Chapter 9

Production of Sorbitol from Biomass

José R. Ochoa-Gómez and Tomás Roncal

Abstract  Sorbitol is a natural occurring sugar alcohol with a current industrial

demand of about 2,000,000 t.year−1 showing its huge worldwide commercial inter-
est, encompassing uses in chemical, food, textiles, pharmaceutical, and health care
and cosmetic industries. The current interest for substituting oil derived chemicals
by biomass derived ones has boosted the interest in sorbitol production because in
2004 it was identified by US Department of Energy as one of the 12 top chemicals
derived from carbohydrates which could be potentially used as platform chemicals
for producing valuable chemical intermediates and materials for industry, and inclu-
sion of sorbitol in the listing currently remains. This review analyzes both sorbitol’s
current market and its potentiality as a platform chemical. Subsequently, current
state of sorbitol production by chemical, electrochemical and biotechnological
methods is revised, and includes a key issue for industrial success: its recovery and
purification. Finally, some prospects about the direction of future research for over-
coming current bottlenecks for further development are discussed.

Keywords Sorbitol chemical production • Sorbitol as a platform chemical •

Sorbitol electrosynthesis • Sorbitol biotechnological production • D-glucose
reduction

Abbreviations

BET Brunauer–Emmett–Teller, authors of the BET theory that is the basis

for an analysis technique for the measurement of the specific surface
area of a material
ATP adenosine 5′-triphosphate
ED electrodialysis
GFOR D-glucose-fructose oxidoreductase

J.R. Ochoa-Gómez (*) • T. Roncal

TECNALIA, Division of Energy and Environment, Biorefinery Department,
Parque Tecnológico de Álava, Leonardo Da Vinci, 11, 01510 Miñano, Spain
e-mail: jramon.ochoa@tecnalia.com

© Springer Nature Singapore Pte Ltd. 2017 265

Z. Fang et al. (eds.), Production of Platform Chemicals from Sustainable
Resources, Biofuels and Biorefineries 7, DOI 10.1007/978-981-10-4172-3_9
266 J.R. Ochoa-Gómez and T. Roncal

GHSV gas hourly space velocity, h−1

HRTEM high-resolution transmission electron microscopy
NAD+ oxidized form of nicotinamide adenine dinucleotide
NADH reduced form of nicotinamide adenine dinucleotide
NADP Nicotinamide adenine dinucleotide phosphate
NCNT nitrogen doped carbon nanotubes
NP nanoparticles
RPBR continuous recycle packed-bed reactor
TOF turnover frequency, s−1 or h−1
TOS time on stream
WHSV weight hourly space velocity, h−1

9.1  Introduction

Sorbitol (D-glucitol, D-sorbitol, D-glucohexane-1,2,3,4,5,6-hexol, CAS Number

50-7-4) was discovered in 1872 by the French chemist Boussingault in the berries
of the mountain ash (Sorbus aucuparia L.) and is now known to occur naturally in
a wide range of fruits and berries. During several decades it was only obtained from
natural sources and scarcely consumed at prices corresponding to a fine chemical
with almost no practical interest outside of the medicinal practice. However, during
the Great Depression, the Atlas Powder Co. (since 1971 ICI Americas Inc.) under-
took a program for investigating production methods and industrial uses of the
higher polyhydric alcohols. This led to the operation of a pilot plant during 1935
and 1936 and finally, as a consequence of the rapidly growing demand, to the con-
struction of a facility for manufacturing 1400 t.year−1 of sorbitol and mannitol by
electroreduction of D-glucose resulting from corn starch [1]. A few years later the
electrochemical process was discontinued and displaced by the more economical
high-pressure catalytic hydrogenation, which is currently the single worldwide
operated chemical process for sorbitol manufacturing.
Since that time, the industrial importance of sorbitol has been increasingly higher
becoming a commodity as shown by a current worldwide demand of about 2000
kt.year−1 and a price around 0.55–0.65 $.kg−1 as a 70% syrup, the more usual com-
mercial form. This importance is likely to grow in the coming future due to the
current worldwide trend for moving from the petro-economy to the bioeconomy,
requiring the use of biomass-derived chemicals as building blocks for manufactur-
ing the intermediates and materials needed for maintaining the standard of living. In
fact, sorbitol was identified by US Department of Energy as a top 12 chemical
derived from carbohydrates that could be used as a platform chemical [2], and inclu-
sion of sorbitol in the list currently remains [3].
Undeniably, the growing industrial importance of sorbitol is attracting strong
interest in improving methods in its production as well as in looking into new
­processes. The objective of this chapter is to give an overview of sorbitol production
9  Production of Sorbitol from Biomass 267

processes. Section 9.2 is devoted to justify the interest in continuous development

of sorbitol production methods as well as in improvement of the current one by
highlighting the industrial importance of sorbitol from two perspectives: the current
one and the future one. The former is set forth by describing its market and applica-
tions (Sect. 9.2.1) while the latter (Sect. 9.2.2) by means of its growing use as a
platform building block for manufacturing valuable chemicals such as, e.g., isosor-
bide, lactic acid, sorbitan esters, lower glycols such as ethylene glycol, propylene
glycols and glycerol, hydrocarbons and aromatics.
Section 9.3 is devoted to describe the different sorbitol production methods:
hydrogenation of D-glucose, electrochemical reduction of D-glucose and biotech-
nological conversion of fructose. Chemical production of sorbitol by hydrogenation
of D-glucose (Sect. 9.3.1) is by far the most studied process, because it is the pri-
mary industrial process. After a short process description, the three main pillars
supporting further developments of this production technology are addressed: cata-
lysts, operation mode and alternative biomass raw materials. Catalyst performance
and stability play a key role in process feasibility. Hydrogenation conditions are
reviewed using catalyst nature as leitmotif, from the oldest nickel Raney catalysts to
the more recent supported Ruthenium-based catalysts. While the more usual opera-
tion mode is still the batch one (actually semi-batch because H2 pressure is kept
constant by continuous feeding of hydrogen) due to both good use of the catalyst
and good temperature control, two driving forces are pushing for use of continuous
processes: (i) avoid the catalyst removal for recycling needed in batch methods
leading to increased recovery costs and progressive catalyst deactivation, and (ii)
the growing sorbitol market requiring high productivities. Finally, the use of alter-
native biomass raw materials, mainly the direct conversion of cellulose into sorbitol,
is discussed due to its very likely future industrial importance because: (a) the grow-
ing demand for sorbitol requires increasing amounts of D-glucose which currently
is obtained from starch, a key polysaccharide for food production, and consequently
alternative non-food D-glucose containing raw materials are needed, with cellulose
being an obvious alternative because it can be obtained from the broadly available
lignocellulosic materials; and (b) boosting the use of sorbitol as a platform chemical
will require a further decrease in its manufacturing cost (for instance, for manufac-
turing fuels) and one way to lower the cost is by manufacturing sorbitol directly
from cellulose in a one-pot process, avoiding the step of hydrolysis to D-glucose
and purification of the resulting syrup.
Section 9.3.2 discusses the electrochemical production of sorbitol from
D-glucose. Bibliography (both papers and patents) devoted to this issue is negligible
in comparison with that related to chemical synthesis. However, while currently not
used industrially, it could have some industrial importance in the future taking
advantage of development of new cell designs, ion exchange membranes, anodes
and decorated porous cathodes in other fields such as in chlor-alkali industry and
fuel cells. These advancements could allow working at high apparent current
­densities while keeping both a low cell voltage and a very high sorbitol selectivity
resulting in both lower CAPEX and OPEX, thus making the electrochemical
268 J.R. Ochoa-Gómez and T. Roncal

p­ roduction of sorbitol not only competitive at some production capacities, but also
attractive due to safe issues as absence of hydrogen handling and operation at ambi-
ent pressure.
Section 9.3.3 deals with the biotechnological production of sorbitol, which could
be used at industrial scale in the future, provided that high space-time-yields, high
sorbitol concentration in culture broth and a straightforward downstream procedure
for sorbitol recovery are achieved. Production methods using the bacterium most
extensively studied, Zymomonas mobilis, are described but also those using Lactic
Acid Bacteria such as Lactobacillus plantarum and Lactobacillus casei.
Recovery and purification are essential points in industrial production.
Consequently, this key issue for the economic feasibility of a chemical process is
discussed in Sect. 9.3.4. Finally, the chapter ends with Sect. 9.4 devoted to conclu-
sions and prospects highlighting the key issues for further process development,
such as catalyst improvement and implementation of continuous processes.

9.2  Sorbitol Industrial Importance

9.2.1  Sorbitol Market

The sorbitol market was about 1830 kt in 2013 and is expected to grow at 3.6%
CAGR from 2014 to 2020, until a total of about 2337 kt affording a global market
of around USD 3.9 billion by 2020 [4]. It is sold in both liquid and solid forms, with
liquid sorbitol marketed as a 70 wt% aqueous syrup accounting for about 83% of
demand in 2013. Shares by end-use segments in 2013 were: cosmetics and personal
care, 32.8%; food, 29.4%; chemical end-use segment, 24%; and pharmaceuticals,
6%, with chemical end-use segment expected to have above average growth rates up
to 2020 as a consequence of its increasingly use as a platform chemical [5].
By application, “diabetic and dietetic food and beverages” is the largest use and
was valued over USD 400 million in 2013 (24% market share), due to the increasing
importance of functional foods with low calorie and sugar free ingredients.
Toothpaste was the second largest application segment accounting for more than
20% of the total volume, and is expected to grow significantly because the high
refractive index of sorbitol allows its use as crystals in transparent gels. Vitamin C
accounted for about 15% of the total market. Other applications include surfactants
(in the form of wetting and foaming agents, dispersants, detergents, and emulsifi-
ers), tobacco for providing a mild effect in sniff, softener and color stabilizer in
textiles, softener in leather industries and rigid polyurethane foams manufacturing.
Key players are Roquette, Cargill and Archer Daniels Midland which together
accounted for a market share of over 70% in 2013.
9  Production of Sorbitol from Biomass 269

Fig. 9.1  Sorbitol platform. Green: currently manufactured. Blue: potentially obtainable in the
future on an industrial scale

9.2.2  Sorbitol as a Platform Chemical

Chemicals that are currently manufactured or could be manufactured from sorbitol,

and their applications, are depicted in Fig. 9.1.
Interest in using sorbitol as a platform chemical was shown as early as 1928 by
IG Farbenindustrie AG [6] which developed a process for dehydrating sorbitol.
Although the chemicals obtained were not identified it is obvious in the light of cur-
rent knowledge that isosorbide was one of them. A few of the important chemicals
currently being produced from sorbitol are:
–– Vitamin C has been produced until recently through the Reichstein-Grüssner
process, wherein D-sorbitol is converted to L-ascorbic using a first fermentation
step resulting in sorbose followed by several chemical steps. Industrial produc-
tion involves the use of two first fermentation steps leading to the key intermedi-
ate 2-keto-L-gulonic acid that is finally chemically converted into vitamin C [7].
–– Isosorbide is obtained by a double dehydration of sorbitol via sorbitan using
acidic catalysts. Isosorbide is also a chemical platform with increasing industrial
interest as shown by the new manufacturing facility put into full-scale operation
by Roquette Group in Lestrem (France) with a 20,000 t.year−1 production capac-
ity [8]. Isosorbide applications encompassing the synthesis of isosorbide dini-
trate, used in pharmaceuticals as a vasodilator, dimethyl isosorbide used today
as  a solvent in cosmetics, isosorbide diesters (plasticizers) and dicarbonates
270 J.R. Ochoa-Gómez and T. Roncal

(polymer synthesis), and polymers such as PEIT (poly(ethylene terephthalate-

co-­isosorbide terephthalate)) for packaging, other polyesters for inks, toners,
powder coatings, packaging and durable goods, polycarbonates, in substitution
of bisphenol A, for durable goods and optical media, epoxy resins for paints and
polyurethanes for foams and coatings [9–12].
–– Sorbitan, as a mixture of 1,4-anhydrosorbitol, 1,5-anhydrosorbitol and
1,4,3,6-dianhydrosorbitol with the first being the predominant product, is manu-
factured by single dehydration of sorbitol but at a lower temperature than that
leading to isosorbide. It is used as raw material for manufacturing non-ionic
surfactants: sorbitan esters (marketed with the trade mark of Spans) and ethoxyl-
ated sorbitan esters (marketed with the trade name of Tweens), both useful as
solubilizers and emulsifiers in food, pharmaceuticals and cosmetics.
–– Sorbitol-based polyether polyols, which are considered to be the universal poly-
ols for rigid polyurethane foams, useful in applications such as thermoinsulation,
wood imitations, packaging, flotation materials and so on [13].
Other chemicals could be produced in the future according to increasing litera-
ture related to the production of industrially valuable sorbitol derivatives:
–– Glycols, such as ethylene glycol, glycerol and propylene glycol. Hydrogenolysis
of sorbitol resulting in ethylene glycol and propylene glycols was described as
early as 1940 [14]. This reaction proceeds better in alkaline medium [15] prob-
ably because C-C cleavage by retro-aldols reactions is a key step to lower glycols
[16]. Catalysts based in many metals, such as Ni, Cu, Ru, Pt and Pd, have been
used, with Ru showing the highest catalytic activity [17]. Reported reactions are
carried out at 423–523 K and 1.4–8 MPa [18] leading to sorbitol conversions of
70–100% with total C2-C3 glycols selectivities of 60–80%. In fact, Global
BioChem Technology Group produces industrially 1,2-propylene glycol, ethyl-
ene glycol, 1,2-propanediol and 2,3-butanediol from sorbitol since 2008 in a 200
kt.year−1 facility at Changchun (China) [19].
–– Lactic acid is used for manufacturing green solvents, such as ethyl lactate, and in
the synthesis of poly(lactic acid) (PLA), which is increasingly used for produc-
ing biodegradable packaging. Likewise, it is also a chemical platform from
which other valuable chemicals such us acrylic acid, acetaldehyde,
2,3-­
pentanedione, pyruvic acid and 1,2-propanediol can be obtained [20].
Racemic lactic acid (non-useful for PLA synthesis but useful for the other above
mentioned uses) can be obtained by alkaline hydrothermal conversion of sorbitol
in 39.5% yield at 553 K using a NaOH/sorbitol molar ratio of 2 [21].
–– Alkanes, ranging from C1 to C6, can be produced by aqueous-phase dehydration/
hydrogenation involving a bi-functional pathway in which sorbitol is repeatedly
dehydrated by a solid acid (SiO2–Al2O3) or a mineral acid (HCl) catalyst and
then hydrogenated on a metal catalyst (Pt or Pd) [22]. A yield of hexane as high
as 56% at 50 g.L−1 of aqueous sorbitol, 523  K and 4  MPa of H2 using a Pt/
NbOPO4 catalyst has been reported [23].
–– Aromatics, together with cyclic-hydrocarbons are the key components of jet fuel.
Conventional technologies need further aromatization and isomerization for
9  Production of Sorbitol from Biomass 271

Fig. 9.2  Schematic flow diagram for sorbitol production by D-glucose hydrogenation

p­ roducing these chemicals. However, a 40.4 wt% yield oil containing 80.0% of
­aromatics (69.7%: Toluene, xylenes, alkylbenzenes, indenes and naphtalenes)
and cyclic-hydrocarbons (10.30%: Methyl-cyclopentane, cyclohexane, methyl-
cyclohexane, ethyl-cyclohexane) is produced via aqueous catalytic hydrodeoxy-
genation of sorbitol over a Ni-HZSM-5/SBA-15 catalyst in a fixed-bed reactor at
593 K, WHSV of 0.75 h−1, GHSV of 2500 h−1 and 4.0 MPa of hydrogen pressure
[24].

9.3  Sorbitol Production from Biomass

9.3.1  Chemical Production of Sorbitol

Sorbitol production by hydrogenation of aqueous solutions of D-glucose using

metal-based reducing catalysts such as Ni is a well-established process [25, 26]. A
schematic diagram of a typical process is depicted in Fig.  9.2. The D-glucose,
resulting e.g. from starch hydrolysis, for hydrogenation must be of the highest
purity to prevent the catalyst becoming poisoned. Ion exchange, carbon treatment
and/or crystallization are techniques normally used to achieve the necessary purity.
In a typical process [27] the sugar syrup containing 30–60 wt% D-glucose is
reacted in a batch slurry reactor at pH 7–9 with hydrogen at high temperature (typi-
cally, 373–443 K) and high pressure (typically, 3–15 MPa) in the presence of a
suitable hydrogenation catalyst, with Raney-type nickel catalysts being the most
used until recently, in an amount of 2.5 to 12 wt% relative to the D-glucose solution.
For batch production, reaction times are in the order of 1–3 h depending on reaction
conditions. When using a Raney-nickel catalyst suitable reaction pH range is 8.0–
9.0, but a pH of about 7.5 is used in industrial production for preventing D-glucose
isomerization into mannose leading to mannitol by hydrogenation. The resulting
crude sorbitol aqueous solution is treated with activated charcoal for decolorization,
subjected to an ion exchange treatment to remove dissolved metal catalyst, and then
vacuum evaporated to obtain a 70 wt% sorbitol syrup, which may be spray dried or
further concentrated to be crystallized to obtain crystalline sorbitol. Handling and
storage after hydrogenation must be carried out according to good manufacturing
practice to ensure microbiological problems are avoided. Continuous processes
have been tested since the 1980s [28]. A D-glucose conversion of 100% and a selec-
tivity to sorbitol higher than 99% are typically obtained.
272 J.R. Ochoa-Gómez and T. Roncal

Catalysts  Nickel-based catalysts are still those mostly industrially used due to
their low costs in comparison with other suitable metal-based catalysts later herein
discussed, such as ruthenium-based catalysts. However, soon after its industrial use
its drawbacks became apparent: deactivation by sintering, leaching into the reaction
mixture leading to enhanced purification costs, and poisoning. The industry required
more stable and active catalysts. Strategies for achieving such a goal have been: (a)
Addition of promoters to Raney nickel catalysts; (b) Use of supports for nickel cata-
lysts; and (c) New catalysts, with Ru-based ones being largely preferred.
(a) Addition of Promoters to Raney Nickel Catalysts  It has been found that the
addition of some metal promoters enhances notably the activity and stability of
Raney Nickel catalysts. Thus, Court et al. [29] prepared catalysts from Ni2-xMxAl3,
(where M = Cr, Fe, Co, Cu, Mo and x ≤ 0.4) and found that the promoter metal,
except cobalt, favors aluminum retention in the catalyst and ascribed at least par-
tially the increase of the nickel activity observed to this high residual aluminum
content. The activity of Cr or Mo promoted catalysts was about twice that of the
unpromoted Raney nickel. Cobalt and copper present in small amounts did not
change the properties of the Ni-Al system.
Gallezot et al. [30] used Mo, Cr, Fe and Sn as promoters for Raney nickel cata-
lysts. No data about conversions and selectivities were reported because the study
was aimed at studying the activity of catalysts by measuring initial conversion rates.
The activities of Mo- and Cr-promoted catalysts decrease only very slightly with
recycling due to surface poisoning by cracking products formed in side reactions.
Fe- and Sn-promoted catalysts deactivated very rapidly due to leaching away from
the surface, but while iron was leached to the liquid phase tin remained in the cata-
lyst micropores. They suggested an acid-base catalyst concluding that promoters in
a low-valent state on the nickel surface act as Lewis adsorption sites for the oxygen
atom of the carbonyl group, which is then polarized favoring its hydrogenation via
a nucleophilic attack on the carbon atom by hydride ions. The same conclusion was
reached by Li et al. [31] studying the activity of a Ni–B/SiO2 amorphous catalyst,
without and with metal promoters, in the liquid phase hydrogenation of a 50 wt%
aqueous solution of D-glucose at 373 K, 4.0 MPa and 6 h in a stainless steel auto-
clave containing 1 wt% of catalyst vs. D-glucose. The addition of W, Mo and Cr
increased conversion from 30% to 35%, 42% and 49%, respectively, with the opti-
mum contents of W, Mo and Cr being 10, 5 and 10 wt%, respectively. Above said
concentrations a decrease in the activity was observed since too many Ni active sites
were covered by these oxides. Selectivities were closed to 100%.
Hoffer et al. [32] studied the influence on Raney-type Ni catalysts of Mo and Cr/
Fe as promoters in the hydrogenation of an aqueous solution of D-glucose (10 wt%)
using a three-phase slurry reactor at 4.0 MPa of H2 and 393 K. Selectivity to sorbitol
was >99%. The promoter concentrations (wt%) were different at bulk (1.4% Mo,
2.4%/3.6% Cr/Fe) and surface (8.7% Mo, 3.8%/5.9% Cr/Fe) catalyst, being essen-
tially as oxides. It was shown that promoters enhanced and stabilized the BET sur-
face area leading to an increase of the catalysts activity. The BET surface areas
increased from 56 m2 gcat−1 in the Raney nickel catalyst to 77 and 112 m2 gcat−1 in the
9  Production of Sorbitol from Biomass 273

Mo and Cr/Fe promoted catalysts, respectively. The activity of catalysts increased

accordingly, from 0.35  kg−1 s−1 to 0.50 and 0.90  kg−1.s−1, respectively. However,
even when the performance of catalyst is normalized to its surface area, the pro-
moted systems still exhibit an enhanced activity [33]. Consequently, the enhance-
ment of the reaction rate by the promoters was ascribed to the promoters being more
electropositive than Ni and acting as adsorption sites for D-glucose, which gener-
ates an ionized species that is susceptible to attack by hydrogen. Catalyst stability
was also enhanced by promoters as shown by a lower loss in activity (30% and 16%
for Mo- and Cr/Fe-promoted catalysts, respectively) than that in unpromoted cata-
lyst (48%) after 3 recycles. However, promoters were unable to keep long term
stability as shown by the continuous loss in activity, which after 5 recycles were
45.6% and 40.1% for Mo- and Cr/Fe-promoted catalysts, respectively. Mo was not
leached while Fe was severely leached (27%) while Cr was slightly leached (1.7%),
although according to authors loss in activity was not mainly due to Fe loss but to
poisoning of the active sites by D-gluconic acid, which was formed via a Cannizzaro-­
type reaction of D-glucose induced by an alkaline environment in the presence of
nickel.
(b) Use of Supports for Nickel Catalysts  Use of supports both to increase metal
dispersion and to have large exposed surface area has revealed to be a good strategy
for improving Ni activity and stability. Thus, a number of carriers such as C, SiO2,
TiO2, Al2O3, ZrO2, and mixtures thereof, have been tested. Kusserow et al. [34] stud-
ied both catalyst preparation method (precipitation, impregnation, sol-gel and tem-
plate syntheses) and support nature (SiO2, TiO2, Al2O3 and carbon) in the batch
hydrogenation of a 40–50 wt% aqueous D-glucose solution to sorbitol at 393 K and
12 MPa H2, for 5 h using a catalyst load of 1 wt% vs. D-glucose, comparing the
results with a Ni commercial catalyst (Ni(66.8 wt%)/SiO2, KataLeuna GmbH
Catalysts, Leuna, Germany). Regarding the specific activities (μg of sorbitol.g−1
nickel.s−1), they found that the activity follows the sequence Al2O3 (100–210) >
TiO2 (80–170) > SiO2 (5–50) ≈ commercial catalyst > C (0–15). Generally, cata-
lysts prepared by impregnation had a higher activity than those prepared by incipi-
ent wetness. On the other hand, the selectivity (expressed as the amount of
by-products) followed the sequence SiO2 (2.1–4.1 wt%) > C (3.4–6.3 wt%) > Al2O3
(5.5–7.5 wt%) > TiO2 (~8 wt%). The same research group [35] demonstrated the
importance of particle size for preventing Ni leaching. They prepared Ni catalysts
with different metal loadings (5, 10, and 20 wt%) by impregnation with nickel eth-
ylenediamine complexes, which had small nickel particles (mean diameter: 2–3 nm)
even for high metal loadings (20 wt%). These catalysts showed almost no nickel
leaching compared with the commercial Ni/SiO2 above described in D-glucose
hydrogenation. Unfortunately, they were slightly less active (TOF (number of
D-glucose molecules converted per second and per surface Ni site): 2 to 10 × 10−3
s−1) than the commercial nickel/silica catalyst (TOF: 14 × 10−3 s−1) and gave lower
yields to sorbitol (3–42% compared to 60%). Catalyst activity was strongly depen-
dent on pretreatment conditions performed. Calcination before reduction led to
higher conversion (19–45%) and selectivity to sorbitol (81–92%) than direct
274 J.R. Ochoa-Gómez and T. Roncal

r­eduction without calcination pretreatment (conversion: 10–16% and selectivity:

21–59%) after 5 h of reaction time, respectively. The difference was ascribed to
complete decomposition of the nickel ethylenediamine precursor being achieved
only when calcination pretreatment was performed.
Geyer et  al. [36] studied the performance of Ni catalysts supported on ZrO2,
TiO2, ZrO2/TiO2, ZrO2/SiO2 and MgO/Al2O3/SiO2, in comparison with a Ni/SiO2
catalyst, all of them prepared by precipitation. Batch hydrogenation tests were car-
ried out with a 50 wt% D-glucose aqueous solution for 4 h at 393 K and 12 MPa H2
using a catalyst concentration of 1.5 wt%. Ni contents in catalysts ranged between
39.9 and 51.2 wt%. No conversion data were reported. However, sorbitol yields
were as follows: Ni/ZrO2/SiO2 (97.8%) > Ni/TiO2 (97.5%) > Ni/ZrO2/TiO2 (96.7%)
> Ni/ZrO2 (93.2%) > Ni/SiO2 (92.4%) > Ni/MgO/Al2O3/SiO2 (83.9%), while TOFs
(10−3 s−1): Ni/TiO2 (44.8) > Ni/ZrO2/TiO2 (35.3%) > Ni/ZrO2 (28.2) > Ni/ZrO2/SiO2
(26.5) > Ni/SiO2 (18.2) > Ni/MgO/Al2O3/SiO2 (14.3). Assuming that yields between
96.7% and 97.8% were equal due to the analytical uncertainty it can be concluded
that only the TiO2 containing catalysts had a better catalytic performance than the
commercial Ni/SiO2 catalyst. Surprisingly, no correlation between the metal disper-
sion and specific hydrogenation activity was found which was ascribed to the prob-
able hydrogenation activity of the catalysts being determined not only by dispersion
but also by a metal support interaction. The stability of the metal dispersion under
reaction conditions was investigated with selected catalysts by a treatment in the
reaction mixture over a period of 100 h. Ni/ZrO2 catalyst had a higher stability of
the Ni dispersion than Ni/SiO2 due to a reduced leaching of the ZrO2 support.
High activity and stability for the continuous hydrogenation of 10 wt% D-glucose
to sorbitol in aqueous solution in a stainless steel fixed-bed reactor were reported by
Li et al. [37] using 60%Ni/AlSiO catalysts prepared by the co-precipitation method,
in which AlSiO were the composite supports with different Al2O3/SiO2 mass ratios.
They found that the catalyst with an Al2O3/SiO2 mass ratio of 4 (Ni/AlSiO-4) in the
support hydrothermically treated at 423 K exhibited the high hydrothermal stability
when the supported nickel particles were small (about 5.2 nm) and highly dispersed,
confirming the importance of particle size in catalyst stability. D-glucose conver-
sions and sorbitol selectivities were 100% between 373 and 413 K at 4 MPa of H2
pressure and a WHSV of 1 h−1. In comparison, a Ni/SiO2 catalyst (particle size
15–30 nm) deactivated quickly due to the fast aggregation of supported Ni particles.
Increase in activity was attributed to the higher H2 uptake of smaller Ni particles.
(c) New Catalysts  Due to their higher activity and better selectivity in various
hydrogenation processes, the Ni-metalloid amorphous alloy catalysts have been
studied as potential substitutes for Raney Ni catalysts. Li et al. [38] prepared a novel
skeletal Ni-P amorphous alloy catalyst (Raney Ni-P, containing 68 wt% Ni, 25 wt%
Al and 7 wt% P) by alkali leaching a Ni-P-Al amorphous alloy obtained by the rapid
quenching technique of a melting solution (1673 K) containing 48.2 wt% Ni metal,
48.7 wt% Al metal, and 3.1 wt% red P. The performance of this catalyst compared
with a Raney Ni one in the liquid phase D-glucose hydrogenation was tested under
the following experimental conditions: catalyst loading, 3 wt% vs. D-glucose;
9  Production of Sorbitol from Biomass 275

D-glucose concentration, 50 wt%; 393 K; PH2 4.0 MPa; stirring rate, 1200 rpm;
reaction time, 6.0 h. While selectivity to sorbitol was 99.5% for both catalysts, con-
versions and TOFs were 55.8% and 0.40 s−1 for Raney Ni-P catalyst, and 17.2% and
0.11 s−1 for the Raney Ni one. Authors concluded that this huge increase in perfor-
mance was apparently the result of promotion of Ni-active sites by phosphorus lead-
ing to a significantly higher density of Ni surface atoms. The amount of Ni dissolved
in the reaction mixture was less than 1.0 ppm, < 7.3 × 10−3 wt% of the initial Ni
content in the catalyst. The catalyst was used repetitively until an abrupt decrease in
the activity was observed. No significant decrease in activity was observed in the
first 5 cycles of the hydrogenation. It was also found that the deactivated catalyst
could be easily regenerated by treating it again in a 6.0 M NaOH solution, suggest-
ing the idea that the Ni leaching was not the main cause of deactivation. Unfortunately,
no data about Ni leaching with the number of reaction cycles were given.
As mentioned above in (a), Li et al. [31] reported the activity of a Ni–B/SiO2
amorphous catalyst (6.4 wt% Ni, no data about B content) in the liquid phase hydro-
genation of a 50 wt% aqueous solution of D-glucose at 373 K, 4.0 MPa of H2 and 6
h in a stainless steel autoclave containing 1 wt% of catalyst vs. D-glucose. The
amorphous catalyst exhibited much higher activity than other Ni-based catalysts,
such as the corresponding crystallized Ni–B/SiO2 catalyst, the Ni/SiO2 catalyst and
the commercial Raney Ni one. Thus, TOF was two times higher for amorphous
Ni–B/SiO2 (0.024 s−1) than for Raney Ni (0.0135 s−1). While selectivities were near
to 100%, conversions were strongly dependent on catalyst characteristics and reac-
tion conditions, ranging from 30% to 82%. Catalyst activity was kept almost con-
stant for 5 cycles and only less than 1.0 ppm Ni in the product mixture was detected
by ICP, 0.027% of the initial Ni content in the catalyst, showing the increase in
stability of the Ni–B/SiO2 amorphous catalyst. Besides the high dispersion of the
Ni–B/SiO2 amorphous catalyst, its high activity was mainly attributed to its favor-
able structural characteristics, such as the high concentration of coordinately unsat-
urated sites and the strong union between Ni active sites, and the electronic
interaction between Ni and B making Ni electron-rich. The increased electronic
density on Ni weakens the adsorption of D-glucose on catalyst surface, via the
donation to Ni of one electron pair from the oxygen of the carbonyl group, thereby
favoring the competitive adsorption of hydrogen against the C = O group. Therefore,
more hydrogen could be adsorbed on the Ni–B/SiO2 amorphous catalyst which in
turn enhanced its hydrogenation activity.
However, substitution of Ni-based catalysts by other typical hydrogenation sup-
ported noble metal-based catalysts seems to be the more promising way for further
process development from an industrial standpoint. Catalysts based on platinum,
palladium, rhodium, and ruthenium have been tested [34, 39]. Taking into account
that, on one hand, the order of activity for D-glucose hydrogenation found was Ru
> Ni > Rd > Pd, with specific activities of Ru catalysts being about 50 times higher
than those of Ni catalysts; and, on the other hand, that ruthenium is not leached
under the usual reaction conditions for D-glucose hydrogenation [39], Ru based
catalysts have been by far the more studied with the aim of designing more active
Ru-based catalysts with the lowest possible cost to compensate the high cost of Ru.
276 J.R. Ochoa-Gómez and T. Roncal

The superior performance of supported Ru-based catalysts in comparison with

other Pt-group metals for hydrogenating carbonyl groups into the corresponding
alcohols in water solutions has been discussed by Michel and Gallezot [40]. They
identified two mechanisms, both involving water. The first mechanism is related to
the interactions, via hydrogen bonds, between the C=O group adsorbed on the Ru
surface and adjacent adsorbed water molecules, lowering the energy barriers lead-
ing to an easier hydrogenation of carbonyl groups by dissociated hydrogen. The
second mechanism is that the dissociation of adsorbed water on the ruthenium sur-
face increases the surface concentration in hydrogen atoms, thus favoring the hydro-
genation reaction. They concluded that more studies are needed for determining the
respective importance of these two mechanisms and, mainly, to confirm the role of
water as a source of hydrogen.
Hoffer et al. [33] studied carbon supported Ru catalysts as better alternatives for
Raney-type Ni catalysts. They found that the Ru/C catalysts had higher activities,
Ru does not leach, and the activity is proportional to the Ru surface area and inde-
pendent of the preparation method. A selectivity higher than 98% (no conversions
given) was obtained in a three-phase slurry reactor at 393 K and 4.0 MPa hydrogen
pressure using a 10 wt% D-glucose solution. All Ru catalysts were at least two times
more active than the Ni catalysts per kg of catalyst. The strong stability of Ru-based
catalysts in comparison with Ni-based ones was unequivocally demonstrated by
Kusserow et al. [34] by operating a continuous process for 1150 h time on stream
using a Ru(0.47 wt%)/Al2O3 catalyst, as later discussed in the next section. No Ru
losses were detected.
Guo et al. [41] prepared a Ru-B amorphous alloy catalyst in the form of ultrafine
particles. The Ru-B catalysts were more active than that of Co-B, Ni-B amorphous
catalysts as well as Raney Ni catalysts for the D-glucose hydrogenation. The Ru-B
amorphous catalyst exhibited higher activity than its corresponding crystallized
Ru-B and pure Ru powder catalysts, showing the promoting effects of both the
amorphous structure and the electronic interaction between the metallic Ru and the
alloying B. A 50 wt% aqueous solution of D-glucose was converted in 2 h into sor-
bitol in 95.1% conversion and 100% selectivity to sorbitol using an amorphous
Ru88.9B11.1 (amounts by weight) catalyst (1 wt% vs. D-glucose) at 353 K and 4.0 MPa
of H2 pressure. Under same conditions, Ni-B, Co-B, crystallized Ru-B, pure Ru, and
a commercial Raney Ni (3 wt% vs. D-glucose) catalysts gave conversions of 58.8%,
47.8%, 23.1%, 22.2% and 16.7%, respectively.
Catalyst development until 2013 has been reviewed by Zhang et al. [42], includ-
ing Ru catalysts. Studies have been devoted to determine the influence on Ru-based
catalyst activity of preparation methods, supports (SiO2, γ-Al2O3, MCM-41) and Ru
precursors (Ru acetate and Ru trichloride). After analyzing the bibliography Zhang
et al. concluded there are three key points that can serve for further development: (a)
pore and BET surface area (related to catalyst dispersion – size): catalyst perfor-
mance increases as both increase; (b) preparation methods, having an important
influence of catalyst performance; and (c) catalyst life: a critical issue due to the
high cost of Ru, so that studies should be focused on improving catalyst life through
developing effective Ru carriers and cheap catalyst regeneration methods.
9  Production of Sorbitol from Biomass 277

Table 9.1  Hydrogenation of D-glucose to sorbitol over ZSM-5, Ru/C and Ru/ZSM-5 catalysts
[43]
SBET Db Acid Conversion Sorbitol TOF
Catalystsa (m2.g−1) (nm) (mmol.g−1) Dispersion (%) yield (%) (h−1)
Ru(5 wt%)/C – – – 0.23 81.3 61.7 13
ZSM-5 358 0.55 3.6 – 30.2 1.2 2
Ru/ 383 0.52 3.1 0.54 99.6 99.2 32
ZSM-5-TF
Ru/ 339 0.26 2.1 0.46 98.4 95.8 15
ZSM-5-MS
Ru/ 345 0.66 3.4 0.48 98.8 97.4 18
ZSM-5-AT
Hydrogenation conditions: 25 wt% D-glucose solution, 2 h, 4.0 MPa H2, 393 K, catalyst concen-
tration: 4 wt% vs. D-glucose
a
SiO2/Al2O3 ratio: 38, except for Ru/ZSM-5-AT: 30
b
Average pore diameter

Thus, Guo et al. [43] studied the influence of preparation method and the support
nature in the hydrogenation of D-glucose to D-sorbitol over Ru/ZSM-5 catalysts.
An incipient wetness impregnation method and a one-step template-free process
(Ru/ZSM-5-AT catalyst) were used. For the conventional impregnation method
both mesoporous ZSM-5 supports created by desilication in alkaline medium (Ru/
ZSM-5-AT catalyst) and commercial microporous (Ru/ZSM-5-MS catalyst) were
tested. The hydrogenation of a 25 wt% D-glucose aqueous solution was performed
for 2 h at 4.0 MPa of H2 and 393 K using a catalyst concentration of 4 wt% relative
to D-glucose, and compared with those obtained using the HZSM-5 zeolite and a
commercial Ru(5 wt%)/C as catalysts. Catalyst characteristics and hydrogenation
results are given in Table 9.1.
As shown in Table 9.1, Ru catalyst prepared by the one-step template-free pro-
cess was highly dispersed in the ZSM-5 framework and exhibited higher catalytic
activity than others, with conversions and yields exceeding those obtained with
catalysts prepared by the conventional impregnation method with microporous or
desilicated ZSM-5 supports, and also with the support alone and the commercial
Ru/C catalyst. The catalytic activity increased with increasing Ru loading up to 4.1
wt% but, more importantly, the formation of undesirable by-products, i.e. D-fructose
and D-mannitol, was significantly minimized with increased metal loading, as
shown by the increase in sorbitol selectivity from 15.9% when the content of Ru
was 1.2 wt%, to 99.6 when it was 4.1 wt%. From an industrial standpoint two fea-
tures shown by Ru/ZSM-5-TF are very significant. On one hand, it leads to a selec-
tivity of 99.6%, meaning a simpler purification procedure, while that obtained with
Ru/ZSM-5-MS and Ru/ZSM-5-AT catalysts were 97.4% and 98.6% respectively.
On the other hand, it showed high stability against leaching and poisoning and could
be reused several times, meaning a longer life span than that of the other catalysts.
The catalyst regeneration proved to be key for maintenance of catalyst performance.
When Ru/ZSM-5-TF was only dried at 393  K in air, and reused without further
278 J.R. Ochoa-Gómez and T. Roncal

treatment, the sorbitol yield decreased from 99.2 to 89.2% after five runs. However,
when the spent catalyst was washed with water and subsequently washed with etha-
nol or acetone three times each, there was no obvious loss of the catalytic reactivity.
Consequently, the deactivation was probably due to the accumulation of organic and
inorganic species adsorbed on the catalyst surface. These results clearly showed that
Ru/ZSM-5-TF was effective for the hydrogenation of D-glucose and had good sta-
bility during the reaction process. Their excellent catalytic behavior and stability
were ascribed to the extensive dispersion of the Ru particles, the strong interaction
between the Ru species and the ZSM-5 support, as well as its suitable surface
acidity-­basicity balance. That is to say, the catalytic performance is strongly depen-
dent on the catalyst preparation method and its method of maintenance.
Mishra et al. [44] studied the performance of HY zeolite supported ruthenium
nanoparticles catalysts using a HY zeolite with Si/Al ratio = 80 (referred to as
HYZ). Catalysts were prepared by using conventional impregnation–reduction
method using NaBH4 in ethanol as reducing agent. The activity tests were carried
out with a 20 wt% aqueous solution of D-glucose for 20 min at 393 K and 5.5 MPa
H2 using a Ru(1 wt%)/HYZ catalyst in a concentration of 2.5 wt% relative to
D-glucose and comparing results with Ru(1 wt%)/NiO-TiO2 and Ru(1 wt%)/TiO2
catalysts previously studied by the same research group [45] and a Ru(5 wt%)/C
commercial catalyst. Metal dispersions were 23.6, 8.6, 4.4 and 6.7 for Ru(1 wt%)/
HYZ, Ru(1 wt%)/NiO-TiO2, Ru(1 wt%)/TiO2 and Ru(5 wt%)/C, respectively.
Hydrogenation tests were carried out with a 20 wt% aqueous solution of D-glucose
for 20 min at 393 K and 5.5 MPa H2 with a catalyst level of 2.5 wt% relative to
D-glucose. Ru(1 wt%)/HYZ showed a higher conversion (19.4%), selectivity
(97.6%) and TOF (1275 h−1) than the other catalysts. TOF was 1.08-, 1.19- and 5.4-­
fold higher that with Ru(1 wt%)/NiO-TiO2, Ru(1 wt%)/TiO2 and Ru(5 wt%)/C cata-
lysts, respectively. After further optimization a 100% conversion and a 98.7%
selectivity were achieved under the same experimental conditions by increasing
reaction time up to 2 h. It was concluded that the acidity (mild acidity) of zeolite
support plays an important role in increasing both the selectivity and the activity to
sorbitol.
Aho et al. [46] have studied the influence of metal dispersion in the semi-batch
hydrogenation of a 18 g.L−1 D-glucose aqueous solution at 393  K and 1.9  MPa
hydrogen pressure over several Ru/C catalysts. Those with ruthenium particle sizes
between 1.2 and 10 nm were investigated. All were active with selectivity to sorbitol
being 87–96%, except for the 10  nm catalyst (28.8%). TOF was maximum for
ruthenium nanoparticles of 2.9 nm.
Lazaridis et  al. [47] reported the hydrogenation/hydrogenolysis of D-glucose
over Pt and Ru catalysts supported on activated micro/mesoporous carbon (AC), at
1.6  MPa and 453  K and low D-glucose concentration (2.7 wt%). The effects of
metal content (1–5 wt%), method of metal pre-treatment/reduction (H2 at 623 K or
NaBH4) and reaction time (1–12 h) were studied. All the Ru and Pt/AC catalysts
were very active (conversion ≥97%) with the Pt/AC catalysts being also very selec-
tive toward sorbitol (selectivity ≥90%) irrespective of metal content, method of
reduction and reaction time. However, the maximum selectivity was 95%, a low
9  Production of Sorbitol from Biomass 279

value in comparison with that industrially desired due to the high temperature used
causing hydrogenolysis resulting in various lower sugar alcohols, such as
1,2,5,6-hexanetetrol, arabinitol, threitol, glycerol and 1,2-propanediol. Conversely,
the Ru/AC catalysts exhibited a wide range of sorbitol selectivity values (55–93%)
depending on metal loading, method of reduction and reaction time. Thus, the 93%
selectivity was only achieved with a metal content of 3 wt% and a reaction time of
1 h using NaBH4 as a metal reducing agent. Under the same experimental condi-
tions, the sorbitol selectivity increased when the reaction time decreased from 12 h
to 1 h and when Ru loading increased from 1 wt% to 3–5 wt%. Moreover, the sor-
bitol selectivity was higher for the hydrogen reduced catalysts than for those reduced
with NaBH4. The higher selectivity of the Pt/AC catalysts toward sorbitol could be
related to the abundance of well formed, single crystal Pt nanoparticles (2–6 nm)
compared to the less crystalline Ru/AC catalysts. Thus, using HRTEM measure-
ments the co-existence of small crystalline metallic Ru nanoparticles together with
amorphous Ru(O)xδ+ species within relatively larger aggregates (20–30 nm) was
identified. Based on this Ru morphology, the authors proposed a model of formation
of protonic acid sites Ru(O)xH+, as well as of protons released in the aqueous solu-
tion from the Ru(O)xH+ species, in the presence of hydrogen. Thus, Ru(O)xH+ spe-
cies are formed by spillover of the formed hydrogen atoms by the dissociative
adsorption of molecular H2 on the metallic Ru(0) nanoparticles to adjacent Ru(O)xδ+
species, with simultaneous electron transfer from the H atoms to the positively
charged Ru-O species. The authors proposed that hydrogenation of glucose to sor-
bitol occurred by the dissociative adsorption of molecular H2 on the metallic Ru(0)
nanoparticles and the subsequent addition of the pair of hydrogen atoms to the
hemiacetal group of glucose, leading to the cleavage of the C-O bond and the forma-
tion of the hydroxyl group, as it generally accepted [18, 42]. However, the protonic
Ru(O)xH+ acidic sites as well as the released H+ in solution induced the conversion
of the sorbitol to 1,2,5,6-hexanetetrol via dehydration and hydrogenation reactions.
Likewise, this acidic environment induced hydrogenolysis of the C3-C4 bond of
sorbitol resulting in glycerol, which in turn can be then converted to 1,2-­propanediol
via dehydration and hydrogenation, while this molecule can be also formed via
hydrogenolysis of the C3-C4 bond of 1,2,5,6-hexanetetrol. According to this mech-
anism, the selectivity to sorbitol of the Ru/AC catalysts could be strongly improved
by means of a catalyst preparation method preventing the formation of Ru(O)xδ+
species.
The good performance and industrial possibilities of Ru-B amorphous alloys as
catalysts for sorbitol synthesis from D-glucose previously shown by Guo et al. [41]
and discussed in this section have been recently confirmed by Wang et al. [48] as
well as the importance of a suitable carrier. They synthesized a highly dispersed
Ru–B amorphous alloy catalyst by loading Ru–B amorphous alloy nanoparticles
(NPs) on a matrix consisting of a highly ordered mesoporous silica nanospheres
(MSNS) externally covered by methyl groups but internally grafted by aminopropyl
groups. The amino and methyl groups acted synergistically as effective functional-
ities for highly dispersing Ru–B NPs within the pore channels of the mesoporous
host. Such catalyst was used in the liquid-phase D-glucose hydrogenation to
280 J.R. Ochoa-Gómez and T. Roncal

D-sorbitol under the following conditions: 50 wt% D-glucose aqueous solution,

3.0 MPa H2, 373 K, 3 wt% catalyst relative to D-glucose, reaction time of 80 min.
The amorphous Ru89–B11/NH2&CH3-MSNS catalyst with a Ru load of 4.4 wt% pro-
vided a 100% D-glucose conversion with 100% selectivity to sorbitol while conver-
sions were 4.5%, 23% and 22% for a commercial Raney Ni catalyst, a commercial
Ru(5 wt%)/C catalyst and the crystallized Ru89–B11/NH2&CH3-MSNS catalyst,
respectively. The amorphous catalyst kept its activity for 9 cycles, showing good
potential for industrial use.
These researchers also compared the activity of this catalyst with those of amor-
phous Ru-B alloys deposited on pure MSNSs and mono-functionalized MSNSs
(NH2-MSNS and CH3-MSNS). Activities decreased in the order Ru–B/NH2&CH3-­
MSNS > Ru–B/NH2-MSNS > Ru–B/CH3-MSNS > Ru–B/MSNS, i.e. in the order in
the same order than the particles dispersion as shown by the surface areas. Moreover,
TEM images showed as Ru-B NPs were located differently depending on the matrix
type. Thus, in the Ru–B/MSNS catalyst Ru–B nanowires within the channels of
MSNSs were observed probably due to the agglomeration of the Ru–B NPs located
inside the pore channels together to form nanowires during the preparation process.
Furthermore, a portion of Ru–B NPs was situated on the external surface of MSNS
because some particles were larger than the pore size of MSNS. For Ru–B/NH2-­
MSNS, nanowires were absent, which was attributed to the coordination effect of
amino groups in the interior of pores on Ru3+ ions. Reduction of these coordinated
Ru3+ ions with BH4− led to the location of Ru–B NPs inside the mesoporous chan-
nels. However, some large NPs were still observed on the external surface. Large
NPs located on the external surface were absent for Ru–B/CH3-MSNS due to the
high hydrophobic external surface, but Ru–B nanowires appeared between the walls
of MSNSs, which was ascribed to the absence of stabilizing groups inside the pore
channels. Moreover, Ru–B NPs-aggregates were observed in Ru–B/CH3-MSNS
due to the fall of Ru–B NPs from the pore channels of support. For Ru–B/NH2&CH3-­
MSNS no aggregated Ru–B NPs were observed, which was explained as due to
both the stabilizing effect of amino groups on Ru3+ ions and the high hydrophobic
external surface deriving from methyl groups. Consequently, the superior activity of
the amorphous Ru89–B11/NH2&CH3-MSNS catalyst was attributed to this difference
in Ru-B location. The location of non-aggregated Ru–B NPs inside the pore chan-
nels together with the increased D-glucose concentration in the channels owing to
the microreactor effect [49, 50] would lead to an increase in the collision frequency
between reactants and Ru active sites and, therefore, to an increase in catalyst activ-
ity. As pointed out by authors, the results of their study demonstrate that region-­
selectively functionalizing the surface of ordered mesoporous materials with diverse
groups allows for molecular-level fine-tuning of catalytic performance of the intro-
duced guest materials.
Dabbawala et al. [51] studied the aqueous phase hydrogenation of D-glucose to
sorbitol using Ru supported on amine functionalized nanoporous hypercrosslinked
polystyrene polymer based catalysts (Ru/AFPS) prepared by simple impregnation-­
chemical reduction method. The interest of this study relies also in that it represents
one of the first studies about utilization of porous polymeric supports to stabilize Ru
9  Production of Sorbitol from Biomass 281

nanoparticles in D-glucose hydrogenation. Catalyst performance tests were carried

out using a 4 wt% D-glucose aqueous solution at 353–383 K and 3.5–6.0 MPa H2
for 60 min, with a catalyst concentration of 5 wt% relative to D-glucose. At 373 K
and 5.5 MPa H2, the catalytic activity of Ru(5 wt%)/AFPS catalyst was much higher
than those of Ru(5 wt%)/Polystyrene, Ru(5 wt%)/C, Ru(5 wt%)/TiO2, Ru(5 wt%)/
Al2O3 and Ru(5 wt%)/SiO2 as shown by a TOF of 230 h−1, about 2.5-fold higher
than that of the unfunctionalized support (Ru(5 wt%)/PS) and about 1.3 times higher
than the second more active catalyst (Ru(5 wt%)/C). Likewise, the Ru(5 wt%)/
AFPS catalyst led to a 67.6% sorbitol yield, the higher of all catalysts studied, with
a selectivity to sorbitol of 98%, equal to that obtained with Ru(5 wt%)/PS and
higher than those obtained with the remaining catalysts. This higher catalytic per-
formace was attributed to the functional amine groups present in AFPS polymer
surface which stabilize Ru nanoparticles and enhance the dispersion of Ru nanopar-
ticles, provide a better wettability thereby improving the mixing with water com-
pared to Ru/PS catalyst. On the other hand, the AFPS catalyst nanoporous structure
with its wide range of pores was considered to be of key importance. Thus, the
presence of micropores in AFPS catalyst leads to a better control of Ru nanoparti-
cles size, while the presence of mesopores allows substrate molecules to easily
approach the Ru active metal center without diffusion limitations. Ru(5 wt%)/AFPS
was reused up to five times without significant loss in activity and selectivity. This
stability was attributed to the strong interaction between support and active Ru
species.
Operation Mode  It is worth mentioning that although most production methods
described in the literature are referred to as batch processes, most of them operate at
constant hydrogen pressure such that hydrogen is continuously fed as it is consumed
which makes them semi-batch processes. Batch (semi-batch) processes have a good
use of the catalyst as well as good temperature control as main advantages which
explain their wide industrial use for sorbitol production. However, the catalyst
removal for recycling is a great disadvantage leading to increased costs and progres-
sive catalyst deactivation. The importance of the reactor type on catalysts stability
in D-glucose hydrogenation has been highlighted by Doluda et al. [52] who have
shown that the use of common batch and shaker type reactor systems results in high
losses of the initial catalysts, due to catalysts grinding on the reactor impeller and
reactor walls and during the catalysts separation and washing. Besides, selectivity is
also negatively affected by formation of gluconic acid due to the presence of trace
oxygen. On the other hand, the continuous increase in sorbitol demand requires
more productive processes. Consequently, the development of continuous methods
is imperative due to the higher space-time yields and the absence of an expensive
catalyst separation step.
Boyers and Flushing [53] disclosed a method for the catalytic hydrogenation of
carbohydrates in the presence of suspended finely divided ruthenium carrier cata-
lysts, method which could be continuously carried out although no example with
this operation was given. However, the use of a finely divided catalyst suspension
282 J.R. Ochoa-Gómez and T. Roncal

catalyst involved a complicated filtration separation and made the quantitative

recovery of the precious catalyst difficult.
Lepper and Schütt [54] described a continuous method for the hydrogenation of
carbohydrates in the presence of a catalyst solid bed of Ru(2 wt%)/C in lumps
(cylindrical form, diameter 2 mm, length 2 to 5 mm). The process involved the use
of two series-connected reactors with a capacity of 4.2 liters and each having an
inside diameter of 70 mm. Working a 423 K and a hydrogen pressure of 25.0 MPa,
10,400 kg of a 50 wt% aqueous solution of D-glucose were converted into sorbitol
in 1550 h (1.7 kg-D-glucose.h−1.m−2) with a conversion of 99% and a selectivity of
99.7%. The catalyst activity was 0.30 mol-D-glucose.h−1.gRu−1, corresponding to a
turnover frequency (TOF) of 30.3, 3.4 times larger than that obtained using a ruthe-
nium catalyst in suspension.
Selectivity of Lepper and Schütt process is within the desired one but conversion
while being as high as 99% was below the target of 100% for decreasing purifica-
tion costs. Looking for this goal Déchamp et  al. [55] studied the continuous
D-glucose hydrogenation to sorbitol in a trickle-bed reactor in the presence of
kieselguhr-­supported nickel catalysts with cocurrent downflow mode in the tem-
perature range 343–403 K and in the pressure range 4–12 MPa, but catalyst perfor-
mance decreased with time due to continuous leaching of both nickel and support.
Similar catalyst deactivation was observed by Tukac [56] with same reactor type in
the temperature range 388–438 K at 0.5–10 MPa starting from a 40 wt% aqueous
solution of D-glucose and using commercial supported nickel catalysts.
Ruthenium-based catalysts are a clear alternative to nickel for achieving 100%
conversions but process conditions must be carefully chosen to avoid deactivation
as shown by Arena [57] who used Ru/Al2O3 catalysts that became deactivated
because of the presence of iron and sulfur impurities and because the physical prop-
erties of the alumina support were modified. To overcome this problem, Gallezot
et al. [58] studied the same process at 373 K under 8 MPa of hydrogen but using
ruthenium catalysts (1.6–1.8 wt%) supported on active charcoal pellets (cylinders of
0.8 mm-diameter for minimizing internal diffusion limitation) instead of nickel
catalysts, with the metal under the form of 1-nm particles homogeneously distrib-
uted throughout the support. The reactor was a trickle-bed one of stainless steel
(length, 330 mm; internal diameter, 15.8 mm; internal volume, 60 cm3) and the feed
was a 40 wt% aqueous solution of D-glucose (0.17 kg-D-glucose.h−1.m−2).
D-glucose conversion was 100% and sorbitol selectivity 99.2% while catalyst activ-
ity was 1.1 mol-D-glucose h−1 gRu−1, corresponding to a TOF of 111.2, 3.67 times
larger than that reported by Lepper and Schütt with the Ru(2 wt%)/C in lumps [54]
and under milder conditions, very likely due to the lower catalyst particle size and
its better dispersion on the support. It was found that selectivity strongly depended
on residence time, with the highest one obtained just when 100% conversion was
achieved. Above this point, mannitol was formed by epimerization of sorbitol, as
well as iditol and galactitol (both by C-2 and C-4 epimerization of sorbitol, respec-
tively) and arabitol (by hydrogenolysis of sorbitol). The catalyst activity was stable
over several weeks and no leaching of ruthenium was detected.
9  Production of Sorbitol from Biomass 283

Longer tests on continuous hydrogenation of D-glucose to sorbitol have been

reported by Kusserow et al. [34] who in a study took more than 1100 h time on
stream (TOS) and investigated the deactivation of an industrial Ni(66.8 wt%)/SiO2
catalyst (1100 TOS) and a Ru(0.47 wt%)/Al2O3 catalyst (1150 TOS) with a particle
size of 1.9 nm. Experiments were carried out in a trickle bed reactor (Catatest, Vinci
Technologies, France, no dimensions data provided), under the following condi-
tions: 40 wt% D-glucose solution was delivered at a flow rate of 40 mL.h−1 to the
reactor packed with 40 mL of catalyst, H2 flow was 23 L.h−1 at a pressure of 8.0
MPa, reaction temperature was 353 K. At the beginning of the experiment, the con-
version of D-glucose was 99.3% for the nickel catalyst and 99.9% for the ruthenium
catalyst. Selectivity to sorbitol was 98.8% and 99.0%, respectively and remained
unchanged during the experiments for both catalysts. The conversion dropped below
98% after 770 h TOS for the Ni catalyst, and after 1080 h TOS for the Ru catalyst.
After 1100 h TOS for Ni catalyst and 1150 h for Ru catalyst, conversions were 95%
and 97%, respectively. No leaching of Ru could be detected by ICP-OES (detection
limit 80 μg.L−1). Conversely, the Ni level in the product production ranged between
0.005 and 0.05 wt% in the first 250 h TOS, while it was constant (~0.008 wt%) over
the remaining reaction time. The nickel content of the catalyst decreased by ~25%
(from 66.8% to 50.5 wt%) during the experiment with a visible disintegration of the
support, not surprising for silica under hydrothermal conditions. The BET surface
decreased by 60% (from 180 to 85  m2.g−1), the fraction of mesopores increased
while that of micropores decreased. It was concluded that the permanent drag-out of
nickel and silica with the product solution was caused by the hot and acidic reaction
media (pH 4 in the product solution) and the chelating properties of D-glucose and
sorbitol. The small deactivation observed in Ru catalyst was attributed to poisoning
by metal impurities leaching out of the working material (stainless steel) of the
reactor.
Aho et al. [59] studied the continuous hydrogenation of D-glucose over carbon
supported ruthenium catalysts in trickle flow reactors (12.5 mm inner diameter and
120 mm length) operated in co-current mode at 402 K and 2 MPa of hydrogen pres-
sure. Long-term (100 h) stability testing of a commercial Ru/C catalyst and Ru/
nitrogen doped carbon nanotubes (NCNT) catalysts was performed. It was found
that Ru/NCNT worked without significant deactivation for more than 100 h, while
the commercial catalyst underwent substantial deactivation within this period. The
increased stability of the NCNT supported Ru catalyst was ascribed to stronger sup-
port–Ru interactions as compared to the Ru/C catalyst. No ruthenium leaching was
observed for the Ru/NCNT catalyst, while leaching from the commercial activated
carbon catalyst was found to be significant. Catalyst activities were similar to those
reported by Gallezot et al. [58] but this process needs further optimization because
maximum D-glucose conversion and selectivity were low relative to the industrial
requirements: 91.4% and 98.2%, respectively, with mannitol and fructose as the
main impurities. Likewise, D-glucose concentration in the aqueous feed was very
low, 3.5 wt%, 11–14.5-fold less than in current industrial batch processes, leading
to an economic penalty in the evaporation step needed for achieving the sorbitol
concentration (70 wt%) in the commercial product. Conversion could be increased
284 J.R. Ochoa-Gómez and T. Roncal

by increasing both the hydrogen pressure and the residence time, but care must be
taken with the increase of the latter because could lead to a higher amount of
impurities.
Alternative Biomass Raw Materials  Although D-glucose is the only raw material
industrially used for production of sorbitol, it is obvious that a large cost reduction
could be achieved by starting directly from a D-glucose source that is directly con-
verted into sorbitol in a one-pot reaction. The idea is to carry out simultaneously the
hydrolysis of the D-glucose-containing polysaccharide and the hydrogenation of
the free D-glucose, a process requiring the combined use of both acid and metal
catalysts, for hydrolysis and hydrogenation, respectively.
Starting directly from lignocellulosics would be an ideal scenario for economic
production of sorbitol. However, due to the compositional complexity of this bio-
mass, selectivity is not high and a mixture of sugar alcohols is obtained leading to
enhanced separation costs. Thus, milled Japanese cedar wood chips containing 40.9
wt%, 24.8 wt%, and 33.4 wt% of cellulose, hemicellulose and lignin, respectively,
were directly converted into sugar alcohols using 4%Pt/C (93 wt% relative to chips
mass) as a catalyst in water without any acid catalysts at 463 K under a H2 pressure
of 5 MPa [60]. Total sugar conversion was 94.1% in 16 h while sugar-alcohol yields
were 36%, 12.6%, 2.3%, 6.9% and 4.3% for sorbitol, mannitol, galactitol, xylitol
and arabitol, respectively. Total sugars conversion decreased to 44% when unmilled
chips were used, showing the huge importance of both surface area, as the hydroly-
sis step to sugars is a heterogeneous process, and the decrease of the cellulose crys-
tallinity due to the milling.
Therefore, attempts have been focused on polymeric D-glucose precursors, i.e.,
starch and cellulose. Cellobiose (4-O-β-D-glucopyranosyl-D-glucose) has been
also converted directly to sorbitol, but it is more a simple model compound resem-
bling cellulose because it is the repeat unit of cellulose ((1,4-O-β-glucopyranosyl)n-­
1-­
D-glucose). Consequently, its conversion will not be discussed herein. Readers
can, e.g., consult references [61, 62].
Starch  The simultaneous hydrolysis and hydrogenation of starch has been known
for a long time, Thus, in 1950 Hartstra et al. [63] reported a process in which a sor-
bitol yield of 60% was obtained in 75  min by hydrogenating a 37.5 wt% starch
water suspension in the presence of 0.01–1 wt% relative to the amount of starch of
a Lewis acid such as magnesium chloride, nickel sulfate or stannous chloride, and a
nickel-kieselguhr hydrogenation catalyst (10 wt% based on starch amount) under a
hydrogen pressure of 5–20 MPa at a temperature of 433–473 K. The sorbitol was
purified by crystallization to 97% purity.
A much more effective process than that of reference [63] for sorbitol production
from starch was developed by Jacobs and Hinnekens [64]. Aqueous 10–30 wt%
corn starch was 100% batch-wise converted within 1 h into sorbitol with a selectiv-
ity higher than 95% at 403–453 K and 5.5 MPa of H2 using Ru (3 wt%) supported
on H-USY zeolite as a catalyst. The amount of Ru metal was 0.12–0.36 wt% rela-
tive to starch. The acidic zeolite support catalyzes the starch hydrolysis to D-glucose
9  Production of Sorbitol from Biomass 285

which is then hydrogenated to sorbitol on the ruthenium surface. Catalyst was

17-fold recycled with no loss in activity.
Some processes have also been developed starting from starch hydrolysates
mainly composed of D-glucose but also containing D-glucose oligomers.
Hydrogenation is carried out with both nickel [65] and ruthenium catalysts [66]
under the same conditions used starting from D-glucose. However, a mixture of
sorbitol, maltitol and hydrogenated oligomers, such as maltotriitol, is produced
boosting the sorbitol separation and purification costs. In fact, hydrogenated starch
hydrolysates are marketed as such and used as sweeteners and humectants.
Cellulose  Cellulose is an especially attractive alternative raw material due to its
huge abundance and because it does not interfere with the food chain, unlike starch.
The hydrolytic hydrogenation of cellulose to sorbitol has been reviewed by Van de
Vyver et al. [67], Zhang et al. [42], Yabushita et al. [68] and Li et al. [69]. The reac-
tion consists of two steps: (i) hydrolysis of cellulose to D-glucose, and (ii) hydroge-
nation of glucose to sorbitol, with the step (i) being the limiting step of the overall
reaction. Consequently, it is not strange that, on one hand, efforts have been focused
on developing catalytic systems with an acidic component able to speed up the
hydrolysis of cellulose, and, on the other hand, the performance of heterogeneous
catalysts is very dependent on the presence of acidic functional groups on the sup-
ports. An obvious strategy is that used by Balandin et al. in its pioneering work [70,
71], consisting of using mineral soluble acids together with a hydrogenation catalyst
such as Ru/C. They reported a total 82% yield of mannitol and sorbitol (no specific
sorbitol yield was reported) under 7 MPa of hydrogen and 433 K in 2 h in the pres-
ence of sulfuric acid. Although this strategy has been followed by several research
groups [72, 73], using both mineral acids and heteropolyacids together with hydro-
genation catalysts containing Pd, Pt or Ru, soluble acids result in a more difficult
separation procedure, generate a large amount of waste sludge in the acid neutral-
ization step and also can cause corrosion of the reactor.
Therefore, the use of heterogeneous catalysts combining a metal hydrogenation
catalyst and an acidic support is being investigated as a more industrially efficient
and sustainable alternative. Thus, microcrystalline cellulose (0.8 wt%) was hydro-
genated into sorbitol in a 25% yield in 24 h using Pt(2.5 wt%)/γ-Al2O3 as a catalyst
at 5 MPa and 463 K [74, 75]. The ability of the catalyst for hydrolyzing cellulose to
D-glucose was thought to the due to the acidic surface of the support. Also, the
metal seems to participate directly in the cellulose hydrolysis by increasing the H+
concentration by heterolytic H2 dissociation on its surface [76]. Han and Lee [77]
reported one-pot conversion of cellulose to sorbitol using catalyst combining metal
nanoparticles (Ni, Pd, Pt or Ru) and sulphonic groups on the surface of an activated
charcoal (M/AC-SO3H) support in a neutral aqueous solution. Using a 4.2 g.L−1
aqueous suspension of ball-milled cellulose and a 40 wt% catalyst load relative to
cellulose, a 95% cellulose conversion and a 71.1% sorbitol yield were obtained at
438 K in 36 h with Ru(10 wt%)/AC-SO3H. The Ru-to-S ratio and the metal load
were shown to be important parameters for effective conversion. No deactivation
was observed even after 5 repeated reactions. Pt/AC-SO3H showed a comparable
286 J.R. Ochoa-Gómez and T. Roncal

catalytic activity to Ru/AC-SO3H while Pd/AC-SO3H and Ni/AC-SO3H gave poorer

sorbitol yields.
Water has been reported as a key solvent because experiments with Ru/C, i.e.
with a non-acidic support, as a catalyst as well as substitution of water by other
solvents have shown the need of hot water for hydrolyzing cellulose which is
thought to be due to the in situ production of H+ [78]. Nevertheless, the key role of
acidic supports has been highlighted by Deng et al. [79] who found that Ru sup-
ported on carbon nanotubes (CNT) was the best catalyst for obtaining sorbitol in a
36% yield from cellulose. By using NH3-TPD and H2-TPD characterizations the
authors concluded that plenty of acid sites and unique hydrogen species over the
Ru/CNT were important for sorbitol formation.
The insolubility of cellulose in the aqueous reaction media means that reaction
proceeds in heterogeneous phase and consequently cellulose particle size and crys-
tallinity are important factors as shown by Ribeiro et al. [80] who using Ru/AC cata-
lysts obtained a cellulose conversion of 36% in 5 h with a sorbitol of 40% when
microcrystalline cellulose was the raw material, while a conversion of 90% and a
sorbitol selectivity close to 80% were obtained if the catalyst was ball-milled
together with cellulose. The catalyst showed excellent stability after repeated use.
Disruption of cellulose crystallinity can be achieved by several methods such as
ball-milling, rod-milling and planetary ball-milling in less than 1 h while jet-milling
is not effective [81].
All reported processes starting from cellulose lead to sorbitol yields below 70%,
with selectivities generally well below 80%. Typically, lower alcohols, such as xyli-
tol, glycerol, propylene glycol, ethylene glycol and methanol, resulting from hydro-
genolysis of sorbitol and soluble oligomers are also formed, together with some
amounts of mannitol, coming from hydrogenation of mannose which in turn in
formed by epimerization of D-glucose, and sorbitan and isosorbide, both resulting
from sorbitol dehydration. Consequently, cellulose is still far away of being a good
raw material for industrial production of sorbitol due, on one hand, to its extremely
low water solubility and low reactivity leading to a low productivity, and, on the
other hand, to the low sorbitol selectivity obtained so far relative to the industrial
processes currently used.
To overcome these drawbacks attempts have been focused on using ionic liquids
(IL) as reaction media because a number of them solubilize cellulose [82] by break-
ing the hydrogen bonds preventing cellulose solubilization, thereby increasing the
cellulose hydrolysis rate and consequently the productivity of the process. For
example, 1-butyl-3-methylimidazolium chloride (BMImCl) dissolves up to 25 wt%
of cellulose, while 1-allyl-3-methylimidazolium chloride and 1-ethyl-3-­
methylimidazolium acetate dissolve 14.5 wt% and 16 wt% at 353  K and 363 K,
respectively. Thus, Ignatyev et al. [83] demonstrated that a combination of a hetero-
geneous Pt or Rh catalyst and a homogeneous Ru catalyst can completely convert
cellulose to a sorbitol/D-glucose mixture in 1-butyl-3-methyl imidazolium chloride
([BMIM]Cl). A full cellulose conversion with a 74% sorbitol yield were achieved at
423 K, 3.5 MPa of H2 in 48 h with a cellulose concentration of 5 wt% using water
(60 μL.g−1 cellulose) and KOH (0.34 wt% relative to cellulose) as co-catalysts and
9  Production of Sorbitol from Biomass 287

a mixture of a heterogeneous (Rh(5%)/C catalyst, 7.6 wt% relative to cellulose, and

a homogeneous one (HRuCl(CO)(PPh3)3, 6.6 wt% relative to cellulose. The homo-
geneous catalyst was needed due to the low solubility of hydrogen in ionic liquids.
No reaction was observed in its absence, even when pure D-glucose was used as raw
material. Important drawbacks of this process are the use of a homogeneous pre-
cious metal based complex catalyst which difficult product separation and the low
hydrogen solubility, inherent to IL, leading to long reaction times. These problems
together with the added difficult for separation the polar sorbitol from the polar IL
represent strong pitfalls for industrial use in the short and medium term.

9.3.2  Electrochemical Production of Sorbitol

The electrochemical production of sorbitol was one of the first industrial processes
for producing an organic chemical by electrosynthesis. Anodic reaction was the
water oxidation resulting in oxygen and protons while cathodic reaction was the
reduction of D-glucose to sorbitol as main product and mannitol as a byproduct.
Anodic and cathodic compartments were separated each other by a diaphragm
(Fig. 9.3). In 1937 the Atlas Powder Company (since 1971 ICI Americas Inc.) built
a facility for manufacturing 1400 t.year−1 of sorbitol and mannitol by electroreduc-
tion of D-glucose (Fig. 9.4) resulting from corn starch [1].
The electrochemical reactor (cell) was an open-top rectangular tank 3.96 m long,
1.83  m wide and 0.91  m deep, fitted with amalgamated-lead cathodes and lead
anodes. 35 anodes and 36 cathodes, each one of 1.66  m2, were connected in
­monopolar mode in each cell. Anodic and cathodic compartments were separated by
a diaphragm of unglazed porcelain. The anolyte was a diluted aqueous solution of
sulfuric acid while the catholyte was an aqueous solution of sodium hydroxide,
sodium sulfate which was needed for increasing electrical conductivity for lowering
power consumption without increasing pH excessively, and D-glucose. The plant
contained 12 cells working batchwise at 85 A.m−2 and room temperature at a volt-
age of 20 V.cell−1. In the alkaline medium D-glucose isomerized partially to man-
nose and both D-glucose and mannose were reduced to sorbitol and mannitol,
respectively. For both improving sorbitol mass transfer from the bulk solution to the
cathode surface and controlling reaction temperature, the D-glucose solution was
continuously circulated through the cell and a double pipe counter-current cooling
coil. The productivity of the cell was later improved by working at 400 A.m−2 at
358–373 K [84].
After batch completion, water was partially removed from the catholyte by vac-
uum evaporation leaving a residue of sorbitol, mannitol and sodium sulfate. The
residue was treated with hot ethyl alcohol for dissolving sorbitol and mannitol. The
sulfate was filtered out the hot solution and the filtrate was sent to a crystallizer to
remove the mannitol by crystallization and subsequent centrifugal separation.
Mannitol was purified by repeated recrystallization from distilled water. The
­alcoholic sorbitol solution was evaporated to remove the ethyl alcohol resulting in
288 J.R. Ochoa-Gómez and T. Roncal

Fig. 9.3  Anodic and

cathodic reactions in the
electrosynthesis of sorbitol
from D-glucose using
divided cells

Fig. 9.4  Flow diagram for Atlas Powder Company electrochemical production of sorbitol from
cornstarch-derived D-glucose
9  Production of Sorbitol from Biomass 289

85 wt% water solution of sorbitol which was decolorized with activated carbon,
filtered and drummed for shipment. D-glucose and sodium sulfate were present in
small amounts as impurities.
As known, the electrochemical process was displaced a few years later by the
more cost-effective high-pressure catalytic hydrogenation. Since then, the bibliog-
raphy (both papers and patents) devoted to electrochemical synthesis is extremely
low in comparison with that related to chemical synthesis and even microbiological
synthesis, with efforts being focused on reducing the manufacturing costs. Cost
reduction requires decreasing power consumption and/or increasing cell productiv-
ity by obtaining simultaneously a valuable chemical at the anode (paired
synthesis).
Reduction in power consumption could be dramatic if an undivided cell could be
used due to the decrease in inter-electrode gap as well as the removal of the voltage
drops due to the anolyte and the diaphragm. However, the aldehyde group in
D-glucose is more easily oxidizable than water which requires the use of an anodic
depolarizer of lower oxidation potential than that of D-glucose. Thus, Hefti and
Kolb [85] developed a method in which sorbitol was synthesized by electroreducing
an aqueous solution of D-glucose and sodium sulphite in an undivided cell with an
amalgamated lead cathode and a graphite anode. The anodic oxidation of sulphite to
sulfate prevents the D-glucose oxidation. Chemical and current yields were 95%
and 90%, respectively. However, 0.8 kg of sodium sulphite.kg−1 of D-glucose was
needed increasing significantly the raw materials costs. Other strong drawback of
this procedure was the low current density used (50 A.m−2) requiring an electrode
area 8-fold higher than that in the Atlas Powder process.
The paired electro-oxidation and electroreduction of D-glucose for manufactur-
ing gluconic acid and sorbitol, respectively, has been reported [86, 87]. As depicted
in Fig. 9.5 for a divided cell, the cathodic reaction is the direct electroreduction of
D-glucose to sorbitol while the anionic one is an indirect electro-oxidation in which
the bromine, electrogenerated by electro-oxidation of bromide, oxidizes chemically
D-glucose to gluconic acid regenerating at the same time the bromide ion, which in
this way is used cyclically. Both undivided and divided cells can be used. Thus, an
undivided packed-bed electrode flow reactor fitted with a Raney Ni powder cathode
and a graphite chip anode was reported by Park el al [86]. Electrolyte was an aque-
ous solution of 1.6 M D-glucose and 0.4 M CaBr2. Working at 333 K, pH 5–7, a
current of 250 A.kg−1 of nickel powder and a flow rate of 6 L.h−1 current yields were
100% for both gluconic acid and sorbitol. However, these favorable results were
obtained by circulating a low electrical charge corresponding to 20% of the theoreti-
cal one leading to a low D-glucose conversion of 20%. On the other hand, this type
of cell is difficult to scale-up for industrial purposes.
Li et al. [87] used a divided filter press cell in which the Pb sheet cathode was
separated from a dimensionally stable anode (DSA) by a cation exchange mem-
brane. Gluconic acid and sorbitol were obtained in 90% chemical yields and over
80% current efficiencies by paired electrolysis of D-glucose at 333 K, 500 A.m−2
and an electrical charge equal to 110% of the theoretical one. The anolyte was 66.7
wt% D-glucose and 2 wt% NaBr in water, and the catholyte 66.7 wt% D-glucose,
290 J.R. Ochoa-Gómez and T. Roncal

Fig. 9.5  Paired electrosyntheis of gluconic acid and sorbitol from D-glucose. CEM Cation
exchange membrane

2.5 wt% NaOH and 2.5 wt% Na2SO4 in water, respectively. The anodic process is an
three-step indirect one in which bromide anion is electro-oxidized to bromine in a
first step, which oxidizes chemically D-glucose to gluconolactone in a second step,
step in which the bromide anion is regenerated. Finally, gluconolactone is hydro-
lyzed to gluconic acid in a third step. The cathodic reaction is the direct electrore-
duction of D-glucose to sorbitol. A low 4 V cell voltage leads to a power consumption
of 1.6 kWh.kg−1-sorbitol.
The process of Li et al. [87] is easy to scale-up but about 1.1 kg of gluconic acid
is produced per kg of sorbitol, which is a major drawback because the small glu-
conic acid market (about 5%) compared to that of sorbitol. Therefore, this process
is only suitable for low volume sorbitol facilities able to produce an amount of
gluconic acid easily absorbed by the market. An anodic reaction yielding a chemical
of similar market volume than sorbitol is needed for large scale sorbitol production
through this procedure.

9.3.3  Biotechnological Production of Sorbitol

An interesting alternative to the chemical production methods is based on biotech-

nology, and involves the use of enzymes and microorganisms as (bio)catalysts to
convert biomass-derived feedstocks into sorbitol. It is generally accepted that bio-
technological methods show a number of advantageous properties compared with
chemical methods, such as ambient temperature and pressure operation, which
9  Production of Sorbitol from Biomass 291

reduces energy costs, and a high selectivity and specificity, which minimizes by-­
product generation.
A complete review about biotechnological production of sorbitol was previously
published by Silveira and Jonas [88], containing the knowledge up to 2002 on this
topic.
Zymomonas mobilis, a Bacterium Able to Produce Sorbitol  The possibility of
producing ethanol using other microorganisms different to Saccharomyces cerevi-
siae, which is the yeast used on the industrial scale, has been considered. Among
potential candidates, special attention has been paid to Z. mobilis, a gram-negative
bacterium that can be found in sugar rich plant materials and fermented plant juices.
Z. mobilis shows several advantages over yeasts as a bioethanol producing microor-
ganism: (1) a higher sugar uptake rate; (2) a higher growth rate; (3) a higher ethanol
yield; (4) a lower biomass production; (5) a higher ethanol tolerance; (6) it does not
require controlled addition of oxygen during the fermentation; and (7) it is amend-
able to genetic manipulation. These advantageous features have allowed this bacte-
rium to be considered as a potential platform for future biorefineries [88]. The only
limitation of Z. mobilis compared to the yeast is that its range of usable carbon
sources is restricted to D-glucose, fructose and sucrose.
Z. mobilis shows an unusual property regarding carbohydrate metabolism: it uses
the Entner-Doudoroff pathway anaerobically to degrade D-glucose [89]. In this pro-
cess, 1 mol D-glucose is converted to almost 2 mol, the one of ethanol and the other
of CO2, that is, it is able to produce ethanol, at near theoretical levels. Fructose and
sucrose, the latter following hydrolysis to D-glucose + fructose, are metabolized
through the same pathway. Carbohydrate metabolism also generates 1 mol ATP and
results in a limited biomass formation, accounting for not more than 2–5% of the
carbohydrate consumed.
It is known that Z. mobilis converts up to 95% of added D-glucose or fructose to
ethanol, but when grown on sucrose the alcohol yield is greatly reduced. As some
strains produced levan when cultured on sucrose, it was first considered that part of
the sugar was directed to the synthesis of this fructose polysaccharide. However, the
amount of levan produced accounted for only a small proportion of the missing
carbon. Further studies showed that the fraction of sucrose not dedicated to ethanol
or levan synthesis resulted in the production of an additional product: sorbitol [90,
91].
Sorbitol was shown to be solely derived from the fructose, but it was only pro-
duced when both fructose and D-glucose were present. Very interestingly, sorbitol
formation was accompanied by an equimolar production of gluconic acid, which
suggested that two enzymes coupled together by an unknown cofactor to catalyze
these conversions [92].
The actual mechanism was soon elucidated, showing that a previously unknown
enzyme was responsible for both steps, the oxidation of D-glucose to glucono-δ-­
lactone and the reduction of fructose to sorbitol [93]. The enzyme, named as
D-glucose-fructose oxidoreductase (GFOR; EC1.1.199), contains a tightly bound
NADP as a cofactor and acts through a classical ping-pong mechanism, where
292 J.R. Ochoa-Gómez and T. Roncal

Fig. 9.6  Mechanism of sucrose conversion into sorbitol and gluconic acid by Z. mobilis. 1:
Invertase; 2: GFOR; 3: gluconolactonase

D-glucose is first converted to gluconolactone, which leaves the enzyme, and then
fructose is reduced to sorbitol. GFOR is a homotetramer located at the periplasm,
whose physiological function is proposed to be the regulation of osmotic stress of
the cell when grown on high sugar concentrations. Additionally, it was found that
gluconolactone leaving GFOR was converted to gluconic acid by the enzyme glu-
conolactonase, although the hydrolysis reaction may also occur spontaneously,
which makes GFOR-catalyzed reaction being almost irreversible in vivo and in
vitro.
A schematic representation of the steps and enzymes involved in sugar conver-
sion into sorbitol and gluconic acid is shown in Fig. 9.6.
The principal advantage of GFOR catalyzing D-glucose/fructose conversion into
gluconic acid/sorbitol is twofold. On the one hand, it is a self-regenerating redox
enzyme system and, on the other hand, it does not require the exogenous addition of
expensive and unstable adenine nucleotide cofactors due to the NADP being tightly
bound to the enzyme.
Biotechnological Production of Sorbitol Using Z. mobilis  In usual co-­
fermentation of D-glucose and fructose by Z. mobilis, sorbitol accumulates in the
culture medium (the bacterium cannot metabolize it) while gluconic acid is metabo-
lized through the Entner-Doudoroff pathway, resulting finally in the production of
ethanol and CO2. So, it was proposed that in a cell-free system or with the purified
GFOR it would be possible to avoid gluconic acid metabolism and ethanol forma-
tion, and only obtain the primary products, that is, sorbitol and gluconic acid.
According to this strategy, a toluene-permeabilized suspension of Z. mobilis cells,
previously grown and concentrated, was used to treat mixtures of D-glucose and
fructose [94, 95]. Cell permeabilization was intended to increase their permeability
and remove soluble co-factors and high energy compounds needed to carry out
gluconic acid metabolism, so avoiding its conversion into ethanol and other prod-
9  Production of Sorbitol from Biomass 293

ucts. With this system, concentrations as high as 290 g.L−1 sorbitol were achieved in
batch, reaching yields close to 95%. The use of Ca-alginate immobilized cells
instead of free cells rendered similar results in batch and, for the first time, a con-
tinuous process was developed showing only a small loss of enzyme activity (less
than 5%) after 120 h operation.
Alternative methods to toluene treatment of cells were evaluated to block ethanol
production, finding that a drying treatment of cells appeared to selectively inactivate
enzymes responsible for the conversion of D-glucose or fructose to ethanol, while
GFOR and gluconolactonase retained their activities [96].
Other different permeabilization methods for Z. mobilis cells were developed.
One of them was described by Bringer-Meyer and Sahm [97], involving freezing at
253 K and thawing at room temperature, and resulting in a production of sorbitol of
233 g.L−1 from a mixture of 234 g.L−1 D-glucose and 234 g.L−1 fructose, that is, an
almost quantitative conversion was achieved (98.5% yield), with an excellent pro-
ductivity of 46.6 g.L−1.h−1.
Permeabilization of Z. mobilis cells using cationic detergents such as cetyltri-
methylammonium bromide (CTAB), increased sorbitol production from 240 to 295
g.L−1 from equimolar fructose and D-glucose mixtures up to 60% (w/v), and avoided
ethanol production. With these permeabilized cells, in a two-stage continuous pro-
cess with κ-carrageenan-immobilized and polyethylenimine-hardened cells, no sig-
nificant decrease in the conversion yield (>98%) was observed after 75 days [98].
Both the use of Z. mobilis CTAB-permeabilized cells in the production of sorbitol
and the method for cell immobilization were later patented by Rehr and Sahm [99,
100]. Operational stability of the system could be enhanced by either drying or add-
ing polyols to κ-carrageenan beads, which increased their rigidity, achieving the
best results with a two-stage continuous packed bed reactor, where a sorbitol con-
centration of 178.6 g.L−1 with a productivity of 25 g.L−1.h−1 was obtained [101].
In the course of reaction, gluconic acid formation leads to a pH decrease causing
enzyme inhibition, an event that must be avoided through alkali addition to the
medium. As a feasible alternative, the coupling of an electrodialysis unit to a system
containing permeabilized and immobilized Z. mobilis cells, was evaluated in order
to efficiently remove gluconic acid and prevent acidification and enzyme inhibition,
allowing an evident improvement in the stability of the enzyme, which maintained
its reaction rate unchanged for 60 h of operation [102].
Most of the previous work on biological sorbitol production involves the use of
permeabilized Z. mobilis cells as biocatalyst. However, with the aim of reducing
production costs, the use of intact, non-permeabilized cells, was also proposed [103,
104]. Applying this method, Silveira et al. [104] found that the use of up to 650
g.L−1 of an equimolar mixture of D-glucose and fructose resulted in an almost com-
plete bioconversion to sorbitol and gluconic acid, without ethanol formation and
reaching yields over 91% for both products. This result was explained by the effects
that the high substrate concentration could cause on the cells, mainly the loss of cell
viability due to the high osmotic pressure and the inhibition of the normal ethanolo-
genic metabolism, resulting in a preferential utilization of substrates via GFOR.
294 J.R. Ochoa-Gómez and T. Roncal

Therefore, the use of free-untreated cells of Z. mobilis, despite their slightly lower
yields compared to those obtained with permeabilized cells, appeared as an attrac-
tive option.
The biotechnological production of sorbitol using Z. mobilis suffers a problem
related to the relatively high cost of fructose compared to product value. As a result,
the use of alternative, low cost, substrates has been considered and evaluated.
A feasible alternative is sucrose, a disaccharide composed of two monosaccha-
rides that are co-substrates of GFOR. Toluene-permeabilized Z. mobilis cells were
co-immobilized in calcium-alginate beads with invertase, and applied to the produc-
tion of gluconic acid and sorbitol from sucrose in a RPBR [105]. Maximum produc-
tivities for sorbitol of 5.20 g.L−1.h−1 were achieved at a dilution rate of 0.053 h−1 and
a sucrose concentration of 20% when recirculated at the rate of 1200 mL.h−1. The
co-immobilized enzymes were reported to remain stable for 250 h in the RPBR
without any loss of activity.
Another substrate checked was Jerusalem artichoke, an inulin-rich plant source.
Inulin is a fructan, that is, a fructose-rich polysaccharide, so it could substitute pure
fructose as GFOR co-substrate. Therefore, toluene permeabilized Z. mobilis cells,
co-immobilized with inulinase (the inulin depolymerizing enzyme) in calcium-­
alginate beads were used to convert D-glucose and Jerusalem artichoke into glu-
conic acid and sorbitol [106]. In a RPBR, the maximum productivity for sorbitol
was found to be 26 g.L−1.h−1, showing the co-immobilized enzymes full stability for
250 h without any loss of activity.
Sugar cane molasses, which is a by-product of the sugar industry, has been con-
sidered as an attractive substrate for sorbitol production [107]. Molasses contains
not only a high sucrose concentration, but in addition other important substances
beneficial for the fermentation process and it has low cost. Cazetta et al. [107] car-
ried out an optimization study for sorbitol production by Z. mobilis in sugar cane
molasses, finding that the best conditions for sorbitol production were 300 g.L−1
total reducing sugars, where around 14 g.L−1 sorbitol were produced, only repre-
senting a 23% of the fructose utilized. This reduced sorbitol production was attrib-
uted to the high concentration of salts present in sugar cane molasses that may have
raised the osmotic pressure above acceptable levels, reducing cell viability and sup-
pressing sorbitol production.
Two low-cost feedstocks, inulin and cassava starch, have been tested to produce
sorbitol [108]. The process involved the use of a commercial glucoamylase enzyme
for the simultaneous saccharification of inulin and starch into high titer D-glucose
and fructose hydrolysate, replacing the expensive and not commercially available
inulinase enzyme. Conversion was carried out using immobilized whole cells of a
recombinant GFOR over-expressing Z. mobilis strain, achieving a titer of 180 g.L−1
sorbitol in batch [108].
Following a strategy of introducing metabolic engineering technologies into
biotechnological production of sorbitol, another recombinant Z. mobilis strain
­
­over-­expressing GFOR was constructed, showing a specific activity at least twofold
greater than that in the wild type strain [109]. Addition of divalent metal ions,
9  Production of Sorbitol from Biomass 295

­especially Zn2+, to freezing and thawing permeabilized recombinant cells was found
to improve the bioconversion process, by inhibiting the Entner-Doudoroff pathway
enzymes, resulting in a drastic reduction in ethanol and a significant increase in
sorbitol yields. Using an equimolar D-glucose/fructose mixture as substrate, e­ thanol
production was reduced from 16.7 to 1.8 g.L−1 and sorbitol yield was increased to
virtually 100%, up to 161.1 g.L−1.
Besides D-glucose, GFOR can also accept other alternative donor substrates,
including monosaccharides as xylose and galactose, or disaccharides as maltose and
lactose, which are oxidized to their corresponding aldonic acids. Conversion of a
mixture of lactose and fructose into lactobionic acid and sorbitol by permeabilized
Z. mobilis cells carrying GFOR and gluconolactonase enzymes was studied, using
either mobilized or Ca-alginate immobilized cells. Under optimal operating condi-
tions, an almost quantitative conversion of 350 mM fructose into sorbitol can be
achieved [110].
As an alternative to the use of Z. mobilis cells, either permeabilized or not, as
biocatalyst, some attempts were made to develop a process employing cell-free
GFOR [111]. This strategy was explained by the higher enzyme concentrations that
could be applied and by the lack of mass transfer limitations that could be expected
using free GFOR, which might increase productivity. However, the free enzyme
was found to rapidly inactivate during the time course of its own catalytic action,
requiring supplementation with thiol-protecting agents to increase GFOR stability.
A continuous process in an ultrafiltration membrane reactor was developed using
a crude cell extract of Z. mobilis, where enzyme inactivation was almost completely
avoided by adding to reaction medium weak bases, such as Tris or imidazol, to neu-
tralize gluconic acid produced, and dithiothreitol to protect thiol groups [112]. This
system was reported to operate over a time period of more than 250 h without sig-
nificant decrease in substrate conversion or enzyme activity, resulting in a sorbitol
productivity of 4.37 g.L−1.h−1 and a yield close to 40%. The use of a tangential
ultrafiltration loop reactor allowed increasing substrate conversion and productivity
up to more than 85% and 5 g.L−1.h−1, respectively, from 3 M sugar [113].
A compilation of the main achievements in biotechnological production of sor-
bitol using Z. mobilis is shown in Table 9.2. Sorbitol synthesis with Z. mobilis is a
biocatalytic process, and not a fermentation, contrary to some reports. Biocatalyst is
GFOR enzyme, which is bound to bacterial cells, so it can be considered as a whole-­
cell biocatalyst. The process comprises two separate steps: (1) biocatalyst genera-
tion and (2) biocatalytic reaction. The step of biocatalyst generation is carried out
by culture of the microorganism to obtain as much biomass as possible to be used
thereupon in the biocatalysis. Therefore, the first step is certainly a true
fermentation.
Space-time yields shown in Table 9.2 actually correspond to productivities of the
second step, the biocatalytic step, and not to that of the whole process. As substrates
and time consumed in biocatalyst generation are not taken into account, the actual
productivities of the whole process could be considered to be lower than those val-
ues shown in Table 9.2.
 296

Table 9.2  Examples of biotechnological production of sorbitol using Zymomonas mobilisa

Sorbitol Space-time Specific
concentration Sorbitol yield productivity
Biocatalyst Permeabilization Bioreactor Substrate (g.L−1) yield (%) (g.L−1.h−1) (g.g−1.h−1) Ref.
Free cells Toluene Batch, stirred tank 300 g.L−1 290 94.3 18.9 – [95]
D-glucose, 300
g.L−1 fructose
Free cells Freezing and Batch, stirred tank 234 g.L−1 233 98.5 46.6 1.08 [97]
thawing D-glucose, 234
g.L−1 fructose
Free cells CTABb Batch, stirred tank 300 g.L−1 295 98 – – [98]
D-glucose, 300
g.L−1 fructose
Ca-alginate Toluene Continuous 200 g.L−1 sucrose 98 92 5.20 – [105]
co-immobilized recycle
cells and invertase packed-bed
Ca-alginate Toluene Continuous 200 g.L−1 74.3 – 26 – [106]
co-immobilized recycle D-glucose and
cells and inulinase packed-bed Jerusalem artichoke
Free cells CTABb Batch, stirred tank 281 g.L−1 285 97.0 47.5 0.6 [99]
D-glucose, 292
g.L−1 fructose
κ-Carrageenan CTABb Two-stage 300 g.L−1 178.6 – 25 1.6 [101]
immobilized cells continuous packed D-glucose, 300
bed g.L−1 fructose
Crude cell extract – Continuous 3 M sugar – >85 5.25 – [113]
tangential
ultrafiltration loop
J.R. Ochoa-Gómez and T. Roncal
 Free cells – Batch, stirred tank 300 g.L−1 279 92.1 – 1.0 [103]
D-glucose, 300
g.L−1 fructose
Free cells – Batch, stirred tank 325 g.L−1 300 91 45 1.5 [104]
D-glucose, 325
g.L−1 fructose
Free cells – Batch, stirred tank Molasses, 300 g.L−1 13.87 23 0.38 1.9 [107]
reducing sugars
Free recombinant Freezing and Batch, stirred tank 157.1 g.L−1 161.1 100.3 61.2 3.06 [109]
GFOR over-­ thawing D-glucose, 158.9
expressing cells; g.L−1 fructose
Zn2+ suppl.
Immobilized – Batch, stirred tank Inulin and cassava 180 – – – [108]
9  Production of Sorbitol from Biomass

recombinant starch,
GFOR over-­ glucoamylase
expressing cells
a
Temperatures for bioconversions between 33 and 312 K, with most at 312 K. pH: 6.2–6.5
b
CTAB Cetyltrimethylammonium bromide
297
298 J.R. Ochoa-Gómez and T. Roncal

Largest space-time yields are associated, not surprisingly, to high cell (biocata-
lyst) loads, greater than 15 g cells (dry weight).L−1 [97, 99, 101, 104, 109] as occurs
in any catalytic reaction. In turn, specific productivities of Z. mobilis catalyst, which
give an idea of its specific activity, depend primarily on the strain used, but also on
reaction conditions, including whether cells are permeabilized or not, and the per-
meabilization method. Specific productivities, when available, are reported to be
between 0.6 and 1.9 g.g−1.h−1 (Table 9.2). As expected, overexpression of the GFOR
enzyme results in an increased specific productivity, amounting to more than 3
g.g−1.h−1 [108].
Other Microorganisms Producing Sorbitol  In addition to Z. mobilis, only a few
microorganisms have been described as natural sorbitol producers. An example is
the methanol-utilizing yeast Candida boidinii, which was shown to produce sorbitol
from D-glucose using an intact cell system, with methanol as the energy source for
generating NADH for the reduction of D-glucose to sorbitol. With this system, max-
imum amounts of sorbitol of 8.8 and 19.1 g.L−1 were obtained from 20 g.L−1 of
D-glucose and fructose, respectively [114]. Sorbitol was directly produced from
fructose in a reduction catalyzed by a sorbitol dehydrogenase enzyme [115].
Another yeast, Saccharomyces cerevisiae strain ATCC 36859, has also been
shown to be able to produce sorbitol when cultured in Jerusalem artichoke juice
[116]. The yeast produced ethanol from the beginning of the process, and sorbitol
production occurred only after D-glucose depletion. When the juice was supple-
mented with 3% yeast extract, sorbitol concentration reached a 4.6%, with a yield
of 0.259 g.g−1 sugar consumed.
Engineering strategies have been applied to some lactic acid bacteria in order to
obtain new sorbitol producers. All of these strategies involve a few manipulations of
carbon and energy metabolism, including overexpression of the key enzymes for
converting a substrate to sorbitol, blocking re-utilization of the produced sorbitol,
decreasing other by-product synthesis, and improving redox balance [117–119].
Lactic acid bacteria do not produce sorbitol at detectable levels, but some of
them have the ability to use sorbitol as a carbon source through the genes contained
in the sorbitol operon. In Lactobacillus casei and L. plantarum two D-sorbitol-6-­
phosphate dehydrogenase-encoding genes have been found, which catalyze oxida-
tion of sorbitol-6-phosphate to fructose-6-phosphate. As this reaction is reversible,
the possibility of sorbitol production from fructose exists, although it is very
unlikely because both genes are tightly controlled by catabolite repression and sub-
strate induction.
So, in order to avoid this and to revert the sorbitol catabolic pathway toward
sorbitol synthesis, a strain of L. casei was constructed by integration of a D-sorbitol-­
6-phosphate dehydrogenase-encoding gene (gutF) in the chromosomal lactose
operon, and the resulting recombinant strain was shown to produce small amounts
of sorbitol from D-glucose with a yield of 2.4% [117]. Subsequent inactivation of
the L-lactate dehydrogenase gene led to an increase in sorbitol production (yield,
4.3%), suggesting that the engineered route provided an alternative pathway for
NAD+ regeneration. However, the recombinant L. casei strain suffered from two
9  Production of Sorbitol from Biomass 299

drawbacks that diminished sorbitol production: after D-glucose depletion sorbitol

was reutilized, and mannitol was also produced. The first problem was avoided by
deleting the gutB gene, responsible for uptake and reutilization of the synthesized
sorbitol, and the second one by inactivating the mtlD gene, encoding the enzyme
catalyzing the conversion of fructose-6-phosphate to mannitol-1-phosphate. The
resulting strain was able to convert lactose into sorbitol with a yield of 9.4% using
an optimized fed-batch system and whey permeate as a substrate [118].
A similar strategy was used with L. plantarum, which was metabolically engi-
neered to produce sorbitol by constitutive overexpression of the two sorbitol-­6-­
phosphate dehydrogenase genes (srlD1 and srlD2) in a mutant strain deficient for
both L- and D-lactate dehydrogenase activities [118]. Using resting cells under pH
control with D-glucose as substrate, sorbitol yield approached to 65%, which is
close to the maximal theoretical value of 67%.
Nevertheless, reported sorbitol production by all these microorganisms, though
possible, falls far below that obtained with Z. mobilis, so they are not currently
alternatives.

9.3.4  Recovery and Purification of Sorbitol

Main by-products in sorbitol production [33, 120] are depicted in Fig. 9.7. D-glucose
can isomerize to fructose and mannose, especially in alkaline medium. Hydrogenation
of mannose yields mannitol while hydrogenation of fructose yields mannitol as well
as sorbitol. L-iditol can be formed by sorbitol isomerization, mainly at the end of
the reaction due to the high concentration of sorbitol. Gluconic acid can be formed
by D-glucose oxidation but its formation can be prevented by deoxigenating the
reaction mixture before reaction. Additionally, xylitol and lower sugar alcohols
could be formed by hydrogenolysis of sorbitol mainly at high temperatures, but they
are not present under usual industrial conditions.
Formation of such a by-products can be largely prevented by carefully selecting
the reaction conditions as shown by the fact herein reported that in the catalytic
hydrogenation of D-glucose a 100% D-glucose conversion and a higher than 99%
selectivity to sorbitol are obtained. The typical procedures for recovery and purifica-
tion of sorbitol manufactured by catalytic hydrogenation and electroreduction of
D-glucose have been depicted in Figs. 9.2 and 9.3, respectively. Also, chromato-
graphic methods are currently industrially used for separating mixtures of sorbitol
and mannitol, such as the PuriTech’s ION-IX continuous countercurrent technology
[121] by means of which a typical blend of mannitol/sorbitol with a content of man-
nitol between 20% and 50% is separated to produce both chemicals in a purity
≥98%.
Relative to the biotechnological production, the sorbitol recovery and purifica-
tion is only addressed in a few papers. Ferraz et al. [122] reported the use of an
electrodialysis (ED) system coupled to the bioreactor to simultaneously remove
300 J.R. Ochoa-Gómez and T. Roncal

Fig. 9.7  Main by-products in sorbitol production

gluconic acid from the medium as it is produced. Gluconic acid is a weak acid (pKa
3.86) meaning that at the pH (6.2–6.5) of the medium it is fully dissociated and,
therefore, it is obvious that it can be separated and concentrated in the concentration
compartment of a 2-compartments ED cell, while sorbitol remains in the diluted
compartment. However, we consider that full removal of gluconic acid by ED is not
economically feasible. Below a concentration of 1–5 g.L−1 the sorbitol solution
must be treated by ion exchange for separating the residual gluconic acid. Chun and
Rogers [94] reported the separation of the products using a basic anion exchange
resin and a solution of Na2B4O7/H3BO3 as eluent, and Silveira et al. [123] proposed
a method for recovering sorbitol and sodium gluconate by selective precipitation of
sodium gluconate with organic solvents like methanol and ethanol. Anyway, it is
obvious that a full recovery procedure of sorbitol from the complex biotech reaction
media has not developed yet, and further and intensive research must be carried out
to achieve this goal. One possibility to investigate is the use of the aqueous two-­
phase extraction technique, taking advantage of the salting-out effect created by
saturating the aqueous solution with a highly soluble salt. This method has given
very good results for recovering and concentrating 2,3-butanediol from a fermenta-
tion broth using an ethanol/phosphate system [124]. A 98.1% 2,3-butanediol recov-
ery was achieved.
9  Production of Sorbitol from Biomass 301

9.4  Conclusions and Future Outlook

Three reported processes for sorbitol production have been reviewed: (i) catalytic
hydrogenation of D-glucose, (ii) electroreduction of D-glucose and, (iii) biotechno-
logical production from fructose using the enzyme D-glucose-fructose oxidoreduc-
tase (GFOR) from Z. mobilis.
While the electroreduction of D-glucose was the first process industrially oper-
ated, it was soon replaced by the catalytic hydrogenation of the same raw material
using Raney Ni catalysts due to both the high power consumption and low produc-
tivity of the former. In fact, the latter is the process exclusively used currently at
industrial level due of course to its higher productivity, but especially because a
100% D-glucose conversion is achieved with a selectivity of sorbitol higher than
99%, which simplified strongly the purification process. The early problems related
to nickel deactivation by sintering, leaching into the reaction mixture and poisoning
have been largely solved by using some promoters, such as Cr and Mo, developing
Ni supported catalysts to increase metal dispersion and large exposed surface area
resulting in improved Ni activity and stability, and using new catalysts, with
ruthenium-­based ones being the most used due to the much higher activity of Ru (up
to 50-fold higher than Ni) and because Ru is not leached. These features make the
catalytic hydrogenation of D-glucose an extremely solid production process which
will be difficult to be displaced. However, in the next 5 years sorbitol worldwide
demand will exceed 2300 kt and it is expected to grow continuously which repre-
sents a challenge for increasing the productivity at lower costs. This challenge will
be the driving force for further improving the catalytic chemical production of sor-
bitol through three ways in which research efforts should be focused on:
1. Developing more efficient continuous processes, which potentially are also more
environmentally friendly as shown by LCA analysis [27];
2. Designing new more efficient and stable catalysts for working both at lower
temperatures and pressures than present. It is worth to mention that hydrogena-
tion pressures used in sorbitol production are much higher (10–15 MPa) than
those usually employed in most industrial hydrogenation processes (<1.5 MPa).
Ru-based catalysts with enhanced activity and stability by introducing metal
promoters and modified supports appear to be the more promising way; and
3. Developing industrially cost-effective methods for using directly the polymeric
sources of D-glucose as raw materials, mainly cellulose because it does not inter-
fere with the food chain. Taking into account both the low cellulose concentra-
tions in the current reported processes and that cellulose hydrolysis is the limiting
step the one pot option does not appear to be the appropriate choice. Probably, a
two-step process involving a first step of cellulose fast hydrolysis and a second
one of hydrogenation of the hydrolysate. The fast hydrolysis step could be car-
ried out by the method developed by Fang [125] consisting of putting lignocel-
lulose biomass in pure water in concentrations up to 51.5% and heating rapidly
the mixture to 603–676 K at 19–42 MPa. As a result, 89 ~ 99% of the lignocel-
lulose biomass is dissolved and rapidly hydrolyzed to saccharides in 3.38 ~
302 J.R. Ochoa-Gómez and T. Roncal

21.79 s, or by the Ultra-Fast Hydrolysis (UFH) method reported by Cantero et al.

[126] using a continuous facility with instantaneous heating and cooling and
with reaction residence times as low as 20 milliseconds. At 673 K and 23 MPa
cellulose was hydrolyzed into soluble sugars with a selectivity of 96% on a car-
bon basis at a residence time of 30 milliseconds with an extremely low selectiv-
ity to 5-hydroxymethylfurfural, a known inhibitor of fermentations, of 0.01%.
Reducing the residence time of a continuous process to milliseconds is a key
breakthrough in process intensification opening the possibility of moving from
the conventional reactors (m3 volume) to microreactors (cm3 volume).
Since its replacement by the catalytic hydrogenation of D-glucose, the research
efforts devoted to the electrochemical production of sorbitol from D-glucose have
been scarce and basically devoted to look for an anodic reaction producing a valu-
able chemical, ascorbic acid, for reducing production costs. However, the sorbitol
and ascorbic acid market volumes are so unbalanced that this is not a suitable
approach for revival of sorbitol electrochemical production. However, the electro-
chemical method has several important advantages. Namely, it operates at ambient
pressure, it does not use hydrogen and it works at a lower temperature than the
chemical hydrogenation. In other words, it is attractive due to safety issues associ-
ated with hydrogen handling and the electrochemical process operates at ambient
pressure. Moreover, the strong drawbacks that led to its displacement by the cata-
lytic hydrogenation, i.e. high power consumption and low productivity, could be
overcome by the great advances in electrochemical technology since that time.
Thus, the use of decorated (with Ni and Ru) porous cathodes has shown to be very
effective in the reduction of carbonyl moieties to alcohol ones [127] and could allow
to work at higher apparent current densities, i.e. higher productivities, while keep-
ing both a low cathode potential and a very high sorbitol selectivity resulting in both
lower cost and OPEX. On the other hand, much lower power consumption could be
achieved by using the zero gap cells [128] in which both anode and cathode are in
contact with the membrane, i.e. inter-electrode gap is equal to membrane thickness
(between ~150 and 250 μm, depending on membrane type). Consequently, the elec-
trochemical production of sorbitol could have a niche in low volume production
plants (below 10,000 t.year−1) where the facility needed in a catalytic hydrogenation
process for in situ hydrogen production is not economically feasible.
Regarding the sorbitol biotechnological production, the use of GFOR from Z.
mobilis constitute currently the best method. However, it still bears some weak-
nesses that make it uncompetitive with respect to chemical methods. One of them is
that the enzyme requires a co-substrate in addition to the substrate (fructose) to be
converted into sorbitol. Therefore, when D-glucose is the co-substrate, as usually
occurs, one mole of gluconic acid is obtained per mole of sorbitol produced. This
means that, in order to favor the cost-effectiveness of the biotechnological process,
it would be necessary to find a suitable outlet for such gluconic acid, which allows
increasing its market well above its current volume. Perhaps, other donor substrates
alternative to D-glucose could be favorably introduced, providing that their result-
ing aldonic acids have a larger market than that of gluconic acid. Another factor that
9  Production of Sorbitol from Biomass 303

hinders progress in the application of this biotechnological process is that the sub-
strate is fructose, a monosaccharide having a relatively high cost compared to prod-
uct value. So, the use of low cost substrates, such as molasses or inulin/fructan-rich
feedstocks, alternative to pure or semipure fructose preparations, must be a priority.
Likewise, the application of recombinant DNA technologies could drive progress in
this field, through the improvement of the catalytic properties of GFOR and its
large-scale production to allow its use as an efficient biocatalyst. Finally a critical
issue so that the biotechnological production of sorbitol has an industrial possibility
is the development of a cost-effective separation and purification procedure. This is
a current major bottleneck for further development. Few papers are devoted to
aspects of separation and purification of sorbitol. Researchers in the field need to be
aware of its key importance and increase dramatically their efforts to develop an
industrially acceptable solution.

References

1. Taylor RL.  Sorbitol from D-glucose by electrolytic reduction. Chem Met Eng.
1937;44:588–91.
2. Werpy T, Petersen G, editors. Top value added chemicals from biomass; Vol. 1: results of
screening for potential candidates from sugars and synthesis gas. Oak Ridge: U.S. Department
of Energy; 2004.
3. Bozell JJ, Petersen GR. Technology development for the production of biobased products
from Biorefinery carbohydrates – the US department of Energy’s “Top 10” revisited. Green
Chem. 2010;12:539–54.
4. Grand View Research. Global sorbitol market. https://www.grandviewresearch.com/press-­
release/global-sorbitol-market (2015). Last accessed 15 May 2016.
5. Radiant Insights Inc.. Sorbitol market size, price analysis, research report, 2020. http://www.
radiantinsights.com/research/sorbitol-market (2015). Last accessed 15 May 2016.
6. The manufacture and production of valuable products from sorbitol. GB Patent 301655
(1928).
7. Pappenberger G, Hohmann HP.  Industrial production of L-ascorbic acid (vitamin C) and
D-isoascorbic acid. Adv Biochem Eng Biotechnol. 2014;143:143–88.
8. Laird K. Roquette brings world's largest isosorbide production unit on stream. Plastics today.
http://www.plasticstoday.com/roquette-brings-worlds-largest-isosorbide-production-unit-­
stream/85076483422068 (2015). Last accessed 15 May 2016.
9. Rose M, Palkovits R.  Isosorbide as a renewable platform chemical for versatile applica-
tions—Quo Vadis? ChemSusChem. 2012;5:167–76.
10. Bersot JC, Jacquel N, Saint-Loup R, Fuertes P, Rousseau A, Pascault JP, Spitz R, Fenouillot
F, Monteil V. Efficiency increase of poly (ethylene terephthalate-co-isosorbide terephthalate)
synthesis using bimetallic catalytic systems. Macromol Chem Phys. 2011;212:114–2120.
11. Ochoa-Gómez JR, Gil-Río S, Maestro-Madurga B, Lorenzo-Ibarreta L, Gómez de Miranda
O.  Improved method for manufacturing 1,4:3,6-dianhydrohexitol di(alkyl carbonate)s. US
Patent Application 20150336978A1 (2015).
12. Fenouillot F, Rousseau A, Colomines G, Saint-Loup R, Pascault JP. Polymers from renew-
able 1,4:3,6-dianhydrohexitols (isosorbide, isomannide and isoidide): A review. Prog Polym
Sci. 2010;35:578–622.
13. Ionescu M. Polyether polyols for rigid polyurethane foams. In:Chemistry and technology of
polyols for polyurethanes. Shawbury: Rapra Technology Limited; 2005. p. 343–4.
304 J.R. Ochoa-Gómez and T. Roncal

14. DuPont de Nemours. Improvements in or relating to the manufacture of polyhydric alcohols.

GB Patent 528064 (1940).
15. Zhao L, Zhou JH, Sui ZJ, Zhou XG. Hydrogenolysis of sorbitol to glycols over carbon nano-
fiber supported ruthenium catalyst. Chem Eng Sci. 2010;65:30–5.
16. Du W-C, Zheng L-P, Shi J-J, Xia S-X, Hou Z-Y.  Production of C2 and C3 polyols from
D-sorbitol over hydrotalcite-like compounds mediated bi-functional Ni–Mg–Al–Ox cata-
lysts. Fuel Process Technol. 2015;139:86–90.
17. Guo X, Guan J, Li B, Wang X, Mu X, Liu H. Conversion of biomass-derived sorbitol to gly-
cols over carbon materials supported Ru-based catalysts. Scientific Reports 5: Article number
16451 (2015).
18. Ruppert AM, Weinberg K, Palkovits R. Hydrogenolysis goes bio: from carbohydrates and
sugar alcohols to platform chemicals. Angew Chem Int Ed. 2012;51:2564–601.
19. Pang J, Zheng M, Sun R, Wang A, Wang X, Zhang T. Synthesis of ethylene glycol and tere-
phthalic acid from biomass for producing PET. Green Chem. 2016;18:342–59.
20. Dusselier M, Wouwe PV, Dewaele A, Makshina E, Sels BF. Lactic acid as a platform chemi-
cal in the biobased economy: the role of chemocatalysis. Energy Environ Sci.
2013;6:1415–42.
21. Ramírez-López C, Ochoa-Gómez JR, Gil-Río S, Gómez-Jiménez-Aberasturi O, Torrecilla-­
Soria J. Chemicals from biomass: synthesis of lactic acid by alkaline hydrothermal conver-
sion of sorbitol. J Chem Technol Biotechnol. 2011;86:867–74.
22. Huber GW, Dumesic JA. An overview of aqueous-phase catalytic processes for production of
hydrogen and alkanes in a biorefinery. Catal Today. 2006;111:119–32.
23. Xi J, Xia Q, Shao Y, Ding D, Yang P, Liu X, Lu G, Wang Y. Production of hexane from sorbi-
tol in aqueous medium over Pt/NbOPO4 catalyst. Appl Catal B-Environ. 2016;181:699–706.
24. Weng Y, Qiu S, Ma L, Liu Q, Ding M, Zhang Q, Zhang Q, Wang T. Jet-fuel range hydrocar-
bons from biomass-derived sorbitol over Ni-HZSM-5/SBA-15 catalyst. Catalysts.
2015;5:2147–60.
25. Müller J, Hoffmann U. Verfahren zur Darstellung von hochwertigen Alkoholen durch kataly-
tische Reduktion von Zuckerarten mit Wasserstof. DE Patent 544 666 (1925).
26. Müller J, Hoffmann U. Verfahren zur Darstellung von hochwertigen Alkoholen durch kataly-
tische Reduktion von Zuckerarten mit Wasserstof. DE Patent 554 074 (1926).
27. Gericke D, Ott D, Matveeva VG, Sulman E, Aho A, Murzin DY, Roggan S, Danilova L,
Hessel V, Loebg P, Kralisch D. Green catalysis by nanoparticulate catalysts developed for
flow processing? Case study of D-glucose hydrogenation. RSC Adv. 2015;5:15898–908.
28. Schiweck H, Bär A, Vogel R, Schwarz E, Kunz M, Dusautois C, Clement A, Lefranc C,
Lüssem B, Moser M, Peters S.  Sugar alcohols. In: Ullmann’s Encyclopedia of Industrial
Chemistry. Wiley-VCH Verlag GmbH & Co. KGaA; 2012.
29. Court J, Damon JP, Masson J, Wierzchowski P. Hydrogenation of D-glucose with bimetallic
catalysts (NiM) of Raney type. Stud Surf Sci Catal. 1988;41:189–96.
30. Gallezot P, Cerino PJ, Blanc B, Flèche G, Fuertes P. D-glucose hydrogenation on promoted
Raney-nickel catalysts. J Catal. 1994;146:93–102.
31. Li H, Li H, Deng J-F. D-glucose hydrogenation over Ni–B/SiO2 amorphous alloy catalyst and
the promoting effect of metal dopants. Catal Today. 2002;74:53–63.
32. Hoffer BW, Crezee E, Devred F, Mooijman PRM, Sloof WG, Kooyman P, van Langeveld
AD, Kapteijn F, Moulijn JA. The role of the active phase of Raney-type Ni catalysts in the
selective hydrogenation of D-glucose to d-sorbitol. Appl Catal A-General.
2003;253:437–52.
33. Hoffer BW, Crezee E, Mooijman PRM, van Langeveld AD, Kapteijn F, Moulijn JA. Carbon
supported Ru catalysts as promising alternative for Raney-type Ni in the selective hydrogena-
tion D D-glucose. Catal Today. 2003;79–80:35–41.
34. Kusserow B, Schimpf S, Claus P. Hydrogenation of D-glucose to Sorbitol over Nickel and
Ruthenium Catalysts. Adv Synth Catal. 2003;345:289–99.
9  Production of Sorbitol from Biomass 305

35. Schimpf S, Louis C, Claus P. Ni/SiO2 catalysts prepared with ethylenediamine nickel precur-
sors: Influence of the pretreatment on the catalytic properties in D-glucose hydrogenation.
Appl Catal A. 2007;318:45–53.
36. Geyer R, Kraak P, Pachulski A, Schödel R. New catalysts for the hydrogenation of D-glucose
to sorbitol. Chem Ing Tech. 2012;84:513–6.
37. Li S, Chen H, Shen J. Preparation of highly active and hydrothermally stable nickel catalysts.
J Colloid Interface Sci. 2016;447:68–76.
38. Li HX, Wang WJ, Deng JF. D-glucose hydrogenation to sorbitol over a skeletal Ni-P amor-
phous alloy catalyst (Raney Ni-P). J Catal. 2000;191:257–60.
39. Wisniak J, Simon R.  Hydrogenation of D-glucose, fructose, and their mixtures. Ind Eng
Chem Prod Res Dev. 1979;18:50–7.
40. Michel C, Gallezot P. Why is ruthenium an efficient catalyst for the aqueous-phase hydroge-
nation of biosourced carbonyl Compounds? ACS Catal. 2015;5:4130–2.
41. Guo H, Li H, Zhua J, Ye W, Qiao M, Dai W.  Liquid phase D-glucose hydrogenation to
d-­glucitol over an ultrafine Ru-B amorphous alloy catalyst. J  Mol Catal A Chem.
2003;200:213–21.
42. Zhang J, Li J-B, Wu S-B, Liu Y. Advances in the catalytic production and utilization of sor-
bitol. Ind Eng Chem Res. 2013;52:11799–815.
43. Guo X, Wang X, Guan J, Chen X, Qin Z, Mu X, Xian M. Selective hydrogenation of D-D-­
glucose to D-sorbitol over Ru/ZSM-5 catalysts. Chin J Catal. 2014;35:733–40.
44. Mishra DK, Dabbawala AA, Parka JJ, Jhung SH, Hwang J-S.  Selective hydrogenation of
D-glucose to D-sorbitol over HY zeolite supported ruthenium nanoparticles catalysts. Catal
Today. 2014;232:99–107.
45. Mishra DK, Lee J-S, Chang JS, Hwang J-S.  Liquid phase hydrogenation of D-glucose to
D-sorbitol over the catalyst (Ru/NiO–TiO2) of ruthenium on a NiO-modified TiO2 support.
Catal Today. 2012;185:104–8.
46. Aho A, Roggan S, Simakova OA, Salmi T, Murzin DY.  Structure sensitivity in catalytic
hydrogenation of D-glucose over ruthenium. Catal Today. 2015;241:195–9.
47. Lazaridis PA, Karakoulia S, Delimitis A, Comanc SM, Parvulescu VI, Triantafyllidis
KS. D-glucose hydrogenation/hydrogenolysis reactions on noble metal(Ru, Pt)/activated car-
bon supported catalysts. Catal Today. 2015;257:281–90.
48. Wang S, Wei W, Zhao Y, Li H, Li H. Ru–B amorphous alloy deposited on mesoporous silica
nanospheres: an efficient catalyst for D-glucose hydrogenation to D-sorbitol. Catal Today.
2015;258:327–36.
49. Santiso EE, George AM, Turner CH, Kostov MK, Gubbins KE, Buongiorno-Nardelli M,
Sliwinska-Bartkowiak M. Adsorption and catalysis: The effect of confinement on chemical
reactions. Appl Surf Sci. 2005;252:766–77.
50. Pan X, Fan Z, Chen W, Ding Y, Luo H, Bao X. Enhanced ethanol production inside carbon-­
nanotube reactors containing catalytic particles. Nat Mater. 2007;6:507–11.
51. Dabbawala AA, Mishra DK, Hwang J-S. Selective hydrogenation of D-glucose using amine
functionalized nanoporous polymer supported Ru nanoparticles based catalyst. Catal Today.
2016;265:163–73.
52. Doluda V, Grigorev M, Matveeva V, Sulman E, Sulman M, Lakina N, Molchanov V, Rebrov
EV. Evaluation of D-glucose hydrogenation catalysts stability in different reactor systems.
WSEAS Transact Biol Biomed. 2016;13:44–51.
53. Boyers GG, Flushing NY. Hydrogenation of mono- and disaccharides to polyols. US Patent
2,868,847 (1959).
54. Lepper H, Schütt H. Process for the continuous preparation of polyhydric alcohols. US Patent
4,520,211 (1985).
55. Déchamp N, Gamez A, Perrard A, Gallezot P.  Kinetics of D-glucose hydrogenation in a
trickle-bed reactor. Catal Today. 1995;24:29–34.
56. Tukac V.  D-glucose hydrogenation in a trickle-bed reactor. Collect Czechoslov Chem
Commun. 1997;62:1243–8.
306 J.R. Ochoa-Gómez and T. Roncal

57. Arena BJ. Deactivation of ruthenium catalysts in continuous D-glucose hydrogenation. Appl

Catal A General. 1992;87:219–29.
58. Gallezot P, Nicolaus N, Flèche G, Fuertes P, Perrad A. D-glucose hydrogenation on ruthe-
nium catalysts in a trickle-bed reactor. J Catal. 1998;180:51–5.
59. Aho A, Roggan S, Eränen K, Salmia T, Murzin DU. Continuous hydrogenation of D-glucose
with ruthenium on carbon nanotube catalysts. Cat Sci Technol. 2015;5:953–9.
60. Yamaguchi A, Sato O, Mimura N, Hirosaki Y, Kobayashi H, Fukuoka A, Shirai M. Direct
production of sugar alcohols from wood chips using supported platinum catalysts in water.
Catal Commun. 2014;54:22–6.
61. Li J, Soares HSMP, Moulijn JA, Makkee M. Simultaneous hydrolysis and hydrogenation of
cellobiose to sorbitol in molten salt hydrate media. Cat Sci Technol. 2013;3:1565–72.
62. Almeida JMAR, Da Vià L, Carà D, Carvalho Y, Romano PN, Peña JAO, Smith L, Sousa-­
Aguiar EF, Lopez-Sanchez JA. Screening of mono- and bi-functional catalysts for the one-­
pot conversion of cellobiose into sorbitol. Catal Today. 2016; In press, corrected proof
available.
63. Hartstra L, Bakker L, van Westen HA. Hydrogenation of carbohydrates. US Patent 2518235
(1950).
64. Jacobs P, Hinnekens H. Single-step catalytic process for the direct conversion of polysac-
charides to polyhydric alcohols. EP Patent 0329923 (1989).
65. Kasehagen L. Hydrogenation of carbohydrates. US Patent 2,968,680 (1961).
66. Kruse WM, Wright LW. Polyhydric alcohol production using ruthenium zeolite catalyst. US
Patent 3,963,788 (1976).
67. Van de Vyver S, Geboers J, Jacobs PA, Sels BF. Recent advances in the catalytic conversion
of cellulose. ChemCatChem. 2011;3:82–94.
68. Yabushita M, Kobayashi H, Fukuoka A. Catalytic transformation of cellulose into platform
chemicals. Appl Catal B Environ. 2014;145:1–9.
69. Li Y, Liao Y, Cao X, Wang T, Ma L, Long J, Liu Q, Xua Y. Advances in hexitol and ethylene
glycol production by one-pot hydrolytic hydrogenation and hydrogenolysis of cellulose.
Biomass Bioenergy. 2015;74:148–61.
70. Balandin AA, Vasyunina NA, Barysheva GS, Chepigo SV, Dubinin MM. Letters to the editor.
Bull Acad Sci USSR Div Chem Sci. 1957;6:403–4.
71. Balandin AA, Vasyunina NA, Chepigo SV, Barysheva GS. Hydrolytic hydrogenation of cel-
lulose. Dokl Akad Nauk SSSR. 1959;128:941–4.
72. Geboers J, Van de Vyver S, Carpentier K, Blochouse K, Jacobs P, Sels B. Efficient catalytic
conversion of concentrated cellulose feeds to hexitols with heteropolyacids and Ru on car-
bon. Chem Commun. 2010;46:3577–9.
73. Palkovits R, Tajvidi K, Procelewska J, Rinaldi R, Ruppert A. Hydrogenolysis of cellulose
combining mineral acids and hydrogenation catalysts. Green Chem. 2010;12:972–8.
74. Fukuoka A, Dhepe PL. Catalytic conversion of cellulose into sugar alcohols. Angew Chem.
2006;118:5285–7.
75. Dhepe PL, Fukuoka A. Cracking of cellulose over supported metal catalysts. Catal Surv Jpn.
2007;11:186–91.
76. Jollet V, Chambon F, Rataboul F, Cabiac A, Pinel C, Guillon E, Essayem N. Non-catalyzed
and Pt/γ-Al2O3-catalyzed hydrothermal cellulose dissolution-conversion: influence of the
reaction parameters and analysis of the unreacted cellulose. Green Chem.
2009;11:2052–60.
77. Han JW, Lee H. Direct conversion of cellulose into sorbitol using dual-functionalized cata-
lysts in neutral aqueous solution. Catal Commun. 2012;19:115–8.
78. Luo C, Wang S, Liu HC. Cellulose conversion into polyols catalyzed by reversibly formed
acids and supported ruthenium clusters in hot water. Angew Chem. 2007;119:7780–3.
79. Deng W, Tan X, Fang W, Zhang Q, Wang Y. Conversion of cellulose into sorbitol over carbon
nanotube-supported ruthenium. Catal Lett. 2009;133:167–74.
80. Ribeiro LS, Órfão JJM, Pereira MFR. Enhanced direct production of sorbitol by cellulose
ball-milling. Green Chem. 2015;17:2973–80.
9  Production of Sorbitol from Biomass 307

81. Suzuki T, Nakagami J.  Effect of crystallinity of microcrystalline cellulose on the com-
pactability and dissolution of tablets. Eur J Pharm Biopharm. 1999;47:225–30.
82. Isik M, Sardon H, Mecerreyes D. Ionic liquids and cellulose: dissolution, chemical modifica-
tion and preparation of new cellulosic materials. Int J Mol Sci. 2014;15:11922–40.
83. Ignatyev IA, Van Doorslaer C, Mertens PGN, Binnemans K, De Vos DE. Reductive splitting
of cellulose in the ionic liquid 1-butyl-3-methylimidazolium chloride. ChemSusChem.
2010;3:91–6.
84. Creighton HJ, Hales RA.  Electrolytic process for reducing sugars. US Patent 2,458,895
(1949).
85. Hefti HR, Kolb W. Electrolytic reduction of sugars. US Patent 2,507,973 (1950).
86. Parl K, Pintauro PN, Baizer MM, Nobe K. Flow reactor studies of the paired electro-­oxidation
and electroreduction of D-glucose. J  Electrochem Soc Electrochem Sci Technol.
1985;132:1850–5.
87. Li H, Li W, Guo Z, Gu D, Cai S, Fujishima A. The paired electrochemical synthesis of glu-
conic acid and sorbitol. Collect Czechoslov Chem Commun. 1995;60:928–34.
88. Silveira MM, Jonas R.  The biotechnological production of sorbitol. Appl Microbiol
Biotechnol. 2002;59:400–8.
89. Sprenger GA. Carbohydrate metabolism in Zymomonas mobilis: a catabolic highway with
some scenic routes. FEMS Microbiol Lett. 1996;145:301–7.
90. Viikari L.  Formation of levan and sorbitol from sucrose by Zymomonas mobilis. Appl
Microbiol Biotechnol. 1984;19:252–5.
91. Barrow KD, Collins JG, Leight DA, Rogers PL, Warr RG. Sorbitol production by Zymomonas
mobilis. Appl Microbiol Biotechnol. 1984;20:225–32.
92. Leigh D, Scopes RK, Rogers PL. A proposed pathway for sorbitol production by Zymomonas
mobilis. Appl Microbiol Biotechnol. 1984;20:413–5.
93. Zachariou M, Scopes RK. D-glucose-fructose oxidoreductase, a new enzyme isolated from
Zymomonas mobilis that is responsible for sorbitol production. J  Bacteriol.
1986;167:863–9.
94. Scopes RK, Rogers PL, Leigh DA. Method for the production of sorbitol and gluconate. US
Patent 4,755,467 (1988).
95. Chun UH, Rogers PL. The simultaneous production of sorbitol from fructose and gluconic
acid from D-glucose using an oxidoreductase of Zymomonas mobilis. Appl Microbiol
Biotechnol. 1988;29:19–24.
96. Ichikawa Y, Kitamoto Y, Kato N, Mori N. Preparation of gluconic acid and sorbitol. European
Patent Application 322,723 (1989).
97. Bringer-Meyer S, Sahm H. Process for obtaining sorbitol and gluconic acid by fermentation,
and cell material suitable for this purpose. US Patent 5,017,485 (1991).
98. Rehr B, Wilhelm C, Sahm H. Production of sorbitol and gluconic acid by permeabilized cells
of Zymomonas mobilis. Appl Microbiol Biotechnol. 1991;35:144–8.
99. Rehr B, Sahm H. Process for obtaining sorbitol and gluconic acid or gluconate. US Patent
5,102,795 (1992).
100. Rehr B, Sahm H.  Process for obtaining sorbitol and gluconic acid or gluconate using
Zymomonas mobilis. US Patent 5,190,869 (1993).
101. Jang KH, Jung SJ, Chang HS, Chun UH. Improvement of the process for sorbitol production
with Zymomonas mobilis immobilized in κ-carrageenan. Process Biochem.
1996;31:485–92.
102. Ferraz HC, Alves TLM, Borges CP. Coupling of an electrodialysis unit to a hollow fiber bio-
reactor for separation of gluconic acid from sorbitol produced by Zymomonas mobilis per-
meabilized cells. J Membr Sci. 2001;191:43–51.
103. Wisbeck E, Silveira MM, Ninow J, Jonas R. Evaluation of the flocculent strain Zymomonas
mobilis Z1–81 for the production of sorbitol and gluconic acid. J  Basic Microbiol.
1997;6:445–9.
308 J.R. Ochoa-Gómez and T. Roncal

104. Silveira MM, Wisbeck E, Lemmel C, Erzinger GS, Lopes da Costa JP, Bertasso M, Jonas
R. Bioconversion of D-glucose and fructose to sorbitol and gluconic acid by untreated cells
of Zymomonas mobilis. J Biotechnol. 1999;75:99–103.
105. Ro H, Kim H. Continuous production of gluconic acid and sorbitol from sucrose using inver-
tase and an oxidoreductase of Zymomonas mobilis. Enzym Microb Technol. 1991;13:920–4.
106. Kim DM, Kim HS.  Continuous production of gluconic acid and sorbitol from Jerusalem
artichoke and D-glucose using an oxidoreductase of Zymomonas mobilis and inulinase.
Biotechnol Bioeng. 1992;39:336–42.
107. Cazetta ML, Celligoi MPC, Buzato JB, Scarmino IS, Da Silva RSF. Optimization study for
sorbitol production by Zymomonas mobilis in sugar cane molasses. Process Biochem.
2005;40:747–51.
108. An K, Hu F, Bao J.  Simultaneous saccharification of inulin and starch using commercial
glucoamylase and the subsequent bioconversion to high titer sorbitol and gluconic acid. Appl
Biochem Biotechnol. 2013;171:2093–104.
109. Liu C, Dong H, Zhong J, Ryu DD, Bao J. Sorbitol production using recombinant Zymomonas
mobilis strain. J Biotechnol. 2010;148:105–12.
110. Pedruzzi I, da Silva EAB, Rodrigues AE. Production of lactobionic acid and sorbitol from
lactose/fructose substrate using GFOR/GL enzymes from Zymomonas mobilis cells: a kinetic
study. Enzym Microb Technol. 2011;49:183–91.
111. Gollhofer D, Nidetzky B, Fürlinger M, Kulbe KD. Efficient protection of D-glucose-fructose
oxidoreductase from Zymomonas mobilis against irreversible inactivation during its catalytic
action. Enzym Microb Technol. 1995;17:235–40.
112. Nidetzky B, Fürlinger M, Gollhofer D, Scopes RK, Haltrich D, Kulbe KD. Improved opera-
tional stability of cell-free D-glucose-fructose oxidoreductase from Zymomonas mobilis for
the efficient synthesis of sorbitol and gluconic acid in a continuous ultrafiltration membrane
reactor. Biotechnol Bioeng. 1997;53:623–9.
113. Silva-Martinez M, Haltrich D, Novalic S, Kulbe KD, Nidetzky B. Simultaneous enzymatic
synthesis of gluconic acid and sorbitol: continuous process development using D-glucose-­
fructose oxidoreductase from Zymomonas mobilis. Appl Biochem Biotechnol.
1998;70–72:863–8.
114. Tani Y, Vongsuvanlert V.  Sorbitol production by a methanol yeast, Candida boidinii
(Kloeckera sp.) No. 2201. J Ferment Technol. 1987;65:405–11.
115. Vongsuvanlert V, Tani Y. Characterization of D-sorbitol dehydrogenase involved in D-sorbitol
production of a methanol yeast, Candida boidinii (Kloeckera sp.) No. 2201. Agric Biol
Chem. 1988;52:419–26.
116. Duvnjak Z, Turcotte G, Duan ZD. Production of sorbitol and ethanol from Jerusalem arti-
chokes by Saccharomyces cerevisiae ATCC 36859. Appl Microbiol Biotechnol.
1991;35:711–5.
117. Nissen L, Perez-Martinez G, Yebra MJ. Sorbitol synthesis by an engineered Lactobacillus
casei strain expressing a sorbitol-6-phosphate dehydrogenase gene within the lactose operon.
FEMS Microbiol Lett. 2005;249:177–83.
118. De Boeck R, Sarmiento-Rubiano LA, Nadal I, Monedero V, Pérez-Martínez G, Yebra
MJ. Sorbitol production from lactose by engineered Lactobacillus casei deficient in sorbitol
transport system and mannitol-1-phosphate dehydrogenase. Appl Microbiol Biotechnol.
2010;85:1915–22.
119. Ladero V, Ramos A, Wiersma A, Goffin P, Schanck A, Kleerebezem M, Hugenholtz J, Smid
EJ, Hols P. High-level production of the low-calorie sugar sorbitol by Lactobacillus planta-
rum through metabolic engineering. Appl Environ Microbiol. 2007;73:1864–72.
120. van Gorp K, Boerman E, Cavenaghi CV, Berben PH. Catalytic hydrogenation of fine chemi-
cals: sorbitol production. 1999 52;349–61.
121. Puritech. Sorbitol/mannitol Separation. http://www.puritech.be/food-processing/sorbitol-­
mannitol (2016). Last accessed 15 July 2016.
9  Production of Sorbitol from Biomass 309

122. Ferraz HC, Borges CP, Alves TLM. Produção de sorbitol e ácido glicônico por células per-
meabilizadas e imobilizadas de Zymomonas mobilis com separação simultânea dos produtos
por eletrodiálise. In: Proceedings of the 13th national symposium on fermentation.
Terezópolis, Brazil; (2000)
123. Silveira MM, Lopes da Costa JP, Jonas R. Processo de produção e recuperação de sorbitol e
ácido glucônico ou gluconato. Brazilian Patent PI 9.403.981-0 (1994).
124. Jiang B, Li Z-G, Dai J-Y, Zhang D-J, Xiu ZL. Aqueous two-phase extraction of 2,3-­butanediol
from fermentation broths using an ethanol/phosphate system. Process Biochem.
2009;44:112–7.
125. Fang Z. Method for the dissolving and rapid hydrolyzing of lignocellulose biomass, device
thereof and use of the same. US Patent 9,243,303 (2016).
126. Cantero DA, Bermejo D, Cocero MJ. High glucose selectivity in pressurized water hydrolysis
of cellulose using ultra-fast reactors. Bioresour Technol. 2013;135:697–703.
127. Ochoa-Gómez JR, García-Luis A, Fernández-Carretero FJ, Lorenzo-Ibarreta L, Prieto
S.  Method for manufacturing 2,3-butanediol. European Patent Application EP14198812
(2014).
128. Pletcher D, Li X, Wang S. A comparison of cathodes for zero gap alkaline water electrolysers
for hydrogen production. Int J Hydrog Energy. 2012;37:7429–35.