Lactic acid production from hydroxyacetone on dual metal/base heterogeneous catalytic systems

Article abstract of DOI:10.1039/c5gc00736d

Oxidative aqueous-phase catalytic conversion of hydroxyacetone to lactic acid was investigated under mild and green reaction conditions. Dual metal/base catalyst systems were applied to replace the alkali hydroxide and address the separation requirements and corrosiveness issues usually involved in similar processes. Pt/Al₂O₃ was taken as the supported metal catalyst while Mg-Al layered double hydroxides (LDH) and their derived 3D mixed oxides were exploited as basic solid catalysts. The anion embedded within the LDH interlamellar domain (nitrate or hydroxyl) and the LDH topology (lamellar or exfoliated 2D nanosheets) were also examined. Activity and chemoselectivity of the dual system were shown to be tailored by the decisive to the catalyst performance; the proximity of metal and basic centres on the same surface rendered highly selective bifunctional catalysts. A remarkable 100% selectivity to lactic acid was reached and the undesired cascade oxidation of lactic to pyruvic acid was hindered.

Full text of DOI:10.1039/c5gc00736d

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: E. M.
Albuquerque, L. E. P. Borges and M. A. Fraga, Green Chem., 2015, DOI: 10.1039/C5GC00736D.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/greenchem

Page 1 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
Lactic acid production from hydroxyacetone on dual metal/base heterogeneous
catalytic systems
Elise M. Albuquerquea,b, Luiz E. P. Borgesb, Marco A. Fragaa,b*
aInstituto Nacional de Tecnologia, Laboratório de Catálise, Av. Venezuela, 82/518,
20081-312, Rio de Janeiro/RJ, Brazil.
bInstituto Militar de Engenharia, Praça Gen. Tibúrcio 80, Praia Vermelha, Urca,
22290-270, Rio de Janeiro/RJ, Brazil.
Corresponding author: marco.fraga@int.gov.br
Abstract
Oxidative aqueousꢀphase catalytic conversion of hydroxyacetone to lactic acid was
investigated under mild and green reaction conditions. Dual metal/base catalyst systems
were applied to replace the alkali hydroxide and address the separation requirements
and corrosiveness issues usually involved in similar processes. A Pt/Al2O3 was taken as
the supported metal catalyst while MgꢀAl layered double hydroxides (LDH) and their
derived 3D mixed oxides were exploited as basic solid catalysts. The anion embedded
within the LDH interlamellar domain (nitrate or hydroxyl) and the LDH topology
(lamellar or exfoliated 2D nanosheets) were also examined. Activity and
chemoselectivity of the dual system were shown to be tailored by the morphology,
topology and basicity of the base catalyst. Catalytic activity was up to five times higher
by adding a heterogeneous base solid to the metal catalyst. Surface sites distribution
was also shown to be quite decisive to the catalyst performance; the proximity of metal
and basic centres on the same surface rendered highly selective bifunctional catalyst. A
remarkable 100% selectivity to lactic acid was reached and the undesired cascade
oxidation of lactic to pyruvic acid was hindered.
Keywords: acetol; layered double hydroxide; hydrotalcite; basic catalysts; nanosheets;
2D catalysts; 3D oxides.
1

Green Chemistry
Page 2 of 40
View Article Online
DOI: 10.1039/C5GC00736D
Introduction
As a consequence of the everꢀgrowing global concern about minimizing
environmental impacts caused by the use of fossil fuelꢀderived products, much effort
have been done focusing on the development of new technologies based on renewable
sources. Those relying on processing sidestream compounds from the bioenergy
industry are increasingly more interesting since they may make biofuels more costꢀ
competitive by coꢀproducing high addedꢀvalue chemicals. Processes focused on
glycerol transformation, a coꢀproduct from biodiesel production, match such approach.
The continuous large availability of glycerol worldwide has been motivating its use as a
green feedstock for the synthesis of many valuable compounds, particularly diols,1ꢀ3
ethers4,5 and carboxylic acids.6ꢀ11
Amongst the wide range of carboxylic acids, lactic acid comes up due to its large
industrial use in different sectors, from the conventional ones (food and cosmetics) to
solvents, chemical specialties and biodegradable green plastics. In this later field of
application, the upsurging market of poly(lactic acid) – PLA stands out.
PLA is a biodegradable, compostable and biocompatible polymer, which enables its
use in ecoꢀfriendly packages and medical materials such as sutures, controlled drug
delivery systems and orthopaedic implants.12,13 Moreover, once disposed, PLA
degradation releases CO2 and water, affording a carbon neutral balance. All these
advantages have boosted the demand of PLA and have consequently motivated the
search for new competitive industrial processes to obtain lactic acid.
Lactic acid can be produced by several different routes as the traditional biological
process (fermentative),14,15 chemical synthesis16,17 and homogeneous10,11 or
heterogeneous18,19 catalytic reactions.
2

Page 3 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
The fermentative synthesis corresponds to approximately 90% of the worldwide
production14 and it is based on fermentation of sugars.15 Despite its current industrial
application, it is a complex process because the microorganisms used require suitable
conditions for their growth, nutrients for its cultivation and, more importantly, the
process is time consuming taking two to four days to accomplish profitable yields.20,21
Homogeneous catalyzed reactions make use of saccharides (cellulose, sucrose,
frutose and glucose) as the major feedstocks and the process occurs in an aqueous
medium with metal salts.22,23 Transformation of glycerol to lactic acid with large excess
of inorganic hydroxides as homogenous catalysts has also been reported.10,11 However,
high yields are only achieved under hydrothermal conditions; that is to say that severe
conditions (temperature and pressure) are demanded in addition to the need of strong,
aggressive hydroxides, which are neither recovered nor recycled. Consequently, the
costs involved in these processes are expected to be quite high, discouraging their
technological exploitation.
Alternatively, lactic acid can be obtained from single or multiꢀstep conversion of
glycerol24 over heterogeneous catalytic systems. Oneꢀpot aerobic oxidation was reported
over monometallic (Au, Pd and Pt) and bimetallic AuꢀPt supported catalysts and
selectivities as high as 85% were reached.24 Pt rendered the more active monometallic
catalyst but irrespective of the support, the presence of a soluble metal hydroxide
(NaOH:glycerol = 4:1 molar ratio) was found to be mandatory. These results seemed to
be conflicting with the widespread literature reporting glyceric acid as the main product
from glycerol oxidation under alkaline conditions.25ꢀ30 Nevertheless, it was claimed that
competitive parallel reactions leading to either glyceric acid (via oxidative
dehydrogenation and oxidation steps) or lactic acid (via oxidative dehydrogenation and
3

Green Chemistry
Page 4 of 40
View Article Online
DOI: 10.1039/C5GC00736D
intramolecular rearrangement steps) may occur and the adjustment of the reaction
conditions may determine product distribution.
More recently, a biochemicalꢀchemical twoꢀstep reaction approach was proposed at
which glycerol could be bioconverted to dihydroxyacetone through microbiological
oxidation, followed by its chemical rearrangement to lactic acid in alkaline medium.31
For such second reaction step, alkalineꢀearth metal hydroxides were exploited as
catalysts, as an alternative to alkali hydroxides, due to their fairly low solubility,
allowing thus mechanical separation. Lactic acid yields of around 60% were achieved
by using Ca(OH)2 and the formation of Ca2+ chelates is suggested to drive the chemical
conversion.31,32 Alternatively, dihydroxyacetone can be converted to lactic acid over
solid Lewis acid catalysts and yields varying from 30 to 100% were reached over
boehmite γꢀAlO(OH) 33 and Snꢀcontaining MFI zeolite34 as catalysts, respectively.
Other multiꢀstep proposals for producing lactic acid from glycerol can be assembled.
Initial glycerol conversion to 1,2ꢀpropanediol via hydrogenolysis35,36 followed by diol
selective oxidation18,37 or glycerol dehydration to hydroxyacetone (acetol)38 succeeded
by oxidation/intramolecular rearrangement39 are both feasible approaches.
In a previous work, we have reported the chemoconversion of hydroxyacetone to
lactic acid over Ptꢀsupported catalyst in the presence of NaOH at atmospheric
pressure.39 Remarkable selectivities (>96%) were accomplished and the results pointed
to the crucial role of controlling the reaction conditions to avoid further oxidation of
lactic acid to pyruvic acid.
This study evaluates the oxidative aqueousꢀphase catalytic conversion of
hydroxyacetone to lactic acid under milder and greener reaction conditions. Dual
catalyst systems composed of a supported metal catalyst and layered double hydroxideꢀ
4

Page 5 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
derived basic solids were applied to replace the alkali hydroxide and address the
separation requirements and corrosiveness issues usually involved in similar processes.
Experimental
Catalyst Synthesis
Layered double hydroxides (LDH) and their derived mixed oxides were all
synthesized to be exploited as basic solid catalysts. A commercial available catalyst
5 wt.% Pt/Al2O3 (Sigma Aldrich) was used in this work.
The LDH sample containing magnesium and aluminium was synthesized by
coprecipitation method. In this procedure, 100 mL of an aqueous solution of
Mg(NO3)2.6H2O and Al(NO3)3.9H2O, with a Mg/Al molar ratio of 3, along with a
solution of NaOH (1.0 molLꢀ1), as precipitating agent, was added dropwise into a
reactor containing 200 mL of ultrapure water at 70 °C and pH = 10, which was
maintaining constant during the whole synthesis. The obtained gel was aged at 70 °C for
12 h under nitrogen flow to minimize carbonate contamination. The resulting
suspension was filtered and washed with ultrapure water to remove the excess of the
precipitating agent. The solid was finally dried at 60 °C for 16 h.
ꢀ
A second sample was prepared by substituting nitrate anion NO3 , and eventual
2ꢀ
carbonate species CO3 , embedded between the lamellae of the previously synthesized
LDH, for hydroxyl (OHꢀ) groups. In this protocol, a part of the original material (LDH)
was firstly turned into a mixed oxide by calcination at 800 °C for 4 h. The lamellar
structure reconstruction was accomplished by adding the oxide solid into ultrapure
water in a mass:volume ratio of 2 g:15 mL. The suspension was stirred for 8 h and then
filtered; the recovered solid was dried at 60 °C for 16 h.40 This sample was named
LDHꢀOH.
5

Green Chemistry
Page 6 of 40
View Article Online
DOI: 10.1039/C5GC00736D
A third sample was prepared by exfoliating the initially precipitated LDH. In this
procedure, the bruciteꢀlike layers were detached from each other providing single 2D
nanosheets of MgꢀAl double hydroxide. For preparing such colloidal suspension,
another part of the original LDH was dispersed in ultrapure water with the aid of an
ultrasound for 10 min, maintaining a mass of solid/volume ratio of 0.005. After, the
suspension was transferred to a Teflonꢀlined stainless steel autoclave and then submitted
to hydrothermal treatment at 100 °C for 16 h.41 This 2D samples was designated as
nanoLDH.
Three more samples were prepared by calcining the original LDH under synthetic
air flow (50 mLminꢀ1) for 4 h. These 3D MgOꢀAl2O3 mixed oxides were obtained at
different calcination temperatures (500, 600 and 800 °C), leading to the formation of
MgOꢀAl2O3ꢀ500, MgOꢀAl2O3ꢀ600 and MgOꢀAl2O3ꢀ800 samples, respectively.
Solids characterization
Xꢀray diffraction analyses (XRD) were made in a Miniflex Rigaku diffractometer
using a CuKα radiation (1.5406 Å, 30 kV and 40 mA). The diffraction patterns were
collected at a rate of 0.01°/step between 2° and 90°. The samples were analysed as
powder without any preꢀtreatment.
LDH samples were analysed by thermogravimetry in a TA Instruments SDT Q600.
The samples were heated up to 1000 °C at a rate of 20 °Cminꢀ1 under a synthetic air
stream (30 mLminꢀ1). The decomposition products were analysed in an Ametek
Dymaxion quadrupolar mass spectrometer coupled on-line (TGꢀMS). Signals at m/z =
+
+
+
18, 30, 32, 44 and 46, related, respectively, to H2O+, NO+, O2 , CO2 and NO2 were
monitored.
The morphology of LDH and LDHꢀOH samples were examined by field emission
6

Page 7 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
scanning electron microscopy (FEꢀSEM) in a Quanta FEG 450 FEI microscope
operating with an accelerating voltage of 20 kV. As for the 2D double hydroxide
nanosheets, it was examined by transmission electron microscope (TEM) in a FEI
Tecnai microscope operating at 200kV. Prior to examination, nanoLDH sample was
ultrasonically dispersed in isopropyl alcohol and then deposited on a copper grid coated
with carbon film.
The chemical composition of the samples was determined by Xꢀray fluorescence
spectrometry (XRF) in a Bruker S8 Tiger equipment.
The porosity analysis was performed for the calcined LDH samples (MgOꢀAl2O3
mixed oxides). The nitrogen adsorption and desorption isotherms were obtained at ꢀ196
°C in a Micromeritics ASAP 2020 equipment. Specific surface areas were determined
by BET formalism.
Temperatureꢀprogrammed desorption of CO2 (TPDꢀCO2) was used to characterize
the surface basic sites of mixed oxides MgOꢀAl2O3 samples. The solids were pretreated
in situ under a flow of 8%O2/He (vol.%) at 50 mLminꢀ1 in order to remove carbonate
species engendered by air exposure. Different pretreatment temperatures were applied
in accordance with the respective calcination temperature of each solid. Samples were
then cooled down to room temperature and a flow of pure CO2 was admitted into the
reactor for 30 min. After purging the reactor, desorption was carried out by heating up
the sample under He flow (50 mLminꢀ1) following a rate of 20 °Cminꢀ1 until the
temperature at which the material was calcined.
Catalytic tests
Catalytic chemoconversion of hydroxyacetone was performed in liquid phase in a
semiꢀbatch glass reactor. The experiments were carried out at 40 °C, using a
7

Green Chemistry
Page 8 of 40
View Article Online
DOI: 10.1039/C5GC00736D
hydroxyacetone solution at 0.20 molLꢀ1 as starting material. Synthetic air was bubbled
into the aqueous phase at 30 mLminꢀ1 throughout the whole experiment and the reaction
medium was stirred at 1000 rpm to keep the content perfectly mixed, favour gas
diffusion and ensure kinetic control. Pt/Al2O3 catalyst was previously reduced at
o
o
hydrogen flow at 350 C (10 Cminꢀ1) for 1 h and then mixed with the basic solid at a
1:1 mass ratio. The reaction was monitored for a period of 6 h and aliquots were
collected at every 30 min for the determination of hydroxyacetone conversion and
products distribution.
The concentration of the reaction products was determined by high performance
liquid chromatography (HPLC) in a Waters Alliance e2695 equipped with a refractive
index detector (RID) and a photodiode array detector (PDA). A Biorad Aminex HPXꢀ
87H column was used at 65 °C in isocratic elution mode with a 5 mmolLꢀ1 H2SO4
solution as the mobile phase at a flow of 0.7 mLminꢀ1. The temperature of RID was kept
constant at 50 oC and the analyses at PDA were performed at 210 nm.
The results of the reaction runs were analyzed in terms of hydroxyacetone
conversion and selectivity
ꢎꢏꢐꢑ
ꢀꢁ = ꢂꢃꢄꢃꢅꢆꢇꢈꢇꢆꢉ ꢆꢊ ꢁꢋꢊꢌꢍꢅꢆ ꢁ = 100 ×
∑
ꢎꢏꢐꢑ
The nC are the number of carbon atoms in the products from hydroxyacetone
chemoconversion and Cp are the concentrations of the main identifiable compounds,
namely lactic acid, pyruvic acid and pyruvaldehyde.
The catalyst reusability was evaluated by performing two consecutive reaction runs
of 6 h each. Right after the first run, the catalyst was filtered, dried overnight, calcined
o
o
at 500 ºC and finally reduced at hydrogen flow at 350 C (10 Cminꢀ1) for 1 h. The
second catalytic run was then carried out following exactly the same protocol describe
before.
8

Page 9 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
Results and discussion
A detailed discussion on the main physicoꢀchemical characteristics of the
commercial Pt/Al2O3 catalyst can be found elsewhere.39 This commercial sample
essentially presents platinum nanoparticles (4.6 nm) highly dispersed on a high surface
area transition γꢀAl2O3.39
The asꢀsynthesized LDH basic catalysts were initially analysed by XRD in order to
guarantee the formation of a layered structure. Figure 1 shows the diffractograms
obtained for LDH and LDHꢀOH samples and they both exhibit diffraction patterns
typical of the aimed lamellar hydrotalciteꢀtype structure (JCPDS nr. 22ꢀ700). This
finding revealed that the lamellar structure destroyed upon calcination to release the
nitrate/carbonate anions was indeed reconstructed into LDHꢀOH after rehydration. Such
structural restoration is indeed wellꢀknown in the literature40,42 and it is a consequence
of the memory effect of hydrotalciteꢀlike materials.42 Nonetheless a new diffraction at
42.84° can be seen in the reconstructed LDHꢀOH sample, which can be ascribed to a
reminiscent magnesium oxide cubic structure (JCPDS nr. 45ꢀ0946) formed before
rehydration.
The interlayer spaces were estimated from the (003) plane at ~11o Bragg angle and it
was found to decrease from 0.794 nm (LDH) to 0.772 nm (LDHꢀOH). It matches indeed
ꢀ
the expected shortening promoted by replacing larger NO3 to OHꢀ anions in the
interlamellar domain as it is wellꢀknown that such space is determined by the
dimensions of the ion species embedded in between.42 It should be mentioned though
that the LDH interlayer spacing as assessed by XRD falls in an intermediate value if one
ꢀ
consider the sole occurrence of either NO3 (0.879 nm) or CO32ꢀ (0.765 nm).42 Similarly,
the value for LDHꢀOH is higher than the theoretical one (0.755 nm).42 These trends
suggest that some CO2 was also able to find its way into the interlamellar domain during
9

Green Chemistry
Page 10 of 40
View Article Online
DOI: 10.1039/C5GC00736D
coprecipation and reconstruction procedures despite the nitrogen flushing. This
discussion was indeed corroborated by identifying the products released during thermal
decomposition in TGꢀMS analyses (Figure 2).
Some differences can be seen in the thermograms; while four thermal events are
registered for LDH, decomposition of LDHꢀOH occurs in three stages. The thermal
o
events occurring below 240 C are associated with the elimination of water (adsorbed
and that present between the lamellae of the materials) and the mass loss in intermediate
temperatures (240 ꢀ 445 °C) is related to the dehydroxylation of the hydrotalciteꢀtype
structure along with the release of interlamellar carbonate species as identified by the
CO2 (m/z = 44) profile registered in the mass spectrometer. A fourth event at high
temperature (>450 °C) is only seen by the LDH sample and corresponds to the removal
ꢀ
of NO3 species from the interlayer space. These data are quite consistent with some
reported elsewhere42,43 and the absence of such thermal event in LDHꢀOH thermogram
confirms the replacement of the compensating nitrate anions during the lamellar
material reconstruction process.
In order to verify the morphology of the lamellar samples and observe the occurrence
of eventual morphological changes from LDH to LDHꢀOH, both samples were
examined by FEꢀSEM and representative images are displayed in Figure 3. It can be
seen that they exhibit similar morphological characteristics, being formed by aggregates
of plentiful thin plates as typically observed for lamellar materials.44,45
The nanoLDH sample was characterized only by TEM to assure the efficient
formation of double hydroxide nanosheets. One should bear in mind that this material
can only be obtained as a 2D nanostructure in a colloidal suspension, since the original
nanosheets stacking immediately occurs by removing the solvent, restoring the layered
structure.46 A representative TEM image of the exfoliated LDH is also depicted in
10

Page 11 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
Figure 3, revealing that the isolated lamellae are indeed irregular hexagonal nanoplates
with a wide size distribution ranging from 30 to 100 nm.
The three activated LDH samples (calcined solids) were analysed as regarding their
crystalline structure, porosity and surface basicity. As expected, their diffractograms
show that the lamellar structure is completely destroyed after calcination at temperature
as low as 500 oC. As a matter of fact, the diffractograms are quite similar regardless the
calcination temperature (Figure 4), exhibiting peaks related to the cubic structure of
MgO periclase phase (space group Fm3m). The lattice parameters collected in Table 1
are consistent with a bare MgO structure, indicating that a segregated alumina phase is
obtained upon calcination. However, no other crystalline phase was identified in the
XRD patterns, which suggests that an amorphous or lowꢀcrystalline alumina is formed.
Finally, MgO mean particle size is seen to increase gradually by raising the calcination
temperature (Table 1).
BET surface areas are also summarized in Table 1. All materials present high surface
areas as expected for mixed oxides obtained from layered double hydroxide precursors.
An important increase (~40%) in surface area is seen when calcination temperature is
raised from 500 to 600 oC and then a slight decrease is detected by further increasing the
o
temperature to 800 C. Such trend can be rationalized by the thermal decomposition
dynamics of LDH as previously discussed herein. As the total decomposition of the
o
double hydroxide precursor only occurs at above 600 C (Figure 2), the development of
a porous structure, and as a consequence higher surface area, can be expected from
samples MgOꢀAl2O3ꢀ500 to MgOꢀAl2O3ꢀ600; once it is generated, the onset of porous
structure collapse or sintering may occur, gradually leading to a drop in surface area
(MgOꢀAl2O3ꢀ800).
11

Green Chemistry
Page 12 of 40
View Article Online
DOI: 10.1039/C5GC00736D
Temperatureꢀprogrammed desorption of CO2 was used to qualitatively and
quantitatively estimate the distribution of surface basic sites of the calcined materials;
the profiles are depicted in Figure 5 and the total density of basic sites are listed in
Table 1. All samples present similar patterns, a desorption peak at low temperature
(<200 °C) and a long tail extending to higher temperatures, evidencing a wide
distribution of distinct basic sites. The experimental curves were thus mathematically
decomposed considering multiple Gaussian contributions. The position of each peak
was determined from an initial estimative and the calculations were conducted based on
a nonlinear routine to minimize the deviations of the squares.47 The desorption
temperature is intimately associated with the basic site strength; the higher the
desorption temperature, the strongest the basic sites. Thus, the basic sites were divided
into weak (up to 170 °C), medium (between 170 and 270 °C) and strong (above
270 °C). Though empirical, these temperature ranges are commonly used in the
literature and are very useful as a rough guide to assign basic strength and compare and
contrast surface site distribution.
Despite exhibiting similar desorption profiles, the density of each basic site is
different. The solid activated at 500 °C has the highest density of sites roughly equally
distributed among the three types. The increase of the calcination temperature caused a
very sharp decrease in the total concentration of basic sites but keeping a similar
strength distribution. At the highest calcinations temperature, the density is dropped
even further and the presence of weak base sites is dominant (~80%).
All base solids were evaluated in the aqueousꢀphase conversion of hydroxyacetone to
lactic acid as a part of a dual metal/base heterogeneous catalytic system. In order to
obtain such a system, base solids were mixed with a previously reduced Pt/Al2O3
commercial catalyst in a 1:1 mass ratio. Catalytic conversion of hydroxyacetone in an
12

Page 13 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
oxidative atmosphere may lead to a wide variety of organic compounds and even carbon
dioxide if completely oxidized. Figure 6 depicts a simplified reaction scheme
highlighting only the more relevant C3 chemicals.
Hydroxyacetone oxidation/oxidative dehydrogenation on the metal catalyst would
lead to pyruvaldehyde, which may be further oxidised to pyruvic acid. It has also to be
taken into consideration the intramolecular disproportionation (Cannizzaro reaction) of
the intermediate pyruvaldehyde to lactic acid on the base catalyst. It has indeed been
reported that such rearrangement easily occurs in alkaline medium19,33,39. Finally, the
undesired oxidation of lactic acid to pyruvic acid is also possible.
Initially, some control experiments were performed in order to examine the role
played by each component of the dual metal/base catalytic system. It is firstly worth
mentioning that no conversion of hydroxyacetone occurs without either catalyst. By
using only the Pt/Al2O3 metal catalyst (Table 2 entry 1), pyruvic acid and
pyruvaldehyde are the major products and only around 5% of hydroxyacetone is
converted within 6 h. On the other hand, lactic acid is the only product detected when
MgOꢀAl2O3 mixed oxide base catalyst is used instead (Table 2, entries 2). It suggests
that the acid is formed from hydroxyacetone dehydrogenation to pyruvaldehyde, which
is rearranged to lactic acid; all steps occurring on the base surface sites. LDH catalyst,
however, showed no activity at all, which is probably related to its lower basicity when
compared to the mixed oxides. All in all, these control experiments show that neither
the metal catalyst nor the base solids alone are kinetically and/or selectively effective
for hydroxyacetone chemoconversion to lactic acid. At the same time, they outline the
key role played by the base centres on selectivity.
As for the dual catalytic systems, it can be easily noticed that they are definitely
more active than the sole metal or base catalysts (Table 2). However they behave quite
13

Green Chemistry
Page 14 of 40
View Article Online
DOI: 10.1039/C5GC00736D
differently and can indeed be divided in two sets; the less active systems, leading
preferentially to the formation of pyruvic acid, are formed by the ones containing
double hydroxides as base catalysts (Table 2, entries 3, 4 and 5), while those comprising
mixed oxides are significantly more active and selective to lactic acid (Table 2, entries
6, 7 and 8).
Irrespective of the anion embedded within the LDH interlamellar domain (nitrate,
entry 3 or hydroxyl, entry 4 in Table 2) or the double hydroxide topology (lamellar,
entry 3 or exfoliated 2D nanosheets, entry 5 in Table 2) the reaction rates are higher
than that estimated for the bare metal catalyst (Table 2, entry 1). Lactic acid is formed
with selectivities around 20 – 30% and pyruvaldehyde is also detected; however,
product distribution is defined by the selective formation of pyruvic acid (55 – 75%).
Despite the similar behaviour at first sight, the timeꢀresolved selectivity profiles
displayed in Figure 7 reveal that these double hydroxideꢀbased dual systems exhibit
significantly different reaction dynamics according to the topology of the base solids.
On the systems composed of lamellar structured base catalysts (LDH and LDHꢀOH),
lactic acid is initially formed along with pyruvic acid and pyruvaldehyde (Figure 7a and
b). It indicates that oxidative dehydrogenation of hydroxyacetone to pyruvaldehyde on
the metal catalyst and its subsequent disproportionation on the base solid take place.
The presence of a base catalyst may also catalyse the interconversion of hydroxyacetone
to lactaldehyde; this prototropic tautomer may be oxidised on metal sites to lactic acid.
This tautomerization route as a first reaction step cannot be ruled out. The formation of
pyruvic acid, however, increases steadily over time and the selectivity trends allow the
conclusion that it may be formed from lactic acid oxidation.
As for the dual system using nanoLDH colloidal suspension as heterogeneous
catalyst (Figure 7c), a contrasting selectivity to pyruvaldehyde can be observed. Such
14

Page 15 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
striking high selectivity (Table 2, entry 5) indicates the limited activity of 2D double
hydroxide nanosheets in the baseꢀinduced intramolecular disproportionation of
pyruvaldehyde to lactic acid. This fashion might also be associated with the presence of
acidic nitrate anions that are set free in the reaction medium upon exfoliation of the
original double hydroxide lamellar structure. It is interesting to notice, however, that the
selectivity to lactic acid is roughly the same within the whole reaction period while an
interdependence between pyruvaldehyde and pyruvic acid is clearly established in the
selectivity profiles displayed in Figure 7c. This result suggests that the formation of
lactic acid on this Pt/Al2O3 + nanoLDH dual system might not involve the Cannizzaro
reaction as previously observed for the lamellar systems but proceeds only via
tautomerization of hydroxyacetone followed by the tautomer oxidation instead.
These findings allow the proposition that simultaneous conversion of
hydroxyacetone via parallel reactions can take place over LDHꢀcontaining dual catalytic
systems but following distinct routes according to the base catalyst topology. A general
schematic reaction pathway considering the lamellar and 2D nanosheets structures for
the base solid is presented in Figure 8a and b, respectively.
The second set of dual catalytic systems is constituted of mixed oxides as the base
solid component (Table 2, entries 6, 7 and 8). These systems are meaningfully more
active, presenting reaction rates approximately 4 – 5 times higher than that valued for
the sole Pt/Al2O3 catalyst (Table 2, entry 1). Lactic acid is always the primary product
with selectivities above 50%, notably higher than those reached with the previous
catalytic systems reported herein. Unlike the double hydroxideꢀcontaining systems,
pyruvic acid is the only other product formed and no traces of pyruvaldehyde were
detected whatsoever. Therefore hydroxyacetone tautomerization and the subsequent
oxidation could be thought at first as the key reaction route for lactic acid production.
15

Green Chemistry
Page 16 of 40
View Article Online
DOI: 10.1039/C5GC00736D
This different behaviour can be associated with the different basicity of the solid base
catalysts as already mentioned. The more basic medium provided by using the MgOꢀ
Al2O3 mixed oxides is evident by the higher pH value of the reaction aqueous phase as
summarized in Table 2. Nevertheless, complementary experiments carried out with
pyruvaldehyde as the initial substrate instead of hydroxyacetone showed that the
aldehyde is promptly converted to lactic acid through Cannizzaro reaction. Similar
findings have indeed been reported elsewhere.19,33,39 These outcomes lead us also to
consider the intermediate formation of pyruvaldehyde and its immediate baseꢀinduced
disproportionation to lactic acid.
Timeꢀresolved trends for the distribution of products (Figure 9) indicate that these
3D oxideꢀcontaining dual systems behave likewise and the cascade oxidation of lactic
acid to pyruvic acid is also seen to proceed on these systems. Nonetheless, this
chemoconversion interdependence is kinetically shifted to longer times (or higher
hydroxyacetone conversion) correlating well with the estimated reaction rates as
normalized by metal catalyst weight (Table 2), suggesting that the more active the
catalytic systems the faster the further oxidation of lactic to pyruvic acid. Such a trend is
in close agreement with the results reported elsewhere on the use of a homogeneous
alkaline catalyst.39
Lastly, the timeꢀresolved curves allow to better differentiate the performance of these
dual systems and to tune the operation conditions in order to optimize lactic acid yields.
Despite the over oxidation of lactic acid upon reaction time, the outstanding behaviour
of Pt/Al2O3 + MgOꢀAl2O3ꢀ800 easily comes out as 100% selectivity toward lactic acid
can be accomplished along the first hour. It brings up the possibility of settling a highly
selective process by adjusting reaction time and developing a more suitable process set
up with series reactors and/or continuous recirculation of liquid feed.
16

Page 17 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
All results presented herein reveal that the production of lactic acid from
hydroxyacetone proceeds via cascade reactions and that a multitude of different
combinations may lead to lactic acid formation. Since the dual systems investigated in
this work comprise two separate solid catalysts, i.e. the metal and base surface sites are
set utterly apart from each other, it is conceivable that the reaction pathway would
depend on each of those two catalysts the hydroxyacetone molecule would first meet in
the reaction medium and adsorb on. Likewise it is also reasonable that once the lactic
acid formed from any cascade combination desorb from the catalyst surface and diffuses
into liquid phase it can be oxidized to pyruvic acid on the metal catalyst since they are
both kept in the same reactor. A different arrangement of the surface active sites could
thus disturb this reaction network dynamics and bring some benefit to the
chemoselective formation of the desired acid, driving the design of a more robust solid
catalyst and/or a more costꢀeffective catalytic process.
Therefore, to assess the relevance of such surface sites disposal, a bifunctional
catalyst was synthesized taking the MgOꢀAl2O3ꢀ800 as a basic support for platinum
metal nanoparticles (Pt/MgOꢀAl2O3ꢀ800). This catalyst was characterized as concerning
chemical composition and metal dispersion by X ray fluorescence spectrometry and
transmission electron microscopy, respectively. It was seen that it presented practically
the same platinum content (4.7 wt%) as the Pt/Al2O3 used in the dual systems but with
much smaller metal nanoparticles (2.4 nm). A representative TEM image is depicted in
Figure 10.
The reaction rate obtained for this sample (Table 2, entry 9) disclosed that it is as
active as its parent dual system (Table 2, entry 8). Nevertheless, this bifunctional
catalyst revealed to be remarkably selective to lactic acid (99%) with only small
amounts of a side product (pyruvic acid) within the whole period of the reaction, as
17

Green Chemistry
Page 18 of 40
View Article Online
DOI: 10.1039/C5GC00736D
presented in Figure 11. This notable performance is likely related to the active sites
assembly, which are now set together on the same surface (vicinal sites). Such vicinity
may facilitate the oneꢀstep conversion of hydroxyacetone to lactic acid; that is to say
that after the adsorption of hydroxyacetone molecule, the intermediate pyruvaldehyde
formed through oxidative dehydrogenation on a metal centre is immediately rearranged
to lactic acid on the vicinal basic site before leaving the catalyst surface. The subsequent
oxidation of the desorbed lactic acid is likely avoided due to the preferential adsorption
of hydroxyacetone on metal sites. It is definitely a noteworthy result for a
heterogeneous catalyst since such chemoselectivity is usually reached over biochemical
systems.
Due to the expressive result, the spent bifunctional catalyst was recovered by
filtration right after reaction, dried and characterized as regarding its crystalline
structure. Its XRD pattern is compared and contrasted with that obtained from the fresh
sample in Figure 12.
It is seen that the mixed oxide 3D structure modifies upon hydroxyacetone aqueousꢀ
phase processing as new diffraction peaks were detected after reaction. As a matter of
fact, the XRD analysis revealed that the catalyst turns back into a lamellar structure,
which is quite clear for the rising diffraction peaks at 11.4º and 22.5º. This behaviour
can indeed be expected due to the wellꢀknown memory effect of such solids as
discussed hereinbefore, which favours the oxide rehydration to a double hydroxide and
the consequent regeneration of the layered topology.
To evaluate the bifunctional catalyst reusability, the dried spent catalyst was firstly
calcined in order to reconstruct the 3D structure of the original mixed oxideꢀsupported
sample. This reassembling is rather important to perform a second reaction cycle
18

Page 19 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
assuring that it starts with the same catalyst, i.e. with a sample with the same chemical
and structural features.
The catalyst revealed to keep as active as in the first run as disclosed by the reaction
rate (3.7 molgꢀ1hꢀ1 in the second run versus 3.5 molgꢀ1hꢀ1 in the first run). Accordingly,
the selectivity to lactic acid was also high (90%) at the beginning of the reaction with
only a minor formation of pyruvic acid. However, contrarily to what was seen in the
first run, lactic acid selectivity dropped gradually upon time (Figure 13a) along with an
everꢀincreasing selectivity to pyruvic acid (Figure 13b). This fashion indicates some
change in the distribution of surface sites during recycling which must be taken into
consideration to further optimize the design of a bifunctional heterogeneous catalyst
bearing metal and base centres.
The results reported herein appoints for the establishment of a promising green
process that would meet the technological challenges of catalyst separation and
corrosiveness while keeping biochemicalꢀlike high selectivity to lactic acid.
Conclusions
The results presented herein show that it is possible to selectively produce lactic acid
from aqueousꢀphase conversion of hydroxyacetone over a heterogeneous catalyst dual
system. For a single Pt metal catalyst, activity and chemoselectivity of the dual system
can be tuned by the morphology, topology and basicity of the base catalyst. Catalytic
activity was up to five times higher by adding a heterogeneous base solid to the metal
catalyst. Lactic acid selectivity was raised from 20ꢀ30% to 50ꢀ70% by changing from
lamellar double hydroxide to 3D mixed oxide base catalysts. Surface sites distribution
was also shown to be quite decisive to the catalyst performance; the proximity of metal
19

Green Chemistry
Page 20 of 40
View Article Online
DOI: 10.1039/C5GC00736D
and basic centres on the same surface rendered highly selective bifunctional catalyst. A
remarkable 100% selectivity to lactic acid was reached and the undesired cascade
oxidation of lactic to pyruvic acid was hindered.
Acknowledgements
The authors acknowledge the financial support from CNPq and FAPERJ. Mr.
Francisco Rangel (INT) is also thanked for his valuable assistance in the FEꢀSEM
analyses. Ms. Nataly Amorim (CETENE) and Dr. A.M Duarte de Farias (INT) are also
acknowledged for their support in TEM examination. EMA also thanks CAPES for the
scholarship.
20

Page 21 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
References
1 J. Chaminand, L. Djakivitch, P. Gallezot, P. Marion, C. Pinel and C. Rosier, Green
Chem., 6, 359ꢀ361, 2004.
2 M. A. Dasari, PꢀP. Kiatsimkul, W. R. Sutterlin, G. J. Suppes, Appl. Catal. A, 281,
225ꢀ231, 2005.
3 T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A, 318, 244ꢀ251,
2007.
4 K. Klepáčová, D. Mravec, M. Bajus, Appl. Catal. A, 294, 141ꢀ147, 2005.
5 R. S. Karinen, A. O. I. Krause, Appl. Catal. A, 306, 128ꢀ133, 2006.
6 N. Dimitratos, F. Porta, L. Prati, Appl. Catal. A, 291, 210ꢀ214, 2005.
7 N. Dimitratos, F. Porta, L. Prati, A. Villa, Catal. Lett., 99, 181ꢀ185, 2005.
8 C. L. Bianchi, P. Canton, N. Dimitratos, F. Porta, L. Prati, Catal. Today, 102ꢀ103,
203ꢀ212, 2005.
9 S. Demirel, K Lehnert, M. Lucas, P. Claus, Appl. Catal. B, 70, 637ꢀ643, 2007.
10 H. Kishida, F. Jin, Z. Zhou, T. Moriya, H. Enomoto, Chem. Lett., 34, 1560ꢀ1561,
2005.
11 Z. Shen, F. Jin, Y. Zhang, B. Wu, A. Kishita, K. Tohji, H. Kishida, Ind. Eng. Chem.
Res., 48, 8920ꢀ8925, 2009.
12 W. Amass, A. Amass, B. Tighe, Polym. Int., 47, 89ꢀ144, 1998.
13 M. Jamshidian, E. A. Tehrany, M. Imran, M. Jacquot, S. Desobry, Compr. Rev. Food
Sci F., 9, 552ꢀ571, 2010.
14 K. Hofvendahl, B. HahnꢀHägerdal, Enzyme Microb. Technol., 26, 87ꢀ107, 2000.
15 SꢀQ. Liu, Int. J. Food Microbiol., 83, 115ꢀ131, 2003.
16 R. P. John, Anisha G. S., K. M. Nampoothiri, A. Pandey, Biotechnol. Adv., 27, 145ꢀ
152, 2009.
21

Green Chemistry
Page 22 of 40
View Article Online
DOI: 10.1039/C5GC00736D
17 R. P. John, K. M. Nampoothiri, A. Pandey, Appl. Microbiol. Biotechnol., 74, 525ꢀ
534, 2007.
18 T. Tsujino, S. Ohigashi, S. Sugiyama, K. Kawashiro, H. Hayashi, J. Mol. Catal., 71,
25ꢀ35, 1992.
19 N. Dimitratos, J. A. LopezꢀSanchez, S. Meenakshisundaram, J. M. Anthonykutty, G.
Brett, S. H. Taylor, D. W. Knight, G. J Hutchings, Green Chem., 11, 1209ꢀ1216, 2009.
20 R. Datta, SꢀP. Tsai, P. Bonsignore, SꢀH. Moon, J. R. Frank, FEMS Microbiol. Rev.,
16, 221ꢀ231, 1995.
21 R. Datta, M. Henry, J. Chem. Technol. Biotechnol., 81, 1119ꢀ1129, 2006.
22 M. Bicker, S. Endress, L. Ott, H. Vogel, J. Mol. Catal. A, 239, 151ꢀ157, 2005.
23 L. Kong, G. Li, H. Wang, W. He, F. Ling, J. Chem. Technol. Biotechnol., 83, 383ꢀ
388, 2008.
24 Y. Shen, S. Zhang, H. Li, Y. Ren, H. Liu, J. Chem. Eur., 16, 7368ꢀ7371, 2010.
25 R. Garcia, M. Besson, P. Gallezot, Appl. Catal. A., 127, 165ꢀ176, 1995.
26 M. Besson, P; Gallezot, Catal. Today, 57, 127ꢀ141, 2000.
27 F. Porta, L. Prati, J. Catal., 224, 397ꢀ403, 2004.
28 C. L. Bianchi, P; Canton, N. Dimitratos, F. Porta, L. Prati, Catal. Today, 102ꢀ103,
203ꢀ212, 2005.
29 S. DemirelꢀGülen, M. Lucas, P. Claus, Catal. Today, 102ꢀ103, 166ꢀ172, 2005.
30 S. Carrettin, P. McMorn, P. Johnston, K. Griffin, G. J. Hutchings, Chem. Commun.,
696ꢀ697, 2002.
31 S. Lux, M. Siebenhofer, Catal. Sci. Technol., 3, 1380ꢀ1385, 2013.
32 T. Okuyama, K. Kimura, T. Fueno, Bull. Chem. Soc. Jpn., 55, 2285ꢀ2286, 1982.
33 A. Takagaki, J. C. Jung, S. Hayashi, RSC Adv., 4, 43785ꢀ43791, 2014.
22

Page 23 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
34 M. Morales, P.Y. Dapsens, I. Giovinazzo, J. Witte, C. Mondelli, S.
Papadokonstantakis, K. Hungerbuhler, J. PerezꢀRamírez, Energy Environ. Sci., 8, 558ꢀ
567, 2015
35 S. Bolado, R. E. Treviño, M. GarcíaꢀCubero, G. GonzálezꢀBenito, Catal. Commun.,
12, 122ꢀ126, 2010.
36 Z. Huang, F. Cui, H. Kang, J. Chen, C. Xia, Appl. Catal. A, 366, 288ꢀ298, 2009.
37 H. H. C. M. Pinxt, B. F. M. Kuster, G. B. Marin, Appl. Catal. A, 19, 45ꢀ54, 2000.
38 A. K. Kinage, P. P. Upare, P. Kasinathan, Y. K. Hwang, JꢀS. Chang, Catal.
Commun., 11, 620ꢀ623, 2010.
39 E. M. Albuquerque, L. E. P. Borges, M. A. Fraga, J. Mol. Catal. A, 400, 64ꢀ70,
2015.
40 K. Ebitani, K. Motokura, K. Mori, T. Mizugaki, K. Kaneda, J. Org. Chem., 71,
5440ꢀ5447, 2006.
41 Z. P. Xu, g. S. Stevenson, CꢀQ. Lu, G. Q. Lu, P. F. Barlett, P. P. Gray, J. Am. Chem
Soc., 128, 36ꢀ37. 2006.
42 F. Cavani, F. Trifiró, A. Vaccari, Catal. Today, 11, 173ꢀ301, 1991.
43 T. Kameda, Y. Fubasami, N. Uchiyama, T. Yoshioka, Thermochim. Acta, 499, 106ꢀ
110, 2010.
44 M. Wang, W.J. Bao, J. Wang, K. Wang, JꢀJ. Xu, HꢀJ. Chen, XꢀH. Xia, Nature, 4, n.
6606, 2014.
45 MꢀQ. Zhao, Q. Zhang, JꢀQ, Huang, GꢀL. Tian, JꢀQ. Nie, HꢀJ. Peng, F. Wei, Nat.
Commun., 5, n. 3410, 2014.
46 C. R. Gordijo, V. R. L. Constantino, D. O. Silva, J. Solid State Chem., 180, 1967ꢀ
1976, 2007.
23

Green Chemistry
Page 24 of 40
View Article Online
DOI: 10.1039/C5GC00736D
47 A. O. Menezes, P. S. Silva, E. P. Hernández, L. E. P. Borges, M. A. Fraga,
Langmuir, 26, 3382ꢀ3387, 2010.
24

Page 25 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
Figure 1. Diffraction patterns of synthesized LDH samples.
25

Green Chemistry
Page 26 of 40
Figure 2. TGꢀMS profiles of (a) LDH and (b) LDHꢀOH samples.
26

Page 27 of 40
Green Chemistry
Figure 3. Representative SEM images of (a) LDH and (b) LDHꢀOH samples and a TEM image of nanoLDH sample.
27

Green Chemistry
Page 28 of 40
View Article Online
DOI: 10.1039/C5GC00736D
Figure 4. Diffraction pattern of MgOꢀAl2O3 mixed oxides.
28

Page 29 of 40
Green Chemistry
View Article Online
DOI: 10.1039/C5GC00736D
Figure 5. TPDꢀCO2 profiles of MgOꢀAl2O3 mixed oxides.
29