1980.
Rabelo, J.; Batista, E.; Cavaleri, F.W.; Meirelles, A.J.A. Viscosity Prediction for Fatty Systems.
J. Am. Oil Chem. Soc. 2000, 77, 1255-1261.
Literature Cited

125
Silva, L. H. M.; Meirelles, A. J. A. PEG + Potassium Phosphate + Urea Aqueous Two-Phase
Systems: Phase Equilibrium and Protein Partitioning. J. Chem. Eng. Data 2001, 46, 251-
255.
Zaslavsky, B. Y. Aqueous Two-Phase Partitioning; Marcell Dekker: New York, 1995.
The authors gratefully acknowledge the financial support of FAPESP (Fundação de Amparo
à Pesquisa do Estado de São Paulo: 01/10137-6) and CNPq (Conselho Nacional de
Desenvolvimento Científico e Tecnológico - 521011/95-7, 466680/00-7, 145941/99).

Capítulo 6 127

Capítulo 6
Thermodynamic properties of maltodextrin
aqueous solutions
Luciana Ninni
1
, Antonio J. A. Meirelles
1
, Gerd Maurer
2
1
Faculdade de Engenharia de Alimentos – Universidade Estadual de Campinas (UNICAMP)
C.P. 6121 CEP 13083-970 Campinas – SP, Brazil
2
Lehrstuhl für Technische Thermodynamic, Universität Kaiserlautern, D-67653
Kaiserslautern, Germany.










Trabalho a ser submetido na revista Carbohydrate Polymers.
Capítulo 6 128

Introduction 129
6. Thermodynamic properties of maltodextrin aqueous solutions
6.1 Abstract
The thermodynamic properties of maltodextrins (MD) in aqueous solutions were studied by
isopiestic method, laser-light scattering and calorimetry. Three different MD samples with
nominal molar masses between 1000 and 3000 g mol
-1
were investigated at various
concentrations. The experimental results of water activity and light scattering were described by
the osmotic virial equation. It was found in this procedure that the second and third virial
coefficients depend on the molar mass of the different compounds. The VERS model was also
used to correlate experimental water activity and calorimetric data for the binary mixtures and to
predict water activity of some low molar mass compounds and other thermodynamic properties
reported in the literature. Mean relative deviations of 0.40% were obtained in the correlations of
water activities of maltodextrin solutions when the maltodextrin molecule was treated as one
component, and 0.93% when it was approximated by four components.
6.2 Introduction
Water-soluble polymers have been used in a wide range of industrial products in the food and
pharmaceutical areas. This group of compounds comprises molecules from natural (renewable
resource raw material) to synthetic (from petrochemical base feed stocks) origin. At the present
time, the need for environmental protection favors the use of biodegradable polymers, such as
polysaccharides, for industrial purposes (Swift, 1998). Polysaccharides occur with a molar mass
distribution, usually represented by the average molar masses: number-average (M
n
) or weight
average (M
w
). Among the polysaccharides, maltodextrins are polymers of great interest due to
their high solubility in water (Gliksmann, 1986; Marchal, 1999) and potential application in
aqueous -two - phase systems (Silva and Meirelles, 2000a,b). They are starch hydrolysates
consisting of α-D-glucose units bounded by (1→4) glycosidic linkages (primarily) as well as by
(1→6) linkages. These polymers exhibit broad molar mass distributions (with number-average
molar masses varying between ~10
3
and 10
5
gmol
-1
) due to the process of starch hydrolysis that is
carried out with enzymes and/or acids at elevated temperatures. Alpha-amylase, commonly
employed in starch hydrolysis is capable of hydrolyze α-1→4 linkages but has little effect on α-
1→6 linkages. Moreover, α-amylase does not readily hydrolyze α-1→4 linkages in maltose and
maltotriose. Thus, maltose, maltotriose and other low molar mass saccharides can be found in
Capítulo 6 130
the final hydrolyzates. These characteristics confer them a wide range in solubility, viscosity,
and chemical stability (Kasapis et al., 1993; Mothé and Rao, 1999).
Thermodynamic properties of polymeric aqueous solutions have been reported in literature
(Hasse et al., 1995; Kany et al., 1999; Groβmann et al., 1995; Gaube et al., 1993, Silva and Loh,
2000). Thermodynamic data determinations are important under the industry and academic
research viewpoints but only few data sets for aqueous maltodextrin solutions can be found in the
literature. Among some of the works available, it can be cited Cesàro et al. (1999), which studied
thermodynamic properties of two biopolymers (maltodextrin and gelatin) by using calorimetry.
Heats of dilution of the single polymers and of the mixed polymer solutions were used to
evaluate the Flory interaction parameters (Flory, 1953). Radosta et al. (1989 a, b, c) presented a
comprehensive study about physical-chemical properties of maltodextrins in aqueous solutions.
In that work a gelling maltodextrin was experimentally fractionated in 6 different fractions with
different average molar masses. It was then investigated the sorption and desorption isotherms
and the amount of non freezable water in the fractions. Many works about maltodextrins in the
literature deal with experimental investigations of gelling maltodextrins. The gelation of low DE
maltodextrins and phase equilibria in gelatin/maltodextrin systems were investigated by Kasapis
et al. (1993a, b, c, d). Schierbaum et al. (1992) also presented a study of the gelation process of
maltodextrin-water systems using low resolution H-NMR. They observed the time required for
the beginning of gelation as a function of temperature and concentration, and the influence of the
addition of amylose on the gelation process.
This paper reports experimental results for aqueous maltodextrin systems involving three
different techniques, isopiestic, laser-light-scattering and calorimetric methods, for determining
thermodynamic properties. Moreover, the osmotic virial equation is used to describe the
experimental results from isopiestic and laser-light-scattering measurements and a semi-empirical
group contribution model, the VERS equation, is going to be tested in correlating and predicting
the investigated thermodynamical properties for water activity and heats of dilution in
maltodextrin and sugar systems.
Experimental 131
6.3 Experimental
6.3.1 Materials
Aqueous solutions of three different maltodextrins purchased from Aldrich Chemical and
designated here according to their dextrose equivalent range as MD 4 – 7, MD 13 – 17 and MD
16 – 19 were studied in the present work. The polymers were used as received and the water
content of each sample was determined by Karl Fischer titration using a Methrom device. The
results of titration for the maltodextrins MD 4 – 7, MD 13 – 17 and MD 16 – 19 are, respectively:
3.31±0.20; 4.42±0.22 and 3.31±0.18 mass % of water.
6.3.2 Methods
6.3.2.1 Molar mass distributions
The maltodextrins were characterized by gel permeation chromatography (GPC) with an
aqueous eluent containing 0.05 M sodium nitrate. A constant flow rate of 1 mL⋅min
-1
was
maintained by a HPLC pump (Spectra Physics, type P 1000). The eluent was degassed in-line
(ERMA, type ERC-3512). About 100 µL of the maltodextrin solution was injected in the eluent
using a Reodyne valve. A column type MCX 1000 Å (PSS Polymer Systems) was utilized for
separation. Detectors were a multiangle laser-light scattering (MALLS, Wyatt Technology, type
DAWN DSP) and a interferometric refractometer (Wyatt Technology, type Optilab 903). Details
on the calibration of the detectors are described by Hasse et. al. (1995).
The results for the number-averaged molar mass (M
n
), weight-averaged molar mass (M
w
) and
polydispersity ratio (M
w
/M
n
) determined in these GPC experiments are presented in Table 6.1 and
the gel permeation profiles of the maltodextrins are shown in Figure 6.1. For each maltodextrin
sample, the determination of molar mass distributions were performed at least three times. Mean
relative deviations between different determinations varied in the range of 3 and 13%.
Tabela 6.1. Experimental data on the number- and mass-average molar mass of the maltodextrins
studied in this work.
M
n
M
w
M
w
/M
n

MD 4 – 7 2683 38360 14.31
MD 13 – 17 1475 13710 8.38
MD 16 – 19 1140 8283 7.31

Capítulo 6 132
As can be seen in Table 6.1 and in Figure 6.1, MD 4 – 7 comprises more high molar mass
saccharides while the maltodextrins MD 13 – 17 and 16 – 19 have higher content of low molar
mass saccharides. This indicates that the MD 4 –7 was less converted to low saccharides in the
process of starch hydrolysis. The presence of more higher molar mass components makes
MD 4 –7 capable of forming thermally reversible gels in aqueous solutions. Further details on
the properties of gelling maltodextrins will appear in the following items.
5 6 7 8 9 10 11
0.0
5.0x10
-5
1.0x10
-4
1.5x10
-4
2.0x10
-4
2.5x10
-4
3.0x10
-4
3.5x10
-4
4.0x10
-4
MD 4-7
MD 13-17
MD 16-19
C
a
r
b
o
h
y
d
r
a
t
e

c
o
n
c
e
n
t
r
a
t
i
o
n

(
g

m
L
-
1
)
Volume (mL)

Figure 6.1. Gel permeation profiles of maltodextrins.
After the separation in the GPC columns, the light scattering from each eluted fraction of
very low concentration is detected simultaneously at 16 angles between 28.7 and 143.8°. Figure
6.2 presents the experimental data of molar masses vs. elution volumes calculated by the software
ASTRA, from the extrapolation of the light scattering to zero angle at each slice, according to the
following relation:
c A
MP R
Kc
2
2
) (
1
+
θ
=
θ
(6.1)
with
Experimental 133
4
0
0
2 2
) d (d 4
λ
π
=
A
N
n c / n
K
2
2
2 2
0
2
3
) 2 ( 16
1
) (
1
g
R
/ sin n
P








λ
θ π
+ =
θ
(6.2)
where K is an optical constant, c the concentration of the polymer solution, R(θ) the excess
Rayleigh ratio at angle θ , M is the weight-average molar mass, P(θ) a form factor describing the
angular and size dependence of the scattered light intensity,
2
g
R the mean-square radius of
gyration, A
2
the second virial coefficient, n
0
the refractive index of the solvent, dn/dc the
differential refractive index increment of the polymer in solution (0.1486, 0.1510, 0.1522 cm
3
g
-1
,
for MD 4 – 7, 13 –17 and 16 – 19, respectively), λ is the laser wave length (633 nm), N
A
the
Avogrado number, and λ
0
the wavelength of the incident light under vacuum (Huglin, 1972).
When c→0 and θ→0, equation 6.1 becomes:
M R
Kc
, c
1
0 0
=
→ θ →
θ
(6.3)
In Figure 6.2, it is observed a linear dependence of the molar masses on the elution volumes.
However, it can be also noted that the scattering of the data is greater in the region of low molar
mass molecules. The reason is that the light scattering detector of the MALLS device is less
sensitive at low molar masses. The correct slope in the obtained curves is important to calculate
molar mass distributions of the polymers as shown below:
) (
) (
) (
) (
) (
) (
) (
V f
V h
M log d
dV
dV
M dW
M log d
M dW
M x − = = =
(6.4)
where x(M) is the differential molar mass distribution also called dξ/d(logM); W(M) is the
cumulative molar mass distribution; f(V)=d(logM)/dV the slope of the curve molar mass (in a
logarithmic scale) vs. volume; h(V) is the normalized concentration defined as: c
i
/


pico
i
V c .

Capítulo 6 134

1
1.0x10
2
1.0x10
3
1.0x10
4
1.0x10
5
1.0x10
6
1.0x10
7
1.0x10
4.0 6.0 8.0 10.0 12.0
M
o
l
a
r

M
a
s
s

(
g
/
m
o
l
)
Volume (mL)
Molar Mass vs. Volume
MD 13 - 17
MD 16 - 19
MD 4 - 7

Figure 6.2. Results of the experiment GPC/MALLS for the maltodextrins studied in this work.

In order to verify whether the slope of the adjusted curve was correct, the GPC results for the
molar mass distribution of each maltodextrin were approximated by the sum of Gauss-Lorentz
functions. The purpose of using this method is to separate the whole distribution in more
monodisperse fractions and also to evaluate if the peaks that elute later in the chromatographic
runs correspond to the lower molar mass compounds: glucose, maltose and maltotriose. The
Gauss-Lorentz (G-L) function is given by:
( ) ( )
2 2
2 2
) ( 1
i i
i i
i
i
c M d exp
c M b
a
GL − −
− +
=
(6.5)
For each maltodextrin, the parameters a
i
, b
i
, c
i
and d
i
(i=1, 2, ..., N; N is the number of G-L
i

functions) were fitted to the molar mass distribution. Figure 6.3 presents the adjusted G-L
i

functions for MD 4 – 7. Now, for each fraction represented by a G-L
i
function, it is known the
correspondent elution volumes. Plotting a diagram of molar mass versus volume with data points
obtained by laser-light scattering experiments up to volume of about 8 mL and assuming that the
adjusted volumes by G-L
i
functions for the first three fractions correspond to glucose, maltose
Experimental 135
and maltotriose, it is possible to verify if the extrapolation of the curve logarithmic of the molar
mass vs. volume obtained for the higher molar mass molecules, as shown in Figure 6.2, is
consistent and consequently if the slope of the straight line is correct.
5 6 7 8 9 10 11
0.0
5.0x10
-5
1.0x10
-4
1.5x10
-4
2.0x10
-4
2.5x10
-4


C
o
n
c
e
n
t
r
a
t
i
o
n

(
g

m
L
-
1
)
Volume (mL)

Figure 6.3. Results of Gauss-Lorentz functions for the molar mass distribution of MD 4 – 7.

This extrapolation can be seen in Figure 6.4 indicating that the previous assumption was
correct and that the adjusted elution volumes (volumes on the maximal peak point) really
correspond to glucose, maltose and maltotriose. The last three points in Figure 6.4 were plotted
using the adjusted elution volumes and the respective molar masses for glucose, maltose and
maltotriose, that are 180, 342 an 504 g mol
-1
, respectively. Concentrations of the lower molar
mass fractions estimated by the Gauss-Lorentz method were assumed to be glucose, maltose and
maltotriose, and they can also be compared to values presented in literature. Values of 1.8 , 4.3
and 8.1 mass % for glucose, maltose and maltotriose, respectively are reported in literature
(Defloor, 1998) for a maltodextrin with approximately 1300 g mol
-1
, and 2.6 , 5.9 and 7.9 were
found in this work for MD 16 – 19, which has a molar mass of 1140 g mol
-1
. Therefore, it can be
considered that the results from GPC/MALLS analysis are reliable and they can be used to
estimate molar mass distribution of maltodextrins.
Capítulo 6 136
5 6 7 8 9 10 11
100
1000
10000
100000
1000000
MD16-19
Regression
M
o
l
a
r

m
a
s
s

(
g

m
o
l
-
1
)
Volume (mL)

Figure 6.4. Molar mass as a function of elution volume, using the MALLS data and elution
volumes for glucose, maltose and maltotriose from the Gauss-Lorenz functions.

6.3.2.2 Isopiestic Method
The water activity in aqueous solutions of maltodextrins was measured with the isopiestic
method at 298.15 K for all three maltodextrins, and at 318.15 K for MD 13 –17 and 16 –19. The
experimental arrangement consisted of a glass multilegged manifold with nine standard taper
female ground-glass joints for the attachment of solution flasks and a high vacuum stopcock for
isolation of the system after evacuation. In order to increase the mass transfer area and enhance
sample mixing during the equilibration period, the glass was allowed to rotate around an axis
inclined at 45°. The system was immersed in a constant temperature water bath during the
equilibration times. The temperature of the thermostat fluctuated by less than ±0.1 K. In a
experiment, the flasks were filled with about a gram of an aqueous solution of non volatile
substance. The nine flasks of the experimental arrangement were used as follows: three flasks
were filled with standard NaCl solutions in a concentration named here as NaCl(1), three flasks
also with NaCl solutions but in other concentration, NaCl(2), and three for maltodextrin
solutions. The two NaCl concentrations (NaCl(1) and NaCl(2)) were chosen in such a way that
would be expected these concentrations were above and below the final equilibrium
concentration. All the solutions were prepared with bidistilled water. In the glass, mass transfer
Experimental 137
of water between the sample and the reference solutions took place through the vapor phase, until
the activity of water in the sample was the same as that in the reference solutions. The water
activity of the samples was estimated from gravimetric analysis of the sample and reference
solutions before and after the experiment. The water activity in the aqueous sodium chloride
solutions were calculated using the correlation of Pitzer and Peiper (1984). The experimental
water activity results for the three maltodextrins at 298.15 K are given in Tables 6.2 – 6.4
together with standard deviations (SD) in concentration of non volatile substance and in water
activity. The w
NaCl
values in the following Tables correspond to the average concentration
between NaCl solutions (NaCl(1) and NaCl(2)) at equilibrium. These equilibrium concentrations
were always the same within the experimental error. The experimental data points at 318.15 K
presented in Table 6.5 were measured also by isopiestic method but with a different experimental
arrangement. It consists of a stainless steel apparatus that can hold up to 30 glass cells placed in
a copper block that is mounted in a thermostated chamber. In a experiment, the glass cells are
filled with about 1 mL of aqueous solution, the chamber is sealed, evacuated and thermostated.
The process of mass transfer and the evaluation of water activities are the same as those
described above.
The apparatus and experimental procedures used for the isopiestic experiments in this work
were described before by Großmann et al. (1995), Kany (1998), and Lammertz and Maurer
(2002).
Table 6.2. Experimental results for the water activities in aqueous solutions of MD 13 – 17 at
298.15 K
w
MD
/ g g
-1
SD/ g of MD a
w
SD/ a
w
units w
NaCl
/g g
-1
SD/ g of NaCl
0.1750 3.971e-3 0.9962 1.500e-3 0.0073 2.630e-3
0.2234 8.740e-4 0.9948 1.04e-4 0.0091 1.822e-4
0.2419 4.190e-4 0.9942 8.792e-5 0.0102 1.539e-4
0.3029 4.665e-4 0.9917 1.011e-4 0.0145 1.756e-4
0.3408 1.036e-3 0.9898 1.028e-4 0.0178 1.773e-4
0.3796 8.680e-4 0.9875 3.462e-4 0.0217 5.910e-4
0.4286 6.087e-4 0.9838 7.795e-5 0.0280 1.308e-4
0.4408 5.525e-4 0.9829 1.659e-4 0.0294 2.771e-4
0.4872 4.817e-4 0.9779 7.274e-5 0.0377 1.185e-4
0.4867 5.135e-4 0.9781 8.069e-5 0.0373 1.316e-4
0.5307 1.593e-3 0.9724 7.494e-5 0.0465 1.186e-4
0.5671 1.427e-3 0.9664 1.777e-4 0.0559 2.724e-4
0.5898 2.494e-4 0.9622 1.293e-4 0.0622 1.937e-4

Capítulo 6 138
Table 6.3. Experimental results for the water activities in aqueous solutions of MD 16 – 19 at
298.15 K
w
MD
/ g g
-1
SD/ g of MD a
w
SD/ a
w
units w
NaCl
/g g
-1
SD/ g of NaCl
0.2598 2.386e-4 0.9928 1.050e-4 0.0086 1.827e-4
0.2743 9.011e-4 0.9922 1.116e-4 0.0134 1.943e-4
0.3434 5.117e-4 0.9886 7.555e-5 0.0198 1.296e-4
0.3887 1.541e-3 0.9854 1.060e-4 0.0254 1.797e-4
0.3844 1.310e-4 0.9860 1.497e-4 0.0243 2.538e-4
0.4197 7.735e-4 0.9831 4.272e-5 0.0292 7.143e-5
0.4492 1.565e-3 0.9802 5.450e-4 0.0339 8.977e-4
0.4559 1.183e-4 0.9797 4.126e-4 0.0348 6.782e-4
0.4824 7.963e-4 0.9763 2.426e-4 0.0403 3.920e-4
0.5357 6.175e-4 0.9689 1.228e-4 0.0521 1.909e-4
0.5850 1.036e-3 0.9594 2.296e-4 0.0664 3.390e-4

Table 6.4. Experimental results for the water activities in aqueous solutions of MD 4 – 7 at
298.15 K
w
MD
/g g
-1
SD/ g of MD a
w
SD/ a
w
units w
NaCl
/g g
-1
SD/ g of NaCl
0.2859 1.546e-3 0.9967 4.286e-4 0.0048 7.518e-4
0.2644 1.026e-3 0.9970 2.842e-4 0.0052 4.984e-4
0.2914 1.310e-3 0.9966 1.028e-4 0.0059 1.803e-3
0.3167 9.982e-4 0.9960 1.620e-4 0.0069 2.843e-4
0.3365 1.822e-3 0.9953 1.789e-4 0.0081 3.138e-4
0.3399 2.520e-4 0.9957 2.031e-4 0.0075 3.563e-4
0.3379 9.311e-4 0.9957 2.306e-4 0.0075 4.046e-4
0.3488 1.992e-3 0.9956 1.877e-3 0.0077 3.291e-3
0.3802 2.348e-3 0.9945 9.010e-4 0.0099 1.496e-3
0.4099 1.111e-3 0.9939 1.906e-4 0.0106 3.333e-4
0.4590 1.367e-3 0.9919 2.373e-3 0.0140 4.125e-3
0.4615 1.097e-3 0.9922 1.393e-3 0.0136 2.424e-3

Experimental 139
Table 6.5. Experimental results for the water activities in aqueous solutions of MD 13 – 17 and
MD 16 –19 at 318.15 K.
w
MD
/g g
-1
SD/ g of MD a
w
SD/ a
w
units w
NaCl
/g g
-1
SD/ g of NaCl
MD 13 – 17
0.2767 7.206e-4 0.9927 1.390e-4 0.0128 2.421e-4
0.3284 1.498e-3 0.9903 1.148e-4 0.0169 1.984e-4
0.4452 2.126e-3 0.9821 2.200e-4 0.0308 3.681e-4
0.4968 1.820e-4 0.9767 7.011e-5 0.0396 1.135e-4
MD 16 – 19
0.2683 2.941e-4 0.9927 1.390e-4 0.0128 2.421e-4
0.3156 2.827e-3 0.9903 1.148e-4 0.0169 1.984e-4
0.4327 1.895e-3 0.9821 2.210e-4 0.0308 3.679e-4
0.4852 4.570e-5 0.9767 7.023e-5 0.0369 1.137e-4

6.3.2.3 Calorimetric investigations
Aqueous solutions of single maltodextrins were mixed with pure water at 313.15K in a batch
calorimeter (model MS 80, Setaram, Lyon, France). Specially designed cells were used to enable
isothermal and isobaric mixing. In the upper compartment a small cylinder with a piston which
separates the solution in the cell from compressed air allowed pressurization at 2 bar. The cells
were filled with syringes in such a way that no gas bubbles were present. The cell is separated in
two parts (containing the polymer solution in the upper compartment and the pure water in the
lower compartment) by a Teflon seal. To start the dilution process, the seal was cut off and the
heat upon dilution was determined from the integrated signal of the thermocouple of the
calorimeter. To enhance mixing due to high viscosity of the maltodextrin solutions, metal balls
and a rotation of the entire calorimeter was performed. To check the accuracy of the
experimental data, an electrical heater was immersed in the experimental cell that was filled only
with water. A constant current was sent through the resistor by a current source. The heat
evolved was estimated from voltage, resistance and the time during which the current was
applied. These values were compared with those from the integrated signal of the thermocouple
of the calorimeter. Relative mean deviations of 1% were encountered between these two values.

Capítulo 6 140
The experimental results are presented in Tables 6.6 and 6.7 for the MD 13 –17 and 16 – 19.
Superscripts bottom and top designate the solutions placed in the lower and upper compartment
of the cells. Note that the measured heats of dilution are negative, corresponding to an
exothermic behaviour upon dilution and that the heat involved in the mixing process is small.
Dextran, another polysaccharide, presents the same exothermic behaviour as reported by
Großmann et al. (1995), who in determined heats of dilution, and Silva and Loh (2000) for heats
of solution. It must be stressed that for each experiment, the heat dissipated on cutting off the
Teflon seal was determined after the dilution process by once again pushing and pulling the
knife. The heat evolved in this procedure was lower than 1 J and its values were taken into
account in the experimental values reported in Tables 6.6 and 6.7.
For the maltodextrin MD 4 –7 it was observed two different effects upon dilution: an
endothermic in the beginning of the experiment, and another one exothermic about 90 minutes
after the cutting off the Teflon seal. In this case, it can be supposed that in the beginning an
endotermic effect takes place due to dilution of the polymer solution, that is a mixture of gel and
solution, in pure water, and the second, due to the rearrangement of the molecules to build a new
gel network. Some authors studied the formation of thermally reversible potato starch
maltodextrin gels by H-NMR, viscometry and electronic microscopy (Schierbaum et al., 1984,
1992). The main characteristic of these maltodextrins in aqueous solutions is that the progress of
the gelling process is time-dependent and the transition takes place from a homogeneous
structureless solution to a two-phase system. The last one consists of the solid aggregation
network and the liquid solution of non-structured MD low molar mass components. A rapid
increase in the gelling process is encountered at high concentrations and low temperatures. Due
to difficulties in analysing the experimental data, because of the errors on pushing the pushing
knife of the cells and the mixing itself involving a gel system, it was decided that these systems
are deserving of much experimental work that are not to be considered in the present paper.
Therefore, the experimental heats of dilution are presented only for the other two maltodextrins
that do not form gel. The capacity of gel formation is enabled by the presence of enough long
chains in solution, as in the case of MD 4 – 7, which has a significant high molar mass fraction as
pointed out in Figure 6.1. According to Schierbaum et al. (1984), products with dextrose
equivalent (DE) up to 5 – 8 are able to form thermally reversible gels at concentration above 10
mass %.
Experimental 141
Table 6.6. Experimental heats of dilution of maltodextrin 13 – 17 at 313.15 K
a

m
w
top
/g m
w
bottom
/g w
MD
top
/g g
-1
Q/J
20.480 15.938 0.290 -1.193
19.416 15.857 0.290 -1.234
20.161 15.902 0.300 -1.220
17.911 15.916 0.398 -1.682
18.230 15.855 0.404 -2.051
17.970 15.905 0.398 -1.825
19.618 14.805 0.508 -3.157
20.492 14.976 0.481 -2.805
18.277 14.783 0.548 -3.640
18.335 14.822 0.548 -3.662
17.405 14.979 0.579 -4.490
17.479 15.000 0.579 -4.492
14.686 14.866 0.638 -5.887
14.592 14.772 0.638 -5.464
12.904 14.854 0.694 -6.686
13.173 14.855 0.694 -6.773
a
w
MD
bottom
/g g
-1
was zero for all experiments.

Table 6.7. Experimental heats of dilution of maltodextrin 16 – 19 at 313.15 K
a

m
w
top
/g m
w
bottom
/g w
MD
top
/g Q/J
22.939 15.014 0.404 -3.502
22.867 14.953 0.393 -2.616
19.857 14.779 0.501 -4.487
18.863 15.472 0.511 -3.899
18.259 15.057 0.548 -6.432
18.406 15.030 0.543 -6.026
16.432 15.038 0.608 -8.938
16.195 15.039 0.608 -8.478
14.924 14.871 0.649 -10.340
15.312 14.953 0.637 -9.610
12.543 14.801 0.700 -12.210
12.818 15.021 0.700 -12.189
a
w
MD
bottom
/g g
-1
was zero for all experiments.

Capítulo 6 142
6.3.2.4 Densimetric data
In order to convert polymer mass fractions (w in g g
-1
) into concentrations (c in g cm
-3
), the
specific densities (ρ* in g cm
-3
) of the maltodextrin solutions were determined at 298.15 K.
Density measurements were carried out in triplicate using a digital densimeter (DMA 58, Anton
Paar) which was calibrated with water and air as standards at the corresponding working
temperature. The accuracy of the density measurements was estimated as 3×10
-5
g cm
-3
. The
experimental results for MD 13 – 17 and MD 16 – 19 are shown in Table 6.8 below.
Table 6.8. Specific density of aqueous MD 13 –17 and MD 16 – 19 solutions at 298.15 K
MD 13 –17 MD 16 –19
w
MD
/g g
-1
ρ*/ g cm
-3
w
MD
/g g
-1
ρ*/ g cm
-3

0.0499 1.01573 0.0500 1.01591
0.0999 1.03516 0.0995 1.03525
0.1487 1.05472 0.1416 1.05224
0.2004 1.07628 0.1988 1.07614
0.2497 1.09761 0.2506 1.09869
0.2996 1.11995 0.2878 1.11537
0.3512 1.14387 0.3489 1.14380
0.4025 1.16867 0.3980 1.16776
0.4992 1.21786 0.4507 1.19390
0.5491 1.24457 0.4938 1.21650
0.5392 1.24056
0.5941 1.27196

No significant difference between the two sets of experimental data of specific density was
observed for both maltodextrins at the studied temperature.
6.3.2.5 Laser-light scattering experiments
The multiangle laser-light scattering device utilized for the determinations of the light
scattering from maltodextrin solutions was the same as that in the GPC experiments. The
difference is that multiple concentrations of the same non-fractionated polymer sample are
pumped in the optical cell to find how the scattering intensity varies with concentration. A flow
rate of 0.5 cm
3
⋅min
-1
was employed using a syringe pump (Infors, type Predicor). Filters
(Sartorius) with pore size 0.20 µm were used in the line between the syringe and the flow cell to
Experimental 143
remove dust from the aqueous solutions and avoid great noises in the measurements. The
experiments were performed at 298.15 ±0.2 K and at least 6 different concentrations were
pumped into the optical cell. The results of the laser-light scattering experiments are presented in
Tables 6.9 and 6.10 for the maltodextrins MD 13 –17 and 16 –19, respectively.
Table 6.9. Excess Rayleigh ratios of MD 13 – 17 in water at 298.15 K
100 c
s
/g cm
-3
0.720 1.438 2.183 3.001 3.617 4.341 5.076 5.805
θ/
o

8.117 8.444 8.847 9.268 9.605 10.030 10.340 10.700
28.71
7.909 8.227 8.882 9.177 9.547 9.951 10.160 10.530
8.155 8.387 8.860 9.212 9.600 9.959 10.320 10.690
36.2
7.903 8.252 8.878 9.093 9.503 9.918 10.250 10.670
8.088 8.375 8.846 9.234 9.597 10.010 10.340 10.710
44.5
7.99 8.294 8.858 9.153 9.529 9.926 10.240 10.590
8.038 8.313 8.759 9.141 9.506 9.863 10.230 10.590
54.0
7.976 8.296 8.777 9.055 9.430 9.835 10.180 10.550
8.135 8.436 8.890 9.278 9.632 10.020 10.380 10.750
64.9
8.137 8.429 8.909 9.230 9.578 9.983 10.330 10.680
8.077 8.354 8.785 9.158 9.511 9.869 10.240 10.610
77.1
8.092 8.396 8.817 9.108 9.464 9.865 10.220 10.610
8.313 8.591 9.021 9.416 9.742 10.130 10.500 10.870
90.0
8.355 8.629 9.048 9.381 9.725 10.120 10.480 10.840
8.242 8.520 8.922 9.294 9.618 9.991 10.360 10.740
115.1
8.304 8.577 8.951 9.289 9.615 10.000 10.360 10.730
8.170 8.445 8.842 9.208 9.535 9.883 10.270 10.640
125.4
8.234 8.534 8.890 9.209 9.537 9.924 10.290 10.680
8.220 8.509 8.909 9.281 9.593 9.968 10.330 10.720
135.5
8.336 8.611 8.969 9.298 9.626 10.020 10.360 10.740
8.068 8.388 8.793 9.150 9.467 9.826 10.210 10.600
1
0
5
K

c
s
(
R
E
θ
)
-
1

m
o
l

g
-
1

143.8
8.202 8.512 8.896 9.199 9.544 9.939 10.250 10.670

Capítulo 6 144
Table 6.10. Excess Rayleigh ratios of MD 16 – 19 in water at 298.15 K
100 c
s
/g cm
-3
1.408 2.111 2.823 3.530 4.230 4.919 5.644
θ/
o

1.051 1.073 1.124 1.188 1.243 1.295 1.379
36.2
1.043 1.107 1.199 1.214 1.285 1.344 1.382
1.061 1.087 1.133 1.191 1.245 1.312 1.359
44.5
1.068 1.102 1.196 1.222 1.279 1.329 1.354
1.058 1.083 1.128 1.198 1.244 1.298 1.374
54.0
1.079 1.116 1.200 1.224 1.283 1.342 1.375
1.083 1.111 1.153 1.217 1.265 1.332 1.382
64.9
1.116 1.124 1.210 1.244 1.295 1.345 1.375
1.079 1.107 1.151 1.217 1.262 1.319 1.387
77.1
1.115 1.135 1.214 1.244 1.297 1.352 1.388
1.118 1.149 1.189 1.253 1.298 1.361 1.413
90.0
1.159 1.162 1.243 1.280 1.329 1.379 1.411
1.087 1.117 1.158 1.223 1.265 1.321 1.387
102.9
1.131 1.138 1.216 1.251 1.298 1.352 1.389
1.118 1.148 1.185 1.249 1.292 1.354 1.403
115.1
1.156 1.155 1.233 1.274 1.317 1.367 1.400
1.112 1.14 1.179 1.246 1.288 1.341 1.405
125.4
1.151 1.158 1.233 1.269 1.316 1.368 1.407
1.121 1.152 1.188 1.250 1.294 1.354 1.400
135.5
1.160 1.152 1.234 1.273 1.318 1.366 1.400
1.108 1.138 1.176 1.242 1.285 1.337 1.401
1
0
5
K

c
s
(
R
E
θ
)
-
1

m
o
l

g
-
1

143.8
1.149 1.148 1.230 1.264 1.315 1.366 1.409

Modeling 145
6.4 Modeling
6.4.1 Osmotic virial equation
The thermodynamic data of aqueous maltodextrin and mono-, di- and trisaccharides were
correlated by the osmotic virial equation:








+ + +
ρ
= ... c A c A
M
c
a
s s
s 2 osm osm
2
n 1
1
3
1
- ln (6.6)
where a
1
is the water activity, c
s
is the concentration of solute in mass/volume, ρ
1
is the
molar density of pure water, M
n
is the molar mass of the solute and A
2
osm
and A
3
osm
are the
second and third osmotic virial coefficients of the solute in the solvent, respectively. If the
experimental reduced water activity (-ρ
1
ln a
1
/c
s
) is plotted over c
s
, it is possible to evaluate at
c
s
=0 the number average molar mass M
n
and the second and third virial coefficients. For
polydisperse polymer samples the second and third virial coefficients are given by (Kurata,
1982):
∑ ∑
= =
=
N
i
N
j
ij j i
A w w A
2 2
osm
2
(6.7)
∑ ∑ ∑
= = =
=
N
i
N
j
N
k
ijk k j i
A w w w A
2 2 2
osm
3
(6.8)
and

=
=
N
i
i s
c c
2

where N is the number of components in a polydisperse sample, c
i
is the concentration of
component i, w
i
is the weight fraction of component i, and A
ij
and A
ijk
are the second and third
mixed virial coefficients.
In a solution of a polydisperse sample it is possible that the virial coefficients are molar mass
dependent, and in this case, it is necessary to consider the polydispersity of the sample in the
osmotic virial equation. The description of polydispersity can be performed by a molar mass
Capítulo 6 146
distribution function or by splitting the distribution in more monodiperse components named
pseudocomponents. This procedure was employed to polydisperse dextran samples in the work
of Kany et al. (1999), assuming that the molar mass distributions of these polymers were
approximated by 8 pseudocomponents. The results of such evaluation showed that the virial
coefficients do not depend on the molar masses of the polymer components. In this case, the
molar masses varied between 3000 and 1.5×10
5
g mol
-1
. The same observation was found by
Hasse et al. (1995) for the virial coefficients of poly(ethylene glycol)s, that are much more
narrowly distributed than dextrans.
As it was seen before, the studied maltodextrins are very polydisperse samples containing
components varying from mono-, di- and trisaccharides such as glucose, maltose and maltotriose
to oligo- and polysaccharides. The great polydispersity can confer the maltodextrins varied
properties in aqueous solutions and this is the objective here: the study of thermodynamic
properties of low molar mass saccharides and maltodextrins in diluted (laser-light scattering) and
concentrated (isopiestic method) solutions.
Thus, it was assumed that the maltodextrins in this work consist of glucose (G), maltose (M),
maltotriose (MT) and a polydisperse fraction (PF) which was splitted in 8 components, as shown
in Figure 6.3. The weight fractions and number- and weight average molar masses of these
components are given in Table 6.11
a
.
Table 6.11. Division of components in the maltodextrin samples
Component w
i
/g g
-1
M
n
(von GPC) M
w
(von GPC)
MD 13 –17
Glucose (G) 0.0184 180.1 180.1
Maltose (M) 0.0606 342.2 342.2
Maltotriose (MT) 0.0449 504.4 504.4
Polydisperse
fraction (PF)
0.8760 2552.0 1.4098×10
4

MD 16 –19
Glucose (G) 0.0266 180.1 180.1
Maltose (M) 0.0594 342.2 342.2
Maltotriose (MT) 0.0792 504.4 504.4
Polydisperse
fraction (PF)
0.8348 1909.2 9.161×10
3


a
as frações mássicas dos oito componentes da fração polidispersa, bem como suas massas
molares são apresentadas na Tabela 6.15 no final deste capítulo.
Modeling 147
Equation 6.6 is used in the correlation of water activity data. For the laser-light scattering
data the evaluation is based on the following equation:
s
E
s
c RT
R
Kc

µ ∂ ρ
− =
θ
1 1
(6.9)
where the terms on the left side are experimental and are given in Tables 6.9 and 6.10, and
the derivative in the right side can be evaluated based on equation 6.6 as follows:
1
pure 1,
1
ln a
RT RT
+
µ
=
µ
(6.10)
... c A c A
M
R
Kc
s s
E
s
+ + + =
θ
2 LS LS
2
w
3
3 2
1
(6.11)
M
w
is the weight averaged molar mass; and
LS
2
A and
LS
3
A are second and third virial
coefficients, respectively given by (Kurata, 1982):
∑ ∑
= =
=
N
i
N
j
ij j i j i
A M M w w
M
A
2 2
2
w
LS
2
1
(6.12)
∑ ∑ ∑ ∑
∑ ∑ ∑
= = = =
= = =

− =
N
i
N
j
N
k
jl ik jk ij l k j
N
l
i l k j i
N
i
N
j k
ijk j i k j i
A A A A M M M M w w w w
M
A M M w w w
M
A
2 2 2 2
3
w
2 2 2
2
w
LS
3
) (
3
4

1
(6.13)
The mixed virial coefficients A
ij
and A
ijk
in equations 6.7-6.8 and 6.12-6.13 were
approximated by geometric mixing rules as given below:
1/2
) (
jj ii ij
A A A = (6.14)
1/3
) (
kkk jjj iii ijk
A A A A = (6.15)
Capítulo 6 148
It can be noted from equations 6.7-6.8 and 6.12-6.13 that, for polydisperse polymers, the
light-scattering virial coefficients
LS
A
2
and
LS
A
3
differ from osmotic virial coefficients
osm
A
2

and
osm
A
2
, whereas for monodisperse polymers there is no difference between osmotic and light-
scattering virial coefficients.
A simultaneous evaluation of laser-light scattering and isopiestic data was performed using a
data bank comprising experimental water activity and light scattering data for maltodextrins
determined in this work, and experimental water activity data for glucose, maltose and
maltotriose at 298.15 K obtained from the literature (Taylor, 1955; Uedaira, 1969; Myiajima,
1983a, b; Cooke, 2002). The assumption that the second and third virial coefficients do not
depend on the molar mass of the solute components did not result in a good agreement between
experimental and correlated values. Thus it was assumed that the second virial coefficients for
glucose (A
GG
), maltose (A
MM
), maltotriose (A
MTMT
) and polydisperse fraction (A
PFPF
) would be a
function of the molar mass of these components and that the third virial coefficient was the same
for all components, i.e., A
GGG
= A
MMM
= A
MTMTMT
= A
PFPFPF
. The results of such correlation are
given in Figure 6.5 for reduced water activities vs. c
s
. Note that in this figure, the results are
getting worse at high solute concentrations of the low molar mass compounds and that the model
underestimate the reduced water activities of the maltodextrins.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.000
0.005
0.010
0.015
0.020
Glucose (Myiajima, 1983)
Glucose (Taylor, 1955)
Glucose (Cooke, 2002)
Maltose (Miyajima, 1983)
Maltose (Uedaira, 1969)
Maltose (Cooke, 2002)
Maltotriose (Miyajima, 1983)
MD 13-17
MD 16-19
correlated
-
ρ

c
s
-
1

l
n

a
1

/

m
o
l

g
-
1
c
s
/ g cm
-3

Figure 6.5. Reduced water activity in aqueous solutions of saccharides at 298.15 K.
Modeling 149
Another attempt to improve the correlation results was to consider the dependence of the
third virial coefficient on the molar mass of the different components (A
GGG
≠ A
MMM
≠ A
MTMTMT

A
PFPFPF
). The obtained results showed that this dependency is clearly verified for the low molar
mass saccharides glucose, maltose and maltotriose and that for the defined polydisperse fraction
the values assume a constant value, as shown in Figure 6.6 and Table 6.12. It is also noteworthy
to observe that the considered dependency leads to better agreement between experimental and
correlated values of reduced water activity of the studied systems, as presented in Figure 6.7. To
evaluate the light scattering data of maltodextrin systems, which were simultaneously correlated
with reduced water activity data, the low molar mass fractions of these polymers corresponding
to the components glucose, maltose and maltotriose were not considered in the calculations. This
means that both concentration and molar mass of such compounds were discounted from raw
data in the calculation of light scattering. This assumption, as it was commented before, is
supported by the fact that the signal provided by the device MALLS for low molar mass
compounds is almost null. Thus, it is assumed that the Kc
s
/R
θ
E
data corresponds to the scattering
of the higher molar mass molecules. Besides virial coefficients, the weight molar masses (M
w
) of
the maltodextrins were also simultaneously correlated. The correlation provided values of M
w

within the experimental error of the GPC determinations. The results can be seen in Figure 6.8.
5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0
0.0
2.0x10
-3
4.0x10
-3
6.0x10
-3
8.0x10
-3
1.0x10
-2
1.2x10
-2
1.4x10
-2
PF
MT
M
G

A
ii
A
iii
smoothed
A
i
i

o
r

A
i
i
i
ln M

Figure 6.6. Dependence of second and third virial coefficients on the molar mass.
Capítulo 6 150
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.000
0.005
0.010
0.015
0.020
Glucose (Myiajima, 1983)
Glucose (Taylor, 1955)
Glucose (Cooke, 2002)
Maltose (Miyajima, 1983)
Maltose (Uedaira, 1969)
Maltose (Cooke, 2002)
Maltotriose (Miyajima, 1983)
MD 13-17
MD 16-19
correlated
-
ρ

c
s
-
1

l
n

a
1

/

m
o
l

g
-
1
c
s
/ g cm
-3

Figure 6.7. Reduced water activity in aqueous solutions of saccharides at 298.15 K.

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
0.0
2.0x10
-5
4.0x10
-5
6.0x10
-5
8.0x10
-5
1.0x10
-4
1.2x10
-4

MD 16-19
MD 13-17
correlated
K
c
s
/
R
θ
,
0
E
/
m
o
l

g
-
1
c
s
/ g cm
-3

Figure 6.8. Laser-light scattering data of aqueous maltodextrin solutions.

Modeling 151
Table 6.12. Second and third virial coefficients of the saccharides and polydisperse fraction
studied in this work
A
GG
= 0.003353 A
GGG
= 0.011556
A
MM
= 0.001315 A
MMM
= 0.007468
A
MTMT
= 0.000572 A
MTMTMT
= 0.005230
A
PFPF
= 0.000213 A
PFPFPF
= 0.003075

6.4.2 VERS Model
Another equation used in this work to evaluate experimental data of water activities and
enthalpies of dilution of the maltodextrins and low molar mass saccharides is the VERS model.
The VERS model (Virial Equation with Relative Surface Fractions), developed by
Grossmann et al. (1995) and used successfully to correlate and predict some thermodynamic
properties of aqueous PEG and dextran solutions, was used in the modelling of the systems in this
work. This equation is based on a semi-empirical group contribution approach for the excess
Gibbs energy with a structure similar to Pitzer’s virial equation for electrolyte solutions (Pitzer,
1991). The activity coefficients are normalized according to the asymmetric convention: for
water the standard state follows Raoul’s law and for the other components it follows Henry’s law.
In this model, the solute concentration is expressed by its surface fraction per 1000 g of water
normalized by the surface fraction of water. The equations for the activities of water (a
w
) and
solute (
*
a
i,m
) are given below:
ijk
k
j
i j k
i
j
j
i j
i
i
i
B
M
i
A
M
m
M
a
1 1 1 1 1 1 1 1 1 1 1
1
1
1
2
1000
2
1000
1000
ln
Θ
Θ
Θ
Θ
∑ ∑ ∑
Θ
Θ









Θ
Θ
∑ ∑
Θ
Θ
− − =
≠ ≠ ≠ ≠ ≠

(6.16)
ijk
j k
k
j
i
ij
j
j
i
i
m i
B
q
q
M
A
q
q
M
m a ∑ ∑
Θ
Θ
Θ
Θ








+ ∑
Θ
Θ








+ =
≠ ≠ ≠ 1 1 1 1 1 1 1 1 1
1
,
2
1000
3
1000
2 ln
*
ln (6.17)

= Θ
j
j j
i i
i
q m
q m

components all
(6.18)
Capítulo 6 152
∑ =
l
l
) i (
l i
Q v q

nts allcompone
(6.19)
where, Θ
1
and Θ
i
are, respectively, the surface fractions of water and component i, m
i
is the
molality of component i, Q
l
the group l surface parameter,
) i (
l
v the number of groups l in
component i, and q
i
the surface parameter of component i. The interaction parameters A
ij
and B
ijk

are given by:
∑ ∑ Θ Θ =
l m
lm
) j (
m
) i (
l ij
a A

groups all

groups all
(6.20)
∑ ∑ ∑ Θ Θ Θ =
l m n
lmn
) k (
n
) j (
m
) i (
l ijk
b B

groups all

groups all

groups all
(6.21)
i
l ) i (
l
) i (
l
q
Q
ν = Θ (6.22)
for the partial molar excess enthalpies of water (
E
h
1
) and solute i (
E
i
h ) the equations are:
j
E
i E
n , p
T
) T / (
T h











µ ∂
− =
2
1
(6.23)
T
A
M
T
RT
h
ij
i j
j
i
E


∑ ∑
Θ
Θ
Θ
Θ








=
≠ ≠ 1 1 1 1
1
1
1000
(6.24)
T
A
q
q
M
T
RT
h
ij
j
j
i
E
i



Θ
Θ








− =


1 1 1 1
1
1000
2 (6.25)










β
+ β ∑ ∑ Θ Θ =


T T
A
) (
lm ) (
lm
) j (
m
(i)
l
ij
2
1
l
groups all
m
groups all
(6.26)
Modeling 153
To account for the influence of temperature on the equilibrium, the following empirical
expression was used:
) ln( )) ( 1 K)( (
0
(2)
0
(1) (0)
T / T T / T / T a
lm lm lm
lm
β + − β + β = (6.27)
where, a
lm
is a binary group site interaction parameter and T
0
=298.15 K is a reference
temperature. All parameters are assumed to be symmetric, i.e.
) ( ) ( j
ml
j
lm
β = β .
For the assignment of groups, water is treated as a single group and the maltodextrins consist
of four structural groups: n pyranose rings, n-1 osidic bonds (O), n alkane (CH
2
) and (3n+2)
hydroxyl groups. This division of groups was proposed by Catté et al. (1995) to describe
carbohydrate molecules. Here, n is calculated as:
142 162
0152 18
.
. M
n
n

= , where M
n
is the number
average molar mass of the solute that is subtracted from the molar mass of one molecule of water
and divided by the molar mass of one molecule of glucose minus one water. The group surface
parameters of the structural groups are estimated by the method of Bondi (1964) and equal to:
1.5620 for the pyranose rings; 0.442 for osidic bonds; 0.540 for CH
2
and 1.200 for hydroxyl
groups. The molecules of the maltodextrins studied in the present work were treated in the
modelling in two different ways: one, considers the maltodextrin as only one component i.e., the
polydispersity is neglected and the low and high molecular mass compounds are considered
together in a same fraction (method 1); another one, considers the polydispersity of these
polymers assuming that the maltodextrins consist of four different components: glucose, maltose,
maltotriose and a polydisperse fraction (method 2). The concentration of the low cited molar
mass components were estimated by the Gauss-Lorentz functions as presented in Figure 6.3. The
polydisperse fraction is the fourth component called in this work as PF, which concentration and
molar mass is estimated from the subtraction of concentrations and molar masses of the low
compounds. The concentrations and molar masses used in method 2 were already presented in
Table 6.11. Number-average molar masses for method 1 are the same as those presented in Table
6.1.
The estimation of the model parameters presented in equation 6.21 was performed as
following: the coefficient
(0)
OH - pyr
(0)
2
CH - pyr

(0)
o - pyr
(0)
pyr - pyr
(0)
or or or β β β β = β
lm
were fitted to
experimental results for the water activities of saccharides and maltodextrin solutions at 298.15
Capítulo 6 154
K; the parameters
(1)
lm
β and
(2)
lm
β with l or m=pyr, O, CH
2
 and HO were fitted to the
experimental results for the heat of dilution of aqueous maltodextrin solutions in pure water at
313.15 K. These parameters were estimated by minimizing the sum of squares as follows:
2
exp
calc exp
) - ( SSQ

=
N
w w
a a
(6.28)
2
exp
calc exp
) - ( SSQ

=
N
Q Q
(6.29)
The values of the estimated parameters for both methods are given in Table 6.13 below and
the results of correlation in Table 6.14.
Figure 6.9 shows the results of correlation obtained for water activity in water – maltodextrin
solutions by method 1 as well as the prediction results for the water activities of mono-, di-, and
trisaccharides in aqueous solutions reported in literature. The predictive capabilities of the VERS
model were also tested using other two kinds of experimental data for glucose and maltose:
freezing point depression and boiling point elevation.
A data base including experimental water activities of glucose, maltose, maltotriose and
maltodextrins was used in the correlation for the determination of interaction parameters. As can
be seen in Figure 6.9, the calculated values are in close agreement with experimental data even at
high polymer concentrations. The same occurred to the predictions of other thermodynamic
properties for these carbohydrate systems except for very high solute concentrations, where the
predictions fail (see Figures 6.10 and 6.11, for method 1). The studied maltodextrins reduce the
activity of water by only about 4 % at 60% mass fraction, which is not so much in comparison
with glucose that, at the same concentration, reduces the a
w
by 16%.


Modeling 155
Table 6.13. Interaction parameters of VERS model for both methods
Binary interaction parameters Method 1 Method 2
) 0 (
pyr pyr−
β

6.53039×10
-1
3.0352×10
-1

(0)
O - pyr
β

-8.85105×10
-1
-3.1716×10
-1

(0)
CH - pyr
2
β

8.64279×10
-1
2.7212×10
-1

(0)
OH - pyr
β

-1.60260×10
-1
-6.2193×10
-2

(1)
pyr - pyr
β

7.2612×10
-5
2.6464×10
-3

(1)
O - pyr
β

-1.1812×10
-4
-1.8649×10
-2

(1)
CH - pyr
2
β

3.5878×10
-4
3.2687×10
-4

(1)
OH - pyr
β

-5.3399×10
-5
-3.6652×10
-3

(2)
pyr - pyr
β

9.7974×10
-4
9.9689×10
-2


0.0 0.1 0.2 0.3 0.4 0.5 0.6
0.86
0.88
0.90
0.92
0.94
0.96
0.98
1.00
MD 13-17 (this work)
MD 16-19 (this work)
glucose (Miyajima, 1983)
maltose (Miyajima, 1983)
maltotriose (Miyajima, 1983)
corr. VERS model
a
1
w
i
/ g g
-1

Figure 6.9. Experimental and calculated water activities of water – maltodextrin and some
mono- and oligosaccharides at 298.15 K (method 1).
Capítulo 6 156
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
225
230
235
240
245
250
255
260
265
270
275
280

glucose (Young, 1957)
maltose (Uedaira & Uedaira, 1969)
predicted VERS model
F
r
e
e
z
i
n
g

t
e
m
p
e
r
a
t
u
r
e

/

K
w
i
/ g g
-1

Figure 6.10. Experimental and predicted freezing temperatures for glucose and maltose at
various concentrations (method 1).
0.0 0.2 0.4 0.6 0.8 1.0
372
374
376
378
380
382
384
386
388
390
392
experimental (Abderafi & Bounahmidi, 1994)
predicted VERS model
B
o
i
l
i
n
g

p
o
i
n
t

e
l
e
v
a
t
i
o
n
/

K
w
i
/ g g
-1

Figure 6.11. Experimental and predicted boiling temperatures for glucose (method 1).
Modeling 157
The influence of temperature on the activity of water is almost negligible i.e., the water
activity remains nearly unchanged when temperature is increased from 298.15 to 318.15 K. This
fact is confirmed by the calorimetric measurements – the heats of dilution for the aqueous
maltodextrin solutions are very small – and also by experimental data on water activity at
318.15K. The heat on diluting an aqueous solution of maltodextrin in pure water measured in the
present work were also correlated by the VERS model. Figure 6.12 shows the calculated and
experimental results for MD 13 – 17 and 16 –19. As can be seen in this figure, the experimental
enthalpies of dilution for both maltodextrins have a slightly dependency on the molar mass and
the heat effects involved are small. Even so, the model can distinguish between the maltodextrins
in calculating the enthalpies involved in the process of dilution. Small values for heats of dilution
were also observed in experiments with dextran solutions, as reported by Großmann et al. (1995).
0.0 0.1 0.2 0.3 0.4 0.5 0.6
-18
-16
-14
-12
-10
-8
-6
-4
-2
0


MD 16-19
VERS model
MD 13-17
VERS model
Q

/

J
w
i
/ g g
-1

Figure 6.12. Experimental and calculated heat of dilution in aqueous solutions of MD 13–17
and MD 16–19 at 313.15 K. (method 1).

The average absolute deviations (%AAD) between experimental and estimated values for all
thermodynamic properties and methods 1 and 2 are presented in Table 6.14.
As can be seen in Table 6.14, the lower global AAD result is provided by method 1, which
does not consider the polydispersity of the maltodextrins, but the difference between both
Capítulo 6 158
methods is not so large. To verify the applicability of the adjusted parameters by both methods to
predict water activities of solutions containing high and low molar mass compounds, the
experimental data from Radosta et al. (1989) for different fractions of a maltodextrin sample was
compared with calculated a
w
from VERS model. The predictions show that the model was able
to distinguish the a
w
lowering behaviour even in the systems with very high molar mass
components (n=120 indicates a molar mass of approximately 19 500 g mol
-1
). The calculation of
these water activities by VERS model with the parameter adjusted by method 1 is slightly better
than that adjusted by method 2. The predictions resulted from method 1 are shown in Figure
6.13. The AAD% for the two attempts are 0.05 and 0.09 % for methods 1 and 2, respectively.
0.0 0.1 0.2 0.3 0.4 0.5
0.975
0.980
0.985
0.990
0.995
1.000

n
120
30 (MD)
16
9
predicted VERS model
a
w
w / g g
-1

Figure 6.13. Results of the predictions (method 1) for a
w
of maltodextrin fractions (n is the
polymerisation number). Experimental data from Radosta et al. (1989).











Modeling 159
Table 6.14. Average relative deviations between experimental and calculated data by VERS
model
Saccharide
Conc. Range/
g g
-1

Temperature/K
%AAD
(method 1)
%AAD
(method 2)
References
Water activity
glucose 0.0177-0.7478 298.15 – 317.99 0.65 0.80
1, 2, 3, 4
maltose 0.0331-0.8372 298.15 – 317.99 0.63 3.42
1, 3, 5
maltotriose 0.0218-0.6567 298.15 – 317.99 0.14 0.37
1, 3
maltotetraose 0.0904-0.7952 317.99 0.44 2.30
3
maltopentaose 0.1382-0.7750 317.99 1.55 1.72
3
MD 16 – 19 0.2598-0.5850 298.15 – 318.15 0.04 0.09 this work
MD 13 –17 0.1750-0.5898 298.15 – 318.15 0.13 0.13 this work
MD 4 –7 0.2859-0.4615 298.15 0.19 0.61 this work
Freezing point depression
glucose 0.005-0.7 0.38 0.55
6
maltose 0.005-0.44 0.02 0.01
7
Boiling point elevation
glucose 0.23 0.22
8
global AAD 0.40 0.93

1
Miyajima et al., 1983;
2
Taylor, 1955;
3
Cooke, 2002;
4
Velezmoro et al., 2000,
5
Uedaira and Uedaira, 1969;
6
Young, 1957;
7
Weast, 1973;
8
Abderafi and Bounahmidi, 1994.

At the beginning, it was expected that the results for the method 2 would provide better
results for calculating the thermodynamic properties. The results show, on the contrary, that
method 1 was better to estimate the studied properties. Probably the reason for this behaviour is
the low concentration of the low molecular saccharides in the maltodextrin samples possibly the
influence of those saccharides on the thermodynamic properties of maltodextrins measured in the
present work is small and to consider their influence as an independent factor do not bring any
improvement to the correlation of the properties.
6.5 Conclusions
The water activity, light scattering and heats of dilution of aqueous solutions of three
different maltodextrins were investigated and evaluated by a osmotic virial equation and a semi-
empirical group contribution model. This data covered a wide concentration range from diluted
to concentrated solutions. Therefore it was possible to evaluate the virial coefficients in the
Capítulo 6 160
systems containing molecules varying from saccharides to polysaccharides. It was found that the
second and third virial coefficients were molar mass dependent up to maltotriose with (n=3) and
that they are constant for higher molar mass components. For the VERS model a good agreement
between experimental and correlated values of water activities was found for maltodextrins and
low molar mass saccharides systems. The model was also able to predict other thermodynamic
properties for glucose and maltose such as freezing point depression and boiling point elevation.
6.6 Acknowledgments
The authors gratefully acknowledge the financial support of CNPq (Conselho Nacional de
Desenvolvimento Científico e Tecnológico - 290011/01-9 + 46668/00-7 + 521011/95-7) and
DAAD (Deutsche Akademische Austauschdienst).
6.7 References
Abderafi, S.; Bounahmidi, T. Measurement and Modeling of Atmospheric Pressure Vapor-
Liquid Equilibrium Data for Binary, Ternary and Quaternary Mixtures of Sucrose, Glucose,
Fructose and Water Components. Fluid Phase Equilibria 93 (1994) 337-351.
Bondi, A. Van der Waals volumes and Radii. J. Phys. Chem. 68 (1964) 441-451.
Catté, M; Dussap, C-G; Gros, J-B. A Physical Chemical UNIFAC Model for Aqueous
Solutions of Sugars. Fluid Phase Equilibria 105 (1995) 1-25.
Cesàro, A.; Cuppo, F.; Fabri, D.; Sussich, F. Thermodynamic Behavior of Mixed
Biopolymers in solution and in Gel Phase. Thermochimica Acta 328 (1999) 143-153.
Cooke, S. A.; Jónsdóttir, S. Ó; Westh, P. A Thermodynamic Study of Glucose and Related
Oligomers in Aqueous solution: Vapor Pressures and Enthalpies of Mixing. J. Chem. Eng. Data
47 (2002) 1185-1192.
Defloor, I.; Vandenreyken, V.; Grobet, P. J.; Delcour, J. A. Fractionation of Maltodextrins by
Ethanol. J. Chromatography 803 (1998) 103-109.
Flory, P. J. Principles of Polymer Chemistry, Cornell University Press, Ithaca, 1953.
References 161
Gaube, J., Pfennig, A., Stumpf, M. Vapor-Liquid Equilibrium in Binary and Ternary
Aqueous Solutions of Poly(ethylene glycol) and Dextran. J. Chem. Eng. Data 38 (1993) 163-166.
Glicksman, M. Food Hydrocolloids v III, CRC Press Inc., New York, 1986.
Groβmann, C., Tintinger, R., Zhu, J., Maurer, G. Aqueous Two-Phase Systems of
Poly(ethylene glycol) and Dextran – Experimental Results and Modeling of Thermodynamics.
Fluid Phase Equilibria 106 (1995) 111-138.
Hasse, H., Kany, H-P, Tintinger, R., Maurer, G. Osmotic Virial coefficients of Aqueous
Poly(ethylene glycol) from Laser-light Scattering and Isopiestic Measurements. Macromolecules
28 (1995) 3540-3552.
Huglin, M. B. Light Scattering from Polymer Solutions. Academic Press, London, 1972.
Kany, H-P, Hasse, H., Maurer, G. Thermodynamic Properties of Aqueous Dextran Solutions
from Laser-light Scattering, Membrane Osmometry, and Isopiestic Measurements. J. of Chem.
Eng. Data 44 (1999) 230-242.
Kany, H-P. Thermodynamische Eigenschaften Wässriger Polymer-Lösungen. Ph Thesis,
University of Kaiserslautern, Germany, 1998.
Kasapis, S., Morris, E. R., Norton, I. T., Clark, A. H. Phase Equilibria and Gelation in
gelatin/maltodextrin Systems – Part I: Gelation of Individual Components. Carbohydrate
Polymers 21 (1993a) 243-248.
Kasapis, S., Morris, E. R., Norton, I. T., Clark, A. H. Phase Equilibria and Gelation in
gelatin/maltodextrin Systems – Part II: Polymer Incompatibility in Solution. Carbohydrate
Polymers 21 (1993b) 249-259.
Kasapis, S., Morris, E. R., Norton, I. T., Clark, A. H. Phase Equilibria and Gelation in
gelatin/maltodextrin Systems – Part III: Phase Separation in Mixed Gels. Carbohydrate Polymers
21 (1993c) 261-268.
Kasapis, S., Morris, E. R., Norton, I. T., Clark, A. H. Phase Equilibria and Gelation in
gelatin/maltodextrin Systems – Part IV: Composition-Dependence of Mixed-Gel Moduli.
Carbohydrate Polymers 21 (1993d) 269-276.
Capítulo 6 162
Lammertz, S.; Maurer, G. Thermodynamic Properties of Aqueous solutions of
Polyelectrolytes and an Inorganic Salt. ESAT Congress, Greek, 2002.
Marchal, L. M.; Beeftink, H. H.; Tramper, J. Towards a Rational Design of Commercial
Maltodextrins. Trends in Food Science & Technology 10 (1999) 345-355.
Miyajima, K., Sawada, M., Nakagaki, M. Studies on Aqueous Solutions of Saccharides. I.
Activity Coefficients of Mmonosaccharides in Aqueous Solutions at 25 °C. Bull. Chem. Soc. Jpn.
56 (1983a) 1620-1623.
Miyajima, K., Sawada, M., Nakagaki, M. Studies on Aqueous Solutions of Saccharides. II.
Viscosity B-Coefficients, Apparent Molar Volumes, and Activity Coefficients of D-Glucose,
Maltose, and Maltotriose in Aqueous Solutions. Bull. Chem. Soc. Jpn. 56 (1983b) 1954-1957.
Mothé, C. G., Rao, M. A. Rheological Behavior of Aqueous Dispersions of Cashew Gum and
Gum Arabic: Effect of Concentration and Blending. Food Hydrocolloids 13 (1999) 501-506.
Pitzer, K. S. Activity Coefficients in Electrolyte Solutions. CRC Press, Boca Raton, 1991.
Pitzer, K. S.; Peiper, C. Thermodynamic Properties of Aqueous sodium Chloride Solutions.
J. Phys. Chem. Ref. Data 13 (1984) 1-102.
Radosta, S.; Schierbaum, F. H. Polymer-water Interaction of Maltodextrins. Part III: Non-
Freezable Water in Maltodextrin solutions and Gels. Starch 41 (1990a) 142-147.
Radosta, S.; Schierbaum, F.; Reuther, F.; Anger, H. Polymer-water Interaction of
Maltodextrins. Part I: Water Vapour Sorption and Desorption of Maltodextrin Powders. Starch
41 (1989b) 395-401.
Radosta, S.; Schierbaum, F.; Yuriev, W. P. Polymer-water Interaction of Maltodextrins. Part
II: NMR Study of Bound Water in Liquid Maltodextrin-Water Systems. Starch 41 (1989c) 428-
430.
Schierbaum, F., Kettlitz, B., Radosta, Reuther, F., Richter, M., Vorweg,W. Acta Alimentaria
Polonica 10 (1984) 69-99.
References 163
Schierbaum, F., Kettlitz, B., Radosta, S., Reuther, F., Richter, M., Vorweg, W. Zum Stand
der Kenntnisse über Struktur-Eigenschaftsbeziehungen von Maltodextrinen. Acta Alimentaria
Polonica 10 (1984) 69-99.
Schierbaum, F., Radosta, S., Vorweg, W., Ynriev, V. P., Brando, E. E., German, M. L.
Formation of thermally reversible maltodextrin gels as revealed by low resolution H-NMR.
Carbohydrate Polymers 18 (1992) 155-163.
Silva, L. H. M., Meirelles, A. J. A. Phase Equilibrium and Protein Partitioning in Aqueous
Mixtures of Maltodextrin with Polypropylene Glycol. Carbohydrate Polymers 46 (2001), 267-
274.
Silva, L. H. M., Meirelles, A. J. A. Phase equilibrium in polyethylene glycol/maltodextrin
aqueous two-phase systems. Carbohydrate Polymers 42 (2000) 273-278.
Silva, L. H. M.; Loh, W. Calorimetric Investigation of the Formation of Aqueous Two-Phase
Systems in Ternary Mixtures of Water, Poly(ethylene oxide) and Electrolytes (or Dextran). J.
Phys. Chem. B 104 (2000c) 10069-10073.
Swift, G. Requirements for Biodegradable Water-Soluble Polymers. Polymer Degradation
and Stability 59 (1998) 19-24.
Taylor, J. B.; Rowlinson, J. S. The Thermodynamic Properties of Aqueous Solutions of
Glucose. Trans. Faraday Soc. 51 (1955) 1183-1192.
Uedaira, H.; Uedaira, H. Miyajima, K., Sawada, M., Nakagaki, M. Activity Coefficients of
Aqueous Xylose and Maltose Solutions. Bull. Chem. Soc. Jpn. 42 (1969) 2137-2140.
Velezmoro, C. E.; Oliveira, A. L.; Cabral, F. A.; Meirelles, A. J. A. Prediction of Water
Activity in Sugar Solutions using Models of Group Contribution and Equation of State. J. Chem.
Eng. Jpn. 33 (2000) 645-653.
Weast, R. C. CRC Handbook of Chemistry and Physics, 53
rd
. edn., The Chemical Rubber,
Cleveland, 1973.
Young, F. E. D-Glucose-Water Diagram. J. Phys. Chem. 61 (1957) 616-619.
Capítulo 6 164

Table 6.15. Components of the polydisperse fraction (PF) of maltodextrins MD 13 – 17 and MD
16 – 19.
MD 13 – 17 MD 16 – 19
w / g g
-1
M / g mol
-1
w / g g
-1
M / g mol
-1

0.0539 780.83 0.1400 809.31
0.1146 1239.40 0.1504 1221.18
0.1575 1457.61 0.1190 1632.33
0.1085 2773.14 0.1484 2864.48
0.1146 5107.10 0.0290 6990.53
0.3009 25062.33 0.2304 19197.2
0.1415 114665.5 0.0097 102304.5
0.0120 157650.2 0.0078 141695.5

Conclusões Gerais 175
Conclusões Gerais

A seguir, são apresentadas conclusões gerais a respeito de cada capítulo apresentado neste
trabalho.
1) Soluções aquosas contendo diferentes polióis (D-sorbitol, D-manitol, xylitol, eritritol e
glicerol) foram estudadas, como apresentado no capítulo 2, determinando-se suas atividades
de água com o emprego de um higrômetro elétrico nas temperaturas entre 10 e 35 °C a
várias concentrações. Os dados referentes às determinações de a
w
, bem como dados de
solubilidades desses compostos em água, foram empregados na predição e correlação com
os modelos de contribuição de grupos ASOG e UNIFAC. O emprego dos valores dos
parâmetros de interação dos modelos disponíveis na literatura resultou em predições ruins,
principalmente a altas concentrações. Por essa razão, optou-se por um reajuste de
parâmetros considerando que um grupo hidroxila presente, por exemplo, numa molécula de
algum álcool, tem um outro comportamento em comparação aos grupos hidroxilas que estão
presentes nas moléculas dos polióis. No polióis, os grupos hidroxilas estão ligados a
carbonos consecutivos na molécula, e por isso, pode haver um efeito de proximidade
intramolecular. O reajuste, baseado nesse efeito de proximidade, e também realizado com
um banco de dados abrangendo altas concentrações dos solutos em solução e diferentes
moléculas de polióis (com número de carbonos variando de 3 a 6), forneceu bons resultados
nas correlações e também nas predições de a
w
em sistemas multicomponentes, depressão do
ponto de congelamento e ponto eutético quando o modelo UNIFAC-Larsen foi utilizado.
2) O emprego do método de contribuição de grupos UNIFAC combinado ao termo de Debye-
Hückel mostrou-se uma ferramenta útil no cálculo da atividade de água, pH e solubilidade
de aminoácidos em soluções aquosas considerando a dissociação parcial desses compostos
em água. Como mostrado no Capítulo 3, o modelo forneceu praticamente os mesmos
desvios para os pH reais e ideais das soluções de aminoácidos estudadas em três diferentes
condições de pH: em tampão ácido, básico e em água. Mas foi também verificado que,
apesar da diferença não ser tão significativa para o pH, para as concentrações a diferença
entre o caso real e o ideal é mais significativa (como se vê na Figura 7). Nos cálculos de
Conclusões Gerais 176
solubilidade dos aminoácidos em água e em soluções contendo sais ou ácidos, notou-se
também que o modelo foi capaz de fornecer melhores resultados em comparação aos
resultados calculados no caso ideal. O que reforça essa afimação é que, para os aminoácidos
valina e glicina, os valores de ∆s e ∆h requeridos para o cálculo de solubilidade não foram
ajustados neste trabalho e sim retirados da literatura.
3) As viscosidades de misturas de polietileno glicóis foram determinadas e correlacionadas
pelo método de contribuição de grupos GC-UNIMOD. O reajuste de parâmetros da
literatura mais uma vez foi necessário, a fim de melhorar os resultados do cálculo da
viscosidade pelo modelo, assim como a predição em sistemas multicomponentes. Foi
observado que, especialmente para misturas binárias que continham polímeros muito
diferentes com relação à massa molar, os resultados de predição com valores de parâmetros
da literatura eram ruins. Isso poder ser o resultado do ajuste desses parâmetros a um banco
de dados que continham moléculas semelhantes entre si. Com o reajuste dos parâmetros de
interação foi também melhorada a capacidade preditiva do modelo para o cálculo de
viscosidades de sistemas contendo PEGs.
4) O emprego de um banco de dados abragendo soluções aquosas de polietileno glicóis com
massas molares entre 200 e 10 000 g mol
-1
tornou possível ajustar parâmetros de uma
equação semi-empírica para a viscosidade de soluções (equação de Kumar) e relacioná-los
ao grau de hidratação desses polímeros, comparando-os, dentro do possível, com dados
disponíveis na literatura. Os resultados da correlação e das predições são bastante razoáveis,
o que sugere o uso dessa equação para a estimativa de viscosidades de soluções aquosas
contendo PEGs. Já o modelo de contribuição de grupos GC-UNIMOD não foi capaz de
descrever bem os dados experimentais, mesmo quando todos os parâmetros de interação de
grupos foram reajustados. Tem-se duas suposições para tal resultado: 1) extrapolação das
viscosidades dos componentes puros (PEGs que se apresentam sólidos nas temperaturas de
estudo); 2) maior complexidade da viscosidade dinâmica dos sistemas estudados, não
corretamente descrita pelo modelo.
5) No estudo de propriedades termodinâmicas de sistemas contendo maltodextrinas foram
realizadas determinações experimentais e avaliação e cálculo das propriedades pela equação
osmótica virial e o método de contribuição de grupos VERS. A avaliação dos coeficientes
Conclusões Gerais 177
viriais pela equação osmótica (que foram ajustados simultâneamente a dados de atividade de
água e espalhamento de luz e, portanto, soluções concentradas e diluídas) mostrou a
dependência desses coeficientes com a massa molar de glicose, maltose e maltotriose,
atingindo valor constante para moléculas acima de massa molar 504 g mol
-1
. O modelo
VERS, que tem estrutura similar à equação virial de Pitzer para eletrólitos, foi empregado
com sucesso no cálculo de atividade de água e outras propriedades em misturas contendo
maltodextrinas e açúcares.
Assim, pode-se concluir que os métodos de contribuição de grupos podem ser uma ferramenta útil
no cálculo de propriedades físico-químicas em sistemas contendo alguns compostos presentes em
produtos alimentícios, como os estudados neste trabalho, assim como de polímeros empregados em
separações de biomoléculas. O que deve ser ressaltado é que, muitas vezes, os valores dos
parâmetros de interação binários já disponíveis na literatura não são capazes de oferecer uma boa
predição de propriedades, e que um reajuste de tais parâmetros torna-se necessário para sistemas
contendo moléculas muito diferentes dos compostos orgânicos mais tradicionais ou que apresentem
determinadas características na molécula, como por exemplo, as hidroxilas no caso dos polióis.

BANCA EXAMINADORA

Prof: Dr. Antonio José de Almeida Meirelles (FEA/UNICAMP)

Profa. Dra. Maria Alvina Krähenbühl (FEQ/UNICAMP)

Prof. Dr. Pedro de Alcântara Pessôa Filho (POLI/USP)

Prof. Dr. Martín Aznar (FEQ/UNICAMP)

Prof. Dr. Fernando Antonio Cabral (FEA/UNICAMP)

Prof. Dr. Eduardo Augusto Caldas Batista (DZT/UEPG)

Profa. Dra. Maria Ângela de Almeida Meireles (FEA/UNICAMP)

Aos meus dois amores, Dirk e nosso filho.

Cristina e Roberta pelos bons momentos que passamos juntas. Aos meus amigos Raquel. Vanderlei e Hicham. Wong e Helcio pela ajuda nos trabalhos experimentais. Aos membros da banca examinadora pelas observações e sugestões feitas para a melhor elaboração da versão final deste trabalho. . Ao CNPq pela bolsa de estudos e pelo financiamento da pesquisa no Brasil e ao convênio CNPq/DAAD pelo financiamento da bolsa de estudos na Alemanha. À minha vó que me faz rir e que tanto amo. Às minhas amigas Cíntia. que estão na Alemanha. À Melissa. Aos meus pais e minhas irmãs que sempre me incentivam e apóiam. pelas conversas alegres e apoio quando precisei. À Silke Lammertz e Viktor Ermatchkov pela ajuda inicial nos laboratórios na Universidade Kaiserslautern. Dr. pela orientação. Ao Prof. incentivo e presença sempre. Christiane. Dr.AGRADECIMENTOS Ao Prof. Antonio José de Almeida Meirelles. Sueli. Gerd Maurer pela orientação em parte deste trabalho e auxílio para a realização do mesmo.

...........................1 Fundamentos e modelos termodinâmicos ........2....................3 Compostos orgânicos de interesse na área de alimentos e biotecnologia .. 12 1................ 26 1.........5 Nomenclatura .........2 Aminoácidos....................................... 22 1.....................4 Propriedades físico-químicas estudadas.............................................................1................................................. 23 1...4......................................................................1.............................................................................................................4.....................................................................................................5 pH ................................................4........................ 7 1.............................................................................................3............ 30 1.......................................................3 Polietileno glicóis ............................................................................................................................ 27 1..4 Maltodextrinas (MD).................................................................. 1 Introdução............ xxiii Abstract . 7 1.................................................................................2 Para cálculo de viscosidades de misturas ...................... 28 1............ 3 Objetivos ..................................................................................4......................................... 30 1.......................................................................... xiii Índice de Figuras............................................................... xviii Resumo ............. 24 1................................................................................................1 Bases termodinâmicas ........... 34 ....................................4 Elevação do ponto de ebulição.........................................................2....................................4......................................................................7 Entalpia de diluição.....................................2 Métodos de contribuição de grupos ......................................................3 Depressão do ponto de congelamento............................................. 6 1................................................. 29 1.............................................................1 Atividade de água...................... 26 1.................ix ÍNDICE GERAL Índice de Tabelas........................................... 12 1............1 Polióis........................................................................................................ 23 1............xxv CAPÍTULO 1 1........................................................6 Referências bibliográficas ...........................................................................................3........4....... 29 1....................................................................................4.............2 Solubilidade.....6 Espalhamento de luz.......................................................................................................................... REVISÃO BIBLIOGRÁFICA...............3...... 9 1.......................................................................................................3... 28 1...................................................................................................... 25 1....1 Para cálculo de coeficientes de atividade .........................................2 Modelos para energia livre de Gibbs.... 18 1.........................................................................................................

............................4....2 Experimental Procedure .................................................................2 Equilibrium data .....................................................................................................................6 Conclusions ........................ 61 3....................4................................... 41 2........................3 Experimental Section ... 43 2.................1 Abstract ...................3................1 Density data.........1 Modified UNIFAC Model for Electrolytes ................................................................................................................................................................................ 41 2...3...................................................................................5 Literature Cited .....................................4......... 62 3................................. 74 3.................................................................................... 75 3........5 Results and Discussion................... PH AND DENSITY OF AQUEOUS AMINO ACIDS SOLUTIONS ...............2 Readjustment of group interaction parameters............................................................................................................................. 84 3.................................................................................................... 47 2..................................................................................................................1 Water activity .............................2 Introduction ........................................................................... 77 3......................... WATER ACTITITY.....................................................2 Solubility of Amino Acids in Aqueous Solutions............................................................................................................................................................................... WATER ACTIVITY IN POLYOL SYSTEMS.............5.................................................................................. 54 2...............................x CAPÍTULO 2 2.................................................... 63 3.................... 59 3........ 84 3.......4 Results and Discussion.......................... 61 3............6 Acknowledgment ..............5............................................. 42 2...................................................... 53 2.......................... 62 3....................................... 57 CAPÍTULO 3 3.........................................1 Abstract ...................3 Experimental Results.......1 Materials.......................................3.............................. 62 3................3 Materials and methods ............................4 Thermodynamic Modeling........... 75 3..................................................................7 References ............................................................................................8 Acknowledgment .........................................4..... 39 2.................................................................. 87 .............................................................................................................................................................................. 68 3............................................... 43 2..............................2 Introduction ................... 68 3.................................................................................................................3 Predictions with the new set of parameters ......................4................

......................4 Results and Discussion.........................2 Introduction .......................................................................................................................................... 89 4....2 Introduction ...1 Abstract ...................................................xi CAPÍTULO 4 4..............................2 Apparatus and procedures ......................................2 Apparatus and Procedures ............................................................................................ KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL) AQUEOUS SOLUTIONS........................................................................... 103 CAPÍTULO 5 5........................................... 107 5..4................................................................3..... 107 5.........3......................1 Excess molar Gibbs energy of activation for viscous flow ...........................................6 Literature Cited ..................... 105 5....................................... 93 4................................................................................................ 95 4.................................... 92 4.................................. 101 4.....................................1 Materials.. 112 5......................4...........................2 Viscosity prediction in multicomponent systems............ 122 ..........................1 Abstract ...............................................................................................................5 Conclusions ..1 Materials.......................................................1 Kumar’s equation for viscosity correlation in PEG mixtures ...............................4 Results and Discussion......................................................................................................................................2 GC-UNIMOD model.............................................3 Experimental Section ..................................................................................... 92 4......... 101 4.........4.7 Acknowledgments .................................................... 108 5...................4..................................................................... 109 5..............................................................................................................................................................................................................3...........5 Literature Cited ..................... 93 4.......................................................................................................................................... 109 5...........................................................................3 Experimental .............. 91 4................................................................................ 91 4...................... 108 5..... 121 5....... 96 4.......................... KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL) BLENDS........................3..................................................

....2 Introduction ................................. 129 6...........................3..........................................................6 Acknowledgement.....................................................................................................................................................................5 Conclusions ..................................................................... 129 6......................2 VERS Model ..........................................................3 Experimental ......................................3.................................. 130 6......... 165 Conclusões Gerais ..................... 145 6............................. 175 .............................................................................4.......................................................................................................................................................... 127 6..................................................................................................................................................................................................................................................................................................................................................................4......................................................................................1 Osmotic virial equation ....................................xii CAPÍTULO 6 6. 145 6....................................................................................................................................................... 159 6.................................................... 130 6.................7 References ..............................................1 Materials........................................ 131 6......4 Modeling ........................2 Methods ....................................................... 151 6. THERMODYNAMIC PROPERTIES OF MALTODEXTRIN AQUEOUS SOLUTIONS....1 Abstract ......... 160 6........... 160 ANEXOS ANEXO 1 ............................................................

................................................................... 48 Mean relative deviations between experimental and calculated aw data points...3 Table 2....................................................................10 Thermodynamic data of polyols and water ................................................. 53 CAPÍTULO 3 Table 3.............................................................. 52 Table 2...1 Table 2........2 Table 2. 47 Systems used for adjusting the group interaction parameters ................... 43 Table 2........ dl-alanine............................8 Mean relative deviations between experimental and calculated solubility data..................................................................... l-arginine and l-proline as a function of amino acid mass fraction (wa) at 25.....................................................................11 Water activity prediction in polyol mixtures......xiii Índice de Tabelas CAPÍTULO 2 Table 2.............................1 Water activity in binary polyol solutions as a function of mass fraction of polyol (w2) .......................................................... 51 Table 2..................0±0.............1 Experimental water activities of glycine...................1 °C................................... 45 Group assignment for polyols .......................................... 46 UNIFAC-Lyngby interaction parameters......................................................9 Table 2.............................................................................................................7 (cont............................ 52 Values of ∆A and ∆B for the calculations of ∆Cp with linear temperature dependency...6 Table 2..................) ................. 50 Table 2............... 64 ........................................................... 44 Water activity in ternary polyol solutions at 25 °C.........5 Table 2.................. 47 ASOG interaction parameters ....4 Table 2.....................................

..................2 Table 4....... 68 Coefficients for correlating pK values with temperature ................... 66 Table 3.................0±0.... 70 ∆s and ∆h values used in this work (in bold script) and obtained from literature ..... l-arginine and l-proline as a function of total solute mass fraction (wt) at 25...........................................................................3 °C ...................... acid and base in aqueous solutions .............3 Densities of aqueous solutions of glycine............4 Table 3................ dl-alanine............................................. 82 CAPÍTULO 4 Table 4.................................................................... l-arginine and l-proline in water and in basic and acid buffers (wa=amino acid mass fraction) 24............................................................................... 93 Table 4...........................................................1 °C ............................ 94 ......................................................... 92 Viscosity of binary blends of poly(ethylene glycol)s at various temperatures...........7 Linear fitting for the densities of amino acids............................................................................2 Poly(ethylene glycol) characterization.......................xiv Table 3..................... salts....7±0.. 74 Table 3..........................9 Table 3................................................................................... 65 Table 3.....5 Table 3...................... dl-alanine.............................. 76 Table 3.... 94 Viscosity of binary blends containing poly(ethylene glycol) 400 at various concentrations.............8 Group interaction energies uij (K) between ions and amino acid characteristic groups ........................................2 Experimental pH of glycine.....6 Division of groups for the amino acids used in this work............................... 78 Table 3. 78 Average relative deviations between calculated and experimental data ..............1 Table 4...............................................10 Group interaction energies uij between like groups....3 (continued) .............

................................. 98 Energy interaction parameters (in Kelvin) for the GC-UNIMOD model ... 110 Table 5.......... 114 Adjusted coefficients of eqs 5..............7 Table 5...................... ν........... 97 Table 4..................................... 111 Kinematic viscosities................... 110 Table 5...... ν.....................4 Viscosities of ternary mixtures of poly(ethylene glycol)s at various temperatures.............. 115 Table 5............. 112 Table 5................6 Coefficients of the Redlich-Kister equation (Ap) and average absolute deviations (AAD) between experimental and calculated G*E values............3 Kinematic viscosities.........................1 Average relative molar masses (Mn)...4 (cont.. 5.......................8 Constants for the generalized equation ............................................ 99 CAPÍTULO 5 Table 5.. w ................................... of multicomponent poly(ethylene glycol) aqueous solutions at various temperatures ................... ν....................................7 Table 4.........8 Average absolute deviations (AAD) for the viscosity estimation............................... 119 .........5 Table 5............ 118 Results of the generalized correlation for viscosity calculations ...................................5 for calculating kinematic viscosities of aqueous PEG solutions......) .............................4 and 5............................ 109 Table 5.....................xv Table 4.......5 Viscosities of a multicomponent mixture of poly(ethylene glycol)s at various temperatures..............................15 K and various mass fractions........ polydispersity index and water content of PEGs used in this work ............2.............6 Linear fittings for the densities of poly(ethylene glycols) in aqueous solutions................. 95 Table 4.................................. w....................................2 Kinematic viscosities..3 Table 5.......... in aqueous solutions of PEGs a 293................. in aqueous solutions of PEGs at various temperatures and mass fractions............................... 95 Table 4...............................................

........................................................... 138 Table 6.........................15 K ............................ 159 ........14 Interaction parameters of VERS model for both methods ........15 K...... 141 Specific density of aqueous MD 13 – 17 and MD 16 – 19 solutions at 298........4 Experimental results for the water activities in aqueous solutions of MD 4 – 7 at 298.................... 141 Experimental heats of dilution of maltodextrin 16 – 19 at 313....................... 151 Table 6............................ 155 Average relative deviations between experimental and calculated data by VERS model ....................12 Excess Rayleigh ratios of MD 13 – 17 in water at 298..............6 Table 6........................... 144 Division of components in the maltodextrins samples............1 Experimental data on the number........................5 Experimental results for the water activities in aqueous solutions of MD 13 – 17 and 16 – 19 at 318......15 K....................................and mass-average molar mass of the maltodextrins studied in this work ...................8 Experimental heats of dilution of maltodextrin 13 – 17 at 313............ 137 Table 6................................................................. 131 Table 6....................... 146 Second and third virial coefficients of the saccharides and polydisperse fraction studied in this work......................xvi CAPÍTULO 6 Table 6.........15 K......................................................................... 139 Table 6.........................10 Table 6..........3 Experimental results for the water activities in aqueous solutions of MD 16 – 19 at 298....................................... 143 Excess Rayleigh ratios of MD 16 – 19 in water at 298.........................15 K .............................................2 Experimental results for the water activities in aqueous solutions of MD 13 – 17 at 298.................15 K .............. 142 Table 6....15 K .......13 Table 6....9 Table 6................. 138 Table 6.........................................................7 Table 6................15 K.......15 K .......11 Table 6..

.........................4 ..6 Tabela A................................2 – A...1 .....8 Viscosidades cinemáticas experimentais dos PEGs 8000 e 10 000 ..............................5 Tabela A.................. 170 Resultados do ajuste de parâmetros de acordo com a opção 3 para o cálculo de viscosidades de PEGs em solução aquosa ............................ 167 Parâmetros das equações A....4 Tabela A...................... 169 Parâmetros de interação binária para o modelo UNIMOD (Opção 2)................................................... 166 Parâmetros da equação A..... 169 Parâmetros de interação binária para o modelo UNIMOD (Opção 3)................................... 169 Parâmetros de interação binária para o modelo UNIMOD (Opção 1)................... 172 .......1 Tabela A..........................7 Tabela A..... 167 Parâmetros para o modelo UNIMOD.2 Tabela A..................3 Tabela A.........xvii ANEXO 1 Tabela A....

....................................................2 Comparison between experimental densities from literature and those determined in the present work................................6 com possibilidade de ramificação. 50 Figure 2..................................5 Predictions of water activities for the ternary system water – xylitol – sorbitol at 25 °C .....................................................................1 Water activities determined in the present work and data taken from ref 5 for the amino acid glycine .....1 Figure 2.......2 Prediction of water activities in the glycerol – water system................................... ......... 54 CAPÍTULO 3 Figure 3......4 Figure 3........... 45 Experimental and calculated water activities for xylitol solutions at 25 °C................................... 67 Figure 3.......................................... 26 CAPÍTULO 2 Figure 2.................................... 80 ................................ 79 Experimental and calculated solubility of some amino acids ..xviii Índice de Figuras CAPÍTULO 1 Figura 1.................................................................................. 49 Figure 2... y= ligação α-1.......... 79 Figure 3..... 49 Figure 2.4 Experimental and calculated solubilities for D-mannitol at different temperatures.................................................................................... 63 Figure 3...........3 Solid-liquid equilibria for sorbitol aqueous solutions (solubility and freezing point depression) .........4 para a formação de cadeia linear).....................3 Experimental and calculated aw of aqueous solutions containing l-proline .................................5 Experimental and estimated pH of glycine solutions........................................................................1 Estrutura química da maltodextrina (x=ligação α-1........

.................. ┉ 400+600+1000+1500+3350.... 119 ....... 1000+3350.. 96 Experimental and correlated viscosities of binary PEG mixtures (experimental data of Table 4.. PEG 600.......... PEG 400.........15 K.......... ○ 1000+1500+3350........ ○............. 400+1000+3350....................... 83 CAPÍTULO 4 Figure 4..................15 K and various concentrations: ■.... ○ 400+1500..........xix Figure 3....... PEG 1000.....6 Experimental and estimated solubilities of glycine at various pH values and NaCl concentrations ............ — GC-UNIMOD/ option 2) ............ 100 CAPÍTULO 5 Figure 5........ □..... Figure 4....... 116 Href values as a function of PEG molecular mass .............. — Redlich-Kister) .7 Comparison between the distribution of the ionic species of glycine as a function of real () and ideal (----) pH .............. 600+1000+1500.............. ┉ 200+600.................. — GC-UNIMOD/ option 2)............1 Excess molar Gibbs energy of activation for viscous flow of PEG binary mixtures (■ 400+600 at 293............. 99 Experimental and predicted viscosities of multicomponent PEG mixtures (∇ 400+600+1000........... —.... ■ 400+600.......... calculated values ... 117 Experimental and calculated viscosities at 293.... PEG 3350.....................2 Figure 5. ▼ 600+1000.... □ 400+600 at 333.......................................1 Figure 5..3 Bref values as a function of PEG molecular mass ......15 K... Figure 4. PEG 1500......... •.. + 600+3350..2: ▲ 200+400............... 81 Figure 3....................... ▲ 400+1500.......3 400+3350..2 400+3350... ▲..

...............2554...........................15 K (method 1) ... 156 ............... 155 Figure 6........... w=0......6 Figure 6......... PEG 1500...............15 K...................8 Figure 6.........7 Figure 6...................................3 Results of Gauss-Lorenz functions for the molar mass distrbution of MD 16-19......... •................ w=0...and oligosaccharides at 298.... 134 Figure 6.2487........... 135 Figure 6.........2441 .................................................................. 132 Results of the experiment GPC/MALLS for the maltodextrins studied in this work... w=0... 156 Figure 6. w=0.. □......... maltose and maltoriose from the Gauss-Lorenz functions.... ■.... 136 Figure 6.......4899.... ∆.........15 K..................................... 150 Laser-light scattering data of aqueous maltodextrin solutions.. PEG 400..9 Reduced water activity in aqueous solutions of saccharides at 298................4 Experimental and calculated viscosities at various temperatures and concentrations: ▲............4 Molar mass as a function of elution volume................5 Figure 6.. 148 Dependence of second and third virial coefficients on the molar mass ........xx Figure 5... 150 Experimental and calculated water activities of water – maltodextrin and some mono........ PEG 600............... ○..................1 Figure 6............................ using the MALLS data and elution volumes for glucose. PEG 10 000..... PEG 1000...... 149 Reduced water activity in aqueous solutions of saccharides at 298..................11 Experimental and predicted boiling temperatures for glucose (method 1) ................... w=0.10 Experimental and predicted freezing temperatures for glucose and maltose at various concentrations (method 1)...............4534.. w=0.......... PEG 8000.......................4777...........2 Gel permeation profiles of maltodextrins......................... 120 CAPÍTULO 6 Figure 6............

................. 173 Figura A.............................2 Viscosidades calculadas e experimentais a 293......... ○...................... (1989). 12 000 e 15 000 ................. 173 Figura A......... PEG 3350.................................................xxi Figure 6............. PEG 400..................12 Experimental and calculated heat of dilution in aqueous solutions of MD 13 – 17 and MD 16 – 19 at 313.....15 K e várias concentrações: ■. 158 ANEXO 1 Figura A..... 174 ................4 Predição das viscosidades de uma mistura quaternária.......... Experimental data from Radosta et al..... •. PEG 1500.. ▲.......... —...................3 Viscosidades calculadas pelo modelo GC-UNIMOD e experimentais a 298.................. PEG 600......15 K e várias concentrações ...... PEG 1000.................................... 157 Figure 6.... valores calculados .........1 Viscosidades cinemáticas experimentais e calculadas para os PEGs 8000 e 10 000 e simulação de viscosidades calculadas para PEGs 6000............................................. 168 Figura A.. □................................15 K (method 1)...........................................13 Results of the predictions (method 1) for aw of maltodextrin fractions (n is the polymerization number)..

aminoácidos. Os compostos estudados foram: polióis. VERS e UNIMOD foram empregados em diferentes sistemas para a correlação e predição de propriedades físico-químicas. maltodextrinas e polietileno glicóis (PEG)s. de posse dos parâmetros de interação entre grupos. e mais especificamente naqueles que possuem compostos biológicos como açúcares. espalhamento de luz. Também são apresentados resultados da correlação e predição de viscosidades de sistemas aquosos contendo PEGs por um modelo semi-empírico que não trata da contribuição de grupos mas considera a hidratação das moléculas desses polímeros em meio aquoso. Neste trabalho. depressão do ponto de congelamento. o uso de métodos de contribuição de grupos para a correlação e predição do equilíbrio de fases tem crescido nos últimos anos.xxiii RESUMO Métodos de contribuição de grupos têm sido ferramenta útil no cálculo de coeficientes de atividade nos mais variados sistemas. pH. testar os métodos de contribuição de grupos para o cálculo de propriedades físico-químicas tratando também das particularidades de cada sistema estudado. . e o PEG é bastante utilizado em sistemas aquosos bifásicos para o estudo do equilíbrio e separação de biomoléculas. Propriedades físico-químicas como atividade de água. sais orgânicos.. como a dissociação parcial dos aminoácidos. UNIFAC. entalpia de diluição e viscosidade. é possível calcular propriedades de sistemas mais complexos a partir de parâmetros obtidos da correlação a dados experimentais de sistemas mais simples. Em sistemas aquosos. os métodos de contribuição de grupos ASOG. solubilidade. A intenção deste trabalho foi então. sendo os três primeiros compostos geralmente encontrados em sistemas alimentícios. são algumas das propriedades determinadas e/ou retiradas da literatura usadas para a correlação e predição empregando os modelos acima citados. etc. aminoácidos. a aplicação de métodos de contribuição de grupos torna-se atraente partindo-se do princípio que. os sistemas são geralmente multicomponentes e por essa razão. o efeito de proximidade dos grupamentos hidroxilas nos polióis e a polidispersão das maltodextrinas. Em alimentos e em biotecnologia. polímeros.

xxiv .

the systems are generally multicomponent. In the food and in the biotechnology areas.. The studied compounds were: polyols. proximity effect of hydroxyl groups in polyols. . and polydispersity of maltodextrins. freezing point depression. the objective of this work was to test the group-contribution models for calculating physical-chemical properties considering the particularities of each system such as partial dissociation phenomena in aqueous amino acid systems. In aqueous systems. UNIFAC. Thus. and PEGs are wide used in aqueous-two-phase systems to study equilibrium and separation of biomolecules. the results of viscosity correlation and prediction in aqueous PEG solutions by a semi-empirical equation that does not consider group contributions but take into account the hydration of the polymer molecules in aqueous media. the group-contribution models ASOG. pH. amino acids. maltodextrins and polyethylene (glycols) (PEG)s. the use of group-contribution methods becomes an attractive tool considering the possibility of calculation of properties in complex systems with parameters adjusted for simple ones. specifically those containing biological compounds such as sugars. organic salts. the use of group contribution models for correlating and predicting phase equilibrium has increased in the last years. polymers. etc. VERS and UNIMOD were utilized in different types of systems for the correlation and prediction of physical chemical properties. and for this reason. amino acids. It is also presented in this work. solubility. enthalpy of dilution and viscosity are some of the properties experimentally determined and/or found in literature used in this work for the correlation and prediction using the above cited models. Physical-chemical properties such as water activity.xxv ABSTRACT Group contribution methods have been used as a useful tool for calculating activity coefficients in different types of systems. which the first three compounds are found in food systems. In this work. light scattering.

Capítulo 1 1 Capítulo 1 REVISÃO BIBLIOGRÁFICA .

2 Capítulo 1 .

Métodos de contribuição de grupos vêm sendo bastante empregados para descrever as propriedades termodinâmicas de misturas de compostos de interesse da indústria alimentícia. etc. têm sido empregados em simulação de processos industriais como a destilação e na modelagem de propriedades termodinâmicas de sistemas mais complexos contendo compostos orgânicos. Dentre os modelos de contribuição de grupos descritos na literatura. . diferenças de tamanhos de moléculas entre compostos homólogos. dissociações. são apresentados os resumos dos capítulos deste trabalho que se referem a trabalhos publicados e a serem submetidos a revistas científicas da área. como por exemplo: açúcares. da presença de compostos muito diferentes nas misturas.Introdução 3 Introdução Métodos de contribuição de grupos são ferramentas úteis na modelagem e predição do equilíbrio de fases em variados sistemas. Apesar de serem encontrados na literatura valores de parâmetros desses modelos para vários grupos que formam as moléculas de compostos orgânicos e inorgânicos. além de parâmetros estruturais que podem ser prontamente calculados ou retirados de tabelas na literatura. se existem parâmetros disponíveis para os grupos presentes em uma mistura. Dessa forma. e também das particularidades dos compostos biológicos. efeitos de proximidade de grupos polares. A seguir. a de necessitar apenas de um pequeno número de parâmetros ajustáveis a serem conhecidos. Também segue o resumo de um anexo com alguns resultados que não foram publicados mas que são também parte deste trabalho. Esses modelos possuem vantagens. polímeros e aminoácidos. Por essa razão. bem como polímeros empregados em separação de compostos biológicos através de sistemas aquosos bifásicos (polietileno glicóis e maltodextrinas). A proposta deste trabalho foi justamente aplicar métodos de contribuição de grupos para descrever propriedades de sistemas aquosos contendo compostos presentes em alimentos (polióis. como por exemplo. farmacêutica e de processos biotecnológicos. é possível realizar a predição dos coeficientes de atividade dos compostos nessa mistura ou de sistemas mais complexos. sais orgânicos. ASOG e VERS que são baseados em expressões para a energia livre de Gibbs. o emprego desses parâmetros em sistemas contendo compostos biológicos muitas vezes não gera bons resultados por causa da complexidade dos sistemas envolvidos. Por isso. como hidratação. pode-se citar os modelos UNIFAC. aminoácidos e maltodextrinas). um ajuste dos parâmetros dos modelos torna-se muitas vezes necessário para uma melhor descrição das propriedades físico-químicas de tais sistemas.

Algumas propriedades de soluções aquosas contendo aminoácidos como atividade de água. 3. ao contrário dos outros capítulos. 2. Capítulo 1. de viscosidades cinemáticas de misturas contendo polietileno glicóis.4 Capítulo 1 1. temperatura e concentração. Capítulo 2. assim como outros dados de equilíbrio apresentados na literatura. pelo modelo GC-UNIMOD. A justificativa de reajuste de parâmetros é que a presença de grupos polares muito próximos (grupos hidroxilas consecutivos ligados à cadeia de carbonos da molécula) podem provocar um efeito de proximidade intramolecular. sendo que para o modelo UNIFAC foram testadas duas diferentes formas de divisão de grupos para as moléculas envolvidas. Na correlação. As viscosidades de misturas binárias foram também empregadas no cálculo da energia de ativação para o escoamento viscoso. mas sim uma correlação semi-empírica que descreveu bem as viscosidades considerando o grau de hidratação das diferentes moléculas estudadas . Revisão Bibliográfica. Capítulo 3. determinadas neste trabalho como função da temperatura. O trabalho apresenta resultados de correlações empregando os modelos ASOG e UNIFAC. dados de solubilidade dos poliálcoois foram também utilizados. Determinação experimental e modelagem da atividade de água em sistemas contendo quatro diferentes polióis. Determinação de viscosidades cinemáticas de misturas aquosas contendo polietileno glicóis numa ampla faixa de massas molares. pH e densidade foram determinadas em três diferentes tipos de solventes. e os modelos de coeficiente de atividade e de outras propriedades de interesse neste trabalho. apresentando os fundamentos termodinâmicos do equilíbrio de fases e de outras propriedades físico-químicas. Capítulo 4. Na correlação. Capítulo 5. A dissociação parcial desses compostos em água e em dois tampões foi considerada nas correlações. Correlação e predição. 5. não foi empregado um modelo de contribuição de grupos. 4. O modelo UNIFAC-Lyngby de contribuição de grupos combinado com a equação de Debye-Hückel foram empregados na correlação e modelagem das propriedades determinadas neste trabalho. As atividades de água de misturas ternárias foram empregadas em teste para avaliar a capacidade preditiva dos modelos com os novos parâmetros ajustados.

apresentada no Capítulo 5. contendo polietileno glicóis. Os parâmetros do modelo VERS foram também empregados na predição da atividade de água. que gerou resultados consideravelmente melhores. maltose. Determinação de atividade de água. Anexo 1. optou-se por outra abordagem.Introdução (correlação de Kumar). Por esse motivo. Apesar das diversas tentativas para melhorar os resultados de correlação e predição. espalhamento de luz e entalpia de diluição em sistemas contendo três diferentes maltodextrinas e correlação dos dados por equação osmótica virial e o modelo VERS (Virial Equation with Relative Surface Fractions). maltotriose. componentes foi satisfatória. depressão do ponto de congelamento e de ebulição em sistemas contendo sacarídeos como glicose. etc. Capítulo 6. pelo modelo GC-UNIMOD. Resultados da correlação e predição das viscosidades cinemáticas de misturas aquosas. o resultado geral foi insatisfatório. 5 A predição das viscosidades de misturas com até cinco 6. . 7.

Com base em um banco de dados que incluiu os dados experimentais determinados neste trabalho e obtidos na literatura. entalpia de diluição. Foi estudado o comportamento de sistemas contendo polióis. Esses compostos possuem diversas aplicações industriais. depressão do ponto de congelamento. poucos dados físicos e termodinâmicos para esses compostos estão disponíveis na literatura. os métodos de contribuição de grupos UNIFAC. incluindo usos como modificadores de textura. elevação do ponto de ebulição e viscosidade de soluções contendo componentes presentes em alimentos e empregados em processos biotecnológicos. plasticizantes e outros.6 Capítulo 1 Objetivos Este trabalho tem por objetivos determinar. aminoácidos e polímeros naturais em misturas com dois ou mais componentes. . modelar e predizer propriedades físico-químicas como atividade de água (aw). ASOG e o modelo VERS foram utilizados na modelagem de propriedades termodinâmicas através do ajuste de parâmetros dos modelos – quando necessário – e tratamento matemático adequado para cada conjunto de dados. pH. umectantes. além de suas propriedades em solução serem importantes no projeto de equipamentos e processos. solubilidade. Apesar de importantes.

e nos capítulos correspondentes a cada composto.n1 . A energia livre de Gibbs G depende das variáveis temperatura.1. . uma quantidade de estado extensiva do sistema:  ∂G  −S =   ∂T  P . 1.1 Fundamentos e modelos termodinâmicos Neste capítulo serão comentados os fundamentos termodinâmicos e modelos que neles se baseiam para melhor compreensão da modelagem. pressão e número de mol dos componentes que formam a mistura.1 Bases Termodinâmicas Os modelos baseados na energia livre de Gibbs de excesso são normalmente empregados para a descrição do equilíbrio de fases.n N ) (1.. aminoácidos e polímeros – os procedimentos no tratamento e cálculo das propriedades serão apresentados em detalhe para cada sistema nos itens abaixo.3) e o volume. P . Como foram estudados diferentes compostos – poliálcoois. Revisão Bibliográfica 1. assim como do cálculo do equilíbrio de fases dos sistemas aquosos estudados. n j (1.1) Na forma diferencial a equação fundamental de Gibbs é dada por: dG = − SdT + VdP + ∑ µ i dni i =1 N (1.2) em que S é a entropia.Fundamentos e modelos termodinâmicos 7 1.. n2 .4) e o potencial químico µi é a energia livre de Gibbs parcial molar do componente i: .n j (1. G = G( T . também uma quantidade extensiva é dado por:  ∂G  V =  ∂p    T .

7) Das equações 1.5) Já que a função G é homogênea de primeira ordem em massa ou número de mol. pode-se aplicar o teorema de Euler para obter a seguinte função: N  ∂G   = ∑ ni µ i = G  i =1  i T .6) e. P = constante ⇒ G = mínima (1.9) N Existe equílibrio termodinâmico de fases em um sistema – que não possui interações com a vizinhança – quando não se observa macroscopicamente mudanças nas suas variáveis num determinado período e a inexistência de fluxos líquidos.n j ≠i (1.7 obtém-se a equação de Gibbs-Duhem.10) Partindo-se da condição de equilíbrio de fases apresentada na equação 1.2 e 1. tem-se que: . A condição para o equilíbrio de fases de um sistema à temperatura e pressão constantes pode ser escrita com o auxílio da segunda lei da termodinâmica como: T .8) e considerando um sistema a temperatura e pressão constantes.8 resume-se a: i =1 ∑ ni dµ i = 0 (1.P . a equação 1.8 Capítulo 1  ∂G  µi =   ∂n    i T .P . dG = ∑ µ i dni + ∑ ni dµ i i =1 i =1 N N (1.n j ≠i i =1 ∑ ni  ∂n  N (1. uma relação fundamental da termodinâmica: SdT − VdP + ∑ ni dµ i = 0 i =1 N (1.10 acima.

. .15) E em que γi é o coeficiente de atividade do componente i e µ i = ∂G E ∂ni P .Fundamentos e modelos termodinâmicos dG = 0 e 9 (1. O termo de excesso – que inclui os coeficientes de atividade – descreve o comportamento real do componente. a energia de Gibbs é função do potencial químico.. Numa solução ideal.2 Modelos para energia livre de Gibbs Como foi apresentado na equação 1.1. um resumo sobre os modelos para o cálculo de coeficientes de atividade. Já o potencial químico de um componente i numa solução real é calculado como a soma de duas contribuições. = P π π µ i ′ = µ i ″ = µ′′′ = . o potencial químico do componente i é dado por: o µ id = µ i + RT ln xi i (1.. à mesma temperatura e pressão da solução e xi é a fração molar do componente i. = µ i i em que o sobrescrito refere-se a fase e o subescrito ao componente do sistema. os coeficientes de atividade possuem um significado decisivo na modelagem do equilíbrio de fases..12) d 2G > 0 que permite concluir que o equilíbrio corresponde a: T ′ = T ′′ = T ′′′ = . Por isso será apresentado a seguir. uma ideal e outra de excesso..T .n o termo de j ( ) excesso.. Neste trabalho serão investigados alguns modelos que relacionam a energia livre em excesso para o cálculo de propriedades físico-químicas em sistemas aquosos contendo diferentes componentes.2. = T π P′ = P′′ = P′′′ = . 1.13) 1. como a seguir: o µ i = µ i + RT ln xi + RT ln γ i 14 244 1 24 4 3 4 3 id µi µ iE (1.14) o onde µ i é o potencial químico do componente i puro. (1. Dessa forma.11.

o volume e a entropia de excesso podem ser considerados desprezíveis em misturas apolares constituídas por compostos semelhantes. outro a teoria da solução atérmica e um terceiro grupo que se utiliza das duas teorias anteriores como base para compor os modelos. Extensões desse tipo de abordagem são as equações de Margules (Prausnitz. uma solução regular pode exibir apenas desvio positivo da idealidade segundo a Lei de Raoult. Abordagem teórica Aqui pode-se citar três diferentes grupos de modelos: um que emprega a teoria da solução regular. considerações da teoria da solução regular. isto é. 2. de acordo com suas abordagens fundamentais. Essas são VE =0 E S4 =3 0 1 24 G E ⇓ =U E (1. Abordagem empírica Capítulo 1 Os modelos de energia livre de Gibbs em excesso podem ser divididos entre empíricos e Com relação à abordagem empírica poder-se-ia citar uma série de trabalhos que se utilizaram de ajuste polinomial aos dados experimentais para obter uma equação simples que estime os coeficientes de atividade.10 teóricos.17) A teoria de soluções regulares chamada também teoria de Scatchard-Hildebrand (Prausnitz. 1969) e de Redlich e Kister (Redlich and Kister. Nessa abordagem . (a) Teoria da solução regular A energia livre de Gibbs pode também ser escrita – com base nas tranformadas de Legendre – da seguinte forma: G E = H E − TS E = U E + PV E − TS E (1. 1. Um exemplo desse tipo de ajuste é a abordagem de Porter (1920) que leva ao cálculo de coeficientes de atividade de compostos bem semelhantes numa mistura. 1986) fornece coeficientes de atividade maiores ou iguais a 1.16) De acordo com a experiência. 1948).

Alguns desses modelos foram estendidos para compor os chamados modelos de contribuição de grupos. É denominada atérmica. os coeficientes de atividade são independentes da temperatura. (c) Abordagem combinada Essa abordagem utiliza-se das teorias das soluções regulares e atérmicas. pois o excesso de entalpia é desprezado. H E = 0 ⇒ G E = −TS E (1.19) H E (1. (b) Teoria da solução atérmica 11 considera-se que a interação entre duas moléculas em solução depende apenas da distância Ao contrário da teoria de soluções regulares. a teoria da solução atérmica não considera interações entre moléculas ou espécies químicas em soluções e sim efeitos de configuração. Tais modelos consideram que o comportamento dos compostos em solução é o resultado das interações e das diferenças de tamanho e forma dos grupos funcionais dessas substâncias. Pode-se citar também nesse grupo o modelo VERS (Virial . Um exemplo típico desse grupo é o modelo UNIQUAC (Universal Quasi- Chemical Equation) proposto por Abrams and Prausnitz (1975).Fundamentos e modelos termodinâmicos entre essas moléculas sendo aleatória a distribuição das moléculas em posição e orientação. e dessa forma.18)  ∂G E / RT  HE   =−   ∂T RT 2   p. Um exemplo desse tipo de abordagem é o modelo de Flory-Huggins (Flory. n j N GE =0⇒ = ∑ ni ln γ i ≠ f (T ) RT i =1 (1. considerando então efeitos de interação e configuração nos modelos para o cálculo dos coeficientes de atividade.20) A entropia de excesso pode ser calculada com o auxílio de modelos que consideram as diferenças de tamanho entre as moléculas num sistema. Os modelos desse tipo mais conhecidos são o ASOG (Analytical Solution Of Groups) de Kojima and Tochigi (1979) e o modelo UNIFAC (UNIQUAC Functional-Group Activity Coefficient) de Fredenslund and Jones (1975). 1941) que prediz desvios negativos da idealidade segundo a lei de Raoult.

2. 1.21) combinatorial residual I. para o componente i de uma mistura: lnγ i = ln γ C i + ln γ iR (1. Modelo UNIFAC Nesse modelo. estão apresentadas as equações dos modelos de contribuição de grupos empregados neste trabalho: UNIFAC. VERS e GC-UNIMOD (esse último baseado no modelo de contribuição de grupos UNIFAC. são apresentadas as equações desses modelos e também do modelo VERS (Virial Equation with Relative Surface Fractions). o cálculo do coeficiente de atividade de um componente é dividido em duas partes. Parte combinatorial No modelo UNIFAC. A seguir.. 1993). utilizado para o cálculo de viscosidades de misturas e desenvolvido por Cao et al.22) . Apenas propriedades dos compostos puros entram nesta parte da equação: φ θ φ z ln γ C = ln i + qi ln i + li − i ∑ x j l j i xi 2 φi xi j (1. o termo combinatorial da equação UNIQUAC é usado diretamente.12 Capítulo 1 Equation with Relative Surface Fractions) desenvolvido por Großmann (1994) e baseado no desenvolvimento virial da equação de Pitzer (Pitzer. como no modelo UNIQUAC: uma parte fornece a contribuição devido às diferenças no tamanho e na forma molecular e a outra está relacionada às interações energéticas entre grupos como mostrado abaixo.1 Métodos de contribuição de grupos Para cálculo de coeficientes de atividade O cálculo de coeficientes de atividade por contribuição de grupos foi primeiramente sugerido em 1925 por Langmuir (Prausnitz. 1986) mas o emprego de tais métodos expandiu-se com o aumento do número de dados experimentais e o surgimento de modelos como o ASOG (Analytical Solutions of Groups) e o UNIFAC (Universal Functional Activity Coefficient).2 1. Abaixo. 1991) para soluções de eletrólitos. ASOG.

Métodos de contribuição de grupos 13 z li = ( ri − qi ) − ( ri − 1 ) 2 (1. para líquidos em condições moderadas.24. o valor de z é próximo de 10. 1. Os fatores de normalização 15. Geralmente.30) Γk é o coeficiente de atividade residual do grupo k e Γki o coeficiente de atividade residual do grupo k numa solução referência contendo apenas moléculas do tipo i.17 Qk = Ak / 2. II.26.17 e 2.25) em que θi e φi são frações de área e volume do componente i. Parte Residual A parte residual do modelo UNIQUAC é substituída no modelo UNIFAC pelo conceito de solução de grupos. dados por Bondi (1968): i ri = ∑ vk Rk k i qi = ∑ vk Qk k (1. 1962): i i lnγ iR = ∑ vk [lnΓk − lnΓk ] k (1.29) em que vki é um número inteiro que representa o número de grupos do tipo k na molécula i. Os parâmetros ri e qi do componente puro são calculados como a soma dos parâmetros de volume (Rk) e de área (Qk) do grupo k na molécula i. respectivamente. Γk pode ser calculado como: .5×109 são dados por Abrams and Prausnitz (1975). θi = qi xi ∑q jx j j φi = ∑rj x j j ri xi (1. são obtidos a partir do volume e área de van der Waals Vk e Ak.5 × 109 (1.27) Rk = Vk / 15. 1. A solução é tratada como uma mistura de grupos funcionais com interações energéticas (Wilson and Deal.28. 1. por sua vez.23) onde z é chamado de número de coordenação e representa o número de locais nos quais pode haver interação entre moléculas. Estes.

é dada por: . e amn representa os parâmetros binários de interação de grupos com unidades em Kelvin e amn≠anm.34) Umn é uma medida da energia de interação entre os grupos m e n.14      ln Γk = Qk 1 − ln ∑ θ m ψ mk  − ∑  θ m ψ km ∑ θ m ψ nm       n m  m   i Capítulo 1 (1. (1986).    v iFH ln γ iFH = 1 + ln  n FH  ∑ x jv j  j =1     viFH − n  FH ∑ xi vi   j =1  (1.32. e é dado por:   Umn − Unn  ψ nm = exp −   = exp (− a nm T ) RT    [ ] (1.36) A parte residual.33) em que θm é a fração de área do grupo m. Modelo ASOG No modelo ASOG o coeficiente de atividade de um componente i na solução é calculado pela soma das contribuições devido às diferenças nos tamanhos moleculares (γiFH) e das interações entre grupos (γiG) que descrevem as diferenças de forças intermoleculares: lnγ i = ln γ iFH + lnγ G i (1. que também contribui para o cálculo do coeficiente de atividade. 1.31) θm = Qm X m ∑ Qn X n n ∑ vm xi Xm = i i ∑ ∑ vk xi i k (1. Xm é a fração do grupo m na mistura e ψnm é o parâmetro de interação entre os grupos m e n.35) A contribuição devido às diferenças nos tamanhos das moléculas é calculada aplicando-se equação similar à equação de Flory e Huggins citados por Prausnitz et al.

37) O coeficiente de atividade do grupo k (Γk) é função da temperatura e da fração de cada grupo na solução. O excesso é definido pela normalização assimétrica como mostrado abaixo: xw → 1 xi → 0 γw →1 γi → 1 para o solvente para os solutos (1.39) n ln akn = mkn + kn T (1. vki o número de átomos não hidrogênio no grupo k da molécula i.41) (1. xj a fração molar do componente j.40) em que viFH é o número de átomos não hidrogênio na molécula i. Γk e Γki são os coeficientes de atividade do grupo k no sistema e no estado padrão (componente i puro).42) . mkn e nkn são parâmetros independentes da temperatura característicos das interações entre os grupos k e n. O parâmetro akn representa as interações energéticas entre os grupos k e n (akn≠ank).38) ∑ x j vn Xn = j j ∑ x j ∑ vk j k i (1. Modelo VERS Neste trabalho.Métodos de contribuição de grupos i i ln γ G = ∑ vk [ln Γk − ln Γk ] i k 15 (1. Como de costume. ele é definido analiticamente pela equação de Wilson (1964). respectivamente. o modelo começa com a energia livre de Gibbs em excesso GE. um modelo semi-empírico é empregado para descrever os dados de atividade de água e entalpia de diluição. como a seguir: ln Γk = − ln ∑ X n akn + 1 − ∑ n n X n ank ∑ X m anm m (1.

Então. Esse coeficiente é multiplicado pela probabilidade dessa interação. que é considerada como sendo o quadrado da concentração de soluto.44) resulta: 2  1000 Θ1  GE  A = nw RT  M w Θ w  11   (1.44) Θi = ∑ todos os componentes j m jq j (1.47) devem ser feitas para todas as espécies de solutos presentes. Das equações (1. A concentração do soluto é expressa como: conc.43) A11 representa a interação entre duas moléculas do componente 1 dissolvidas em água.16 Capítulo 1 A influência da pressão sob GE é desprezível. o parâmetro de área qi é calculado por método de contribuição de grupos: .45) em que Θ1 é a fração de área do componente 1. da seguinte forma: 3  1000   1000  Θi Θ j Θi Θ j Θ k GE  ∑ ∑  ∑ ∑ ∑ = Aij +  Bijk  M  nw RT  M w  i ≠ w j ≠ w Θ w Θ w   w  i ≠ w j ≠ w k ≠ w Θw Θw Θw 2 (1.46) A expressão (1. quando um componente 1 não iônico é dissolvido em água. mi a molalidade do componente i e qi o parâmetro de área do componente i.46) pode ser facilmente estendida para interações ternárias assim como para mais solutos.47) As somas na equação (1.43)-(1. o excesso da energia livre de Gibbs é expressa da maneira mais simples: GE = A11[conc.(1) = 1000 Θ1 M w Θw mi qi (1. Para considerar a influência da massa molar do soluto.(1)]2 nw RT (1. Θw a fração de área da água.

Para descrever os dados calorimétricos e a influência da temperatura na atividade de água. Esse modelo foi denominado VERS (Virial Equation with Relative Surface Fractions). ou seja.Métodos de contribuição de grupos 17 qi = (i ) ∑ vl Ql todos os grupos l (1. Assume-se que todos os parâmetros são (j simétricos.49) Bijk = ∑ ∑ ∑ ( Θl i ) Θ ( j ) Θ ( k ) blmn m n todos grupos todos grupos todos grupos l m n (1. pode-se então escrever equações para a atividade de água (aw) e atividade do soluto ( a* i. ml A partir das equações acima. alm e blmn são parâmetros de interação de grupos. que apresenta a influência da temperatura nos parâmetros de interação binários alm foi empregada: (0 (1) (2 alm = βlm) + βlm (T / K )(1 − (T0 / T )) + βlm) ln(T / T0 ) (1. Qt é o parâmetro de área do grupo l.51) O parâmetro vl(i) representa o número de grupos l na molécula de soluto i.m ): .52) em que T0 é a temperatura de referência igual a 25 oC. a seguinte expressão empírica.48) e para os parâmetros Aij e Bijk: Aij = ∑ ∑ ( Θ l i ) Θ ( j ) alm m todos grupos todos grupos l m (1.50) ( ( Q Θl i) = ν l i) l qi (1. que βlm) = β ( j ) e blmn = bln m = bm ln = bmnl = bnlm = bnml .

2.2 Para cálculo de viscosidade de misturas Modelo GC-UNIMOD O modelo GC-UNIMOD proposto por Cao et al.56) e do soluto: hiE≠ w 2 Θ j ∂Aij  1000  qi   = −2T ∑ M  q RT  w  w j ≠ w Θ w ∂T (1.55) E hw 2  1000  Θi Θ j ∂Aij   ∑ ∑ =T M  RT  w  i ≠ w j ≠ w Θ w Θ w ∂T (1. Nesse .54) assim como para a entalpia molar de excesso da água:  ∂( µ E / T )    E i hw = −T 2   ∂T     p .58) 1.53) 2 Θj  1000   1000  qi   = ln mi + 2  M  q ∑ Θ Aij + 3 M   i.n j (1. (1993) é baseado no método de contribuição de grupos UNIFAC. m  1  1 j ≠1 1  1  ln a* Θj Θ kB q1 j ≠ 1 k ≠ 1 Θ1 Θ1 ijk qi ∑ ∑ (1.57) com: ∂Aij ( 2)  (i ) ( j )  (1) β lm = Θ Θ β + ∑ ∑ ∂T todos grupos todos grupos l m  lm T   l m      (1. O GC-UNIMOD é também um modelo de contribuição de grupos funcionais empregado no cálculo da viscosidade de misturas líquidas multicomponentes.18 Capítulo 1 2  1000  Θi Θ j Θ Θ Θ M1 1000   ∑ ∑ ∑ i j k B ln a w = − mi − Aij − 2 ∑ ∑ ∑ ijk  M  1000 i M 1 i ≠ 1 j ≠ 1 Θ1 Θ1  1  i ≠ 1 j ≠ 1 k ≠ 1 Θ1 Θ1 Θ1 (1.

ξiC à contribuição combinatorial e ξiR à residual do componente i. Rk.59) I. são normalmente Segue modelo. 19 Para os cálculos. parâmetro de volume do grupo k. vki. vi e φi são respectivamente. fração molar do componente i na mistura. II. ln( ν ) = ∑ (ξ C + ξ iR ) i i =1 n (1. viscosidade cinemática do líquido i e fração de volume do componente i.62) em que Mi. massa molar da mistura. M. ri. . xi. abaixo uma apresentação das equações matemáticas correspondentes ao modelo GC-UNIMOD para o cálculo da viscosidade cinemática em que: ν corresponde à viscosidade cinemática da mistura. número de segmentos na molécula i. número de grupos k na molécula i. a viscosidade de uma mistura pode ser calculada como a soma de duas partes como no utilizados parâmetros do UNIFAC para o equilíbrio líquido-vapor (UNIFAC-VLE). Parte Residual i ξ iR = ∑ vk [Ξ ki − Ξ i ] ki k (1. massa molar do componente i.63) Ξki e Ξiki correspondem respectivamente à viscosidade residual do grupo k para o componente i quando numa mistura e à viscosidade residual do grupo k no componente i quando em uma solução contendo apenas o componente i.60) (1.Métodos de contribuição de grupos modelo UNIFAC: uma combinatorial e uma residual.61) xi ri φi = ∑ x jrj j =1 n (1. Parte Combinatorial x  M  C ξ i = φi ln ν i i  + 2φi ln i  M  φ    i i ri = ∑ vk Rk k     (1.

Chumpitaz et al. fração de área do grupo k e composição local dos grupos.65) a  Ψmn = exp mn   T  X m Qm (1. parâmetro de interação binária entre os grupos m e n.20 Q visc Ξ mi = − m N mi φi ∑ θ km ln ψ km Rm k θ mn = θ m Ψmn Capítulo 1 ( ) (1. Xk ψmn. θk. fração molar do grupo k. Alguns trabalhos do grupo de pesquisa em que esta fase foi realizada estão relacionados com a determinação experimental e modelagem de propriedades físicas.67)  q − r 1 − ri  visc  N ki = Qk  i i −  2 z    i qi = ∑ vk Qk k (1.68) (1.70) As viscosidades dos compostos puros requeridas pelo modelo GC-UNIMOD podem ser retiradas da literatura (quando houver dados) ou calculadas por modelos do tipo proposto por Joback and Reid (1987). n  x  M  ln(ν) = ∑ φi ln ν i i  + 2φi ln i  M  φ   i = 1  i     + ∑ v i  Ξ − Ξ i   k  ki ki     k   (1. (1999) e Cruz et al. (2000). θmn correspondem a: parâmetro de interação binária entre os grupos m e n. com as expressões matemáticas para as contribuições combinatorial e residual. qi. parâmetro de área para molécula i. A equação abaixo representa o cálculo global da viscosidade. como por exemplo: Valeri and Meirelles (1997).69) amn. Trata-se . parâmetro de área superficial para o grupo k.64) ∑ θ k Ψkn k (1.66) θm = ∑ X k Qk k (1. Qk.

73) Seguindo a proporcionalidade entre viscosidade e concentração proposta por Harned and Owen (1965): (η rel . 1993) e posteriormente.74 na equação 1. o número de moles de sal. também empregada no cálculo de viscosidades de misturas de sais e açúcares (Pereira.73: xc = x1 (1 − x1 h) (1.74) (1. empregadas nas investigações da viscosidade de misturas de polietileno glicóis no Capítulo 5. Primeiramente a correlação foi proposta para o cálculo de viscosidades de soluções de sais (Kumar.76) . A equação de Kumar para viscosidade foi desenvolvida assumindo que um sal AB. Se N1. Equação de Kumar 21 agora de estender tal abordagem a misturas multicomponentes que de fato são encontradas na Apesar de não se tratar de um modelo de contribuição de grupos. pode-se obter N2 da equação 1. forma um complexo solvatado como.x1h) (1. 2001). solvente e complexo solvatado.75) (1.72) (1. respectivamente. as equações da correlação de Kumar.1) ∝ xc = B x1 (1 . N2 e Nc são.Métodos de contribuição de grupos prática industrial. são apresentadas neste item. quando dissolvido em um solvente S. as frações molares do sal (x1) e do complexo solvatado (xc) podem ser escritas como: x1 = N1 ( N1 + N 2 ) xc = N c [N c + ( N 2 − h N c )] Assumindo que N1=Nc. a seguir: AB + h S ⇔ AB ⋅ hS (1.72 como: N 2 = N c (1 .x1 ) x1 Substituindo N2 da equação 1.71) em que h é o número de moléculas de solvente.

Dentre muitos materiais comumente encontrados em biotecnologia e na indústria de alimentos estão polímeros. recuperação de biomoléculas ou mesmo desenvolvimento de produtos é geralmente necessário o conhecimento de propriedades físicas dos mesmos para a otimização de operações na indústria. ηrel=η/η0. Para o projeto de processos de purificação. B. O . o estudo do equilíbrio de fases desempenha papel importante em processos de extração líquido-líquido como nos sistemas aquosos bifásicos. O conhecimento de propriedades termodinâmicas de sistemas aquosos simples pode levar a uma melhor compreensão de biomoléculas em seus ambientes naturais. em particular na área biotecnológica e na indústria de alimentos. também interessante e muito utilizado nas mais diversas áreas industriais e investigações científicas – como. triglicerídeos. e também na tecnologia de alimentos já que tanto os produtos naturais quanto os processados apresentam microestrutura complexa e multifásica. 1. açúcares. por exemplo. etc.76 pode também ser reescrita como: 1/(ηrel . Além de compostos orgânicos. A equação 1. estudos da atividade de água (equilíbrio líquido-vapor) como parâmetro de controle de qualidade em alimentos.1) = −h / B + 1 / (B x1 ) (1. Dentro desse contexto. ácidos orgânicos. O parâmetro h indica a solvatação ou hidratação. deve-se enfatizar a importância do estudo do comportamento de compostos alimentícios em soluções aquosas. respectivamente e B é uma constante. processos de cristalização e crioconcentração que envolvem equilíbrio sólido-líquido. Dessa forma. representa as interações entre íon e solvente como discutido por Feakins and Laurence (1966). se o solvente é água e o outro parâmetro. aminoácidos.3 Compostos orgânicos de interesse na área de alimentos e biotecnologia O crescente avanço tecnológico tem despertado interesse para o desenvolvimento de processos eficientes na indústria.77) A equação acima é a expressão proposta para correlacionar viscosidade com fração molar do sal na solução. SABs (técnica comumente empregada para separação e purificação de biomoléculas).22 Capítulo 1 em que. em sistemas aquosos bifásicos – são os polietileno glicóis. η e η0 são as viscosidades da solução e do solvente. outro material estudado neste trabalho.

pH. incluindo a de alimentos.. Existe atualmente um grande interesse em se produzir xilitol (poliálcool) por via microbiológica (Azuma et al. polietileno glicóis e maltodextrinas – para esclarecimento de suas estruturas químicas. por exemplo. são produzidos industrialmente a partir da hidrogenação catalítica de carboidratos comestíveis ou são produzidos por processos biológicos (Billaux et al. 2000.. o emprego de métodos para a estimativa de dados como solubilidade apresenta-se como uma ferramenta útil no planejamento de processos de recuperação de compostos . 1998). O poder adoçante é outra importante característica desses compostos. 1. etc. aminoácidos. soluções de sorbitol são usadas como umectantes e plasticizantes em algumas formulações de alimentos e filmes comestíveis (Moreton and Armstrong.1 Polióis Polióis são polihidróxi-álcoois empregados em diversas áreas. Silva et al. propriedades físico-químicas e empregos na indústria e pesquisa.3. A seguir. os aminoácidos são de grande interesse para a pesquisa devido à sua simplicidade e importância para o entendimento do comportamento de moléculas como peptídeos e proteínas em soluções mais complexas.. produtos de higiene e fármacos (Kirk and Othmer.como xilitol – produzidos por fermentação.Compostos orgânicos de interesse na área de alimentos e biotecnologia 23 conhecimento de propriedades físico-químicas tais como solubilidade. massa molar e parâmetros de interação soluto-solvente de misturas que incluem esses compostos é necessário sob o ponto de vista de projeto de equipamentos e processos e também para o melhor entendimento da formação da estrutura no alimento.2 Aminoácidos Dentre vários compostos bioquímicos. Em geral. Nos próximos itens. 1998). os aminoácidos são produzidos por microorganismos em meio aquoso contendo solutos como sais.3. ácidos orgânicos. ponto de congelamento. 1. são também descritos os métodos de cálculo das propriedades físico-químicas que se deseja estudar. atividade de água. 1991). . Na indústria de alimentos. que são utilizados na confecção de doces dietéticos. Dessa forma. a farmacêutica e a de cosméticos. Em geral. 1983). segue-se uma descrição rápida dos compostos que são utilizados neste trabalho – polióis. Do ponto de vista da simulação de processos. o uso de modelos matemáticos existentes e consolidados na literatura torna-se uma ferramenta útil para o cálculo das propriedades de interesse.

Coimbra et al. atividade de água e a influência do pH sobre essas propriedades em misturas contendo biomoléculas. O polímero é então caracterizado por um grau de polimerização médio. pode alcançar 90% do custo total de obtenção de bioprodutos. A estrutura química dos PEGs é representada por: HO − (CH 2 CH 2 O) n − H onde n representa o grau de polimerização. Dessa forma. Na área biotecnológica são empregados em sistemas aquosos bifásicos para a separação e purificação de biomoléculas em meios biocompatíveis. 1989).. 2000. Silva e Meirelles.3 Polietileno glicóis Polietileno glicóis (PEGs) são polímeros de cadeia linear formados por unidades de oxietileno. cosmética.. nos quais eles são geralmente produzidos. 2002). 1980. é interessante o estudo de propriedades físico-químicas de sistemas contendo tais compostos a várias concentrações e pHs. apresentam-se sólidos (MM 3000 a 20000). otimização e também projeto de equipamentos. b. cujas características principais são alta solubilidade em água. Estas propriedades têm favorecido o uso dos PEGs em diversas aplicações comerciais das indústrias farmacêuticas. 1980). o conhecimento do comportamento reológico de sistemas contendo PEGs é importante para avaliação. 1.24 Capítulo 1 Processos de separação baseados na precipitação e cristalização têm sido largamente utilizados para concentração e purificação de biomoléculas (Cussler et al. Comercialmente são identificados por . De acordo com Eyal and Bressler (1993). não é mais possível a separação de compostos puros por destilação devido à baixa volatilidade dos mesmos e. química e de alimentos. semi-sólidos (MM 1000 a 2000) ou líquidos viscosos (MM 200 a 700) à temperatura ambiente (Davidson. Dependendo do tamanho da cadeia. o custo de separação e concentração de biomoléculas a partir de meios aquosos diluídos. À medida que o grau de polimerização aumenta. baixa toxicidade e boa estabilidade. O polímero de mais alta massa molar produzido comercialmente como um composto puro é o tetraetilenoglicol com n=4. Sé e Aznar. já que o PEG possui baixa toxicidade (Chirife e Ferro Fontan. dessa forma. Dessa forma.3. No projeto de equipamentos e processos é necessário o conhecimento de algumas propriedades físico-químicas como solubilidade. 2000a. 1995.. Alves et al. são produzidas misturas de moléculas com várias massas molares.

78) ∑ mi M i Mw = i ∑ mi i (1. geralmente representada pelas massas molares médias: nominal (Mn) ou mássica (Mw). maltodextrina é apresentada a seguir: A estrutura química da . PEG Maltodextrinas são polissacarídeos solúveis em água que têm sido empregados numa ampla variedade de produtos industriais nas áreas de alimentos e farmacêutica. 1998). a necessidade de proteção do meio ambiente coloca os polímeros biodegradáveis. A ampla distribuição de massas molares confere às maltodextrinas variadas características relacionadas com propriedades físico-químicas como viscosidade. capacidade de formação de gel. 1993. Os polissacarídeos apresentam distribuição de massa molar.b). etc. as maltodextrinas são polímeros de grande interesse comercial devido à sua alta solubilidade em água (Gliksmann. 1. Marchal. (Kasapis et al.79) em que ni é número de moles e mi a massa da espécie i. Mothé e Rao. Essas massas molares são calculadas como: ∑ ni M i Mn = i ∑ ni i (1. 1986. Suas massas molares variam de 900 a 18000 g/mol. pressão de vapor. etc. 1999). As MD apresentam-se como uma mistura de sacarídeos com uma ampla distribuição de massas molares entre polissacarídeos e oligossacarídeos..4 Maltodextrinas (MD) 25 números que indicam suas massas molares aproximadas. como os polissacarídeos. 2000 a. Atualmente. PEG 400.3.Compostos orgânicos de interesse na área de alimentos e biotecnologia 1000. Dentre os polissacarídeos. A partir desses valores pode-se também definir o quociente Mw/Mn denominado índice de polidispersão que é uma medida da amplitude da distribuição. como por exemplo. 1999) e aplicação potencial em sistemas aquosos bifásicos (Silva e Meirelles. em vantagem para fins industriais (Swift.

a fase vapor comporta-se como um gás ideal e a razão de fugacidades pode então ser aproximada pela razão entre a pressão parcial da água numa amostra (Pw) e a pressão de vapor da água pura na mesma temperatura (Pwo). entalpia de diluição e viscosidade. em que o coeficiente de atividade γw pode ser calculado pelos métodos de contribuição de grupos. As equações termodinâmicas para o cálculo dessas propriedades são apresentadas a seguir. elevação do ponto de ebulição. O produto γwxw é a atividade da água numa mistura. e o tratamento termodinâmico mais detalhado nos capítulos seguintes referentes a cada um dos compostos estudados. P. métodos de contribuição de grupos foram empregados para a estimativa de propriedades como atividade de água. pH.1 Estrutura química da maltodextrina (x=ligação α-1.26 Capítulo 1 Figura 1.4 Propriedades físico-químicas estudadas No presente trabalho.1 Atividade de água O conceito termodinâmico de atividade de água pode ser relacionado à razão entre a fugacidade da água numa amostra (fw) e a fugacidade da água pura (fwo) à mesma temperatura e pressão: f (T . 1. solubilidade. .6 com possibilidade de ramificação. y= ligação α-1. xw ) Pw aw = xw γ w = w = o o f w (T .4 para a formação de cadeia linear). P ) Pw (1.4. depressão do ponto de congelamento. 1. espalhamento de luz.80) Em condições moderadas de temperatura e pressão.

Considerando a dependência linear de ∆Cp com a temperatura podese calcular a atividade do soluto como:  ∆H f ∆A − ∆BTref ∆B 2  1 1 ln γ i xi = − + Tf + T f  − R R 2R   T T f   ∆A − ∆BTref  T  ∆B + ln (T −T f )  T  2R R  f   +   (1. xi ) (1. Peres and Macedo. 1994. a equação resultante para predizer a concentração de saturação do sólido i num líquido pode ser expressa como: ln γ i xi = − ∆H f  T f  ∆C p  − 1 +  RT f  T R   Tf  ∆C p  T f  − 1 − ln  T   T R        (1.84 calcula-se a solubilidade xi.4.82) A consideração que ∆Cp é uma função linear da temperatura já foi utilizada em trabalhos prévios (Catté et al. 1996. pode-se escrever: f iS (T . sabendo-se as propriedades do componente i puro e o coeficiente de atividade de i em solução. Esse cálculo requer um procedimento iterativo .83) em que Tref é uma temperatura de referência.81) Assumindo que a em que. P) = f iL (T . A expressão representando a diferença entre os calores específicos do líquido e do sólido (∆Cp) é dada por: ∆C p = ∆A + ∆B(T − Tref ) (1. e T a temperatura do sistema. Considerando que não há solubilidade do solvente na fase sólida. 1997) para sistemas contendo açúcares. temperatura do ponto triplo é aproximadamente a temperatura de fusão do sólido (Tf) e que ∆Cp apresenta dependência linear com a temperatura numa faixa entre T e Tf.2 Solubilidade 27 O critério de cálculo da solubilidade de um componente sólido i em solução é a igualdade de fugacidades de i nas fases líquida (L) e sólida (S) no equilíbrio. P.Propriedades físico-químicas estudadas 1. ∆A e ∆B são dois parâmetros ajustáveis e ∆Hf a entalpia de fusão do sólido..84) Com a equação 1. xi é a fração molar do soluto.

o coeficiente de atividade da água (calculado por método de contribuição de grupos) e as propriedades da água pura. w  1 1  T − T R  w mist  ∆C p.w é a entalpia de fusão da água na temperatura Tw. Dessa forma. Nesse ponto. xw a fração molar da o água e Pw a pressão de vapor da água pura que pode ser calculada usando a equação de Antoine com as constantes dadas por Gmehling (1977). Um parâmetro geralmente estudado em equilíbrio sólido-líquido de soluções é o ponto eutético. existem em equilíbrio uma fase líquida com o(s) componente(s) sólido(s).86) em que P é a pressão total do sistema. pode-se estimar a temperatura de congelamento da mistura (Tmist). γw o coeficiente de atividade da água. Capítulo 1 para a estimativa simultânea da solubilidade e do coeficiente de atividade. que neste trabalho será 1. a relação básica para o cálculo da temperatura de ebulição deste sistema é: o P = γ w xw Pw (1.w é a diferença entre os calores específicos da água líquida e do gelo à temperatura Tmist. sabendo-se a concentração da mistura. Tw e Tmist os pontos de congelamento da água pura e da mistura com concentração conhecida. está simultâneamente em equilíbrio com o sólido puro A e o sólido puro B.4. assumindo que ∆Cp. ou seja. 1. e ∆Cp. no ponto eutético. R a constante dos gases.28 obtido dos métodos de contribuição de grupos. uma solução de composição xe formada por exemplo por dois componentes A e B. respectivamente.85) em que aw é a atividade de água.3 Depressão do ponto de congelamento Pode-se calcular a depressão do ponto de congelamento de uma mistura empregando a seguinte expressão (Ferro Fontan and Chirife.4 Elevação do ponto de ebulição Em um sistema soluto-solvente em equilíbrio líquido-vapor. ∆Hf. assumindo que a fase de vapor tem o comportamento de um gás ideal. 1981): ln a w = ln( xw γ w ) = ∆H f .4. w  T − 1 − w   T R  mist      (1. w  Tw − ln  T R   mist  ∆C p .w é independente da temperatura entre Tw e Tmist. .

uma estimativa inicial para a Teb e o coeficiente de atividade da água que pode ser calculado por método de contribuição de grupos. um novo valor para a Teb pode ser obtido e comparado ao valor inicial estimado. H simultaneamente.89) . A constante K é dada por: 2 4π 2 n1 (∂n ∂cs )2 K= N Aλ4 (1.88 são experimentais.4. 1982).88) Os termos da parte esquerda da equação 1.6 Espalhamento de luz Em vários livros-texto pode-se encontrar a derivação da equação básica para a avaliação dos dados de espalhamento de luz laser (Kurata. 1.87) O procedimento numérico desenvolvido por Achard et al.Propriedades físico-químicas estudadas 29 A temperatura de ebulição (Teb) pode então ser calculada iterativamente empregando a equação 1.4. Dessa forma. Segundo essa derivação pode-se encontrar uma expressão que relaciona os dados medidos nos experimentos de espalhamento de luz laser com a concentração do soluto (cs) e o potencial químico do solvente (µ1). 1. Kcs ρ ∂µ1 =− 1 RE RT ∂cs (1.5 pH O pH de uma solução contendo eletrólitos fracos pode ser estimado pela seguinte equação que considera a não idealidade do sistema: pH = − log  γ c + c +   H H    (1.86 e conhecendo-se a composição da mistura (xw). (1994) que considera o fenômeno de dissociação parcial das espécies em solução apresenta-se como uma metodologia eficiente para o cálculo da concentração de íons H+ ( c H + ) e do coeficiente de atividade ( γ c + ). Este procedimento é apresentado em detalhe no capítulo sobre soluções de aminoácidos (Capítulo 3).

interferométrico para cada polímero e temperatura de trabalho.4.90) em que hiE é a entalpia parcial molar do componente i menos a entalpia molar do componente i puro a mesma temperatura e pressão.7 Entalpia de diluição A entalpia parcial de excesso de um componente i está relacionada com o coeficiente de atividade da seguinte maneira: h E  ∂ ln γ i − i =  RT 2  ∂T     p . 1.n j (1. considera a diferença entre a intensidade de luz espalhada pela solução polimérica e a intensidade espalhada pelo solvente puro. NA é o número de Avogrado e λ é o comprimento de onda do laser.5 Nomenclatura Símbolos latinos A A a a B b c Cp Área de van der Waals Coeficiente virial Atividade Parâmetro de interação Coeficiente virial Parâmetros de interação concentração molar (g mol-1) Capacidade calorífica . 1.30 Capítulo 1 O incremento (∂n/∂c) é determinado por refratômetro O excesso Rayleigh (RE) em que n é o índice de refração do solvente puro.

31441 J (mol K)-1) Parâmetro de volume Relação de Rayleigh Entropia .Nomenclatura f G H h K M m n NA N N n n P q Q r R R R S 31 Fugacidade Energia de Gibbs Entalpia Entalpia Constante ótica Massa molar Parâmetros de interação independente da temperatura Índice de refração Número de Avogrado Número de componentes Número de viscosidade Número de moles Parâmetros de interação independente da temperatura Pressão Parâmetro de área de componente Parâmetro de área de grupo Parâmetro de volume de componente Constante universal dos gases (R=8.

32 T U U v V V X x z Símbolos gregos Capítulo 1 Temperatura Energia de interação Energia interna Número de grupos Volume Volume de van der Waals Fração de grupo Fração molar Número de coordenação β γ ∆ θ Θ λ µ ν Ξ Ψ Г Parâmetro de interação entre grupos no modelo VERS Coeficiente de atividade Medida de variação Ângulo Fração de área Comprimento de onda de luz laser Potencial químico Viscosidade Viscosidade residual Parâmetro de interação binária Coeficiente de atividade de grupo .

(1). (2) o C Índices dos parâmetros do modelo VERS Componente puro Termo combinatorial . j e k Componente j Componente k no modelo VERS Grupo k Grupo l no modelo VERS Grupo m mistura Grupo n Soluto Água H+ i ij ijk j k k l m mist n s w Sobrescrita (0).Nomenclatura Subscrita 33 0 1 11 f Referência Componente 1 Coeficiente virial binário entre duas moléculas do componente 1 Fusão Íon hidrogênio Componente i Coeficiente virial binário dos componentes i e j Coeficiente virial ternário dos componentes i.

Jacquemin C. Dussap. Sugar alcohols. L. B 2000.. Piggott... J. H..S. J.M..B. Partitioning of whey proteins. G. Achard. M. New York. N. J. Meirelles... . Alves. T. USA. Chumpitaz. S. Ikeuchi. 40. 1975. AIChE J. C. 19. A. C.. J.34 E FH Capítulo 1 Medida de excesso Termo de Flory-Huggins ideal Termo residual Viscosidade id R visc Siglas MD MM PEG Maltodextrina Massa molar Polietileno glicol 1. Wiley. G. and porcine insulin in aqueous two-phase systems. Prediction of pH in complex aqueous mixtures using a group contribution method. 1994. D.R.. M. t. A. AIChE J. Em Handbook of sweeteners. 1210 – 1222. Statistical thermodynamics of liquid mixtures: a new expression for the excess Gibbs energy of partly or completely miscible systems. A. Physical Properties of Molecular Crystals. B. 1968. Billaux. da Silva. Kiritami.. A. 116 – 128. bovine serum albumin. Edit. Franco. F. Marie.Y..6 Referêcias bibliográficas Abrams. 743. Bondi. 235 – 239. L.S. R. Prausnitz. Increase in xylitol production by Candida tropicalis upon addition of salt.. L. Azuma. Ooshima.. M. T. 21. Biomass & Bioenergy 2000. D. J. Glasgow:Blackie. H. J. Liquids and Glasses. 1991. Messing. Chrom. Flourie.. 129 – 135.. Kato.

259 – 268. J. Mini review industrial separation of carboxylic and amino acids by liquid membranes: applicability. 379 – 382.-R.F. S. L. Biotech. Rasmussen. Eng. D. A 1966. Meirelles. C. R... Eng. J. 16. Lawrence. 35 and 45 degrees. J. Surface tension of fatty acids and triglycerides. Data 2000. 76 (3). J. Cussler. Fredenslund. 96.. Food Technology 1981. Kula.. Kinematic viscosities of poly(ethylene glycols). process considerations and potential advantages. Alves. 1717 – 1719. Chumpitaz. Ind. Handbook of water-soluble gums and resins.. J. E. 32. New York. K. M.. Soc. Ferro Fontan. 1980. 41. W. A Study of the water activity lowering behavior of polyethylene glycols in the intermediate moisture range. 5. A. 1980. mass transfer and efficiency. J. 45. A. L. Chirife. 212 – 216. C. 45. Knudsen. S. . 61 – 63. E. Performance of a Graesser contactor in the continuous extraction of whey proteins: mixing. Feakins. Eyal. A. Chem. J. P. McGraw-Hill Book Company. P. Meirelles. K. 2088 – 2092. L. L. Davidson. Chem.. J. F. G. 287 – 291. A. A. J.. Chirife. 2. 1993. S.. Coimbra. Chemtech. Bressler.. Achard. Coutinho. Crystalization for bioseparations. Cruz. Am. Hu. Fluid Phase Equilibria 1994. The evaluation of water activity in aqueous solutions from freezing point depression. Chem. D. J. Catté. Belter. Ferro-Fontan.. L. Society 1999.. M. A. F.. Meirelles. 1993. C-G. R. A. J... A. A. 19. 21 – 30.. A.. 376 – 384.M. Food Sci. Bioseparation 1995. Oil Chem. Bioeng. W.Referências bibliográficas 35 Cao. 1989. D. Thommes. 33 – 50. Dussap. Chumpitaz. J.. J-B. A.. Gros. Res. Relative viscosities of solutions of sodium and potassium chlorides and bromides in n-methylformamide at 25. J. L. C. G. Excess properties and solid-liquid equilibria for aqueous solutions of sugars using UNIQUAC model. M. Group-contribution viscosity predictions of liquid mixtures using UNIFAC-VLE parameters.

The Physical Chemistry of Electrolytic Solutions.. Beeftink. Kumar A. 1965. Trends in Food Science & Technology 1999. R. Clark. E. Thermodynamics of high polymer solutions. Kirk. Onken. U. Harned H. 1981. Gmehling. A. Tramper. Carbohydrate Polymers 1993.E. Armstrong. non-aqueous and mixed solvents applicable to high concentration.. The effect of film composition on the diffusion of ethanol through soft gelatin films.. 123 – 131. 1993. 243 – 248. 1982.. Group contribution estimation of activity coefficients in nonideal liquid mixtures. I. J. Food Hydrocolloids v III. Thermodynamics of Polymer Solutions. 1941. 1983. Chem.. Norton. 21. Kasapis. OThmer. B. 9. AIChE J. . J. Phase equilibria and gelation in gelatin/maltodextrin systems – Part I: Gelation of individual components. M. Encyclopedia of Chemical Technology. J. John Wiley & Sons: New York. Morris. New York. 948 – 954. Moreton. Pharm 1998. A simple correlation for estimating viscosities of solutions of salts in aqueous. Jones R. C. Ed. H. 10. 345 – 355. A. Chem.. M. H. Phys. temperature and pressure. 660 – 661. J. D. L. Kurata. J. Owen. Prausnitz. J.M. 21. A. 1994. Untersuchungen zur Verteilung von Aminosäuren und Peptiden auf wässrige Zwei-Phasen-Systeme. M. Frankfurt: Dechema.. Groβmann. New York.. Int. R. Dissertação de Doutorado da Universidade de Kaiserslautern. Glicksman. H. CRC Press Inc.F. Reinhold Publications Co. J. 1975.. 161. S. T. 1986.L... 3rd. C. Can. Fredenslund A. Marchal. Towards a rational design of commercial maltodextrins. P.. Eng. R. Alemanha.36 Capítulo 1 Flory. B. Harwood Academic Publishers: New York. N. Vapor-Liquid Equilibirum Data Collection. 1086 – 1099. 71...

A. Chem.Referências bibliográficas gum arabic: effect of concentration and blending. A modified unifac model for the calculation of thermodynamic properties of aqueous and non-aqueous solutions containing sugars. G. Kister. Prausnitz. A. 14. M. M. Algebraic representation of thermodynamics properties and the classification of solutions.. R. Chem. 71 – 95. Vázquez. J. 40. Thermodynamic properties of sugars in aqueous solutions: correlation and prediction using a modified uniquac model. A. M.N. Molecular thermodynamics of fluid phase equilibria. New Jersey. 335 – 345. J.. A. Azevedo.. Food Hydrocolloids 1999. 35 – 40. Chenlo. G. Lichtenthaler. Molecular thermodynamics of fluid phase equilibria. Rheological behavior of aqueous dispersions of cashew gum and Pereira. Faraday Soc. Fluid Phase Equilibria 1996... 267 – 274. Rao.. A. Moreira.. H.. Fluid Phase Equilibria 1997. 81. CRC Press. 47 – 74. 1401 – 1405.. Meirelles. M. Peres. W. C. R. Silva. Activity Coefficients in Electrolyte Solutions.M.: Prentice-Hall. F. Sé. 139. L. H. E. 13. Macedo.A. Eng. 1920. On the vapour pressures of mixtures. 123. J.. 1986. 1948. S. Carbohydrate Polymers 2001. L. N. Eng. A.. J. 47. E. 345 – 348. Boca Raton. Peres. A. G.. Macedo. Carbohydrate Polymers 2000. Porter. 1991.M. Ind. E. Pitzer. . M. 42. Redlich. K. Kinematic viscosity prediction for aqueous solutions with various solutes. J. Trans. Prentice Hall Inc. 501 – 506. Data 2002. R. M. Meirelles. M. J. Aznar. C. Silva. Chem. 273 – 278. A. 37 Mothé. Eng. A.. 46. J. Englewood Cliffs. J. E.G. Phase equilibrium in polyethylene glycol/maltodextrin aqueous two-phase systems. Phase equilibrium and protein partitioning in aqueous mixtures of maltodextrin with polypropylene glycol. A. O. Prausnitz. 1969. 2001. Liquid-liquid equilibrium of the aqueous two-phase system water + peg 4000 + potassium phosphate at four temperatures: experimental determination and thermodynamic modeling.

Oil Chem. Soc. 1964. Requirements for biodegradable water-soluble polymers. Ind. Polymer Degradation and Stability. Valeri. 1998. Appl. M. S. . A. Vapor-liquid equilibrium XI. C. 1221 – 1226. J. Eng. . J. Wilson. 86. M. G.. Am. M. G. 20 – 24. Activity coefficients and molecular structure. D. J. A new expression for the excess free energy of mixing.M. G. 74.38 by xylose-fermenting yeasts. A. 19 – 24. H. Factors that affect the biosynthesis of xylitol Swift... S. Biotech. Capítulo 1 Silva. Meirelles. Am. triglycerides and their binary mixtures. Viscosities of fatty acids. Chem. Fundam. G. 59. Chem. I. Deal. 1962. Society 1997. 1998. Mancilha. 331 – 339. A. Biochem. 127 – 134. 70 – 72. Felipe. Wilson.

2000. 654 – 660.O. Antonio J. p. M.Capítulo 2 39 Capítulo 2 WATER ACTIVITY IN POLYOL SYSTEMS L. Box 6121 Zip Code 13083-970 Campinas – SP. P. v. Camargo. . 45. Brazil Trabalho publicado na revista Journal of Chemical Engineering and Data. State University of Campinas – UNICAMP. Meirelles Department of Food Engineering (DEA/FEA).S. Cidade Universitária “Zeferino Vaz”. A. Ninni.

40 Capítulo 2 .

sugars. Results were compared with the group contribution-based models ASOG and UNIFAC. 1998. is an important support for modeling and designing industrial processes like concentration and purification in a separation unit. probably as a consequence of the strongly polar hydroxyl groups bounded to consecutive carbon atoms in the polyol molecule.. meso-erythritol and glycerol). and microbiological reactions that influence food stability are dependent on the availability of water. 1979) are considered to provide good results for estimating activity coefficients in the liquid phase. and cosmetics. enzymatic. Group contribution-based models such as UNIFAC (Fredenslund et al. Velezmoro et al. D-mannitol. Better agreement was obtained by readjusting some of the interaction parameters. 2000). The best results were achieved using the UNIFACLarsen model with an average relative deviation of 0. . 2. The knowledge of the phase equilibria in systems containing biological products. xylitol. New specific groups were defined: glucose and fructose rings. organic salts and other solutes (Velezmoro and Meirelles. Correa et al. Water activity in polyol systems 2.. urea. Ninni et al. pharmaceuticals. since many chemical.in food systems containing amino acids.Introduction 41 2. Water activity is an important physical chemical property in food engineering.1 Abstract Water activities of binary and ternary mixtures containing polyols were measured using an electronic hygrometer with temperature ranging from (10 to 35) °C. The concentrations of the mixtures varied according to the solubility limit for each polyol (D-sorbitol. The data bank used in this procedure included water activity data as well as polyol solubility data taken from the literature. such as polyols. Thermodynamic models have been used for calculating some physicochemical properties – such as water activity.9% for water activity and solubility data. 1999b. 1975) and ASOG (Kogima and Tochigi. 1999a.2 Introduction Polyols are polyhydroxy alcohols or sugar alcohols used in many diverse fields including foods. (1994) studied the behavior of some aqueous polyol solutions: water activities (aw) of binary (polyol + water) and ternary (urea + polyol + water and urea + sugar + water) solutions were measured and results compared with those of the ASOG group contribution model. The predictions using parameters from the literature were poor..

In this work.01% approximately in mass fraction. The amount of water varied from 0.64 mass %. Herisau. new parameters for UNIFAC-Larsen and ASOG models were determined. xylitol. . Before this solution were prepared. and it was considered for calculating the water concentration in solutions. The compositions were accurate to ±0. the water content in the solid polyols was determined by Karl Fischer titration (Metrohm.1 mg accuracy.001 aw units. allowing the calculation of water activity. An electronic hygrometer AQUA-LAB CX-2 (DECAGON. Recently. 2. Furthermore. solubility. They were analytical grade reagents from SIGMA with purity > 99%. The solutions were prepared by mass percent with distilled water using an analytical balance (Sartorius. USA). Goettingen. Chicago. Germany) with ±0.08 to 0. The following polyols used were: D-sorbitol.42 Capítulo 2 polyalcohol. Switzerland). Device Inc. D-mannitol. and freezing point data of aqueous polyol solutions with low deviations between experimental and calculated values.. Pullmann. Measurements were made in triplicate with a reproducibility of ±0. and cyclic polyalcohol.3 Experimental Section Water activities of binary and ternary aqueous polyol systems were determined from (10 to 35) ºC. meso-erythritol and glycerol. water activity has been measured for the binary (polyol + water) and ternary (polyol + polyol + water) systems using an electronic hygrometer. USA) previously calibrated with saturated salt solutions was used for measuring the aw. The results showed that readjustment of binary interaction parameters provided better agreement between the model and experimental aw values when compared with the predictions using the original parameters from Kojima and Tochigi (1979)..1 °C by circulation of thermostated water from water bath (Cole Parmer Instruments Co. The temperature inside the hygrometer was regulated at the desired ±0. Peres and Macedo (1997) have shown that the UNIFAC-Larsen model can be successfully used for calculating thermodynamic properties of aqueous and nonaqueous solutions containing sugars.

991 0.6457 0.0°C w2 aw 0. 1938) which were measured using the isopiestic method.2.937 0.1986 0.0992 0.1750 0.978 0. the experimental data for the system containing glycerol are in very good agreement with the available literature data (Scatchard et al.3974 0.992 0.992 0.1491 0. respectively. The work of Roa and Daza (1991) also emphasizes a good performance of a prior version of this electronic hygrometer (AQUA-LAB CX-1) for measuring water activity for various kinds of food systems.2982 0. (1999a) for sugar solutions and poly(ethylene glycol) solutions.991 0. For the same mass concentration.0510 0.2980 0.4468 0. Such result confirms the accuracy of experimental aw data measured using the AQUA-LAB CX-2 (DECAGON Devices Inc.1.934 0. Pullman.1%. The change of aw with temperature is small.5706 0.1745 0.986 0. it was observed that the polyols with low molecular weight are better water activity depressors than those with high molecular weight.1. The mean relative deviation between these two data sets is 0..1246 0.997 0.935 0.995 0.980 .977 0. Similar results were also reported by Velezmoro et al.0495 0.1263 0.901 0..4 2.983 0.897 0.1036 0.996 0.6946 aw 0.1 and 2.5952 aw 0.0519 0.919 0.862 0.987 0.5466 0.949 0.989 0.0°C 35.970 0.1501 0.899 0. This difference is well visualized at high solute concentrations.755 water (1) + mannitol (2) 25. Water activity in binary polyol solutions as a function of mass fraction of polyol (w2) 10.4966 0.0995 0. (2000) and Ninni et al.5955 0.0497 0.985 0.0°C w2 aw w2 aw 0.0°C w2 0. respectively.844 water (1) + sorbitol (2) 25.996 0.982 0.2965 0.4944 0.1978 0.960 0. USA).803 35. Table 2.Results and Discussion 2.4.1467 0.979 0.960 0. As can be seen in Figure 2.1035 0.3966 0.996 0.0993 0.0°C w2 0.3953 0.1 Results and Discussion Water activity 43 The experimental water activity data obtained in this work for the binary and ternary mixtures are given in Tables 2.0751 0.3478 0.961 0. The concentrations are in mass fraction.843 0.987 0.2485 0.872 0.992 0.1985 0.0501 0.4918 0.

981 0.0966 0.980 0.816 0.816 0.985 0.0502 0.44 Table 2.4489 0.996 0.880 0.3009 0.847 0.1498 0.7489 0.1095 0.962 0.951 0.628 0.938 0.6487 0.812 0.733 0.967 0.5423 0.844 0.973 0.974 0.972 0.0498 0.2502 0.5993 0.897 0.993 0.994 0.0°C w2 0.1020 0.7986 0.2996 0.1997 0.4986 0.979 0.4491 0.483 0.0501 0.6987 0.6490 0.3002 0.919 0.5021 0.0°C 35.854 0.989 0.0°C w2 aw w2 aw 0.1510 0.0999 0.2022 0.5970 0.0997 0.921 0.951 0.3494 0.2914 0.1 (cont.2002 0.936 0.876 0.2033 0.951 0.901 0.1011 0.4992 0.3995 0.2496 0.878 0.992 0.3690 0.4492 0.980 0.918 0.775 0.3345 0.0°C 35.3470 0.1001 0.1547 0.937 0.4995 0.850 0.896 0.2334 0.940 0.5993 aw 0.557 0.2524 0.399 10.0502 0.774 water (1) + glycerol (2) 25.3994 0.5491 0.5991 0.922 0.2995 0.0°C w2 aw w2 aw 0.0506 0.952 0.818 Capítulo 2 water (1) + erythritol (2) 25.0500 0.874 0.) water (1) + xylitol (2) 25.780 0.3990 0.3906 0.964 0.993 0.5489 0.971 0.1501 0.3495 0.3995 0.1999 0.3496 0.963 0.987 0.4969 0.0°C w2 aw w2 aw 0.1499 0.903 0.6495 0.2005 0.922 0.925 0.5533 0.2504 0.950 0.689 0.936 0.982 0.840 0.5489 0.1996 0.740 0.2983 0.990 0.988 0.632 0.872 0.926 0.0486 0.937 0.991 0.0°C 30.5670 0.683 0.953 0.2996 0.4497 0.1499 0.967 0.3479 0.8487 0.912 .956 0.6989 0.849 0.963 0.975 0.2496 0.

871 0.7 0.2974 0.1485 0.973 0.1503 0. Prediction of water activities in the glycerol – water system.0249 0.0254 0. 1938) exp.965 0.6 0.992 0.0491 0.0505 0. (this work) UNIFAC-Larsen with parameters from Larsen et al.993 0.988 0.0494 0.970 45 water(1)+glycerol(2)+sorbitol(3) w2 w3 aw 0.4 0.1399 0.1977 0.993 0.8 0.0 0.1911 0.895 0.0499 0.972 0.2998 0.961 0. aw .0383 0.1958 0.1002 0.0250 0.0498 0.0999 0.0765 0.0251 0.897 0.0ºC water(1)+xylitol(2)+sorbitol(3) w2 w3 aw 0.0999 0.1259 0.0 exp.981 0.1904 0.987 0.979 0.2960 0.908 0.0872 0.0376 0.2.979 0.0254 0.0750 0.0264 0.974 0.3262 0.980 0. (Scatchard et al.2494 0.0511 0.1398 0.982 0.887 0.0610 0.9 0.1490 0.2266 0.0499 0.1245 0.0525 0.0255 0.984 0.6 0.954 0.772 0.932 0.1004 0.1764 0.940 0.684 water(1)+xylitol(2)+mannitol(3) w2 w3 aw 0.1246 0.0999 0.3226 0.994 0.993 0.787 water(1)+glycerol(2)+xylitol(3) w2 w3 aw 0.1248 0.964 0.3249 0.2 0.0866 0.0756 0.0750 0.0770 0.3480 0.Results and Discussion Table 2.1245 0.0250 0.0 0.1493 0.991 0. (1994) ASOG (this work) Figure 2.848 0.0500 0.3013 0.3039 0.959 0.1397 0.960 0.1779 0.2482 0.8 1.2040 0.1.957 water(1)+erythritol(2)+mannitol(3) w2 w3 aw 0.0250 0.2988 0.938 0.0761 0.1023 0.0749 0.3501 0.2229 0.2389 0.0249 0.971 0.1029 0.2105 0.953 1.1249 0.987 0.1152 0.0605 0.1498 0.726 water(1)+glycerol(2)+mannitol(3) w2 w3 aw 0.0511 0.1398 0.988 0.3246 0.0249 0.0501 0.5 0.0994 0.982 0.1005 0.0509 0.0749 0.1491 0.2626 0.4 w 0. Water activity in ternary polyol solutions at 25. (1987) UNIFAC-Larsen (this work) ASOG with parameters from Correa et al.780 0.827 0.0254 0.

CHOH are groups proposed by Wu and Sandler (1991a.. 4CH. An average relative deviation.875) but a high deviation (15.46 Capítulo 2 The prediction of water activity using different versions of the UNIFAC model (Fredenslund et al.1).2% for the original UNIFAC model (Skjold-Jorgensen et al. 1987) was used in the predictions. was found as 2. 2CH2OH 4CHOH.1 % for the ASOG model (with parameters from Correa et al. (1994). OH are groups proposed by Skjold-Jorgensen et al. 1987) and the ASOG model (Kojima and Tochigi.998 to 0. Bower and Robinson. 1979) and 1. This was well observed in the system containing glycerol. Robinson and Stockes..8CH2. This could be a consequence of using a restricted range of solute concentrations for the adjustment of the parameters. For the ASOG model with parameters from Correa et al. 2CH. 5POH 5. This suggests that there is an intramolecular proximity effect between these constituent groups. 4OH 2CH2. 1979) was carried out as a first estimate. including all experimental aw values from this work and from the literature (Scatchard et al. 4CH. which presented a mean deviation of 0. 1991b) and Abildskow et al. 3OH 2CH2. the predictions were similarly poor. (1979). some of the UNIFAC and ASOG interaction parameters were readjusted. 1. 5OH 2CH2. also relative to chemical structure of the polyols. 6POH viFH 6 8 10 12 12 glycerol meso-erythritol xylitol D-mannitol D-sorbitol a CH2OH. It was observed that the models were not sufficiently accurate to predict values of aw at high solute concentrations.. 2CH2OH 2CHOH. Group assignment for polyols UNIFACa vki 1CHOH.4CH2. c POH is a group proposed by Correa et al. 1CH. CH. 3POH 3.6CH2. 6OH ASOGc vki 2.8% when the UNIFAC-Larsen model (Larsen et al. (1994).3. 3CH. 1975. The group assignments utilized for the polyols in the various attempts are summarized in Table 2.2CH2. 2CH2OH 4CHOH. vki is the number of groups k in molecule i. 6POH 5.2CH2.. Another reason. On the basis of these results and considering the proximity effect of the hydroxyl groups. 1938.6%) for the whole concentration range studied in this work (see Figure 2. b CH2.. (1994)). viFH is the number of atoms (other than hydrogen atoms) in molecule i. Larsen et al. 4POH 4. 2CH2OH UNIFACb (vki) vki 2CH2. 2CH2OH 3CHOH. 1961. (1996). 6OH 2CH2. b). 1963). Table 2.8% in a restricted range of water activity (0.3. . This poor estimation is attributed to the effect of strongly polar hydroxyl groups bounded to consecutive carbon atoms in the molecule. as proposed by Wu and Sandler (1991a. could be referred to the models being unable to distinguish between the molecular structure of isomers (mannitol and sorbitol).

5a -0.0b 278. Table 2.b) and the second one suggested by Skjold-Jorgensen et al. 1979.7a Parameters readjusted in this work.5 for the UNIFAC-Larsen and ASOG models.0831a -2. and they are also given in Tables 2.4. we have used the original equation and the UNIFAC-Larsen version. Table 2.5 for the UNIFAC-Larsen and ASOG models. different strategies were used. (1994).4 and 2.2868a 257.5b 410.7b b OH 972.3184a -3. the Marquardt method (Marquardt.9a H2O 1857. 1979. UNIFAC-Larsen interaction parameters CH2 CH2 CH OH H2O a CH 0.2 Readjustment of group interaction parameters 47 To readjust some of the interaction parameters of the UNIFAC and ASOG models.3b -0. To readjust some of the parameters. Skjold-Jorgensen. Parameters obtained from Kogima and Tochigi (1979).0 637. Parameters obtained from Larsen et al.6705a n m n m n m Parameters readjusted in this work. In the UNIFAC model. (see Table 2.0b 1857.Results and Discussion 2.7b 0.5b 410. ASOG interaction parameters CH2 CH2 POH H2O a POH 2434.8b -175. For the ASOG model we have used the group assignment proposed by Correa et al. (1987).4.2727b -42.7b 0.76a -0.5. 1963) was used for minimizing the following objective function (OF) a w calc − a w exp a w exp sol cal − sol exp sol exp OF = ∑ n +∑ k (2.0 637. 1987).3).4 and 2. respectively.5045b b H2O -277.1) .2184a -2382. Larsen et al. (1979). The other group interaction parameters were set equal to the values available in the literature (Kogima and Tochigi.7a -9. The binary interaction parameters readjusted in this work are given in Tables 2..8b 972. For each version we have assumed two alternatives for the division of groups: the first one proposed by Wu and Sandler (1991a.

Moreover. In these cases. 1991 Mullin. D-sorbitol. 1991 Röper et al. (1994).4 for the binary systems of xylitol.48 Capítulo 2 where n and k are the number of experimental water activity and solubility data. and the subscripts calc and exp mean calculated and experimental values. 1993... The UNIFAC-Larsen model was capable of providing good results even at high solute concentrations. The results calculated by both the UNIFAC-Larsen and ASOG models are shown in Figures 2. Billaux et al.6. Billaux et al.. the use of a wide range of solute concentrations and different types of polyols makes it possible to attain significant interaction parameters for use in calculation of the thermodynamical properties studied. The systems used for readjusting the interaction parameters are given in Table 2. the readjustment of the ASOG parameters between the POH group and the other mixture constituent groups improved the calculated aw values for these systems. Systems used for readjusting the group interaction parameters aqueous system D-sorbitol D-sorbitol D-mannitol D-mannitol xylitol meso-erythritol D-sorbitol D-mannitol xylitol erythritol temp range/oC Water Activity 10-35 25 25-35 25 10-35 25-30 Solubility 0-50 0-100 20-50 20-25 reference data this work Bower and Robinson. 1961 this work this work Mullin. 1993. . respectively.6.2 – 2. 1963 this work Robinson and Stokes. Table 2. (1979). in comparison with the obtained results from Correa et al. 1993 The best results were obtained using the UNIFAC-Larsen model with the group assignment proposed by Skjold-Jorgensen et al.. 1991 Billaux et al.mannitol. and D. sol is the solubility.

40 0.3.3 w 0.5 0. solubility ASOG Figure 2. Experimental and calculated water activities for xylitol solutions at 25 °C. o . 1991) exp.Results and Discussion 1.65 0.75 0.80 0. solubility UNIFAC-Larsen pred.85 0.solubility (Mullin.. freezing point (Uraji et al.7 Figure 2. 60 50 40 temperature / C 30 20 10 0 -10 -20 0.0 0.90 solubility (w/w) exp. freezing point ASOG calc.55 0.00 49 0.70 0. Solid-liquid equilibria for sorbitol aqueous solutions (solubility and freezing point depression). solubility (Billaux et al.4 0..2 0.90 aw 0.80 0.45 0.2.1 0.75 0.50 0. 1997) exp.60 0.85 experimental (this work) UNIFAC-Larsen with parameters from Larsen et al.95 0. 1993) pred. freezing point UNIFAC-Larsen calc. (1987) UNIFAC-Larsen (this work) ASOG with parameters from Correa et al. (1994) ASOG (this work) 0.6 0.

19 0.42 0.50 0.23 0.26 2.7 shows a comparison between the average relative deviations obtained in this work and those from the literature for water activities.30 0. 1963 this work Robinson and Stockes.76 0.04 0. this work Correa et al.51 0.35 0. this work (1987) (1994) 2.5 0.0 -10 0 10 20 30 40 50 60 o Capítulo 2 experimental (Billaux et al.38 reference D-sorbitol D-sorbitol D-mannitol D-mannitol xylitol meso-erythritol avg deviation this work Bower and Robinson.07 0.01 0.06 0.12 0. Table 2.68 0. Experimental and calculated solubilities for D-mannitol at different temperatures.24 1.17 0.1 0.69 0.3 0. 1961 this work this work .6 solubility (w/w) 0. Tables 2. The deviations between experimental and calculated solubilities using the sets of parameters adjusted in this work are given in Table 2.11 0.8.08 0.60 1.11 0.4 0.2 0.51 3.7 0. Mean relative deviations between experimental and calculated aw data points aqueous system % deviation UNIFAC-Larsen ASOG Larsen et al.19 0. 1993) UNIFAC-Larsen ASOG 70 80 90 100 110 temperature / C Figure 2.4..64 0.16 0.02 2.8 0.7.01 0. 1991) experimental (Mullin.

86 20.23 4. 1993). These are for the solid and liquid phases of erythritol and mannitol (Spaght et al. Moreover. 1993. 1997) was used.25 reference Mullin. This assumption was already used in previews works from Catté et al.79 Table 2.9 as well as the data used for water in the calculations of the freezing point depression of sorbitol and xylitol solutions....80 0. Mullin. For these reasons we opted for estimating ∆C p as a linear function of temperature.63 6.50 0 .9 also indicates published values for ∆C p from the literature.1993... 1993 For the calculation of polyol solubility in water the same equation adopted by Peres and Macedo (1996. The other cited literature (Barone et al. The values in bold type were used in this work. The thermodynamic properties needed for the determination of solubilities are reported in Table 2.8. the experimental values taken from distinct literature sources are expressively different. It should be observed that the ∆A values estimated in this work are relatively close to the experimental ∆C p . Mean relative deviations between experimental and calculated solubility data aqueous system D-sorbitol D-mannitol xylitol erythritol avg temperature range/ºC 0 . 1991.80 3.100 20 .22 1. 1997) for systems containing sugars. and excluding the data for mannitol and erythritol.2) where Tref is a reference temperature which was set equal to 25°C and where ∆A and ∆B are two adjustable parameters. There is only one literature source in which experimental Cp values were measured for a wide range of temperatures. 1990) used a group contribution approach to estimate the Cp values for the liquid polyol. (1994) and Peres and Macedo (1996. in some cases.91 1.Results and Discussion 51 % deviation UNIFACASOG Larsen 0.51 0. Röper. The values for ∆A and ∆B are presented in Table 2. 1993. Billaux et al. Table 2.50 20 . 1991 Röper et al. 1991 Billaux et al. The parameters ∆A and ∆B were adjusted for each polyol using the experimental solubilities at various temperatures taken from the literature (Billaux et al.. 1932). 1991 Mullin. The expression for representing the difference between the heat capacities of the pure liquid and those of the pure solid polyols ( ∆C p ) is given below: ∆C p = ∆A + ∆B ( T − Tref ) (2.10. Billaux et al.32 1. values were measured at only one temperature.

c Raemy and Schweizer (1983).9. (1932). d Fassman (1975). for the ASOG model.002e ∆Cp (25ºC)/ J⋅K-1⋅mol-1 191a 191a 290.03e water a Barone et al. a mean relative deviation of about 16% was found between the calculated values and the ∆C p experimental data at 25 °C. Tm/K 366. Values of ∆A and ∆B for the calculations of ∆Cp with linear temperature dependency polyol D-sorbitol D-mannitol xylitol meso-erythritol UNIFAC-Larsen ∆A ∆B J⋅K-1⋅mol-1 J⋅K-2⋅mol-1 214. OH.7 225.2 1.15e enthalpy of fusion. CH.2 56.8 124. and H2O).9a 391.0 0. Table 2. .36b 157a 122a 155. as can be seen in Table 2.9712 215.58b. In general. 52. Table 2.8 ∆B J⋅K-2⋅mol-1 -4. Thermodynamic data of polyols and water polyol sorbitol mannitol xylitol erythritol melting temp.9598 It must be stressed that. the minimum values for the deviation concerning solubility data were obtained for the UNIFACLarsen model (with the groups CH2.0924 ASOG ∆A J⋅K-1⋅mol-1 215. Otherwise. For instance.0 3. (the values for ∆Cp are results from regression of the experimental data). CH. ∆fusH/kJ⋅mol-1 30.36b 40.0d 390.5a 439.3261 117.0c 39. Furthermore.2073 2.4a 42. the deviation between the ∆A and ∆C p values amounts to approximately 35%. better results were achieved only after the readjustment of four pairs of temperature-dependent interaction parameters. For example. we readjusted only two temperatureindependent interaction parameters between the groups OH/H2O and H2O/OH. these deviations were higher for the other models used in this work.52 Capítulo 2 values at 25 °C.3c 6. OH.5 115.7a 367.9 -2.1643 -0.0618 128.6b 273.42b 38.2b 365. for the UNIFAC-Larsen model.10.1a 433. (the ∆Cp for water was considered constant with temperature). b Spaght et al.8c 37.4a 38. in the case of the original UNIFAC model. and H2O). in the case of the UNIFAC-Larsen model (with the groups CH2.7738 1. e Daubert and Danner (1985).1a 53. (1990).8.

0.11 indicates the average mean deviations obtained for the water activities of the ternary mixtures and the binary systems containing glycerol.1).09 0.63 3.65 0. The experimental data are from Uraji et al.787 0.4.0.18 2. this work xylitol+sorbitol glycerol+mannitol glycerol+sorbitol glycerol+xylitol xylitol+mannitol erythritol+mannitol glycerol glycerola avg a 0.60 2.30 2. Figure 2.65 0.27 0.54 1.14 0. the predictions of freezing point depression did not present the same accuracy.23 0.88 0.992 .990 – 0.16 15. (1938).991 .5 shows calculated and experimental aw values for the ternary mixture xylitol-sorbitol-water at 25°C. (1987) this work (1994) ASOG Correa et al.94 4.12 6. (1994) have already commented about the difficulties in obtaining good results for the systems with glycerol.10 0.99 4.55 0.970 0.13 2.35 0. Note that a good prediction for the eutectic point concentration was obtained for both systems.0.16 0. For freezing point calculations.998 – 0.762 2.12 for the ASOG and UNIFAC-Larsen models.37 6.399 0.957 0. However.993 . .0.684 0.3.85 2.3 Predictions with the new set of parameters The experimental freezing point data available in the literature as well as the aw values of the ternary systems and of the binary system containing glycerol were used only for comparison with predicted values.08 0.95 1.48 0.20 2. The prediction of the eutectic points for xylitol and sorbitol aqueous solutions was also performed.40 2.87 0. with mean deviations of 10% for the UNIFAC-Larsen model and 47% for the ASOG model.Results and Discussion 53 2. Table 2.49 0.16 3.993 .994 . Table 2.726 0.99 Scatchard et al. an expression proposed by Ferro Fontan and Chirife (1981) was used.17 0. (1997). These results are given in Table 2.0.953 0.993 .0.11.08 2.63 3. Correa et al. Water activity prediction in polyol mixtures % deviation polyol mixture aw range UNIFAC-Larsen Larsen et al.21 0. It can be noted that the deviations for glycerol were higher than those for the ternary systems but significantly lower in comparison to the predictions with parameters from the literature (see Figure 2. and the results achieved for sorbitol are also represented in Figure 2.

5 0.80 0. G. 1987) ASOG (with parameters from Correa et al.. Fluid Phase Equilib.. . 86...5. 1996.. Enthalpies and entropies of sublimation. Faraday Trans. Della Gatta. Predictions of water activities for the ternary system water – xylitol – sorbitol at 25.0 °C. Sugar alcohols.3 0. Experimental and calculated eutetic points of xylitol and sorbitol aqueous solutions experimental T/°C conc/(w/w) -12. V. Gani.54 ASOG T/°C conc/(w/w) -5.8 0.85 experimental (this work) UNIFAC-Larsen (with parameters from Larsen et al. Table 2. S. Soc. Chem.5 Literature Cited Abildskov.95 0.44 -13.54 Capítulo 2 1. D. J. Piggott. Marie..43 -8.5 0. 75 – 79. 1991...53 UNIFAC-Larsen T/°C conc/(w/w) -10.12.S. J. In: Handbook of sweeteners.2 0. Billaux. G..75 0. 1990. Eds. Constantinou. Jacquimin. Glasgow: Blackie: New York.00 0. J.0 0. vaporization and fusion of nine polyhydric alcohols. C. L.4 0.53 xylitol sorbitol 2. R.90 aw 0.R. Towards the development of a second-order approximation in activity coefficient models based on group contributions.5 0.6 0. Barone.. 1994) UNIFAC-Larsen (this work) ASOG (this work) 0.7 concentration (w/w) Figure 2. 118.9 0.. Ferro.. Flourie. B. M.2 0. Piacente.1 0.43 -15.3 0. 1 – 12.

Gros.L. Soc. C. 1987.E. Rasmussen. Sereno. An algorithm for least-squares estimation of nonlinear parameters. J. 1979. OH. B. . 431 – 441. Appl. 1963... Meirelles. Robinson. C-G. V. Crystallization. J.. 11. Prediction of Vapour-Liquid Equilibria by the ASOG Method. Correa. Chirife. Water activity in poly(ethylene glycol) aqueous solutions.D.M. Res. 1540 – 1541. Isopiestic vapour pressure measurements of the ternary system: sorbitol-sodium chloride-water at 25°C.E. A. Camargo.. CRC Press: Cleveland. Dussap. J. Tochigi. R. Mullin J. 2274 – 2286.S. Indust. Chem. 189 – 199.W. Handbook of biochemistry and molecular biology.. Ferro Fontan. Marquardt. J-B. 1975. Acta 1999a. P. M. Eng. AIChE J. 328.. 67... Catté. Data Compilation Tables of Properties of Pure Compounds. Achard. Fluid Phase Equilib. 21 – 30. 98.. Kojima. 26. Butherworth – Heinemann: New York. A. Fluid Phase Equilib.J.Literature Cited 55 Bower. A modified UNIFAC group contribution model for prediction of phase equilibria and heats of mixing.. Measurement of water activity in water-urea-sugar and water-urea-polyol systems. J..L. Danner. Fassman. K. Jones R.. Food Technol. Excess properties and solid-liquid equilibria for aqueous solutions of sugars using UNIQUAC model. Comesaña. 1963. 1975. 33 – 50. Fredenslund. A.. L. 1985. Math.A. R. Fredenslund A. Prausnitz. 16. 1086 – 1099.M. The evaluation of water activity in aqueous solutions from freezing point depression. Ninni. 169 – 176. Phys. Chem. Daubert.A. A. J. 1993.. K. Group contribution estimation of activity coefficients in nonideal liquid mixtures. Thermochim. Larsen. Elsevier: Tokyo.F. T. AIChE/DIPPR: New York.W. and its prediction by the ASOG group contribution method. D. G. J. 1994. 1981. Ind. C. 1994. 96. M. 21.P.

sulphuric acid. Stockes. L. S. Robinson. J. Phys..B. 71 – 95.. M. 714 – 722. 18. mannitol and their mixtures at 25°C. Daza. 3061 – 3070. 1979.. Anal. A.. Eng. Fluid Phase Equilib. Roa. M.-Wiss. Chem. Modeling and prediction of pH and water activity in aqueous amino acid solutions. Therm. Skjold-Jorgensen. Starch 1993. A. 208 –213.56 Capítulo 2 Ninni.F.S.A. Chem. M. Computers and Chemical Engineering 1999b. Am. 1961. potassium chloride. Fluid Phase Equilib. 400 – 405.. S. Lebensm. sucrose..T. Soc. Scatchard. Rasmussen.. G..A. Macedo. Activity coefficients in aqueous solutions of sucrose. Thermodynamic properties of sugars in aqueous solutions: correlation and prediction using a modified UNIQUAC model. 1983. H. 23 (supplement) S383 – S386. Parks. 28. 1938.A. Peres.. 1. Schweizer. T. Revision and extension. A. B. J. E. Röper. . Phys.M. 11.H. 1996.M. Peres. J. Raemy. Chem. 47 – 74. R.S..J. a new material for food and non-food applications. J. Hamer. Some heat capacity data on organic compounds obtained with a radiation calorimeter. 24. Process Des. Isotonic solutions. Thermal behaviour of carbohydrates studied by heat flow calorimeter. Meirelles. Gmehling. E. V.E. Camargo.60. Technol. 882 – 888. Dev. 1991. Spaght. J. urea and glycerol at 25°C.. Ind. Thomas.S.. Vapor-liquid equilibria by UNIFAC group contribution. R. S. Kolbe. A. A modified UNIFAC model for the calculation of thermodynamic properties of aqueous and non-aqueous solutions containing sugars. W. Macedo. Evaluation of water activity measurements with a dew point electronic humidity meter. 65. 36. Goossens. 1954 – 1958.. The chemical potencial of water in aqueous solutions of sodium chloride. Erythritol.. Chem. 123. J.E.J.A. 1997. P. G. 139. 1932. 95 – 108. Wood.

K. 1991b. 380 – 385.. 2. Tech. 1997. 521011/95-7).. Wu. Res.. C.. 44. C. A. 881 – 897.A. Brighnton. Water activity in solutions containing organic acids. 95/02617-5) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq . Velezmoro.. M. H. S. Nakashima. 30. Part 1. Food Sci. Test of new groups in UNIFAC.. Thermal Analyses of polyol-aqueous solutions below 0°C by differential scanning calorimetry. Kohno.I. 881 – 889. A. Eng. Chem. Vitali. Drying Technol. S. Chem. Theory and the basis for group identifications. 1991a. Watanabe.. 1789 –1805. 30. Shimoyamada. UK.6 Acknowledgment This work was supported financially by research grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP . Ind. Meirelles.S. H. T.J. Sandler. Japanese Soc. 1998. K.I. Engineering and Food at ICEF 7. Velezmoro. .Literature Cited 57 Uraji.Proc. H. J. Meirelles. Eng. Use of ab initio quantum mechanics calculations in group contribution methods. Res.. Ind.J. p. Sandler. Wu.A. A145.S. 1. 2.. 16. Use of ab initio quantum mechanics calculations in group contribution methods.Proc. 1997. Proceedings of the Seventh International Congress on Engineering and Food. A.

Capítulo 3 59 Capítulo 3 WATER ACTIVITY. State University of Campinas (UNICAMP). Zip Code 13083970. Meirelles Department of Food Engineering (DEA/FEA). Cidade Universitária "Zeferino Vaz". Antonio J. v. Campinas. 17 (4). Brasil Trabalho publicado na revista Biotechnology Progress. PH AND DENSITY OF AQUEOUS AMINO ACIDS SOLUTIONS Luciana Ninni. p. 2001. 703 – 711. . PO Box 6121. SP. A. Faculty of Food Engineering.

60 Capítulo 3 .

it is relevant to study the physical-chemical properties of systems containing such compounds at various concentrations and pH values. In general the amino acids are produced by microorganisms in aqueous media containing solutes like salts. considering the partial dissociation of the amino acids and the other solutes that form the aqueous systems. In this work we present experimental data on water activity. The UNIFAC-Larsen model and the Debye-Hückel equation were associated with this procedure to allow for the non-ideality of the systems studied. pH and density of some aqueous amino acid solutions were determined at 25 °C in three different types of solvents. water activity (aw) and the influence of different pH values on these properties in mixtures containing biomolecules. dl-alanine. etc. amino acids are of much research interest because of their simplicity and importance for understanding the behavior of molecules such as peptides and proteins in more complex solutions. The activity coefficients were estimated by the UNIFAC-Larsen model combined with the Debye-Hückel equation. and solubility). In the design of equipments and processes it is necessary the knowledge of some physical-chemical properties such as solubility.Introduction 61 3. Among the biochemicals. pH. Interaction energies between the charged species Na+ and Cl. Thus.and the specific groups of amino acids (COOH and NH2) were adjusted using experimental solubility data. l-arginine and l-proline in three different types of solvents: water. The iterative procedure proposed by Achard et al. Previous published experimental data on water activity and solubility of amino acids in aqueous solutions were used together with data from this work to test the applicability of a group contribution model. tartaric and malic acids for the prediction . taking also into account the partial dissociation phenomena of species in solution. organic acids. (1994a) was utilized for modeling the properties (aw. 3. Water Activity. pH and Density of Aqueous Amino Acids Solutions 3. pH and density of aqueous solutions containing the amino acids glycine. acid and basic buffers. concentration and purification processes of biomolecules.1 Abstract The water activity. Such approach has been successfully applied to other aqueous solutions of weak electrolytes like citric.2 Introduction The crescent advances in the biotechnological industry have raised interest for the development of efficient separation.

resulting in 3. 3.. In contrast to prior works the present study takes into account a more comprehensive data bank on different physical-chemical properties of amino acid solutions and model these properties considering the partial dissociation of weak electrolytes. . respectively.0038 glycine/0. The pH was measured with a Chem Cadet 5986-50 pH meter (Cole-Parmer Instrument Co. The buffers were prepared with citric acid/sodium citrate and glycine/NaOH.0084 citric acid/0.22 pH values. citric acid and sodium citrate) were also purchased from Merck and used to prepare the buffer systems.0043 (ratio: 0.09 and 9. The objective of the work reported here is to test and extend the procedure proposed by Achard et al.7±0.3. 2000.3 °C) at various solute concentrations.62 Meirelles. No further purification was employed. dl-alanine. 1998).0±0. Data on solubility of amino acids in water. The influence of pH on water activity was also verified by the use of acidic and basic buffer solutions.3 Materials and Methods 3. Germany) with precision of 1 × 10-4 g. USA) calibrated with buffer solutions (pH=4 and 7) supplied by Merck. Capítulo 3 of water activity and pH of those solutions (Maffia and Meirelles.1 Materials Samples of the amino acids glycine. Other chemicals of >99 mass % purity (NaOH.3. USA) using an analytical balance (Sartorius. and 0. and the results used later in the modeling of this property. 3. Velezmoro and Moreover.2 Experimental Procedure Water activity was measured at 25.1 °C and pH at room temperature (24. The compositions were accurate to ± 0.0029 sodium citrate) and 0. (1994a.0005 NaOH) total solid mass fractions.0001 approximately in mass fraction.02 pH units. l-arginine and l-proline were supplied by Merck with purity greater than 99 mass %.0113 (ratio: 0. The mean standard deviation obtained for the whole set of experimental data was 0. buffers and salt solutions were included in this data bank. An electric hygrometer AQUA-LAB CX-2 (Decagon.b) to the estimation of physical-chemical properties of amino acid solutions. the procedure mentioned above is further tested by using a comprehensive data bank on other physical-chemical properties of amino acid solutions available in the literature. Solutions were prepared gravimetrically with deionised water (Millipore.

1 to 3.14 0.0±0. Kuramochi et al. The corresponding experimental water activity data for glycine were compared to the data determined in the present work (see Figure 3.0 ± 0.08 0.1 °C by circulation of thermostated water from a water bath (Cole Parmer Instrument Co.12 0. . Measurements were made in triplicate with a reproducibility of ± 0. The accuracy of the density measurements was estimated as 3 × 10-5 and 3 × 10-4 g cm-3 for the digital densimeter and picnometer. The digital densimeter was calibrated with air and water as standards while the picnometer was calibrated with water at the corresponding working temperature.18 0. 1997b) 0.1. >99 mass % purity: potassium sulfate. Anton Paar.02 0.99811 this work (Kuramochi et al.3 Experimental Results Experimental water activity.95 2 R =0.04 0. 3.3. resulting in an average value of 0. USA).96 0. The mean relative deviations between both data sets were 0.001 aw units. 1. respectively.06 0. Austria) (for the systems amino acid + buffer) and a picnometer (for the systems amino acid + water) at 25.00 0.3.98 aw 0.101% for both amino acid solutions.Experimental Results 63 USA) previously calibrated with saturated salt solutions (analytical grade reagents from Merck.00 0.051% for alanine. sodium chloride and potassium carbonate) was used for measuring aw.10 0.1 °C. The temperature inside the hygrometer was regulated at 25. potassium chloride..99 0.1). pH and density data for the amino acid solutions are summarized respectively in Tables 3. Experimental density measurements were performed in the systems cited above for the conversion of solute mass fractions into molar concentrations.151% for glycine and 0.97 0.. Density measurements were carried out in triplicate using a digital densimeter (DMA 58.20 0.16 0. Water activities determined in the present work and data taken from ref 5 for the amino acid glycine. (1997b) measured the partial pressure of water for glycine and alanine aqueous solutions.22 wa Figure 3.

957 dl-alanine + acid buffer 0.0597 0.0253 0.866 0.988 0.0755 0.1481 0.2501 0.2002 0.0 ±0.0498 0.971 l-proline + acid buffer 0.987 0.991 0.1244 0.0721 0.0376 0. Experimental water activities of glycine.997 0.992 0.1503 0.995 0.0714 0.991 0.983 0.0252 0.0498 0.0309 0.1237 0.1000 0.971 l-proline + basic buffer 0.0500 0.988 0.991 0.1110 0.987 0.973 0.905 l-arginine + acid buffer 0.1770 0.980 0.4900 0.998 0.0748 0.3799 0.983 0.2021 0.994 0.0498 0.2942 0.0247 0.966 0.3355 0.998 0.996 0.1.990 0.0981 0.64 Capítulo 3 Table 3.1712 0.1242 0.1484 0.0997 0.1233 0.0573 0.993 0.0615 0.2951 0.988 0.978 0.988 l-arginine + basic buffer 0.1127 0.0494 0.983 0.1315 0.0545 0.923 0.0772 0.0873 0. dl-alanine.958 0.901 0.1195 0.0996 0.995 0.3497 0.994 0.0677 0.974 0.1227 0.0504 0.996 0.1239 0.0258 0.971 l-proline + water 0.981 0.0918 0.994 0.995 0.989 0.992 0.984 0.1353 0.959 0.964 0.993 0.1006 0.972 0.990 0.1233 0.970 0.0616 0.960 dl-alanine + water 0.975 0.0876 0.986 . at 25.982 0.979 0.0889 0.997 0.0530 0.990 0.985 0.944 0.0496 0.1 °C wa aw glycine + water 0.1123 0.0241 0.993 0.1210 0.1500 0.990 0.0997 0.992 0.0251 0.0945 0.0503 0.1713 0.0943 0.975 0.0996 0.989 wa aw glycine + acid buffer 0.886 l-arginine + water 0.985 0.0243 0.994 0.0863 0.0744 0.0640 0.956 dl-alanine + basic buffer 0.970 0. l-arginine and proline as a function of amino acid mass fraction (wa).799 wa aw glycine + basic buffer 0.4025 0.972 0.924 0.0997 0.993 0.957 0.981 0.996 0.0752 0.2003 0.927 0.983 0.985 0.0766 0.963 0.0961 0.992 0.0247 0.978 0.0495 0.992 0.2936 0.990 0.979 0.0759 0.990 0.0865 0.0505 0.982 0.

22 0.0258 0.0772 0.01 7.0640 0.3799 0.2942 0.99 5.0744 0.0766 0.0252 3.25 8.20 0.99 7.67 0.10 0.0498 9.25 11.20 11.99 0.25 10.0876 0.46 6.0759 0.79 5.09 0.13 11.0500 0.17 8.0309 0.12 10.04 0.61 8.15 6.0505 0.75 5.1127 0.78 5.0243 0.0530 0.64 0.08 10.58 8.0748 9.27 0.0573 0.0943 0.82 0.0918 3.0981 4.0863 9.1195 4. l-arginine and l-proline in water and in basic and acid buffers (wa=amino acid mass fraction) 24.91 5.32 6.0616 9.93 0.86 5.0997 0.71 5.1210 4.31 7.0945 0.2501 0.1713 4.1500 0.32 8.2021 4.0498 0.0496 3.92 0. dl-alanine.1244 10.81 8.2002 0.19 0.1239 6.0996 4.1481 4.84 0.0503 0.97 5.27 0.09 11.78 7.21 6.49 0.23 11.0498 3.0504 0.0755 0.58 8.81 5.0253 0.3355 8.52 .00 0.52 6.06 0.7 ±0.90 0.88 0.0615 3.17 11.1227 0.69 7. Experimental pH of glycine.0677 0.1233 9.0997 0.25 wa wa 0.96 10.01 11.2951 4.0996 0.95 acid buffer glycine 0.0247 3.0494 0.95 0.1503 0.0889 4.3497 0.92 5.0997 9.1233 0.Experimental Results 65 Table 3.0545 0.1006 0.4900 5.39 6.0721 0.1242 basic buffer 8.1353 pI 5.0873 0.46 dl-alanine 0.65 6.1712 0.2.0597 0.1000 0.11 8.36 0.0961 0.38 10.0865 0.0251 0.0752 3.80 5.14 l-arginine 0.75 0.3 °C wa 0.62 0.1110 10.96 5.88 5.1484 0.42 10.77 0.21 l-proline 0.1770 0.1123 0.52 0.4025 4.1237 0.21 8.0495 3.76 11.44 10.65 7.1315 4.0714 4.0241 0.04 8.2003 0.89 7.0376 0.84 0.2936 0.03 11.

0683 1.07590 dl-alanine + basic buffer 0.029 0.1094 1.01281 0.01572 0.0360 1.1352 1.14700 wt ρ (g cm-3) glycine + basic buffer 0.02576 0.5013 1.0022 0.02512 0.02597 0.07421 dl-alanine + acid buffer 0. 25.00991 0.03221 0.04393 l-proline + acid buffer 0.0861 1.0514 1.1031 1.03055 0.0961 1.01641 0.2942 1.0729 1.0048 0.0365 1.0616 1.0573 1.1527 1.e.0376 1.1239 1.0315 a l-arginine + acid buffer 0.01249 0.12111 0.0865 1.0551 0.0077 0.0371 0.05341 0.0250 0.02450 0.3.0360 1.0345 0.03661 l-proline + basic buffer 0.0219 0.00578 0.1308 1.02750 0.1109 1.08500 0.02283 0.0133 0.0118 0.1276 1.03332 0.0255 1.02796 0.0069 0.01783 0.1276 1.1008 1.0533 1.66 Capítulo 3 Table 3.04262 0.0250 0.1039 1.3799 1.2979 1. wt=wa+wb.0178 0.06098 0.0611 1.3064 1.01846 0.01896 0.0241 1.02453 0.06379 0.0283 0.1000 1.2756 1.2046 1.0286 1.1097 l-arginine + water 0. i.0137 0.04010 0.0352 1.0597 1.1500 1.1237 1.1002 1.1223 1.0505 0. l-arginine and l-proline as a function of total solute mass fraction (wt).1022 0.1040 1.0755 1.0588 1.05320 0.0718 dl-alanine + water 0.0283 0.1712 1.02043 0. .0986 1.0505 1.1826 1.1227 1.02777 0.1285 1.00626 0.0195 0.06450 0.09632 l-proline + water 0.0988 1.0609 1.0409 wt ρ (g cm-3) glycine + acid buffer 0.0728 1.03288 wt corresponds to total solute mass fraction considering the concentration of the amino acid (wa) and the solutes of the buffers (wb).08851 0.3398 1.0398 0.2002 1.0280 1.1157 1.1 °Ca wt ρ (g cm-3) glycine + water 0.0787 1.02144 0.0908 1.01379 0.4138 1.1323 1.1357 1.05680 0.0766 0.0876 1.1594 1.0 ±0.0759 1. dl-alanine.0301 1.0976 1. Densities of aqueous solutions of glycine.02052 0.0764 1.0053 0.03185 0.0997 1.0827 1.0187 1.0611 1.03467 0.1813 1.0590 0.01283 0.02416 0.01458 0.03870 l-arginine + basic buffer 0.0815 1.04395 0.0841 0..111 1.1127 1.1428 1.0608 1.0402 0.03282 0.2134 1.3497 1.0541 1.1503 1.00848 0.

base or salt. . The addition of buffer solutions containing a strong electrolyte (NaOH) or an acid/salt (citric acid/sodium citrate) solution gives rise to interactions between the amino acid groups and the ionic species from the acid. Dalton and Schmidt.02 1. As a consequence. 1998.0043 for acid and basic buffers. 1933) and determined in this work. 1998) this work dl-alanine: (Dalton and Schmidt. there was no significant interference of the buffer solutions upon the aw data measured in the present work through short-range type interactions. proline and alanine.00 0.004% for aqueous solutions of glycine and dl-alanine. it was observed that higher concentrations of the cationic and anionic species in solution lead to a slightly decrease in the aw values.04 1.10 wa 0. the agreement is very satisfactory (mean relative deviations equal to 0.98 0.0 °C reported in literature (Soto et al.20 Figure 3. Therefore.0113 and 0.07% and 0. 1933) this work linear regression 1.05 0. 1. at 25..10 1. Figure 3.. The presence of buffers containing ionizable species contributes to changes in the net charges of the amino acids according to pH value. As can be seen. glycine.0 °C.2 shows a comparison between density data for aqueous solutions of dl-alanine and glycine at 25. respectively) yielding aw values equal to 0. Comparison between experimental densities from literature and those determined in the present work.2.999 measured by the electric hygrometer. respectively). The buffer solutions were used in very low concentrations (total buffer solutes weight fractions: 0.00 0.15 0.08 1.06 density (g cm ) -3 glycine: (Soto et al.Experimental Results 67 A very slight influence of pH on aw values was noted for arginine.

and for the cationic form: α-CH2. 1997a. NH2. 1990. COO-. NH2 and COOH. 2×scCH2 sc=side chain. combined with the Debye-Hückel term. 1994). 2×scCH2.4 Thermodynamic Modeling 3. (1998) using a numerical treatment of the infrared spectra data. C(NH2)NH. α=group bounded to the α carbon in the molecule.. In Table 4. (1996.when dissolved in water according to the pH of the solution. Division of groups for the amino acids used in this work amino acid glycineb l-arginine dl-alanine dl-valine serine l-proline a groupsa α-CH2.. Gupta and Heidemann. the distribution of the different ionic forms of the amino acid glycine as a function of pH was obtained by Max et al. 1994). α-CH. scCH3 NH2. COOH. the nonideality of such systems was represented either by local composition models (Nass. Kuramochi et al. sc-CH2 (side chain alkane). 1994.. Table 3. b For the zwitterion glycine the groups are α+ + CH2. OH CH2NH.4. based on the UNIFAC-Larsen model. COOH α-CH.. To account for the electrostatic forces in the amino acid solutions. the constituent groups of the amino acids considered in this work are given. Pinho et al. COOH. anionic and zwitterion . In several publications. NH 3 . 1997a). For example. CH2NH NH2.. 1999).. attention has been given to the modeling of phase behavior in aqueous systems containing biochemicals like amino acids (Khoshkbarchi and Vera. NH2.4. 1980) has been added to the activity coefficient models in some reported works in the literature (Pinho et al.cationic. The following species were . 1997. Pradhan and Vera. Peres and Macedo.1 Modified UNIFAC Model for Electrolytes Capítulo 3 Recently. 1989) or by group contribution models (Pinho et al. for the anionic: α-CH2. Chen et al. NH 3 . Ninni et al. COOH. The amino acids appear in various ionic forms . COO-. COOH. 2×scCH3 NH2. NH2. 1998.68 3. α-CH. 1994. 1996. COOH. Amino acids molecules were divided into several main groups including new group assignments proposed by the cited authors: α-CH2 (alkane bounded to the α carbon atom in the amino acid molecule). α-CH. scCH. scCH2. α-CH. 1988. the DebyeHückel term (Pitzer. In this work we used the model presented by Kuramochi et al. COOH. 1996.

1) (3. King. To estimate the concentrations of the different ionic forms present in solution.106). The method takes into account the partial dissociation phenomena of the amino acids and combines equilibrium relations.8) + Reaction 3.6) (3.1 corresponds to the formation of the zwitterion ( NH 3 RCOO − ) with dual electric charges..⇔ H + + citric --- Na 3 citric ⇒ 3 Na + + citric --H 2 O ⇔ H + + OH − Kw (3. which can originate the amino acid cationic and anionic species (reactions 3.) can be obtained as: 10 − pK . 1937.3) (3. Values for pKi and pKw used in this work were found in (Greenstein and Winitz.7) H 3 citric ⇔ H + + H 2 citric H 2 citric . The equilibrium constants (Ki and KAi. the following series of reactions takes place simultaneously when an amino acid is dissolved in an acidic buffer containing citric acid ( H 3 citric )/sodium citrate ( Na 3 citric ): + NH 2 RCOOH ⇔ NH 3 RCOO − + + NH 3 RCOOH ⇔ NH 3 RCOO − + H + + NH 3 RCOO − ⇔ NH 2 RCOO − + H + KD K1 K2 KA1 KA2 KA3 (3. 69 observed for glycine in water: cationic (pH 0 to 5). This means that the uncharged amino acid is almost completely converted to the zwitterion form.2 and 3. as well as equilibrium constants describing the chemical equilibrium. In aqueous solution.4) (3. (1994a) was utilized in the present work.5) and anionic (pH 7 to For modeling the amino acid systems it is assumed that the constituent groups of the anionic and cationic species are the same as those of the zwitterion form (see Table 3. Izatt and . mass and electroneutrality balances for calculating the true concentrations of the species and their activity coefficients.⇔ H + + Hcitric -Hcitric -. 1961.3. 1951. the iterative procedure proposed by Achard et al. Smith et al.2) (3.4).Thermodynamic Modeling 14). The equilibrium constant KD is generally very large (105 . 2. This set of equations requires initial solute and water concentrations. respectively).. zwitterion (pH 0 to 12.5) (3.. Structural parameters Rk and Qk for the ionic groups of the amino acids were considered equal to the neutral ones. i=1.

009 0.7 -37.7 1077. the effect of the temperature on the equilibrium constants were taken into account by adjusting eq 3.5 deviationb 0.2 -44.2641 3. exp  n  n pK n    ∑ Reactions 3. B and C are adjustable parameters and T the temperature in Kelvin.2839 3.5.9489 3. b Mean relative deviation (%) between estimated and exp cal  pK n − pK n  100  × experimental values calculated by:  . For the amino acids and water.1010 1594. pK i = A + B ( T + C) (3. temperature in K.where n=number of experimental data points.2125 2.6254 1663. Dissociation constants (pKAi) for citric acid at 25 °C were found in (Izatt and Christensen.4 to 3. 1970) and their values are following: pKA1=3.011 According to eq 3.011 0.1443 1.762. pKA3=6.11 -229. 1970).071 0.3074 1.003 0.5850 8.0618 B 3.73 -116.131. Values of A.9823 2.4 -55.9 to the available experimental data. Eleven species are present in the mixture mentioned above.9) where A.9 0. mass and electroneutrality balances) that allows to calculate simultaneously the activity coefficients and the concentration of species is given below: .14 -47.88 -252.1 0. pKA2=4.010 0. The system of equations (equilibrium relations.397.008 0.3 -0.7 correspond to citric acid and sodium citrate dissociations in the aqueous system.1900 2232.5.2012 -276.9 in the text. Table 3. Coefficients for correlating pK values with temperaturea substance glycine dl-alanine l-proline dl-valine dl-serine water a A pK1 pK2 pK1 pK2 pK1 pK2 pK1 pK2 pK1 pK2 pKw 2.2984 3.4 13. B and C are given in Table 3.015 0.04 -271.272 0.004 0.0430 1650.944 1724.1 3.155 0.3 C -256.1602 2.70 Capítulo 3 Christensen.

.0 = c H 2O + c OH − L oL where c in the equations above are the molar concentrations (mol L-1) of the species.18) 2 RCOO − c Na 3citric.13) K A3 = c H + c citric .2 (3.10) K2 = c NH 2 RCOO − cH+ γ * (NH 2 RCOO − ) γ * H+ L vm oL vw c NH + RCOO − 3 γ * H+ * + (NH 3 RCOO − ) (3.0 = c NH + RCOO − + c NH + RCOOH + c NH 3 3 3 (3.15) electroneutrality balance c H + + c NH + RCOOH + c Na + = c OH − + c NH 3 2 RCOO − + c H citric − + 2 c Hcitric − − + 3 c citric − − − 2 (3..12) K A2 = c H + c Hcitric .γ c H citric 2 γ * Hcitric . ..Thermodynamic Modeling equilibrium relations 71 K1 = c NH + RCOO − c H + γ 3 * + (NH 3 RCOO − ) γ * H+ L vm oL vw c NH + RCOOH 3 γ * + (NH 3 RCOOH) (3.14) K w = c H + c OH − * L γ OH  v m  − 2 aw  oL  v   w  (3.16) mass balances c H 3citric. v m and v w are the molar volumes of the mixture and water (L mol-1).0 + c Na 3citric.+ c Hcitric .- γ * citric .- L vm oL vw γ * H+ * H 2 citric (3.11) K A1 = c H + c H citric − γ 2 γ * H 2 citric − L vm oL vw c H 3citric γ * H+ * H 3 citric (3.- L vm oL vw γ γ * H+ * Hcitric (3.γ c Hcitric .20) c H 2O.17) (3..0 = c H 3citric + c H citric .19) (3.0 = c Na + c NH 2 RCOOH = c NH + RCOO − . respectively and γ i* refers to the activity coefficient of specie i in the asymmetric convention (infinite dilution basis).+ c citric .. The subscript 0 refers to the initial concentration of solute.

and activity coefficients are estimated considering the occurrence of solvation. mole fractions of hydrated species xw − x H w = ∑ Nh x 1 − ∑ Nh x j j. Nhk is the hydration number of specie k. the activity coefficient (in the asymmetric convention) of a specie i is expressed by long-range (electrostatic interactions) and short-range (physical interactions) contributions: ln γ * = ln γ *S − R + ln γ *L − R i i i (3. structural parameters of solvated species.23) (3. concentrations.and long-range interactions. . and Rw and Qw stands for volume and area parameters of water. the short-range contribution combines the UNIFAC-Larsen model with solvation equations for chemical interactions between water and ionic species.25) xiH = 1− ∑ xi Nh j x j (3. j≠w j j j. j≠w j (3.25 and 3.22) In addition. j≠w H where x w and x iH are the mole fractions of water and ionic species i. the activity coefficient is calculated in the symmetric convention as follows: L ln γ w = ln γ S − R + ln γ w− R w (3. respectively. For the solvent (water). As a result.26) j. The equations corresponding to this phenomenon are presented below: structural parameters of species k RkH = Rk + Nhk Rw QkH = Qk + Nhk Qw (3.24) where superscript H refers to the hydrated state.26 should be performed upon all species j in solution except for water. the superscript S-R and L-R refer respectively to short. respectively.72 Capítulo 3 In this work. The summations in eq 3.21) where.

The solution non-ideality is considered in the eqs 3. as suggested by Achard et al. activity of H+ ions in molar concentration scale can be obtained from the calculated activity in molar fraction scale by the equation below:  *  γ * c  = − log  γ H + xH + pH = − log c + H + oL  H     vw      (3. H and γ iS − R . (1987) gives symmetric activity coefficients in the mole fraction scale ( γ i ) that have to be normalized to the standard state for the ionic species (asymmetric convention): ln γ i* = ln γ i − ln γ i• where γ i• is the activity coefficient at infinite dilution of all solutes in the mixture: γ • = lim γ i i x w →1 (3.H w w H xw xw 73 (3. H xiH S − R . H are calculated by the UNIFAC-Larsen equation using mole fractions and w structural parameters of the hydrated species.30) where xw is the water mole fraction.29) (3.15 in which true concentration ratios can be estimated using the framework developed by Achard et al.1. (1994a). (1994a).Thermodynamic Modeling conversion between hydrated and non-hydrated activity coefficients γ S − R = γ S − R . with the closest approach parameter equal to 17. The long-range contribution to the activity coefficients was calculated using the DebyeHückel equation. In the particular case of pH.31) . It follows that the pH and aw values can be estimated using the true concentrations of the ionic species (aw = γw * xw and pH = − log[γ c + c H H+ ] ).28) γ S − R .27) γ iS − R = γ iS − R . The selected UNIFAC model proposed by Larsen et al.10 to 3. H H xw γw xi [ ] − Nhi (3.

7173b.77b.96c 1476. 2.1c. oL and v w is the molar volume of water. 1115.003a 1654. 1217.33 were adjusted by regression from the solubility data of the amino acids dl-alanine. it is necessary to perform an iterative method to obtain the solubility of that molecule using eqs 3. 2717. The activity coefficient γ A± can be estimated by the UNIFAC-Larsen model combined with the Debye-Hückel equation. 885. 1990. 1107.042d 1. and f AS is the fugacity of the pure solid amino acid.32 and 3.6). 1996. l-proline and dl-serine.1305b. 1711. 2. 0. 1.96b.97d. 654. 754. The ratio f AS /f oL can be related to the temperature of the system through the following relation: S fA  ∆s ∆h  = exp −  L fo  R RT  (3. Therefore.305d 2.33) where ∆s and ∆h are the change in molar enthalpy and entropy of the amino acid from the reference state to the solid state (Prausnitz et al. 2641.3628a. b Khoshkbarchi and Vera. 959.6a.65a 5. Table 3. d Gupta and Heidemann.4. 2. cFasman.. 1996).74 Capítulo 3 where c H + and x H + stand for molar concentration and molar fraction of ions H+.96a. 4. These values are generally obtained from * solubility studies. .61b.2c Adjusted in this work.9c.689d 1390.7c.6235b.2127b. The values of ∆s and ∆h for the other amino acids studied in this work (glycine and dl-valine) were maintained equal to the literature values (see Table 3. ∆s e ∆h values used in this work (in bold script) and obtained from literature amino acid glycine dl-alanine dl-valine dl-serine l-proline a ∆s/R ∆h/R (K) 2.55b. can be written in terms of the zwitterion form: xA± γ * A± S f oL = f A (3.33.35b.2c 1264. respectively. 3. 2. 1976.07b.946d 2265.2 Solubility of Amino Acids in Aqueous Solutions The solid-liquid equilibrium of an aqueous amino acid solution and the pure solid amino acid.4184b.6. 2.060d -0. once the mole fraction of the amino acid is known.1835a.32) f oL is the reference fugacity in the asymmetric convention. ∆s and ∆h in eq 3.1110.

and NO3-) on the solubility of dl-alanine at 25 °C within the pH range of 2 to 10.are the mole fractions of the zwitterion. Cl-. the solubility of amino acids in aqueous electrolyte solutions has been estimated using the UNIFAC-Larsen model combined with the Debye-Hückel equation.35: ρi (t ) = ρi (25o C ) ρ w (25o C ) ρ w (t ) (3.35) .7. it is necessary to estimate the densities of the amino acid solutions. (1971) shows the influence of acid or base on the solubility of amino acids in two different solvents: water and ethanol.5 Results and Discussion 3. Pradhan and Vera (1998) reported the effect of the addition of strong acids and bases on the solubility of amino acids at various pH values. 3.1 Density data In the calculations of activity coefficients and true concentration of species. 1996) experimental solubility data of some amino acids at various pH and NaCl concentrations. In the present work. bases and salts on the aqueous solubility of amino acids within a certain pH range. The pH was varied from 0 to 13 using HCl or NaOH. Linear equations were fitted to experimental density data at 25 °C for the amino acid systems considered in the present work. There are several literature data reporting the influence of acids. It is also reported in literature (Carta and Tola. We used the following approach to estimate such densities: the densities of the amino acid (i) solutions at 25 °C relative to water (subscript w) at the same reference temperature were multiplied by the density of water at the desired temperature. The equations are given in Table 3. To estimate solubilities of the amino acids in water. x A ± . as can be seen by eq 3.34) where. cationic and anionic species. bases and acids. Na+. x A + . x A . The experimental data shows the dependence of the solubility of amino acids on the addition of salts.5. density data at temperatures other than 25 °C were also required. It was observed that there was no significant influence of different ions (K+.Thermodynamic Modeling The total solubility (xA) can be written as: xA = xA± + xA+ + xA- 75 (3. Experimental data reported by Needham et al.

9970 1. The prediction capacity of eq 3.0599wt +1.9970 0.9972 0.9971 0. (1998).9974 0.9998 0. salts.9970 0.9962 0.9975 0.48 4.9977 0. acid and base in aqueous solutionsa system equation: ρ=ρ(wt) g cm-3 glycine + water glycine + acid buffer glycine + basic buffer dl-alanine + water dl-alanine + acid buffer dl-alanine + basic buffer l-proline + water l-proline + acid buffer l-proline + basic buffer l-arginine + water l-arginine + acid buffer l-arginine + basic buffer l-serine + water dl-valine + water NaOH + water HCl + water NaCl + water a Table 3.64 3.9970 0.20 mass % for KCl).9970 0.76 Capítulo 3 Linear fitting for the densities of amino acids.3241wt +0.80 4.3072wt +0.60 4. 1989) at various temperatures (20 .9967 0.52 3.4327wt+0.60 3.100 °C for KCl) and solute concentrations (1 .9994 0.53 3.42 this work this work this work this work this work this work this work Yan et al.80 4.999980 given as 4.7585wt +0.3258wt +0.7 .32 4.7.4229wt +0.20 3.16 3.80 3.3073wt +0.80.9980 Soto et al.36 4.4450wt +0.4349wt +0.2781wt +0. .62 3.5037wt +0.35 was tested using experimental densities of binary aqueous solutions containing a non-electrolyte solute (sucrose) (Norrish.2988wt +0.2782wt +0.2982wt +0.2935wt +0.85 mass % for sucrose and 0.88 4. this work this work this work Dalton and Schmidt (1933).9974 0. reference R2 0. 1967) or an electrolyte solute (KCl) (Lobo and Quaresma.2276wt +0.0013 0. (1999) Dalton and Schmidt (1933) Lobo and Quaresma (1989) Lobo and Quaresma (1989) Lobo and Quaresma (1989) Correlation coefficients such as 0.60 4.90 °C for sucrose and 25 . this work this work 3.

1999). The interaction parameters used were found in Achard et al.3% for the ternary systems: glycine + NaCl + water and glycine + KCl + water. Kuramochi et al.2 Equilibrium data The UNIFAC-Larsen model was then used to calculate the short-range contribution to the activity coefficients of the different species in solution considering the hydration of ions H+ and Na+. ρw. 1997a) (for interaction parameters involving new groups such as sc-CH. respectively (Achard et al. or amino acid + salt + water) and the quaternary (amino acid + base/acid + salt + water) systems it was utilized a semi-empirical equation for calculating the corresponding densities.and the specific groups of the amino acids (COOH and NH2) were .35 were equal to 0. 3. 1989.1% for sucrose and KCl solutions. Interaction energies (uij) between the charged species Na+ and Cl.5.i ≠ w w ) (3. (1994a) (for interaction parameters involving charged species). Such result clearly shows the good performance of the selected approach to estimate densities at various temperatures. The relative mean deviation was found to be 0..606. respectively. For the ternary (amino acid + base/acid + water.35 is based on the assumption that dρi/dwi (the derivative of solution density in relation to solute concentration) does not dependent on temperature. so that the solution density at some specified solute concentration changes with temperature only as a consequence of the change in the solvent density. Component i could be a salt. This result confirms the applicability of the equation for calculating the densities of the multicomponent systems involved in this work.959 and 2. It requires densities of the binary solutions (solute + solvent) as is shown below (Zafarani-Moattar et al. Soto et al. with hydration numbers (Nh) equal to 2.36 should be performed upon all the components i in solution..36. (1987). The summation in eq 3.36) ρmix.09% and 0. 1994a). density of pure water and density of the binary aqueous mixtures containing component i. α-CH) and Larsen et al.Results and Discussion 77 In the case of sucrose solutions the reference temperature was 20 °C and for KCl solutions it was 25 °C. Mean relative deviations between experimental density data and those calculated by eq 3. 1995): ρ mix − ρ w = ∑ (ρ − ρ i i . Eq 3. (1996.. calculated densities for aqueous solutions of glycine and salts were compared with experimental ones from literature (Lobo and Quaresma. To test the accuracy of the eq 3. respectively. except for water. an acid or a base. and ρi represent the densities of the multicomponent mixture.

.4 depicts the estimated pH values for mixtures containing glycine in three different solvents (water. 1987). a NH 2 .37. and the values for u w . Values for a COOH . w .38) were obtained in ref (Achard et al.36 -1521. allows to reduce the number of parameters to be fitted. Group interaction energies ujj between like groups ujj (K) COOH-COOH NH2-NH2 -622. As can be observed.8. w . 1994b) and are reported in Table 3. Group interaction energies uij (K) between ions and amino acid characteristic groups Na+ COOH NH2 442. Table 3. Cl - (3.COOH and uNH2. The option for adjusting the interaction energies uij.78 Capítulo 3 adjusted using the available solubility data. 1994b). a w. l-arginine and dl-alanine are given in Table 3.724 Cl-1459. u Na + .NH2) required for aij calculations had been estimated by the following expression (Achard et al..18 -878..10. the UNIFAC-Larsen model combined with the Debye-Hückel term yields a good fit to these data. given below: a ij = u ij − u jj (3.37) Adjusted group interaction energies (uij) between the amino acid groups (COOH and NH2) and the ions Na+ and Cl.785 357.are given in Table 3.9: u jj = a jw − a wj + u ww where the subscripts w and j stand for water and specie j.9. . The ujj values (uCOOH. respectively. NH 2 were taken from Larsen's tables (Larsen et al. Interaction parameters aij are related to the interaction energies uij and ujj according to eq 3. a w.3 it is shown a comparison between experimental and predicted aw values for the amino acid l-proline in water and in very diluted buffer solutions.w .COOH . lproline. Figure 3. acid buffer and basic buffer)..10 In Figure 3. Na + and u Cl.50 Table 3.8. instead of the interaction parameters aij. Mean relative average deviations for the systems with glycine.

5.14 0.4. The deviations for the solubility data are also given in Table 3.12 -1 pI acid buffer basic buffer model 0.5 4.5 5.5 0. Experimental and estimated pH of glycine solutions.0 3.00 79 0.3.0 7.5 7.80 Figure 3.3 0.0 2.5 3.18 0. .04 0.02 0.0 6. The model reproduces well the observed experimental data.4 0.5 8.85 0.0 pH 5.90 aw pI model acid buffer model basic buffer model 0. Experimental and calculated aw of aqueous solutions containing l-proline.0 0.08 0.95 0.5 6.5 0.2 w 0. and some of the results are shown in Figure 3.1 0.06 0.0 4.16 0.Results and Discussion 1.10 0. 8.20 wa (g g ) Figure 3.10.

0 0. 1998) report the final pH of a mixture containing the amino acid.0 0 20 40 dl-valine 60 t ( C) o Capítulo 3 l-proline glycine dl-alanine dl-serine 80 100 120 Figure 3. the concentrations of acid or base were utilized to estimate solubility of some amino acids in a wide range of pH values. the concentrations of acid or base were selected when the pH estimated by the model reached the experimental pH.6.6 wa 0.3 0. when the pH values of the initial solutions were reported.6 shows the results of the calculated solubilities of the amino acid glycine at various pH values and NaCl concentrations. while the . Most experimental data are from the work of Carta and Tola (1996). Figure 3. In the other case. 1996) provide the pH of the initial solutions – before the addition of the amino acid.4 0.5. For example.6. Results concerning solubility of glycine in salt solutions (glycine + NaCl + water + HCl/NaOH) at different pH values are presented in Figure 3.7 0. it can be seen that good agreement is observed between experimental and calculated solubilities at high NaCl concentrations (5 and 15%). the amount of acid or base in the present work was selected – through analysis of the estimated pH by the model – to obtain the desirable pH or when pHcalc=pHexp.80 1. the total acid or base concentrations were selected from the analysis of the experimental and estimated pH for aqueous solutions containing the acid or the base only. Finally.2 0. It must be stressed that studies about pH influence on solubility of amino acids do not report the concentrations of acid and base utilized. From Figure 3.1 0. while others (Carta and Tola. when experimental solubility data as a function of final pH is given for a mixture containing the amino acid + acid or base.9 0. Some works (Pradhan and Vera. the acid or the base.5 0. In both cases.8 0. Experimental and calculated solubility of some amino acids.

1933).26 0. Dalton and Schmidt. Such deviations are outside the range of the experimental error for aw measurements provided by the manufacturer (0. 81 calculated results for the amino acid at the isoelectric point agree well with experimental data Table 3. that is activity coefficients for all species equal to 1 ( γ i . the aw values calculated for the ideal case had an expressive difference from experimental results for the systems containing the amino acid proline (mean relative deviations of about 1.10 also shows the mean relative deviations for aw. Experimental and estimated solubilities of glycine at various pH values and NaCl concentrations. 1971.22 wa 0. .1%)..3%) and also much higher than the standard deviation observed in the experimental measurements (0.4%).28 Carta and Tola (1996) Carta and Tola (1996) Carta and Tola (1996) Needham et al. (1971) Dalton and Schmidt (1933) 0% NaCl 5% NaCl 15% NaCl 0. alanine and arginine this difference was less pronounced but greater than that obtained considering the non-ideality of the systems.Results and Discussion from (Needham et al.6. pH and solubility of the amino acids in aqueous solutions considering the ideal case.∀i =1). For the other systems containing the amino acids glycine. It was observed that. 0.16 0 2 4 6 pH 8 10 12 14 Figure 3.24 0.18 0.20 0.

10. 1997b this work this work 3.43 5.40 this work Kuramochi et al.55 6.63 2. 1997 Khoshkbarchi and Vera.26 4.99 3.73 6. 1997 0.60 11.47 7.13 0.90 14.99 7.94 solubility proline + water alanine + water glycine + water valine + water serine + water glycine + NaCl + HCl or NaOH alanine + NaCl valine + NaCl serine + NaCl alanine + HCl or NaOH glycine + HCl or NaOH total deviation 0.. 1998 Needham et al.61 3.63 Khoshkbarchi and Vera. 1996 Pradhan and Vera. 1976 Fasman.15 pH pI acid buffer basic buffer 3.12 0. 1971 .96 0.06 17.76 17.26 14.05 8. 1976 Fasman.91 this work this work this work 4. Average relative deviations between calculated and experimental data property % deviation real aw reference ideal pI this work literature acid buffer basic buffer 0.44 28. 1997 Khoshkbarchi and Vera.76 12.51 0. 1976 Fasman.93 0.12 0.11 0.08 0.77 2.32 Fasman.18 0.18 6..64 12. 1976 Carta and Tola.51 4.41 59. 1976 Fasman.82 Capítulo 3 Table 3.66 3.

it was verified slight differences between them. In the case of glycine and valine the values of ∆s and ∆h were taken from literature and then were not adjusted in the present work even though the results of the calculated solubility considering the ideal case presented very high deviations from experimental data.Results and Discussion 83 Concerning the deviations between real and ideal pH values. 1.0 relative concentration 0. numerical data for the distribution of the species is not available in the literature.5 and 8-11. This case can be further visualized in Figure 3.5 and anionic pH 7 to 14).0 0 2 4 6 pH 8 10 12 14 Figure 3. Large discrepancies can also be noted between calculated (ideal case) and experimental data for solubility (see Table 3. Comparison between the distribution of the ionic species of glycine as a function of real () and ideal (----) pH.8 0.2 0.7 that the ranges in which such species appear in the various forms agree well with those reported by (Max et al.4 0.7. Unfortunately. Although it is not possible to verify which of the curves (real or ideal pH) represent well the ionic distribution of glycine.10).6 NH 3RCOOH + NH 3RCOO + - NH2RCOO - 0.. 1998) (cationic pH 0 to 5. but it can be noted in Figure 3. those curves yield different concentrations of the species within the pH range of 1-3.7 in which the distribution of the three ionic species of glycine as a function of pH is shown. zwitterionic pH 0 to 12. .

G. 1994a.-B. Achard.89. J.and the amino acid specific groups were adjusted. C. Y. l-Tyrosine. and Glycine in Aqueous Solutions at Various pHs and NaCl Concentrations. the Densities of their Solutions at Twenty-Five Degrees.7 References Achard. Prediction of pH in Complex Aqueous Mixtures Using a Group Contribution Method. The interaction parameters between Na+ and Cl. 414-417.84 Capítulo 3 3... Greenstein. .. The model provided good results for such properties. pp 486-491.. CRC Press: Cleveland. Gupta. R. Physical and Chemical Data. pH and solubility of amino acids. J. C. Dussap. Data 1996. Fluid Phase Equilibria 1994b. C.. Evans. Dalton.. 40. 103. B. Chemistry of the Amino Acids Vol.. A. R. 36. J. 549-578. AIChE J. Phase Partitioning of Biomolecules: Solubilities of Aminoacids. C. Prog. AIChE J. 1933. Chem. l-Leucine. 333-341. 41. L. John Wiley & Sons: New York. pH and density of mixtures containing amino acids. M.. P. 1976. G. D. 1. 1989. Biol. Tola. Gros. (1994a) combined with the Debye-Hückel term. Zhu. Chem. 3rd ed. 111-118... was extended to estimate physical-chemical properties such as water activity. pp 115. Solubility Models for Amino Acids and Antibiotics. Dussap. The use of the UNIFAC group contribution model with solvation equations as proposed by Achard et al. B. Fasman. J. J. 5. The Solubilities of Certain Amino Acids in Water. Representation of Vapour-Liquid Equilibria in WaterAlcohol-Electrolyte Mixtures with a Modified UNIFAC Group-Contribution Method. yielding reasonable results concerning pH influence on the solubility of the amino acids in more complex mixtures. Solubilities of l-Cystine. R. 98. Handbook of Biochemistry and Molecular Biology.. Carta. 3. and the Calculated Heats of Solution and Partial Molal Volumes. 1961. Winitz.-G. 1210-1222. C. L.-G. Eng. B. Gros. J. 71. C. Biotechnol. B. Schmidt. C. Heidemann. 1990. Chen.6 Conclusions This work has presented experimental data on water activity. A.

M. Res. The Ionization Constants of Glycine and the Effect of Sodium Chloride upon its Second Ionization. Khoshkbarchi. J. Chem. Effect of NaCl and KCl on the Solubility of Amino Acids in Aqueous Solutions at 298. 1970.. 582-587. Data 2000. Khoshkbarchi. 1818-1819. Elsevier Sci.. E.46. Nagahama. Kuramochi. Applied Spectroscopy 1998. Am. J. D. A. H. Prog. H.. Eng. 470-474. J.. Noritomi. D. Sober. 1989. pp J 79. A. H. Rubber Co. D. Maffia. 371-379. J. 1996. A. Res.. H. pp 1643. M. M. 35. Measurements of Vapor Pressures of Aqueous Amino Acids Solutions and Determination of Activity Coefficients of Amino Acids. R. 2nd ed. Chem.. 1987.References 85 Izatt. Noritomi. Ind. Soc. Infrared Titration of Aqueous Glycine. 130. 4319-4327.. 36. 2274-2286. Nagahama. H. Chem. K. L. Vera. J. Eng. 42. Heats of Proton Ionization. K. J. Water Activity and pH in Aqueous Polycarboxylic Acid Systems. D. In Handbook of Biochemistry: Selected Data for Molecular Biology. K. Eng. Eng. J.. Lobo.. Measurement of Solubilities of Two Amino Acids in Water and Prediction by the UNIFAC Model. Max.. C. Chem. ..: Cleveland. M. 1951. 2445-2451. 52. 12. 155-159. A Simplified Pertubed Hard-Sphere Model for the Activity Coefficients of Amino Acids and Peptides in Aqueous Solutions. Fluid Phase Equilibria 1997a. pK and Related Thermodynamic Quantities. Chem. Chem. Kuramochi. 1997. Ind. J-J.. Nagahama.. Christensen. Fredenslund. Hoshino. Publishers: New York. K. 1996.. H. 73. Kuramochi. Trudel M. Eng. L. M. Data 1997b.. K. The Chem. Rasmussen. V. Vera. B. Res. Quaresma. Biotechnol. 226-233. Larsen. M. King. Ed. J. J. Meirelles. C. Noritomi. Hoshino.. 117-132. H.. Representation of Activity Coefficients of Fundamental Biochemicals in Water by the UNIFAC Model. 26. Handbook of Electrolyte Solutions. Ind.. H. Chapados. P. parts A and B. A Modified UNIFAC Group-Contribution Model for Prediction of Phase Equilibria and Heats of Mixing. H. A.2 K: Measurement and Modeling.

Chem. E. C. 1967. M. Effect of Cation and Anion of an Electrolyte on Apparent Molar Volume. Peres. A. Journal of Pharmaceutical Sciences 1971. A. C. Water Activity in Solutions Containing Organic Acids. 60. J. Soto. Vera. Scientific and Technical Surveys.. The British Food Manufacturing Industries Research Association: Leatherhead. S. 33. Khoshkbarchi. Chem. 1998. Biol.86 34. P. R. K. Soc. Azevedo. Alberto. 565-567. R. 1789-1805.. Biophysical Chemistry 1999. pp 26. Modeling and Prediction of pH and Water Activity in Aqueous Amino Acid Solutions. Silva. 74. P. 165-173. E. Comp. 73-82. E. K. Solubility of Amino Acids: a Group Contribution Model Involving Phase and Chemical Equilibria. 1996... Pradhan. Velezmoro. A. Electrolytes: From Dilute Solutions to Fused Salts. A. Eng. Eng. 1999. Taylor. 1937. Sci. AIChE J. 1988. A. Needham. A. N. S. A. Macedo.. Soto. Representation of the Solubility Behavior of Amino Acids in Water. A. 1980. 76.. Smith. E. Solubility of Amino Acids in Pure Solvent Systems. Paruta. J. M. Eng. 109-123. A.. T. 1994. S. 2902-2906. Drying Technol. A. L. A. H. Representation of Solubilities of Amino Acids Using the UNIQUAC Model for Electrolytes. Khoshkbarchi.. Experimental Data and Modelling of Apparent Molar Volumes. Meirelles. 1257-1266. E. J. M. Englewood Cliffs: New York. Chem. 1994. Macedo.. Gerraughty. G. Ninni. Molecular Thermodynamics of Fluid Phase Equilibria. Meirelles. Smith.. J. pp 418-420. M. B. A. 121-132. Am. A. K. Lichtenthaler. J.. A.. 1341-1347.. Prausnitz. Biophysical Chemistry 1998. Chem. Isoentropic Compressibility and Refractive Index of Glycine in Aqueous Solutions. Chem. 49. Norrish. C. 23 (Supplement). K. J. M. Capítulo 3 Nass. 152. Fluid Phase Equilibria 1998. Res.. M. Camargo. . Pitzer. Thermodynamic Properties of Solutions of Amino Acids and Related Substances. J. Pinho. 102. A. 383-386... Effect of Acids and Bases on the Solubility of Amino Acids. 16. R. Ind. Alberto. Isoentropic Compressibilities and Refractive Indices in Aqueous Solutions of Glycine + NaCl. R.. K. K. E. Selected Tables of Physical Properties of Sugar Solutions. N. 3803-3812. S.

Liu.. M. J. 145941/99. M. Thermochimica Acta 1999. Wang.8 Acknowledgment This work was supported financially by research grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP . Volumetric Properties of PEG + Salt + Water. Kabiri-Badr. Zafarani-Moattar.Proc. J.. 17-27. S. 3.15 K. 334.Proc. 466680/00-7 and 521011/95-7). J.. 559-562.References 87 Yan. Salabat. T.. 40. Apparent Molar Volumes and Viscosity BCoefficients of Some Alpha-Amino Acids in Aqueous Solutions from 278.. W. 98/12302-3) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq . A. N. Lu. . Z.15 to 308. Journal of Chemical Engineering Data 1995. B.

Brazil Trabalho publicado na revista Journal of Chemical & Engineering Data. .Faculty of Food Engineering . Fung. Meirelles LASEFI . Antonio J. 47. v.Introduction 89 Capítulo 4 KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL) BLENDS Luciana Ninni. 2002. A.O. Wong H.State University of Campinas (UNICAMP) P. 835 – 838. p. Box 6121 Zip Code 13083-970 Campinas – SP.

90 Capítulo 4 .

2000. the . there are many empirical or semi-empirical equations in the literature (Domínguez et al. For instance. 2000) for estimating viscosity data.5% was obtained for the viscosity calculations using the GC-UNIMOD model. Nhaesi and Asfour...1 Abstract Kinematic viscosities of binary and multicomponent mixtures containing poly(ethylene glycols) were measured as a function of temperature. 4. a generalized correlation based on the number of carbon atoms was developed for calculating kinematic viscosities of pure PEGs in a molecular mass range between 200 and 3350 as a function of temperature.. The GC-UNIMOD has been used to predict the viscosity of systems containing mixtures of organic solvents (Domínguez et al. In a previous paper (Cruz et al. Hence. Average absolute deviation around 3. 1993).. 2000). low toxicity. which requires binary interaction parameters between constituent groups of molecules in the mixture and viscosities of pure components. PEGs are also used as processing aids in making other products. 2000) and working fluid pairs such as methanol and some poly(ethylene glycol) dimethyl ether (PEGDME) mixtures (Herraiz et al. The group-contribution viscosity model GCUNIMOD was employed to correlate the viscosity of the binary systems and then to predict the viscosities of the multicomponent mixtures.. In the case of mixtures. lubricity. 1999). Viscosities of the binary mixtures were used to calculate the excess molar Gibbs energy of activation for viscous flow G*E. because of the industrial interest. 2000. Among them there is the group-contribution viscosity model GC-UNIMOD (Cao et al. experimental viscosity data and models for estimating viscosities of PEG mixtures are important. blends of PEG 300 and 1450 are available from most suppliers for use in pharmaceutical and cosmetic products.2 Introduction Poly(ethylene glycol) (PEG) belongs to a class of synthetic polymers that finds several industrial applications due to their availability in a wide range of molecular masses besides having unusual combination of properties such as water solubility. etc. Nhaesi and Asfour.Introduction 91 4. Kinematic Viscosities of Poly(ethylene glycol) Blends 4. An example is the use of molten PEGs in heat transfer baths (Powell. In the case of the last mentioned system. 1980). The G*E values are positive over the entire composition range and they increase as the difference in the molecular masses of the two polymers increases. Blends of different molecular masses are desired to obtain combinations of properties.

They were analytical grade reagents used without further purification. based on the average molecular mass determined by GPC. In Table 4.1. The water content of the PEG samples was previously determined by Karl Fischer titration using a Metrohm device (Switzerland).02 0. True carbon number .1 Experimental Materials The samples of poly(ethylene glycol) of different average molecular masses.75 ± 0. while the true carbon number was evaluated from the results of the GPC molecular mass distribution. 2000) to binary and multicomponent blends of PEGs using the GC-UNIMOD model. and the GC-UNIMOD model was used to correlate and predict these experimental data.01 0. Their polydispersity index was determined by gel permeation chromatography (GPC) using a Ultrahydrogel column Waters device.3.067 1. PEG 3350.09 ± 0.01 NCN a 8 18 26 44 68 152 b TCN b 8 18 28 44 66 126 Nominal carbon number based on the molecular mass according to the PEG denomination.086 1.069 1.23 ± 0. Table 4. ranging from (200 to 3350) g mol-1.3 4.095 1.. Poly(ethylene glycol) Characterization PEG 200 400 600 1000 1500 3350 a avg molecular mass (Mn) 202 400 616 987 1468 2806 polydispersity index 1. were supplied by Sigma. except PEG 1500.54 ± 0.8 mL min-1.92 Capítulo 4 authors reported an average absolute deviation around 20 % for viscosity predictions and observed that these predictions were getting worse with the increase of the molecular size of PEGDME.069 1. The denomination nominal carbon number refers to carbon numbers calculated from the molecular mass given by PEG manufacturers such as PEG 200.15 K. The following experimental conditions were employed for the GPC runs: water as the mobile phase at a rate of 0.02 1. which was purchased from Fluka. and a refractive index detector. the kinematic viscosities of blends of PEGs covering a wide range of mixture molecular masses were measured at various temperatures. injection temperature of 313. 4.02 0. The aim of this paper is to extend our prior studies (Cruz et al. volume of the sample injected equal to 100 µL.1 the characteristics of the polymers used in this work are given together with the nominal and true carbon numbers for each PEG.20 ± 0.02 1. Therefore.27 ± 0.073 water content (mass %) 0.

4999 77.15 298.4999 0.2.15 0.58 600(1)+1000(2) 333.5007 0.9 × 10-9 m2 s-1. measured viscosities was 1. 1. The viscometers were placed in a constant temperature water bath controlled to ± 0.4999 0.3 %.5031 0.27 15.15 308.01 s accuracy was used for measuring the efflux time.5007 0.4999 0.62 200(1)+600(2) 0.52 37.11 18.GmbH) with ± 0.Results and Discussion 4.44 21.80 400(1)+600(2) 0.4999 0. Calibrated Cannon-Fenske type viscometers were used to measure the kinematic viscosities at different temperatures (Cannon Instrument Co.5007 127.2 to 4.5001 0.5031 93.4999 0.5031 0.5001 0.99 29.20 600(1)+3350(2) 0.5031 0.1 mg accuracy.3 × 10-7 m2 s-1 and the minimal.15 303.95 58.5031 0. 4.4836 0.66 60.3. Table 4.4999 0.5043 0. The estimated error in mass fraction was 1 in 10 000.27 41. 300 and 350. The experiments were replicated at least five times for each PEG mixture and the results reported are the average values. The kinematic viscosities ν were calculated from the efflux time and the The maximal standard deviation of the instrument constant provided by the manufacturer.5001 36.61 77. appropriate for the range of viscosity values measured in the present work.4999 0.86 49. The viscometer sizes were 150.18 30.73 46.4999 0.4918 0. Viscosity of Binary Blends of Poly(ethylene glycol)s at Various Temperatures T/K w1 ν/10-6 m2 s-1 w1 ν/10-6 m2 s-1 w1 ν/10-6 m2 s-1 200(1)+400(2) 293.5 as a function of mass fraction (w) and temperature T.5007 0.81 23.15 0.97 .44 99.15 318.89 72.4999 97.15 353.1 °C.4999 0.49 37.5031 0.15 343.5031 0.15 323.53 59. The average uncertainty of the viscosity values was estimated to be equal to 0. resulting in an average deviation of 2.4 Results and Discussion The viscosity data are presented in Tables 4.4999 0.73 20.49 33.).94 25.15 363. 200. An electronic timer with 0.90 30.4836 40.41 46.4918 0.87 400(1)+1500(2) 0.13 24.15 313.5007 0.32 70.19 26.2 Apparatus and procedures 93 The solutions were prepared by mass on an analytical balance (Sartorius Analytic .5001 0.25 40.85 54.6 × 10-8 m2 s-1.

5000 88.5000 0.3992 0.5000 0.15 363.15 0.66 ν/10-6 m2 s-1 w1 ν/10-6 m2 s-1 Capítulo 4 Table 4.2500 0.69 63.0513 0.78 400(1)+1500(2) 400(1)+3350(2) 0. (continued) T/K w1 1000(1)+3350( 400(1)+3350(2) 2) 333.2000 0.9482 0.61 121.07 49. Viscosity of Binary Blends Containing Poly(ethylene glycol) 400 at Various Concentrations T/K w1 ν/10-6 m2 s-1 w1 ν/10-6 m2 s-1 w1 ν/10-6 m2 s-1 400(1)+600(2) 293.82 110.0511 0. the GC-UNIMOD model was selected for viscosity calculations.0273 0.14 52.1513 0.15 333.3496 0.2001 0.15 293.02 31.00 196.5000 0.15 293.2000 0.91 68.49 56.15 293.42 133.15 333.15 333.25 139.15 293.8002 224.15 333.1010 0.96 146.5000 0. .82 219.76 21.7000 0.8998 0.70 162.0996 0.09 63.15 333.2500 0.94 Table 4.7006 0.27 0.80 20.83 50.15 293.10 62.92 38.5001 0. Moreover.04 19.2999 0.2.2999 0.63 24.5001 0.98 147.84 23.17 23.9003 70.15 343.5001 114.3.1029 0.04 26.94 0.62 In this work.15 0.77 28.6998 0.15 333.9482 154. the viscosity data for the systems containing PEG 400 have been used to calculate the excess molar Gibbs energy of activation for viscous flow.8998 0.76 48.98 105.15 293.3992 0.79 84.5001 0.7000 0.48 39.0997 0.15 333.15 353.87 25.15 333.

The pure component viscosities were calculated using the equation proposed by Cruz et al.01 16. T/K Viscosities of a Multicomponent Mixture of Poly(ethylene glycol)s at Various Temperatures w1 w2 w3 w4 ν/10-6 m2 s-1 400(1)+600(2)+1000(3)+1500(4)+3350(5) 333. G*E.1999 0.59 47.3335 0.3334 0.15 353.3332 0.3332 0.2002 0.27 79.15 353.3335 0.10 34.3339 0. R is the gas constant.3330 0. xi is component i mole fraction.3331 0.85 31.2000 0.5.77 13.77 4.23 1000(1)+1500(2)+3350(3) 0.3334 0.1997 0.59 Table 4.15 0.2000 0.3335 0.3334 30.3332 0.67 21.3334 0.21 600(1)+1000(2)+1500(3) 0.1.25 52.Results and Discussion Table 4. respectively.1999 0.2000 0.3339 0.84 34.3339 0.3331 0. (2000) and the true carbon numbers given in Table 4.15 363.26 22.1997 0.3334 0.3334 109.20 26.15 343.3334 0.4.15 363.15 0.00 400(1)+1000(2)+3350(3) 333.2002 0.29 38.3334 0.3334 0.2002 61. T is the absolute temperature.28 60.15 343.1999 0.3334 0. 400+1500 and 400+3350) have been used to calculate the excess molar Gibbs energy of activation for viscous flow.3330 0.3331 0. .3339 0.15 363.3330 0.1999 0.3334 0.3330 0.15 0.3331 71.3334 0.15 343. and n is the number of components in the mixture.1) where M is the mixture molar mass.2002 0.3332 0.4. Viscosities of Ternary Mixtures of Poly(ethylene glycol)s at Various Temperatures T/K w1 w2 ν/10-6 m2 s-1 w1 w2 95 ν/10-6 m2 s-1 400(1)+600(2)+1000(3) 333.1 Excess molar Gibbs energy of activation for viscous flow Viscosities of binary mixtures (PEGs 400+600. ν and νi are the kinematic viscosities of the mixture and of the pure components.1997 0.3335 47.40 44.2000 0.15 353. through the following equation: n   G *E = RT  ln(ν M ) − ∑ x i ln(ν i M i )  i =1   (4.1997 0.13 26.

0 3. 4.5 2.0 2. and Ap are adjustable parameters.2 0.5 0 0.5 5.8 1.0 4. respectively.4. The G*E values are positive over the entire composition range and they increase as the difference in the molecular masses of the two polymers increases.6 0. 5. 4.0 0.6. like the UNIFAC (Fredenslund et al. ▲ 400+1500. The coefficients Ap and the average absolute deviations (AAD) between experimental and calculated G*E values are given in Table 4.15 K.2 GC-UNIMOD model The GC-UNIMOD viscosity equation. is expressed as the sum of combinatorial and residual terms..96 Capítulo 4 The G*E (in J mol-1) for the binary mixtures was fitted to a Redlich-Kister’s equation: G * E = xi x j p =0 ∑ A p ( xi − x j ) p (4. — Redlich-Kister). Excess molar Gibbs energy of activation for viscous flow of PEG binary mixtures (■ 400+600 at 293.5 1.0 1. 1975) model.1.4 x1 0.0 Figure 4.0 0. The G*E experimental values and the Redlich-Kister’s fitted polynomials are plotted in Figure 4. 400+3350.15 K.1). □ 400+600 at 333. The increase of the temperature has a similar effect (Fig. The combinatorial term of both the viscosity equation and the UNIFAC model is dependent on the molecular size and shape.1. while .5 G*E/ 103 J mol-1 4.2) 3 where xi and xj denotes the mole fractions of components i and j.5 3.

1984). or CH2OH.925 5.. CH2CH2O. CH2 (option 1) (Skjold-Jorgensen. 1989).30 5. CH2O. CH2 (option 3) (Herskowitz and Gottlieb. A readjustment of the energy interaction parameters becomes necessary to improve the predictive capability of the model. the GC-UNIMOD model. with energy interaction parameters obtained from UNIFAC-VLE tables (Skjold-Jorgensen. Coefficients of the Redlich-Kister Equation (Ap) and Average Absolute Deviations (AAD) between Experimental and Calculated G*E values T/K 400 (1)+600 (2) 293.32 8805. the mixture of PEGs 600 and 3350) the model predictions were poor. three pairs of interaction parameters have to be adjusted in each one of the above options. the viscosities of the pure components must be known.855 70.9 % was obtained. 1999.119 4.027 5960. different possibilities of dividing the polymer molecules in functional groups were considered.01 495. The equation for pure PEGs (Cruz et al.127 16811.076 2049.20 A0 A1 A2 A3 AAD As a first attempt. The new parameters were obtained using a nonlinear estimation method (Marquardt.02 22896.1809 409. As suggested by Herraiz et al. (1999). OH. For these predictions the PEG molecule was divided in three functional groups: CH2. OH and CH2O. The experimental data of the binary mixtures (Tables 4. Table 4.3292 1172.Results and Discussion 97 the residual term takes into account the interaction energy between different groups present in the mixture.6134 73. An average absolute deviation of 15. 1979). 2000) was used for calculating the polymer viscosities required in the GC-UNIMOD model.15 400 (1)+3350(2) 333.63 6549. 1963). CH2O. In order to estimate the kinematic viscosities of the PEG mixtures..6.8 4.3) were used for this readjustment. CH2 (option 2) (Ninni et al.15 333. As can be seen.15 15771. 1979). PEG molecules can be divided in the following groups: OH. The average absolute deviations (AAD) between experimental and estimated values were calculated according to the following equation: . was tested for estimating the viscosities of the binary mixtures measured in this work.3651 1249.15 400 (1)+1500 (2) 333.193 3957. It was also verified that specially for the binary mixtures containing polymers with very different molecular masses (for example.2 and 4. Rasmussen and Rasmussen.

89 1.54 3.7).03 1500+3350 1. Table 4. Better agreement between experimental and calculated data for the binary mixtures was obtained using options 1 and 2.50 1.94 4.36 3.i − ν calc.09 0.07 400+600+1000 600+1000+1500 400+1000+3350 1000+1500+3350 400+600+1000+ 5.75 4.90 0.53 1.53 3.98  m  ν exp .62 4.8.20 2.18 2.76 4.09 1.09 2. but in this case options 2 and 3 result in slightly better AAD global values.76 %.90 .86 2.77 1.14 3.74 8.62 % and 3.91 6.97 4. and m is the number of experimental points. around 2.04 4.i and νcalc. The comparison for the ternary and quinary mixtures is shown in Figure 4.59 2.12 1.3) where νexp.2 shows the viscosity values calculated by the GC-UNIMOD model compared with the experimental binary data given in Table 4.7.75 1.54 3.i  i =1     100  ×  m  Capítulo 4 (4.2. with AAD global values of 3.00 0.10 0. Figure 4.82 5. The GC-UNIMOD model was also capable of predicting rather well the viscosities of the multicomponent mixtures. Average Absolute Deviations (AAD) for the Viscosity Estimation AAD Binary Mixtures Option 1 Option 2 Option 3 Multicomponent Mixtures Option 1 Option 2 Option 3 AAD 200+400 200+600 400+600 600+1000 400+1500 600+3350 1000+3350 400+3350 global 4.52 5.28 3.43 2.27 2.86 2.46 4.i are the experimental and calculated kinematic viscosities.74 2.45 3.9 % (Table 4. respectively. The energy interaction parameters for option 2 are given in Table 4.3. respectively (see Table 4.20 2.i AAD = ∑  ν exp .16 7.7).20 7.03 5.

— GC-UNIMOD/ option 2).2.8. + 600+3350.8155 85. ▼ 600+1000. As can be observed in Table 4.4514 OH -250. . ■ 400+600.7 the highest AAD values were obtained for those binary mixtures containing PEGs 200 and/or 400. ┉ 200+600.7). 400+3350.Results and Discussion 99 CH2CH2O CH2 -93.3 -101.3 - Table 4. In the case of the multicomponent systems no experimental values were measured for mixtures containing PEG 200. Probably this is the reason why the AAD values for the predictions (multicomponent mixtures) were slightly lower than those obtained for the correlation of the binary mixtures (Table 4.7182 41. Energy Interaction Parameters (in Kelvin) for the GC-UNIMOD Model CH2CH2O CH2 OH 52. Experimental and correlated viscosities of binary PEG mixtures (experimental data of Table 4.2: ▲ 200+400. ○ 400+1500. 1000+3350.4333 140 120 100 2 -1 ν/ 10 m s 80 60 40 20 290 300 310 320 330 340 350 360 370 380 390 -6 T/ K Figure 4.

3 % for predictions of mixture viscosities using energy interaction parameters from UNIFAC-VLE tables. The authors of the GC- UNIMOD reported deviations in the range of 2.8 %. In this case the AAD values varied in the range of 0. Gündüz.7 to 5. 1999) predictions of the GC-UNIMOD model for mixtures of some polyethylene glycol dimethyl ethers and methanol over a wide range of temperatures. (2000) employed the GC-UNIMOD model for predicting the viscosity of fatty systems. Mei et al. Experimental and predicted viscosities of multicomponent PEG mixtures (∇ 400+600+1000.0 %.. some correlation and prediction results reported in the literature should be mentioned. The comparison can also include some empirical equations used for correlating experimental viscosities of aqueous systems containing PEGs. 1996. Rabelo et al. 1995. 1994. The GC-UNIMOD model. The reported average absolute deviations were around 20 %.3.100 120 100 80 ν/ 10 m s 2 -1 Capítulo 4 60 40 20 0 330 -6 335 340 345 350 355 360 365 T/ K Figure 4. 600+1000+1500. For comparison purposes.. with new set of interaction parameters. . — GC-UNIMOD/ option 2).8 to 14. works well for estimating the viscosity data measured in the present work.. ○ 1000+1500+3350. González-Tello et al. Some of these works (Gündüz. The results showed that the predictions were getting worse with the increase of the polymer molecular mass. 2000) reported results for concentrated and/or diluted solutions with AAD values varying in the range of 1. It is also depicted in the literature (Herraiz et al. 400+1000+3350. ┉ 400+600+1000+1500+3350.2 to 6.

Gündüz. M. 4. 2000. J. 4. G. 1993. K. 680. Fluid Phase Equilib.. A. 1975. Eng. Cruz. G.15 and 313.. Res. L. Chem. Chumpitaz. Chromatogr. it should be noted that for this prediction we had to use the nominal carbon number. P. U. Rasmussen. Data 2000. F. A. M.. 1980) for PEG mixtures is the viscosity of a blend of PEGs 300 and 1500 at equal mass proportion of both compounds. A. L. not the true carbon number. Jones. 277-292. Although such deviation is comparatively greater than those obtained for the correlation and prediction of our experimental data. Fredenslund. 61-63. R. . A. Viscosities of the Ternary Mixture (2-butanol + n-hexane + 1-butylamine) at 298. Urieta. In most cases a good agreement between experimental and calculated values was obtained. Kinematic Viscosities of Poly(Ethylene Glycols). F. yielding an AAD value of 10. F. J. Eng. M. 39. M. S. Blázquez. This data was compared to the value predicted by the GC-UNIMOD model (option 2). The GC-UNIMOD model was tested for correlating and predicting viscosity data. W. AIChE J. J. Density and Viscosity of Concentrated Aqueous Solutions of Polyethylene Glycol... 21. J.. J.15 K. 169. Chem. I. 1086-1099.6 %.. Domínguez.. Meirelles. Chem. Eng. Ind.. S. to estimate the pure PEG viscosities. Prausnitz. D. 2088-2092.. Evaluation of Viscosities of Polymer-Water Solutions used in Aqueous TwoPhase Systems. A. B 1996. I.5 Conclusions In this work the kinematic viscosities of blends of poly(ethylene glycol)s with nominal molecular masses ranging from (200 to 3350) g mol-1 were determined at various temperatures.. 611-614. Gascón. Data 1994. 263-266. Fredenslund. Pardo.. Knudsen. Group Contribution Estimation of Activity Coefficients in Nonideal Liquid Mixtures.Conclusions 101 The unique experimental data available in the literature (Powell. J. J. Camacho. Alves. González-Tello. 32.. J..6 Literature Cited Cao. 45. P. Royo. Group-Contribution Viscosity Predictions of Liquid Mixtures Using UNIFAC-VLE Parameters. L.

Cavaleri. H. Rasmussen. 2. E.. Gottlieb. Chem. Rabelo. Eng..-H. In Handbook of Water Soluble Gums and Resins. J. Z. Process Des.. Kolbe. A.-Q. Fluid Phase Equilib. D. J. Thermochim. Marquardt. Capítulo 4 Gündüz. Appl. Sci.. M. J. S. An Algorithm for Least-Square Estimation of Nonlinear Parameters. 743. R. Progress 1989.. L. Z. S. Math. F.. Viscosity Prediction of Polyethylene Glycol . Chem. M. Chem. Acta 1999. Skjold-Jorgensen. Han. Lin... P. Viscosity Prediction for Fatty Systems. Asfour. 2000. Ind. 50-56. Zhu. Eng.. . A. 1979. G.. S. 2000. 1963. J. Nhaesi. 328. Densities and Viscosities of Polyethylene Glycol + Salt + Water Systems at 20°C. McGraw-Hill Book Company: New York. Batista. Mei. Water Activity in Poly(Ethylene Glycol) Aqueous Solutions. J. M. J. J. J. Am. 431-441. Meirelles. Rasmussen. Shen. Chromatogr.-Q. Eng. Data 1984. Eng. Soc. 714-722. Polyethylene Glycol. J. B 2000. Chapter 18.. Gmehling. Indust. Coronas. Davidson. Tetraethylene Glycol and Estimation of UNIFAC Parameters. Data 1995. 1168-1171. Herskowitz. B.. Meirelles. 169-176. M. Revision and extension. 77. 55. W. 11. A. Rasmussen. Eng. A. Ninni. D. A. A-F A. Thermophysical Properties of Methanol + Some Polyethylene Glycol Dimethyl Ether by UNIFAC and DISQUAC Group-Contribution Models for Absorption Heat Pumps... P. D. 18..-X. A. Camargo. 1999. U. Ed. 155. J. Phase Equilibria in Aqueous Polymer Solutions.Dextran . Oil Chem. L. L. Vapor-Liquid Equilibrium in Aqueous Solutions of Various Glycols and Poly(ethylene glycols). 85.. Dev. 1980. Vapor-Liquid Equilibria by UNIFAC Group-Contribution.Water Solutions used in Herraiz. 181-185. 2861-2873. 1255-1261. 450-452. Chem. Chem. Powell.102 Aqueous Two-Phase Systems. 327-337.. W.. Soc. 40. Fernández.. J. Prediction of the Viscosity of Multi-Component Liquid Mixtures: a Generalized McAllister Three-Body Interaction Model. 29.

.521011/95-7. 145941/99).Literature Cited 103 4. 466680/00-7.7 Acknowledgments The authors gratefully acknowledge the financial support of FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo – 01/10137-6) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico .

Meirelles Department of Food Engineering (DEA/FEA). Cidade Universitária “Zeferino Vaz”. 324 – 329. Wong H. Zip Code 13083-970. p.Capítulo 5 105 Capítulo 5 KINEMATIC VISCOSITIES OF POLY(ETHYLENE GLYCOL) AQUEOUS SOLUTIONS Luciana Ninni. 2003. State University of Campinas – UNICAMP. Antonio J. v. . Campinas. Brazil Trabalho publicado na revista Journal of Chemical & Engineering Data. PO Box 6121. SP. Helcio Burd. 48. Fung. A.

106 Capítulo 5 .

1 Abstract Kinematic viscosities of aqueous mixtures containing poly(ethylene glycol)s (PEG) with nominal molecular masses ranging from (200 to 10.. etc. 1995. 2001. In addition to the knowledge of phase equilibrium.2 Introduction Poly(ethylene glycol) (PEG) is a linear. 2000. 2000). However.. Among them one can cite the formation of two-phase systems with aqueous solutions of other polymers or salts. its non-toxicity. or of one polymer (PEG) and a salt (usually a phosphate or sulfate salt). data on the properties of the phases are necessary for the design of ATPS extraction processes in large-scale applications. available in a variety of molecular masses. 1998).Introduction 107 5.5% were observed between calculated and experimental results. The adjusted parameters also allowed the viscosity prediction for multicomponent mixtures. 1999). The use of PEGs in aqueous two-phase systems (ATPS) has been reported by several authors (Zaslavsky.. concentrations . The binary experimental data were used for adjusting the parameters of a Kumar-like equation. (Harris. is important because these solutions are used in many industrial and biotechnological applications (Albertson. e. Gongxi et al. 5. high capacity and fast separation. and the formation of complex with metal cations. neutral polyether. methods for the estimation of viscosities at various temperatures. Extraction using ATPS can be performed continuously in available commercial separators (Coimbra et al. Relative errors around 5. dextran and poly(ethylene glycol) (PEG). Kinematic Viscosities of Poly(ethylene glycol) Aqueous Solutions 5. with an overall deviation of 9.9%. since such systems allow the separation and purification of biomolecules in a gentle environment. The advantages of using aqueous two-phase extraction lie in volume reduction. 1992) Aqueous two-phase systems are increasingly used for separation of biomolecules. PEG has a variety of properties that make it suitable for applications in the biomedical and biotechnological areas.g. Grossmann et al..000) g⋅mol-1 have been determined at various concentrations and temperatures. Gündüz. and consequently. 1999. Silva et al. The knowledge of physical properties. cells and cell particles (Bradoo et al. Porto et al.. Such systems are composed of two incompatible polymers.. moreover. 1995. soluble in water and most organic solvents (Powell. 1980). it is relatively straightforward to scale up a separation process (Persson et al. it is not viable to measure viscosities at all conditions of interest.. 1999). 1971). such as the viscosity of aqueous mixtures containing PEGs.

concentration and temperature was also developed. The estimated error in the mass fractions was 2 in 10 000. 2001) has been employed to correlate and predict kinematic viscosities of aqueous solutions containing PEGs in a wide range of polymer concentrations and at different temperatures. Numerous empirical and semi-empirical methods have been proposed to calculate liquid viscosity of mixtures (Mehrotra et al.. 1996.. deionized with a Mili-Q water system (Millipore. and up to 25 mass % for PEGs 8000 and 10 000.3. In the present work. Gonzaléz-Tello et al. 1994) is that viscosity of concentrated PEG solutions were employed together with literature viscosity data at various temperatures. .3 5. 1993. The difference between the data treatment from this work and other related works (Gündüz. Nevertheless. Pereira et al. Gündüz. Gündüz. the reformulated Kumar’s equation (Pereira et al. 1996. Moreover. 2001.0001 g. 2000.. GmbH) accurate to ±0. 1000 and 1500. 2000). up to 30 mass% for PEG 3350. The water content in the polymer samples was taken into account for calculating the solution concentrations.000) g⋅mol-1 were purchased from SIGMA and used without further purification. PEG concentrations in the solutions varied up to 50 mass % for systems containing PEGs 200. 1992. 2001. Bieze et al. 1989.. Bisal et al. 2000). some equations are inappropriate for concentrated and/or multicomponent mixtures. Cao et al. 1993. A generalized correlation for estimating kinematic viscosities of PEG aqueous solutions as a function of polymer molecular mass. 1996.. hydration numbers were estimated and compared with available values in the literature (Antonsen and Hoffman. Cao et al. for each polymer employed. Kirinčič and Klofutar.. 600. Mei et al. Gündüz. PEGs were characterized according to their molecular masses and water content using the procedures described in a previous article (Cruz et al. 1995.. Pereira et al.1. The solutions were prepared on a mass basis using an analytical balance (Sartorius Analytic..108 Capítulo 5 and solute types are of great practical interest. 1990). 1994. USA) and used for preparing the PEG solutions. The characteristics of the polymers used in the present work are given in Table 5...1 Experimental Section Materials Analytical grade PEGs with nominal molecular masses ranging from (200 to 10. 5. concentrations and PEG molecular masses. 1999. Moulik and Gupta.. 400. Water was distilled.

02 0.15) K. . 150 and 200) were used to measure the kinematic viscosities (Cannon Instrument Co.2 Apparatus and Procedures Calibrated Cannon-Fenske glass capillary viscometers (sizes 50.02 1.1 °C.01 0.3. 75. 5. An electronic timer with 0.05 5.02 1..069 1. The variation coefficient varies within the range 0.08 0. Austria) for holding the temperature constant within ±0. being the lowest figures obtained for the lowest viscosity values.1.54 ± 0.086 1. The standard deviations of the viscosity determinations varied within the range 1.095 1.15 and 323.067 1. 100. The measurements were performed at temperatures varying between (293. polydispersity index and water content of PEGs used in this work PEG 200 400 600 1000 1500 3350 8000 10 000 avg rel molar mass (Mn) 202 400 616 987 1529 2806 7975 10 475 polydispersity index 1.1% to 0.4.073 1.106 water content (mass %) 0.2 – 5.Experimental Section 109 Table 5.9×10-7 m2⋅s-1 to 1. Average relative molar masses (Mn). The experiments were replicated at least three times for each PEG mixture and the results given below are the average values.75 ± 0. The viscometers were placed in a water bath (Anton Paar.01 0.7%.23 ± 0.09 ± 0.7%.02 0.66 ± 0.36 ± 0. so that the uncertainty of the experimental measurements can be estimated as not higher than 0.069 1.099 1.4 Results and Discussion The kinematic viscosities for the binary and multicomponent PEG solutions determined in this work are given in Tables 5.1×10-9 m2⋅s-1.01 s accuracy was used for measuring the efflux time. USA).20 ± 0.27 ± 0.

854 0.365 1.15 293.1489 4.356 0.0501 1.415 0.4777 0.684 2.0492 1.971 3.745 2.121 0.2490 0.900 PEG 3350 0.4899 0.2.15 313.0985 1.0492 1.4899 0.089 1.2399 0.2490 0.15 293.994 0.656 0.0737 1. in aqueous solutions of PEGs at various temperatures and mass fractions.1994 2.4902 0.239 0.2953 6. Kinematic viscosities.4902 0.4902 PEG 600 0.2490 0.237 2. w T/K w ν / 10-6 m2 s-1 w ν / 10-6 m2 s-1 PEG 200 293.696 15.543 0.649 2.588 0.15 323.534 10.0246 1.4899 0.15 303.0996 1.539 2.0741 1. w ν / 10-6 m2 s-1 PEG 400 0.1529 1.345 0.311 0.835 0.2967 8.2487 0.0744 2.256 7.15 K and various mass fractions.0248 1.669 0.470 0.2399 0.215 0.4899 3.304 w Table 5.2993 4.252 0.2490 0.169 0.941 0.425 3.180 .0243 1.501 0.0749 1.216 0.4902 0.693 w ν / 10-6 m2 s-1 PEG 1000 0.0248 1.15 313.2487 0.148 7.828 10.1002 1.416 1.0498 1.1491 2.379 0.1482 2.0249 1.1980 4.443 w ν / 10-6 m2 s-1 PEG 600 0.793 0.143 0.2275 0.510 0.956 PEG 1500 0.15 303.4777 0.094 0.389 7.4777 0.231 4.2275 0.823 14.425 4.4958 PEG 1000 4.15 303.1998 2.2987 16.143 0.267 0.751 0.2275 0.0995 2.917 7.15 0.936 0.456 0.15 323.3.15 313.406 0.502 PEG 400 3. ν. in aqueous solutions of PEGs at 293.15 313.1477 2.0748 1.110 Capítulo 5 Table 5.15 293.361 5.829 1.2487 0.1960 3.15 323.1983 7.4777 1.524 10.15 323.028 21. Kinematic viscosities.15 303.167 2. ν.4958 0.0497 1.0992 2.2998 3.747 0.2275 0.2399 0.4958 0.

2349 0.2349 0.15 323.119 2.2492 0.312 22.15 323.655 24.0499 0.716 28.1000 0.15 303.0530 0.15 313.15 303. (cont.0499 0.1223 0.15 293.191 6.15 323.1223 0.15 303.1000 0.376 1.15 293.3.2492 0.2441 0.15 313.128 1.664 33.0530 0.15 303.125 8.757 5.15 0.476 18.338 2.965 2.623 4.15 313.1000 0.0499 0.046 4.4534 PEG 8000 0.473 3.967 0.364 5.314 PEG 3350 2.457 1.Results and Discussion Table 5.329 1.904 14.15 303.0969 0.387 1.15 313.0969 0.824 16.281 111 .3389 0.1223 0.833 1.2492 0.496 12.15 293.15 313.519 5.715 0.748 2.496 19.250 12.15 293.444 2.2349 0.3389 0.4534 0.037 2.15 293.0977 0.15 323.0977 0.396 3.0530 0.1223 0.980 31.15 323.0969 0.838 1.679 1.2554 2.2492 0.2441 2.15 323.922 1.2349 0.0499 0.2441 0.228 9.784 19.2554 0.977 24.15 303.0977 0.3389 0.1000 0.2554 0.15 313.863 6.3389 PEG 10 000 3.) PEG 1500 293.196 4.0969 0.2554 0.086 9.4534 0.0530 0.

15 0.166 5.430 8. mass fraction of solute .1000 11. Kinematic viscosities. Kumar’s expression was formulated to estimate viscosities in aqueous and non-aqueous salt solutions. The Kumar expression was later reformulated (Pereira et al.0495 0.248 293.1003 8. b and h are the solute-solvent interaction coefficient and the hydration number (number of molecules of bound solvent per molecule of solute).1 Kumar’s equation for viscosity correlation in PEG mixtures The Kumar expression (Kumar.112 Capítulo 5 Table 5.15 323.0999 0. 2001) to allow the calculation of kinematic viscosities of aqueous solutions containing salts or organic solutes.0999 0.1000 0.0495 0.1005 0.0495 PEG 400 (1) + PEG 8000(2) 0. Initially.675 3.1012 0.1032 0.221 6. ν. the hydration number and the ion-solvent interaction coefficient. x is the solute mole fraction.170 13.1006 0.15 303.15 313.1012 0.15 323.0997 0.1000 0. as given below: η bx =1+ η0 1 − hx (5.0997 0.1005 0.1006 0.0999 0.0495 0. respectively.1003 0.15 303.1012 0. Such correlation involves two parameters.1003 0.051 10.1032 0.0997 0.1032 0.1006 0.657 293. η0 is the solvent dynamic viscosity..1032 PEG 3350 (1) + PEG 8000(2) + PEG 10 000(3) 0.4.468 17.1012 0.15 23.1005 0.1003 0.0999 0.4. of multicomponent poly(ethylene glycol) aqueous solutions at various temperatures T/K w1 w2 w3 w4 ν / 10-6 m2 s-1 293.1000 0.15 323.1005 0.15 303.15 313.15 313.15 PEG 400 (1) + PEG 600(2) + PEG 1000(3) + PEG 1500(4) 0. The resulting equation is a function of relative density between solution and solvent.936 4.1) where η is the solution dynamic viscosity.1006 0. 1993) is one of the equations presented in the literature for calculating dynamic viscosities of solutions as a function of solute concentration.0997 0.702 5.078 4.

Results and Discussion ternary, quaternary and quinary solutions.

113

and temperature. Pereira et al. (2001) investigated the predictive capacity of the equation for

The reformulated Kumar equation was used in this work to correlate and predict the kinematic viscosities of the aqueous PEG mixtures. multicomponent systems are given as follows: The equations for binary and

ν rel =

1 ρ rel

Bw   1 + 1 − Hw   

(5.2)

ν rel

 ∑ Bwi  1   = 1+ i  1 − Hw  ρ rel ∑ i  i  

(5.3)

where the parameters B and H are interpreted, respectively, as a coefficient of solute-solvent interaction and as the number of solvent molecules solvating the solute, both functions of temperature; w is the mass fraction of solute; νrel and ρrel are the ratios between the kinematic viscosities and densities of the solution and the corresponding values for the pure solvent, respectively and i represents the solutes in the multicomponent mixture. Dependence of the parameters B and H on temperature is given by polynomials in (T - Tref), as follows: B = Bref + ∑ Bn (T − Tref ) n
n

(5.4)

H = H ref + ∑ H n (T − Tref )n
n

(5.5)

where Tref is the reference temperature equal to 298.15 K. As can be seen, eqs 5.2 and 5.3 require the information of densities of pure water and of aqueous PEG solutions. For this reason, linear equations were fitted to experimental density data of PEGs in water (González-Tello et al., 1994; Kirinčič and Klofutar, 1998) at 298.15 K. The coefficients of the linear equations are given in Table 5.5.

114

Capítulo 5

Table 5.5. Linear fittings for the densities of poly(ethylene glycol)s in aqueous solutions equation: ρ /(103 kg m-3) = a + b w
PEG 300 400 600 900 1000 1500 2000 3000 3350b
a b

a 0.996 04 0.996 11 0.996 23 0.995 68 0.995 70 0.995 63 0.995 77 0.995 74 0.996 48

b 0.166 97 0.169 85 0.172 33 0.176 62 0.176 30 0.175 87 0.178 62 0.179 43 0.175 64

R2 3,66a 3,74 3,68 3,76 3,76 3,72 3,75 3,71 3,67

PEG 4000 6000 8000b 10 000 12 000 15 000 20 000 35 000

a 0.996 01 0.995 75 0.996 67 0.995 42 0.995 86 0.995 86 0.996 04 0.996 17

b 0.17836 0.18059 0.17701 0.18295 0.18234 0.18140 0.18068 0.18117

R2 3,71 3,72 3,49 3,75 3,68 3,76 3,63 3,80

Correlation coefficients such as 0.999 80 given as 3,80. Experimental density data from González-Tello et al. (1994). All other sets of density data were taken from Kirinčič and Klofutar, 1998.

In addition, to estimate the densities of the PEGs in water at temperatures other than 298.15 K the following approach was used, since these data are not available in the literature: the densities of the PEG solutions at 298.15 K relative to water (subscript w) at the same reference temperature were multiplied by the density of water at the desired temperature:

ρ i (T ) =

ρ i (298.15 K ) ρ (T ) ρ w (298.15K ) w

(5.6)

In dealing with multicomponent systems, a semi-empirical equation was used for calculating the corresponding densities. It requires densities of the binary solutions (solute + solvent), and is given by:

ρ mix − ρ w =

i , i≠ w

∑ (ρ

i

− ρw )
(5.7)

where ρmix, ρw, and ρi represent the densities of the multicomponent mixture, of pure water and of the binary aqueous mixtures containing component i, respectively. The summation in eq 5.7 should be performed upon all the components i in solution, except for water. Equations 5.6 and 5.7 were used in a prior work (Ninni and Meirelles, 2001) for estimating the densities of aqueous solutions containing different solutes with very good agreement between experimental and calculated values.

Results and Discussion

115

By nonlinear regression (Marquardt, 1963) of eq 5.2, the values of Bref, B1, B2 and Href, H1, H2 were obtained for each PEG + water system. A data bank comprising experimental viscosities of binary systems reported in this work and those from literature (Mei et al., 1995; Kirinčič and Klofutar, 1999), in a wide range of PEG molecular masses, concentrations and temperatures was employed in the correlations. The data bank comprises a total of 259 experimental points. The adjusted parameters as well as the average absolute deviations (AAD) between experimental and calculated viscosities are presented in Table 5.6.

Table 5.6. Adjusted coefficients of eqs 5.2, 5.4 and 5.5 for calculating kinematic viscosities of aqueous PEG solutionsa
PEG 200 300 400 600 900 1000 1500 2000 3000 3350 4000 6000 8000 10 000 12 000 15 000 20 000 35 000 global
a

w ≤0.49 ≤0.29 ≤0.50 ≤0.49 ≤0.21 ≤0.50 ≤0.45 ≤0.40 ≤0.12 ≤0.34 ≤0.10 ≤0.25 ≤0.26 ≤0.24 ≤0.04 ≤0.04 ≤0.03 ≤0.03

T/K 303 – 323 298 293 – 323 293 – 323 298 293 – 323 293 – 323 298 298 293 – 323 298 298 293 – 323 293 – 323 298 298 298 298

Bref 3.4444 4.5680 5.4186 6.7086 7.3506 7.6304 10.3957 11.3580 13.1893 16.2610 16.6385 24.3935 29.8474 34.1380 41.0170 45.3602 48.9736 70.2625

B1/10-3 55.3 0.0 37.8 -1.05 0.0 2.0 -1.89 0.0 0.0 -0.02 0.0 0.0 -4.56 5.36 0.0 0.0 0.0 0.0

B2/10-3 -2.38 0.0 -2.38 -0.002 0.0 1.03 1.10 0.0 0.0 1.27 0.0 0.0 -4.60 -10.3 0.0 0.0 0.0 0.0

Href 1.3862 1.4256 1.4487 1.5693 1.5891 1.6762 1.7942 1.9760 2.2100 2.2050 2.3424 2.7669 2.9778 3.2098 3.2500 3.4566 3.6566 4.0611

H1/10-3 -5.44 0.0 -12.7 -9.62 0.0 0.385 -7.76 0.0 0.0 1.94 0.0 0.0 -8.34 3.66 0.0 0.0 0.0 0.0

H2/10-3 0.106 0.0 0.29 0.068 0.0 -0.36 0.0 0.0 0.0 0.039 0.0 0.0 0.253 0.0 0.0 0.0 0.0 0.0

AADb 5.9 2.4 3.4 4.9 3.7 5.7 7.9 4.7 2.7 10.8 2.6 7.8 4.3 7.0 2.5 3.4 2.3 3.8 5.4

Experimental viscosity data from the present work and from the literature (Mei et al., 1995; Kirinčič and Klofutar, 1999).

b

 n  ν exp , i − ν calc, i AAD = ∑  ν exp , i  i =1   

 100  ×  n 

I. 1989). since there are reports about the variation of hydration numbers with molecular mass of different PEG molecules. and viscometry (Kirinčič and Klofutar.. 1992).1. and a minor temperature dependence of hydration has been observed.20 water molecules may bind a single oxygen center for PEGs with molecular masses in the range of (200 and 1000) g⋅mol-1. 1999). According to Moulik and Gupta (1989). varies widely according to the experimental measurement techniques. For this reason. it is interesting to compare Href values obtained in the present work with the hydration numbers reported in the literature. Values between 0. Bref values as a function of PEG molecular mass.1 and 5.116 Capítulo 5 It was observed that Href and Bref exhibit a dependence on the polymer molecular mass as depicted in Figures 5. 70 60 50 40 ref B 30 20 10 0 0 5000 10000 15000 20000 25000 PEG molecular mass 30000 35000 Figure 5.16 to 3.53 for PEG 400 and 1.93 for PEG 600.2. 1994). 1992). The same authors obtained hydration numbers (molecules of bound water per monomer unit) of 1. which include conductometry (Bisal et al. 1990).9 and 6 water molecules per monomer have been quoted (Bieze et al. Bahri and Guveli (1988) have reported that 0. The number of water molecules. differential scanning calorimetry (DSC) (Antonsen and Hoffman. R. spectrophotometry (Moulik and Gupta. 4 to 6 water molecules per oxygen for a PEG of molecular mass 20 000 have been obtained. Hydration numbers obtained from . It is also reported that PEGs of higher molecular masses show increased hydration (Antonsen and Hoffman. which are thought to be bounded in the PEG molecules..

(1990) for PEGs 200. forming segment-segment interactions as it traps additional. the physical meaning of eq 5. Hence. 4.5 1. Hydration numbers determined by viscometry are also reported by Kirinčič and Klofutar (1999).0 1. 1. the amount of water bound per polymer repeated unit varied from 2. it seems that in the present work the Href values might be interpreted as the number of molecules of bound water per oxyethylene unit.07.21 and 2.5 2. it can be observed that the hydration numbers at 298.15 K (Href) obtained in this work are in agreement with part of the data reported in the literature. 1992).44 for PEGs with molecular masses between (300 and 35 000) g⋅mol-1.0 0 5000 10000 15000 20000 25000 30000 35000 PEG molecular mass Figure 5. The values varied from 2.6 and the hydration numbers reported in the literature.3 to 3. more loosely bound water between the segments. 1.0 Href 2. respectively.95.79. 600 and 1000 with values equal to 0. based on the figures presented previously.3 × 105 g⋅mol-1.2 for calculating viscosities of aqueous PEG solutions becomes more consistent despite . 300.Results and Discussion 117 viscometry were presented by Bisal et al. Href values as a function of PEG molecular mass. comparing the Href values from Table 5. Therefore.5 3. 400. as the molecular mass increases. The explanation given for such high values was that.2. In general.0 3.8 for PEG molecular masses between 200 and 2. In the work of Antonsen and Hoffman (1992).45. They also observed that the amount bound at higher molecular masses is greater than at low molecular masses. A different result was obtained by DSC determinations of bound water (Antonsen and Hoffman.06 to 28. 0. the polymer chain begins to fold in on itself.

such information is not available for the viscosity data taken from the literature. The parameters of this generalized equation were estimated employing the entire set of experimental data.1 gives the average molecular masses for the PEGs used in the present work.8 and Figures 5. a generalized correlation could be obtained using eq 5. such as PEG 1000 or PEG 3350.3 and 5.1424 B2 = 6.7. The molecular masses used in this correlation were the values given by the PEG denomination. Taking Bref and Href as functions of PEG molecular mass.2152 C3 = 22077. . as described above. Substituting eq 5.9974 and 0.2).1 B1 = -0.5 H1 = 1. Mw is the molecular mass of the PEG molecule and Ci's are adjusted constants. respectively. Table 5.4. such values were regressed to an exponential type equation as given below: Yref = C1 + C 2 exp(− M w / C3 ) (5. In any case. Results of the regression are given in Tables 5.4 and 5.5. Considering that the values of Bref and Href are functions of the polymer molecular masses (as shown in Figures 5.99096 C2 = -2.32 × 10-3 a Ba C1 = 83. the data presented here are closely related to the general conclusion that between one and four water molecules are bound per repeated oxyethylene unit of the polymer. and that hydration numbers rise with the PEG molecular mass.1 and 5. Correlation coefficients of 0.2 for calculating the viscosities of the systems studied.8 in eqs 5.9795 C2 = -80.8) where Yref is Bref or Href. Constants for the generalized equation Ha C1 = 3.9977 were obtained in the regression for estimating Bref and Href values.118 Capítulo 5 some discrepancies.59861 C3 = 7891.7 and 5.2.32 × 10-3 Href and Bref are calculated by eq 5. a generalized expression was obtained for the H and B parameters of eq 5.8 using the parameters Ci given above.08 × 10-3 H2 = -0. Although Table 5.

PEG 1500. PEG 3350.8 5.6 2.4 5. □. —. values. Results of the generalized correlation for viscosity calculations PEG 200 300 400 600 900 1000 1500 2000 3000 global 20 15 ν / 10-6 m2 s-1 10 5 0 0.7 4.8.15 w 0.30 Figure 5. PEG 1000.1 3.5 3.05 0.6 3.0 2.9 6.15 K and various concentrations: ■.0 7.3. ▲. PEG 400.20 0. Experimental and calculated viscosities at 293.Results and Discussion 119 AAD% 16.5 5.8 6. ○.6 5.25 0.4 6.2 2.8 3. PEG 600.00 0. calculated .6 3.10 0. •.5 Table 5.1 PEG 3350 4000 6000 8000 10 000 12 000 15 000 20 000 35 000 AAD% 7.

■.2554. In case of PEG 3350 the main reason for such deviation is probably the expressive difference between the nominal and the average molecular masses (Table 5. 1999). The systems containing PEG 200 or 3350 are those with the higher deviations between experimental and calculated viscosities.8). This difference observed in the polymer behavior according to its molecular mass can eventually explain the larger deviation obtained for aqueous solutions of PEG 200. In case of PEG 200 the relatively low experimental viscosity values might have contributed for the large deviation observed. 1992. trapping additional water between the segments of the polymer chain (Antonsen and Hoffman. PEG 1000. The AAD-value for the whole set of experimental data was 5. w=0. Kirinčič and Klofutar.4899. PEG 400. For polymers of larger molecular mass the chain begins to fold in itself. PEG 10 000.4777. Some authors suggested that for poly(ethylene glycol)s of low molecular mass. w=0.4. ∆. only tightly bound water is associated to the polymer chain. w=0.120 50 45 40 35 Capítulo 5 ν / 10-6 m2 s-1 30 25 20 15 10 5 0 290 295 300 305 310 T/K 315 320 325 Figure 5.2487.2441. ○.5% (Table 5. w=0. . w=0. PEG 8000. Experimental and calculated viscosities at various temperatures and concentrations: ▲. □. It is also important to contrast the results of the correlation obtained in the present work with those reported in the literature for the viscosity of aqueous PEG systems.1). but it also seems that such values do not fit so well in the suggested dependence of parameters Bref and Href on the polymer molecular mass. PEG 1500. PEG 600. González-Tello et al. such as PEG 200. a difference not so large for the other PEG denominations. w=0.4534. •.

in comparison with other equations.4.0)% for some multicomponent fatty mixtures. a Grunberg equation was used for correlating the viscosities. the equation proposed to calculate viscosities of aqueous PEG solutions takes into account the influences of poly(ethylene glycol) concentration and temperature in the range (277 to 313) K. Probably there is an additional effect upon viscosity attributable to the relatively high concentration of polymers with different molecular masses.8% (Günduz.15 K. ketones. (1993) reported deviations in the range of (2. It is also mentioned in the literature predictions of the GC-UNIMOD model for mixtures of some polyethylene glycol dimethyl ethers . 5. with a resultant AAD of 1. The values of viscosity calculated by eq 5.8 and 14. Cao et al. An AAD of 9. For a mixture of both polymers (PEG + dextran) in water.5% for viscosities lower than 10 mPa·s. The predictive capacity of eq 5. (2000) reported deviations in the range of (0.3 were compared with the experimental data from Table 5. ethers. respectively.Results and Discussion 121 (1994) reported an AAD-value of 3. In that work. including mostly binary mixtures of alkanes. They used a large data bank.9% was obtained. resulting in an AAD of 6.4. This effect is not properly described by the parameters adjusted to aqueous binary mixtures containing one polymer. it should be noted that such AAD values obtained in this work are similar to those reported in literature for viscosity prediction of multicomponent mixtures. It should be observed that the predicted values are in all cases lower than the experimental viscosities. Concluding the above remarks. quaternary and quinary aqueous PEG systems was investigated.15) K were considered. A subsequent work of the same author (Günduz.3)% for viscosity predictions by the GC-UNIMOD model.2 Viscosity prediction in multicomponent systems. alcohols.5% for dynamic viscosities (η) greater than 10 mPa·s. The same model equation was tested by Günduz (1996) for PEG 8000 and dextran 580 000 in water at 293. It must be stressed that in this latter case the results were obtained by adjustment of four parameters for each polymer separately. Polymer concentrations up to 7 mass % and temperatures between (303.4% for aqueous mixtures of PEG + dextran of different molecular masses. In spite of this. is that it has a generalized form in relation to PEG molecular mass. and 17.15 and 343.5%. Using the same model Rabelo et al. it can be observed that the results achieved in the present work are in good agreement with the AAD values usually presented in the literature. The advantage of the equation presented here.9% and 1. 1996). 2000) presented AAD-values up to 9. concentration of polymer and temperature.3 for estimating viscosities of the ternary.7 to 5. and esters.

R. H. Sci. etc) were employed at relatively low solute concentrations (up to 15 mass %). W. 266. 1993.. Saxena. Partitioning and Resolution of Mixture of Two Lipases from Bacillus stearothermophilus SB-1 in Aqueous Two-Phase System.. S. Harris. J. In general. K. A.E. glycerin. 35. J. 1990. Eng.. Phys. Huige. J. either in correlation or in prediction.S. Chem. M. Knudsen. Bahri. Barnes. A. predictions with good accuracy were obtained by Pereira et al. On the other hand. 65686576. Plenum Press: New York. Bradoo.. 1988.. J. Cao. However.. P. P. Bieze.. Viscosity B Coefficients of Polyethylene Glycols in Water. 1992. were satisfactory. Wiley: New York. 1994. K. Guveli.P.N. 57-62. Bisal.. S.M. 4212-4216. Antonsen.. Enderby. only low molecular mass solutes (salts.C. the results obtained in this work. Gupta. The reported average absolute deviations were around 20%.. Moulik. Process Biochem. Distribution of Water around Poly(ethylene oxide): A Neutron Diffraction Study. Leyte. (2001) employing the modified Kumar equation for estimating viscosities of multicomponent mixtures. 94. In Poly(ethylene glycol) Chemistry. S. Res.C. Moreover. 32. R. Bhattacharya. the obtained generalized equation based on the polymer molecular mass is an useful tool since only few parameters are required for viscosity estimations in systems containing PEGs with molecular masses between 200 and 35 000 g⋅mol-1. T.J. Chem.5 Literature Cited Albertson P. . Fredenslund.W.. Group-Contribution Viscosity Predictions of Liquid Mixtures Using UNIFAC-VLE Parameters. 98. Rasmussen. A. J.. Colloid Polym. Partition of Cell Particles and Macromolecules. Hoffman. 20882092. 5. C. Phys. P. Ind. 1999. Chem. D. monosaccharides.122 Capítulo 5 and methanol over a wide range of temperatures... 141-144.E..K. A. Conductivity Study of Microemulsions: Evaluation of Hydration of Oil/Water Microemulsions Applying Bruggeman Equation. Ed. K. 1971.

7. A. Yixin. R..F.. H. Chem. M. Blázquez.. 1998. Alves. Biotechnol. Z. Ziqiang. Ed. J. 233-247. 155. 263-266.G. Measurement of Mass Transfer Coefficients and their Modelling of Continuous Countercurrent Aqueous Two-Phase System in a Packed Extraction Column. Grossmann. U. 1999. J. M.15 K. S.. González-Tello. Cruz. B 1996. F. Gündüz. J. L.L. Data 2000. C.. Optimization of Bovine Serum Albumin Partition Coefficient in Aqueous Two-Phase Systems. Partitioning of Low Molecular Combination Peptides in Aqueous Two-Phase Systems of Poly(Ethylene Glycol) and Dextran in the Presence of Small Amounts of K2HPO4/KH2PO4 Buffer at 293 K: Experimental Results and Predictions. J. Fluid Phase Equilib. Camacho.. Biosseparation 1995. Meirelles.. S. Data 1994. 196-204. 9 . Chem. .A. Günduz.J... Maurer.. Harris. 181-185. U. Y. Density and Viscosity of Concentrated Aqueous Solutions of Polyethylene Glycol. Chromatogr. Viscosity of Aqueous Solutions of Poly(Ethylene Glycol)s at 298.15 K. J. Harris. 5. G. In Poly(ethylene glycol) Chemistry.. D.. Kinematic Viscosities of Poly(Ethylene Glycols). J. Meirelles. Chem. 45. Zhu. Kirinčič. C.S. B 2000. Viscosity Prediction of Polyethylene Glycol-Dextran-Water Solutions Used in Aqueous Two-Phase Systems.. Chromatogr. J. P. Bioseparation 2000. J. M. 680.. Eng. Eng. Performance of Graesser Contactor in the Continuous Extraction of Whey Proteins: Mixing. R. J. Tintinger. U. 277-281.. 39. R. Gündüz. J. Klofutar.. Kula. 1998. 743. C. Evaluation of Viscosities of Polymer-Water Solutions used in Aqueous Two-Phase Systems.. 311-325. A. 259-268. 1992. J. Bioeng. 699-711. Klofutar. M. 149.A. Eng. Plenum Press: New York. Chumpitaz. Mass Transfer and Efficiency. 60. Fluid Phase Equilib. 611-614. Kirinčič. 1999. G.. Thömmes.D. A. Zhaoxiong. A Volumetric Study of Aqueous Solutions of Poly(etheylene glycol)s at 298. Chin. J. Gongxi. 61-63. G. S.Literature Cited 123 Coimbra.

A. Ninni. Progr. Marquardt. In Handbook of Water Soluble Gums and Resins. 81. Soc. Zhu.. 71.. 117. Chem. 1168-1171. A. Appl.W. Chromatogr. Johansson.. J. Ed. Y.J. Aires-Barros.A. W. Porto. F. Lin. . Kinematic Viscosity Prediction for Aqueous Solutions with Various Solutes. G. A 1999. pH and Density of Aqueous Amino Acids Solutions. L. D.. An Algorithm for Least-Square Estimation of Nonlinear Parameters. J.-H. J. 31-48. Persson. J. Can. J. J.. 344-355.. Batista. H-O. S. Chem. J. Chem. Meirelles. Davidson. Monnery. Can. 35-40. M. Hydration Studies on Some Polyhydroxy Non-Electrolytes and NonIonic Surfactants. Oil Chem. F. Han. Mei. Eng. Rabelo. Hydrodynamics and Mass Transfer in Aqueous Two-Phase Protein Extraction Using a Continuous Perforated Rotating Disc Contactor. 11.. R.. A.. 431-441.. Biotechnol. Math.. 77. Chenlo. Water Activity.-Q. R. 1980. B.-Q. Indust. A Review of Practical Calculation Methods for the Viscosity of Liquid Hydrocarbons and their Mixtures. 1989. 2000. 17. 1963. 2001.. J. Sarubbo. L. Purification of Protein and Recycling of Polymers in a New Aqueous Two-Phase System Using Two Thermoseparating Polymers. P. Vázquez. J. S. Gupta. Moulik. Fluid Phase Equilib. McGrawHill: New York. Temperature and Pressure. Z. Viscosity Prediction for Fatty Systems. A. G. L. W. Soc.. Lima. L. Svrcek. 1996. J. Powell. Tambourgi.J. Moreira.A.. Mehrotra. A Simple Correlation for Estimating Viscosities of Solutions of Salts in Aqueous. Eng. E. A.124 Capítulo 5 Non-aqueous and Mixed Solvents Applicable to High Concentration. Data 1995. Pereira. L. 356-363. R. L. 22. 2000. Eng. Meirelles. Cavaleri. Am. Bioprocess Eng. Densities and Viscosities of Polyethylene Glycol + Salt + Water Systems at 20°C. M. 703-711. D. 864. F. Z. E.... F. K. 948-954. Kumar A. Tjerneld. 1255-1261. Cabral... 67.W.. J. D. M. 2001. M. 215-218.. J. Chem. 1993. S. 40.-X.

466680/00-7. Eng. B. A. Aqueous Two-Phase Partitioning. The authors gratefully acknowledge the financial support of FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo: 01/10137-6) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico . Meirelles. A. J. Marcell Dekker: New York. M. Chem. 1995. J.521011/95-7. 46. Y.. . 145941/99). PEG + Potassium Phosphate + Urea Aqueous Two-Phase Systems: Phase Equilibrium and Protein Partitioning. 251255. Zaslavsky. H.Literature Cited 125 Silva. L. Data 2001.

Gerd Maurer2 1 Faculdade de Engenharia de Alimentos – Universidade Estadual de Campinas (UNICAMP) C. . A.Capítulo 6 127 Capítulo 6 Thermodynamic properties of maltodextrin aqueous solutions Luciana Ninni1. D-67653 Kaiserslautern.P. Meirelles1. Trabalho a ser submetido na revista Carbohydrate Polymers. Antonio J. Brazil 2 Lehrstuhl für Technische Thermodynamic. Universität Kaiserlautern. 6121 CEP 13083-970 Campinas – SP. Germany.

128 Capítulo 6 .

b). These polymers exhibit broad molar mass distributions (with number-average molar masses varying between ~103 and 105 gmol-1) due to the process of starch hydrolysis that is carried out with enzymes and/or acids at elevated temperatures. Thus. It was found in this procedure that the second and third virial coefficients depend on the molar mass of the different compounds.phase systems (Silva and Meirelles. maltose. Among the polysaccharides. and 0. α-amylase does not readily hydrolyze α-1→4 linkages in maltose and maltotriose. 1998). 6. Moreover. 2000a. commonly employed in starch hydrolysis is capable of hydrolyze α-1→4 linkages but has little effect on α1→6 linkages. usually represented by the average molar masses: number-average (Mn) or weight average (Mw). Alpha-amylase.2 Introduction Water-soluble polymers have been used in a wide range of industrial products in the food and pharmaceutical areas. Thermodynamic properties of maltodextrin aqueous solutions 6. Mean relative deviations of 0. the need for environmental protection favors the use of biodegradable polymers. The VERS model was also used to correlate experimental water activity and calorimetric data for the binary mixtures and to predict water activity of some low molar mass compounds and other thermodynamic properties reported in the literature. maltotriose and other low molar mass saccharides can be found in . Marchal.1 Abstract The thermodynamic properties of maltodextrins (MD) in aqueous solutions were studied by isopiestic method. such as polysaccharides. The experimental results of water activity and light scattering were described by the osmotic virial equation. laser-light scattering and calorimetry. At the present time.40% were obtained in the correlations of water activities of maltodextrin solutions when the maltodextrin molecule was treated as one component.Introduction 129 6. Three different MD samples with nominal molar masses between 1000 and 3000 g mol-1 were investigated at various concentrations.93% when it was approximated by four components. maltodextrins are polymers of great interest due to their high solubility in water (Gliksmann. 1986. They are starch hydrolysates consisting of α-D-glucose units bounded by (1→4) glycosidic linkages (primarily) as well as by (1→6) linkages. 1999) and potential application in aqueous -two . for industrial purposes (Swift. This group of compounds comprises molecules from natural (renewable resource raw material) to synthetic (from petrochemical base feed stocks) origin. Polysaccharides occur with a molar mass distribution.

d). 1993. the VERS equation. (1992) also presented a study of the gelation process of maltodextrin-water systems using low resolution H-NMR. isopiestic. Thermodynamic properties of polymeric aqueous solutions have been reported in literature (Hasse et al.. Silva and Loh. 1995. is going to be tested in correlating and predicting the investigated thermodynamical properties for water activity and heats of dilution in maltodextrin and sugar systems. Mothé and Rao. the osmotic virial equation is used to describe the experimental results from isopiestic and laser-light-scattering measurements and a semi-empirical group contribution model. Gaube et al.130 and chemical stability (Kasapis et al. 1993. b. Schierbaum et al. it can be cited Cesàro et al.. Among some of the works available. In that work a gelling maltodextrin was experimentally fractionated in 6 different fractions with different average molar masses.. The gelation of low DE maltodextrins and phase equilibria in gelatin/maltodextrin systems were investigated by Kasapis et al. (1999). Moreover.. Capítulo 6 the final hydrolyzates. Many works about maltodextrins in the literature deal with experimental investigations of gelling maltodextrins. (1993a. Kany et al. 1999). Heats of dilution of the single polymers and of the mixed polymer solutions were used to evaluate the Flory interaction parameters (Flory. . They observed the time required for the beginning of gelation as a function of temperature and concentration. c. for determining thermodynamic properties. laser-light-scattering and calorimetric methods. (1989 a. 1953). 1995. It was then investigated the sorption and desorption isotherms and the amount of non freezable water in the fractions.. Thermodynamic data determinations are important under the industry and academic research viewpoints but only few data sets for aqueous maltodextrin solutions can be found in the literature. This paper reports experimental results for aqueous maltodextrin systems involving three different techniques. Radosta et al. c) presented a comprehensive study about physical-chemical properties of maltodextrins in aqueous solutions. Groβmann et al. 2000). 1999. and the influence of the addition of amylose on the gelation process. b. which studied thermodynamic properties of two biopolymers (maltodextrin and gelatin) by using calorimetry. viscosity. These characteristics confer them a wide range in solubility.

MD 13 – 17 and MD 16 – 19 are. A constant flow rate of 1 mL⋅min-1 was maintained by a HPLC pump (Spectra Physics. type DAWN DSP) and a interferometric refractometer (Wyatt Technology.05 M sodium nitrate. type ERC-3512). Tabela 6.31 8. Details on the calibration of the detectors are described by Hasse et.and mass-average molar mass of the maltodextrins studied in this work. type P 1000).2. weight-averaged molar mass (Mw) and polydispersity ratio (Mw/Mn) determined in these GPC experiments are presented in Table 6.1. The polymers were used as received and the water content of each sample was determined by Karl Fischer titration using a Methrom device.1 Experimental Materials 131 Aqueous solutions of three different maltodextrins purchased from Aldrich Chemical and designated here according to their dextrose equivalent range as MD 4 – 7. Wyatt Technology. MD 13 – 17 and MD 16 – 19 were studied in the present work. 6. For each maltodextrin sample. Detectors were a multiangle laser-light scattering (MALLS.42±0. About 100 µL of the maltodextrin solution was injected in the eluent using a Reodyne valve.22 and 3. The eluent was degassed in-line (ERMA. type Optilab 903).38 7. al. the determination of molar mass distributions were performed at least three times.20.3. MD 4 – 7 MD 13 – 17 MD 16 – 19 Mn 2683 1475 1140 Mw 38360 13710 8283 Mw/Mn 14.3.1 Molar mass distributions The maltodextrins were characterized by gel permeation chromatography (GPC) with an aqueous eluent containing 0. respectively: 3. The results of titration for the maltodextrins MD 4 – 7.Experimental 6.3.3 6. The results for the number-averaged molar mass (Mn). 4.18 mass % of water.31 . Mean relative deviations between different determinations varied in the range of 3 and 13%.2 Methods 6. A column type MCX 1000 Å (PSS Polymer Systems) was utilized for separation. (1995).31±0. Experimental data on the number.31±0.1.1 and the gel permeation profiles of the maltodextrins are shown in Figure 6.

1.2 presents the experimental data of molar masses vs. according to the following relation: Kc 1 = + 2 A2 c Rθ MP(θ) with (6.1.132 Capítulo 6 As can be seen in Table 6.0x10 5. This indicates that the MD 4 –7 was less converted to low saccharides in the process of starch hydrolysis. After the separation in the GPC columns. the light scattering from each eluted fraction of very low concentration is detected simultaneously at 16 angles between 28.5x10 -4 -4 Carbohydrate concentration (g mL ) 3. from the extrapolation of the light scattering to zero angle at each slice.0x10 3.8°.5x10 1. Further details on the properties of gelling maltodextrins will appear in the following items. The presence of more higher molar mass components makes MD 4 –7 capable of forming thermally reversible gels in aqueous solutions. Figure 6.7 and 143.1) . Gel permeation profiles of maltodextrins.1 and in Figure 6. elution volumes calculated by the software ASTRA.0x10 1.0x10 2.5x10 2. 4.0x10 -4 MD 4-7 MD 13-17 MD 16-19 -1 -4 -4 -4 -4 -5 0.0 5 6 7 8 9 10 11 Volume (mL) Figure 6. MD 4 – 7 comprises more high molar mass saccharides while the maltodextrins MD 13 – 17 and 16 – 19 have higher content of low molar mass saccharides.

n0 the refractive index of the solvent. P(θ) a form factor describing the angular and size dependence of the scattered light intensity. for MD 4 – 7. respectively). R(θ) the excess Rayleigh ratio at angle θ . volume. The reason is that the light scattering detector of the MALLS device is less sensitive at low molar masses.θ → 0 M (6. 1972). it is observed a linear dependence of the molar masses on the elution volumes. pico . A2 the second virial coefficient.3) 2 In Figure 6. M is the weight-average molar mass.1 becomes: Kc 1 = Rθ c → 0 .1510. R g the mean-square radius of gyration. W(M) is the cumulative molar mass distribution.1522 cm3 g-1. h(V) is the normalized concentration defined as: ci/ ∑ c i ∆V .2) where K is an optical constant. The correct slope in the obtained curves is important to calculate molar mass distributions of the polymers as shown below: x(M ) = dW ( M ) dW ( M ) dV h(V ) = =− d (log M ) dV d (log M ) f (V ) (6.2. NA the Avogrado number.4) where x(M) is the differential molar mass distribution also called dξ/d(logM). equation 6. λ is the laser wave length (633 nm). 0. c the concentration of the polymer solution.Experimental 133 K= 4π 2 (dn / dc) 2 n0 N Aλ4 0 2  16π 2 n0 sin 2 (θ / 2)  2 1 R g = 1+    P(θ) 3λ2   (6. and λ0 the wavelength of the incident light under vacuum (Huglin.1486. When c→0 and θ→0. However. 0. dn/dc the differential refractive index increment of the polymer in solution (0. it can be also noted that the scattering of the data is greater in the region of low molar mass molecules. 13 –17 and 16 – 19. f(V)=d(logM)/dV the slope of the curve molar mass (in a logarithmic scale) vs.

134 Capítulo 6 Molar Mass vs. N. for each fraction represented by a G-Li function. The purpose of using this method is to separate the whole distribution in more monodisperse fractions and also to evaluate if the peaks that elute later in the chromatographic runs correspond to the lower molar mass compounds: glucose.. 2. maltose ..0 10. Plotting a diagram of molar mass versus volume with data points obtained by laser-light scattering experiments up to volume of about 8 mL and assuming that the adjusted volumes by G-Li functions for the first three fractions correspond to glucose. N is the number of G-Li functions) were fitted to the molar mass distribution. Figure 6.0 Volume (mL) Figure 6.2. maltose and maltotriose.5) For each maltodextrin.0 8.19 MD 4 . Now.0x101 4.3 presents the adjusted G-Li functions for MD 4 – 7.7 1.0x106 MD 16 . ci and di (i=1.0 12.0 6. Volume 1. the GPC results for the molar mass distribution of each maltodextrin were approximated by the sum of Gauss-Lorentz functions.0x105 Molar Mass (g/mol) 1. bi.. it is known the correspondent elution volumes.17 1. .0x102 1.0x104 1.0x107 MD 13 . In order to verify whether the slope of the adjusted curve was correct.0x103 1. Results of the experiment GPC/MALLS for the maltodextrins studied in this work. the parameters ai. The Gauss-Lorentz (G-L) function is given by: GLi = ai exp − d i2 (M − ci )2 1 + bi2 ( M − ci ) 2 ( ) (6.

respectively.Experimental 135 and maltotriose. Results of Gauss-Lorentz functions for the molar mass distribution of MD 4 – 7. and 2. volume obtained for the higher molar mass molecules. maltose and maltotriose.8 . respectively are reported in literature (Defloor. it can be considered that the results from GPC/MALLS analysis are reliable and they can be used to estimate molar mass distribution of maltodextrins. maltose and maltotriose. as shown in Figure 6. Concentrations of the lower molar mass fractions estimated by the Gauss-Lorentz method were assumed to be glucose.6 . This extrapolation can be seen in Figure 6. Therefore.0x10 -5 0.5x10 -4 1.0x10 -4 Concentration (g mL ) -1 1.3.0 5 6 7 8 9 10 11 Volume (mL) Figure 6. maltose and maltotriose. 342 an 504 g mol-1. it is possible to verify if the extrapolation of the curve logarithmic of the molar mass vs.1 mass % for glucose. 1998) for a maltodextrin with approximately 1300 g mol-1. The last three points in Figure 6. which has a molar mass of 1140 g mol-1.5x10 2.3 and 8. and they can also be compared to values presented in literature. is consistent and consequently if the slope of the straight line is correct.0x10 -4 5.4 were plotted using the adjusted elution volumes and the respective molar masses for glucose.9 and 7. 4. that are 180. 5. -4 2. .2.9 were found in this work for MD 16 – 19.4 indicating that the previous assumption was correct and that the adjusted elution volumes (volumes on the maximal peak point) really correspond to glucose. Values of 1. maltose and maltotriose.

All the solutions were prepared with bidistilled water. The temperature of the thermostat fluctuated by less than ±0. using the MALLS data and elution volumes for glucose. The two NaCl concentrations (NaCl(1) and NaCl(2)) were chosen in such a way that would be expected these concentrations were above and below the final equilibrium concentration. Molar mass as a function of elution volume. The nine flasks of the experimental arrangement were used as follows: three flasks were filled with standard NaCl solutions in a concentration named here as NaCl(1).136 1000000 Capítulo 6 100000 MD16-19 Regression Molar mass (g mol ) -1 10000 1000 100 5 6 7 8 9 10 11 Volume (mL) Figure 6. maltose and maltotriose from the Gauss-Lorenz functions.3.15 K for MD 13 –17 and 16 –19. In the glass. and at 318. 6. In a experiment. the glass was allowed to rotate around an axis inclined at 45°.2.4.15 K for all three maltodextrins. mass transfer . In order to increase the mass transfer area and enhance sample mixing during the equilibration period. NaCl(2). and three for maltodextrin solutions. the flasks were filled with about a gram of an aqueous solution of non volatile substance.2 Isopiestic Method The water activity in aqueous solutions of maltodextrins was measured with the isopiestic method at 298. The system was immersed in a constant temperature water bath during the equilibration times. three flasks also with NaCl solutions but in other concentration. The experimental arrangement consisted of a glass multilegged manifold with nine standard taper female ground-glass joints for the attachment of solution flasks and a high vacuum stopcock for isolation of the system after evacuation.1 K.

525e-4 4.740e-4 4.4286 0.0622 SD/ g of NaCl 2.3796 0.069e-5 7.0280 0.756e-4 1.500e-3 1.665e-4 1. The experimental data points at 318. evacuated and thermostated.Experimental 137 of water between the sample and the reference solutions took place through the vapor phase.9898 0.777e-4 1.593e-3 1.3408 0.5671 0.0091 0. the glass cells are filled with about 1 mL of aqueous solution.630e-3 1.0373 0.3029 0.724e-4 1.185e-4 1.937e-4 .539e-4 1.0465 0.9942 0.9664 0.15 K are given in Tables 6.9829 0.190e-4 4.5898 SD/ g of MD 3.0073 0.1750 0.4408 0.135e-4 1.9962 0. The experimental water activity results for the three maltodextrins at 298.494e-5 1. (1995).462e-4 7.0145 0.4 together with standard deviations (SD) in concentration of non volatile substance and in water activity. The apparatus and experimental procedures used for the isopiestic experiments in this work were described before by Großmann et al. These equilibrium concentrations were always the same within the experimental error. until the activity of water in the sample was the same as that in the reference solutions.773e-4 5.494e-4 aw 0.2 – 6.186e-4 2.316e-4 1.822e-4 1.4872 0. the chamber is sealed.9838 0.2419 0.028e-4 3. In a experiment.427e-3 2.817e-4 5.792e-5 1.15 K wMD/ g g-1 0.9622 SD/ aw units 1.9779 0. The water activity in the aqueous sodium chloride solutions were calculated using the correlation of Pitzer and Peiper (1984).2. The water activity of the samples was estimated from gravimetric analysis of the sample and reference solutions before and after the experiment.036e-3 8.2234 0. The process of mass transfer and the evaluation of water activities are the same as those described above.659e-4 7.087e-4 5.910e-4 1.9917 0. Kany (1998).0178 0.15 K presented in Table 6. and Lammertz and Maurer (2002).0377 0.771e-4 1.5 were measured also by isopiestic method but with a different experimental arrangement.293e-4 wNaCl/g g-1 0.0294 0. The wNaCl values in the following Tables correspond to the average concentration between NaCl solutions (NaCl(1) and NaCl(2)) at equilibrium.9781 0.011e-4 1.0217 0.0102 0.9948 0.9875 0.308e-4 2.04e-4 8. Table 6. It consists of a stainless steel apparatus that can hold up to 30 glass cells placed in a copper block that is mounted in a thermostated chamber.9724 0.274e-5 8.0559 0.4867 0.795e-5 1.971e-3 8.680e-4 6.5307 0. Experimental results for the water activities in aqueous solutions of MD 13 – 17 at 298.

138

Capítulo 6

Table 6.3. Experimental results for the water activities in aqueous solutions of MD 16 – 19 at 298.15 K
wMD/ g g-1 0.2598 0.2743 0.3434 0.3887 0.3844 0.4197 0.4492 0.4559 0.4824 0.5357 0.5850 SD/ g of MD 2.386e-4 9.011e-4 5.117e-4 1.541e-3 1.310e-4 7.735e-4 1.565e-3 1.183e-4 7.963e-4 6.175e-4 1.036e-3 aw 0.9928 0.9922 0.9886 0.9854 0.9860 0.9831 0.9802 0.9797 0.9763 0.9689 0.9594 SD/ aw units 1.050e-4 1.116e-4 7.555e-5 1.060e-4 1.497e-4 4.272e-5 5.450e-4 4.126e-4 2.426e-4 1.228e-4 2.296e-4 wNaCl/g g-1 0.0086 0.0134 0.0198 0.0254 0.0243 0.0292 0.0339 0.0348 0.0403 0.0521 0.0664 SD/ g of NaCl 1.827e-4 1.943e-4 1.296e-4 1.797e-4 2.538e-4 7.143e-5 8.977e-4 6.782e-4 3.920e-4 1.909e-4 3.390e-4

Table 6.4. Experimental results for the water activities in aqueous solutions of MD 4 – 7 at 298.15 K
wMD/g g-1 0.2859 0.2644 0.2914 0.3167 0.3365 0.3399 0.3379 0.3488 0.3802 0.4099 0.4590 0.4615 SD/ g of MD 1.546e-3 1.026e-3 1.310e-3 9.982e-4 1.822e-3 2.520e-4 9.311e-4 1.992e-3 2.348e-3 1.111e-3 1.367e-3 1.097e-3 aw 0.9967 0.9970 0.9966 0.9960 0.9953 0.9957 0.9957 0.9956 0.9945 0.9939 0.9919 0.9922 SD/ aw units 4.286e-4 2.842e-4 1.028e-4 1.620e-4 1.789e-4 2.031e-4 2.306e-4 1.877e-3 9.010e-4 1.906e-4 2.373e-3 1.393e-3 wNaCl/g g-1 0.0048 0.0052 0.0059 0.0069 0.0081 0.0075 0.0075 0.0077 0.0099 0.0106 0.0140 0.0136 SD/ g of NaCl 7.518e-4 4.984e-4 1.803e-3 2.843e-4 3.138e-4 3.563e-4 4.046e-4 3.291e-3 1.496e-3 3.333e-4 4.125e-3 2.424e-3

Experimental

139

Table 6.5. Experimental results for the water activities in aqueous solutions of MD 13 – 17 and MD 16 –19 at 318.15 K.
wMD/g g-1 MD 13 – 17 0.2767 0.3284 0.4452 0.4968 MD 16 – 19 0.2683 0.3156 0.4327 0.4852 2.941e-4 2.827e-3 1.895e-3 4.570e-5 0.9927 0.9903 0.9821 0.9767 1.390e-4 1.148e-4 2.210e-4 7.023e-5 0.0128 0.0169 0.0308 0.0369 2.421e-4 1.984e-4 3.679e-4 1.137e-4 7.206e-4 1.498e-3 2.126e-3 1.820e-4 0.9927 0.9903 0.9821 0.9767 1.390e-4 1.148e-4 2.200e-4 7.011e-5 0.0128 0.0169 0.0308 0.0396 2.421e-4 1.984e-4 3.681e-4 1.135e-4 SD/ g of MD aw SD/ aw units wNaCl/g g-1 SD/ g of NaCl

6.3.2.3 Calorimetric investigations

Aqueous solutions of single maltodextrins were mixed with pure water at 313.15K in a batch calorimeter (model MS 80, Setaram, Lyon, France). Specially designed cells were used to enable isothermal and isobaric mixing. In the upper compartment a small cylinder with a piston which separates the solution in the cell from compressed air allowed pressurization at 2 bar. The cells were filled with syringes in such a way that no gas bubbles were present. The cell is separated in two parts (containing the polymer solution in the upper compartment and the pure water in the lower compartment) by a Teflon seal. To start the dilution process, the seal was cut off and the heat upon dilution was determined from the integrated signal of the thermocouple of the calorimeter. To enhance mixing due to high viscosity of the maltodextrin solutions, metal balls and a rotation of the entire calorimeter was performed. To check the accuracy of the

experimental data, an electrical heater was immersed in the experimental cell that was filled only with water. A constant current was sent through the resistor by a current source. The heat evolved was estimated from voltage, resistance and the time during which the current was applied. These values were compared with those from the integrated signal of the thermocouple of the calorimeter. Relative mean deviations of 1% were encountered between these two values.

140

Capítulo 6 The experimental results are presented in Tables 6.6 and 6.7 for the MD 13 –17 and 16 – 19.

Superscripts bottom and top designate the solutions placed in the lower and upper compartment of the cells. Note that the measured heats of dilution are negative, corresponding to an

exothermic behaviour upon dilution and that the heat involved in the mixing process is small. Dextran, another polysaccharide, presents the same exothermic behaviour as reported by Großmann et al. (1995), who in determined heats of dilution, and Silva and Loh (2000) for heats of solution. It must be stressed that for each experiment, the heat dissipated on cutting off the Teflon seal was determined after the dilution process by once again pushing and pulling the knife. The heat evolved in this procedure was lower than 1 J and its values were taken into account in the experimental values reported in Tables 6.6 and 6.7. For the maltodextrin MD 4 –7 it was observed two different effects upon dilution: an endothermic in the beginning of the experiment, and another one exothermic about 90 minutes after the cutting off the Teflon seal. In this case, it can be supposed that in the beginning an endotermic effect takes place due to dilution of the polymer solution, that is a mixture of gel and solution, in pure water, and the second, due to the rearrangement of the molecules to build a new gel network. Some authors studied the formation of thermally reversible potato starch

maltodextrin gels by H-NMR, viscometry and electronic microscopy (Schierbaum et al., 1984, 1992). The main characteristic of these maltodextrins in aqueous solutions is that the progress of the gelling process is time-dependent and the transition takes place from a homogeneous structureless solution to a two-phase system. The last one consists of the solid aggregation network and the liquid solution of non-structured MD low molar mass components. A rapid increase in the gelling process is encountered at high concentrations and low temperatures. Due to difficulties in analysing the experimental data, because of the errors on pushing the pushing knife of the cells and the mixing itself involving a gel system, it was decided that these systems are deserving of much experimental work that are not to be considered in the present paper. Therefore, the experimental heats of dilution are presented only for the other two maltodextrins that do not form gel. The capacity of gel formation is enabled by the presence of enough long chains in solution, as in the case of MD 4 – 7, which has a significant high molar mass fraction as pointed out in Figure 6.1. According to Schierbaum et al. (1984), products with dextrose

equivalent (DE) up to 5 – 8 are able to form thermally reversible gels at concentration above 10 mass %.

Experimental
Table 6.6. Experimental heats of dilution of maltodextrin 13 – 17 at 313.15 Ka
mwtop/g 20.480 19.416 20.161 17.911 18.230 17.970 19.618 20.492 18.277 18.335 17.405 17.479 14.686 14.592 12.904 13.173
a

141
mwbottom/g 15.938 15.857 15.902 15.916 15.855 15.905 14.805 14.976 14.783 14.822 14.979 15.000 14.866 14.772 14.854 14.855 wMDtop/g g-1 0.290 0.290 0.300 0.398 0.404 0.398 0.508 0.481 0.548 0.548 0.579 0.579 0.638 0.638 0.694 0.694 Q/J -1.193 -1.234 -1.220 -1.682 -2.051 -1.825 -3.157 -2.805 -3.640 -3.662 -4.490 -4.492 -5.887 -5.464 -6.686 -6.773

wMDbottom/g g-1 was zero for all experiments.

Table 6.7. Experimental heats of dilution of maltodextrin 16 – 19 at 313.15 Ka
mwtop/g 22.939 22.867 19.857 18.863 18.259 18.406 16.432 16.195 14.924 15.312 12.543 12.818
a

mwbottom/g 15.014 14.953 14.779 15.472 15.057 15.030 15.038 15.039 14.871 14.953 14.801 15.021

wMDtop/g 0.404 0.393 0.501 0.511 0.548 0.543 0.608 0.608 0.649 0.637 0.700 0.700

Q/J -3.502 -2.616 -4.487 -3.899 -6.432 -6.026 -8.938 -8.478 -10.340 -9.610 -12.210 -12.189

wMDbottom/g g-1 was zero for all experiments.

142
6.3.2.4 Densimetric data

Capítulo 6

In order to convert polymer mass fractions (w in g g-1) into concentrations (c in g cm-3), the specific densities (ρ* in g cm-3) of the maltodextrin solutions were determined at 298.15 K. Density measurements were carried out in triplicate using a digital densimeter (DMA 58, Anton Paar) which was calibrated with water and air as standards at the corresponding working temperature. The accuracy of the density measurements was estimated as 3×10-5 g cm-3. The experimental results for MD 13 – 17 and MD 16 – 19 are shown in Table 6.8 below.
Table 6.8. Specific density of aqueous MD 13 –17 and MD 16 – 19 solutions at 298.15 K
MD 13 –17 wMD/g g-1 0.0499 0.0999 0.1487 0.2004 0.2497 0.2996 0.3512 0.4025 0.4992 0.5491 ρ*/ g cm-3 1.01573 1.03516 1.05472 1.07628 1.09761 1.11995 1.14387 1.16867 1.21786 1.24457 wMD/g g-1 0.0500 0.0995 0.1416 0.1988 0.2506 0.2878 0.3489 0.3980 0.4507 0.4938 0.5392 0.5941 MD 16 –19 ρ*/ g cm-3 1.01591 1.03525 1.05224 1.07614 1.09869 1.11537 1.14380 1.16776 1.19390 1.21650 1.24056 1.27196

No significant difference between the two sets of experimental data of specific density was observed for both maltodextrins at the studied temperature.
6.3.2.5 Laser-light scattering experiments

The multiangle laser-light scattering device utilized for the determinations of the light scattering from maltodextrin solutions was the same as that in the GPC experiments. The difference is that multiple concentrations of the same non-fractionated polymer sample are pumped in the optical cell to find how the scattering intensity varies with concentration. A flow rate of 0.5 cm3⋅min-1 was employed using a syringe pump (Infors, type Predicor). Filters (Sartorius) with pore size 0.20 µm were used in the line between the syringe and the flow cell to

Experimental

143

remove dust from the aqueous solutions and avoid great noises in the measurements. The experiments were performed at 298.15 ±0.2 K and at least 6 different concentrations were pumped into the optical cell. The results of the laser-light scattering experiments are presented in Tables 6.9 and 6.10 for the maltodextrins MD 13 –17 and 16 –19, respectively.
Table 6.9. Excess Rayleigh ratios of MD 13 – 17 in water at 298.15 K
100 cs/g cm-3 θ/o 28.71 8.117 7.909 36.2 8.155 7.903 8.088 7.99 54.0 8.038 7.976 64.9 105K cs(REθ)-1 mol g-1 8.135 8.137 8.077 8.092 90.0 8.313 8.355 115.1 8.242 8.304 125.4 8.170 8.234 135.5 8.220 8.336 143.8 8.068 8.202 8.444 8.227 8.387 8.252 8.375 8.294 8.313 8.296 8.436 8.429 8.354 8.396 8.591 8.629 8.520 8.577 8.445 8.534 8.509 8.611 8.388 8.512 8.847 8.882 8.860 8.878 8.846 8.858 8.759 8.777 8.890 8.909 8.785 8.817 9.021 9.048 8.922 8.951 8.842 8.890 8.909 8.969 8.793 8.896 9.268 9.177 9.212 9.093 9.234 9.153 9.141 9.055 9.278 9.230 9.158 9.108 9.416 9.381 9.294 9.289 9.208 9.209 9.281 9.298 9.150 9.199 9.605 9.547 9.600 9.503 9.597 9.529 9.506 9.430 9.632 9.578 9.511 9.464 9.742 9.725 9.618 9.615 9.535 9.537 9.593 9.626 9.467 9.544 10.030 9.951 9.959 9.918 10.010 9.926 9.863 9.835 10.020 9.983 9.869 9.865 10.130 10.120 9.991 10.000 9.883 9.924 9.968 10.020 9.826 9.939 10.340 10.160 10.320 10.250 10.340 10.240 10.230 10.180 10.380 10.330 10.240 10.220 10.500 10.480 10.360 10.360 10.270 10.290 10.330 10.360 10.210 10.250 10.700 10.530 10.690 10.670 10.710 10.590 10.590 10.550 10.750 10.680 10.610 10.610 10.870 10.840 10.740 10.730 10.640 10.680 10.720 10.740 10.600 10.670 0.720 1.438 2.183 3.001 3.617 4.341 5.076 5.805

44.5

77.1

217 1.210 1.317 1.083 1.409 54.230 1.191 1.111 1.405 1.919 Capítulo 6 5.233 1.156 125.1 105K cs(REθ)-1 mol g-1 1.408 2.116 1.644 1.112 1.530 4.135 1.294 1.354 1.243 1.269 1.131 115.312 1.298 1.9 1.315 1.199 1.233 1.216 1.316 1.244 1.329 1.153 1.043 44.341 1.185 1.116 77.374 1.319 1.107 1.2 1.367 1.285 1.387 1.279 1.9 1.298 1.144 Table 6.15 K 100 cs/g cm-3 θ/o 36.283 1.111 2.1 1.401 1.295 1.361 1.318 1.158 1.345 1.107 1.375 1.389 1.214 1.400 1.160 143.138 1.058 1.249 1.179 1.250 1.0 90.188 1.14 1.118 1.224 1.087 1.198 1.344 1.061 1.400 1.079 1.087 1.149 1.051 1.068 1.342 1.217 1.359 1.262 1.280 1.379 1.073 1.230 4.0 .382 1.264 1.083 1.297 1.128 1.411 1.117 1.387 1.5 1.366 1.079 64.329 1.115 1.152 1.234 1.245 1.138 1.155 1.274 1.265 1.133 1.176 1.366 1.413 1.243 1.368 1.4 1.148 1.352 1.246 1.321 1.108 1.5 1.214 1.188 1.10.285 1. Excess Rayleigh ratios of MD 16 – 19 in water at 298.354 1.223 1.244 1.375 1.162 1.148 1.242 1.196 1.152 1.251 1.379 1.124 1.158 1.102 1.354 1.337 1.388 1.352 1.332 1.382 1.8 1.244 1.200 1.222 1.265 1.124 1.403 1.298 1.118 1.121 1.288 1.149 1.189 1.295 1.253 1.151 1.407 1.151 135.273 1.823 3.292 1.400 1.159 102.

4 6.Modeling 6. In a solution of a polydisperse sample it is possible that the virial coefficients are molar mass dependent. The description of polydispersity can be performed by a molar mass . di. it is possible to evaluate at cs=0 the number average molar mass Mn and the second and third virial coefficients. 1982): osm A2 = i =2 j =2 ∑ ∑ wi w j Aij (6. wi is the weight fraction of component i.7) N N N osm A3 = i =2 j =2 k =2 ∑ ∑ ∑ wi w j wk Aijk (6.8) N N cs = and i =2 ∑ ci N where N is the number of components in a polydisperse sample.s   M + A2 cs + A 3 cs + .. Mn is the molar mass of the solute and A2osm and A3osm are the second and third osmotic virial coefficients of the solute in the solvent.  ρ1  n  (6.4. For polydisperse polymer samples the second and third virial coefficients are given by (Kurata. ci is the concentration of component i.6) where a1 is the water activity. cs is the concentration of solute in mass/volume. If the experimental reduced water activity (-ρ1 ln a1/cs) is plotted over cs. ρ1 is the molar density of pure water. and in this case. and Aij and Aijk are the second and third mixed virial coefficients.1 Modeling Osmotic virial equation 145 The thermodynamic data of aqueous maltodextrin and mono-.. respectively. it is necessary to consider the polydispersity of the sample in the osmotic virial equation.and trisaccharides were correlated by the osmotic virial equation:  c  1 osm osm 2 ln a1 = .

maltose and maltotriose to oligo.4 1. the studied maltodextrins are very polydisperse samples containing components varying from mono-.0792 0. Table 6.3.8348 Mn (von GPC) 180.1 342.0184 0.8760 0. as shown in Figure 6. (1995) for the virial coefficients of poly(ethylene glycol)s. it was assumed that the maltodextrins in this work consist of glucose (G).0266 0.2 504.and weight average molar masses of these components are given in Table 6.4098×104 180.1 342. the molar masses varied between 3000 and 1.0594 0.15 no final deste capítulo. The great polydispersity can confer the maltodextrins varied properties in aqueous solutions and this is the objective here: the study of thermodynamic properties of low molar mass saccharides and maltodextrins in diluted (laser-light scattering) and concentrated (isopiestic method) solutions. The results of such evaluation showed that the virial coefficients do not depend on the molar masses of the polymer components.0 180.2 504.1 342.and trisaccharides such as glucose. The same observation was found by Hasse et al.4 1909. maltose (M). Thus. bem como suas massas molares são apresentadas na Tabela 6.161×103 a as frações mássicas dos oito componentes da fração polidispersa. di.and polysaccharides.2 504.4 2552. .11a. assuming that the molar mass distributions of these polymers were approximated by 8 pseudocomponents. As it was seen before.11.1 342.2 504.0449 0. maltotriose (MT) and a polydisperse fraction (PF) which was splitted in 8 components.2 Mw (von GPC) 180.4 9. This procedure was employed to polydisperse dextran samples in the work of Kany et al.0606 0. that are much more narrowly distributed than dextrans.5×105 g mol-1. (1999). Division of components in the maltodextrin samples Component MD 13 –17 Glucose (G) Maltose (M) Maltotriose (MT) Polydisperse fraction (PF) MD 16 –19 Glucose (G) Maltose (M) Maltotriose (MT) Polydisperse fraction (PF) wi/g g-1 0. The weight fractions and number. In this case.146 Capítulo 6 distribution function or by splitting the distribution in more monodiperse components named pseudocomponents.

7-6.Modeling data the evaluation is based on the following equation: Kc s ρ ∂µ1 =− 1 E RT ∂c s Rθ 147 Equation 6. For the laser-light scattering (6.13) The mixed virial coefficients Aij and Aijk in equations 6.9 and 6.6 as follows: µ1 µ1. 1982): LS A2 = ww M M A 2 ∑ ∑ i j i j ij M w i =2 j =2 1 N N (6.12) LS A3 = ww w M M A − 2 ∑ ∑ ∑ i j k i j ijk M w i =2 j =2 k =2 4 1 N N 3 3M w i = 2 j = 2 k = 2 l = 2 ∑ ∑ ∑ ∑ wi w j wk wl M i M j M k M l ( Aij A jk − Aik A jl ) N N N N (6. and the derivative in the right side can be evaluated based on equation 6.6 is used in the correlation of water activity data.13 were approximated by geometric mixing rules as given below: Aij = ( Aii A jj )1/2 (6. 3 Mw (6. and A2 and A3 are second and third virial coefficients...12-6.9) where the terms on the left side are experimental and are given in Tables 6.11) LS LS Mw is the weight averaged molar mass.8 and 6.10.14) Aijk = ( Aiii A jjj Akkk )1/3 (6. respectively given by (Kurata. pure = + ln a1 RT RT Kc s E Rθ (6.15) .10) = 1 LS 2 + 2 A2 cs + 3 ALSc s + .

e. The results of such correlation are given in Figure 6. i. Note that in this figure.015 -1 -ρ cs ln a1 / mol g 0. Thus it was assumed that the second virial coefficients for glucose (AGG). b.7 0. 2002) Maltose (Miyajima. A simultaneous evaluation of laser-light scattering and isopiestic data was performed using a data bank comprising experimental water activity and light scattering data for maltodextrins determined in this work. for polydisperse polymers. 2002) Maltotriose (Miyajima. 1955. 2002).3 0.020 0.148 Capítulo 6 It can be noted from equations 6. Reduced water activity in aqueous solutions of saccharides at 298. The assumption that the second and third virial coefficients do not depend on the molar mass of the solute components did not result in a good agreement between experimental and correlated values..6 -3 0. 1955) Glucose (Cooke.2 0.7-6.5 for reduced water activities vs. 1983) Glucose (Taylor. 1983) MD 13-17 MD 16-19 correlated -1 0.005 0. . Myiajima. 1983) Maltose (Uedaira.8 and 6.5 0.1 0.12-6.4 0. 1983a.15 K obtained from the literature (Taylor. the LS LS osm light-scattering virial coefficients A2 and A3 differ from osmotic virial coefficients A2 osm and A2 . maltotriose (AMTMT) and polydisperse fraction (APFPF) would be a function of the molar mass of these components and that the third virial coefficient was the same for all components. and experimental water activity data for glucose.000 0. maltose (AMM). Cooke. 1969) Maltose (Cooke.9 1. 1969. AGGG = AMMM = AMTMTMT = APFPFPF. the results are getting worse at high solute concentrations of the low molar mass compounds and that the model underestimate the reduced water activities of the maltodextrins. whereas for monodisperse polymers there is no difference between osmotic and light- scattering virial coefficients.8 0. maltose and maltotriose at 298.0 cs / g cm Figure 6.15 K.010 Glucose (Myiajima.0 0. cs. Uedaira. 0.5.13 that.

the low molar mass fractions of these polymers corresponding to the components glucose. To evaluate the light scattering data of maltodextrin systems. Dependence of second and third virial coefficients on the molar mass. This means that both concentration and molar mass of such compounds were discounted from raw data in the calculation of light scattering.0x10 -3 Aii or Aiii MT 6. which were simultaneously correlated with reduced water activity data. Thus.5 10. This assumption. as shown in Figure 6. is supported by the fact that the signal provided by the device MALLS for low molar mass compounds is almost null.12. maltose and maltotriose and that for the defined polydisperse fraction the values assume a constant value.7.0 ln M Figure 6.5 9. The obtained results showed that this dependency is clearly verified for the low molar mass saccharides glucose. The results can be seen in Figure 6. .0 5.5 7.8.0 7.2x10 -2 1.5 8. as it was commented before. The correlation provided values of Mw within the experimental error of the GPC determinations.0x10 -3 0. maltose and maltotriose were not considered in the calculations. It is also noteworthy to observe that the considered dependency leads to better agreement between experimental and correlated values of reduced water activity of the studied systems.5 12.Modeling 149 Another attempt to improve the correlation results was to consider the dependence of the third virial coefficient on the molar mass of the different components (AGGG ≠ AMMM ≠ AMTMTMT ≠ APFPFPF).6.0 6. as presented in Figure 6.6 and Table 6.4x10 -2 G 1. Besides virial coefficients.0x10 -2 Aii Aiii smoothed M 8.0 8.5 11.0x10 -3 PF 4. it is assumed that the Kcs/RθE data corresponds to the scattering of the higher molar mass molecules.0 10. 1.5 6. the weight molar masses (Mw) of the maltodextrins were also simultaneously correlated.0 9.0 5.0 11.0x10 -3 2.

0x10 -4 Kcs/Rθ.0 /mol g -1 8.07 cs/ g cm Figure 6.05 0.0 0. 1983) Glucose (Taylor.7.010 0.8 0.0x10 -5 E 4. Laser-light scattering data of aqueous maltodextrin solutions. 1983) Maltose (Uedaira. Reduced water activity in aqueous solutions of saccharides at 298.005 0.03 0.5 0.0 0. 2002) Maltose (Miyajima.2x10 -4 1. 1.020 Glucose (Myiajima.150 0.15 K.2 0.01 0. 2002) Maltotriose (Miyajima.06 0. 1983) MD 13-17 MD 16-19 correlated Capítulo 6 0.1 0.0 cs / g cm Figure 6.4 0.7 0. 1955) Glucose (Cooke. .02 0. 1969) Maltose (Cooke.0x10 -5 0.8.000 0.04 -3 0.6 -3 0.0x10 -5 6.3 0.0x10 -5 MD 16-19 MD 13-17 correlated 2.9 1.00 0.015 -1 -1 -ρ cs ln a1 / mol g 0.

developed by Grossmann et al.m 2  1000  Θi Θ j Θ Θ Θ M1 1000  ∑ ∑ ∑ i j k B (6.Modeling 151 Table 6.16) ln a1 = − Aij − 2 ∑ ∑ ∑ mi − ijk  M  M 1 i ≠ 1 j ≠ 1 Θ1 Θ1 1000 i  1  i ≠ 1 j ≠ 1 k ≠ 1 Θ1 Θ1 Θ1 2 Θj  1000   1000  qi   = ln mi + 2  M  q ∑ Θ Aij + 3 M   i. This equation is based on a semi-empirical group contribution approach for the excess Gibbs energy with a structure similar to Pitzer’s virial equation for electrolyte solutions (Pitzer.2 VERS Model Another equation used in this work to evaluate experimental data of water activities and enthalpies of dilution of the maltodextrins and low molar mass saccharides is the VERS model.12. was used in the modelling of the systems in this work.000572 APFPF = 0.17) Θi = mi qi ∑ all components j m jq j (6.011556 AMMM = 0. The activity coefficients are normalized according to the asymmetric convention: for water the standard state follows Raoul’s law and for the other components it follows Henry’s law. (1995) and used successfully to correlate and predict some thermodynamic properties of aqueous PEG and dextran solutions.005230 APFPFPF = 0. 1991). m  1  1 j ≠1 1 1   Θj Θ kB q1 j ≠ 1 k ≠ 1 Θ1 Θ1 ijk qi ∑ ∑ ln a* (6. The VERS model (Virial Equation with Relative Surface Fractions).4. In this model. Second and third virial coefficients of the saccharides and polydisperse fraction studied in this work AGG = 0.003075 6.18) .000213 AGGG = 0.003353 AMM = 0. The equations for the activities of water (aw) and solute ( a* ) are given below: i.001315 AMTMT = 0.007468 AMTMTMT = 0. the solute concentration is expressed by its surface fraction per 1000 g of water normalized by the surface fraction of water.

22) E for the partial molar excess enthalpies of water ( h1 ) and solute i ( hiE ) the equations are:  ∂( µ E / T )    E i h1 = −T 2   ∂T     p . respectively.23) j Θ Θ j ∂Aij  1000  ∑ ∑ i = T  M  RT  1 i ≠ 1 j ≠1 Θ1 Θ1 ∂T Θ j ∂Aij  1000  qi  = −2T  ∑  M q RT  1  1 j ≠ 1 Θ1 ∂T  β( 2 )    Θ(i)Θ( j )  β( 1 ) + lm  m l lm T   all groups all groups   ∑ l (6.19) where.n E h1 (6. the surface fractions of water and component i. The interaction parameters Aij and Bijk are given by: Aij = ∑ ∑ Θ( i )Θ( j ) alm m l (6.24) hiE≠ 1 (6. and qi the surface parameter of component i.21) all groups all groups all groups l m n Q Θ( i ) = ν ( i ) l l l q i (6. Θ1 and Θi are. Ql the group l surface parameter.20) all groups all groups l m Bijk = ∑ ∑ ∑ Θ( i )Θ( j )Θ( k )blmn m n l (6. vl the number of groups l in component i.26) m . mi is the (i ) molality of component i.152 Capítulo 6 qi = ∑ allcomponents l ( vl i )Ql (6.25) ∂Aij ∂T = ∑ (6.

maltotriose and a polydisperse fraction (method 2).200 for hydroxyl groups.15 .5620 for the pyranose rings. Number-average molar masses for method 1 are the same as those presented in Table 6. where Mn is the number 162. 0. The estimation of the model parameters presented in equation 6.142 average molar mass of the solute that is subtracted from the molar mass of one molecule of water and divided by the molar mass of one molecule of glucose minus one water.3. Here. n is calculated as: n = M n − 18. n alkane (CH2) and (3n+2) hydroxyl groups. considers the polydispersity of these polymers assuming that the maltodextrins consist of four different components: glucose. i. The concentrations and molar masses used in method 2 were already presented in Table 6.e. n-1 osidic bonds (O).0152 .21 was performed as (0) following: the coefficient βlm = β (0) .1. the polydispersity is neglected and the low and high molecular mass compounds are considered together in a same fraction (method 1).442 for osidic bonds. maltose. the following empirical (6.Modeling expression was used: (0) (1) (2) alm = βlm + βlm (T / K)(1 − (T0 / T )) + βlm ln(T / T0 ) 153 To account for the influence of temperature on the equilibrium.540 for CH2 and 1.pyr or β (0) . All parameters are assumed to be symmetric.o or β (0) . βlm) = β ( j ) .15 K is a reference (j temperature. (1995) to describe carbohydrate molecules. This division of groups was proposed by Catté et al. alm is a binary group site interaction parameter and T0=298.27) where. 0. The concentration of the low cited molar mass components were estimated by the Gauss-Lorentz functions as presented in Figure 6.e.. another one. ml For the assignment of groups. water is treated as a single group and the maltodextrins consist of four structural groups: n pyranose rings. considers the maltodextrin as only one component i.11. The group surface parameters of the structural groups are estimated by the method of Bondi (1964) and equal to: 1. which concentration and molar mass is estimated from the subtraction of concentrations and molar masses of the low compounds. The molecules of the maltodextrins studied in the present work were treated in the modelling in two different ways: one. The polydisperse fraction is the fourth component called in this work as PF.CH or β (0) .OH were fitted to pyr pyr pyr pyr 2 experimental results for the water activities of saccharides and maltodextrin solutions at 298.

29) The values of the estimated parameters for both methods are given in Table 6. Figure 6. which is not so much in comparison with glucose that. As can be seen in Figure 6. maltotriose and maltodextrins was used in the correlation for the determination of interaction parameters. where the predictions fail (see Figures 6. for method 1).13 below and the results of correlation in Table 6. These parameters were estimated by minimizing the sum of squares as follows: SSQ = exp calc ∑ ( aw .aw ) Nexp 2 (6. at the same concentration.15 K. the calculated values are in close agreement with experimental data even at high polymer concentrations. The same occurred to the predictions of other thermodynamic properties for these carbohydrate systems except for very high solute concentrations. di-. maltose. A data base including experimental water activities of glucose.28) 2 SSQ = ∑ (Q exp . O.10 and 6. reduces the aw by 16%. and trisaccharides in aqueous solutions reported in literature. The studied maltodextrins reduce the activity of water by only about 4 % at 60% mass fraction. The predictive capabilities of the VERS model were also tested using other two kinds of experimental data for glucose and maltose: freezing point depression and boiling point elevation.9 shows the results of correlation obtained for water activity in water – maltodextrin solutions by method 1 as well as the prediction results for the water activities of mono-. CH2 and HO were fitted to the experimental results for the heat of dilution of aqueous maltodextrin solutions in pure water at 313.Q calc ) Nexp (6. the parameters βlm and βlm with l or m=pyr.14.11.154 Capítulo 6 (1) (2) K.9. .

CH pyr 2 β (1) .2193×10-2 2.2612×10-5 -1.and oligosaccharides at 298.3399×10-5 9.CH pyr 2 β (0) .6652×10-3 9.88 0.85105×10-1 8.64279×10-1 -1.O pyr β (0) .94 a1 0.1812×10-4 3.15 K (method 1).96 0.9.4 -1 0. .2687×10-4 -3.OH pyr β (1) . Experimental and calculated water activities of water – maltodextrin and some mono.5 0.92 MD 13-17 (this work) MD 16-19 (this work) glucose (Miyajima.7212×10-1 -6.pyr pyr β (1) .5878×10-4 -5.6464×10-3 -1. VERS model 0.0352×10-1 -3.pyr pyr 1.13.60260×10-1 7.00 0.Modeling 155 Method 1 6.53039×10-1 -8.3 0.8649×10-2 3. Interaction parameters of VERS model for both methods Binary interaction parameters β ( 0) − pyr pyr β (0) .1 0.90 0.98 0.9689×10-2 Table 6. 1983) corr.7974×10-4 Method 2 3.1716×10-1 2.6 wi / g g Figure 6.86 0.2 0.0 0. 1983) maltotriose (Miyajima.O pyr β (1) .OH pyr β (2) . 1983) maltose (Miyajima.

156
280 275 270

Capítulo 6

Freezing temperature / K

265 260 255 250 245 240 235 230 225 0.0 0.1 0.2 0.3 0.4 0.5
-1

glucose (Young, 1957) maltose (Uedaira & Uedaira, 1969) predicted VERS model

0.6

0.7

0.8

wi / g g

Figure 6.10. Experimental and predicted freezing temperatures for glucose and maltose at
various concentrations (method 1).
392 390 388 experimental (Abderafi & Bounahmidi, 1994) predicted VERS model

Boiling point elevation/ K

386 384 382 380 378 376 374 372 0.0 0.2 0.4 0.6 0.8 1.0

wi / g g

-1

Figure 6.11. Experimental and predicted boiling temperatures for glucose (method 1).

Modeling

157

The influence of temperature on the activity of water is almost negligible i.e., the water activity remains nearly unchanged when temperature is increased from 298.15 to 318.15 K. This fact is confirmed by the calorimetric measurements – the heats of dilution for the aqueous maltodextrin solutions are very small – and also by experimental data on water activity at 318.15K. The heat on diluting an aqueous solution of maltodextrin in pure water measured in the present work were also correlated by the VERS model. Figure 6.12 shows the calculated and experimental results for MD 13 – 17 and 16 –19. As can be seen in this figure, the experimental enthalpies of dilution for both maltodextrins have a slightly dependency on the molar mass and the heat effects involved are small. Even so, the model can distinguish between the maltodextrins in calculating the enthalpies involved in the process of dilution. Small values for heats of dilution were also observed in experiments with dextran solutions, as reported by Großmann et al. (1995).
0 -2 -4 -6 -8

Q/J

-10 -12 -14 -16 -18 0.0 0.1 0.2 0.3 0.4
-1

MD 16-19 VERS model MD 13-17 VERS model

0.5

0.6

wi / g g

Figure 6.12. Experimental and calculated heat of dilution in aqueous solutions of MD 13–17 and MD 16–19 at 313.15 K. (method 1).
The average absolute deviations (%AAD) between experimental and estimated values for all thermodynamic properties and methods 1 and 2 are presented in Table 6.14. As can be seen in Table 6.14, the lower global AAD result is provided by method 1, which does not consider the polydispersity of the maltodextrins, but the difference between both

158

Capítulo 6

methods is not so large. To verify the applicability of the adjusted parameters by both methods to predict water activities of solutions containing high and low molar mass compounds, the experimental data from Radosta et al. (1989) for different fractions of a maltodextrin sample was compared with calculated aw from VERS model. The predictions show that the model was able to distinguish the aw lowering behaviour even in the systems with very high molar mass components (n=120 indicates a molar mass of approximately 19 500 g mol-1). The calculation of these water activities by VERS model with the parameter adjusted by method 1 is slightly better than that adjusted by method 2. The predictions resulted from method 1 are shown in Figure 6.13. The AAD% for the two attempts are 0.05 and 0.09 % for methods 1 and 2, respectively.
1.000

0.995

0.990

aw

0.985

0.980

n 120 30 (MD) 16 9 predicted VERS model
0.0 0.1 0.2 0.3 0.4 0.5

0.975

w/gg

-1

Figure 6.13. Results of the predictions (method 1) for aw of maltodextrin fractions (n is the polymerisation number). Experimental data from Radosta et al. (1989).

0904-0.2598-0. 2. Probably the reason for this behaviour is the low concentration of the low molecular saccharides in the maltodextrin samples possibly the influence of those saccharides on the thermodynamic properties of maltodextrins measured in the present work is small and to consider their influence as an independent factor do not bring any improvement to the correlation of the properties. 5 1. 1983.0177-0. 6. 3. 4 Velezmoro et al.15 298.30 1.7952 0.1750-0.15 0.09 0.7478 0. on the contrary.4615 298. 8 Abderafi and Bounahmidi.65 0. 2002. 1994.13 0.6567 0.1382-0.93 5 8 Miyajima et al.Modeling 159 Table 6. it was expected that the results for the method 2 would provide better results for calculating the thermodynamic properties.04 0.44 1.23 0. 4 1.5850 0. 1957. Range/ g g-1 Temperature/K %AAD (method 1) %AAD (method 2) References Water activity glucose maltose maltotriose maltotetraose maltopentaose MD 16 – 19 MD 13 –17 MD 4 –7 0.8372 0.99 317.01 0. 3 3 3 this work this work this work 6 7 Freezing point depression glucose maltose 0..37 2.15 – 318.55 0.005-0.13 0. 1969.02 Boiling point elevation glucose global AAD 1 6 0. 2000. At the beginning.80 3.55 0. 7 Weast. The results show.42 0. 3.15 – 318.99 298. Uedaira and Uedaira.15 – 317.15 – 317.19 0.5898 0.61 1.0218-0.99 298. Average relative deviations between experimental and calculated data by VERS model Saccharide Conc.63 0. This data covered a wide concentration range from diluted to concentrated solutions.99 317.0331-0.22 0.7 0.2859-0. Therefore it was possible to evaluate the virial coefficients in the . light scattering and heats of dilution of aqueous solutions of three different maltodextrins were investigated and evaluated by a osmotic virial equation and a semiempirical group contribution model.15 – 317. 1973.44 0.15 298.14 0.5 Conclusions The water activity.40 0. Young. 2 Taylor.14.38 0. 1955. that method 1 was better to estimate the studied properties.005-0.99 298.72 0.7750 0. 3 Cooke..

J. Cesàro. Defloor. Bounahmidi. J. Catté. The model was also able to predict other thermodynamic properties for glucose and maltose such as freezing point depression and boiling point elevation. Glucose. Thermochimica Acta 328 (1999) 143-153. Fructose and Water Components. Chromatography 803 (1998) 103-109. A. Gros. A Physical Chemical UNIFAC Model for Aqueous Solutions of Sugars. J. C-G. Eng. Thermodynamic Behavior of Mixed Biopolymers in solution and in Gel Phase. J.. P. I. Delcour. Van der Waals volumes and Radii. S. M. 6. Fabri. Fluid Phase Equilibria 93 (1994) 337-351. P. Phys. Westh. Sussich.. 68 (1964) 441-451. J-B. 6.. Measurement and Modeling of Atmospheric Pressure Vapor- Liquid Equilibrium Data for Binary.6 Acknowledgments The authors gratefully acknowledge the financial support of CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico . Grobet. . S. T. Jónsdóttir.. A Thermodynamic Study of Glucose and Related Oligomers in Aqueous solution: Vapor Pressures and Enthalpies of Mixing.. S. Bondi.160 Capítulo 6 systems containing molecules varying from saccharides to polysaccharides. Cuppo. Ó.. Fluid Phase Equilibria 105 (1995) 1-25. J. P. A. Ternary and Quaternary Mixtures of Sucrose. F. A. Fractionation of Maltodextrins by Ethanol. D. Ithaca. F. It was found that the second and third virial coefficients were molar mass dependent up to maltotriose with (n=3) and that they are constant for higher molar mass components. A. Dussap. Flory. Data 47 (2002) 1185-1192.7 References Abderafi.290011/01-9 + 46668/00-7 + 521011/95-7) and DAAD (Deutsche Akademische Austauschdienst). Cooke. 1953. For the VERS model a good agreement between experimental and correlated values of water activities was found for maltodextrins and low molar mass saccharides systems. Principles of Polymer Chemistry. Chem. Chem. V.. Cornell University Press. J.. Vandenreyken.

and Isopiestic Measurements. A. Clark. Glicksman.. T. Clark. A. Kany.. Carbohydrate Polymers 21 (1993c) 261-268. Maurer. Chem. Kasapis. University of Kaiserslautern. Clark. Academic Press. Macromolecules 28 (1995) 3540-3552. H. New York. Morris. Thermodynamische Eigenschaften Wässriger Polymer-Lösungen.. S.. R.. Pfennig. S. H-P. Data 38 (1993) 163-166... I. G. Ph Thesis. Phase Equilibria and Gelation in gelatin/maltodextrin Systems – Part IV: Composition-Dependence of Mixed-Gel Moduli. Phase Equilibria and Gelation in gelatin/maltodextrin Systems – Part II: Polymer Incompatibility in Solution. H. H. I. Kany.. J. 1998. Carbohydrate Polymers 21 (1993b) 249-259. . Osmotic Virial coefficients of Aqueous Poly(ethylene glycol) from Laser-light Scattering and Isopiestic Measurements. H. R. R. Morris. J. H.. Aqueous Two-Phase Systems of Poly(ethylene glycol) and Dextran – Experimental Results and Modeling of Thermodynamics. A. Kany. Kasapis. T. Norton. Thermodynamic Properties of Aqueous Dextran Solutions from Laser-light Scattering. B. London. Hasse. Kasapis. E. Norton. Tintinger. Light Scattering from Polymer Solutions. Phase Equilibria and Gelation in gelatin/maltodextrin Systems – Part I: Gelation of Individual Components... Morris. H. S. CRC Press Inc. Norton. M. E. Clark. R..References 161 Gaube. G. J. G. Eng. Kasapis. E.. R. 1972. T. Carbohydrate Polymers 21 (1993d) 269-276.. C. Vapor-Liquid Equilibrium in Binary and Ternary Aqueous Solutions of Poly(ethylene glycol) and Dextran.. Data 44 (1999) 230-242. Tintinger. Maurer.. Zhu. Fluid Phase Equilibria 106 (1995) 111-138. Membrane Osmometry. Groβmann.. Morris.. I. Carbohydrate Polymers 21 (1993a) 243-248. A. T. Stumpf. H-P. Norton. I. Phase Equilibria and Gelation in gelatin/maltodextrin Systems – Part III: Phase Separation in Mixed Gels. M. Eng. Hasse. H-P. 1986. Huglin. of Chem. E. A. Maurer.. Food Hydrocolloids v III. S. J. R.. M. Germany..

and Maltotriose in Aqueous Solutions. Nakagaki. Boca Raton. H.. Vorweg. M. Rheological Behavior of Aqueous Dispersions of Cashew Gum and Gum Arabic: Effect of Concentration and Blending. Studies on Aqueous Solutions of Saccharides. Studies on Aqueous Solutions of Saccharides. Starch 41 (1989b) 395-401. II. Kettlitz.. I. CRC Press. F. F. Schierbaum.. Sawada. Apparent Molar Volumes. Bull. G. Yuriev. Part II: NMR Study of Bound Water in Liquid Maltodextrin-Water Systems. H.. Activity Coefficients of Mmonosaccharides in Aqueous Solutions at 25 °C. H. Data 13 (1984) 1-102. Acta Alimentaria Polonica 10 (1984) 69-99. Miyajima. Greek.. Food Hydrocolloids 13 (1999) 501-506. M. Chem. and Activity Coefficients of D-Glucose. M.. Pitzer. Ref. Radosta. Radosta... Maltose. S. Trends in Food Science & Technology 10 (1999) 345-355. Reuther. J.. Thermodynamic Properties of Aqueous sodium Chloride Solutions. W. Radosta. Mothé. Schierbaum. S. H.. L. Anger. Polymer-water Interaction of Maltodextrins.. Towards a Rational Design of Commercial Maltodextrins... A.. K. P. S. Jpn. S.. Peiper.. F. Phys. Maurer. G. S. Radosta. Rao. 2002. Jpn. 56 (1983a) 1620-1623. K. Polymer-water Interaction of Maltodextrins. Sawada. Chem. Pitzer. M.W.. M. J. Bull. Nakagaki. Soc. M. M.. Tramper. C. ESAT Congress. 1991. B.162 Polyelectrolytes and an Inorganic Salt. Reuther. Starch 41 (1989c) 428430. 56 (1983b) 1954-1957. . Beeftink.. F. Activity Coefficients in Electrolyte Solutions. Polymer-water Interaction of Maltodextrins. Capítulo 6 Lammertz. Chem. S. C. Starch 41 (1990a) 142-147. Part I: Water Vapour Sorption and Desorption of Maltodextrin Powders. K. Thermodynamic Properties of Aqueous solutions of Marchal. Miyajima. Soc. F. Schierbaum. F. Schierbaum. Richter. K. Viscosity B-Coefficients. Part III: NonFreezable Water in Maltodextrin solutions and Gels.

E. V. M.. Uedaira. Calorimetric Investigation of the Formation of Aqueous Two-Phase Systems in Ternary Mixtures of Water. Silva. Miyajima. B. 1973. Phase equilibrium in polyethylene glycol/maltodextrin aqueous two-phase systems. S. Kettlitz. Chem. H. M. Cabral. Carbohydrate Polymers 46 (2001). J. Meirelles. Chem. Phys. CRC Handbook of Chemistry and Physics. Jpn. L. Phys. 51 (1955) 1183-1192. M. L. Requirements for Biodegradable Water-Soluble Polymers. Activity Coefficients of Aqueous Xylose and Maltose Solutions. H. A. E.... Jpn. L.. L. Radosta. Rowlinson. Reuther. The Chemical Rubber.. F. A. Formation of thermally reversible maltodextrin gels as revealed by low resolution H-NMR. S. Prediction of Water Activity in Sugar Solutions using Models of Group Contribution and Equation of State.. F. Soc. M. A. Carbohydrate Polymers 42 (2000) 273-278. Phase Equilibrium and Protein Partitioning in Aqueous Mixtures of Maltodextrin with Polypropylene Glycol. Velezmoro. R. Meirelles. H. Uedaira. edn. J. Cleveland.. 53rd. Radosta. F. K. Brando.. C. E.. S. F.References 163 Schierbaum. G. W. Weast. Bull. Eng. Taylor.. A. E. Carbohydrate Polymers 18 (1992) 155-163. Schierbaum. Poly(ethylene oxide) and Electrolytes (or Dextran). Silva. 61 (1957) 616-619.. Zum Stand der Kenntnisse über Struktur-Eigenschaftsbeziehungen von Maltodextrinen. W. Meirelles. H. J.. C.. Trans. M. 33 (2000) 645-653.. Ynriev. P. Loh. Faraday Soc. A. F. Swift.. B 104 (2000c) 10069-10073. Acta Alimentaria Polonica 10 (1984) 69-99. Young.. 42 (1969) 2137-2140. L. Oliveira. Silva. A. J.. Nakagaki. Vorweg.. W. H. J. Polymer Degradation and Stability 59 (1998) 19-24. Chem. A. B.. Richter. The Thermodynamic Properties of Aqueous Solutions of Glucose. J. J. Vorweg. M. A. J. German. D-Glucose-Water Diagram. Chem.. M. . 267274. Sawada.

33 2864.1504 0.1415 0. MD 13 – 17 w / g g-1 0.2 w / g g-1 0.3009 0.1484 0.5 .40 1457.14 5107.10 25062.0539 0.2304 0.48 6990.5 157650.5 141695.83 1239.53 19197.1400 0.0290 0.1146 0.18 1632.1146 0.61 2773.1190 0.33 114665.164 Capítulo 6 Table 6.1085 0.15.31 1221. Components of the polydisperse fraction (PF) of maltodextrins MD 13 – 17 and MD 16 – 19.0120 M / g mol-1 780.2 102304.0097 0.1575 0.0078 MD 16 – 19 M / g mol-1 809.

.

.

.

.

.

.

.

.

.

.

eritritol e glicerol) foram estudadas. Os dados referentes às determinações de aw. O reajuste. Mas foi também verificado que. pode haver um efeito de proximidade intramolecular. para as concentrações a diferença entre o caso real e o ideal é mais significativa (como se vê na Figura 7). D-manitol. Por essa razão. forneceu bons resultados nas correlações e também nas predições de aw em sistemas multicomponentes. 2) O emprego do método de contribuição de grupos UNIFAC combinado ao termo de DebyeHückel mostrou-se uma ferramenta útil no cálculo da atividade de água. e por isso. bem como dados de solubilidades desses compostos em água. O emprego dos valores dos parâmetros de interação dos modelos disponíveis na literatura resultou em predições ruins. principalmente a altas concentrações. xylitol. apesar da diferença não ser tão significativa para o pH. baseado nesse efeito de proximidade. como apresentado no capítulo 2. Como mostrado no Capítulo 3. depressão do ponto de congelamento e ponto eutético quando o modelo UNIFAC-Larsen foi utilizado. e também realizado com um banco de dados abrangendo altas concentrações dos solutos em solução e diferentes moléculas de polióis (com número de carbonos variando de 3 a 6). optou-se por um reajuste de parâmetros considerando que um grupo hidroxila presente. determinando-se suas atividades de água com o emprego de um higrômetro elétrico nas temperaturas entre 10 e 35 °C a várias concentrações. o modelo forneceu praticamente os mesmos desvios para os pH reais e ideais das soluções de aminoácidos estudadas em três diferentes condições de pH: em tampão ácido. Nos cálculos de . No polióis. por exemplo.Conclusões Gerais 175 Conclusões Gerais A seguir. foram empregados na predição e correlação com os modelos de contribuição de grupos ASOG e UNIFAC. os grupos hidroxilas estão ligados a carbonos consecutivos na molécula. básico e em água. 1) Soluções aquosas contendo diferentes polióis (D-sorbitol. tem um outro comportamento em comparação aos grupos hidroxilas que estão presentes nas moléculas dos polióis. são apresentadas conclusões gerais a respeito de cada capítulo apresentado neste trabalho. pH e solubilidade de aminoácidos em soluções aquosas considerando a dissociação parcial desses compostos em água. numa molécula de algum álcool.

Foi observado que. os resultados de predição com valores de parâmetros da literatura eram ruins. para os aminoácidos valina e glicina. O que reforça essa afimação é que. A avaliação dos coeficientes . Com o reajuste dos parâmetros de interação foi também melhorada a capacidade preditiva do modelo para o cálculo de viscosidades de sistemas contendo PEGs. 5) No estudo de propriedades termodinâmicas de sistemas contendo maltodextrinas foram realizadas determinações experimentais e avaliação e cálculo das propriedades pela equação osmótica virial e o método de contribuição de grupos VERS. Os resultados da correlação e das predições são bastante razoáveis. comparando-os. assim como a predição em sistemas multicomponentes. O reajuste de parâmetros da literatura mais uma vez foi necessário. os valores de ∆s e ∆h requeridos para o cálculo de solubilidade não foram ajustados neste trabalho e sim retirados da literatura. Isso poder ser o resultado do ajuste desses parâmetros a um banco de dados que continham moléculas semelhantes entre si. mesmo quando todos os parâmetros de interação de grupos foram reajustados. notou-se também que o modelo foi capaz de fornecer melhores resultados em comparação aos resultados calculados no caso ideal. o que sugere o uso dessa equação para a estimativa de viscosidades de soluções aquosas contendo PEGs. Tem-se duas suposições para tal resultado: 1) extrapolação das viscosidades dos componentes puros (PEGs que se apresentam sólidos nas temperaturas de estudo). dentro do possível. não corretamente descrita pelo modelo.176 Conclusões Gerais solubilidade dos aminoácidos em água e em soluções contendo sais ou ácidos. 2) maior complexidade da viscosidade dinâmica dos sistemas estudados. especialmente para misturas binárias que continham polímeros muito diferentes com relação à massa molar. com dados disponíveis na literatura. 3) As viscosidades de misturas de polietileno glicóis foram determinadas e correlacionadas pelo método de contribuição de grupos GC-UNIMOD. a fim de melhorar os resultados do cálculo da viscosidade pelo modelo. Já o modelo de contribuição de grupos GC-UNIMOD não foi capaz de descrever bem os dados experimentais. 4) O emprego de um banco de dados abragendo soluções aquosas de polietileno glicóis com massas molares entre 200 e 10 000 g mol-1 tornou possível ajustar parâmetros de uma equação semi-empírica para a viscosidade de soluções (equação de Kumar) e relacioná-los ao grau de hidratação desses polímeros.

maltose e maltotriose. pode-se concluir que os métodos de contribuição de grupos podem ser uma ferramenta útil no cálculo de propriedades físico-químicas em sistemas contendo alguns compostos presentes em produtos alimentícios. e que um reajuste de tais parâmetros torna-se necessário para sistemas contendo moléculas muito diferentes dos compostos orgânicos mais tradicionais ou que apresentem determinadas características na molécula. O que deve ser ressaltado é que. . os valores dos parâmetros de interação binários já disponíveis na literatura não são capazes de oferecer uma boa predição de propriedades. foi empregado com sucesso no cálculo de atividade de água e outras propriedades em misturas contendo maltodextrinas e açúcares. como por exemplo. O modelo VERS. muitas vezes. que tem estrutura similar à equação virial de Pitzer para eletrólitos. portanto. Assim.Conclusões Gerais 177 viriais pela equação osmótica (que foram ajustados simultâneamente a dados de atividade de água e espalhamento de luz e. assim como de polímeros empregados em separações de biomoléculas. as hidroxilas no caso dos polióis. atingindo valor constante para moléculas acima de massa molar 504 g mol-1. soluções concentradas e diluídas) mostrou a dependência desses coeficientes com a massa molar de glicose. como os estudados neste trabalho.