Fluid Phase Equilibria 236 (2005) 125–135
Oligomer distribution in concentrated lactic acid solutions
Dung T. Vu, Aspi K. Kolah, Navinchandra S. Asthana, Lars Peereboom,
Carl T. Lira ∗ , Dennis J. Miller
Chemical Engineering and Materials Science, Michigan State University 2527 Engineering Building, East Lansing (USA), 48824-1226
Received 21 April 2005; received in revised form 1 June 2005; accepted 3 June 2005
Available online 10 August 2005
Abstract
Lactic acid (2-hydroxypropanoic acid) is a significant platform chemical for the biorenewable economy. Concentrated aqueous solutions of
lactic acid (>30 wt.%) contain a distribution of oligomers that arise via intermolecular esterification. As a result, the titratable acidity changes
non-linearly with acid concentration. In this work, the oligomer distribution of lactic acid is characterized using GC, GC/MS, and HPLC to
extend existing literature data, and titratable acidity is measured via titration with NaOH. A thermodynamic model with a single parameter is
proposed that accurately represents oligomer distribution and titratable acidity over the full range of lactic acid concentrations.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Lactic acid; Oligomerization; Chemical theory; Esterification; Alpha-hydroxy acid; 2-Hydroxypropionic acid
1. Introduction
In recent years, there is increasing emphasis on using
biorenewable materials as substitutes for petroleum-based
feedstocks. This paradigm shift is attributable to rising crude
oil prices and the increasing desire to reduce dependence
on petroleum. A major building block for the biorenewable
economy is lactic acid (2-hydroxypropionic acid), an ␣hydroxy acid containing both a hydroxyl and carboxylic acid
functional group. For an excellent review on lactic acid the
reader is referred to Holten [1]. Lactic acid was first isolated
by the Swedish scientist Scheele in 1780 [2], and first
produced commercially in 1881 [3]. Applications for lactic
acid are found in the food (additive and preservative), pharmaceutical, cosmetic, textile, and leather industries. Lactic
acid can be formed either via fermentation of carbohydrate
monomers or via a chemical route, but since about 1990 only
the fermentation route is practiced commercially. The recent
completion of the NatureWorks lactic acid facility for polylactic acid production, with an annual capacity of 140,000
metric tonnes of polylactic acid (PLA) [4], has greatly
∗
Corresponding author. Tel.: +1 517 355 9731; fax: +1 517 432 1105.
E-mail address: lira@egr.msu.edu (C.T. Lira).
0378-3812/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.fluid.2005.06.021
enhanced the stature of lactic acid as a key biorenewable
platform.
Polylactic acid [5] is a versatile thermoplastic polymer
that has useful mechanical properties including high strength
and high modulus. Applications of PLA include household
commodity products, polymers used in food contact,
biomedical materials like surgical sutures, absorbable
bone plates for internal bone fixation, artificial skin, tissue
scaffolds, and controlled release drugs. PLA is one of
the few polymers whose structure and properties can
be modified by polymerizing a controlled composition
of the l- and d-isomers to give high molecular weight
amorphous or crystalline polymers. PLA has a degradation
time of 6 months to 2 years in the environment. For
more details on PLA the reader is referred to Garlotta
[6].
Esters of lactic acid, formed via combination with alcohols like methanol and ethanol, are finding increased use
as environmentally benign solvents. Lactic acid esters are
biodegradable, non-toxic, and have excellent solvent properties, which make them attractive candidates to replace halogenated solvents for a wide spectrum of uses. Esterification
of lactic acid with alcohol can also be used as a highly efficient method for purification of lactic acid from fermentation
126
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
broths, especially when lactic acid is desired in concentrated
solutions.
It has been observed experimentally that dilute (<20 wt.%)
lactic acid solutions contain only lactic acid monomer (LA1 )
[7], an observation that has been verified in this paper.
However, many processes involving lactic acid, including
polymerization and esterification, require concentrated lactic
acid solutions, and lactic acid in these solutions undergoes
intermolecular self-esterification to form higher oligomers.
This oligomerization occurs to an increasing degree at
high acid concentration, low water concentration, and high
temperature.
In oligomerization, two molecules of lactic acid first react
to form a linear dimer, commonly called lactoyllactic acid
(LA2 ), along with a mole of water.
the quantities used to describe the concentration of lactic
acid and its oligomers in solution.
1.1.1. Equivalent monomer lactic acid
In the literature, it has been found convenient to
express the concentration of lactic acid oligomers as a
percent of equivalent monomer lactic acid on a waterfree basis. We abbreviate such a description with the
acronym %EMLAj . To illustrate the concept, consider
a solution consisting of 50 mol water, 9.20 mol LA1 ,
0.343 mol LA2 , and 0.0128 mol LA3 . Upon hydrolysis of the
oligomers, 9.20 + 2 × 0.343 + 3 × 0.0128 = 9.924 mol lactic
acid monomer would be present. The amount of water
present would be 50 − 0.343 − 2 × 0.0128 = 49.63 mol H2 O.
The lactic acid in the original solution is reported as
(1)
Lactic acid also forms a cyclic dimer noted as lactide,
but this compound is known to be unstable in water [1] and
thus is not a concern in this work. Lactoyllactic acid (LA2 )
can further esterify with LA1 to form the trimer lactoyllactoyllactic acid (LA3 ); this process can further continue to
give higher chain intermolecular polyesters LA4 , LA5 and
so on.
9.20/9.924 = 92.7% EMLA LA1 , 2 × 0.343/9.924 = 6.9%
EMLA LA2 , and 3 × 0.0128/9.924 = 0.38% EMLA LA3 .
Introducing the molecular weight of water and oligomers,
the solution has a total mass of 50 × 18.02 + 9.20 ×
90.08 + 0.343 × 162.14 + 0.0128 × 234.21 = 1788.3 g.
(2)
The inherent tendency of aqueous lactic acid to form
intermolecular esters in solution poses a formidable obstacle
in the modeling of its liquid-phase behavior and vapor-liquid
phase equilibria. For design of reaction and separation
processes involving concentrated lactic acid solutions, a
model to predict thermodynamic properties of these complex
chemically reactive mixtures is an indispensable tool. This
paper presents such a model that requires only one parameter
to adequately represent lactic acid solution behavior over
the full range of concentration.
1.1. Definition of concentrations
Experimental work on quantifying concentrations of
lactic acid oligomers in aqueous solution has been previously reported by Montgomery [7], Ueda and Terashima
[8], and Watson [9], but the methods used in reporting these
concentrations and the definitions of concentrations are not
always clearly presented. Therefore, we clearly define here
1.1.2. Superficial weight percent
The superficial weight percent of lactic acid is
expressed as the weight of total monomer with the corresponding water of hydrolysis divided by total solution weight. For the example above, the superficial wt.%
is (9.924 mol LA × 90.08/1788.3 = 0.500) 50.0 wt.% lactic
acid, and (49.63 × 18.02/1788.3 = 0.500) 50.0 wt.% water.
When lactic acid is purchased, the concentrations expressed
in wt.% should be interpreted as superficial wt.%. In this
manuscript, we explicitly label such concentrations superficial wt.% to avoid confusion.
When solutions are very concentrated, the superficial concentration of lactic acid can exceed 100 wt.%. The concept
of 125 superficial wt.% lactic acid arises from the fact that
100 g of a polymer (C3 H4 O2 )n upon hydrolysis gives rise
to 100 × 90.08/72.06 = 125 g of lactic acid, where 90.08 is
the molecular weight of lactic acid monomer, and 72.06 is
the molecular weight of the ester repeat unit in the polymer.
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
When an aqueous solution has a lactic acid content exceeding
100 superficial wt.%, the water of esterification (oligomerization) has been removed from the solution, and the solution is
thus characterized by a negative superficial wt.% of water.
1.1.3. True weight percent
True weight percent utilizes the mass of a particular sample and the total mass of the individual species within the
solution. Using the same example again, the true wt.% values
are 46.3 true wt.% LA1 (9.20 × 90.08/1788.3 = 0.463), 3.1
true wt.% LA2 (0.343 × 162.14/1788.3 = 0.031), 0.17 true
wt.% LA3 (0.0128 × 234.21/1788.3 = 0.0017), and 50.4 true
wt.% H2 O (50 × 18.02/1788.3 = 0.504).
2. Experimental
2.1. Chemicals
Analytical grade aqueous lactic acid solutions were used
in experiments: 85 superficial wt.% was purchased from J.T.
Baker, Inc. and 50 superficial wt.% was purchased from
Purac, Inc. HPLC grade water was purchased from J.T. Baker,
Inc. HPLC grade acetonitrile was purchased from EMD
Chemicals. An aqueous solution of 85 wt.% phosphoric acid
was purchased from J. T. Baker, Inc.
127
For solutions containing more than 20 but less than 85
superficial wt.% lactic acid, the total free acidity of the
solution was determined from titration with standard 0.1N
NaOH In solutions above 85 superficial wt.%, titration with
0.1N NaOH occurred with too little base to accurate determine the endpoint. More reproducible results were found
when using 0.01N NaOH. In addition, titrating the lactic solution in ice yielded more reproducible results due
to decreased probability of hydrolysis. Ester bonds present
in oligomers are susceptible to hydrolysis in the presence
of aqueous NaOH at room temperature. This could lead
to inconsistencies in determination of total acid content
by titration, therefore the solution was titrated in ice to
minimize hydrolysis. After titration of free acidity, excess
NaOH was added and the solution was heated to about
80 ◦ C to hydrolyze the oligomers to monomeric sodium lactate. Hydrolysis was carried out for two hours for solutions
below 100 superficial wt.% and for four hours for solutions above 100 superficial wt.%. The quantity of unreacted
NaOH was determined by back titration of the resultant
solution with standardized 0.1 N H2 SO4 solution (SigmaAldrich). For concentrations where only monomer and dimer
exist, the quantity of LA1 in solution was calculated by
the difference between NaOH consumed for neutralization of total acid and the quantity of NaOH consumed
for the hydrolysis of ester linkage present in oligomers
[11,12].
2.2. Preparation of oligomer solutions
Solutions of lactic acid below 50 superficial wt.% were
prepared by adding water to 50 superficial wt.% lactic
acid, whereas solutions between 50 superficial wt.% and
85 superficial wt.% were prepared by mixing the 50% and
85% solutions. After mixing, the solutions were heated at
80 ◦ C for 1 week to increase the rate of formation of various
oligomers of lactic acid. To concentrate lactic acid above
85 wt.%, water was removed from 85 wt.% lactic acid at
45 mmHg using a vacuum distillation apparatus. At that
pressure, the boiling point temperature started at 30 ◦ C for
90 superficial wt.% solution and rose to 135 ◦ C for solutions
of 120 superficial wt.%. Following evaporation, the solutions
were equilibrated by refluxing at 100 ◦ C for 30 h.
2.3. Analytical methods
The composition of lactic acid and its oligomers in solution was characterized using a combination of three analytical
techniques.
2.3.1. Titration
The composition of dilute solutions containing less than
20 superficial wt.% lactic acid contains >98%EMLA LA1 and
water [1]. Lactic acid solution containing less than 10 superificial wt.% of lactic acid contains 99.6% EMLA LA1 [1,10],
and direct titration with standardized 0.1N NaOH (SigmaAldrich) gave an accurate analysis of LA1 in solution.
2.3.2. GC analysis and GC/MS analysis
Water concentrations in lactic acid standard solutions
were verified using a Varian 3600 gas chromatograph (GC)
equipped with a thermal conductivity detector (TCD). The
GC column was 3.25 mm OD × 4 m long and was packed
with 80/100 mesh Porapak-Q. The oven temperature was
held constant at 413 K for 2 min, ramped at 20 ◦ C/min
to 493 K, and held at 493 K for 6 min. The injector temperature was maintained at 493 K and the TCD block
temperature was held at 523 K. Helium was used as the
carrier gas. HPLC grade acetonitrile was used as an internal
standard.
Qualitative analysis of LA1 and its higher oligomers LA2 ,
LA3 LA4 , etc. by GC–MS was carried out on a JEOL
AX-505H double-focusing mass spectrometer coupled to a
Hewlett-Packard 5890J gas chromatograph via a heated interface. GC separation employed a J&W DB-23 fused-silica
capillary column (30 m length × 0.25 m ID. with a 0.25 m
film coating). Splitless injection was used. Helium gas flow
was maintained at 1 mL/min. The GC temperature program
was initiated at 323 K and was ramped at 10 ◦ C/min to 533 K.
MS conditions were as follows: interface temperature 523 K,
ion source temperature 523 K, electron energy 70 eV, and
scan frequency was l Hz over the m/z range of 45–750. Prior
to its injection for analysis by GC–MS, LA1 , LA2 , LA3 ,
and LA4 were derivatized with TMS {Propanoic acid, 2[(trimethylsilyl)oxy]-trimethyl silyl ester} to enhance their
volatility.
128
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
2.3.3. HPLC analysis
The concentration of LA1 and oligomers in concentrated
lactic acid solutions were quantified using a Hewlett Packard
1090 Liquid Chromatograph equipped with an auto sampler,
gradient flow pump, oven and a Hitachi-L400H UV detector
set at 210 nm. Lactic acid samples below 85 superficial wt.%
were analyzed using a mobile phase of water + acetonitrile in
gradient concentration at a flow rate 1 mL/min on a Novapak
C18 column (3.9 mm × 150 mm). Both water and acetonitrile
were acidified using 2 ml of 85% (w/v) phosphoric acid in 1 L
of solvent. The water was analyzed to be pH 1.3. The column
oven temperature was maintained at 40 ◦ C. Beginning with
a mobile phase of 100% acidified water, the acetonitrile
concentration was ramped linearly to 60 vol.% from zero
to 20 min and then ramped linearly up to 90% from 20 min
to 25 min. The mobile phase composition was maintained
constant at 90% to 28 min and then returned to 100%
water.
For analysis of solution concentrations above 85 superficial wt.% lactic acid, the total flow rate and column temperature were maintained as above, but the gradient was modified.
The mobile phase was ramped linearly from 10% to 100%
acetonitrile from 0 to 25 min. Acetonitrile concentration of
mobile phase was brought back to 10% at 35 min.
2.3.3.1. Response factor for LA1 . Dilute solutions of lactic
acid (<20 superficial wt.%) contain >98% EMLA LA1 ; their
concentrations can be accurately determined by titration as
described in Section 2.3.1. To prepare a standard containing
only LA1 , a dilute solution containing 7–8 superficial wt.%
total lactic acid in water was prepared and heated for 6 h in
presence of Amberlyst-15 cation exchange resin to facilitate
hydrolysis of any LA2 or higher oligomers present. Titration
of this solution with 0.1N NaOH showed a value of 7.3 true
wt.% LA1 . This solution was used to create HPLC calibration
standards for LA1 that spanned the range of LA1 concentrations (0.1–1 true wt.%) used in HPLC analysis. A linear UV
response was observed from the calibration curve obtained
by sample dilution. The response factor for LA1 obtained
from this calibration was used for quantitative determination
of LA1 in concentrated lactic acid solutions.
2.3.3.2. Response factor for LA2 . A 50 superficial wt.%
lactic acid solution, containing LA1 and LA2 , was
titrated/hydrolyzed/back-titrated with standardized 0.1 N
NaOH solution as described in Section 2.3.1. By this method
the composition of LA1 and LA2 were quantified as 46 and
3 true wt.%, respectively. HPLC analysis was performed on
the sample and LA1 was quantified using the response factor
from calibration described in Section 2.3.3.1. GC analysis of
the sample showed the presence of 51 true wt.% water, and
closed the material balance. This standardized solution was
diluted in water to provide a series of calibration standards
that spanned the pertinent range of true wt.% of LA1 (0.1
to 1 wt.% by appropriate dilution with water) and LA2 . A
linear UV response with concentration was observed for
LA2 following prompt analysis. The response factor from
this calibration curve for LA2 was used for quantitative
determination of the superficial LA2 concentration in lactic
acid solutions. The ratio of response factors for superficial wt.% was found to be LA2 /LA1 = 1.43 in all HPLC
analyses.
2.3.3.3. Response factors for LA3 and LA4 . In a solution
with approximately 93 superficial wt.% aqueous lactic acid
solution, the linear oligomers LA3 and LA4 are observed
in significant quantities in addition to LA2 . HPLC analyses of the solution showed compositions of 58 and 22 true
wt.% for LA1 and LA2 , respectively, with the remaining lactic acid in the form of higher oligomers. GC analysis of the
solution showed the presence of 12 true wt.% water. The
presence of lactic acid oligomers up to LA4 was also verified
by GC–MS analysis. The assignment of response factors for
higher oligomers was based on the following premises: (1)
the difference in successively higher oligomers of lactic acid
is the presence of an additional ester group; (2) the UV detector response is related to the presence of carbonyl groups in
the ester functionality; and (3) the ratio of LA2 /LA1 response
factors was 1.43. Therefore, the same ratio of response factors
was assigned to each of the successively higher oligomers of
lactic acid for superficial wt.% (LAj /LA1 = 1.43). Using these
response factor ratios for LA3 and LA4 , the concentrations of
LA3 and LA4 were determined from HPLC to be 6 and 2 true
wt.% respectively. Using these values, the material balance
closed (58 + 22 + 6 + 2 + 12 = 100).
To further test the calibration, a series of dilutions where
prepared from a solution that was determined by titration to
be 73.8 superficial wt.% lactic acid. The dilutions spanned
the range of various wt.% of LA1 , LA2 , LA3 , and LA4
acids (0.1–1 wt.% by appropriate dilution with water), and
the HPLC analysis showed a linear concentration response.
Using the response factors determined above, the total superficial concentration was determined to be 74%, in excellent
agreement with titration and thus verifying the reliability of
the oligomer HPLC response factors.
2.3.3.4. Analysis of higher (>LA4 ) lactic acid oligomers.
High oligomers of lactic acid are insoluble in water, but they
are miscible in acetonitrile. Mixtures of acetonitrile + water
have intermediate solvent strength. To dilute a sample of 115
superficial wt.% lactic acid to an overall concentration of
2 wt.% in a homogeneous phase, a solution of at least 50 wt.%
acetonitrile was needed. However, this composition was not
suitable for injection because HPLC could not provide reliable resolution between LA1 and LA2 if more than 20 wt.%
acetonitrile was present in an injected sample containing large
quantities of LA1 and LA2 . The difficulties did not arise when
the quantities of LA1 and LA2 were small. To provide reliable results, lactic acid solutions greater than 105 superficial
wt.% were analyzed in two fractions. Approximately 0.1 g
lactic acid solution was transferred to a microcentrifuge tube
and weighed. Approximately 1 mL of water was added, the
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
129
Table 1
Summary of HPLC results and comparison with total superficial acid by titration
Overall superficial wt.% LA%
HPLC analysis (%EMLA)
Titration
HPLC
LA1
LA2
LA3
LA4
LA5
LA6
LA7
LA8
LA9
LA10
LA4+
12.24
24.36
44.47
53.43a
59.59
70.60
81.46
87.13b
88.06
96.75
100.18
103.27
106.41
113.61
115.47
119.57
10.81
26.88
47.62
51.25a
62.02
71.93
81.90
89.62b
89.63
96.42
102.05
104.43
105.65
108.07
116.25
120.02
99.63
96.31
94.74
94.53a
89.95
84.61
75.66
65.92b
66.85
54.42
45.19
33.36
33.10
29.29
7.62
2.18
0.37
3.59
5.06
5.28a
9.33
13.58
19.49
25.05b
24.09
28.56
29.03
30.11
25.33
24.20
10.47
4.49
0.00
0.10
0.20
0.19a
0.72
1.65
3.88
6.90b
6.87
11.48
14.69
18.97
17.46
17.83
11.44
5.02
0.00
0.00
0.00
0.00
0.00
0.16
0.69
1.63b
1.72
3.84
6.49
9.68
10.76
11.74
12.06
5.83
0.00
0.00
0.00
0.00
0.00
0.00
0.28
0.49b
0.48
1.38
2.98
4.73
6.30
7.28
12.50
8.25
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.32
1.25
1.87
3.47
4.32
11.86
10.40
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.37
0.81
1.91
2.54
10.84
13.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.33
0.96
1.45
9.01
13.81
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.14
0.43
0.77
6.72
12.75
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.19
0.42
3.96
10.54
0.00
0.00
0.00
0.00
0.00
0.16
0.97
2 12
2.20
5.55
11.10
17.56
24.11
28.69
70.47
88.31
Percentages by HPLC analysis are calculated as explained in the introduction and are also plotted on Fig. 4.
a Commercial LA 50%.
b Commercial LA 85%.
solution was shaken, and then centrifuged at 4000 rpm in
a desktop microcentrifuge for 4 min. The water phase was
carefully removed using a pipette. The water extraction was
repeated four to five times. This water-soluble fraction was
weighed and held for analysis. Next, the water-insoluble high
oligomers were recovered in 100% acetonitrile and this acetonitrile phase was weighed. All steps were done at room
temperature. The oligomer contents in both water and acetonitrile were combined in calculation of superficial wt.%
oligomer distribution in the two fractions, and then combined to calculate the superficial wt.% of the original sample
and %EMLAj . The response factors for the higher oligomers
where assumed to be the same as the values for LA3 and LA4 .
The HPLC results for total lactic acid content determined by
adding the superficial wt.% of the individual oligomers is in
good agreement with the results from titration as shown in
Table 1.
3. Mathematical model
We present here a model of infinite oligomer formation
using chemical theory. There are a few examples in the literature of compounds whose phase equilibria properties have
been described with the help of chemical theory or chemical theory along with physical intermolecular forces. The
most strikingly related example is that of formaldehyde in
aqueous and/or methanolic solutions, which reveals extreme
deviations from ideality caused mainly by chemical reactions.
Formaldehyde in the presence of water gives methylene glycol and polyoxomethylenes; in the presence of methanol it
gives hemiformal and higher hemiformals [13].
VLE for formaldehyde-containing systems has been
described using chemical theory by Kogan [14], Kogan and
Ogorodnikov [15,16], Brandani et al. [17] and Masamoto
and Matsuzaki [18]. Maurer [13] presented for the first time
a model in which chemical reactions together with physical
intermolecular forces were used successfully to describe the
VLE and enthalpy for formaldehyde-containing systems containing both reactive and inert components such as trioxane.
Maurer’s model was subsequently extended and tested using
new data; for an update on the model up to 1992 the reader
is referred to Hahnenstein et al. [19]. This approach has also
been used by Brandani et al. [20–22].
For the system formaldehyde-water, the mole fraction of
compounds in the liquid phase is calculated by modeling
the oligomerization as two equilibrium constants—one for
methylene glycol formation from formaldehyde and water
and the second for subsequent higher methylene glycol
oligomer formation.
γMG
xMG
K1 =
(3)
(xw xFA ) (γw γFA )
xn xw
γn γw
Kn =
2≤n
(4)
(xn−1 xMG ) (γn−1 γMG )
These assumptions are reasonable since methylene glycol
is a chemically different structure than formaldehyde, while
the higher oligomers of methylene glycol are chemically
similar to each other. The formaldehyde–methanol system
is treated in a similar way.
3.1. Literature models for lactic acid based on chemical
theory
Prior modeling work to determine the distribution of lactic acid oligomers in solutions above 20 wt.% concentration
has been performed by Bezzi et al. [23] and reported by
130
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
Holten [1]. In the first modeling approach, only the dimers of
lactic acid (LA2 ) were considered. This approach, however,
becomes inaccurate at higher concentrations of lactic acid
(>50 wt.%), where significant oligomerization occurs. In a
second modeling approach, polylactic acids were taken into
account, giving a more realistic representation at high concentrations. However, this model was limited in that solutions
were characterized only by concentration of free lactic acid
(LA) and total oligomer species; no distributions of oligomers
was generated. This polylactic model works poorly at low
concentrations, and is interpretative rather than predictive in
its application.
We are unaware of published mathematical models, apart
from the ones described above, that attempt to represent the
liquid phase distribution of lactic acid and its oligomers in
solution. Therefore, we propose here a model that is based
on chemical theory and incorporates an infinite series of
oligomer components. The model accurately predicts liquid phase compositions of lactic acid in a method similar
to Maurer’s for formaldehyde systems, and represents a clear
advancement of the characterization of concentrated lactic
acid solutions. In order to compare the present model to those
in the literature, this work utilizes the terminology used by
Montgomery [7] and Ueda and Terashima [8] as clarified in
Section 1.1.
This reasoning is analogous to the treatment of the
formaldehyde model, where all polyoxomethylenes have the
same equilibrium constant since they are chemically very
similar but the formaldehyde to methylene glycol reaction
involves different chemical structures and therefore has a different equilibrium constant [13].
Eq. (9) can be rearranged to the following form
nLAj = nLA(j−1) r
where
r=
nLA1 K
nW
nLAj = nLA1 r(j−1)
(5)
LA2 + LA1 ⇋ LA3 + W
(6)
LA3 + LA1 ⇋ LA4 + W
(7)
Generally, oligomer formation can be written as
LA(j−1) + LA1 ⇋ LAj + W
(8)
The chemical reaction equilibrium constants for the above
reactions in a generalized form by
nLAj nW
j>2
(9)
Kj =
(nLA(j−1) nLA1 )
Note that since the number of moles of products and reactants is equivalent regardless of the degree of oligomerization,
the equilibrium constant written in Eq. (9) is equivalent to an
equilibrium constant written in mole fractions.
Since lactic acid oligomers (LA2 , LA3 , etc.) are all formed
via identical reaction pathways and are themselves chemically similar, it is reasonable to assume that the esterification
reactions (Eqs. (5)–(8) above) have the same value of equilibrium constant.
K = K1 = K2 = K3 = K4 = . . . = Kj
(10)
(13)
A total lactic acid superficial mole balance is given by
jnLAj = nLA1 (1 + 2r + 3r 2 + 4r 3 + · · ·)
niLA =
3.2. Infinite series polymer model
2LA1 ⇋ LA2 + W
(12)
and it is recognized that nLA1 and nW are properties of the
solution, identical for all oligomers at a specific superficial concentration. Because of the recursion, it is possible to
write
=
From a thermodynamic standpoint, the formation of
oligomeric intermolecular esters of lactic acid can be
described as the set of successive reactions shown below,
where W denotes water
(11)
nLA1
(1 − r)2
(14)
where the left hand side is the superficial number of
moles of lactate in solution, the second and third expressions represent the infinite converging series obtained
by inserting Eq. (13), and the final term represents the
closed form solution. The water superficial mole balance is given by taking the difference between the true
moles present, and those consumed by hydrolysis of
oligomers
niW = nW −
(j − 1)nLAj
= nW − nLA1 r(1 + 2r + 3r 2 + 4r 3 + · · ·)
nLA1 r
= nW −
(1 − r)2
(15)
where Eq. (13) is substituted into the summation between the
second expression and the third, and the right hand side is the
closed form solution. The left-most variable in Eq. (15) is the
superficial number of moles of water. Eq. (14) can be inserted
into (15) to give
nW = niW + niLA · r
(16)
Inserting Eqs. (14) and (16) into Eq. (12) provides a relation between K and r in terms of the superficial concentrations
of lactic acid and water
K=r
(niW + niLA r)
niLA (1 − r)2
(17)
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
Free acid and all oligomers contribute to titratable acidity
that can be calculated by
nLA1
nLAj = nLA1 (1 + r + r 2 + r 3 + · · ·) =
(18)
(1 − r)
3.3. Application
To apply the model, an overall superficial number of moles
niW , niLA and K are specified. Eq. (17) is rearranged as a
quadratic in r and solved explicitly for the value of r. The
value of r is then used to calculate nLA1 from Eq. (14), and
subsequently the distribution of oligomers from Eq. (13) as
well as the remaining balances.
The equations can be manipulated to express the various
oligomer concentrations in terms of the overall superficial
wt.% lactic acid.
The %EMLA for LAj is
%EMLAj = jr(j−1) (1 − r 2 )
(19)
The superficial wt.% of LAj is
(Superficial wt.% of LAj )
= (%EMLAj )(overall superficial wt.% LA)
(20)
The true wt.% of water is
(True wt.% water)
= 100 + (overall superficial wt.% LA)(0.2r − 1)
(21)
The true wt.% of a LA is
(True wt.% LAj )
= (0.8j + 0.2)(overall superficial wt.% LA)r (j−1) (1 − r)2
(22)
131
4. Results and discussion
4.1. Analytical results and modeling
Aqueous solutions of lactic acid were prepared and analyzed for oligomer concentrations up to 120 superficial wt.%
lactic acid. Table 1 gives a summary of the HPLC results and
a comparison with total acidity of the solution determined
by titration. The HPLC results for overall superficial wt.%
were calculated by summing the peak areas for the individual oligomers. As a check of the HPLC method, the total acid
content by the HPLC and titration agreed within ±3 wt.% for
solutions up to 105 wt.% lactic acid.
The value of the equilibrium constant K = 0.2023 was
obtained by least squares regression of %EMLA for species
LA1 through LA4 simultaneously. Using this value, the
distribution is modeled with an average deviation of ±0.12%
of the reported %EMLA. For each composition from Table 1,
calculated %EMLA of the oligomers is presented in Table 2.
From the HPLC results, the material balance provided the
superficial number of moles of lactic acid and water. Using the
value of K and the superficial moles, the value of r was determined for each overall composition, and then Eq. (19) was
applied.
Fig. 1 shows a GC/MS result for an 85 superficial wt.%
lactic acid solution, demonstrating by molecular weights that
only linear oligomers of lactic acid are present. All four components, namely LA1 , LA2 , LA3 and LA4 , were identified and
verified by their respective mass fragmentation data obtained
from GC/MS.
Fig. 2 shows an example HPLC chromatograph of a 115
superficial wt.% solution of lactic acid. Fig. 3 shows total
titratable acidity as a function of lactic acid concentration
as summarized by Holten [1] from various sources and
from this work. The titratable acidity reflects a balance
between increasing total acid content and increasing
Table 2
Summary of calculated %ELMA for oligomers at each of the experimental compositions from Table 1
Sample (g)
0.081
0.293
0.213
0.116
0.111
0.115
0.107
0.093
0.086
0.081
0.093
0.079
0.054
0.111
0.198
Acid superficial
(wt.%)
niW
(mmol)
niLA
(mmol)
r
10.8
26.9
47.6
51.25
62.0
71.9
81.9
89.6
96.4
102.0
104.4
105.7
108.1
116.2
120.0
3.99
11.9
6.21
3.14
2.34
1.80
1.08
0.533
0.171
−0.093
−0.228
−0.249
−0.240
−1.01
−2.20
0.097
0.876
1.13
0.660
0.763
0.922
0.977
0.921
0.920
0.923
1.07
0.931
0.642
1.44
2.64
0.005
0.014
0.034
0.039
0.058
0.084
0.126
0.180
0.255
0.348
0.397
0.425
0.484
0.721
0.840
Calculated (%EMLAj )
LA1
LA2
LA3
LA4
LA5
LA6
LA7
LA8
LA9
LA10
LA4 +
99.0
97.1
93.3
92.4
88.8
84.0
76.4
67.3
55.5
42.5
36.3
33.1
26.6
7.80
2.57
0.96
2.80
6.36
7.20
10.2
14.0
19.2
24.2
28.3
29.6
28.9
28.1
25.8
11.2
4.32
0.007
0.061
0.326
0.421
0.884
1.76
3.62
6.51
10.8
15.4
17.2
17.9
18.7
12.2
5.440
0.000
0.001
0.015
0.022
0.068
0.196
0.607
1.56
3.68
7.16
9.11
10.1
12.1
11.7
6.095
0.000
0.000
0.001
0.001
0.005
0.020
0.095
0.350
1.17
3.11
4.52
5.39
7.32
10.5
6.402
0.000
0.000
0.000
0.000
0.000
0.002
0.014
0.075
0.358
1.30
2.16
2.75
4.25
9.10
6.456
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.016
0.107
0.527
1.00
1.36
2.40
7.65
6.329
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.003
0.031
0.209
0.454
0.662
1.33
6.30
6.078
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.009
0.082
0.203
0.316
0.725
5.11
5.746
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.003
0.032
0.090
0.149
0.390
4.09
5.365
0.000
0.001
0.015
0.023
0.073
0.218
0.719
2.00
5.36
12.4
17.6
20.9
28.9
68.8
87.7
The first four columns are from experimental results, and the remaining columns are calculated based on the model using K = 0.2023.
132
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
Fig. 1. GC/MS of 85 wt.% LA. The mass fragments (not shown) were used to verify that linear oligomers of LA are present. No lactide was observed.
Fig. 2. HPLC chromatograph of the water soluble fraction from 115 superficial wt.% lactic acid demonstrating the separation of oligomers.
degree of oligomerization that eliminates free acid groups.
The titratable acidity goes through a maximum at about
90 wt.% lactic acid. The model represents the experimental data with an average deviation of ±2% of titratable
acidity.
Fig. 4 shows the experimental distribution of LA1 ,
LA2 , LA3 and higher oligomers collected in this work and
Fig. 3. Total titratable acidity tabulated from various workers by Holten [1]
and measured in this work compared with the model proposed in this work.
() data compiled by Holten; () this work.
compared to data from Ueda and Terashima [8] and Montgomery [7]. Higher oligomers are denoted by LA4+ , i.e. sum
of tetramers and higher oligomers. The abscissa of Fig. 4
denotes the superficial lactic acid concentration; note that
it runs through 125% as explained in the introduction. The
ordinate of Fig. 4 denotes the %EMLA distribution of lactic
acid between monomer and its oligomers on a water-free
basis. The percentages are calculated as described in the
introduction. The lines shown in Fig. 4 are the calculated
values of LA1 , LA2 , LA3 . LA4 and LA4+ from the model.
Excellent agreement is seen between the experimental values
of this work and the values calculated from the model.
It can be seen from the experimental data of this work
and also from Montgomery [7], that there is a maximum
value of approximately 15% EMLA LA3 occurring at 114
superficial wt.% and a maximum value of 29% EMLA LA2
occurring at 105 superficial wt.%. Experimental data from
Ueda and Terashima [8] are also presented; this set of experimental data runs up to 87% total acidity. Watson’s [9]
experimental data are not plotted because he reports the presence of lactide, which is known to be unstable in aqueous
solutions.
Fig. 5 compares the experimental analysis and model
concentrations of LA5 through LA10 for solutions with
Fig. 4. Experimental oligomer distribution compared with the model
expressed as %EMLA. Solid lines represent the model, solid symbols are
measured in this work and open symbols are from literature as reported by
[7] and [16]. The curve labeled LA4 + indicates the sum of all oligomers
LAj ; where j ≥ 4.
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
Fig. 5. Experimental oligomer distribution compared with the model
expressed as %EMLA. Experimental difficulties in analyzing the two highest
concentrations are discussed in the text.
superficial lactic acid content of 80 to 125 wt.%. The
agreement is excellent for analyzed solutions up to 108
superficial wt.% of acid. The agreement is not as good for
the solutions with superficial concentrations of 116 wt.%
and 120 wt.%. These samples were analyzed in two fractions
as discussed above. Since the total acid content is in good
agreement by HPLC and titration (Table 1), we believe that
the disagreement between the model and HPLC results is due
to the incomplete separation of oligomers in the HPLC, even
though distinct peaks appear on the HPLC chromatogram.
Attempts to refine the HPLC method further for these very
high molecular weight solutions have not been successful.
133
Concentrated solutions of lactic acid (>105 superficial
wt.%) are fluid at 120 ◦ C, but are very viscous at room
temperature. The solutions had a very slight amber tint, but
none of the dark coloration indicated by Montgomery [7].
Our results are in good agreement with those of Montgomery
[7] except at the highest concentration. Montgomery reported
incomplete separation of LA3 and higher oligomers—a
problem that we experienced only for higher oligomers
(>LA5 ). To test for hydrolysis under analysis conditions in
this work, ethyl lactate was analyzed using the same HPLC
method as for the lactic acid oligomers and was found to
be stable. Also, our results are also consistent with those of
Montgomery, who tested extensively for hydrolysis.
In discussion of the distribution of weight percentages in
lactic acid solutions, it is appropriate to express the concentrations in terms of superficial wt.%. The superficial wt.%
for oligomers can be quickly calculated from the values in
Table 1 by multiplying the total acid superficial wt.% by the
% EMLA. A summary of true weight percentages calculated
by the oligomer model is shown in Table 3.
4.2. Implementation of lactic acid model into ASPEN
plus
Implementation of the model is extended to ASPEN Plus,
which is the most widely used simulation software in the
chemical process industry. Use of this model will be shown
in future publications for the esterification of lactic acid with
ethanol from the authors’ laboratories [24].
Table 3
Model calculations of true wt.% of water and lactic acid oligomers for various superficial compositions
Superficial
wt.% LA
Superficial
wt.% water
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
123
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
−5
−10
−15
−20
−23
True weight percent compositions
Water
LA1
LA2
LA3
LA4
LA5
LA6
LA7
LA8
LA9
LA10
LA11+
95.0
90.0
85.0
80.0
75.1
70.1
65.1
60.2
55.3
50.4
45.5
40.6
35.8
31.1
26.4
21.9
17.5
13.3
9.49
6.20
3.61
1.79
0.689
0.149
0.0219
4.98
9.91
14.8
19.6
24.3
29.0
33.6
38.0
42.3
46.3
50.2
53.8
56.9
59.6
61.5
62.5
62.2
60.1
55.4
47.6
36.6
23.7
11.6
3.09
0.506
0.019
0.079
0.187
0.350
0.575
0.874
1.26
1.75
2.35
3.11
4.03
5.18
6.58
8.31
10.4
13.0
16.2
19.8
23.6
26.6
27.0
22.9
14.3
4.67
0.853
0.000
0.001
0.002
0.005
0.011
0.021
0.038
0.064
0.105
0.167
0.260
0.400
0.611
0.931
1.42
2.18
3.37
5.23
8.04
11.9
16.0
17.7
14.1
5.66
1.15
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.002
0.004
0.008
0.015
0.028
0.051
0.094
0.175
0.330
0.636
1.25
2.48
4.83
8.56
12.4
12.5
6.22
1.41
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.002
0.004
0.009
0.020
0.047
0.113
0.282
0.725
1.85
4.34
8.21
10.6
6.45
1.63
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.002
0.007
0.019
0.061
0.204
0.684
2.12
5.24
8.58
6.44
1.82
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.003
0.013
0.056
0.246
1.01
3.25
6.79
6.27
1.97
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.003
0.015
0.087
0.469
1.98
5.27
5.99
2.10
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.004
0.030
0.216
1.19
4.03
5.64
2.20
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.010
0.098
0.708
3.05
5.25
2.29
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.005
0.079
0.989
8.55
44.2
84.1
134
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
Fig. 6. Process flow diagram and results for the truncated ASPEN simulation compared to the complete oligomer model. The comparisons of composition are
for a superficial composition of 92.72 wt.% lactic acid.
The proposed model could be incorporated into the
process simulator via a user-written subroutine. As an
alternative, we assume that oligomerization is adequately
approximated by a truncated series. Fig. 4 implies that
solutions up to 90 wt.% can be represented by monomer
lactic acid and the first four oligomers (LA2 –LA5 ). We
have used this assumption to simulate a distillation column
for the purpose of evaluating its suitability for process
simulation.
Fig. 6 shows the ASPEN Plus simulation to remove
water from a 22 superficial wt.% lactic acid solution (nonequilibrated) and form an equilibrated 92.72 superficial wt.%
solution. The reactive distillation column is assumed to operate with equilibrium stages, so the bottoms product contains
an equilibrium mixture of lactic acid oligomers at an overall
concentration of 92.72 superficial wt.%. The oligomer concentrations obtained from the ASPEN Plus simulation with
the truncated model compare well with those from the nontruncated oligomer model as summarized in the inset table
within Fig. 6. The simulation verifies that the model can be
used to model a distillation column where a dilute solution
of lactic acid is converted to concentrated solution consistent with oligomer distribution represented by the full model.
Other options for comparison of the truncated and full model
could have been used, such as an equilibrium reactor with a
non-equilibrium feed; the selection of a distillation column
was arbitrary.
4.3. Effect of temperature and its effect on equilibrium
constant (K)
There are no experimental reports available on heats of
formation of oligomers of lactic acid. Other esterification
reactions involving carboxylic acids and alcohols are either
thermoneutral or have very low heats of formation in the
range of 2–6 kJ/mol [25–27], resulting in negligible to modest
changes (10–15%) in equilibrium constants with temperature
changes of 80 K. In this work, the series of esterification
reactions leading to formation of oligomers are assumed
to be thermoneutral, resulting in a temperature-independent
K = 0.2023. Also, the oligomerization reactions are extremely
slow at room temperature, which makes it very difficult to
assess the reaction kinetics and time required to reach any
redistribution at room temperature [1]. Experiments over a
period of eight weeks showed no measurable redistribution
of oligomers from the solutions that were prepared at the
elevated temperatures reported above.
5. Conclusions
In this work, we provide new data to complement and
extend literature data for oligomerization of lactic acid in
aqueous solutions. We present a model based on chemical
theory that consists of an infinite sequence of equilibrium
D.T. Vu et al. / Fluid Phase Equilibria 236 (2005) 125–135
homo-esterification reactions between successive oligomers
of lactic acid. We show that a single value of the equilibrium
constant (K = 0.2023) applied to all oligomerization reactions
accurately predicts titratable acidity and oligomer concentrations for solution concentrations ranging from very dilute to
greater than 100 superficial wt.% lactic acid. We demonstrate
that inclusion of oligomers only up to LA5 is suitable for process modeling of lactic acid solutions up to 90 wt.%.
List of Symbols
Kj
chemical reaction equilibrium constant for j order
oligomer
LA1
monomeric lactic acid
LA2
dimer lactic acid, lactoyllactic acid
LA3
trimer lactic acid, lactoyl-lactoyllactic acid
LAj
polymeric lactic acid consisting on j units of lactic
acid
nj
molar concentration of component j
r
defined by Eq. (12)
xj
Mole fraction of component j
␥j
activity coefficient of component j
Superscripts
i
initial (used for superficial number of moles)
Subscripts
FA
formaldehyde
j
component
LAj
polymeric lactic acid consisting of j units of lactic
acid
MG
methylene glycol
MGn
higher polyoxomethylene glycols
n
order of oligomer
W
water
Acknowledgement
The authors extend appreciation to the National Corn
Growers Association and the Department of Energy for
financial support.
135
References
[1] C.H. Holten, Lactic acid: Properties and Chemistry of Lactic Acid
and Derivatives, Verlag Chemie, 1971.
[2] C.H. Scheele, Om Mjolk: Kgl. Vetenskaps-Academiens nya Handlingar, 1 Stockholm (1780) 116–124.
[3] M.H. Hartmann, in: D.L. Kaplan (Ed.), Biopolymers from
Renewal Resources, Springer-Verlag, Berlin, 1998, pp. 367–
411.
[4] S.K. Ritter, Chem. Eng. News 82 (22) (2004) 31–34.
[5] W. Robert, Chem. Week. April 10 (2002) 31.
[6] D. Garlotta, J. Polym. Environ. 9 (2002) 63–84.
[7] R. Montgomery, J. Am. Chem. Soc. 74 (1952) 1466–1468.
[8] R. Ueda, T. Terashima, Hakko Kogaku Zaashi. 36 (1958) 371–
374.
[9] P.D. Watson, Ind. Eng. Chem. 32 (1940) 399–401.
[10] K. Tanaka, R. Yoshikawa, C. Ying, H. Kita, K. Okamoto, Chem.
Eng. Sci. 57 (2002) 1577–1584.
[11] Holten, ibid, pp. 200.
[12] A. Engin, H. Haluk, K. Gurkan, Green Chem. 5 (2003) 460–466.
[13] G. Maurer, AIChE J. 32 (1986) 932–948.
[14] L.V. Kogan, Zhur. Prikl. Khim. 52 (1979) 2149.
[15] L.V. Kogan, S.K. Ogorodnikov, J. Appl. Chem. USSR 53 (1980)
98–101.
[16] L.V. Kogan, S.K. Ogorodnikov, J. Appl. Chem. USSR 53 (1980)
102.
[17] V. Brandani, G.D. Giacomo, P.U. Foscolo, Ind. Eng. Chem. Process.
Res. Dev. 19 (1980) 179–185.
[18] J. Masamto, K. Matsuzaki, Chem. Eng. Jpn. 27 (1994) 6–
11.
[19] I. Hahnenstein, M. Hasse, Y.-Q. Liu, G. Maurer, AIChE Symp. Ser.
298 (1994) 141–157.
[20] S. Brandani, V. Brandani, G.D. Giacomo, Ind. Eng. Chem. Res. 30
(1991) 414–420.
[21] S. Brandani, V. Brandani, G.D. Giacomo, Fluid Phase Equil. 63
(1991) 27–41.
[22] S. Brandani, V. Brandani, G.D. Giacomo, Ind. Eng. Chem. Res. 31
(1992) 1792–1798.
[23] S. Bezzi, L. Riccoboni, C. Sullam, Mem. cl. sci. fis. mat. nat. 8
(1937) 181–200.
[24] N.S. Asthana, A.K. Kolah, C.T. Lira, D.J. Miller, US. Patent
Appl. 2004. Improved Process for Production of Organic Acid
Esters.
[25] J. Gangadwala, S. Mankar, S. Mahajani, Ind. Eng. Chem. Res. 42
(2003) 2146–2155.
[26] W. Song, G. Venimadhavan, J.M. Manning, M.F. Malone, M.F.
Doherty, Ind. Eng. Chem. Res. 37 (1998) 1917–1928.
[27] W. Jiu, C. Tan, Ind. Eng. Chem. Res. 40 (2001) 3281–3286.