Examples of adsorption systems for removal of fermentation inhibitors from lignocellulosic biomass hydrolysate

A major challenge faced in the commercial production of lignocellulosic bio-ethanol is the inhibitory compounds generated during the pretreatment of biomass. During the dilute acid and hot water pre-treatment, the major degradation by-products released are organic acids such as acetic acid (formed by hydrolysis of hemicellulose), furans such as 5-hydroxymethylfurfural (HMF), furfural (compounds derived from degradation of hexose and pentose sugars), and phenols such as vanillin, 4-hydroxybenzaldehyde (4-HB), and syringaldehyde (released by lignin degradation) [60]. These compounds are very toxic to fermenting species [55–57]. Each of these compounds may affect cell growth, sugar uptake, individual metabolic pathways, or all three [61]. A number of studies have been reported to reduce the concentration of these inhibitors using adsorption techniques such as adsorption on activated charcoal, adsorbent resins, ion-exchange resins, and zeolites. In this section, some of these studies are briefly described.

Acetic acid is produced from the hydrolysis of acetyl groups in the hemicellulose [62]. Acetic acid in protonated form can diffuse through the cytoplasmic membrane of cells and detrimentally affect cell metabolism [63]. Berson et al. [64] investigated adsorption of acetic acid from a dilute acid pretreated corn stover hydrolysate on activated carbon. The initial hydrolysate had acetic acid concentration of 16.5 g l⁻¹. The hydrolysate filtered from the corn stover hydrolysate slurry was contacted with activated carbon (Calgon BL) provided by Calgon in a shake flask. The operating conditions were: activated carbon concentration of 80 g l⁻¹; shaker speed 350 rpm; temperature 35^◦C and contact time 10 minutes. At the end of experiment, carbon was removed from hydrolysate using centrifugation. To bring down the concentration of acetic acid from 16.5 to a level below 2 g l⁻¹, the hydrolysate was again brought in contact with activated carbon and the this step was repeated five times. Simultaneous adsorption of glucose and xylose was not measured during these stages. A US patent application [65] reported usage of a weak basic anion resin to adsorb inhibitors and mainly the acid molecules (such as acetic acid).

Wickramasinghe and Grzenia tested an adsorptive microporous membrane, Sartobind Q, and a weak base anion exchange resin, Amberlyst 21, to recover acetic acid from the hemicellulose hydrolysate on lab scale [66]. Sartobind Q is made of cross-linked regenerated cellulose membrane with a strong basic anion exchange group R-CH2-N⁺-(CH3)₃ attached to their internal pores and Amberlyst 21 is made of Polystyrene macroreticular with styrene functional group. They showed that adsorptive membranes had many advantages over resin based system such as

• Reduced pore diffusion because adsorbate transport to the adsorbent site occurs by convection.

• Reduced processing time.

• Lower pressure drop compared to the resin packed bed.

• Scale-up for membrane system is easier than with the packed bed [67–69].

Lee et al. [70] reported a study on detoxification of inhibitors from prehydrolysate produced from auto-hydrolysis of mixed hardwood chips using activated carbon. The acid treated prehydrolysate had:

xylose=13.18 g l⁻¹; glucose=2.42 g l⁻¹; acetic acid=4.58 g l⁻¹; HMF=0.08 g l⁻¹; furfural=0.51 g l⁻¹; and formic acid=4.04 g l⁻¹. The detoxification experiments were carried out in 250 ml flasks kept in a shaking water bath at 50^◦C and 180 rpm speed. The concentrations of activated carbon (05-690A, 50–200 mesh, Fisher Scientific Co., Pittsburgh, PA, USA) used were 1.0, 2.5, 5.0, and 10.0 wt % with incubation time of 1 h. Figure 5.7 shows the percentage removal of toxic compounds against percentage activated carbon used.

From Figure 5.7, it can be seen that activated carbon at 2.5 wt% level in prehydrolysate could remove 96% of hydroxymethylfurfural (HMF), 93% of the furfural, 42% of the formic acid and 14% of the acetic acid. However, it also removed 8.9% of the xylose. The subsequent fermentation with Thermoanaer-obacterium saccharolyticum strain MO1442 gave essentially a 100% yield. This study demonstrated that activated carbon is a good adsorbent to remove inhibitors like HMF and Furfural. Its adsorption capacity for acetic acid is limited probably due to the competitive inhibition of other compounds present in the complex hydrolysate. This is consistent with Bersonet al.’s study [64], which showed that five adsorption cycles were required to remove acetic acid completely from corn stover hydrolysate.

Larsson et al. [57] used polystyrene divinylbenzene-based anion-exchange resin (AG 1-X8, Bio-Rad, Richmond, VA) to detoxify the hydrolysate of dilute acid pretreated Spruce. The adsorption treatment with anion resin at pH 10 showed the highest inhibitor removal compared with other treatments such as pH adjustment to 10 with NaOH or Ca(OH)₂, sulfite treatment, hydrolysate concentration by evaporation, enzymatic degradation using laccase, and microbial degradation usingTrichoderma reesei. Table 5.5 shows reduction in hydrolysate inhibitor concentrations by adsorption with anion resin AG1-X8.

Carvalhoet al. [71] studied detoxification of hemicellulosic hydrolysate produced from dilute sulfuric acid pre-treatment of Eucalyptus shavings using combination of activated carbon and adsorbent resins.

Initially the hydrolysate was concentrated 5.8 times by vacuum evaporation, which helped to remove the

0 0

% removal of species 25 50 75 100

2 4 6

Activated carbon applied (wt %)

8 10

Formic acid Acetic acid HMF Furfural Xylose

Figure 5.7 Percentage removal of inhibitors versus activated carbon charge for 1 hour adsorption time.

Reprinted from [70] c2011, with permission from Elsevier

Table 5.5 Detoxification of Spruce hydrolysate by adsorption using anion resin AG1-X8. Reference [57], with kind permission from Springer Science+Business Media c1999

Treatment glucose+mannose

g l⁻¹

Levulinic acid g l⁻¹

Acetate g l⁻¹

Furfural gl⁻¹

5-HMF g l⁻¹

Total phenolic content g l⁻¹

Feed 32.2 2.6 2.4 1.0 5.9 0.48

Anion resin, pH 5.5 (0.45 g per g of hydrolysate)

29.6 0.36 0.26 0.69 4.37 0.15

Anion resin, pH 10 (0.49 g per g of hydrolysate)

23.8 0.18 0.10 0.27 1.77 0.043

more volatile furfural. The concentrated hydrolysate was treated with combination of different adsorbents as mentioned below:

1. Adsorption on activated charcoal and subsequently on the anionic resin Purolite A-860 S.

2. Adsorption on activated charcoal and subsequently on the adsorbent resin Purolite MN-150.

3. Adsorption on diatomaceous earths and subsequently on the anionic resin Purolite A-860 S.

4. Adsorption on diatomaceous earths and subsequently on the adsorbent resin Purolite MN-150.

The adsorption process on charcoal and diatomaceous earth was carried in a shake flask. The operating conditions were: adsorbent loading=1.2 g adsorbent per 50 ml of hydrolysate; speed=200 rpm; tempera-ture=30^◦C; residence time=34.5 min. The adsorption processes on resins were performed in a packed bed column with a bed volume of 200 cm³ with a feed rate of 0.9 cm³min⁻¹. Figure 5.8 shows the percent toxic compounds removed for each process.

From Figure 5.8, by comparing the four processes, the combination of charcoal and adsorbent resin MN-150 (Process B) was found to be the most effective for removal of acetic acid, HMF and phenolics with minimum sugar loss (5.9 and 1.3% for glucose and xylose respectively).

0

Removal (%)

Glucose Xylose Acetic acid

Furfural Compound

HMF Phenolics 10

20 30 40

A Detoxification Procedures

BC 50 D 60 70 80 90 100

Figure 5.8 Sugar and inhibitors removal from concentrated eucalyptus hydrolysate for different adsorption processes. Reprinted from [71] c2006, with permission from John Wiley & Sons

From the earlier examples it can be seen that anion exchange resins seem to be among the promising adsorbents for detoxifying the cellulosic hydrolysate. However it suffers from some disadvantages: (i) pH needs to be adjusted to 10 thus requiring significant quantities of acid and base chemicals; (ii) a significant amount of fermentable sugar loss up to 26%. Another promising adsorbent is activated charcoal, which does not require pH adjustment. Activated carbon also is found to adsorb sugar molecules to some extent.

Adsorption process with activated charcoal can become expensive sometimes because the powdered form of activated charcoal cannot be regenerated and granular activated charcoal usually incurs a 10% loss during regeneration cycle [72]. Zeolite materials are found to overcome most of these disadvantages, especially the sugar loss.

Ranjan et al. [72] used different zeolite adsorbents (ZSM-5 (framework type MFI), beta, faujasite and ferrierite (framework type FER)) to recover 5-hydroxymethylfurfural (HMF), furfural and vanillin from lignocellulosic biomass hydrolysates. Hydrolysate was prepared from Aspen wood chips using dilute sulfuric acid pretreatment. The hydrolysate had total sugar content of 37.5 g l⁻¹, 1.45 g l⁻¹ of furfural, 0.16 g l⁻¹ of HMF and 0.05 g l⁻¹ of vanillin. The detoxification of hydrolysate was carried out by adding 10 g l⁻¹ β zeolite and the mixture was stirred overnight followed by zeolite removal by centrifugation.

The treated hydrolysate had no detectable amount of inhibitors and the sugar loss was also negligible.

This improved the ethanol yield of the fermentation process significantly (Figure 5.9a). Figure 5.9b shows the single component adsorption isotherm for furfural, vanillin and HMF onβ zeolite with different Si/Al ratios. From Figure 5.9b it can be seen that high silica zeolite more efficiently removed furfural, vanillin and HMF.

Inhibitors such as HMF and furfural can be used as building blocks for the production of fine chemicals and plastics [73, 74]. Hence recovery of these inhibitors as pure compounds can be considered to have a promising future. Ranjan et al. [72] also reported a preliminary study on the selective adsorption of furfural over HMF on MFI/FER zeolite. They followed the screening process developed by Gaounaris et al. [75] to select a suitable adsorbent. Figure 5.10 shows the preferential adsorption of furfural over HMF on FER zeolite from the binary mixture. FER zeolite also showed negligible adsorption of sugars.

0 0.00

Amount adsorbed (g/g of zeolite)

0.05 0.10 0.15 0.20 0.25

4 8

Solution concentration (g/L) (b)

12

Xylose Si/Al = 100 Si/Al = 12

HMF Furfural Vanillin

16 20

0 0

Total Sugar (g/L) Ethanol produced (g/L)

10 20 30 40

0 10 20 30 40 50

10 20

Time (h) (a)

30 40

Figure 5.9 (a) Effect of zeolite pretreatment on ethanol fermentation. (b) Single component adsorption of furfural, vanillin and HMF onβzeolite. Reprinted from [72] c2009, with permission from Elsevier

HMF Furfural

0 0

−1Amount adsorbed (g g zeolite) 0.01 0.02 0.03 0.04 0.05 0.06

10 20

Solution concentration (g L⁻¹)

30 40

Figure 5.10 Adsorption isotherm of HMF and Furfural binary mixture on FER zeolite. Reprinted from [72]

c 2009, with permission from Elsevier

5.7.2 Examples of adsorption systems for recovery of biofuels from dilute

In document Separation and Purification Technologies in Biorefineries (Pldal 145-149)