Bioresource Technology 102 (2011) 10013–10019

Contents lists available at SciVerse ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Cellulose extraction from orange peel using sulfite digestion reagents

Ioan Bicu ⇑, Fanica Mustata
Petru Poni Institute of Macromolecular Chemistry, Aleea Gr. Ghica Voda Nr. 41A, Iasi 700487, Romania

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

Article history: Orange peel (OP) was used as raw material for cellulose extraction. Two different pulping reagents were
Received 8 June 2011 used, sodium sulfite and sodium metabisulfite. The effect of the main process parameters, sulfite agent
Received in revised form 8 August 2011 dosage and reaction duration, on cellulose yield was investigated. A central composite rotatable design
Accepted 10 August 2011
involving two variables at five levels and response surface methodology were used for the optimization
Available online 17 August 2011
of cellulose recovery. Other two invariable parameters were reaction temperature and hydromodulus.
The optimum yields, referred to the weight of double extracted OP, were 40.4% and 45.2% for sodium sul-
Keywords:
fite and sodium metabisulfite digestions, respectively. The crude celluloses were bleached with hypo-
Orange peel
Cellulose
chlorite and oxygen. The physicochemical characterization data of these cellulose materials indicate
Hemicellulose good levels of purity, low crystallinities, good whitenesses, good water retention and moderate molecular
Sulfite pulping weights. According to these specific properties the recovered celluloses could be used as fillers, water
Response surface methodology absorbents, or as raw materials for cellulose derivatives.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Orange peel (OP) produced as a primary byproduct by the fruit-
processing industry, if disposed on the land, can cause environ-
mental pollution (Ververis et al., 2007; Mamma et al., 2008; Wang
et al., 2008; Siles López et al., 2010). For this reason, the correct
management of this byproduct constituted a permanent problem
for the juice, canning, or chemical industries (Ma et al., 1993;
Laufenberg et al., 2003; Siles López et al., 2010). To prevent such
serious pollution problems and still with economic benefits,
numerous studies regarding the establishment of the chemical
composition of OP (Eaks and Sinclair, 1980; Ma et al., 1993; Marin
et al., 2007; Wang et al., 2008), and the conversion of the identified
components into useful value-added products were performed. Ini-
tially, attempts were made to use dried OP as natural feed additive
(Bampidis and Robinson, 2006), and even as medicinal supplement
(Callaway et al., 2008) for animals. But, owing to both their low
nutritional intake, or owing to the diseases caused to the consum-
ing animals (mycotoxicosis, rumen parakeratosis), this end use
proved to be a low profitable one, with a limited extension
(Bampidis and Robinson, 2006; Ververis et al., 2007). Other several
research studies pursued extraction of some natural byproducts as
sugars (Eaks and Sinclair, 1980; Grohmann et al., 1995), pectin (Ma
et al., 1993), carotenoids and flavonoids (Manthey and Grohmann,
1996; Marin et al., 2007; Wang et al., 2008), oils and perfumes
(Braddock et al., 1986), very useful in food industry or in keeping
⇑ Corresponding author. Tel.: +40 232 217454; fax: +40 232 211299.
E-mail address: ibicu@icmpp.ro (I. Bicu).
the human health. Much more research studies were devoted to
the conversion of OP to other valuable products, such as ethanol
(Grohmann et al., 1994; Wilkins et al., 2007), citric acid
(Rivas et al., 2008), succinic acid (Li et al., 2010), methane (Martin
et al., 2010), enzymes (Mamma et al., 2008; Rangarajan et al.,
2010), by hydrolytic and fermentative processes.
Much less numerous studies concerning the recovery of the
polysaccharides from OP as polymer materials, but not as raw
material for the hydrolysis and fermentation processes, were
reported. This, the more so as some of the above mentioned pro-
cesses can produce such polymer materials as fibrous residues
(Aravantinos-Zafiris et al., 1994; Liu et al., 2006; Marin et al.,
2007). Furthermore, the up to date review of the literature con-
cerning the extraction of cellulose from citrus peels shows that it
was entirely carried out according to the alkaline process
(Xuan et al., 2006; Yasar et al., 2007; Li et al., 2008; Ejikeme,
2008). The resulting fibrous materials were used as additives in
the papermaking processes (Ververis et al., 2007), as adsorbent
substrates for retention of metal ions (Xuan et al., 2006; Li et al.,
2008), or as raw material for the preparation of the cellulose deriv-
atives (Yasar et al., 2007). But, no studies have been reported on
the extraction of cellulose from OP by sulfite process. The sulfite
pulping process is known as being that in which the raw material
is cooked in a solution containing sodium salts of sulfurous acid, as
sulfites or bisulfites. The sulfite process is an acid one and under
these conditions, dissolves lignin, hemicelluloses and other cellu-
lose accompanying substances. The resulting cellulose has special
properties such as high alpha-cellulose content, high brightness
and whiteness, uniform molecular weight distribution (Biermann,
1996), and is suitable for the production of regenerated cellulose

0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2011.08.041
10014 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019
and cellulose derivatives, or of other cellulose-based end products
(Elg-Christoffersson, 2005).
That is why the purpose of the present paper is to prepare
chemical grade cellulose with high chemical accessibility and reac-
tivity, starting from OP and using sulfite cooking reagents.
2. Methods
2.1. Plant materials and chemicals
Navel oranges, purchased from the open market, were peeled
(flavedo + albedo). The peels were segmented in the humid state
to a particle size of about 2  2 cm, air-dried to an approximate
10% moisture content, coarsely milled in a laboratory milling ma-
chine, and sieved. Only the peel fraction having a particle size dis-
tribution range of 1.0 up to 3.0 mm was used in the present
experiments. The dried and ground OP was defatted by extraction
with the trichloromethane–hexane azeotrope (72/28, w/w, b.p.
69 °C) in a Soxhlet apparatus for 10 h. This azeotrope was chosen
for the extraction of fats, resins, oils, waxes and coloring maters,
so-called liposoluble fraction, because some preliminary experi-
ments showed that such a mixture is able to ensure high extraction
efficiency. The defatted OP was dried in open air at room temper-
ature (25 °C) and again extracted, this time using the azeotrope
water/ethanol/toluene (21/55/24, w/w/w, b.p. 74 °C), for 32 h. This
ternary azeotrope proved to be very efficient for the extraction of
both the water soluble pectin fraction and flavonoids (especially
hesperidin, the most abundant flavonoid in Navel orange peel) by
the aqueous ethanol component, and of the other compounds, such
as unextracted waxes or dyes, by the aromatic hydrocarbon com-
ponent. The slightly prolonged duration of extraction (up to 32 h)
was required, on the one hand, by the relatively low solubility of
hesperidin even in the hot alcohols and on the other hand, by
the need to entirely extract hesperidin from the processed OP. If
not, the presence of hesperidin in the final bleached cellulose, even
as traces, could cause its yellowing as time goes. It is well known
that hesperidin easily dissolves into the alkaline media. But, be-
cause the sulfite process is an acid one, appears the risk to transfer
hesperidin to the end product if it was not completely removed
from OP by the extraction with the organic solvents.
The chemicals used in the present research were of reagent
grade. Sodium sulfite, sodium metabisulfite, hexane, sodium hypo-
chlorite were bought from Sigma–Aldrich (Steinheim, Germany).
Methanol, ethanol, hydrogen peroxide, trichloromethane and tolu-
ene, were purchased from Chemical Company SA (Iasi, Romania).
2.2. Analytical methods
2.2.1. Chemical analysis of OP and cellulose materials
Orange peel was analyzed for cellulose, hemicellulose, lignin,
ash, crude fats, liposoluble fraction, alcohol soluble fraction, flavo-
noids and protein. The analyses were carried out in duplicate and
the average was reported. All results were calculated on an oven-
dry matter basis.
The elemental analysis (C, H and N) was performed with a Per-
kin-Elmer CHNS-2420 analyzer. Protein content was calculated by
multiplying the nitrogen content by the factor 6.25 (Rivas et al.,
2008). The crude fat content was determined gravimetrically by
extraction with the trichloromethane/hexane azeotrope (72/28,
w/w, b.p. 69 °C) for 10 h. The weight of the dry residue remained
after the recovery of the solvent was referred to the weight of OP
taken for analysis to give the fat content. The amount of soluble
pectin was determined by the extraction of defatted OP with the
water/ethanol/toluene triple azeotrope (21/55/2, w/w/w, b.p.
74 °C) for 32 h, and was expressed as percent of the weight of
OP. Some chemical components were determined on the same
defatted OP by using the standard analytical procedures: ASTM
D1102-84 for ash content, ASTM D1106-84 for Klason lignin con-
tent, ASTM D1104-56 for holocellulose content, ASTM D1103-60
for cellulose content and ASTM D1110-84 for water extractives.
Average degree of polymerization (DPn) of cellulose was deter-
mined from the specific viscosity data using Eq. (1) (Klemm
et al., 1998).
2000gspec
DPn ¼ ð1Þ
cð1 þ 0:29gspec Þ
where gspec = (t1  t0)/t0, t1 is the efflux time of cellulose solution, t0
the efflux time of solvent and c the concentration of cellulose solu-
tion (g/L). Viscosity measurements were performed on the cupram-
monium cellulose solutions at 25 °C with an Ubbelohde viscometer.
2.2.2. Physical analysis
Water retention value (WRV) of cellulose was measured by
immersing the cellulose material in distilled water (0.5% consis-
tency) at 20 °C for 12 h, filtering at a centrifugal force of 1000g
for 10 min, weighing the obtained wet mass, drying it at 103 °C
for 6 h, weighing again the dried mass, and using Eq. (2):
WRVð%Þ ¼ 100ðW 1  W 0 Þ=W 0 ð2Þ
where W0 and W1 were the weights of dry and wet cellulose,
respectively.
Fourier transform infrared spectra (FTIR) were recorded on a
Bruker Vertex 70 spectrophotometer, with KBr pellets. The CIE
whiteness of the bleached cellulose was estimated using CIE
L⁄a⁄b⁄ system. The L⁄, a⁄ and b⁄ color coordinates were determined
with a Pocket Spec Color QA spectrophotometer provided with a
sensor head of 6 mm. Whiteness was determined on the powdered
cellulose placed in a ring and pressed with a plunger by hand. Mea-
surements were made using the CIE D65 illuminant and a 10° stan-
dard observer. The color coordinates were used to calculate CIE
whiteness according to Eq. (3) (Ganz and Pauli, 1995).

W CIE ¼ 2:41  L  4:45  b ½1  0:009ðL  96Þ  141:4 ð3Þ
Loose bulk densities of cellulose powders were determined
according to the standard ASTM D7481-09. Cellulose crystallinity
was studied by means of X-ray diffractometry technique. Wide an-
gle X-ray diffraction (WAXD) was performed with a Bruker AD8
Advance diffractometer at 40 kV and 40 mA. WAXD patterns were
obtained using Cu Ka radiation k = 1.541 Å and a diffraction angle
2h ranging from 4° to 40°. The crystallinity index (CrI) was calcu-
lated using Eq. (4) (Cao and Tan, 2005).
CrIð%Þ ¼ ðI002  Iam Þ=I002  100 ð4Þ
in which I002 is the maximum intensity of the crystalline plane
(0 0 2) reflection (2h = 22.5°) and Iam the maximum intensity of
amorphous phase (2h = 18°).
2.3. Experimental procedure
2.3.1. Sulfite pulping
The sulfite pulping experiments were carried out in a 0.25 L lab-
oratory rotary glass vessel equipped with an air cooled condenser
and a thermometer. The reaction vessel was immersed in a hot oil
bath held at 105 °C. All the above and the rotating system were
tapped from a rotary evaporator. The inclination of mechanical de-
vice was set at 45° and the rotational speed at 40 rpm. OP was
charged into the digester together with the cooking liquor includ-
ing distilled water and sulfite reagent, and was allowed to stand for
30 min to impregnate. Always 20.0 g of oven dry (o.d.) OP and
150 g of cooking solution (liquor to solid OP ratio 7.5:1) were
 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019 10015

added. In all experiments a preheating time of 5 min was counted coded and actual levels of the variables used in the design matrix
sufficient to reach the working temperature (98 °C). In order to as well as the data regarding the experimental yield of cellulose
optimize the cooking conditions, the effect of sulfite reagent dos- pulp for both sodium sulfite and sodium metabisulfite processes.
age and reaction duration, as variable parameters, was studied, A second order polynomial equation was developed to study the
Table 1. After cooking the fibrous product was separated from effects of independent variables on cellulose yield. Its general form
residual green-yellow liquor by filtration on a fiberglass fabric, dis- is as follows:
integrated and washed repeatedly (three times) with fresh tap
water. The wet pulp was dehydrated to 10–12% consistency and fi- Y ¼ b0 þ b1 X 1 þ b2 X 2 þ b11 X 21 þ b22 X 22 þ b12 X 1 X 2 ð5Þ
nally, bleached.
where Y is the dependent variable (percent cellulose yield), b0 the
constant term, b1 and b2 the linear coefficients, b11 and b22 the qua-
2.3.2. Bleaching of pulp dratic coefficients, b12 the cross-product coefficient and X1 and X2
In this study, sulfite pulp was bleached in a three bleaching se- the coded levels of independent variables. Statistica’ 99 Edition soft-
quence, namely: hypochlorite in alkaline medium (H1), hypochlo- ware was used for regression and graphical analysis. The optimal
rite in slightly acid medium (pH = 6.5) (H2) and hydrogen values (stationary points) of selected variables were obtained by
peroxide (HP). The addition of chemicals (% on dry weight pulp) solving the polynomial equations or by analyzing response surfaces
was as follows: H1, 2.30 as active chlorine, H2, 1.2 as active chlorine and contour plots.
and HP, 0.5 as H2O2. It seems that, in the last stage bleaching took
place in presence of active oxygen formed from the reaction be-
tween hydrogen peroxide and hypochlorite. Bleaching process 3. Results and discussion
was conducted at 25 °C and at a pulp consistency of 3.5%. The treat-
ment durations were as follows: 240 min for H1 stage, 120 min for 3.1. Chemical characterization of OP
H2 stage and 60 min for HP stage. Before and after each treatment
stage, pulps were washed with distilled water until a neutral pH Initially, defatted OP was subjected to the elemental analysis.
was reached. After bleaching, cellulose was washed three times According to the results of this analysis the composition of OP
with ethyl alcohol and dried under ambient conditions. was as follows (wt.%): C, 45.82; H, 6.55; N, 1.06. It can observe that
these values are very close to those reported in the specialty liter-
ature (Li et al., 2008). All these characterization data shall be
2.4. Experimental design
deemed to be representative in defining of the chemical character-
ization of OP, but the most interesting of these is the nitrogen con-
Generally, a cooking process is defined by four main parame-
tent, because it is useful for the calculation of protein content
ters, namely chemical dosage, reaction temperature, reaction dura-
(Table 3). The elemental composition of OP extracted additionally
tion and hydromodulus (liquor to solid OP ratio, v/w). In this
with the azeotrope water–ethyl alcohol–toluene has changed as
research, some process parameters as reaction temperature and
follows (wt.%): C, 43.83; H, 6.22; N, 1.72. One can observe that,
hydromodulus were kept constant (98 °C and 7.5:1, respectively),
as a consequence of extraction of soluble sugars and pectin, carbon
the first for reasons of work at atmospheric pressure and the sec-
and hydrogen contents slightly decreased, whereas nitrogen con-
ond for reasons of the minimum value able to ensure a good
tent, and implicitly protein content, slightly increased. The next
homogenization by stirring of the reaction mass. This is why, in
step in the characterization of OP consisted of the determination
this study a 22 factorial design and response surface methodology
of chemical composition in accordance with the methods de-
(RSM) were employed to investigate the influence of the other two
scribed in the Section 2.2.1. This composition is presented in Table
parameters (chemical dosage and reaction duration) on the extrac-
3. The data in Table 3 prove that OP contains high amounts of sol-
tion of cellulose from OP with sulfite reagents. Therefore, the re-
uble sugars, soluble pectin, cellulose and hemicellulose. Also, one
agent dosage (Xd) and reaction duration (Xt) were considered as
can remark the particularly low content of lignin. These properties
independent variables. The percent yield of cellulose pulp was cho-
make OP a valuable raw material suitable for the preparation of
sen as dependent variable. Under these conditions a factorial de-
high purity cellulose.
sign with 13 experimental runs (4 factorial points, 4 star points
The chemical composition presented in Table 3 also reveals that
and 5 central points) of two variables and five levels (1.414,
cellulose, hemicelluloses, lignin and ash contents for the studied
1, 0,+1,+1.414) was fitted to perform the second order polynomial
OP are close to those reported by Aravantinos-Zafiris et al. (1994)
model (Cochran and Cox, 1992). The actual range for each indepen-
and Ververis et al. (2007) for Greek OP. Only the Spanish OP recov-
dent variable was established in accordance with the results of
ered as industrial byproducts had somewhat higher cellulose,
some preliminary experiments. In Tables 1 and 2 are given the
hemicellulose, lignin, ash and protein contents and a lower pectin
content, probably depending on the industrial source (Marin et al.,
Table 1 2007).
The level and the range of experimental variables in the central composite design. For a better extraction of cellulose, and not only, OP was addi-
Variable Level tionally processed with the azeotrope water–ethyl alcohol–toluene
to extract as much as possible pectin and other accompanying sub-
1.414 1 0 +1 +1.414 DX
stances, for example hesperidin. Otherwise, the extracted pectin
Sodium sulfite as digestion reagent
and flavonoids constitute, apart from cellulose, other valuable
Xd (%) 15.86a 20.0 30.0 40.0 44.14 10.0
32.23b 40.64 60.96 81.28 89.69 20.32 byproducts (Manthey and Grohmann, 1996; Wang et al., 2008).
Xt (min) 180 224 330 436 480 106 Moreover, extraction of hesperidin by this route led to a high yield
Sodium metabisulfite as digestion reagent (5.6–5.9%) and a high purity product. On the other hand, if not ex-
Xd (%) 15.0a 21.59 37.50 53.41 60.0 15.91 tracted, hesperidin can cause the yellowing of bleached cellulose as
45.96c 66.16 114.92 163.68 183.88 48.76 time goes. Therefore, only this double extracted OP was used in the
Xt (min) 180 224 330 436 480 106 pulping experiments.
a
Expressed as Na2O. A third remark refers to the high content of ash (2.87%) of defat-
b
Expressed as Na2SO3. ted OP. The ash content exceeded this value after extraction of sug-
c
Expressed as Na2S2O5. ars and pectin from OP (4.66%) and as will be seen later, it will
10016 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019

Table 2
Experimental program of sulfite cooking tests and experimental responses.

Run Coded variables Uncoded variables Response

X1 X2 Reagent dosage Reaction duration (min) Yield of cellulose (%)
SSPa SMPb SSP SMP SSP SMP
RDc RDd RDc RDd
1 1 1 20.0 40.64 17.6 53.93 224 224 56.2 60.1
2 +1 1 40.0 81.28 53.4 163.6 224 224 43.3 53.6
3 1 +1 20.0 40.64 17.6 53.93 436 436 46.3 53.9
4 +1 +1 40.0 81.28 53.4 163.60 436 436 40.6 48.4
5 1.414 0 15.86 32.23 15.0 45.96 330 330 55.2 51.9
6 +1.414 0 44.14 89.69 60.0 183.84 330 330 41.5 47.2
7 0 1.414 30.0 60.96 37.5 114.90 180 180 52.1 59.5
8 0 +1.414 30.0 60.96 37.5 114.90 480 480 39.4 43.7
9 0 0 30.0 60.96 37.5 114.90 330 330 46.1 49.7
10 0 0 30.0 60.96 37.5 114.90 330 330 46.5 48.3
11 0 0 30.0 60.96 37.5 114.90 330 330 45.4 48.7
12 0 0 30.0 60.96 37.5 114.90 330 330 45.1 50.6
13 0 0 30.0 60.96 37.5 114.90 330 330 44.7 49.9
a
SSP, sodium sulfite process.
b
SMP, sodium metabisulfite process.
c
RD, reagent dosage expressed as Na2O, %.
d
RD, reagent dosage expressed as pure substance, %.

Table 3 duration, hydromodulus and temperature. The digestion solution

Composition of dry orange peel.
consisted of sulfite reagent and water. Sulfite cooking was carried
Compounds Value Method out at pH values ranging from 3.5 to 6.7, depending on the chem-
(%) ical nature of the used reagent, its dosage and the phase of the
1 Cold water soluble content 23.6 ASTM D1110-84 digestion process. For example, pH had the lowest value at the
(simple sugars) beginning of sodium metabisulfite cooking and the highest at the
2 Hot water soluble content 52.3* ASTM D1110-84 end of sodium sulfite cooking.
(sugars, soluble pectin)
3 Soluble pectin 28.7 By difference
During sulfite digestion the greater part of extractives from OP
4 Protein 6.62 From nitrogen content as lignin, protein, pectin and the low-molecular-weight fraction of
5 Flavonoids (hesperidin) 5.86 By precipitation of hemicelluloses have dissolved. Under the conditions of sulfite
concentrated, cooled, water– cooking pectin, in fact its water insoluble fraction called protopec-
ethyl alcohol–toluene extract
tin, underwent a hydrolysis reaction and dissolved. More correctly,
6 Holocellulose 27.4 Sodium chlorite
7 Cellulose 14.4 Kurschner in this phase sulfite reagents broke the ester linkages existing be-
8 Hemicellulose 10.9 By difference tween protopectin and cellulose. The presence of such ester link-
9 Lignin 1.33 72% Sulfuric acid ages in the chemical structure of OP is evidenced in Table 4, first
10 Ash 2.87 ASTM D1102-84 column, by the specific absorption band at 1748 cm1. But, accord-
11 Liposoluble fraction (crude 5.53 Hexane–trichloromethane
fats, essential oils, resins, azeotrope
ing to the expectations, this absorption band disappeared in the
coloring matters) spectrum of cellulose derived from this OP (Table 4, second col-
12 Alcohol soluble fraction 51.6* Water–ethyl alcohol–toluene umn) just as a result of the dissolution of ester bond by hydrolysis.
(sugars, pectin, flavonoids) azeotrope The solubilized pectin was separated from waste liquors by precip-
13 Alcohol insoluble fraction 48.4 By difference
itation into much ethyl alcohol and purified by reprecipitation
(holocellulose, lignin,
insoluble pectin, protein, ash) from the same solvent. The FTIR spectroscopy data of precipitate
(Table 4, third column) proves the existence of the reaction prod-
*
Hot water and triple azeotrope proved to be approximately equal as regards their
ucts formed through the reaction between pectin and (bi)sulfites,
ability to extract soluble sugars and pectin.
named as hydroxy sulfonates, by the appearance of specific absorp-
tion bands at 1124 and 618 cm1, both indicating the presence
increase even more in the case of holocellulose (7.25%), sulfite cel-
lulose (5.70%) and metabisulfite cellulose (6.39%), as a result of the
removal of lignin, protopectin, hemicellulose and such like. In sum- Table 4
mary, one can conclude that both sodium sulfite digestion and so- Characteristic FTIR bands of functional groups of OP, cellulose recovered from OP with
dium metabisulfite digestion reduced to some extent the ash sodium metabisulfite and of the reaction product separated from the waste liquors.
content in the delignified OP, but not entirely. However, if desired, Assignments IR vibration frequency (cm1)
a hydrochloric acid solution with a concentration of about 1 mol L
OP Cellulose Reaction product
can be an appropriate treatment for removal of calcium salts con-
H-bonded OH stretching 3420 3426 3430
stituting the ash or mineral matter. For example, such a treatment
CH stretching 2928 2899 2933
carried out at room temperature for 10 h diminished the ash con- C@O stretching in carboxylics 1748 – 1748
tent to less than 0.5%. OH bending of adsorbed water 1633 1631 1617
CH, CH2 bending 1442 1434 –
C–O stretching in CH2-OH – 1031 –
3.2. Cellulose extraction SO2 stretching in SO2-OH – – 1124
C–O–C stretching – 895 –
OH out-of-plane bending 665 670 –
In the process of extracting cellulose from OP, raw material was
S–O stretching in SO2-OH – – 618
reacted with the digestion solution under controlled dosage,
 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019 10017

of –SO3H groups (Mallakpour and Azizallahi, 2007). Such hydroxy

sulfonate compounds might also form by the reaction of sulfites
with sugars, lignin and protein. Based on the above findings, it
should be noted that sulfite process does not degrade pectin to
the same extent as, e.g. soda process does it, so that the pectin
derivatives remain some useful byproducts, probably similar to
carrageenans.

3.2.1. Experimental design

The effect of reagent dosage (Xd) and reaction duration (Xt) on
cellulose yield in the sulfite and metabisulfite pulping of OP was
investigated by using the central composite rotatable design. In
each case, 13 experiments were carried out according to the condi-
tions indicated in Table 2. Response values as cellulose yields are
given in the last columns of this table. The second-order polyno-
mial response surface model in Eq. (5) was fitted to the response
variable (Y) in each case. Estimated regression coefficients of the
quadratic polynomial models for response variables, together with
the corresponding R2 and t-values, are given in Table 5. Neglecting
the insignificant terms, the data in Table 5 allowed to write the Fig. 1. Three-dimensional response surface and contour plots for the effect of
next final predictive equations (variables in coded values): sodium sulfite dosage and reaction duration on cellulose yield.

Y SS ¼ 45:56  4:75X d  3:82X t þ 1:28X 2d þ 1:80X d X t ð6Þ

Y SM ¼ 49:44  2:33X d  4:22X t þ 0:90X 2d þ 1:92X 2t ð7Þ
The goodness of fit of the models was verified by the determi-
nation coefficient (R2). As can be seen in the Table 5 the two values
of the coefficient of determination are quite high. The highest value
(0.98) corresponds to the case of sodium sulfite digestion and the
lowest (0.84) to the case of sodium metabisulfite digestion. These
values indicate that 98% (84%) of the total variations of dependent
variables can be explained by the models.
The relationship between the responses and the experimental
variables were illustrated graphically by plotting three-dimen-
sional response surface plots and contour plots, Figs. 1 and 2. The
examination of the two figures shows that, as expected, cellulose
yield strongly decreases with the increase of both the addition of
pulping agent and the reaction time. In each case a well defined
minimum can be observed. Moreover, from these figures, and espe-
cially from the analysis of Eqs. (6) and (7), one can observe that the
linear terms of independent variables have the most significant
influence on cellulose yield, while the quadratic terms and cross-
product terms exert only a moderate, or even insignificant, one.
As regards the optimal conditions, this time these were calculated Fig. 2. Three-dimensional response surface and contour plots for the effect of
sodium metabisulfite dosage and reaction duration on cellulose yield.
by equating the partial derivatives of dependent variables with re-
spect to independent variables to zero and solving the resulting
system of equations. As a result, the optimum values of test vari-
metabisulfite cooking, with the corresponding values for the
ables in coded units are as follows; Xd = 2.12 and Xt = 0.68 for so-
dependent variables (cellulose yields) YSS = 41.24% and
dium sulfite cooking and Xd = 1.29 and Xt = 1.10 for sodium
YSM = 45.62%. But, the coded optimum value for sodium sulfite dos-
age, found to be 2.12, placed outside the studied area of values. To
Table 5 solve the problem, was calculated another pair of values, placed
Estimated model coefficients and t-value of statistical significance. this time in the inside of the domain, and leading to the same cel-
Coefficient Factor Sodium sulfite process Sodium metabisulfite lulose yield. Such a pair can be: Xd = 0.18 and Xt = 1. In natural units
process the corresponding values for the test variables are as follows:
Estimated t-valuea Estimated t-valuea Xd = 31.8% Na2O and Xt = 436 min for sodium sulfite cooking, and
coefficient coefficient Xd = 58.02% Na2O and Xt = 447 min for sodium metabisulfite cook-
b0 – 45.56 62.15 49.44 53.11
ing. The verification of predicted results was accomplished by
b1 Xd 4.75 18.27 2.33 7.06 employing the optimized conditions in two additional digestion
b2 Xt 3.82 14.70 4.22 12.79 experiments. The experiments ended in cellulose yields of 40.4%
b11 X 2d 1.28 4.62 0.90 2.55 for sodium sulfite cooking and of 45.2% for sodium metabisulfite
b22 X 2t 0.02 0.07* 1.92 5.44 cooking. These results are quite equal or even lower than those
b12 XdXt 1.80 4.92 0.25 0.54* predicted, what emphasizes the effectiveness of the established
– R2 0.98 0.84
model. And finally, taking into consideration the fact that the
a
Student t test was used with the level of significance of 5%. above cellulose yields were calculated in reference to the weight
*
Value is statistically significant if t > 2.228. of double extracted OP, and that the solvent-insoluble fraction
10018 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019

was 48.4% (Table 3), it comes out that the actual cellulose yields in remark the absence from the IR spectra of these cellulose materials
the two cases are: Y 0SS = 19.6% and Y 0SM = 21.9%, both related to the of the absorption bands characteristic of lignin or other accompa-
weight of the starting defatted OP. nying substances. All these spectral data prove that the materials
extracted from OP using sulfite cooking reagents are nearly pure
celluloses.
3.3. Cellulose characterization
As shown in Tables 3 and 6, both the celluloses produced by
digestion with sulfite reagents and the starting OP present some
3.3.1. Chemical characterization
high ash contents of 5.70%, 6.39% and 2.87%, respectively. When
The chemical composition of bleached celluloses extracted from
compared, e.g., with the ash content of wood celluloses (0.1–
OP using sulfite reagents under the optimum pulping conditions is
0.5%), these values could be considered as being very high. But, tak-
tabulated in the first four rows of Table 6. As shown in table, these
ing into consideration an average pulp yield of about 20%, it follows
celluloses contain about 75–80% alpha cellulose, and significant
that the ash content of the two celluloses extracted from OP should
amounts of hemicelluloses (13–18%) and ashes (5.7–6.4%). As re-
be five times higher, namely 28.5% and 31.95%, respectively. The
gards the extraction yields, it can observe that these fall within
fact that the real ash contents are that presented in 5, denotes that
the range from 40% to 45%. These values reveal that cellulose yields
both sulfite reagents reduced to some extent the ash content of the
in both sulfite processes are much higher than those registered, for
delignified OP, but not entirely. However, if desired, hydrochloric
example, when OP was digested with sodium hydroxide (26.4%,
acid at a concentration of about 4% can be an appropriate treat-
unpublished results). On the other hand, when the two sulfite
ment for removal of calcium salts which constitute the ash or min-
digestion processes are compared with each other, one can observe
eral matter. For example, such a treatment carried out at room
that cellulose yields characteristic of sodium metabisulfite pulping
temperature for 10 h diminished the ash content to less than 0.5%.
are constantly superior to those characteristic of sodium sulfite
pulping. This means that sulfite digestion, an acid one, took place
3.3.2. Physical characterization
under some more gentle conditions than alkaline digestion and
The bleached celluloses extracted from OP by the two sulfite
that sulfite reagents, especially sodium metabisulfite, were able
processes are some white granular, or rather powder, materials.
to better protect both cellulose and especially hemicellulose
However, in this case the specific shape of the fibers has been pre-
against the depolymerization action by hydrolysis. All the above
served to some extent. The solid particles are composed of aggre-
prove that OP, despite being a waste product, can be considered
gated and non-aggregated fiber fragments and cellulose
as a promising resource for the production of chemical grade cellu-
microfibrils, and vary in size from less than 10 lm to about
lose under the conditions of sulfite digestion processes.
300 lm. All these presume the existence of some moderate de-
Table 4 shows the characteristic absorption bands of cellulose
grees of polymerization that characterize the studied celluloses. In-
produced using sodium metabisulfite as pulping reagent (second
deed, in Table 6 can be observed values for the degree of
column) and of its corresponding raw material, double extracted
polymerization ranging between 318 and 344. By other hand,
OP (first column), respectively. The examination of the data in
when the degrees of polymerization of the two sulfite celluloses
the first column of Table 4 shows that these are similar to those
are compared with each other, it can be seen that the value corre-
published in literature for this polymer substance (Li et al., 2008;
sponding to cellulose extracted with sodium sulfite as digestion re-
Muresan et al., 2009). The IR spectrum of OP in Table 4, first col-
agent is greater than that corresponding to cellulose extracted by
umn, is also similar to a great extent to the standard IR spectrum
means of sodium metabisulfite. These moderate degrees of poly-
of cellulose but, by comparing the two spectra each other, one
merization can be ascribed both to the relatively high hemicellu-
can observe the appearance of the absorption band at 1748 cm1
lose content of the produced pulps and to the relatively low
in the latter. This absorption band can be attributed to the ester
molecular weight of cellulose in the starting raw material. The
linkage formed between the OH groups in cellulose and hemicellu-
presence of hemicellulose conferred to the cellulose products the
lose and COOH groups in pectins, and its presence in the spectrum
capability to absorb large amounts of water, of up to four times
of OP can prove the existence of protopectin in the chemical struc-
the weight of the starting material, Table 6.
ture of OP. And conversely, because in the digestion process this
Whiteness, as an important factor in defining the optical prop-
ester linkage has been broken by hydrolysis, its corresponding
erties of cellulose products, constitutes a measure of the effective-
absorption band disappeared from the IR spectrum of cellulose.
ness of the bleaching operation (Dence and Reeve, 1996). CIE
Other typical absorption bands characteristic of pure cellulose
whiteness, WCIE%, is the most commonly used whiteness index.
are that at 3426, 1631, 1434, 1031 and at 895 cm1. One can also
The whiteness values for the two celluloses extracted from OP by
sulfite process are given in Table 6. Some CIE whiteness levels of
Table 6
about 75–79%, which can be considered as being approximately
Chemical composition and some characterization data regarding the bleached high, prove that the chosen process is an effective one. These val-
celluloses extracted from OP. ues would have exceeded these limits if the ash contents of the
Sodium sulfite Sodium metabisulfite
analyzed celluloses would have been lower.
process process The crystalline properties of celluloses extracted from OP by
means of sulfite reagents are shown in Table 6, as CrIs. The cellu-
Cellulose yielda 40.4 45.2
Cellulose content (%) 80.8 75.5 lose extracted in presence of sodium sulfite have a CrI of 43.7,
Hemicellulose content 13.5 18.1 whereas that extracted in presence of sodium metabisulfite have
(%) a CrI of 41.8. One can easily observe that CrI value of the first type
Ash content (%) 5.70 6.39
of cellulose material is higher than that corresponding to the sec-
Bulk density (g/cm3) 0.267 0.313
DPn 344 318 ond type. This finding can be explained through the higher hemi-
CrI (%) 43.7 41.8 cellulose content of the last, but not only. On the whole, these
WRV (%) 338 393 data constitute good evidence to conclude that the obtained cellu-
WCIE (%) 77.6 (45.3)b 79.3 (47.8)b loses are partially crystalline celluloses, that the amorphous por-
a
Determined by relating the weight of crude cellulose to the weight of double tion quantitatively exceeds the crystalline portion, and that the
extracted OP. crystallinity indexes of the two celluloses are lower than that of
b
WCIE values of unbleached celluloses are in parentheses. other common cellulose materials, of about 50–70% (Reddy and
 I. Bicu, F. Mustata / Bioresource Technology 102 (2011) 10013–10019
Yang, 2009). Going further with the reasoning, lower cellulose
crystallinity means a greater amorphous portion, and ultimately,
a greater permeability to water and other chemicals. This last valu-
able property of the two celluloses under discussion is also upheld
by their high water retention capability expressed as WRV, Table 6.
Indeed, the data in the table shows that the presence of those
somewhat large amounts of hemicellulose induced to the cellulose
obtained from OP the capability to absorb large amounts of water,
of up to four times the weight of the starting material. For all these
reasons, it can be presumed that these celluloses, likely to be re-
garded as dissolving celluloses, are especially suitable for chemical
processing applications, respectively for the preparation of cellu-
lose derivatives, or as fillers and water absorbents.
4. Conclusions
A sulfite pulping process in which the defatted and partially
depectined orange peel was cooked in aqueous liquors containing
sodium sulfite or sodium metabisulfite is provided. The response
surface methodology proved to be a useful tool in the determina-
tion of the optimum experimental conditions and in establishing
of predictive models for the extraction yield of cellulose. The opti-
mum yields, determined by reference to the dry weight of the dou-
ble extracted OP, were 40.4% and 45.2% for sodium sulfite and
sodium metabisulfite digestions, respectively. The physicochemi-
cal characterization data of these cellulose materials indicate good
levels of purity, low crystallinities, good whitenesses, moderate
molecular weights and good water retention capabilities. In accor-
dance with these specific properties OP celluloses could be used for
applications such as fillers in paper and other composites, water
absorbents, or as raw materials for cellulose derivatives.
References
Aravantinos-Zafiris, G., Oreopoulou, V., Tzia, C., Thomopoulos, C.D., 1994. Fibre
fraction from orange peel residues after pectin extraction. Lebensm. Wiss.
Technol. 27, 468–471.
Bampidis, V.A., Robinson, P.H., 2006. Citrus by-products as ruminant feeds. Animal
Feed Sci. Technol. 128, 175–217.
Biermann, C.J., 1996. Handbook of Pulping and Papermaking, second ed. Academic
Press (Elsevier), San Diego, CA, USA, pp. 72–73.
Braddock, R.J., Temelli, F., Cadwallader, K.R., 1986. Citrus essential oils – a dossier
for material safety data sheets. Food Technol. 40, 114–116.
Callaway, T.R., Carroll, J.A., Arthington, J.D., Pratt, C., Edrington, T.S., Anderson, R.C.,
Galiean, M.L., Ricke, S.C., Crandall, P., Nisbet, D.J., 2008. Citrus products decrease
growth of E. Coli O157: H7 and Salmonella typhimurium in pure culture and in
fermentation with mixed ruminal microorganisms in vitro. Foodborne Pathog.
Dis. 5, 621–627.
Cao, Y., Tan, H., 2005. Study on crystal structures of enzyme-hydrolysed cellulosic
materials by X-ray diffraction. Enzyme Microb. Technol. 36, 314–317.
Cochran, W.G., Cox, G.M., 1992. Experimental Designs, second ed. Wiley, New York,
pp. 335–375.
Dence, C.W., Reeve, D.W. (Eds.), 1996. Pulp Bleaching: Principles and Practice. TAPPI
Press, Atlanta, pp. 697–716.
Eaks, I.L., Sinclair, W.B., 1980. Cellulose–hemicellulose fractions in the alcohol-
insoluble solids of Valencia orange peel. Food Sci. 45, 985–988.
Ejikeme, P., 2008. Investigation of the physicochemical properties of
microcrystalline cellulose from agricultural wastes. I: orange mesocarp.
Cellulose 15, 141–147.
10019
Elg-Christoffersson, K., 2005. Dissolving Pulp-Multivariate Characterization and
Analysis of Reactivity and Spectroscopic Properties, Ph.D.Thesis, Umea
University, Umea, Sweden, 2005, pp. 14–15.
Ganz, E., Pauli, H.K.A., 1995. Whiteness and tint formulas of the Commission
Internationale de l’Eclairage: approximations in the L⁄a⁄b⁄ color space. Appl.
Opt. 34, 2998–2999.
Grohmann, K., Baldwin, E.A., Buslig, B.S., 1994. Production of ethanol from
enzymatically hydrolyzed orange peel by the yeast Saccharomices cerevisiae.
Appl. Biochem. Biotechnol. 45–46, 315–327.
Grohmann, K., Cameron, R.G., Buslig, B.S., 1995. Fractionation and pretreatment of
orange peel by dilute acid hydrolysis. Bioresour. Technol. 54, 129–141.
Klemm, D., Phillip, B., Heinze, T., Heinze, U., Wagenknecht, W., 1998.
Comprehensive Cellulose Chemistry. Vol 1, Fundamentals and Analytical
Methods. Wiley-VCH, Weinheim, Germany, pp. 234–235.
Laufenberg, G., Kunz, B., Nystroem, M., 2003. Transformation of vegetable waste
into value added products. (A) the upgrading concept; (B) practical
implementations. Bioresour. Technol. 87, 167–198.
Li, X., Tang, Y., Cao, X., Lu, D., Luo, F., Shao, W., 2008. Preparation and evaluation of
orange peel adsorbents for effective removal of cadmium, zinc, cobalt, and
nickel. Colloid Surf. A 317, 512–521.
Li, Q., Siles, J.A., Thompson, I.P., 2010. Succinic acid production from orange peel and
wheat straw by batch fermentation of Fibrobacter succinogenes S85. Appl.
Microbiol. Biotechnol. 88, 671–678.
Liu, Y., Shi, J., Langrish, T.A.G., 2006. Water-based extraction of pectin from flavedo
and albedo of orange peels. Chem. Eng. J. 120, 203–209.
Mamma, D., Kourtoglou, E., Christakopoulos, P., 2008. Fungal multienzyme
production on industrial by-products of the citrus-processing industry.
Bioresour. Technol. 99, 2373–2383.
Ma, E., Cervera, Q., Mejia Sanchez, G.M., 1993. Integrated utilization of orange peel.
Bioresour. Technol. 44, 61–63.
Mallakpour, S., Azizallahi, M.A., 2007. Aftertreatment of direct dyes on wool and
nylon 6 with synthans and a synthan/cationic system. Dyes Pigments 74, 258–
261.
Manthey, J.A., Grohmann, K., 1996. Concentrations of hesperidin and other orange
peel flavonoids in citrus processing byproducts. J. Agric. Food Chem. 44, 811–
814.
Marin, F.R., Soler-Rivas, C., Benavente-Garcia, O., Castillo, J., Perez-Alvarez, J.A.,
2007. By-products from different citrus processes as a source of customized
functional fibres. Food Chem. 100, 736–741.
Martin, M.A., Siles, J.A., El Bari, H., Chica, A.F., Martin, A., 2010. Orange peel: organic
waste or energetic resource? http://www.ramiran.net/ramiran2010/docs/
Ramiran2010_0083_final.pdf (accessed 25.02.2011).
Muresan, A., Cerempei, A., Dunca, S., Muresan, R., Butnaru, R., 2009.
Aromatherapeutic characteristics of cotton fabrics treated with rosemary
essential oil. Cellulose Chem. Technol. 43, 435–442.
Rangarajan, V., Rajasekharan, M., Ravichandran, R., Sriganesh, K., Vaitheeswaran, V.,
2010. Pectinase production from orange peel extract and dried orange peel solid
as substrates using Aspergillus niger. Int. J. Biotechnol. Biochem. 6, 445–453.
Reddy, N., Yang, Y., 2009. Natural cellulose fibers from soybean straw. Bioresour.
Technol. 100, 3593–3598.
Rivas, B., Torrado, A., Torre, P., Converti, A., Domínguez, J.M., 2008. Submerged citric
acid fermentation in orange peel autohydrolysate. J. Agric. Food Chem. 56,
2380–2387.
Siles López, J.A., Li, Q., Thompson, I.P., 2010. Biorefinery of waste orange peel. Crit.
Rev. Biotechnol. 30, 63–69.
Ververis, C., Georghiou, K., Danielidis, D., Hatzinikolaou, D.G., Santas, P., Santas, R.,
Corleti, V., 2007. Cellulose–hemicellulose, lignin and ash content of some
organic materials and their suitability for use as paper supplements. Bioresour.
Technol., 296–301.
Wang, Y.C., Chuang, Y.C., Hsu, H.W., 2008. The flavonoid, carotenoid and pectin
content in peels of citrus cultivated in Taiwan. Food Chem. 106, 277–284.
Wilkins, M.R., Widmer, W.W., Grohmann, K., 2007. Simultaneous saccharification
and fermentation of citrus peel waste by Saccharomyces cerevisiae to produce
ethanol. Proc. Biochem. 42, 1614–1619.
Xuan, Z., Tang, Y., Li, X., Liu, Y., Luo, F., 2006. Study on the equilibrium, kinetics and
isotherm of biosorption of lead ions onto pretreated chemically modified
orange peel. Biochem. Eng. J. 31, 160–164.
Yasar, F., Togrul, H., Arslan, N., 2007. Flow properties of cellulose and carboxymethyl
cellulose from orange peel. J. Food Eng. 81, 187–199.