CHEMICAL PRODUCT DESIGN
1
“SELECTION”
Last time…
Concluded the second stage of the Chemical
Product Design 4-step procedure:
• Needs: what needs should the product fulfil? What
specifications do these needs imply?
• Ideas: what products/concepts could satisfy these needs?
• Selection: which ideas are most promising?
• Manufacture: how do we turn the design into a commercial product?
This time…
Introduction to the selection phase
Selection using thermodynamics
• Group contribution models
• Solubility parameters
• Exercise
Introduction
The ideas phase and particularly
the screening process left us
around 5 really good ideas
with a lot of potential.
The selection phase involves
more detailed scientific
scrutiny of the good ideas.
We use all scientific tools at our disposal, and do
research and experiments
100 ideas
20 better ideas
5 really good ideas
Introduction
We also evaluate risks, and
take steps to mitigate them
With these tools we try to
make a decision on which
product to manufacture
We can produce a selection
matrix to help us make
that decision
5 really good ideas
1 workable really good idea
“We’ve considered every potential risk, except the risks of avoiding all risks.”
Selection using science
Two fundamental areas:
• Thermodynamics (the science of the possible)
• Kinetics (the science of the how quickly?)
Also make use of a number of other skills and
techniques in many areas of science and
engineering
Selection by thermodynamics
Often we wish to make improved products by removing
harmful or toxic chemicals in existing processes
Eg. CFCs were used as coolants in
fridges for a long time
When they were discovered to
be harmful to the ozone layer
in the 1980s, they were slowly
phased out
Replacement chemicals had to be
found with similar properties
eg. HCFCs
Other examples:Solvents, Asbestos, BPA in baby bottles etc.
Ingredient substitutions
• Main aim: to duplicate product’s properties
• Most common effort: search for less volatile & less
toxic solvents, example:
– Methylene chloride (CH2Cl2) used for fine chemical
manufacturing is carcinogenic
– Acetone (CH3COCH3) useful in lab – more toxic than
methanol
• Option:
– Replace with solvent which is non-volatile, non-toxic
& cheap
– Equal performance with additional benefits: safety
& cheapness
First step
How to predict the properties of the
potential molecules with out doing
experiments?
Property Prediction Methods
• Motivation
– Experiments are time-consuming and expensive.
– How do we identify the components to investigate?
– Components of similar molecular structure have been found to have similar properties.
• Property Prediction
– Physical properties of molecules depends on their topological structure
– Data keeps increasing- so are the models!!!
Property Prediction Methods
Group Contribution Methods
– Physical properties can be estimated from molecular
building blocks.
– Predominant means of predicting physical properties for
new components.
– Based on UNIFAC group descriptions
– Large amounts of experimental property data has been
fitted to obtain the contributions of individual groups.
UNIFAC UNIQUAC Functional-group Activity Coefficients
– Originally developed for predicting activity coefficients in
non-ideal systems.
– Functional groups in the molecules involved in liquid
mixtures are used to estimate activity coefficients.
– Equivalent to UNIQAC model with further breaking
down into the constituent groups.
Eg. UNIQUAC METHANOL+ WATER
UNIFAC ETHANOL: CH3 CH2 AND OH
UNIFAC – Two terms are involved in UNIFAC model
Combinatorial: Individual contributions from molecular groups. Accounts for the deviation from ideality because of the differences in molecular shapes.
Residual: Contribution from interactions between the groups. Correct for the change in interacting forces between molecules upon mixing.
– UNIFAC groups have extended applications in property prediction models such as group contribution methods.
rci lnlnln
Group contribution models
In the Group Contribution Method (GCM), the property of
a compound is estimated as a summation of the
contributions of molecular groups present in the molecular
structure. A correction term has been added to account
for interactions between the groups.
The property estimation model in GCM is
k
kk
i
ii CNCNXf )(
f(X) Function of property X
i Index for group
N Number of groups
C Group contribution
k Index for second order group (interaction)
β Binary variable to check the existence of second order group
Higher order groups
Important
If one building block is completely overlapped by another
one, choose the bigger one
Example
CH3CO-CH2-CH2-CH3
CH3CO is a building block in group contribution/UNIFAC
tables. So, do not break that into CH3 and CO.
Common properties with GC models
Property Property function Group contribution terms
Normal melting point, Tm 0exp mm TT jm
j
jim
i
i TMTN 21
Normal boiling point, Tb 0exp bb TT jb
j
jib
i
i TMTN 21
Critical temperature, Tc 0exp cc TT jc
j
jic
i
i TMTN 21
Viscosity, η ln(η) 21 j
j
ji
i
i MN
Standard enthalpy of formation, Hf 0ff HH jf
j
jif
i
i HMHN 21
Standard enthalpy of vaporization, Hv 0vv HH jv
j
jiv
i
i HMHN 21
Standard enthalpy of fusion, Hfus 0fusfus HH jfus
j
jifus
i
i HMHN 21
Acentric factor, w Cawb
/exp 21 j
j
ji
i
i wMwN
Liquid molar volume, Vm dVm 21 mj
j
jmi
i
i VMVN
Surface Tension, σ σ 21 j
j
ji
i
i MN
Example
Estimate the boiling point of CH3(CH2)3OH
Property model
Group contributions
Adjusted parameter, Tb0= 222.543 K
CH3 CH2 OH
0.8491 0.7141 2.5670
Calculated value: 382 K Measured value: 391 K
Target property
How to look for a good solvent without
doing many experimental works?
Problem
Solubility parameters
A useful tool in selection using thermodynamics
is the use of Hildebrand solubility parameters
This is based on the idea of chemical potentials. For an
ideal solution,
x is the activity coefficient, RT ln(x) is energy of mixing
But most solutions are non-ideal! In this case
w is a measure of how non-ideal the solution is.
solventsolutesolnin solute ln xRT
2
solutesolventsolutesolnin solute ln xxRT w
Solubility parameters
The Margules equation approximates w to
where Vsolute is the molar volume of the solute, and dsolvent
and dsolute are the Hildebrand solubility parameters.
Ideal solutions have dsolvent = dsolute and w = 0.
The more different dsolvent and dsolute are, the less miscible
they are – the worse the solvent.
2
solutesolventsolutesolnin solute ln xxRT w
2
solutesolventsolute ddw V
Solubility parameters
dsolvent is known for most solvents.
However, in most cases we don’t know what dsolute is!
A good way to look for similar properties is to substitute a
working solvent with a known d for one with a similar d.
2
solute
2
solutesolventsolutesolventsolutesolnin solute ln xVxRT dd
Solubility parameters
Example: An alternative solvent to Methanol
2
solute
2
solutesolventsolutesolventsolutesolnin solute ln xVxRT dd
Solvent Hildebrand solubility
parameter d
Methanol 29.61
Ethanol 26.13
Propanol 24.45
Ethylene glycol 33.7
Propylene glycol 29.52
Solubility parameters
The Hildebrand parameter d was later refined
by Charles Hansen by splitting it into three
parts:
dD for the dispersion
dP for the polar interactions
dH for the hydrogen bonding
These are known as the Hansen solubility parameters
2
H
2
P
2
D
2 dddd
Solvent Hildebrand
solubility
parameter
d
Hansen
dispersion
dD
Hansen
polar
dP
Hansen
hydrogen
dH
Cyclohexane (C6H12) 16.8 16.8 0 0.2
Hexachloroethane (C2Cl6) 22.5 22.0 4.7 0
Hydrazine (N2H4) 18.7 14.2 8.3 8.9
Benzene (C6H6) 18.5 18.4 0 2
Solubility parameters
Some more examples:
• Cyclohexane – non-polar, almost no hydrogen bonding
• Hexachloroethane – polar but no hydrogen
• Sometimes better to use Hansen parameters!
Hansen solubility parameters
Using the Hansen parameters is a good way of predicting
the effectiveness of solvents.
The distance of the solvent from the solubility sphere of the
polymer
with the radius of interaction of the polymer R0.
If the ratio is less than 1, then the polymer is probably
soluble in the polymer.
2p
H
s
H
2p
P
s
P
2p
D
s
Da 4 dddddd R
0
a
R
R
Estimation of Hansen parameters
by Van Krevelen method
• Hansen parameters can be estimated from molecular
groups by Van Krevelen’s method
• Fdi, Fji, and Ehi values for a selection of molecular
building blocks are available.
• Vm is obtained from group contribution model.
Comparison of solubility
• Estimate the Hansen solubility parameter for the solute.
• Estimate the Hansen solubility parameter for the
potential solvents.
• Estimate the radius of interaction for all solvents.
• The best solvent is the one with lowest radius of
interaction.
Solubility parameters
A very useful technique based on thermodynamics
for comparing and selecting solvents
Hildebrand parameters very quick and easy to use
Hansen parameters more sophisticated
Can use solubility parameters to make predictions
and pre-selection of solvents and processes
Must check the final results with experiments