Milox fractionation of empty fruit bunches from Elaeis guineensis Ana Ferrer a , Alberto Vega b , Alejandro Rodríguez a , Pablo Ligero b , Luis Jiménez a,⇑ a Department of Chemical Engineering, University of Córdoba, Spain b Department of Physical Chemical and Chemical Engineering, University of A Coruña, Spain article info Article history: Received 4 April 2011 Received in revised form 5 July 2011 Accepted 13 July 2011 Available online 28 July 2011 Keywords: Empty fruit bunches Pulping Milox process Bio-refinery Sugars abstract Milox pulping of EFB was used to obtain pulps. In the first Milox stage, the influences of operating vari- ables on pulp properties were studied and polynomial and neural fuzzy models that reproduced the experimental results with errors less than 10% were developed. Operating variables were found (93 wt.% of formic acid, 3 wt.% of hydrogen peroxide, and 165 min) that yielded acceptable pulp proper- ties (40.5% yield, 50.3% brightness and 608 mL/g viscosity) at reasonable chemical and energy costs. The second stage was studied by subjecting the liquors of the previously optimized first stage to different treatment times. This time should be 30 min or less, to avoid a negative effect on viscosity. The residual liquor from the first Milox stage contained virtually no precipitable lignin and only low amounts of sugar (wt.%): glucose 0.71, xylose 4.22, galactose 1.19, mannose 0.22, all on original raw material dry. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Since it has become difficult to meet the demand for pulp for paper and other used with classic raw materials such as hard- and softwoods, alternative sources such as non-woody materials, fast growing plants, and residues from agricultural and food and forest industries are being explored (Jiménez et al., 2006). An inter- esting non-woody material is empty fruit bunches (EFB), the resid- ual lignocellulosic material from the palm oil industry. Pulps have been produced from EFB by alkaline processes involving the Kraft method (Ibrahim, 2002, p. 40; Jiménez et al., 2009a), as well as the soda process (Law and Jiang, 2001). Daud et al. (1998) obtained pulps with soda, sodium carbonate and sodium sulfite processes and found that the first process was the most efficient. Ibrahim (2002) compared the composition of EFB pulps obtained after Kraft, Kraft anthraquinone, soda and soda–anthraquinone pro- cesses, and determined that the soda process yielded the highest viscosity and contents of lignin, holocellulose and a-cellulose. Wanrosli et al. (2004) employed a central composite experi- mental design to study the influence of temperature, time and al- kali concentration on the yield, Kappa number, tensile index and tear index of EFB pulps finding that a relatively low temperature (about 160 °C) within the limits of pulping time (60–120 min) and of alkali change between 20% and 30% is generally sufficient. The refining of EFB cellulose pulps obtained by soda–anthraqui- none has been investigated using an experimental design. It was found that operating at 15% soda, 170 °C, 70 min and 2400 number of PFI refining revolutions is possible to obtain cellulose pulps with properties that differ less than 12% of their optimal values (59.63 Nm/g tensile index, 4.17 kN/g burst index and 7.20 mNm 2 / g tear index), and a Shopper–Riegler value of 47.5° SR. These values are acceptable to obtain paper sheets (Jiménez et al., 2009a). In or- der to use the EFB to obtain corrugating medium-grade paper board, Roliadi and Pasaribu (2004) carried out laboratory scale experiments with a semi-synthetic pulp soda–anthraquinone (4– 10% soda, 0–0.15% anthraquinone, 135 °C and 3 h). Daud et al. (2005) and Ghazali et al. (2006) have also used the EFB in the alka- line peroxide mechanical pulping process (APMP), and found that the chemical and mechanical pretreatments removed 95% manga- nese, 45% calcium, 37% extractives, 81% ash and 84% silica initially present in the lignocellulosic material. Organosolv processes involving ethanol (Suleman and Yusoff, 1997; Aziz et al., 2002), the modified IDE process (Quader and Lonnberg, 2005), or high-boiling point organic solvents (Rodríguez et al., 2008) have also been employed, and biopulping with the white rot fungus K14 (isolated from tropical plantation) decreased lignin content more rapidly than with Phanerochaete chrysosporium or Colorius versicolor (Goenadi et al., 1998). Total chlorine free (TCF) bleaching of soda–anthraquinone and diethanolamine pulps from EFB was studied by Jiménez et al. (2009b) who found that dietha- nolamine pulp produced better paper than soda–anthraquinone pulp when an OZP (oxygen–ozone–peroxide) bleaching sequence was employed. Organosolv methods for pulping of lignocellulosic materials are an attractive alternative to the Kraft process (Muurinen, 2000; 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.07.037 ⇑ Corresponding author. Address: Department of Chemical Engineering, Campus of Rabanales, C-3, University of Córdoba, Córdoba, Spain. Tel.: +34 957218658; fax: +34 957218625. E-mail address: iq1jiall@uco.es (L. Jiménez). Bioresource Technology 102 (2011) 9755–9762 Contents lists available at SciVerse ScienceDirect Bioresource Technology journal homepage: www.elsevier.com/locate/biortech