A new physical–chemical process for the efficient production of cellulose fibers from Spanish broom (Spartium junceum L.)

Bioresource Technology 101 (2010) 724–729

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

A new physical–chemical process for the efficient production of cellulose fibersfrom Spanish broom (Spartium junceum L.)

Bartolo Gabriele a,*, Teresa Cerchiara b,*, Giuseppe Salerno b, Giuseppe Chidichimo b, Mabel Valeria Vetere b,Cosimo Alampi b, Maria Caterina Gallucci b, Carmela Conidi c, Alfredo Cassano c

a Dipartimento di Scienze Farmaceutiche, Università della Calabria, 87036 Arcavacata di Rende (Cosenza), Italyb Dipartimento di Chimica, Università della Calabria, 87036 Arcavacata di Rende (Cosenza), Italyc Istituto per la Tecnologia delle Membrane, ITM-CNR, c/o Università della Calabria, 87036 Arcavacata di Rende (Cosenza), Italy

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

Article history:Received 22 May 2009Received in revised form 23 July 2009Accepted 6 August 2009Available online 5 September 2009

Keywords:BiomaterialsFibers extractionMorphological analysisNatural fibersSpanish broom

0960-8524/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.biortech.2009.08.014

* Corresponding authors. Tel.: +39 0984492813; faE-mail addresses: [email protected] (B. Gabriel

(T. Cerchiara).

A novel and efficient method for the extraction of cellulose fibers from Spanish broom (Spartium junceum L.)is presented. The method is based on the sequential combination between an initial chemical stage (alka-line digestion) and a subsequent physical–chemical stage, consisting of compression with hot air in anautoclave followed by rapid decompression (DiCoDe process, digestion–compression–decompression).The alkaline mother liquor deriving from the initial digestion step can be conveniently recycled after cen-trifugation followed by ultrafiltration. The process is characterized by the production of fibers with excel-lent physical–chemical properties, such as high mechanical resistance and high elasticity, and rapidproduction times. The fibers obtained after the DiCoDe process can be further softened and whitenedby means of enzymatic digestion.

Fibers were morphologically characterized by scanning electron microscopy (SEM), while their compo-sition and physical–chemical properties were determined by conventional methods, including colorime-try, TAPPI protocols, IR spectroscopy, and X-ray diffractometry.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Natural fibers play a fundamental role in our life and society.There are many different fibers that are extensively producedand used every day, such as cotton, jute, flax, or hemp. Althoughthe possibility to obtain fibers from broom (Spartium junceum L.)has been known for many years (Angelini et al., 2000; Avellaet al., 1998), the rather low efficiency of the extraction processfrom its branches (known as vermenes) has so far limited the pos-sibility of an extensive application on a large scale of broom fibersin the textile industry, despite the availability of the raw materialand the highly attractive characteristics of the final fibers, such asbiodegradability and high mechanical resistance (Mohanty et al.,2005). The main problems in the extraction of cellulose fibers frombroom vermenes are connected with the difficulty in removing thecortical cuticle, basically made of long-chained alkenes, alcohols,and esters, and in detaching the fibers from the internal ligneousand pectin substances (Trotter, 1941). These difficulties cause anexcessive slowness of the process, usually associated to the pro-duction of coarse fibers, which finds limited applications in thetextile industry (Foster, 2006).

ll rights reserved.

x: +39 0984492044.e), [email protected]

In this work, we wish to report a novel and efficient methodol-ogy for the production of broom cellulose fibers from vermenes.Our method, recently patented (Chidichimo et al., 2007), is charac-terized by the production of fibers with excellent physical–chemi-cal properties, such as high mechanical resistance and highelasticity, and rapid production times, of the order of a few hours.

2. Methods

2.1. Chemicals and materials

Broom vermenes (S. junceum L.) were obtained from a researchfield of the University of Calabria, Arcavacata di Rende (CS), Italy.NaOH and lignin peroxidase were purchased from Sigma–Aldrich.

2.2. Procedure for the extraction of cellulose fibers from broomvermenes (DiCoDe process; the experimental flow sheet is shown inFig. 1)

2.2.1. First stage: (a) alkaline digestion and (b) recycling of the alkalinemother liquor

(a) Alkaline digestion: dry Spanish broom vermenes (5 kg, ca.50 cm length) were immersed in a hot (110 �C) solution (10 L) ofNaOH in water (15% w/w for 15 min or 5% w/w for 1 h). The hot

B. Gabriele et al. / Bioresource Technology 101 (2010) 724–729 725

sprigs were then washed with cold (25 �C) water (5 L) by means ofshower diffusers immediately after their removal from the diges-tion pot. The cold water used for washing the sprigs was allowedto fall into the digestion pot, which was allowed to cool to roomtemperature. After separation of the solid ligneous part by meansof a metallic comb, the fibers obtained were compressed to removethe excess of water. The humid fibers thus obtained had a moisturecontent of ca. 33%.

(b) Recycling of the alkaline mother liquor: the alkaline mother li-quor deriving from the digestion stage (see above) could be recy-cled after suitable removal of the suspended solids. After coolingof mother liquor to room temperature, the precipitated substanceswere removed by centrifugation (at 1500 rpm for 5 min). The li-quid phase was then subjected to ultrafiltration (UF, see Section2.6 for details) to remove the remaining suspended solids. Waterand NaOH were then added to the permeate to obtain ca. 10 L ofa solution at pH ca. 14, which was reutilized as such for a secondalkaline digestion stage.

2.2.2. Second stage: compression–decompressionThe final humid fibers obtained from the first stage (ca. 33%

moisture content) were introduced into a 50 L stainless steel auto-clave in the presence of air. The autoclave was pressurized at roomtemperature with air (10 atm) and then heated at 120 �C for 3 h.While hot, the autoclave was degassed very rapidly by means ofa suitable valve. The collected fibers were then washed with coldwater (25 �C), mechanically dewatered by compression and thendried. The typical yield of the dry fibers thus obtained was 0.8 kg,with a length of ca. 30 cm.

2.3. Enzymatic softening of broom fibers

An optional additional stage of the process consisted in theenzymatic softening of the fibers obtained after the DiCoDe treat-ment. In this case, the broom fibers, obtained as described above(Section 2.2) (ca. 500 mg) were subjected to enzymatic digestionin water (100 mL, pH 3.00 at T = 30 �C) for 1 h in the presence oflignin peroxidase from Phanerochaete crysosporium (0.1 mg/mL,0.1 U/mg). After washing with water and drying, softer andwhither fibers with respect to the original ones were obtained.

2.4. Measurements and physical–chemical characterization of broomfibers

The amount of cellulose in the broom fibers obtained as de-scribed above was determined using a colorimetric method withthe anthrone reagent (Updegraff, 1969). Pentosans and lignin weredetermined according to the TAPPI T 223 cm-84 and TAPPI T222 cm-02 methods, respectively (TAPPI, 1984, 2002). The ash con-tent of the fiber was determined by weighing the residue remain-ing after ignition at 575 �C for 3 h (Han and Rowell, 1997). Pectinswere determined colorimetrically using the carbazole method(McComb and McCready, 1952).

FT-IR spectra were obtained with a BIORAD spectrophotometerusing KBr disks. A LEO 420 scanning electron microscope (SEM)was used to observe the morphological features of the untreatedand treated fibers. The specimens to be observed were mountedon a conductive adhesive tape, sputter coated with gold–palladiumand observed at the microscope using a voltage of 15 kV. X-ray pat-terns were recorded on a Philips PW 3710 diffractometer using CuKa radiation. A Leitz Laborlux 12Pol. Optical microscope (OM) wasused to observe the fibers obtained after the steam explosionprocess.

The tensile properties of the fibers in terms of the tenacity andpercentage of breaking elongation were determined using an In-stron 6021/5500 tensile testing machine according to UNI EN ISO

5079 at room temperature (20 �C) and 65% relative humidity.About 100 fibers were tested and the average is reported.

2.5. Chemical analysis of the solids precipitated from the alkalinemixture

The lignin content of the solids precipitated from the alkalinemixture (see Section 2.2) was estimated gravimetrically. The resi-due was dissolved in dioxane/water (9:1 v/v). The mixture was al-lowed to stand overnight at 4 �C in 1 M acetic acid. The lignin wasthen recovered by centrifugation, washed twice with a minimumamount of cold water and then freeze-dried (Durot et al., 2003).

Pectins were extracted using 0.3 N HCl under reflux at 90 �C for45 min. The extract was centrifugated and the supernatant wascollected and dispersed in 2-propanol. Pectin was precipitatedusing HCl 0.1 N at pH 2. The precipitate was collected, centrifuged,dispersed in 2-propanol and centrifuged. The washing was re-peated three times. The product was dispersed in water, freeze-dried and analyzed by colorimetric method (Kalapathy and Proctor,2001).

Pentosans and the ash content were determined, respectively,by TAPPI T 223 and TAPPI T 244. Fats were extracted with a soxhletapparatus with a toluene–ethanol mixture (2:1 v/v) for 8 h. The ex-tract was dried at 85 �C to constant weight (Ogbonnaya et al.,1997).

Chlorophyll content was determined spectrophotometrically byethanol extraction (Papista et al., 2002) and metal–organic com-plexes were determined by atomic absorbance.

2.6. Experimental details for ultrafiltration (UF) operation

The UF treatment of the liquid phase obtained after centrifuga-tion of the cooled alkaline mixture (see Section 2.2) was performedusing a laboratory pilot unit supplied by Verind SpA (Rodano, Mi-lan, Italy). The equipment consists of a 25 L stainless steel feedtank, a feed pressure pump, a pressure control system, a thermom-eter, two manometers (0–400 kPa) located at the inlet (Pin) and atthe outlet (Pout) of the membrane module and a magnetic flow me-ter for the measure of the axial feed flow rate (Qf). A tube and shellheat exchanger, placed after the feed pump, was used to maintainthe temperature of the liquor constant. A data acquisition system,permitting the continuous monitoring of the transmembrane pres-sure and of the axial feed flow rate, was connected to the UF plant.A digital balance, connected to the system, was used to measurethe quantity of permeate collected per time unit and consequentlythe permeate fluxes. The UF plant was equipped with a hollow fi-ber membrane module (membrane material polysulphone, nomi-nal molecular weight cut-off 100 kDa, maximum operatingtemperature 50 �C, membrane surface area 1.2 m2, inner fiberdiameter 2.1 mm) supplied by China Blue Star Membrane Technol-ogy Co., Ltd. (Beijing, China). The alkaline mother liquor was fed inthe lumen side of the fibers while the permeate was collected inthe tube side of the membrane module. UF experiments were per-formed according the batch concentration mode, collecting sepa-rately the permeate stream and recycling the retentate in thefeed tank of the pilot unit. The UF was operated at room tempera-ture (20 ± 2 �C), with an axial feed flow rate of 490 L/h and a trans-membrane pressure of 42 kPa. The membrane module was rinsedwith tap water for 30 min after the treatment of the alkaline li-quor; then it was submitted to a cleaning procedure with a NaOHsolution at a concentration of 1% (w/w) and at a temperature of40 �C for 60 min. A final rinse of the system with tap water for atleast 20 min was carried out. After each cleaning procedure, thewater flux of the membrane module in fixed conditions (tempera-ture 25 �C, axial feed flow rate 500 L/h) was measured.

726 B. Gabriele et al. / Bioresource Technology 101 (2010) 724–729

2.7. Physical–chemical analyses of liquid samples coming from UFoperation

Samples of feed, permeate and retentate coming from the UFexperiments (see Section 2.6) were collected and stored at�20 �C for further analyses.

Total dry solids (TDS) were measured by drying weighed sam-ples at 105 �C and determining the weight of the residue. The ashcontent was measured by heating the residue from the TDS mea-surement of the liquors to 575 �C for 3 h, and weighing the sampleafterwards. The lignin content was determined by measuring theUV light absorption at 280 nm. (Hill et al., 1988), using a ShimadzuUV-160A spectrophotometer (Shimadzu Scientific Instruments,Inc., Japan). Before measurements, samples were diluted with0.1 M NaOH (Jönsson et al., 2008). The adsorption constant usedwas 21.1 (L/g cm) (Jönsson et al., 2008).

The suspended solid content was determined in relation to thefeed solution (w/w%) by centrifuging, at 2000 rpm for 20 min,45 mL of a pre-weighted sample; the weight of settled solids wasdetermined after removing the supernatant.

FT-IR spectrum of pectins present in the UF permeate was ob-tained with a Jasco 430 spectrophotometer using KBr disks.

3. Results and discussion

3.1. Extraction of cellulose fibers from Spanish broom (Fig. 1) and theirmorphological analysis by scanning electron microscopy (SEM)

Our method for the highly efficient production of broom fibersis essentially based on the sequential combination between an ini-tial chemical stage, consisting of an alkaline digestion, and a subse-quent physical–chemical stage, involving compression with hot airin an autoclave followed by rapid decompression. The whole pro-cess can therefore be indicated with the acronym DiCoDe (diges-tion–compression–decompression).

The alkaline digestion (first stage) consists in the immersion ofvermenes of length of ca. 50 cm in a NaOH solution (5–15% w/w)maintained at 70–110 �C for a period ranging from 15 min to 3 h.As expected, a lower digestion time requires a higher base concen-tration and/or a higher temperature. Typical experiments were car-ried out with 15% w/w NaOH solution at 110 �C for 15 min or with5% w/w NaOH solution for 1 h. During the alkaline digestion, themajor part of the substances that constitute the external cuticleare hydrolyzed, while the pectin substances bound to the fibersof cellulose are in part deprotonated. Thus, after washing of thehot sprigs with cold water immediately after their removal fromthe alkaline digestion bath, followed by separation of the solid lig-neous part by means of a metallic comb, bundles of fibers are ob-tained. These fibers are still rather thick, as the transversalthickness is in the order of hundred of microns. They are also rel-atively stiff, and are basically made of elementary cellulose fibersstill incorporating pectins and lignins. In the hot alkaline motherliquor, substances of low solubility, such as pectins, hemicellu-loses, and lignins, are present in suspension, while partly solublesubstances (such as partly deprotonated pectins, polyalcohols,polyphenols, and deprotonated fatty acids) are in solution. Aftercooling, the separation of low-soluble substances gradually in-

Table 1Chemical composition of untreated broom and of broom fibers after different DiCoDe trea

Treatment Ce

500 mg of raw material 44After treatment with 5% NaOH for 1 h; 3 h autoclave (500 mg of fibers) 77After treatment with 15% NaOH for 15 min; 3 h autoclave (500 mg of fibers) 91

creases until no further precipitation takes place. These substancescan be easily separated by centrifugation. It is worth noting thatthe bio-mass thus obtained, after suitable washing to remove theresidual alkalinity, can potentially find practical application forthe production of fertilizers for agriculture or, after suitable fer-mentation processes, for the production of biofuels. On the otherhand, the liquid phase obtained after centrifugation can be easilyrecycled, after ultrafiltration (to remove the remaining suspendedsolids, see Section 3.5 for details) followed by the addition of thesuitable amounts of water and of the base necessary to adjustthe final alkaline concentration to the value required by theprocess.

The fibers obtained from the alkaline digestion stage, after elim-ination of the excess of water by simple compression, is then sub-jected to a physical–chemical process, in order to further purify theelementary fibers from the pectin and lignin residues and, at thesame time, to obtain the reduction of their diameters. This process(second stage) consists in the compression of the wet fibers in astainless steel autoclave under 5–70 atm of hot air (100–120 �C)for 0.5–4 h (typically, 10 atm at 120 �C for 3 h), followed by a rapiddecompression of the system by opening a suitable valve withoutprevious cooling of the mixture. It is worth noting that our processis different from the classical steam explosion process (Garcia-Jal-don et al., 1998), which employs a violent steam of superheatedwater, and which leads to coarser final fibers, as evidenced by opti-cal microscopy analysis. Under our conditions, the pectin and lig-nin residues are partially decomposed by oxidation, and thedegradation products thus formed are easily removed from the ele-mentary fibers by simple washing with water. Moreover, the com-pression–decompression sequence causes the break of thereticulations inside the fibers. As a consequence, the final elemen-tary fibers, as observed by scanning electron microscopy (SEM)analysis, appear well separated and with very small diameters(7–10 lm). The mean length of the elementary fibers, determinedaccording to literature (Ververis et al., 2004) was 25–27 mm. Thesedimensions are typical of elementary fiber extracted from otherplants (Reddy and Yang, 2007); however, it should be noted thatthe possibility to obtain elementary fibers with diameters of about10 lm from Spanish broom has been achieved so far only by moreexpensive mechanical or physical–chemical processes under dras-tic conditions (Angelini et al., 2000).

3.2. Enzymatic softening of the fibers

The fibers obtained with the DiCoDe process can be further soft-ened by an additional, short enzymatic digestion process (1 h),using ligninolytic enzymes. Whiter and softer fibers with respectthe original ones are obtained in this way. The SEM morphologicalanalysis of the fibers obtained after the enzymatic treatmentshows well separated elementary fibers.

3.3. Physical–chemical properties and composition of the fibers

The FT-IR spectrum of the treated fibers is significantly differentfrom that of the untreated fibers. In particular, the untreated fiberpresents a band centered at 1734 cm�1, which disappears inthe spectrum of the treated fiber. This is in agreement with the

tments.

llulose (%) Lignin (%) Pentosans (%) Pectins (%) Ash (%)

.5 ± 0.2 18.5 ± 0.3 16.3 ± 0.1 13.3 ± 0.1 4.0 ± 0.2

.1 ± 0.4 5.0 ± 0.1 7.7 ± 0.2 9.2 ± 0.3 1.0 ± 0.1

.7 ± 0.1 3.2 ± 0.4 4.1 ± 0.3 0.0 ± 0.0 1.0 ± 0.2

B. Gabriele et al. / Bioresource Technology 101 (2010) 724–729 727

removal of pectins as a consequence of the DiCoDe treatment, sincethe band centered at 1734 cm�1 is attributed to the C@O stretchingof methyl ester and carboxylic groups present in pectins (Ouajai

Table 2Mechanical properties of broom fibers extracted under different DiCoDe treatments.

Treatment Tenacity(cN/tex)

Strain atbreak (%)

5% NaOH for 1 h; 3 h autoclave 20.4 ± 1.5 4.7 ± 1.015% NaOH for 15 min; 3 h autoclave 35.9 ± 1.6 5.8 ± 1.715% NaOH for 15 min; 3 h autoclave; 1 h

lignin peroxidase41.1 ± 1.5 6.5 ± 1.6

Flax 16.9 1.9

Dry vermene

1st Stage: Alkal(10 L NaOH 15% w/

Hot sprigs

Washing with cold water (25°C, ~5 L)

Sprigs at R.T.

Mechanical removal of the solid ligneous part

Solid ligneous part (~ 3Kg)

Humid sprigs (~ 33% moisture)

2nd Stage: Compression in autoclave (10 atm, 120°C, 3 h)

Rapid decompression (with the autoclave still hot)

Hot fibers

Fibers at R.T.

Dry fibers (~0.8 Kg)

Drying

Washing

Washing with cold water (25 °C)

Fig. 1. Flow sheet of the experimental procedure for the

and Shanks, 2005). The band appearing at ca. 1750 cm�1 can beattributed to the acetyl and uronic ester groups of residual hemi-celluloses or to the ester linkage of the ferulic and p-coumaric acidsof lignin (Sun et al., 2005). In addition, in both spectra are presentsome characteristic bands of residual lignin in the region 1500–1600 cm�1, corresponding to the aromatic skeletal vibration(Moràn et al., 2008), as well as some characteristic cellulose bandsaround 1000–1200 cm�1 (Zhbankov et al., 2000; Cao and Tan,2004). In particular, the band near 1160 cm�1 is representative ofthe antisymmetric bridge stretching of C–O–C groups in cellulose(Sekkal et al., 1995).

The broom fibers, obtained after the DiCoDe treatment, have avery similar X-ray diffraction pattern with respect to that of com-

s (~5 Kg)

ine digestion w, 110°C, 15 min)

Separation

water

Hot alkaline mixture

Alkaline mixture at R.T.

Centrifugation to remove the precipitated solids

Precipitated solids (~700 g)

Chlorophyll: ca. 4% Pectins:15-20% Metal-organic complexes: ca. 30% Pentosans: 5-8% Lignin: 15-20% Fats: 10-15% Ash: ca. 15%

Alkaline solution

U.F. at R.T.

Solid retentate (~ 450 g)

Alkaline permeate

Addition of NaOH and water

Alkaline solution to be recycled in the alkaline digestion stage

Cooling at R.T.

extraction of cellulose fibers from broom vermenes.

Table 3Composition of feed, permeate and retentate during ultrafiltration of the cooledalkaline solution.

Sample Suspendedsolids (%)

Ash(%)

Lignin(g/L)

Dissolved solids(g/L)

Feed 5.10 6.30 3.90 81.00Permeate 0.93 6.33 3.15 64.50Retentate 5.72 7.68 4.60 79.50

728 B. Gabriele et al. / Bioresource Technology 101 (2010) 724–729

mercial cellulose, even though the intensity of the peaks is lower.This is due to a lower degree of crystallinity (69%) with respectto the commercial cellulose (100%). Both fibers (DiCoDe fibersand commercial cellulose) have the prominent cellulose peak at2h angle of 22� representing the 0 0 2 plane. The characteristic1 0 1 and 10 �1 peaks (2h between 13� and 18�) are not distinctand are combined into one broad peak. On the other hand, in theraw broom bundles diffractogram no peaks appear at 22� (2h)and in the range from 13� to 18� (2h), respectively. This is due tothe lower amounts of cellulose and higher amounts of lignin as re-ported in Table 1.

The final physical–chemical properties of the fiber can be easilymodulated by varying the conditions used in the alkaline digestionstep. The composition of the fibers obtained under different diges-tion conditions (treatment with 5% NaOH for 1 h or with 15% NaOHfor 15 min) is reported in Table 1, together with the composition ofuntreated broom for comparison. The reduced content of pectinsfrom 13.3% to 0.0%, pentosans from 16.3% to 4.1% and lignin from18.5% to 3.2%, after the DiCoDe treatment, confirms that part ofthe components which make fibers stiff have been removed. As ex-pected, there is a clear tendency for the residual non-cellulosicsubstances to be reduced by increasing the severity of treatment(such as NaOH concentration).

3.4. Mechanical properties of the fibers

The mechanical properties of the fibers obtained under differentdigestion conditions (treatment with 5% NaOH for 1 h and with15% NaOH for 15 min) are reported in Table 2. It can be seen thattenacity and strain at break increase with the severity of treatment(such as NaOH concentration). The maximum tenacity (35.9 cN/tex) was obtained with 15% NaOH. This is in agreement with amore efficient removal of the non-cellulosic substances whenemploying a higher base concentration. In fact, if the removal ofthese substances is more efficient, the inter-fibrillar regions be-come less dense and less rigid.

As shown in Table 2, the tenacity of the fibers results even high-er after the additional enzymatic softening stage (see Section 3.2).Also, the strain at break of the softened fiber is slightly better thanthat of the fibers deriving from the DiCoDe process (Table 2), so wecan conclude that the enzymatic softening does not deteriorate thefiber quality.

We have also compared the obtained broom fibers with one ofthe most common fiber currently used (flax, Table 2). As can beseen from the tensile properties shown in Table 2, the broom fibers

VRF

0 2 4 6 8 10 12 14 16

J p (k

g/m

2 h)

0

2

4

6

8

10

12

14

16

18

Fig. 2. Ultrafiltration of the liquid phase obtained after centrifugation of the cooledalkaline mixture. Permeate flux vs. volume reduction factor (VRF).

obtained after the DiCoDe treatment turned out to be even moreelastic than flax fibers.

3.5. Ultrafiltration (UF) operation

Fig. 2 shows the permeate flux vs. the volume reduction factor(the ratio between the initial volume and the residual retentatevolume, VRF) for a generic UF run of the liquid phase obtained aftercentrifugation of the cooled alkaline mixture (see Section 3.1), per-formed under the selected operating conditions. The initial perme-ate flux of about 16 L/m2 h decreased to 9.7 L/m2 h when the VRFreached a final value of 14.5. Despite the high VRF obtained andthe high osmotic pressure of the solution reached at the end ofthe process, the flux decay observed during the UF treatment canbe considered modest indicating a low fouling index of themembranes.

In Table 3 the composition of the feed solution, permeate andretentate in terms of suspended solids, total dry solids (TDS), ashand lignin is reported. About 82% of suspended solids were re-moved by the UF membrane; this value (lower than 100%) can bedue to low molecular weight deprotonated pectins (<100 kDa) ableto permeate through the membrane. The rejection of the mem-brane towards TDS was 20%.

About 80% of the lignin was recovered in the permeate; it isworth noting that a similar result was obtained by Jönsson et al.(2008) in the ultrafiltration of hardwood black liquor with15 kDa ceramic membrane. The ash content of the initial feedwas not affected by the UF process.

The FT-IR spectrum of recovered pectins in the permeate solu-tion contains characteristic broad bands of carboxylic groups atca. 1600–1730 cm�1 (Monsoor et al., 2001).

The permeate solution was reused for the alkaline digestion ofthe vermenes after addition of the suitable amounts of water andof the base necessary to adjust the final alkaline concentration tothe value required by the process. The solution could then be recy-cled without significantly influencing the physical–chemical prop-erties of the fibers. Moreover, UF membranes can be used toseparate lignin and hemicelluloses from the spent liquor. The per-meate, containing most of the lignin and only a minor amount ofhemicelluloses, can in principle be concentrated by nanofiltration(Jönsson et al., 2008) and the obtained retentate solution can beutilized as a precursor for carbon fiber production (Kadla et al.,2002).

4. Conclusions

In conclusion, we have reported a new and efficient methodol-ogy for the production of cellulose fibers with excellent physical–chemical properties from Spanish broom (S. junceum L.). Thisnew methodology is based on the combination between an initialalkaline digestion stage and a subsequent compression–decom-pression stage with hot air in an autoclave (DiCoDe process). Thefinal fibers obtained have excellent physical–chemical properties(such as high mechanical resistance and high elasticity), and canbe further softened and whitened by means of an additional enzy-matic digestion stage.

B. Gabriele et al. / Bioresource Technology 101 (2010) 724–729 729

The ultrafiltration of the liquid phase obtained after centrifuga-tion of the alkaline liquor mother is a viable method for recyclingand reusing the solution according to the proposed processscheme, thus making the overall process particularly convenient.

Acknowledgements

This work was supported by the Ministero dell’Istruzione,dell’Università e della Ricerca (MIUR, Roma, Italy) (Project No.987, ‘‘Development and optimization of processes for the produc-tion of materials from broom fibers”) and by the University ofCalabria.

References

Angelini, L.G., Lazzeri, A., Levita, G., Fontanelli, D., Bozzi, C., 2000. Ramie (Boehmerianivea (L.) Gaud.) and Spanish broom (Spartium junceum L.) fibres for compositematerials: agronomical aspects, morphology and mechanical properties. Ind.Crops Prod. 11, 145–161.

Avella, M., Casale, L., Dell’Erba, R., Focher, B., Martuscelli, E., Marzetti, A., 1998.Broom fibers as reinforcing materials for polypropylene-based composites. J.Appl. Polym. Sci. 68, 1077–1089.

Cao, Y., Tan, H., 2004. Structural characterization of cellulose with enzymatictreatment. J. Mol. Struct. 705, 189–193.

Chidichimo, G., Alampi, C., Cerchiara, T., Gabriele, B., Salerno, G., Vetere, M.V., 2007.Physical Chemical Process for Production of Vegetable Fibers. World Patent WO2007/102184 A2 (13.09.2007).

Durot, N., Gaudard, F., Kurek, B., 2003. The unmasking of lignin structures in wheatstraw by alkali. Phytochemistry 63, 617–623.

Foster, H.B., 2006. Greek sparto: past and present. Text. J. Cloth Cult. 4, 36–66.Garcia-Jaldon, C., Dupeyre, D., Vignon, M.R., 1998. Fibres from semi-retted hemp

bundles by steam explosion treatment. Biomass Bioenergy 14 (3), 251–260.Han, J.S., Rowell, J.S., 1997. Chemical composition of fibers. In: Rowell, R.M., Young,

R.A., Rowell, J.K. (Eds.), Paper and Composites from Agro-Based Resources. CRCPress, London (Chapter 5).

Hill, M.K., Violette, D.A., Woerner, D.L., 1988. Lowering black liquor viscosity byultrafiltration. Sep. Sci. Technol. 23, 1789–1798.

Jönsson, A.S., Nordin, A.K., Wallberg, O., 2008. Concentration and purification oflignin in hardwood kraft pulping liquor by ultrafiltration and nanofiltration.Chem. Eng. Res. Des. 86, 1271–1280.

Kadla, J.F., Kubo, S., Venditti, R.A., Gilbert, R.D., Compere, A.L., Griffith, W., 2002.Lignin-based carbon fibers for composite fiber applications. Carbon 40, 2913–2920.

Kalapathy, U., Proctor, A., 2001. Effect of acid extraction and alcohol precipitationconditions on the yield and purity of soy hull pectin. Food Chem. 73, 393–396.

McComb, E.A., McCready, R.M., 1952. Colorimetric determination of pecticsubstances. Anal. Chem. 24, 1630–1632.

Mohanty, D.A.K., Misra, M., Drzal, L.T. (Eds.), 2005. Natural Fibers, Biopolymers andTheir Biocomposites. CRC Press, London.

Monsoor, M.A., Kalapathy, U., Proctor, A., 2001. Determination of polygalacturonicacid content in pectin extracts by diffuse reflectance Fourier transform infraredspectroscopy. Food Chem. 74, 233–238.

Moràn, J.I., Alvarez, V.A., Cyras, V.P., Vàzquez, A., 2008. Extraction of cellulose andpreparation of nanocellulose from sisal fibers. Cellulose 15, 149–159.

Ogbonnaya, C.I., Roy-Macauley, H., Nwalozie, M.C., Annerose, D.J.M., 1997. Physicaland histochemical properties of kenaf (Hibiscus cannabinus L.) grown underwater deficit on a sandy soil. Ind. Crops Prod. 7, 9–18.

Ouajai, S., Shanks, R., 2005. Morphology and structure of hemp fibre afterbioscouring. Macromol. Biosci. 5, 124–134.

Papista, E., Acs, E., Boddi, B., 2002. Chlorophyll-a determination with ethanol – acritical test. Hydrobiologia 485, 191–198.

Reddy, N., Yang, Y., 2007. Characterizing natural cellulose fibers from velvet leaf(Abutilon theophrasti) stems. Bioresour. Technol. 99, 2449–2454.

Sekkal, M., Dincq, V., Legrand, P., Huvenne, J.P., 1995. Investigation of the glycosidiclinkages in several oligosaccharides using FT-IR and FT Raman spectroscopies. J.Mol. Struct. 349, 349–352.

Sun, X.F., Xu, F., Sun, R.C., Fowler, P., Baird, M.S., 2005. Characteristics of degradedcellulose obtained from steam-exploded wheat straw. Carbohyd. Res. 340, 97–106.

TAPPI method T 222 cm-02, 2002. Acid-Insoluble Lignin in Wood and Pulp. TAPPIPress, Norcross, GA.

TAPPI method T 223 cm-84, 1984. Pentosans in Wood and Pulp. TAPPI Press,Norcross, GA.

Trotter, A., 1941. La Ginestra. Arte della Stampa, Roma.Updegraff, D.M., 1969. Semimicro determination of cellulose in biological materials.

Anal. Biochem. 32, 420–424.Ververis, C., Georghiou, K., Christodoulakis, N., Santas, P., 2004. Fiber dimensions,

lignin and cellulose content of various plant materials and their suitability forpaper production. Ind. Crops Prod. 19, 245–254.

Zhbankov, R.G., Firsov, S.P., Korolik, E.V., Petrov, P.T., Lapkovski, M.P., Tsarenkov,V.M., Marchewka, M.K., Ratajczak, H., 2000. Vibrational spectra and thestructure of medical biopolymers. J. Mol. Struct. 555, 85–96.