As mentioned before, during the conversion of LCB into monomeric sugars, other type of products are formed and some of them can be strong inhibitors in fermentation bioprocesses. When compared to the fermentation of pure sugars, LCB hydrolysates present slower kinetics with a lower ethanol yield and productivity and in some cases a complete inhibition of growth and ethanol production can be observed. The variety and concentration of toxic compounds in feedstocks depend on both, the raw material and the pretreatment conditions applied for polysaccharides hydrolysis. The maximum concentration allowed for each inhibitor, without losing fermentation efficiency, depends on several factors: the origin of toxic compound, the inhibition mechanism, the microbial strain used and its physiological state, and also the fermentative process technology, the dissolved oxygen concentration in the medium and the pH (Mussatto et al. 2004).

Selection of a detoxification methodology for a specific feedstock is mandatory for attaining good results in 2nd generation bioethanol production. The identification of the main and relevant inhibitors present in the feedstocks is crucial in order to choose a specific, efficient and low-cost detoxification methodology. Besides, this knowledge can helps to establish the best conditions in hydrolysis pretreatment in order to minimize the inhibitors formation.

Fermentation inhibitors are conventionally classified in four groups according to their origin in lignocellulosics and hydrolysis processing: sugar degradation products, lignin degradation products, compounds derived from extractives and heavy metal ions (Parajo et al. 1998; Mussatto et al. 2004). Sugar degradation products are formed during hydrolysis and the main compounds produced are furfural from pentoses and 5-hydroxymethylfurfural (HMF) from hexoses as mentioned above. Furfural can inhibit cell growth, affecting the specific growth rate and cell-mass yield (Palmqvist et al. 2000b). However, it was noticed that some bioethanol-producing microorganisms like Pichia stipitis are not affected by furfural in low concentrations up to 0.5 g. L-1 (Mussatto et al. 2004). Moreover it could have a positive effect on cell growth. Nigam (2001) referred that ethanol yield and productivity were not affected by 0.27 g. L-1 of furfural. However concentrations above 1.5 g. L-1 interfered in respiration and inhibited cell growth almost completely, decreasing ethanol yield in 90% and productivity in 85% (Nigam 2001b). HMF has an inhibitory effect similar to that of furfural, but at a lower extension. Usually HMF is present in lower concentrations than furfural, due to its high reactivity and also due to the experimental conditions in the hydrolysis process that degrades lower amounts of hexoses. It was reported that HMF increases the lag phase extension and decreases cell growth (Delgenes et al. 1996; Palmqvist et al. 2000b). Mussatto et al. (2004) reported that a synergistic effect occurs when these compounds are combined with several other compounds formed during lignin degradation. Different compounds, aromatic, polyaromatic, phenolic, and aldehydic can be released from lignin during hydrolysis of LCB materials, and they are considered more toxic to microorganisms than furfural and HMF, even in low concentrations. Phenolic compounds are the most toxic products for microorganisms present in lignocellulosic hydrolyzates. They promote a loss of integrity in biological membranes, thus, affecting their ability as selective barriers and as enzyme matrices and decreasing cell growth and sugar assimilation (Parajo et al. 1998; Palmqvist et al. 2000b). Syringaldehyde and vanillic acid affect cell growth (Mussatto et al. 2004; Cortez et al. 2010) and the ethanolic fermentative metabolism of several microorganisms, like P. stipitis (Delgenes et al. 1996). In SSL, these compounds are normally present in the sulphonated form, due to the cooking process (Marques et al. 2009).

Extractives (acidic resins, taninic, and terpene acids) and also acetic acid derived from acetyl groups present in the hemicellulose are released during the hydrolytic processes. In terms of toxicity, the extractives are considered less toxic to microbial growth, than lignin derivatives or acetic acid (Mussatto et al. 2004). Gallic acid and pyrogallol are low molecular weight phenolic compounds normally formed from hydrolysable tannins (Marques et al. 2009) and some authors have shown anti-fungal properties of these phenolics (Dix 1979; Panizzi et al. 2002; Upadhyay et al. 2010). Acetic acid is also known as antimicrobial compound and the mechanism of inhibition is well-understood. At low pH, in the undissociated form, it can diffuse across the cell membrane, promoting the decrease of the cytoplasmatic cell activity and even causing cell death (Lawford et al. 1998; Mussatto et al. 2004). It has been reported that acetic acid inhibition degree depends not only on its concentration, but also on oxygen concentration and on pH of fermentation medium (Vanzyl et al. 1991). Another type of inhibitors are heavy metal ions, namely iron, chromium, nickel and copper, which result from reactors corrosion during the acidic hydrolysis pretreatment. Their toxicity acts at metabolic pathways level, by inhibiting enzyme activity (Mussatto et al. 2004).

As previously mentioned, a detoxification step is required before the hydrolysates undergo fermentation. Therefore, after identification of the toxic compounds, the choice of the best hydrolysate detoxification method is crucial for an effective and economical feasible detoxification methodology, in order to improve the fermentative process (Mussatto et al. 2004; Sanchez et al. 2008). Three different approaches have been described to decrease the concentration of inhibitors: (1) prevention of formation of inhibitors during the pretreatment step as mentioned before; (2) detoxification of the raw-material before fermentation; (3) development of microorganisms able to resist to inhibition.

Xavier and co-workers (2010) reported HSSL containing nearly 25 g. L-1 of xylose to P. stipitis for bioethanol production. Four increasing concentrations of HSSL were accessed to evaluate its toxicity. The results showed that increasing HSSL content in the fermentation medium decreased dramatically the maximum cell growth rate (pmax), ethanol yield (Yp/s) and productivity (qpm) attained. It was reported that HSSL content higher than 40% (v/ v) was critical for bioethanol production (Table 3). Acetic acid has been appointed as the main inhibitor of P. stipitis and other microorganisms (Schneider 1996; Lawford et al. 1998; Nigam

2001a). After the removal of acetic acid, ethanol fermentations were still unsuccessful, meaning that other compounds present had a toxic effect (Xavier et al. 2010).

Several biological, physical and chemical detoxification methods were developed in order to reduce inhibitor concentrations. The efficiency of detoxification methodology depends on chemical composition of the hydrolysate, as well as on microorganism chosen for bioethanol production (Mussatto et al. 2004; Helle et al. 2008; Sanchez et al. 2008). For this reason, the detoxification methods cannot be directly compared since mechanisms of inhibition and degree of toxicity removal are completely different (Palmqvist et al. 2000a).

Evaporation with vapour and vacuum evaporation are physical detoxification methods, in order to reduce the concentration of volatile compounds present in the hydrolysates, such as acetic acid, furfural and formaldehyde, and at the same time, to increase sugars concentrations. However, these methods also increase the non-volatile toxic compounds content, such as extractives and lignin derivatives. A balance between these two effects should be achieved or, consequently, the degree of fermentation inhibition will increase. Furthermore, the energy required for these processes should be properly considered to attain a potential economical gain (Lawford et al. 1993; Mussatto et al. 2004). As mentioned above, in the particular case of HSSL, evaporation is already implemented in the pulp production process for liquor concentration, to prepare it to burn for energy and chemical recovery. This is an advantage for HSSL bioconversion, and it is possible to optimise the evaporation stage, in order to get a good balance between volatile and non-volatile toxic compounds and sugar concentration for the fermentation process. Additionally, the condensate obtained in this step is rich in furfural and acetic acid, that can be easily extracted and purified for selling purposes as added-value products (Evtuguin et al. 2010). Alkali treatment, in particular overliming, is the most common detoxification method and is considered one of the best technologies. This method consists on the addition of lime (Ca(OH)2), or other alkali compound such as sodium or potassium hydroxide, until pH 9-10 promoting the precipitation of toxic compounds. Acetic acid, furfural, HMF, soluble lignin and phenolic compounds are mostly removed with this methodology, increasing the fermentability of hydrolysates. Several authors obtained the best results with alkali treatment using calcium hydroxide (Lawford et al. 1993; Martinez et al. 2001; Helle et al. 2008; Sanchez et al. 2008). Martinez et al. (2001) reported for sugarcane bagasse hydrolysate at 60 °C that the addition of Ca(OH)2 to adjust the pH to 9.0, promoted the precipitation of furanic and phenolic compounds. The obtained results showed a removal of nearly 51% and 41% respectively, of furans and phenolics with only 8.7% of sugars loss. Lawford et al. (1993) also used Ca(OH)2 for HSSL treatment at pH 10, followed by neutralisation to pH 7 with 1N of H2SO4. This methodology resulted in the improvement of the volumetric productivity and conversion efficiency, 92%, of bioethanol production by a recombinant strain of Escherichia coli.

Toxic compounds can also be removed by adsorption. Several authors have studied the capacity of removal of toxic compounds using different materials as adsorbents such as, activated charcoal (Dominguez et al. 1996; Lee et al. 1999; Mussatto et al. 2001; Canilha et al. 2004) and ion-exchange resins (Vanzyl et al. 1991; Larsson et al. 1999; Lee et al. 1999; Nilvebrant et al. 2001; Xavier et al. 2010). In particular, a specific strategy of adsorption on ion-exchange resins was employed by Xavier et al. (2010) to toxic compounds removal from HSSL for subsequent sugar purification and then ethanol fermentation with P. stipitis (Fig. 8).

In order to remove the cations added during pulping processing, namely Mg2+, HSSL was initially treated with a cation-exchange resin column. Then free carboxylic acids and polyphenols, including lignosulphonates, were separated from sugars with an anion — exchange resin in the second column. This process provided a transparent solution (sugars faction) containing essentially neutral monomeric sugars with traces of neutral polyphenolics (Table 4). However, this separation process released the sugars with some dilution and a concentration step was required for fermentation. This procedure led to excellent results of ethanol production by P. stipitis: high fermentation efficiency, 96%, productivity, 1.22 g. L-Lh’1, and yield, 0.49 g of ethanol / g of sugar.

Biological methods for detoxification of hydrolysates involve the use of specific enzymes or microorganisms that can degrade or consume the toxic compounds present in the hydrolysates. Jonsson et al. (1998) reported an increasing glucose consumption and ethanol productivity when wood hydrolysates were detoxified with laccase and peroxidase enzymes from Trametes versicolor, a white-rot fungus. These oxidative enzymes have the capability to degrade acid and phenolic compounds (Jonsson et al. 1998). The use of

anot detected

microorganisms was also proposed to remove inhibitors from HSSL. Xavier and co-workers

(2010) presented the first approach for HSSL biological detoxification, specifically for acetic acid removal. Four yeasts commonly used for acetic acid removal from wine were chosen, Candida tropicalis, Candida utilis, S. cerevisiae and Pichia anomala, and results are presented in Table 5.

According to these results, S. cerevisiae was selected for biological deacidification of HSSL. Sequential strategy of deacidification by S. cerevisiae and fermentation by P. stipitis on 60% of HSSL was carried out. Despite the acetic acid consumption by S. cerevisiae, xylose fermentation by P. stipitis produced only cell biomass, and no ethanol was detected in the medium. These results clearly showed the presence of other toxic compounds from HSSL, eventually phenolic compounds, probably inhibiting the sugars conversion to ethanol by P. stipitis (Xavier et al. 2010).

A different approach for performing biological detoxification of HSSL, with better results, was made in the same research group, using the Paecilomyces variotti filamentous fungus. This fungus can be found in air and soils of tropical countries, and has been studied for single cell protein (SCP) production, another important added-value product, normally used in animal feeding (Nigam 1999). Besides, P. variotti presents a good performance to grow in residues like HSSL and consumes substrates, including phenolic compounds, as carbon source. Pereira et al. (2011) showed for the first time the possibility of using this fungus to detoxify HSSL hydrolysates for subsequent ethanol fermentation. The biological treatment with P. variotti yielded HSSL with very low levels of acetic acid. Moreover, toxic compounds like gallic acid, pyrogalol and other low molecular phenolics were completely consumed and metabolized by P. variotti, indicating that this detoxification method can be suitable for treating HSSL into a proper feedstock for further bioprocessing. A successful fermentation of this detoxified HSSL by P. stipitis was performed, attaining an ethanol yield of 0.24 gethanol. gsugars-1. However, more research is required in order to improve the ethanol fermentation yields and productivities (Pereira et al. 2011).

Comparing the four different detoxification methodologies described, ion-exchange resins provided the best results on subsequent bioethanol fermentation (Table 6). High percentages of different toxic compounds from the hydrolysate were removed and provided the highest ethanol yield (0.49 g. g-1) and volumetric productivity (1.22 g. L-1.h-1). However, ion-exchange resins are expensive and difficult to implement and operate in large scale industries. P. variotti treatment, despite the fact of having promoted low ethanol fermentation yields in preliminary results (Table 6), appeared to be a very promising detoxification method. Furthermore the biomass of P. variotti can be used as SCP for animal feeding, increasing the economic potential of the process. More research work is being developed to combine this coupled strategy of biological detoxification of HSSL with simultaneous SCP production (Pereira et al. 2011). Other approaches for detoxification of hydrolysates were proposed and different methods can be used sequentially to improve their own capacity (Mussatto et al. 2004).