Folie 1

1
University of Erlangen, Germany Department of
Chemical Engineering Chair of Separation Science
Technology
Wolfgang Arlt professor and chair Tailor-made
compounds for separation wolfgang.arlt_at_cbi.uni-er
langen.de
?
2

• Nature shows
• There are not so many separation schemes but
  separation or reaction is done via special
  molecules.
• Take distillation as an example
• 1st separation help is the heat introduced for
  evaporation
• 2nd separation help can be an added compound
• Because of limited time, the lecture focuses on
  distillation but also can be made for extraction
  or membranes or reactions.
• I do research on 2 classes of compounds
• ionic liquids
• hyperbranched polymers

?
Why to tailor compounds?
3
Ionic Liquids
?
Part one
4
3rd generation
1st generation
alkylsulfate, phosphate anions
chloroaluminate
corrosive/hydrolysis by water
halogen-free SILP
high halide concentration HCl, HF
low viscosity, high electrical conductivity
?
generations of Ionic Liquids (Wasserscheid,
Chemical Reaction Engineering, University of
Erlangen)
5
An entrainer influences the activity coefficients
of at least 1 component in the separation factor
and is high-boiling
?
world patent by Arlt and coworkers
choice of entrainer in extractive distillation
6
stationary phase with the entrainer (not shown)
Equipment
column
detector
mobile phase
peaks
Component B
Component A
concentration
Chromatogram
time
Start
separation factor
Thermodynamics experiments for the calculation
of ?
7
Gas-Liquid-Chromatography
1 Oven, 2 carrier gas stream (He), 3 heat
exchanger, 4 heat conductivity detector, 5
Injector block, 6 column, 7 Gas bubble flow
meter
Thermodynamics experiments for the calculation
of ?
8
Apparatus for Headspace Analysis
Thermodynamics experiments for the calculation
of ?
9
Ethanol Water Tetrahydrofuran
Water Methylcyclohexane Toluene Hexan
1-Hexene
?
systems under detailed consideration
10
The discovery of ionic liquids was not too
difficult!
use of real salts
11
a smaller side chain in the cation seems to be
favorable
ethanol-water _at_ 1 bar
?
experimental results for polar mixtures
12
IL
phase rich in entrainer
LLE
phase poor in entrainer
comp.1
comp.2
entrainer producing a LLE
13
The rich in entrainer phase produces the
ultimate effect on the azetrope
The poor in entrainer phase produces no effect
on the azeotrope

• An azeotrope comes together with higher activity
  coefficients.
• A LL-demixing comes together with very high
  pronounced coefficients. The activity coefficient
  of the compound showing the demixing is heavily
  influenced. The other phase shows no remarkable
  influence.
• In summary, entrainers causing demixing show
  great effects.
• 2 phases in packed columns can increase the HETP.

The activity coefficient at infinite dilution is
no more the right choice for a property!
?
what influences the impact of the entrainer?
14

• Ionic liquids show effects because of their ionic
  structure.
• Ionic liquids can have long aliphatic side
  chains, reducing the ionic character by shielding
  and turning it more aliphatic
• We believe that the interaction of an IL to a
  solvent is ion-dipole.

schematic data in MMIM Cl
activity coefficient
LLE
LLE
1
??
??
0
0,1
0,5
1
xMethylcyclohexane
?
The dual nature of Ionic Liquids
15
best entrainer among 42 studied ionic liquids
C8Chin R C8H17 (cation)
BTA- (anion)
R
R
?
seperation of hexane 1-hexene
16

• polar mixtures like THF-Water
• selectivity and capacity change in parallel
• shorter side chain in R-MIM increases the
  selectivity for my anions except Cl and
  Acetate
• Cl and Acetate gave the best selectivity with
  any R-MIM
• BTA is very poor
• best suited IL is an optimum of anion and cation
• unpolar mixtures like Methylcyclohexane -Heptane
  or Hexan 1-Hexene
• selectivity and capacity change controversally
• BTA, B(CN)4 gave good results
• BBB, BOB, BSB ) contain aromatic or cyclic
  groups and gave good results

) BBB bis1,2-benzenediolato(2-)-O.O borate
BOB bis(oxalato(2-)) borate BSB bis
(salicylato(2-)) borate
?
are there rules?
17
Hyperbranched Polymers (hyP)
?
Part two
18
VLE THF-water _at_ 1 bar
M.Seiler, J.Rolker, W.Arlt, Phase Behavior and
Thermodynamic Phenomena of Hyperbranched Polymer
Solutions, Macromolecules 36(6) 2003 2085-2092.
hyperbranched polyesteramide ? Mn 1200
g/mol ? Mw/Mn 5 ? 10 -OH groups
?
Enhancing the separation factor by hyP
19
breaking the aceotrop isopropanol-water with
hyperbranched polymers not only a function of
concentration
20

• hyP can be designed by the backbone-molecules and
  by the end groups
• there is certainly an entropic effect (the
  regular structure)
• UNIFAC-FV has been used as a probe for the
  enthalpic effect
• cutting the hyP into linear, dendritic and end
  groups (experimentally known from NMR-data)
• analyzing the effect of the groups
• for polyglycerol hyP all groups contribute
• for poly-ester hyP only e-groups contribute in
  aqueous systems
• poly-amidoamine dendrimers only terminal and
  dendritic groups contribute

M.Seiler, J.Rolker, L.V.Mokrushina, H.Kautz,
H.Frey, W.Arlt, Vapor-liquid equilibria in
dendrimer and hyperbranched polymer solutions
experimental data and modeling using UNIFAC-FV,
Fluid Phase Equil. 221 2004 83-96
UNIFAC-FV seems to be an adequate tool to predict
solvent activities
?
how hyP work?
21
combinatorial part stems from UNIQUAC, takes into
account the size and is not temperature-dependent
the free volume part contains the Flory GE-model
without interaction parameter and contains an
empirical b-parameter what is set to specified
values. It is somehow additional to the
combinatorial part. the residual part contains
the group interactions scaled over the surface.
The group interaction parameters stem from
Only the group definitons were changed.
UNIFAC-FV model
22
groups in a hyperbranched polymer
23
example for the defintion of hyperstructures
24
UNIFAC-FV results for the isothermal
hyperbranched polyglycerol (PG1)ethanolwater
VLE at 363.15 K and different polymer
concentrations C1 lineardendriticterminal C2
linear terminal C3 terminal
The agreement between the experimental and
calculation results is remarkable, when all the
structural subgroups of poly-glycerol (i.e. L13,
L14, T, D) are taken into account for the
calculations of the solvent residual activities.
result of UNIFAC-FV calculation
25
Process Schemes
?
Part three
26
Having solved the thermodynamics, what is the
influence on the distillation process? 1.
Recovery is different 2. no entrainer is expected
in the top product (unless chemical
degradation) 3. no rectifying section above
entrainer feed 4. energy savings
?
27
no rectifying section
no entrainer in the distillate
flash
no distillative recovery
desorber
international patents of use of IL and hyP and on
recovery pending!
?
example of an extractive solvent recovery flash
and desorber
28
Tailor-made design of IL and hyP as selective
solvents continues in Erlangen as a cooperation
between chemists and engineers. Only as an
exception, a company will change a runnig
process. So we wait for the 1st implementation of
an extractive column with non-volatile
entrainers. Prices for IL are coming down since
Degussa/Solvent Innovation and BASF (Basionics)
entered the market.
conclusions
29
?
.