Biochemical Engineering Journal 40 (2008) 500–506

Contents lists available at ScienceDirect

Biochemical Engineering Journal

journal homepage: www.elsevier.com/locate/bej

Release of ferulic acid from corn cobs by alkaline hydrolysis

Paolo Torre a , Bahar Aliakbarian a , Beatriz Rivas a,b , José Manuel Domı́nguez b , Attilio Converti a,∗
a
Department of Chemical and Process Engineering, Genoa University, Via Opera Pia 15, I-16145 Genoa, Italy
b
Department of Chemical Engineering, Vigo University (Campus of Ourense), Polytechnical Building, As Lagoas, E-32004 Ourense, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The process of corn cobs alkaline hydrolysis to produce solutions with high hydroxy-cinnamic acids con-
Received 16 November 2007 tent was investigated. In particular the attention was focused on the solubilisation of ferulic acid (FA)
Received in revised form 16 January 2008 and related compounds, mainly p-coumaric acid (p-CA). Although these compounds have applications as
Accepted 8 February 2008
antioxidants, the purpose of this work was to obtain FA solutions that can be used as feedstock for the
biotechnological production of vanillin in future studies. The effects of different concentrations of NaOH
Keywords:
(0.2 ≤ Ca ≤ 2.0N) and solid/liquid ratios (0.028 ≤ S/L ≤ 0.168 g/g) on the solubilisation of FA versus time have
Corn cobs
been investigated at room temperature. Optimal hydrolysis conditions (Ca = 0.5N, S/L = 0.084 g/g after 6 h)
Alkaline hydrolysis
Ferulic acid
ensured the production of hydrolysates with relatively high contents of both FA (1171 ± 34 mg/L) and p-
Coumaric acid coumaric acid (2156 ± 64 mg/L), which can be used in future studies for the microbial transformation into
Phenolic compounds vanillin.
Oligosaccharides © 2008 Elsevier B.V. All rights reserved.

1. Introduction respectively. Whereas the former is principally associated with

Corn cobs is an important by-product of corn industry: for
every 100 kg of corn grain approximately 18 kg of corn cobs are
produced. A large quantity of corn cobs remains unused as ligno-
cellulosic waste or used as animal feed. Such a feedstock contains
31.7% cellulose, 34.7% hemicellulose, 20.3% lignin and 3.4% acetyl
groups (oven-dry basis) [1], which can be fractioned in separate
streams.
Lignin is a well-known component of secondary (non-growing)
cell walls, where it replaces water and completely prevents fur-
ther growth. It is mainly made of phenolic acids linked through
ester, ether, or acetal bonds to other components of the plant
cell wall [2]. The major phenolic compounds identified in both
primary and secondary cell walls of graminaceous plants, and
in particular of cereals, are cinnamic acids such as ferulic acid
(4-hydroxy-3-methoxycinnamic acid; FA) and p-coumaric acid (4-
hydroxycinnamic acid; p-CA) [3]. These cinnamic acids act in the
cross-linking of plant cell walls and are precursors of a variety of
compounds that play an important role in plant defence responses
[4]. FA and p-CA are very abundant, representing together up to 1.5%
by weight of cereal cell walls [5].
Since p-CA and FA are bifunctional, they are able to form ester
or ether linkages by reaction of their carboxyl or phenolic groups,
∗ Corresponding author. Tel.: +39 010 353 2593; fax: +39 010 353 2586.
E-mail address: converti@unige.it (A. Converti).
lignin, the latter is mainly esterified with hemicellulose [6]. The
linkage between FA and lignin is not yet fully understood, because it
largely depends on the raw material, and the results are influenced
by the fractionation methods employed. Mature wheat endosperm
contains a feruloylated arabinoxylan in the secondary cell wall,
where ferulic acid seems to be bound via ester linkages to the struc-
tural polysaccharides [6,7], while dimers of ferulic acid can serve
to cross-link hemicellulose [8]. In cereals FA is ester-linked to O-
5 of arabinose of cell wall-bound arabinoxylans [6,9], whereas in
sugar beet it is associated mainly with the pectine side chains and
linked to the O-2 of arabinose (50–60%) and the O-6 of galactose
(40–50%) [8,10]. Cell walls of cereal straw such as wheat straw and
grass are characterized by high levels of ester-linked p-CA and FA
[11]. Higher proportions of ferulates are located mainly in cereal
bran, maize bran being an especially rich source of FA [5], while
p-CA is predominant in cereal stems [12].
Alkaline treatments dissolve lignin by cleavage of ester link-
ages in lignin–polysaccharide complexes, thus releasing phenolic
acids [13]. The high solubility of cereal lignin in alkaline solution
might originate either from alkali cleavage of ferulic acid cross-
link between lignin and hemicelluloses or from modification of
lignin polyelectrolyte properties induced by free carboxyl groups
of phenolic acids ethers [14].
Chemical and enzymatic treatments of agro-industrial wastes
or surpluses can allow FA and p-CA to be released and then to use
in further transformations. There is great interest in the potential of
FA and related compounds either as antioxidants, i.e. in food preser-

1369-703X/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.bej.2008.02.005
 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506
vation because of their ability to inhibit fatty acid peroxidation [15],
or as feedstocks for the high-added value vanillin bio-production,
one of the main flavourings and aroma compounds [16], through
an integrated, eco-sustainable process.
In this work we investigated the solubilisation of FA and other
compounds from corn cobs by alkaline hydrolysis employing NaOH
as a catalyst. The effects of NaOH concentration and solid/liquid
ratio on the solubilisation process were assessed. Hydrolysates of
alkaline treatments were characterized in their content of free
and total phenolic compounds, sugars, oligosaccharides (OS) and
organic acids, with the aim of optimizing FA release for future
vanillin production.
2. Materials and methods
2.1. Raw material and chemical characterization
Corn cobs were collected locally (Ourense, Galicia, Spain), milled
to a particle size smaller than 1 mm, homogenized in a single lot
and stored until use. Aliquots from the above homogenized lot
were subjected to moisture determination (ISO 638 method) and
to quantitative acid hydrolysis (TAPPI T13m method) with 72%
sulphuric acid. The solid residue after hydrolysis was expressed
as Klason lignin. Hydrolysates were analyzed for monosaccha-
rides (glucose coming from cellulose; xylose and arabinose from
hemicelluloses) and acetic acid (from acetyl groups) by high per-
formance liquid chromatography (HPLC) as reported elsewhere
[17]. The results allowed the determination of the sample con-
tents of cellulose and starch (based on glucose present in liquors),
hemicellulosic polysaccharide constituents (based on xylose and
arabinose), and acetyl groups (based on acetic acid).
Corn cobs were subjected to non-isothermal treatments with
water (auto-hydrolysis) using 8 g water/g dry solid in a Parr reactor
model 4563 M (Parr Instrument Company, Moline, IL). The reaction
medium was heated according to the standard temperature profile
to reach a final temperature of 202 ◦ C. The solid and liquid phases
were separated by filtration. The solid residue of corn cobs from
auto-hydrolysis treatment (CAR) was subjected to moisture deter-
mination and to quantitative acid hydrolysis with 72% sulphuric
acid following standard methods and was utilized as substrate for
the alkaline hydrolysis experiments performed in this study.
2.2. Alkaline hydrolysis
Alkaline hydrolysis experiments were carried out by employ-
ing either raw corn cobs or CAR. Reactions were performed at room
temperature using NaOH as catalysts in a solid/liquid ratio between
0.028 and 0.084 g of raw material/g of NaOH solution. NaOH con-
centration and reaction times were varied in the ranges 0.2–2.0N
and 1–24 h, respectively. All the treatments were conducted in trip-
licate, and the standard deviation of each determination did not
exceed 3%.
The liquors obtained from alkaline treatments were separated
from the solid fraction by vacuum filtration through common lab-
oratory paper filters. Aliquots from liquors were characterized
through the determination of free and total phenolic compounds,
sugars, oligosaccharides, organic acids and solubilised non-volatile
compounds (SNVC) content. Solid residues from treatments were
washed with water to remove the residual alkalinity, air-dried, and
analyzed for the moisture content to measure the raw material
dissolved fraction, and subjected to the same analytical determi-
nations as for the raw material.
501
2.3. Analytical methods
Free phenolic acids concentrations were determined by HPLC,
model 1100 (Agilent, Palo Alto, CA), using a UV detector (at 280 nm)
and a Vydac 201TP54 C18 reverse-phase column (Grace, Deerfield,
IL) with a C18 guard column. Samples were neutralized and filtered
through 0.45 ␮m-pore membranes (Millipore, Vimodrome, Italy).
Separation was achieved using a linear gradient of two solvents:
solvent A (1.0% acetic acid in water, v/v), solvent B (1% acetic acid,
9% water, 90% methanol, v/v). The linear gradient was run at 35 ◦ C
in 60 min from 0% to 52% of B at a flow rate of 1 mL/min.
Glucose, xylose, arabinose and acetic acid were determined
directly by the same HPLC as above, using a Refractive Index detec-
tor and a Supelcogel H59304-U column (Sigma–Aldrich, Bellefonte,
PA). Samples were neutralized and filtered through 0.45 ␮m-pore
membranes (Millipore). Separation was achieved using as mobile
phase 0.005 M H2 SO4 at a flow rate of 0.5 mL/min at 50 ◦ C.
Oligosaccharides were measured by an indirect method based
on quantitative acid hydrolysis of the liquors. For analytical
purposes, samples of liquors were subjected to quantitative posthy-
drolysis (treatment with 4% sulphuric acid at 121 ◦ C for 20 min), and
the reaction products were assayed by the same HPLC method as
for direct monomers determination. The increase in the concentra-
tions of monosaccharides and acetic acid caused by posthydrolysis
provided a measure of the oligomer concentration, expressed as
xyloligoosaccharides, arabinooligosaccharides, glucooligosaccha-
rides, and acetyl groups concentration [17].
Total phenolic compounds (TPC) concentration was measured
using the Folin–Ciocalteu assay [18]. Briefly, 4.8 mL of distilled
water, 0.2 mL of sample, and 0.5 mL of Folin–Ciocalteu reagent were
added to 25 mL graduated flasks. After mixing, 1 mL of 20% sodium
carbonate solution was added, and distilled water was added to
reach the final volume of 10 mL. Solutions were mixed and allowed
to stand in dark at room temperature for 1 h. Sample aliquots
were used to determine total phenols concentration using a spec-
trophotometer, model Lambda 25 (PerkinElmer, Waltham, MA),
at a wavelength of 725 nm. TPC concentration was standardized
against caffeic acid and expressed in mg/L of caffeic acid equiv-
alents (CAE). To this purpose, we used a calibration straight line
described with the equation ABS725 = 2.06 × 10−3 CAE within the
range 100–1000 mg/L of CAE with R2 = 0.9962.
The content of unidentified phenolic compounds (UPC) was
calculated from the TPC concentration and the free phenolic com-
pounds concentrations using the equation:
UPC = TPC − CFA ˇFA − Cp-CA ˇp-CA (1)
where CFA and Cp-CA are the ferulic and p-coumaric acid con-
centrations (mg/L), while ˇFA (1.2477 mg of CAE mg−1 of FA)
and ˇp-CA (1.4057 mg of CAE mg−1 of p-CA) are the conversion
factors to express the concentrations of ferulic and p-coumaric
acids as mg/L of CAE. These conversion factors resulted from
the ratios between the extinction coefficients of caffeic acid
(εCAF = 2.1 ␮g L−1 cm−1 ), ferulic acid (εFA = 1.7 ␮g L−1 cm−1 ) and p-
coumaric acid (εp-CA = 1.5 ␮g L−1 cm−1 ), respectively, obtained by
the Folin–Ciocalteu assay.
The content of solubilised non-volatile compounds (SNVC) of
the liquor was measured by oven drying until constant weight,
taking into account the contribution of NaOH to the dry weight.
The solid phase from alkaline hydrolysis was recovered by filtra-
tion, washed with distilled water to remove the residual alkalinity,
dried at room temperature and subjected to the same chemical
analysis as for the feedstock material.
502 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506

Table 1
Composition of the liquors from alkaline hydrolysis either of the corn cobs or the
residue of corn cobs auto-hydrolysis (CAR)
Raw material p-CAa (mg/L) FAb (mg/L)
Corn cobs 468 248
CAR 692 194
a
p-CA: p-coumaric acid.
b
FA: ferulic acid.
c
SNVC: solubilised non-volatile compounds.
3. Results and discussion
The alkaline hydrolysis is a well-known technique utilized to
recover the cellulosic fraction of lignocellulosic materials in solid
phase [19,20]. Such a treatment is able to modify the consistency
of cellulose structure, thus making it promptly available for dif-
ferent uses such as its enzymatic hydrolysis [21]. At the same time,
the phenolic compounds present in the lignin fraction of lignocellu-
losics are solubilised. However, the present use of high temperature
and alkali concentration to exalt cellulose recovery leads to the
degradation of a significant portion of these compounds [22,23]
as well as to undesired, partial solubilisation of oligosaccharides.
Therefore, the use of milder conditions is needed to optimize the
release of ester-linked phenolic acids for possible biotransforma-
tion into vanillin.
Different alkalis such as NaOH, Ca(OH)2 , NH4 OH, and H2 O2 are
frequently utilized as catalysts for hydrolysis. Among these, NaOH
was selected in this work because, notwithstanding its relatively
low delignification yield, it was shown to be more selective, with
respect to the other agents, in the release of phenolic compounds
such as p-coumaric and ferulic acids [24].
3.1. Raw material selection
Two different materials, namely corn cobs and CAR, were ini-
tially tested under non-optimized conditions for the release of
ferulic acid as a precursor for vanillin bio-production, i.e. 1.0N
NaOH for 24 h using a solid/liquid ratio of 0.028 g/g of solution.
Table 1 shows the composition of liquors from alkaline hydroly-
sis of both materials, in terms of ferulic acid, p-coumaric acid and
SNVC concentrations, while Table 2 summarizes the experimental
data of weight loss and composition of the corresponding dry solid
residues either before or after treatment.
The alkaline hydrolysis of both materials produced solutions
with relevant concentrations of hydroxy-cinnamic acids resulting
from breaking of ␣-ester bonds present in lignin–polysaccharide
complexes. The ferulic acid content of liquor from corn cobs alkaline
hydrolysis (248.5 mg/L) was about 28% higher than that obtained
responsible also for the release of this ferulic acid portion, thus
making CAR a poorer material.
Despite of these promising results in terms of released FA, the
SNVCc (g/L) alkaline treatment was responsible for the undesired solubilisation
18.5 of no less than 80% of the starting hemicellulose present in corn cobs
15.4 and 74% of the residual 1% in CAR (Table 2). On the contrary, the
higher release of p-CA from CAR (692 mg/L), consequent to alkaline
hydrolysis, confirms its ligninic localization. p-CA accumulation in
the cell wall is in fact mainly correlated to the extent of lignifica-
tion, whereas ferulic acid was reported to be especially associated
with polysaccharide deposition [25]. The results of Table 2 also indi-
cate that only a little portion of corn cobs cellulose was solubilised
by the alkaline treatment, whereas its percentage in CAR remark-
ably decreased. This result suggests that the auto-hydrolysis was
mainly responsible for a significant modification of cellulose struc-
ture, which made it more susceptible to subsequent alkali attack.
On the contrary, as earlier mentioned, hemicellulose was effectively
hydrolyzed by both hydrolysis treatments. Finally, the solubilisa-
tion of about 53% of the ligninic fraction of corn cobs and 33% of
that of CAR suggests the presence of a more soluble fraction effec-
tively removed by the auto-hydrolysis and a more resistant one
susceptible to alkali [26].
These results on the whole suggest that corn cobs is a better
raw material than CAR for ferulic acid release, therefore the sub-
sequent optimization study was carried out using only the former
raw material.
3.2. Effect of solid/liquid ratio
In order to get in the hydrolyzate ferulic acid concentrations
close to those consistent with vanillin bio-production [16], alkaline
hydrolysis of corn cobs was performed with 1.0N NaOH at three
different solid/liquid (S/L) ratios, namely 0.028, 0.084 and 0.168 g/g,
i.e. in a range close to the value proposed by Mussatto et al. [27]
for sugarcane bagasse (0.05 g/g). Table 3 lists the results, in terms
of released p-CA, FA and SNVC versus time, obtained only at the
lowest two ratios, since at the highest ratio a too viscous mixture
was obtained.
The kinetics of p-CA and FA solubilisation was very similar under
both conditions, ensuring maximum concentrations of these prod-
ucts after about 6–16 h. The clear decrease in their levels after
longer hydrolysis time was likely due to their degradation conse-
quent to their well-known antioxidant power [28]. In particular,
an increase in S/L from 0.028 to 0.084 g/g led to a 3-fold increase in
Table 3
Composition of the liquors from alkaline hydrolysis of corn cobs with 1.0N NaOH at
two different solid/liquid ratios (S/L)

with CAR, which is consistent with the ferulated-hemicellulosic Time (h) p-CAa (mg/L) FAb (mg/L) SNVCc (g/L)
structure detected in most cereals including corn cobs [6,8]. (a) S/L = 0.028 g/g
Because the auto-hydrolysis of corn cobs is able to solubilise most of 1 328 209 8.35
the hemicellulose fraction [17], such a treatment could have been 2 450 286 12.5
4 526 278 15.4
6 568 315 17.2
Table 2 16 633 347 18.3
Weight loss and composition of dry samples of the corn cobs and the residue of corn 24 468 248 18.5
cobs auto-hydrolysis (CAR), either before (I) or after (II) alkaline hydrolysis
(b) S/L = 0.084 g/g
Weight loss (%) Composition (%) 1 1012 718 13.6
Cellulose Hemicellulose Lignin Acetyl groups 2 1964 991 17.3
4 2198 1097 26.4
Corn cobs 6 2553 1100 31.2
I – 31.7 34.7 20.3 3.40 16 2278 1015 31.8
II 65.2 37.0 19.4 39.0 0.00 24 1896 892 33.1
CAR a
p-CA: p-coumaric acid.
I – 59.1 9.40 22.6 1.00 b
FA: ferulic acid.
II 54.2 66.7 5.28 23.0 0.00 c
SNVC: solubilised non-volatile compounds.
 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506
Fig. 1. Effect of NaOH concentration (Ca ) on the release of ferulic acid from corn
cobs by alkaline hydrolysis: Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (), Ca = 2.0N ()
employing a solid–liquid ratio of 0.084 g/g.
maximum ferulic acid concentration (up to 1100 mg/L) and a 4-fold
increase in p-coumaric acid concentration (up to 2553 mg/L), corre-
sponding to 1.36 and 3.16% of the starting raw material, respectively.
A comparison of these results with the starting ferulic acid lev-
els previously reported for vanillin bio-production from synthetic
solutions (0.6–1.0 g/L) [16] demonstrates that S/L = 0.084 g/g was by
far the best hydrolysis ratio for our purposes.
On the contrary, the hydrolysate content of SNVC progressively
raised along the whole hydrolysis time, but a 3-fold increase in S/L
led only to a 2-fold increase in their level after 24 h (from 18.5 to
33.1 g/L), which means that a reduction of the solubilised fraction
took place (from 65.2% to 44.7%). Simple mass balances (see later
on) demonstrate that the high solubilised non-volatile compounds
content was the main result of the solubilisation of oligosaccharides
from the hemicellulosic fraction. As an example, at S/L = 0.084 g/g,
after 24 h-treatment, no less than 60% of hemicelluloses were sol-
ubilised, whereas almost all the starting cellulose (98.1%) was kept
in the solid phase.
3.3. Effect of the catalyst concentration and the hydrolysis time
The alkali concentration and the hydrolysis duration are recog-
nized as being key factors in the release of phenolic compounds
[29], because a mild treatment usually leads to limited solu-
bilisation, whereas too severe conditions can provoke products
Fig. 2. Effect of NaOH concentration (Ca ) on the release of p-coumaric acid from
corn cobs by alkaline hydrolysis: Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (), Ca = 2.0N
() employing a solid–liquid ratio of 0.084 g/g.
503
degradation [19]. For these reasons, this part of the work was
focused on the optimum NaOH concentration for lignin solubili-
sation, with particular concern to ferulic acid release. Corn cobs
was treated at S/L = 0.084 g/g at room temperature at different con-
centrations (0.2–2.0N) and hydrolysis time (0–24 h), and the liquid
phase was analyzed for the concentrations of FA and p-CA, total phe-
nolic compounds, monosaccharides, acetic acid, oligosaccharides,
and SNVC.
Figs. 1 and 2 show the main results of FA and p-CA solubilisation
obtained in these experiments. An increased catalyst concentra-
tion accelerated the solubilisation of ferulic acid, whose maximum
concentration (about 1100–1200 mg/L) required no less than 6 h
at 0.5N and only 1–2 h at 2.0N (Fig. 1). A clear decrease in fer-
ulic acid concentration took place after the achievement of this
maximum threshold, as the likely result of oxidative degradation
whose rate increased from 5.3 to 18.3 mg L−1 h−1 when Ca was
raised from 0.5 to 2.0N. No significant degradation was detected
at the lowest catalyst level (0.2N). Among the identified pheno-
lic compounds, p-CA was released at the highest concentration. It
exhibited a time behaviour (Fig. 2) similar to that of ferulic acid and
reached a maximum threshold concentration of 2783 mg/L after 6 h
of hydrolysis with 2.0N NaOH. On the contrary, using the most dilute
catalyst solution ensured only 62% of this level, thus confirming
the low effectiveness of alkaline treatments under too mild condi-
tions. Also this product was subject to time-dependent degradation
that occurred at a rate even higher than that of FA and increasing
with the catalyst level (27.5 and 35.2 mg L−1 h−1 at Ca = 1.0 and 2.0N,
respectively).
As well known, saponification of corn cob at room tempera-
ture, in general, cleaves ester linkages between lignin and cinnamic
acids, polysaccharide and cinnamic acids, and diferulic acid bridges
between polysaccharides or direct ester links between lignin and
polysaccharide [19]. Therefore, the high concentrations of FA and
p-CA in the alkali-treated hydrolysates of corn cobs suggest that
most of these acids were originally present in esterified form in the
raw material. These results agree with those of a number of previ-
ous studies. For example, Jeffries [6] reported that ester-linked p-CA
was likely to be exclusively associated with lignin, and the majority
of its molecules in grass are linked to the ␥-position of the side chain
of lignin molecules. FA, on the contrary, is both etherified to lignin
and arabinoxylans and forms an alkali-labile cross-link between
these two cell wall polymers [30]. Such a relatively higher content
of alkali-labile cross-linkages within the lignin network or between
lignin and polysaccharides might explain the fast and easy solubil-
isation of both phenolic acids by the alkaline treatments shown in
Figs. 1 and 2. Because of its higher level in the raw material, the
content of p-CA was always twice that of FA, irrespective of both
the process time and the hydrolysis conditions. These results are
in close agreement with those obtained for the alkaline hydroly-
sis of sweet sorghum stalk (p-CA/FA = 2.5–2.8) [25] and of maize
stover (p-CA/FA = 2.0–2.8) [30], and the alkaline pre-treatment of
sugarcane bagasse pith for the production of single cell proteins
(p-CA/FA = 2.15) [24]. On the contrary, Lawther et al. [19] obtained
p-CA/FA = 0.38 with wheat straw.
Other phenolic compounds, typically encountered in alka-
line hydrolysates, such as protocatechuic acid, 4-hydroxybenzoic
acid, 4-hydroxybenzaldehyde, syringic acid, catequine, caffeic acid,
vanillin, vanillic acid and vanillic alcohol [14,24,31,32,33], were not
released in significant concentrations under the conditions inves-
tigated in this work. Nevertheless, the total phenolic compounds,
including FA and p-CA, were quantified as caffeic acid equivalents
(Fig. 3). The time behaviour of TPC was qualitatively similar to
those of FA and p-CA, in that Ca > 0.5N lasted about 2–4 h to reach a
maximum concentration of 7.8–8.4 gCAE /L, whereas no more than
45% of this level was detected at Ca = 0.2N. Fig. 4 illustrates the
504 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506
Fig. 3. Effect of NaOH concentration (Ca ) on the release of total phenolic compounds
(TPC) from corn cobs by alkaline hydrolysis: Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (),
Ca = 2.0N () employing a solid–liquid ratio of 0.084 g/g.
behaviour versus time of the fraction of unidentified phenolic com-
pounds (YUPC ), defined as TPC but FA and p-CA expressed as caffeic
acid equivalents to TPC. Under intermediate hydrolysis conditions
(Ca = 0.5N), this fraction was about 0.11 during the first 6 h, while
almost linearly increased up to about 0.45. This behaviour suggests
that the treatment was quite selective for the quick release of an
easily hydrolyzable phenolic fraction, made up mainly of FA and p-
CA, and subsequently allowed solubilising other more recalcitrant
phenolics. At higher Ca values (1.0–2.0N) the hydrolysis conditions
became more drastic, and most of the solubilisable TPC were quickly
released. As a consequence, YUPC remained almost constant around
0.45 since the beginning of the hydrolysis. The apparently strange
behaviour observed at Ca = 0.2N could have been the result of two
simultaneous effects both related to very mild hydrolysis condi-
tions, i.e. a reduction in the total solubilisable fraction of phenolics
(from 3.8 to 9.2 gCAF /L) combined with the release of a little uniden-
tified phenolic concentration (0.5 gCAF /L) even more solubilisable
than FA and p-CA.
Fig. 5 shows the time behaviour of the concentration of solu-
bilised non-volatile compounds (SNVC) released by the alkaline
treatment, which included a significant portion of other unde-
sired substances besides the phenolics. It is noteworthy that, at
the highest Ca values (1.0 and 2.0N), about 40% of the starting raw
material was released as SNVC after 6 h, whereas no more than one
half of this level was achieved after 24 h under milder conditions
Fig. 4. Fraction of unidentified phenolic compounds (UPC) with respect to total
phenolic compounds (TPC): Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (), Ca = 2.0N ()
employing a solid–liquid ratio of 0.084 g/g.
Fig. 5. Effect of NaOH concentration (Ca ) on the release of solubilised non-volatile
compounds (SNVC) from corn cobs by alkaline hydrolysis: Ca = 0.2N (䊉), Ca = 0.5N
(♦), Ca = 1.0N (), Ca = 2.0N () employing a solid–liquid ratio of 0.084 g/g.
(0.2 ≤ Ca ≤ 0.5N). These results taken together suggest that Ca = 0.5N
should be used to obtain ferulic acid-rich solutions from corn cobs
for subsequent vanillin bio-production.
The high content of unidentified compounds present in the
hydrolysates suggested us to determine the concentrations of
monosaccharides, oligosaccharides and acetic acid released by par-
tial alkaline hydrolysis of hemicellulose and cellulose fractions of
corn cobs.
Monosaccharides were undetectable in the hydrolysates
obtained under all the conditions tested for alkaline hydrolysis,
while oligosaccharides were released under different hydrolysis
conditions (see Fig. 6a). Fig. 6b–d shows the composition of these
oligomers, determined according to the procedure described in
Section 2 and expressed as xylooligosaccharides, arabinooligosac-
charides, and glucooligosaccharides, respectively. It should be
noticed that this fraction on the whole accounted for a signifi-
cant portion of SNVC and achieved a maximum concentration of
24 g/L, when the hydrolysis was performed with NaOH 2.0N for
24 h. Under all the conditions tested, the release of oligosaccha-
rides was very fast, becoming almost complete after only 5–6 h and
increasing with NaOH concentration. This behaviour suggests that
this treatment was sufficiently severe to only partially hydrolyze
the ␤-glycosidic bonds present in polysaccharides. In particular, the
release of oligosaccharides was very low (2–6 g/L) at Ca ≤ 0.5N, cor-
responding at the end of hydrolysis to only 3–8% of the starting
material, while, at Ca ≥ 1.0N, it increased up to 16–24 g/L, corre-
sponding to no less than 20–29%, respectively.
Oligosaccharides were mainly constituted by xylooligosaccha-
rides, the remaining fraction consisting mostly of arabinooligosac-
charides and only of a little portion of glucooligosaccharides. These
results are in good agreement with the composition of corn cobs
(Table 1), which is made of 34.7% hemicellulose (about 90% xylan
and 10% arabinan) [17]. Since xylooligosaccharides represented the
major fraction of these oligomers, they showed qualitatively the
same time behaviour (Fig. 6b) as total oligosaccharides. In partic-
ular a Ca increase from 0.2 to 2.0N led to a progressive increase in
the fraction of xylooligosaccharides with respect to total oligomers
released after 24 h treatment from 49% to 84%, which corresponded
to 4.8% and 78% of xylan originally present in the raw material,
respectively. Arabinan showed higher susceptibility to solubilisa-
tion than xylan in alkaline medium. The arabinooligosaccharide
content of the hydrolysate in fact increased from 0.6 to 2.5 g/L
when NaOH concentration increased from 0.2 to 2.0N (Fig. 6c),
which represented even 27% and 100% of the arabinan origi-
nally present in the raw material, respectively. At Ca = 0.2–0.5N,
 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506 505

Fig. 6. Effect of NaOH concentration (Ca ) on the alkaline hydrolysis of (a) total oligosaccharides, (b) xylooligosaccharides, (c) arabinooligosaccharides, (d) glucooligosaccharides
contained in corn cobs: Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (), Ca = 2.0N () employing a solid–liquid ratio of 0.084 g/g.

17–19% of total oligosaccharides released consisted of arabi- Table 4

Composition of the liquors from alkaline hydrolysis of corn cobs with 0.5N NaOH,
nooligosaccharides, while at Ca = 1.0–2.0N this fraction decreased
S/L = 0.084 g/g after 6 h
to 11–12% because of the increased fraction of hydrolyzed xylan.
Notwithstanding the relatively high cellulose content of corn cobs Total polyphenols (mgCAE /L) 5004 ± 218
p-Coumaric acid (mg/L) 2156 ± 63
(Table 1), released glucooligosaccharides were always present in
Ferulic acid (mg/L) 1171 ± 34
the hydrolysates at very low concentrations (0.6–1.0 g/L) (Fig. 6d). Oligomers (g/L) 4.58 ± 0.136
These glucooligosaccharides corresponded to only little percent- Glucooligosaccharides (g/L) 0.588 ± 0.032
ages either of the starting raw material (0.7–1.2%) or of its cellulosic Xylooligosaccharides (g/L) 3.31 ± 0.097
fraction (2.2% and 3.8%). These results confirm the well-known Arabinooligosaccharides (g/L) 0.703 ± 0.019
Acetic acid (g/L) 3.48 ± 0.146
recalcitrance of cellulosic materials to alkaline hydrolysis unless
Glycerol (g/L) 0.78 ± 0.052
they had preliminarily been subjected to specific pre-treatment. SNVCa (g/L) 12.6 ± 0.42
The fast solubilisation of these few glucooligosaccharides within a
SNVC: solubilised non-volatile compounds.
1–4 h could be the result of the release of hemicellulosic glucose or
hydrolysis of residual starchy material present in the corn cobs
germ [34]. as furfural and hydroxymethylfurfural [35] were detected under the
As Fig. 7 shows, acetic acid, which accounted for about 3–4% of alkaline hydrolysis conditions investigated in this study.
the starting raw material, was almost completely released in free
form, whereas no acetyl groups linked to oligosaccharides were 3.4. Hydrolysate selection for vanillin bio-production
identified. On the contrary, no sugar degradation by-products such
In conclusion, under the mildest alkaline hydrolysis condi-
tions investigated in this work (0.2 ≤ Ca ≤ 0.5N), it was possible,
at the same time, to obtain solutions with high ferulic acid
content and to retain most of the hemicellulose and cellulose
fractions in the solid residue. The hydrolysis conditions ensuring
the best compromise in terms of high ferulic acid concentra-
tion (1171 mg/L) and low percentage (9%) of ferulic acid with
respect to overall SNVC (Table 4) were S/L = 0.084 g/g, Ca = 0.5N
and 6 h of hydrolysis time. Whereas such a hydrolysate could
be an excellent candidate to be used for vanillin bio-production,
the resulting solid residue could be employed in specific
treatments addressed to the exploitation of its polysaccharide
content.

Acknowledgements
Fig. 7. Effect of NaOH concentration (Ca ) on the release of acetic acid from corn
cobs by alkaline hydrolysis: Ca = 0.2N (䊉), Ca = 0.5N (♦), Ca = 1.0N (), Ca = 2.0N () The authors are grateful to the Italian Ministry of Education, Uni-
employing a solid–liquid ratio of 0.084 g/g. versity and Research (MIUR), for its financial support (PRIN 2004
506 P. Torre et al. / Biochemical Engineering Journal 40 (2008) 500–506
prot. 2004090003) as well as to the Spanish Ministry of Science and
Technology for the financial support of this work (project CTQ2006-
02241/PPQ, which has partial financial support from the FEDER
founds of the European Union) and for the research grant awarded
to Ms. Beatriz Rivas.
References
[1] B. Rivas, J.M. Domı́nguez, H. Domı́nguez, J.C. Parajó, Bioconversion of posthy-
drolysed autohydrolysis liquors: an alternative for xylitol production from corn
cobs, Enzyme Microb. Technol. 31 (2002) 431–438.
[2] R. Robbins, Phenolic acids in foods: an overview of analytical methodology, J.
Agric. Food Chem. 51 (2003) 2866–2887.
[3] G.X. Pan, J.L. Bolton, G.J. Leary, Determination of ferulic and p-coumaric acids
in wheat straw and the amounts released by mild acid and alkaline peroxide
treatment, J. Agric. Food Chem. 46 (1998) 5283–5288.
[4] M. Bunzel, J. Ralph, J. Marita, R.D. Hatfield, H. Steinhart, Diferulates as structural
components in soluble and insoluble cereal dietary fibre, J. Sci. Food Agric. 81
(2001) 653–660.
[5] M.J. Gasson, Y. Kitamura, W.R. Mclauchlan, A. Narbad, A.J. Parr, E.L.H. Parsons,
J. Payne, M.J.C. Rhodes, N.J. Walton, Metabolism of ferulic acid to vanillin. A
bacterial gene of the enoyl-SCoA hydratase/isomerase superfamily encodes
an enzyme for the hydration and cleavage of a hydroxycinnamic acid SCoA
thioester, J. Biol. Chem. 273 (1998) 4163–4170.
[6] T. Jeffries, Biodegradation of lignin and hemicelluloses, in: C. Ratledge (Ed.), Bio-
chemistry of Microbial Degradation, Kluwer Academic Publishers, Dordrecht,
The Netherlands, 1994, pp. 233–277.
[7] B.G. Smith, P.J. Harris, Ferulic acid is esterified to glucuronoarabinoxylans in
pineapple cell walls, Phytochemistry 56 (2001) 513–519.
[8] L. Saulnier, J. Thibault, Ferulic acid and diferulic acids as components of sugar-
beet pectins and maize bran heteroxylans, J. Sci. Food Agric. 79 (1999) 396–
402.
[9] E. Allerdings, J. Ralph, H. Steinhart, M. Bunzel, Isolation and structural iden-
tification of complex feruloylated heteroxylan side-chains from maize bran,
Phytochemistry 67 (2006) 1276–1286.
[10] S. Levigne, M. Ralet, B. Quemener, J. Thibault, Isolation of diferulic bridges
ester-linked to arabinan in sugar beet cell walls, Carbohydr. Res. 339 (2004)
2315–2319.
[11] T. Higuchi, Biochemistry, Molecular Biology of Wood, first ed., Springer–Verlag,
New York, 1996.
[12] C.B. Faulds, G. Williamson, The role of hydroxycinnamates in the plant cell wall,
J. Sci. Food Agric. 79 (1999) 393–395.
[13] T. Kondo, T. Ohshita, T. Kyuma, Comparison of characteristics of soluble lignins
from untreated and ammonia-treated wheat straw, Anim. Feed Sci. Technol. 39
(1992) 253–263.
[14] F. Xu, R. Sun, J. Sun, C. Liu, B. He, J. Fan, Determination of cell wall ferulic and
p-coumaric acids in sugarcane bagasse, Anal. Chim. Acta 552 (2005) 207–217.
[15] E. Graf, Antioxidant potential of ferulic acid, Free Radical Biol. Med. 13 (1992)
435–448.
[16] D. De Faveri, P. Torre, B. Aliakbarian, J.M. Domı́nguez, P. Perego, A. Con-
verti, Response surface modeling of vanillin production by Escherichia coli
JM109pBB1, Biochem. Eng. J. 36 (2007) 268–275.
[17] G. Garrote, H. Domı́nguez, J.C. Parajó, Kinetic modeling of corncob autohydrol-
ysis, Process Biochem. 36 (2001) 571–578.
[18] T. Swain, W.E. Hillis, The phenolic constituents of Prunus domestica. The quan-
titative analysis of phenolic constituents, J. Sci. Food Agric. 10 (1959) 63–68.
[19] J.M. Lawther, R. Sun, W.B. Banks, Fractional characterization of alkali-labile
lignin and alkali-insoluble lignin from wheat straw, Ind. Crops Prod. 5 (1996)
291–300.
[20] Y. Li, R. Ruan, P.L. Chen, Z. Liu, X. Pan, X. Lin, X. Liu, C.K. Mok, T. Yang, Enzymatic
hydrolysis of corn stover pretreated by combined dilute alkaline treatment and
homogenization, Trans. ASABE 7 (2004) 821–825.
[21] R.A. Silverstein, Y. Chen, R.R. Sharma-Shivappa, M.D. Boyette, J. Osborne, A com-
parison of chemical pretreatment methods for improving saccharification of
cotton stalks, Bioresour. Technol. 98 (2007) 3000–3011.
[22] C. Antoine, S. Peyron, F. Mabille, C. Lapierre, B. Bouchet, J. Abecassis, X. Rouau,
Individual contribution of grain outer layers and their cell wall structure to
the mechanical properties of wheat bran, J. Agric. Food Chem. 51 (2003)
2026–2033.
[23] C. Barron, A. Surget, X. Rouau, Relative amounts of tissues in mature wheat
(Triticum aestivum L.) grain and their carbohydrate and phenolic acid compo-
sition, J. Cereal Sci. 45 (2007) 88–96.
[24] R. Rodrı́guez-Vázquez, D. Dı́az-Cervantes, Effect of chemical solutions sprayed
on sugarcane bagasse pith to produce single cell protein: physical and chemical
analyses of pith, Bioresour. Technol. 47 (1994) 159–164.
[25] E. Billa, D.P. Koullas, B. Monties, E.G. Koukios, Structure and composition of
sweet sorghum stalk components, Ind. Crops Prod. 6 (1997) 297–302.
[26] C. Lapierre, Application of new methods for the investigation of lignin struc-
ture, in: H.G. Jung, D.R. Buxton, R.D. Hatfield, J. Ralph (Eds.), Forage Cell Wall
Structure and Digestibility, American Society of Agronomy, Madison, WI, 1993,
pp. 133–166.
[27] S. Mussatto, G. Dragone, I. Roberto, Ferulic and p-coumaric acids extraction
by alkaline hydrolysis of brewer’s spent grain, Ind. Crops Prod. 25 (2007)
231–237.
[28] S. Ou, Y. Luo, F. Xue, C. Huang, N. Zhang, Z. Liu, Separation and purification of fer-
ulic acid in alkaline-hydrolysate from sugarcane bagasse by activated charcoal
adsorption/anion macroporous resin exchange chromatography, J. Food Eng. 78
(2007) 1298–1304.
[29] J.O. Fritz, K.J. Moore, Separation and quantification of lignin-derived phenolic
monomers using high-resolution gas chromatography, J. Agric. Food Chem. 35
(1987) 710–713.
[30] M.B. Scalbert, Comparison of wheat straw lignin preparations. II. Straw lignin
solubilisation in alkali, Holzforschung 40 (1986) 249–254.
[31] G.C. Galletti, C.I. Muscarella, P. Bocchini, F. Cinti, Changes in the phenolic
composition of maize stovers after blooming, J. Agric. Food Chem. 44 (1996)
2844–3840.
[32] M. Chakraborty, K. Das, G. Dey, A. Mitra, Unusually high quantity of 4-
hydroxybenzoic acid accumulation in cell wall of palm mesocarps, Biochem.
Syst. Ecol. 34 (2006) 509–513.
[33] D. Del Pozo-Insfran, C.H. Brenes, S.O. Serna Saldivar, S.T. Talcott, Polyphenolic
and antioxidant content of white and blue corn (Zea mays L.) products, Food
Res. Int. 39 (2006) 696–703.
[34] E. Chanliaud, L. Saulnier, J. Thibault, Alkaline extraction and characterization of
heteroxylans from maize bran, J. Cereal Sci. 21 (1995) 195–203.
[35] E. Sjostrom, Carbohydrate degradation products from alkaline treatment of
biomass, Biomass Bioenerg. 1 (1991) 61–64.