close

Вход

Забыли?

вход по аккаунту

?

The Effect of Additives and Impurities on the Partition of Ethanol into n-Decanol from Aqueous Solutions.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process., 8{5/6),pp.551-569,2000.
The Effect of Additives and Impurities on the
Partition of Ethanol into n-Decanol from
Aqueous Solutions
N. Carolan* and L.R. Weatherley**
Department of Chemical & Process Engineering, The University of
Canterbury, Christchurch, NEW ZEALAND
* NovoNordisk, Novo Alle, 2880 Bagsvaerd, DENMARK
The recovery of biological products from fermentation media by direct liquid-liquid
extraction is complicated by the presence of impurities. These may comprise either side
products of the bio-transformation or residual species from the pre-jermentation
substrates. As a corollary to the detrimental effect of impurities upon extraction,
addition of reagents, for example salts, can be used to enhance the extraction and
mitigate to some extent against these. This paper describes an experimental study of
the partitioning of ethanol from aqueous solutions into 1-decanol. Specific groups of
impurities and additives were identtjied, comprsing: electrolyte compounds (including
aciak, bases, and salts), ionic and non-ionic sulfactants, commercial de-emulsifiers,
organic acids, polysaccharide compounds (i.e. sugars and starch), protein, biomass
(including yeast and bacteria),and 1-propanol. A number of thermodynamic models
for predicting the effect of each additive on the phase distribution behaviour of ethanol
were screened and tested. The additives considered in this paper are confined to those
of relatively simple chemical structure and included 1-propanol, oleic acid, glucose,
sodium chloride, sodium hydroxide, calcium chloride, ammonium chloride, ammonium
sulphate, and hydrochloric acid. These were added in controlled amounts to the
principal extraction system components and the effect upon the partition coefficients of
ethanol determined.
~
~~
**Authorfor correspondence.
551
N. Carolan and L.R. Weatherley
Introduction
The manufacture of highly specialised products by microbial fermentation and by
using novel biotransformations is becoming well established especially in the
pharmaceutical industry and in the food and speciality chemicals sector. In the longer
term it is likely that there will be an increasing focus on the use of microbial processes
for the manufacture of lower value, low molecular weight species using renewable
feedstocks. Liquid-liquid extraction is one of several important unit operations which
may be used for the primary recovery and separation of such products. It can be
applied as part of a single line continuous process, as in a conventional recovery
operation or as part of a recycle process to facilitate simultaneous product purification
and separation of unreacted substrates. The latter is of importance for successful
extractive fermentation and biotransformation processes. These feature enhanced
single product yield and also recovery of valuable but unstable intermediate products
which may only be present at low concentration in a sequential reaction scheme. In
both types of process the overall process efficiency and productivity are enhanced in
many cases if direct extraction of the untreated fermentation liquor or reactor solutions
can be achieved.
The inhibitory effects of biomass components and impurities upon extraction
kinetics and physical phase disengagement have been documented (1-4).In the wider
context of industrial solvent extraction processes, the effects of surface active agents
upon kinetics of liquid-liquid extraction have been we11 documented (5,6). The key
components in biological media which affect extraction include the biomass solids (7),
polysaccharides which are produced during the fermentation (8), and proteins (9).
There are also other components which may enhance the fermentation including lipids,
fatty acids and buffering electrolytes. Chemical surfactants may also be added to the
fermenter feed solutions as anti-foaming agents. These may include proprietary antifoams and de-emulsifiers. Downstream, compounds associated with de-emulsification
in whole broth extraction systems include polypropylene glycol, sodium lauryl suphate
(SLS) and the surfactant cetyl pyridinium bromide (CPB) (9). Thus there is potentially
a large number of species present in the extraction mixture in addition to the main
solvent and product species.
552
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
In this paper we are focussing on ethanol as a straightforward and reasonably well
understood example of a microbial product. In particular we are concerned with its
distribution behaviour in the presence of some of the above mentioned impurities.
There has been some study of ethanol extractions referred to in the literature, much of
this is in the context of ethanol derived from fermentation as a fuel additive. The effect
of broth constituents on ethanol extraction has been examined by a number of groups.
Souissi and Thyrion (10) studied ethanol partition from a synthetic broth into hexanol
and into methyl isobutyl ketone at 35'C and found that the distribution coefficient
appeared to be higher than for the pure system.
In a later study Crabbe et al. (7) compared the partition of ethanol into l-decanol
from a nutrient broth solution, from a yeast suspension of Succhuromyces cerevisiue
(20 g/L),
and from pure water. The data for water and yeast suspensions were almost
identical indicating that yeast had no effect on partition in that range. Data for the
nutrient solution showed a higher degree of extraction compared to either the pure
mixture or the fermentation broth.
Weatherley et al. (3) also examined the effect of yeast concentration
(Succhuromycescerevisiae ) up to 3 g/L on the partition of ethanol into l-decanol, and
found that there was no obvious effect although they suggested examining higher
concentrations of yeast. This was also consistent with the behaviour observed by
Crabbe et al. (7). Earlier studies by Laughland et al. (11) also examined ethanol
extraction into 1-decanol at 18OC and compared partitioning from pure ethanol-water
mixtures, from both a filtered broth and a whole broth. Comparison of the partition
behaviour of pure ethanol solutions in the aqueous phase concentration range of 09wt%. with that of the filtered broth showed a reduction of ethanol partition in the
latter case. At higher concentrations of ethanol, the distribution coefficient for filtered
broth exceeded that for pure ethanol solutions. The partition for the whole broth system
was the lowest over the ethanol range studied.
In another investigation Varma et al. (12) studied the partition of ethanol into 1decanol in the presence of added glucose and molasses but found that the presence of
these did not affect extractability.
A detailed study was carried out by Tedder (13) where the equilibrium
concentrations of both ethanol and water in the presence of dextrose were measured
553
N. Carolan and L.R.Weatherley
using the solvent TBP in Isopar M diluent. The results suggested an increase in the
activity of the water in the aqueous phase due to the presence of dextrose. It was also
determined that the distribution coefficient increased with respect to the initial ethanol
concentration, and with respect to the dextrose concentration and the temperature.
A review of the published research on partitioning of ethanol from biological
media revealed a rather inconsistent pattern of effects attributable to impurities,
biomass and additives. In this paper we describe a different approach in which known
impurities together with a range of additives were identified and their effect upon
partitioning is systematically characterised. The study in this paper was confined to
examining the effect of single component impurities and additives.
The model extraction system used was the partition of ethanol from aqueous
solution into 1-decanol, which had the advantage of being an essentially immiscible
system.Therefore changes in solvent distribution between phases as a function of
ethanol concentration alone were assumed to be small, and thus any effects upon
ethanol partition alone could be readily determined.
The objective of this study was to determine the scope for predicting phase
partitioning behaviour and distribution of ethanol in the presence of each of a number
of additives. The additives were chosen in light of those reviewed above and in this
initial attempt at modelling it was decided to only examine the effects of those having a
relatively simple structure. The additives studied here were glucose, hydrochloric acid,
sodium chloride, oleic acid, 1-propanol, ammonium chloride, ammonium sulphate and
calcium chloride.
The Modelling Approach
The approach adopted for the modelling of the partition data in the presence of the
various impurities involved the prediction of: (a) the existence of a two phase liquidliquid system; and (b) the value of the distribution coefficient. A number of
thermodynamic predictive models were evaluated. The ability of each to predict the
existence of a two-phase liquid-liquid system was used as an initial screening criterion
before the calculation of the distribution data was undertaken. The models are listed in
Table 1 with appropriate source references.
554
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
Table 1. Thermodynamic Models Evaluated.
Model
Peng-Robinson
The Non-Random
Two-Liquid model
The UNIFAC model
The UNIFAC L-L model
UNIFAC (Dortmund Modified)
Hayden-0' Connell (NRTL)
Hayden-O'Connell eq. of state
Lyngby modified UNIFAC
The UNIQUAC model
Abbreviation
Reference
PENG-ROB
(15)
MTL
(16)
In order to evaluate these models the ASPEN PLUS (Release 9.3) process
optimisation software was used (14). This included a number of models capable of
predicting liquid-liquid equilibria, and the software contains the required parameters.
ASPEN contains option sets, so rather than just applying one model, by choosing an
option set, two models are used, one for the liquid phase and one for the vapour phase
respectively. In the systems investigated here it was assumed that vapour phase
phenomena were unimportant and therefore assumed to have little impact on the results
obtained. Details of the option sets and the models applied are also included in Table
1.
Experimental Details
The objective of the experimental work was to quantify any changes in the partition
behaviour of ethanol into 1-decanol in the presence of each of a number of additives.
The partition coefficient was determined using a fixed volume ratio of aqueous and
organic phases, 15 ml of aqueous and 15 ml of l-decanol. Precise metering of the feed
volumes was achieved by means of an automatic air displacement pipette. The initial
concentration of the aqueous phase was fixed at 100 g/L of ethanol (Sigma >99.99%)
and the decanol was pure and initially ethanol free. Each additive compound was made
up to the appropriate concentration in the aqueous phase prior to contacting with the 1decano!. The additive compounds studied are listed in Table 2 The two phases were
then shaken mechanically at a controlled temperature of 20.0"C until equilibrium was
555
N.Carolan and L.R. Weatherley
confirmed. The phases were then separated and a sample of each phase removed for
analysis.
Table 2. Additives used in this study.
Additive
Glucose
Hydrochloric acid
Oleic acid
1-propano1
Ammonium chloride
Ammonium sulphate
Calcium chloride
Sodium chloride
Concentration Range
(g J 1)
0 - 200
0-40
0-50
0 - 20
0 - 200
0 - 200
0 - 200
0 - 200
Supplier
Grade
BDH
BDH
BDH
Sigma
BDH
BDH
BDH
BDH
Analar
Analar
> 92%
HPLC
Analar
Analar
>98%
>99.9%
The additives, including lipophilic species, were introduced into the system as part
of the aqueous phase. This was designed to simulate real broth systems where any
“additives” present would be part of the aqueous broth phase. The concentration ranges
of the additives investigated were extended to beyond those which would be present in
a typical broth system. In some cases significant changes in the phase separation
behaviour was observed with a tendency to form stable emulsions.
Following
equilibration, all samples were centrifuged for five minutes to ensure complete phase
separation. Samples were taken from each phase using an automatic air displacement
pipette. The sample was taken from the centre of the phase. When sampling was from
the bottom (aqueous) phase, the plunger was initially pushed to a depth of about 75%.
As the pipette was lowered through the top phase the small amount of air was pushed
gently out of the pipette tip using the plunger. This was to ensure no organic phase
entered the pipette. When the pipette tip was withdraw from the sample bottle, the tip
was delicately wiped to remove any organic solvent from the outside of the tip, without
removing any of the aqueous phase from inside the tip. Prior to analysis by gas
chromatography each sample was diluted by a factor of 40 with HPLC grade 1-butanol
using a precision balance (k 0.0001g). Dilutions for the ethanol and organic acid
systems were also checked using a precision balance.
556
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
The volumes of the equilibrated phases were also determined and it was established
that the volume changes occurring during the partition were insignificant.
Results
Modelling
All the option sets gave solutions for this system, as shown in Table 3, and this
indicated that the required parameters for each model were available in the ASPEN
programme. A successful solution was defined when the option set predicted phase
separation.
T d l e 3. Suitabiliry of option sets for the 100 g L ethanol system
Option Set
PENG-ROBINSON
WILSON
NRTL
NRTL-HOC
W A C
UNIF-DMD
UNIF-LL
UNIF-HOC
UNIOUAC
Successful Result?
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Comments
No phase separation
No phase separation
Phase separation
Phase separation
Phase separation
Phase separation
Phase separation
Phase separation
Phase separation
Those option sets which did not predict phase separation were disregarded for further
investigations with the salt system. The distribution coefficient values for option sets
which predicted phase separation in the ethanol system in the absence of salt are shown
in Table 4.
It is evident that the accuracy of the option sets in the ethanol system is poor. This
is to be expected because of the high degree of non-ideality in the system which
includes an aqueous phase and a polar solute. The most successful option sets were
NRTL, NRTL-HOC,UNF-DMD, and UNIQUAC. All option sets predicted a
significant change in the ratio of the phase volumes (V, 1 V,,)
which does not occur in
557
N. Carolan and L.R. Weatherley
Table 4. Option sets applied to the 100 g/L ethanol system.
Option set
D
Vaq I Vorg
(g/L
BASIS)
Experimental
NRTL
NRTL-HOC
UNIFAC
UNIF-DMD
UNIF-LBY
UMF-LL
UMF-HOC
UNIQUAC
0.42~1
0.514
0.507
0.893
0.462
0.712
0.733
0.893
0.508
0.853
0.854
0.803
0.887
0.832
0.820
0.803
0.830
the experimental system. The most successful option set was the Dortmund UNIFAC
set (UNIF-DMD) which estimates a distribution coefficient of 0.46 compared to the
experimental value of 0.42. However, although this option set shows the smallest
change in phase volumes, the predicted change is a significant overestimation of what
actually occurs.
To find an option set which might be able to indicate the possible effect of
additives in the ethanol system a check was made to establish which option set would
provide solutions for a wide range of additives. This was achieved by discovering
which option set would yield successful solutions (i.e.predict phase separation) when
sodium chloride was present in the system (see Table 5).
Only the NRTL and NRTL-HOC option sets yielded a successful solution. These
option sets are very similar, and so it was decided to use the more commonly used
NRTL option set for further study.
In the ASPEN chemical data-bank some of the compounds used in this research
were not available and so modelling of all the systems for which experimental data
were collected was not possible. However, similar compounds were investigated
wherever possible. Iron (111) chloride is an example, because although it was not used
in the experimental studies, the modelled effect of the trivalent metal ion on the
distribution coefficient was deemed to give some indication of the ability of the NRTL
option set to predict salt effects in the presence of tri-valent species. In running the
ASPEN simulations, all additives were assumed to be added initially in the aqueous
phase thus reflecting the procedure adopted in the experiments.
558
Partition of Ethanol into n-Decarwlfrom Aqueous Solutions
Table 5. Suitability of option sets for the 100 giL. ethanol system with 100 giL. sodium
chloride present.
Option set
W A C
Successful result?
X
UNIF-DMD
X
UNIQUAC
x
WILSON
x
NRTL
d
d
NRTL-HOC
Comments
Missing parameters
no solution
Missing parameters
no solution
Missing parameters
no solution
No phase separation
Phase separation
Phase separation
Experimental Results
The distribution coefficients for each system are presented in Figures 1 to 9. In each
case the distribution coefficient was calculated as the ratio of the organic phase
concentration of ethanol to the aqueous phase concentration. The values of distribution
coefficient are plotted against the additive concentration.
Discussion
The data pertaining to the effect of each additive upon ethanol distribution are now
discussed below.
I -propano1
The effect of the additive 1-propanol was to increase the distribution coefficient of
ethanol from 0.40 to 0.45 over a concentration range 0 to 20 g/L, as shown in Figure 1.
1-propanol is known to extract to a higher degree than ethanol into 1-decanol (22, 23)
since it is a more lipophilic molecule than ethanol due to the additional CH3 group. As
the 1-propanol concentration in 1-decanol increases, the organic phase becomes more
polar and thus would tend to solvate further 1-propanol. This increase in polarity is
caused not only by the higher quantity of 1-propanol in the organic phase but is also
due to co-extraction of water with the 1-propanol. This would explain the tendency of
the distribution coefficient of low molecular weight aliphatic alcohols to increase with
559
N. Carolan and L.R.Weatherley
".-
'
0
10
5
20
15
I - R O P ~ O I concentration ( g 1.' )
Figure 1. The effect of 1-propanol on distribution coeficient.
-
0.43
C
0.38 7
0
10
20
40
30
50
Oleic acid concentration ( g I-')
Figure 2. The effectof oleic acid on distribution coeficient.
0
50
100
150
Glucose concentration(g I-')
Figure 3. The effectof glucose on distribution coeficient.
200
Partition of Ethanol into n-DecanolfromAqueous Solutions
increasing solute concentration (23). This effect is clearly evident here for ethanol
extraction into 1-decanol in the presence of propanol.
Oleic Acid
The partition of ethanol decreased significantly in the presence of oleic acid (see
Figure 2.). Oleic acid has a low aqueous solubility and a high solubility in 1-decanol
and thus will compete with ethanol for 1-decanol molecules in the organic phase thus
reducing the distribution coefficient for ethanol. An increase in the non-polar nature of
the organic phase would reduce the distribution coefficient of ethanol.
The oleic acid may also be capable of extracting ethanol and thus act as a cosolvent. Hence the overall distribution coefficient would also depend on the solubility
of ethanol in oleic acid. Oleic acid has a fairly low capacity for ethanol, a distribution
coefficient of approximately 0.2 having been reported (24). If oleic acid and 1-decanol
were used as a solvent mixture to extract ethanol, a lower distribution coefficient for
ethanol would be expected than if 1-decanol alone had been used.
Glucose
Glucose clearly causes a large increase in the distribution coefficient of ethanol, and
increases from 0.40 to 0.53 are observed over the concentration range 0-200 glL (see
Figure 3). The increase in the distribution coefficient appears higher over the
concentration range 0 to 100 g/Land then tails off slightly. In a study by Tedder (13)
the presence of dextrose was shown to cause a large increase in the distribution
coefficient for ethanol partition into TBP in Isopar M diluent. This effect suggests an
increase in the activity of the water in the aqueous phase due to the presence of
glucose. The structure of glucose suggests a number of molecular groups capable of a
strong interaction with both water and ethanol molecules.
The increase in the
distribution coefficient may be due to a strong interaction of glucose with water
molecules, thus rendering them less available for ethanol soivation in the aqueous
phase.
561
N.Carolan and L R. Weatherley
1.10
_I
0
50
100
150
200
Sodium chloride concentration (g I-')
Figure 4. The effect of sodium chloride on distribution coefficient.
-
0.57
-
0
e
0.39 7
0
10
20
30
40
Sodium hydroxide concentration (g 1.')
Figure 5. The effect of sodium hydroxide on distribution coefficient.
"..XI
'
0
50
100
150
Calcium chloride concentration (g 13
Figure 6. The effect of calcium chloride on distribution coeflcient.
562
200
Partition of Ethanol into n-Decanolfiom Aqueous Solutions
Sodium Chloride and Sodium Hydroxide
The addition of sodium chloride increased the distribution coefficient of ethanol from
0.40 to over 0.95 over the concentration range 0 to 200 g/L, as shown in Figure 4. This
strong increase is most likely due to salting out effects. The increase in the distribution
coefficient
appears to be approximately linear with respect to sodium chloride
concentration.
Over the concentration range 0 to 40 g/L sodium chloride, the
distribution coefficient increases from 0.40 to just below 0.50. Other experiments (25)
showed that this was a smaller increase than was observed in the case of sodium
hydroxide addition (see Figure 5). The salting out effect was quantified using the Long
and McDevit relationship (26) in terms of the salting out coefficient (ks). The ks values
for sodium hydroxide (0.256) A d sodium chloride (0.195) indicate that on a molar
concentration scale sodium chloride will salt out ethanol to a lesser degree than sodium
hydroxide. The salting out coefficients do not necessarily reflect the true situation in a
four componen two-phase system such as we are evaluating here. However it would be
reasonable to expect the relative affinity of the solute for the aqueous phase to be
influenced in the same way as in the single-phase three component system.
Another parameter which may be useful for quantifying the relative magnitude of
salting out effects exhibited by different electrolytes is the quotient of free energy of
hydration and ionic radius ( -AGohy&),as given in Table 6. The -AGohY&values (27)
for O H (3301 Id/mol nm) and for C1' (1917 Id/mol nm), suggest that sodium hydroxide
would be expected to be more effective at salting out ethanol compared with sodium
chloride. The greater tendency of the hydroxide ion to hydrate is reflected in the AG0hydrvalue and is consistent with a higher salting-out effect with respect to ethanol
distribution.
Calcium Chloride
The results in Figure 6 show that the addition of calcium chloride at concentrations
inthe range 0-200 g L increased the distribution coefficient of ethanol from 0.40 to
0.75. This large increase can be explained by the presence of the divalent calcium ion,
which has a high charge density. However the overall effect of calcium chloride was
higher than expected when compared with that of ammonium sulphate, as discussed
563
N. Carolan and L.R. Weatherley
-
1.10
0
g 0.90
.
I
0
8
E
0
.-
-
0.70
a
0.50
v)
a 0.30
0
50
100
200
150
Ammonium chloride concentration ( g 1.')
Figure 7. The effect of ammonium chloride on distribution coeficient.
1.10 -
u
E 0.90-8 0.70 --
13
0
50
100
150
200
~mmonium
sulphate concentration (g 1.')
Figure 8. The effect of ammonium sulphate on distribution coeficient.
i
I
20
30
7
P
.s 0.40
.
d
--
n
0.39 1
0
10
Hydrochloric acid concentration ( g 1.')
Figure 9. The effect of hydrochloric acid on distribution coeficient
564
40
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
below. The Ca2' ion has a -AGOhy& value of 15150 kJ/mol nm which compares with
values of 3670 kJ/mol nm for Na' and 1973 kJ/mol nm for Nb'. This is consistent
with the strong ability of the CaZ+ion to salt out non-electrolytes compared with
univalent ions such as Na' and NJ&
Table 6. Ionic Addirives - Ionic Radii and Standard Molar Free Energies of Hydration
Ion
Na+
K+
N&+
Ca2+
OH
Cl-
so?-
r
- AG'hyd
- AGohyd/r
(nm>
kJ/mol
kJ/mol nm
0.102
0.138
0.148
0.100
0.133
0.181
0.230
375
307
292
1515
439
347
3670
2225
1973
15150
3301
1917
4739
1090
Ammonium Chloride and Ammonium Sulphate
The addition of ammonium chloride increases the distribution coefficient of ethanol to
a lesser extent than sodium chloride, as shown in Figure 7. On addition of ammonium
sulphate, the distribution coefficient of ethanol increases from 0.40 to 0.93 over a salt
concentration range of 0 to 200 g/L(see Figure 8). The SO? ion is more effective at
salting out than the C1' ion with a -AGOh,& value of 4739 kJ/mol nm compared to 1917
kJ/mol nm for Cf. As expected, ammonium sulphate has a significantly higher salting
out capacity than ammonium chloride. However, the higher molecular weight of
ammonium sulphate (132.13 glmol) shows that for a given mass concentration, the
number of ions of ammonium sulphate is considerably less than for ammonium
chloride. When compared to sodium chloride this is also true. As a result of its large
molecular weight, ammonium sulphate is slightly less effective than sodium chloride
on a mass concentration scale despite the higher charge of the anion in ammonium
sulphate and the presence of three ions per molecule.
565
N.Carolan and L.R. Weatherley
Hydrochloric Acid
The effects of a number of other electrolytes on the partition of ethanol were also
investigated. Results presented in Figure 9 show the effect of hydrochloric acid
addition, yielding slight increases in the distribution coefficient of ethanol. The likely
mechanism here is the “salting out” of ethanol due to introduction of hydrated ions to
the aqueous phase.
Comparison with the Model
The results listed in Table 7 show that the NRTL option set is not even qualitative in its
prediction of the effect of the salts. Salts have little effect on partition, but show strong
effects on the phase volumes. In the experimental studies the opposite is true.
Addition of oleic acid shows a small increase in partition, and decreases the phase
volumes slightly. The experimental results showed a negligible effect on partition, but
it would be expected to reduce the aqueous phase volume slightly because of its high
solubility in the organic phase.
Table 7. NRTL Option set for the 100 g/Lethanol system.
ADDITIVE
CONCENTRATION
D
Vaq I Vorg
(g/L
basis)
Experimental
None
NaCl
CaCI2
FeCI3
Oleic acid
1-Propano1
Sucrose
566
0.42
0.5 14
50
100
300
50
100
200
50
100
200
20
50
20
50
50
0.517
0.524
0.583
0.537
0.559
0.602
0.529
0.545
0.579
0.524
0.539
0.53 1
0.557
0.542
Z1
0.853
0.852
0.846
0.795
0.644
0.442
0.079
0.836
0.819
0.778
0.843
0.825
0.819
0.768
0.822
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
The predictions for I-propanol effects are qualitatively correct, with a small but
significant increase in partition and a decrease in the aqueous phase volume. The
reduction in phase volume is expected because 1-propanol would tend to increase the
mutual solubility of the phases, but would be smaller in the experimental system due to
the low mutual solubility of 1-decanol and water. Finally, sucrose increases the
partition.
Sucrose was not investigated in the experimental studies but it is a similar molecule
to glucose which does increase the partition of ethanol in the experimental system. So
overall, the NRTL option set has limited application in predicting the effects of
additives in the ethanoYl -decanoYwater system.
Conclusions
The presence of additives and impurities clearly exert an important effect upon the
distribution coefficient observed for the partition of ethanol in n-decanol. Oleic acid
was the only solvating organic liquid additive studied here and its presence exerted a
negative effect upon the partition. It was concluded that the most likely explanation for
this was competition of oleic acid for polar sites in n-decanol thus reducing availability
of extraction capacity for ethanol molecules.
The addition of glucose increased the distribution coefficient for ethanol and this
increased with respect to glucose concentration. This was consistent with other
published partition data showing a positive effect of sugar addition.
Phase separation behaviour was predicted successfully using all but two of the
thermodynamic models tested. The Peng-Robinson model and the Wilson model did
not predict phase separation.
The models which predicted phase separation gave reasonable prediction of the
distribution coefficient for the distribution of ethanol between water and n-decanol, and
in all cases within an order of magnitude.
All the models tested predicted a change in the phase ratio upon equilibration
which was not observed experimentally.
In the presence of salt, only the NTRL and the NTlU-HOC models predicted phase
separation.
567
N.Carolan and L.R. Weatherley
The addition of sodium chloride and sodium hydroxide resulted in salting-out
effects with respect to ethanol distribution, and the observations were consistent with
the predictions of the Long and McDevit relationship for salting-out modifications to
partition.
Oleic acid was the only solvating organic additive studied. The observed negative
influence upon ethanol partition would be readily explained by competition for the
organic by the oleic acid effectively reducing solvent availability for ethanol.
Glucose addition increases the distribution coefficient and this reflected literature
data showing a very similar effect of sucrose addition upon ethanol partition.
The quotient of free energy of hydration and ionic radius proved to be a useful
parameter in determining the relative effects of salting out agents upon the distribution
coefficient.
Acknowledgment
The financial support of the UK Biotechnology and Biological Sciences Research
Council is gratefully acknowledged.
Nomenclature
D
-AGOhyd
ks
r
v,
Vorg
Distribtion coefficient
Standard free energy of hydration
Salting out coefficient,
ionic radius
Aqueous phase volume
Aqueous phase volume
(kJ/mol)
(nm)
(L)
(L)
References
1.
2.
3.
4.
5.
6.
568
Schugerl, K. “Solvent Extraction in Biotechnology. Recovery of Primary and Secondary Metabolites”,
Springer-Verlag. (1994).
Akzo Nobel, “Fermentation Production, Exaaction Demulsifiers”, Company Brochure, (1988)
Weatherley. L.R., Allan, G., Goodwin, J.A.S. “Interfacial Mass Transfer and Partition Studies During
Whole Broth Extraction”. Proceedings of the International Solvent Extraction Conference, (1993),
V01.1, 1003-1009.
Stuckey, D.C. “Solvent Extraction in Biotechnology”, Proceedings of the International Solvent
Extraction Conference, (1996). Vof. 1, 25-34.
Liang T.B. & Slater M.J. “Liquid-liquid extarction drop formation: mass transfer and the influence of
surfactant” Chem.Eng.Sci. ,(1990). vol. 45 (1). 97 105.
Agble. D., Mendes-Tatsis, M.A. ‘‘Mass Transfer with Interfacial Convection and Added Surfactants”,
Proceedings of the International Solvent Extraction Conference, (1996). Vol. 1,267-272.
-
Partition of Ethanol into n-Decanolfrom Aqueous Solutions
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Crabbe. P.G., Tse, C.W., M u m , P.A.. “Effect of Microorganisms on Rate of Liquid Extraction of
Ethanol from Fermentation Broths”. Biotechnology and Bioengineering, (1986). Vo1.28, No.7, 939943.
Weatherley, L.R., Doherty-Spiers, E.A., Goodwin, J.A.S., Slaughter, J.C.. “A Study of Biomass
Morphology on Partition in the System Propanol!Watert/n-Decanol in the Presence of Aureobasidium
pulluluns”. in Separations for Biotechnology Volume 3, ed. D.L. Pyle, SCI Publishing, (1994), 539545.
Lennie, S., Halling, P.J., Bell, G. “Causes of Emulsion Formation During Solvent-Extraction of
Fermentation Broths and its Reduction by Surfactants” Biotechnology and Bioengineering, (1990).
V01.35, NO.9.948-950.
Souissi. A., Thyrion, F.C. “Liquid-Liquid Extraction of Ethanol from Aqueous Solutions”, Proceedings
of the 2nd World Congress of Chemical Engineering, (1981), Vo1.4.443-446.
Laughland, G.J., Millar, M.K., Weatherley, L.R., “‘ElectrostaticallyEnhanced Recovery of Ethanol
from Fermemtation Liquor by Solvent Extraction”, IChemE Symposium Series No.103, Extraction
1987, (1987). EFCE Publication Series 347.263-278.
Varma, R., Sawant, U.D., Karanth, N.G. “Estimation of Ethanol in Fermentation Broth by SolventExtraction and Gas-Chromatography”, Enzyme and Microbial Technology, (1984). Vo1.6, N0.5, 233235.
Tedder,D.W. “Fuel Grade Ethanol Recovery by Solvent Extraction: Final Report”, Progress Report,
Georga Institute of Technology, Atlanta, August. (1984).
ASPEN Technology Inc., “ASPEN operation manual”, Cambridge, England.
Peng, D.Y., Robinson, D.B. “A New Two-Constant Equation of State”, Industrial Engineering. and
Chemistry Fundamentals, (1976), Vo1.15, pp59-64.
Renon, H., Prausnitr J.M. “Local Compositions in Thermodynamic Excess Functions for Liquid
Mixtures”. AIChE Journal, (1968). Vo1.14, No.1. ,pp135-144.
Hansen. H.K., Rasmussen, P., Fredenslund, A,, Schiller, M., Ghemling, J. “Vapor-kquid Equilibria by
UNIFAC Group Contribution. 5 Revision and Extension”, Industrial Engineering. and Chemistry
R-h,
(1991), VO1.30.2352-2355.
Gmehling, M., Li, J., Schiller, M. “A Modified UNIFAC Model. 2. Resent Parameter Matrix and
Results for Different Thermodynamic properties”, Industrial Engineering. and Chemistry Research,
(1993). V01.32,178-193.
Hayden, J.G., O’Connell, J.P. “A Generalized Method for Predicting Second Virial Coefficients”,
Industrial Engineering and Chemistry Process Design and Development, (1974). Vol. 14, No.3, pp209216.
Lanen, B., Rasmussen, P., Fredenslund, A. “A Modified UNIFAC Group-Contribution Model for
Prediction of Phase Equilibria and Heats of Mixing”, lndustrial Engineering. and Chemistry Research,
(1987). V01.26, 2274-2286.
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 Journal, (1975), V01.21.
116-121.
Kertes, AS., King, C.J. “Extraction Chemistry of Fermentation Product Carboxylic-Acids”,
Biotechnology and Bioengineering, (1986). Vo1.28, N0.2.269-282.
Kertes, AS., King, C.J. “Extraction Chemistry of Low-Molecular Weight Aliphatic-Alcohols”,
Chemical Reviews, (1987). Vo1.87, No.4,687-710.
Munson. C.L., ffing, C.J. “Factors Influencing Solvent Selection for Extraction of Ethanol from
Aqueous-Solutions”. Industrial and Engineering Chemistry Process Design and Development, (1984),
V01.23, No.1, 109-115.
Carolan, N. “Partition studies in whole broth extraction”, PhD Thesis, (1997), The Queens University
of Belfast, United Kingdom.
Coetzee, J.F., Ritchie, C.D. “Solute-Solvent Interactions”, Marcel Dekker, (1976).
Marcus, Y., “Ion solvation”, Wiley, (1985).
Received: 16 August 1999; Accepted afer revision: 15 March 2000.
569
Документ
Категория
Без категории
Просмотров
1
Размер файла
789 Кб
Теги
partition, solutions, effect, ethanol, aqueous, impurities, decanol, additive
1/--страниц
Пожаловаться на содержимое документа