Manipulation of ultrasonic effects on lignocellulose by varying the frequency, particle size, loading and stirring

Manipulation of ultrasonic effects on lignocellulose by varying the frequency, particle size, loading and stirring

Accepted Manuscript Manipulation of Ultrasonic Effects on Lignocellulose by Varying the Frequen‐ cy, Particle size, Loading and Stirring Madeleine J. ...

707KB Sizes 0 Downloads 32 Views

Accepted Manuscript Manipulation of Ultrasonic Effects on Lignocellulose by Varying the Frequen‐ cy, Particle size, Loading and Stirring Madeleine J. Bussemaker, Feng Xu, Dongke Zhang PII: DOI: Reference:

S0960-8524(13)01357-6 http://dx.doi.org/10.1016/j.biortech.2013.08.106 BITE 12305

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

16 April 2013 16 August 2013 19 August 2013

Please cite this article as: Bussemaker, M.J., Xu, F., Zhang, D., Manipulation of Ultrasonic Effects on Lignocellulose by Varying the Frequency, Particle size, Loading and Stirring, Bioresource Technology (2013), doi: http:// dx.doi.org/10.1016/j.biortech.2013.08.106

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

MANIPULATION OF ULTRASONIC EFFECTS ON LIGNOCELLULOSE BY

2

VARYING THE FREQUENCY, PARTICLE SIZE, LOADING AND STIRRING

3

MADELEINE J. BUSSEMAKER *, FENG XU , AND DONGKE ZHANG

1

1

4

1,2

1

Centre for Energy (M473), The University of Western Australia, 35 Stirling

5

Highway, Crawley, WA 6009, Australia 2

6

Institute of Biomass Chemistry and Technology, Beijing Forestry University,

7

Beijing 100083, China *

8

Corresponding author email: [email protected]

9

+61 8 6488 8026

10

ABSTRACT 

11

The parameters, including ultrasonic frequency, still versus stirring, biomass

12

particle size and biomass loading were concurrently investigated for the ultrasonic

13

treatment of wheat straw. Experiments were conducted at three different frequencies; 40,

14

376, and 995 kHz using three different solid to liquid ratios, 1/50, 1/20, and 1/15 (g/ml),

15

with and without mechanical stirring. Additional treatments in different particle size

16

ranges, 0 – 0.5, 0.5 – 1, and 1 – 2 mm were performed at the solid to liquid ratio of 1/20

17

(g/ml). Fractionation was improved at 40 and 995 kHz via different mechanisms.

18

Delignification was favored at the ultrasonic treatment frequency of 40 kHz, biomass

19

loading 1/20 (g/ml) with stirring and particle size range of 0.5–1 mm. However at 995

20

kHz carbohydrate solubilization was favored, especially in the particle size range of <

21

0.5 mm. The treatment efficacies highlighted the use of ultrasound for physical and

22

chemical effects at different frequencies.

1

1 2 3 4

Keywords: fractionation; lignocellulose; pretreatment; sonochemistry; ultrasound; wheat straw 1. INTRODUCTION  In order to utilize lignocellulosic biomass for biofuel or biorefinery purposes the

5

biomass must be pretreated to enable either enzymatic or chemical conversion of

6

particular components (Amidon et al., 2008). The conversion of wheat straw presents

7

numerous technical challenges specifically in the pretreatment step. Pretreatment and

8

fractionation must include a chemical mechanism to enhance the separation of the

9

lignocellulosic components in order to facilitate the utilization of cellulose,

10

hemicellulose or lignin (Hon & Shiraishi, 2001). Ideally in a biorefinery all the

11

components of the biomass would be utilized, however the removal of lignin increases

12

hydrolysis yields of sugar components as highlighted in recent reviews (Hendriks &

13

Zeeman, 2009; Kumar et al., 2009). Here the use of ultrasound as a pretreatment

14

technology for a model lignocellulose, wheat straw was considered.

15

The viability of ultrasound as a pretreatment option for lignocellulose was reviewed

16

and it was found that in order to realize the full potential of ultrasound the parametric

17

influences need to be better understood (Bussemaker & Zhang, 2013). The parameters

18

explored in this manuscript were biomass loading, biomass particle size, frequency and

19

stirring. Each of these parameters affects the degree of influence of ultrasound has on

20

the pretreatment. Time was kept uniform in order to evaluate the parametric effects

21

under otherwise constant conditions. In the literature, a majority of ultrasonic

22

pretreatment was less than an hour, with a concentration around the 30 min treatment

23

(Bussemaker & Zhang, 2013). Ultrasonic treatment is energy intensive and hence

24

shorter treatment times are more desirable. For the extraction of lignin and

25

carbohydrates from wheat straw the majority of the initial ultrasonic benefits occurs 2

1

within the first 20-30 min of treatment (Sun & Tomkinson, 2002; Sun et al., 2002). In

2

addition, longer treatment times are known to promote higher lignin condensation on

3

the biomass (Garcia et al., 2011; Sul'man et al., 2011). Therefore, for the purpose of an

4

evaluation of the parametric differences in ultrasonic treatment of wheat straw, a shorter

5

treatment is often chosen.

6

The solid loading determines how much biomass can be processed at one time but

7

also affects the ultrasound attenuation. A range of loadings under the same conditions

8

were investigated for lignocellulose delignification (Sasmal et al., 2012) and hydrolysis

9

(Velmurugan & Muthukumar, 2011). Optimal loadings varied between the studies as

10

the purposes and feedstocks were varied. The effect of biomass loading on ultrasonic

11

treatment was only investigated at low frequencies and there is a lack of good

12

understanding of this effect at high frequencies.

13

Generally ultrasonic treatment in the literature deals with biomass screened to

14

smaller than 1 mm. Ultrasonic treatment was successful in enhancement of hydrolysis

15

(Yunus et al., 2010), and fractionation and delignification (Garcia et al., 2011) for

16

particle size ranges over 1 mm. These studies demonstrate ultrasound can be effective

17

over a range of particle sizes. However delignification (Velmurugan & Muthukumar,

18

2012) and sugar production (Esfahani & Azin, 2012) were more effective at lower

19

particle sizes. Grinding to a small size is an energy intensive step in the consideration of

20

pretreatment on a larger scale. Therefore the efficacy of ultrasound over a range of

21

particle sizes is an important consideration for any treatment purpose.

22

Frequency affects the mechanoacoustic and sonochemical effects of ultrasound. At

23

lower frequencies, (under 100 kHz) the cavitation and collapse are more violent and at

24

higher frequencies, more active bubbles are generated, producing more radicals in

25

solution (Entezari & Kruus, 1994; Kanthale et al., 2008; Mason et al., 2011). Previous 3

1

work demonstrated ultrasound enhanced enzymatic and acidic hydrolysis of cellulose to

2

glucose (Velmurugan & Muthukumar, 2012; Yachmenev et al., 2009; Yunus et al.,

3

2010), delignification of lignocellulose (Baxi & Pandit, 2012; Garcia et al., 2011;

4

Sasmal et al., 2012; Sul'man et al., 2011) and liquefaction of lignocellulose (Garcia et

5

al., 2011; Kunaver et al., 2012). The augmentation of ultrasound for these purposes may

6

occur through chemical or physical effects of ultrasound. However, up until now, no

7

comparison has been made on the effects of ultrasound on lignocellulose over a range of

8

frequencies.

9

Stirring or agitation of an ultrasonic field can also influence the sonochemical yields

10

and acoustic streaming. Stirring and ultrasound was successful in improving enzymatic

11

hydrolysis (Yachmenev et al., 2009), delignification (Baxi & Pandit, 2012) and

12

liquefaction (Kunaver et al., 2012), at low frequencies. However the effect of stirring

13

was not evaluated in these studies.

14

The literature on ultrasonic pretreatment of lignocellulose demonstrated ultrasound

15

could be a viable option for the enhancement of subsequent treatment. Furthermore,

16

ultrasonic theory demonstrated parameters will influence the ultrasonic field and

17

subsequent ultrasonic effects. Hence it was hypothesized that the parameters: frequency,

18

particle size, loading and reactor configuration would have significant differences on the

19

efficacy of the ultrasonic pretreatment. To the best of the authors’ knowledge, no

20

studies have been presented comparing the concurrent effects of, frequency, particle

21

size, and stirring on ultrasonic pretreatment of lignocellulose. These parameters were

22

investigated on an agricultural residue, wheat straw, and evaluated in terms of

23

fractionation, delignification and carbohydrate solubilization, compared to a control

24

treatment.

4

1 2

2. MATERIALS AND METHODS  2.1

Material

3

Wheat straw was sourced from the Wheatbelt region of Western Australia and was

4

manually downsized prior to processing in a Waring Blender for 1 min. The resultant

5

particle size ranged up to 5 mm in diameter with a particle size distribution of 1.8%,

6

15.0%, 30.5% and 49.0% for ranges, > 2 mm, 2 - 1 mm, 1 - 0.5 mm and < 0.5 mm,

7

respectively, utilized for the experiments of different loadings. The sample was then

8

sieved into several size fractions of 0 – 0.5, 0.5 – 1, and 1 – 2 mm, respectively, for

9

treatments with different particle sizes. The raw composition of the feedstock was 35 ±

10

3%, 24 ± 3% and 14.9 ± 0.5% for cellulose, hemicellulose and lignin, respectively, as

11

determined following the National Renewable Energy Laboratory (NREL) procedures

12

(Sluiter et al., 2008b).

13

2.2

Pretreatment methods

14

Pretreatments were conducted using two different reactors sourced from Meinhart

15

Ultraschalltechnik. The first reactor was of diameter of 90 mm and a frequency of 40

16

kHz and the second reactor was of diameter 57 mm and a transducer base of 50 mm

17

with frequencies of 376 and 995 kHz. For experiments with overhead stirring and the

18

control experiment, a Caframo overhead stirrer was used at 300 rpm. The blade

19

diameter was 55 mm, positioned 65 mm from the surface of the transducer in the 40

20

kHz reactor and 138 mm from the transducer plate in the multi-frequency reactor. The

21

power of each reactor was measured for comparison via standard calorimetric

22

techniques (Koda et al., 2003), determined as: 92 ± 1, 76 ± 4, and 72 ± 7 W L-1 for 40,

23

376 and 995 kHz, respectively. This corresponds to a maximum temperature increase of

24

33, 27 and 26 K for 40, 376 and 995 kHz respectively, based on sonication of 500 ml of 5

1

water with specific heat capacity 4.2 J g-1 K-1. The control experiment was conducted in

2

a beaker with overhead stirring. For all pretreatments, 500 ml of air saturated deionized

3

water at room temperature was used and the pretreatment time was kept constant at 25

4

minutes. At each frequency, (control, 40, 376 and 995 kHz), three different solid to

5

liquid ratios (SLR), of 1/50, 1/20 and 1/15 (g/ml) were tested. Additional experiments at

6

different particle sizes were performed at SLR 1/20 (g/ml), at each frequency with

7

stirring. Each pretreatment and analysis was repeated at least twice and the average and

8

standard deviation were determined.

9

2.3

10

Analysis of solid and liquid fractions

After pretreatment the solid fraction was filtered, and the liquor collected for analysis.

11

The solid residue was then washed with deionized water and air-dried. The yield of the

12

solid fraction was measured gravimetrically and the pH recorded. The compositional

13

analysis for acid insoluble lignin, cellulose and hemicellulose in the solid and liquid

14

divisions was determined following standard procedures set out by NREL (Sluiter et al.,

15

2008a; Sluiter et al., 2008b).

16

Scanning electron microscopy (SEM) images of the air-dried solid residues, of the

17

different particle size ranges after treatment at various frequencies, coated with gold,

18

were captured on a Phillips XL30 ESEM.

19

2.4

20

Isolation of milled wood lignin in the solid residue

Treatments at SLR 1/20 (g/ml) of the mixed biomass were repeated six times at each

21

frequency, including the control, and combined for subsequent lignin isolation. After

22

pretreatment the straw was dewaxed with a 2:1 toluene:ethanol mixture via Soxhlet

23

extraction (6 h). The straw was then oven dried for 16 h at 55 °C.

6

1

The lignin was isolated following classical ball milled wood lignin isolation methods

2

(Bjorkman, 1956), with some modification. Briefly, the dewaxed sample (45 g) was ball

3

milled for 5 hours over 10 hours with a planetary ball mill (Fritsch, pulverisette 6) using

4

a zirconium jar and zirconium balls at 500 rpm. The samples, (40 g) were then extracted

5

twice in dioxane (96%) at room temperature for 24 hours with a solid to liquid ratio of

6

1/10 (g/ml) while kept in the dark. The residue was separated from the dioxane solution

7

after each extraction via centrifuge and washed with 96% dioxane. The combined

8

supernatants and washing liquor from each extraction were concentrated to 50 ml or less

9

under reduced pressure at a maximum temperature of 55 °C. The concentrated solution

10

was then added to 150 ml of 95% ethanol and allowed to stand for at least 5 hours for

11

precipitation of hemicelluloses in solution. The solid precipitation was then separated

12

via centrifugation, and the supernatants were combined and concentrated to less than 50

13

ml under reduced pressure at maximum temperature of 55 °C. The concentrated liquor

14

was added dropwise to 500 ml of hydrochloric acid (pH 1.5), with stirring. The acidified

15

solution was allowed to stand for 2 hours prior to separation of the solid and liquid

16

division via centrifugation. The precipitate was washed with hydrochloric acid (pH 1.5)

17

and freeze-dried in a minimal amount of distilled water.

18

2.5

Characterization of isolated lignin

19

The structure of the isolated lignin was characterized using infrared spectrometry.

20

The Fourier transform infrared (FTIR) spectra were collected using a Thermo Scientific

21

Nicolet iN10 FTIR Microscope (Thermo Nicolet Corporation, Madison, WI) equipped

22

with a liquid nitrogen cooled mercury-cadmium-telluride (MCT) detector. Spectra were

23

taken in the range of 4000–650 cm-1 at 4 cm-1 with 64 scans per sample.

7

1

3. RESULTS AND DISCUSSION 

2

3.1

3

3.1.1

4

composition

5

Treatments at different frequencies

Effect of treatment at different frequencies on solid residue yield and

The fractionation was improved at 40 and 995 kHz, as measured by the solid residue

6

yield (Figure 1). At the SLR 1/20 (g/ml) with stirring, the treatment at 376 kHz had a

7

higher solid residue yield (59%) than at 995 and 40 kHz (55.2 and 56.2% respectively).

8

Additionally, without stirring, at the lowest SLR of 1/50 (g/ml) 376 kHz treatment

9

produced a higher solid residue yield (56.8%) as compared to the treatments at 995 and

10

40 kHz (53.2 and 54.1% respectively). This trend was also evident for the 1 – 2 mm

11

fraction, where 995 kHz and 40 kHz treatments resulted in lower solid residue yields

12

than the treatment at 376 kHz: 61.8 and 62.0% compared to 64.3%. Therefore

13

ultrasound at 40 and 995 kHz was solubilizing the biomass more than at 376 kHz.

14

The fact that 376 and 995 kHz had comparable power but different solid yield results

15

indicated there was more than a heating effect which caused the observed differences in

16

the results. Therefore the effect of frequency on the composition of the solid residue

17

should elucidate the mechanisms of solubilization at high and low frequency.

18

Higher frequencies were more effective at carbohydrate solubilization (Figure 2).

19

The hemicellulose content was reduced by 6.2 and 9.1% with treatment of the < 0.5 mm

20

fraction at 995 kHz, compared to treatments at 376 and 40 kHz, respectively. Similarly,

21

the cellulose content was reduced by 6.9 and 10.2% with treatment of the < 0.5 mm

22

fraction at 995 kHz, compared to treatments at 376 and 40 kHz, respectively. Moreover,

23

at 40 kHz, SLR 1/20 (g/ml), with stirring the cellulose remaining was increased by 2.8%

24

compared to the control, however at 995 kHz without stirring the remaining cellulose

25

was reduced by 6.1% compared to the control. Then, the total glucose detected in the 8

1

pretreatment liquors at 995 kHz, SLR, 1/50 (g/ml), with stirring was greater; 9.0%

2

compared to 6.6% for the control and 6.9% for 376 kHz treatment (Figure 3). Therefore

3

the fractionation increase at 995 kHz was attributed to an increase in solubilization of

4

the carbohydrate component of the biomass.

5

The lowest frequency, 40 kHz, was most effective for delignification (Figure 2). At

6

SLR 1/20, with stirring, the lignin remaining in the solid residue after treatment at 40

7

kHz was reduced by 7.2 and 5.7% compared to 376 and 995 kHz treatments,

8

respectively. Then at SLR, 1/50 (g/ml), the lignin remaining in the solid residue after 40

9

kHz was reduced by 1.7 and 3.4% compared to 376 and 995 kHz treatments,

10

respectively. Moreover, the percentage of lignin remaining after treatment of the 0.5 – 1

11

mm fraction decreased by 3.2, 4.7 and 3.9%, compared to the control for 40, 376 and

12

995 kHz, respectively. For the control, 376 and 995 kHz treatments the percentage of

13

hemicellulose and cellulose remaining decreased or remained the same in comparison to

14

the treatment for the 0.5 – 1 mm fraction to the < 0.5 mm fraction. Yet the lignin

15

percentage in the solid fraction was comparable for all pretreatments for the < 0.5 mm

16

fraction. Logically the smaller particle size would enable solubilization due to increased

17

accessibility. However, after 40 kHz treatment, the hemicellulose remaining was not

18

different between the two fractions (both 24.3% of the solid residue) and the cellulose

19

remaining was higher at the lower particle size range (33.4% compared to 32.4% of the

20

solid residue). The different response to the treatment at 40 kHz was indicative of

21

degradation of another biomass component, in this case, lignin. Therefore the

22

composition yields indicated that delignification was most efficient at 40 kHz.

23

The characterization showed a structure typical to milled wood lignin with attached

24

hemicellulose. The FTIR assignment was completed with reference to previous

25

literature (Table 1) (Sun & Tomkinson, 2002; Sun et al., 2000; Sun et al., 2005). 9

1

Aromatic skeletal vibrations were assigned to the bands at 1594, 1508 and 1423 cm-1

2

and the hemicellulose signals of acetyl groups (1732 cm-1), glycosidic linkages v(C-O-C)

3

and C-O and C-C stretching (1048 cm-1) and the arabinosyl side chains (1166 and 977

4

cm-1) were observed (Sun et al., 2005). No significant differences between the spectra of

5

the milled wood lignin isolated from different pretreatments were observed. Therefore,

6

the delignification, favored at 40 kHz did not result in alteration of the structure of

7

native lignin in the solid residue.

8

3.1.2

Mechanisms of ultrasonic interaction with biomass at different frequencies

9

The sonochemical effects at these settings increased with frequency which coincided

10

with an increase in carbohydrate degradation. Cleavage of lignin-carbohydrate linkages

11

was enhanced by ultrasound in previous literature (Sun & Tomkinson, 2002). The

12

detection of increased total glucose amounts in solution at high frequencies indicated

13

the radicals produced by ultrasound were responsible for this cleavage. Furthermore, the

14

decrease in remaining carbohydrates in the solid fraction, only without stirring at high

15

frequency, supports the theory that sonochemical effects were responsible for

16

solubilization of the carbohydrate component of the biomass. Previous observations of

17

work into stirring at high frequencies in this reactor demonstrated the radical yield was

18

not affected at higher input power or decreased at lower input power with an increase in

19

stirring speed. Evidence of increased attenuation of the ultrasonic wave with stirring of

20

a heterogeneous solution was shown with the pitch change at 40 kHz. Consequently

21

stirring at high frequencies in heterogeneous solutions would also attenuate the wave,

22

decreasing the radical production. Therefore at higher frequencies the increased

23

solubilization of carbohydrates was attributed to the sonochemical effects of increased

24

radical production.

10

1

Mechanoacoustic effects are maximized at lower frequencies. The physical

2

disturbance was evidenced in the pits, accompanied by cracking, in the SEM images.

3

The pits were only observed for the 0.5 – 1 mm fraction where lignin content was

4

reduced. The pits were approximately 200 µm in diameter. Acoustic cavitational pits

5

were reported to be larger for aluminium foil (1.5 mm), however, the actual size would

6

depend on the thickness and hardness of the material as well as the occurrence of pit

7

clustering (Dular et al., 2013). Nevertheless, the size of the pits were congruent with

8

pits from hydrodynamic cavitation (20-–220µm) (Franc, 2009), therefore the pits were

9

attributed to cavitation from ultrasound. It should be noted the pits were only observed

10

on the wheat straw from the 0.5 – 1 mm fraction treatments and at 376 kHz, the

11

evidence was difficult to find compared to the numerous pits observed at 40 and 995

12

kHz. Lignin is the strengthening component of the cell wall (Fengel & Wegener, 1984),

13

and disturbance of the biomass structure would increase solubilization. In addition,

14

although the power is in the same order as the higher frequencies, the slight increase in

15

power and corresponding final temperature at 40 kHz was also likely to contribute to

16

the lignin degradation. The similarity of the structures of the isolated milled wood lignin

17

supports the mechanoacoustic solubilization pathway. Likewise, in an ultrasonic

18

environment where radical production was promoted with low frequency ultrasound,

19

lignin degradation within the lignocellulosic biomass was not promoted (Ninomiya et

20

al., 2013). In the work presented here the mechanoacoustic effects were the dominant

21

means of delignification. Mechanoacoustic effects decrease in severity with an increase

22

in frequency, hence delignification; even with stirring was not favored at high

23

frequencies. Previous reports of the physical effects of ultrasound demonstrated

24

pretreatment improved the accessibility lignocellulose, increased pore size and

25

improved surface erosion (Sul'man et al., 2011; Velmurugan & Muthukumar, 2011; 11

1

Yachmenev et al., 2009; Yunus et al., 2010). The delignification of lignocellulose was

2

therefore attributed to the evident physical effects of ultrasound.

3

In summary, at both 995 and 40 kHz fractionation was improved. At 995 kHz the

4

solubilization of carbohydrates was increased, attributed to enhanced radical attack from

5

increased sonochemical effects. Then, at 40 kHz, delignification was enhanced,

6

attributed to the enhanced accessibility from the physical effects of ultrasound such as

7

pits and cracking.

8 9 10

3.2 Stirring effects 3.2.1

Mechanical effects of stirring on fractionation and delignification

Overhead stirring was used to mimic the introduction of fluid flow perpendicular to

11

the direction of ultrasonic wave propagation. In the high frequency reactor very little

12

biomass was moved with ultrasound, without stirring, unlike the stirred experiment.

13

Additionally, at 40 kHz, the biomass seemed to form a stagnant layer at the top in the

14

still experiments which was moved in the stirred experiments. Stirring was necessary to

15

expose the bulk of the biomass to ultrasound as reflected in the solid residue yields. For

16

example, at SLR 1/20 (g/ml) solid residue yields, with stirring were increased by 5.8

17

and 7.6% at 40 and 995 kHz, compared to the control. Also, at SLR 1/50 (g/ml) solid

18

residue yields were increased, with stirring by 5.4% for both 40 and 376 kHz, compared

19

to the control (Figure 1A). Furthermore, ultrasonic treatments for the 1 – 2 mm fraction,

20

decreased the solid residue yield compared to the control, by 6.8, 3.0 and 6.5% for 40,

21

376 and 995 kHz respectively (Figure 1B). Improved fractionation, (lower solid residue

22

yield) compared to the control was only achieved with ultrasound and stirring, rather

23

than ultrasound alone.

24

The lignin content after 40 kHz treatment was reduced by 3.2, 7.2 and 8.1% from

25

stirring for SLRs 1/15, 1/20 and 1/50 (g/ml), respectively (Figure 2A). Stirring at SLR 12

1

1/50 (g/ml) at 376 kHz also reduced the lignin content in the solid residue by 2.1%. The

2

only differences with cellulose and hemicellulose with stirring were increases of the

3

hemicellulose and cellulose content by 8.9 and 7.8%, respectively at 995 kHz, SLR 1/20

4

(g/ml). Therefore stirring tended to improve delignification, but not carbohydrate

5

solubilization.

6

3.2.2

7

Lignin condensation and stirring effects

Increases in the lignin content were observed at SLR 1/15 (g/ml) with 40 kHz.

8

Additionally, across all of the control experiments there was an increase or no notable

9

difference in the lignin composition compared to the raw biomass. Yet no significant

10

differences were observed in hemicellulose and cellulose content (Figure 2). This

11

implies condensation of lignin may have occurred. Lignin condensation was found to

12

occur during non-ultrasonic (Hendriks & Zeeman, 2009) and ultrasonic pretreatments of

13

lignocellulose in previous studies (Garcia et al., 2011; Sul'man et al., 2011; Sun &

14

Tomkinson, 2002).

15

The removal of lignin under the influence of low and high frequency ultrasound was

16

assisted or not affected by stirring. Furthermore, at SLR 1/20 (g/ml) after treatment at

17

995 kHz the percentage of cellulose and hemicellulose remaining in the solid fraction

18

was higher with stirring which indicated the removal of another biomass component, in

19

this case, of lignin. The assistance of lignin removal may be due to radical production

20

and/or a prevention of lignin condensation.

21

The radically induced homolytic cleavage of lignin occurs in solution with

22

ultrasound (Yoshioka et al., 2000). Furthermore, evidence of cleavage of inter-unitary

23

lignin bonds was found after sequential treatments which included an ultrasonic

24

treatment (Sun & Tomkinson, 2002). Stirring increased the radical production at low

25

frequencies, in this reactor, and in other reactors (Hatanaka et al., 2006). The fact 13

1

ultrasound was more efficient in lignin reduction with stirring at 40 kHz, as compared to

2

the control, suggests the increase of radicals may increase lignin reduction. However at

3

the two higher frequencies, 376 and 995 kHz, more radicals were produced. Therefore,

4

if the increase in production of oxidizing radicals from ultrasound caused an increase in

5

delignification, then more evidence of delignification should have been observed at

6

these two frequencies for still and stirred experiments. Hence it is likely other

7

mechanisms were involved in the differences between lignin percentage with and

8

without stirring.

9

Treatment at 40 kHz, without stirring, did not have an effect on the lignin content,

10

even though low frequency ultrasound was shown to enhance delignification (Baxi &

11

Pandit, 2012; Garcia et al., 2011; Sasmal et al., 2012; Sul'man et al., 2011) This

12

supports the theory of lignin condensation and precipitation on the biomass in this case.

13

In comparison to the still experiments, the stirred experiments resulted in a lower

14

percentage of remaining lignin and hence it was likely the influence of stirring on

15

ultrasound prevented lignin condensation. Lignin condensation during ultrasound may

16

be promoted at the bubble interface via radical combination reactions (Bussemaker &

17

Zhang, 2013). Furthermore, fluid flow was hypothesized to decrease spherical

18

cavitation, and hence fluid flow would reduce the collection of species at the interface

19

of spherical bubbles which would then reduce the occurrence of condensation reactions.

20

Therefore, the results indicated the stirred experiments were more effective at

21

delignification, most likely to be due to a prevention of lignin condensation from

22

increased fluid flow.

23 24

3.3 Particle Size 3.3.1

Particle size and increased carbohydrate solubilization

14

1

The hemicellulose composition of the solid residue of the < 0.5 mm fraction was

2

13% less than the 1 mm fraction in the control experiments (Figure 2B). At 40 kHz,

3

treatment of the 1 – 2 mm fraction resulted in 20 and 17% more hemicellulose than the

4

0.5 – 1, and < 0.5 mm fractions, respectively. Furthermore at 995 kHz there was a

5

decrease in percentage composition of hemicellulose with a decrease in particle size;

6

28.6, 26.8 and 22.1% hemicellulose content in the solid residue for the 1 – 2, 0.5 – 1,

7

and < 0.5 mm fractions, respectively. Similarly, at 376 kHz there was 11% more

8

hemicellulose in the solid residue of the 0.5 – 1 mm fraction than the < 0.5 mm fraction.

9

Therefore a decrease in particle size tended to result in an increase in hemicellulose

10

solubilization.

11

The pH was decreased with particle size for the control experiment; 5.61, 5.58 and

12

5.54 for the 1 – 2, 0.5 – 1, and < 0.5 mm fractions, respectively (Figure 4). Similarly,

13

the recorded pH for the 40 kHz treatments also decreased with size; 5.57, 5.54 and 5.47

14

for the 1 – 2, 0.5 – 1, and < 0.5 mm fractions, respectively. The dissolution of

15

hemicellulose, coupled with the decrease in pH indicated acid moieties were produced

16

from hemicellulose. The acidic degradation products may have been caused by

17

oxidation of reducing end-groups of hemicellulose or the hydrolysis of the

18

hemicellulose to acids (Hon & Shiraishi, 2001). The trend was present in the control

19

and the ultrasonic treatments which implied that the increase in acidity comes from the

20

hydrolysis of hemicellulose rather than ultrasonic effects.

21

For the higher frequencies, 995 and 376 kHz, the cellulose in the < 0.5 mm fraction

22

was 3.3 and 11% less compared to the cellulose in the 1 – 2 mm fraction (Figure 2B).

23

Furthermore, treatment at 995 kHz to the < 0.5 mm fraction, reduced the cellulose

24

content by 12% compared to the 0.5 – 1 mm fraction. The decrease in cellulose with

25

particle size supports a trend in increased carbohydrate solubilization with a decrease in 15

1

particle size. Thus, the results indicated a reduction in particle size will promote the

2

hydrolysis and solubilization of hemicellulose and cellulose.

3

3.3.2

Particle size and fractionation and delignification

4

The solid residue yield, for the 1 mm fraction was reduced by 4.9 and 9.9% relative

5

to the < 0.5 and 1 – 2 mm fractions, respectively in the control experiment (Figure 2B).

6

Moreover, after treatment with ultrasound, the solid yield for the 0.5 – 1 mm fraction

7

was reduced by 4.0, 8.2 and 4.3% compared to the 1 – 2 mm fraction for 40, 376 and

8

995 kHz, respectively. However no difference was seen in comparison of the 0.5 – 1

9

mm fraction to the < 0.5 mm fraction, (59-63%) in ultrasonic treatments. Furthermore, a

10

reduction of solid residue yields was observed with ultrasound in the particle size range

11

1 – 2 mm, compared to the control. As the particle size decreased, the surface area of

12

the biomass available increased and the mass transfer environment was improved.

13

Therefore at lower particle size ranges the ability for solubilization of components was

14

enhanced. However, the control results suggest either the stirring was not as effective as

15

ultrasound at enhancing solubilization, and/or condensation of dissolved biomass

16

components was occurring at the < 0.5 mm fraction. All in all, there was no benefit

17

from additional sieving to < 0.5 mm compared to 0.5 – 1 mm to improve fractionation.

18

Ultrasonic treatment for the 0.5 – 1 mm fraction, enhanced delignification at all

19

frequencies, compared to the control. Yet, 40 kHz was previously determined to be the

20

optimal frequency for delignification. In the control experiment, more hemicellulose

21

was dissolved with a decrease in particle size. Hence for the larger particle sizes in the 1

22

– 2 mm fraction, the degradation of hemicellulose was not as efficient. Similarly,

23

delignification was not effective for the 1 – 2 mm fraction, noted at 40 kHz, where

24

lignin degradation was expected. Then for the < 0.5 mm fraction, lower pHs of the

25

treatment liquor at 40 kHz and 995 kHz indicated a similar loss of hemicellulose. Hence 16

1

no difference in lignin percentage at 40 kHz was observed for the < 0.5 mm fraction due

2

to a similar loss of the carbohydrate portion of the biomass. Yet, in alkaline conditions

3

the optimization of delignification found the smallest size, 0.27 mm was more effective

4

for delignification and sugar production (Velmurugan & Muthukumar, 2012). The use

5

of alkali promoted delignification, hence the increased accessibility of the alkaline

6

solution at the smaller particle size could explain the different trends observed.

7

Therefore, with regards to delignification and fractionation, the results indicated that

8

the 0.5 – 1 mm fraction was better, or as good as the < 0.5 mm fraction. However, the

9

carbohydrate solubilization benefited from a reduction in particle size, attributed to an

10

increase in surface area available for radical attack.

11

3.4 The solid loading effect

12

3.4.1

Loading and increased fractionation, delignification and carbohydrate

13

solubilization

14

In general the solid residue yield decreased with a decrease in loading from 1/20 to

15

1/50 (g/ml) (Figure 1A). For example, in the control experiment the solid residue yield

16

was reduced by 5.8% after treatment at 1/50 (g/ml) compared to 1/20 (g/ml). A decrease

17

in solid residue yield in comparison between SLR 1/15 and 1/20 (g/ml) was only

18

observed at 40 kHz, with a 6.5% reduction, attributed to ultrasonic effects. Therefore

19

with regards to fractionation, the most effective loading was 1/50 (g/ml) with and

20

without ultrasound.

21

A 7.0 and 4.7% decrease in the remaining percentage of lignin for 1/20 and 1/50

22

(g/ml) was observed compared to SLR 1/15 (g/ml), with stirring at 40 kHz (Figure 2A).

23

Without stirring, at 40 kHz, the lignin content was decreased by 3.0% after treatment of

24

the 1/20 (g/ml) loading compared to the 1/15 (g/ml) loading. The efficacy of

25

delignification at 1/20 (g/ml) agrees with other ultrasonic effects on the delignification 17

1

of three biomass types, bon bogori trunk, moj trunk and acacia nut husks (Sasmal et al.,

2

2012). Hence from the results of this investigation, delignification was best at either

3

SLR 1/20 or 1/50 (g/ml).

4

For the control pretreatments, the hemicellulose remaining after treatment at SLR

5

1/20 (g/ml) was 7.0% less than the hemicellulose remaining after treatment at SLR 1/50

6

(g/ml) (Figure 2A). Similarly at 40 kHz, without stirring the remaining hemicellulose in

7

the solid fraction after treatment at SLR 1/20 (g/ml) was 12% lower than the

8

hemicellulose remaining after treatment at SLR 1/50 (g/ml). Moreover, the control

9

treatment at SLR 1/20 (g/ml) resulted in a 3.5% reduction in cellulose compared to SLR

10

1/50 (g/ml). This would indicate for the control and at 40 kHz more carbohydrates were

11

released at SLR 1/20 compared to SLR 1/50 (g/ml). However, the total glucose released

12

comparative to the initial biomass loading was favored at SLR 1/50 (g/ml) for all three

13

frequencies and the control experiment, which did not support this theory. Hence the

14

increase in hemicellulose and cellulose observed at 1/50 (g/ml) was likely due to a

15

decrease in lignin, previously discussed. Furthermore, at 995 kHz, where carbohydrate

16

solubilization was favored, a 7.9 and 9.9 % decrease in remaining cellulose and

17

hemicellulose, respectively was observed with a decrease in SLR from 1/15 to 1/20

18

(g/ml). In addition, at 376 kHz, the remaining cellulose was reduced by 3.3% going

19

from SLR 1/15 to 1/20 (g/ml). At the lower loading, although carbohydrate

20

solubilization would be increased, lignin solubilization is also increased, preventing the

21

observation of notable differences. These findings are supported by the work done on

22

the hydrolysis of sugarcane bagasse at a variety of biomass loadings which found 1/20

23

g/ml was the optimal loading for hydrolysis (Velmurugan & Muthukumar, 2011).

24

Furthermore at lower loadings, less biomass was processed in the same time frame.

25

Hence from the results presented here and from previous literature the recommended 18

1

loading for the purpose of solubilization and utilization of cellulose and hemicellulose

2

would be 1/20 (g/ml).

3

3.4.2

Biomass loading and ultrasonic propagation

4

The visual observations indicated there was a lack of movement of biomass when

5

ultrasound was used, without stirring. The biomass could either absorb or reflect the

6

ultrasound in solution, causing attenuation and a reduction of chemical or physical

7

ultrasonic effects which produced fractionation, delignification and carbohydrate

8

solubilization. The delignification and carbohydrate solubilization were hindered at 1/15

9

(g/ml) which indicated that at there was too much wheat straw in the reactor to allow for

10

ultrasound to be effective at producing physical or chemical effects. Then at 1/20 (g/ml)

11

the carbohydrate solubilization and delignification were facilitated, indicating that

12

ultrasonic-induced cavitation mechanisms were present in the reactor. When the loading

13

was further decreased to 1/50 (g/ml) the compositional differences from the alternate

14

ultrasonic effects were not as apparent, reducing the potential selectivity of the

15

parameters.

16

3.5 Degradation and depolymerization of carbohydrates in solution

17

The relative percentage of free glucose in solution was less than half at SLR 1/20

18

(g/ml) than the treatment at SLR 1/15 (g/ml) in the control experiment produced (Figure

19

3B). Then, for treatments with 40 kHz of ultrasound, without stirring, the free glucose

20

in solution increased with a decrease in SLR; 2.5, 4.1 and 5.5% for SLRs 1/15, 1/20 and

21

1/50 g/ml respectively. Similarly at 995 kHz, with stirring, the free glucose yield in the

22

liquor for SLR 1/20 (g/ml) was less than for SLR 1/50 (g/ml); 3.9 compared to 5.1%.

23

Therefore free glucose released into the liquor tended to decrease with solid loading for

24

the control experiments but tended to increase with loading for the ultrasonic treatments.

19

1

The relative percentage for total glucose was higher at SLR of 1/50 (g/ml) (6.6%),

2

compared to 1/20 (g/ml), (4.3%) for the control experiments (Figure 3B). Similarly, at

3

40 kHz without stirring, the total glucose percentage was doubled at SLR 1/50 (g/ml),

4

compared to at SLR 1/15 (g/ml), (7.3 compared to 3.6%). At 376 kHz, with stirring the

5

percentage of total glucose at SLR 1/50 (g/ml), was higher than at to SLR 1/20 (g/ml);

6

6.9% compared to 4.7% respectively. Furthermore, at 995 kHz, the treatment with

7

stirring at SLR 1/50 (g/ml) released more total glucose (9.0%) than both of the other

8

loadings; 5.4 and 5.6% for 1/20 and 1/15 (g/ml) respectively. Hence for the free glucose,

9

the ultrasonic trends were opposite to the control but for the total glucose, the control

10

and ultrasonic treatments increased with a decrease in biomass loading.

11

A comparison of the ratio of free glucose to total glucose is presented in Table 2.

12

Higher free to total glucose ratios demonstrated more glucose was in solution in the

13

monomer form after pretreatment. However when the free to total glucose ratio was

14

decreased, the hydrolysis produced more glucose monomers from polymers in the

15

pretreatment solution. The ratios of free to total glucose for the control experiments at

16

SLR 1/20 and 1/50 (g/ml) (both 0.4) are less than all of the ultrasonic experiments at

17

these two SLRs (ranging from 0.5 to 0.8). Furthermore after treatment of the 1 – 2 mm

18

fraction, the control treatment produced a lower ratio (0.6) than the comparative

19

ultrasonic treatments (0.7–1.4). However at the higher SLR and lower particle sizes the

20

control treatments did not show any trend in difference from the ultrasonic treatments.

21

Therefore, under those conditions there were no ultrasonic effects on depolymerization.

22

The depolymerization without ultrasound was likely to be assisted by increased

23

acidity from increases in acetic acid solubilized with the solubilization of hemicellulose,

24

as indicated by the lower pHs of the pretreatment liquors of SLRs 1/15 (g/ml) and the

25

0.5 – 1 and < 0.5 mm fractions (Figure 4). Therefore at lower SLRs and for the 1 – 2 20

1

mm particle size fraction, the results suggest depolymerization in solution was enhanced

2

by ultrasound.

3

The degradation of carbohydrates in solution was evidenced by the higher percentage

4

of total glucose in solution from treatment at 376 kHz, compared to 40 and 995 kHz of

5

the 0.5 – 1 mm fraction; 6.5% compared to 3.9 and 4.2% for 376, 40 and 995 kHz,

6

respectively (Figure 3B). The reason for the increased total glucose at 376 kHz,

7

compared to treatment at 995 kHz was likely to be due to enhanced degradation

8

reactions of glucose units with the higher radical production. Additionally, the fact there

9

was a lower percentage of total glucose after treatment at 376 kHz for the < 0.5 mm

10

fraction compared to the 0.5 – 1 mm fraction, suggests the glucose, which was more

11

easily solubilized at lower particle sizes, was degraded in solution. Polymers have been

12

found to degrade through the mechanoacoustic and sonochemical effects of ultrasound

13

from a cumulative effect of the hydroxyl radicals, shear forces and degradation of

14

hydrophobic polymers in the hot region around the collapsing bubbles (Henglein et al.,

15

1989; Kardos & Luche, 2001).

16

Solubilization of carbohydrates was favored at higher ultrasonic frequencies, and the

17

sonochemical effects under these conditions produced competing degradation and

18

depolymerization reactions within the pretreatment liquor.

19 20

3.6 Outcomes A majority of studies use low frequency ultrasound to treat biomass, however, here

21

high frequency ultrasound was also effective. Delignification was enhanced from the

22

mechanoacoustic effects of ultrasound at low frequencies, more so than the radicals

23

produced. Then, the treatment of the carbohydrates within the biomass was affected by

24

the ultrasonic radicals at high frequencies, more so than the lignin.

21

1

The ability for stirring to reduce lignin condensation is significant for the treatment

2

of polymers with ultrasound. Lignin condensation can inhibit biomass accessibility and

3

is counterproductive when subsequent processing of delignified material is desired.

4

Hence if condensation is reduced with stirring, while retaining the benefits of ultrasonic

5

processing, the hydrolysis of the remaining carbohydrates can be improved.

6

From this investigation, fractionation and delignification did not benefit from a

7

particle sizes below 0.5 mm. This was attributed to the mechanical effects of ultrasound

8

having more of an impact on larger particle size ranges. Nevertheless, the radically

9

enhanced solubilization of carbohydrates was enhanced by smaller particle size ranges.

10

Therefore if mechanical effects are required, then the benefit of downsizing particles is

11

reduced compared to chemical treatment. This has the potential to reduce the overall

12

cost of the implementation of ultrasound for biomass processing.

13

The biomass loading was generally found to be optimal at 1/20 (g/ml), although in

14

some cases the optimal loading was also 1/50 (g/ml). Processing of higher loadings in

15

the same time frame can improve the desired yields at the same process cost. Therefore

16

it is recommended that ultrasonic processing of biomass should consider 1/20 (g/ml)

17

rather than lower loadings.

18

In summary, based on the experimental results presented here the recommended

19

conditions for delignification are 40 kHz, particle size 0.5 - 1 mm, 1/20 (g/ml) with

20

stirring. Then for carbohydrate solubilization the most suitable conditions were found to

21

be 995 kHz, particle size < 0.5 mm, at loading of 1/20 (g/ml), without stirring.

22

4. CONCLUSIONS 

23

High (995 kHz) and low frequency (40 kHz) treatment enhanced fractionation with

24

stirring. At 40 kHz, delignification was enhanced by up to 7.2 % from the mechanical

25

effect of ultrasound, which was augmented by stirring, due to a reduction of lignin 22

1

condensation. Concomitantly, at 995 kHz the carbohydrate solubilization was increased

2

by up to 9.1% attributed to the sonochemical effect of ultrasound.

3

Carbohydrate solubilization increased with a decrease in particle size. However

4

fractionation and delignification only benefited from particle size below 1 mm. Lastly,

5

the preferred treatment loading of wheat straw was 1/20 (g/ml) for fractionation,

6

delignification and carbohydrate solubilization.

7 8 9

For IR spectra of isolated lignin and SEM images of treated and untreated wheat straw, refer to supplementary material, available online.

10

ACKNOWLEDGEMENT 

11

The authors gratefully acknowledge the financial and other support provided by

12

the Australian Research Council under the ARC Linkage Projects Scheme (ARC

13

Linkage Project LP100200135). The authors acknowledge the facilities, and the

14

scientific and technical assistance of the Australian Microscopy & Microanalysis

15

Research Facility at the Centre for Microscopy, Characterisation & Analysis, The

16

University of Western Australia, a facility funded by the University, State and

17

Commonwealth Governments. The authors acknowledge Mark Thomas of Treth Farm

18

for the provision of the wheat straw. Bussemaker would also like to appreciatively

19

acknowledge the Julian Hunka scholarship provided through The University of Western

20

Australia for provision of living expenses.

21 22 23 24 25 26 27

REFERENCES  Amidon, T.E., Wood, C.D., Shupe, A.M., Wang, Y., Graves, M., Liu, S.J. 2008. Biorefinery: Conversion of woody biomass to chemicals, energy and materials. J. Biobased Mater. Bioenergy, 2, 100-120. Baxi, P.B., Pandit, A.B. 2012. Using cavitation for delignification of wood. Bioresour. Technol., 110, 697-700. Bjorkman, A. 1956. Studies on finely divided wook. Part 1. Extraction of lignin with neutral solvents. Svensk Papperstidning, 59, 477-485.

23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

Bussemaker, M.J., Zhang, D. 2013. Effect of ultrasound on lignocellulosic biomass as a pretreatment for biorefinery and biofuel applications. Ind. Eng. Chem. Res., 52, 3563-3580. Dular, M., Delogosha, O.C., Petkovsek, M. 2013. Observations of cavitation erosion pit formation. Ultrason. Sonochem., 20, 1113-1120. Entezari, M.H., Kruus, P. 1994. Effect of frequency on sonochemical reactions. I: Oxidation of iodide. Ultrason. Sonochem., 1, S75-S79. Esfahani, M.R., Azin, M. 2012. Pretreatment of sugarcane bagasse by ultrasound and dilute acid. AsiaPac. J. Chem. Eng., 7, 274-278. Fengel, D., Wegener, G. 1984. Wood: chemistry, ultrastructure, reactions. Walter de Gruyter, Berlin. Franc, J.-P. 2009. Incubation time and cavitation erosion of work-hardening materials. J. Fluids Eng., 131, 021303-1-14. Garcia, A., Gonzalez Alriols, M., Llano-Ponte, R., Labidi, J. 2011. Ultrasound-assisted fractionation of the lignocellulosic material. Bioresour. Technol., 102, 6326-6330. Hatanaka, S.-i., Mitome, H., Yasui, K., Hayashi, S. 2006. Multibubble sonoluminescence enhancement by fluid flow. Ultrasonics, 44, e435-e438. Hendriks, A.T.W.M., Zeeman, G. 2009. Pretreatments to enhance digestibility of lignocellulosic biomass. Bioresour. Technol. , 100, 10-18. Henglein, A., Ulrich, R., Lilie, J. 1989. Luminescnece and chemical action by pulsed ultrasound. J. Am. Chem. Soc., 111, 1974-1979. Hon, D.N.-S., Shiraishi, N. 2001. Wood and Cellulosic Chemistry. Marcel Dekker, Basel. Kanthale, P., Ashokkumar, M., Grieser, F. 2008. Sonoluminescence, sonochemistry (H2O2 yield) and bubble dynamics: Frequency and power effects. Ultrason. Sonochem., 15, 143-150. Kardos, N., Luche, J.L. 2001. Sonochemistry of carbohydrate compounds. Carbohydr. Res., 332, 115-131. Koda, S., Kimura, T., Kondo, T., Mitome, H. 2003. A standard method to calibrate sonochemical efficiency of an individual reaction system. Ultrason. Sonochem., 10, 149-156. Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P. 2009. Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res., 48, 3713-3729. Kunaver, M., Jasiukaityte, E., Cuk, N. 2012. Ultrasonically assited liquefaction of lignocellulosic materials. Bioresour. Technol., 103, 360-366. Mason, T.J., Cobley, A.J., Graves, J.E., Morgan, D. 2011. New evidence for the inverse dependence of mechanical and chemical effects on the frequency of ultrasound. Ultrason. Sonochem., 18, 226230. Ninomiya, K., Takamatsu, H., Ohnishi, A., Takahashi, K., Shimizu, N. 2013. Sonocatalytic-Fenton reaction for enhanced OH radical generation and its application to lignin degradation. Ultrason. Sonochem., 20, 1092-1097. Sasmal, S., Goud, V.V., Mohanty, K. 2012. Ultrasound assisted lime pretreatment of lignocellulosic biomass toward bioethanol production. Energ. Fuel., 26, 3777-3784. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, A., Templeton, D. 2008a. Determination of sugars, byproducts, and degradation products in liquid fraction process samples. National Renewable Energy Laboratory. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D. 2008b. Determination of structural carbohydrates and lignin in biomass. National renewable energy laboratory. Sul'man, E.M., Sul'man, M.G., Prutenskaya, E.A. 2011. Effect of ultrasonic pretreatment on the composition of lignocellulosic material in biotechnologiacal processes. Biocatalysis, 3, 28-33. Sun, R., Tomkinson, J. 2002. Comparative study of lignins isolated by alkali and ultrasound-assisted alkali extractions from wheat straw. Ultrason. Sonochem., 9, 85-93. Sun, R.C., Sun, X.F., Ma, X., H. 2002. Effect of ultrasound on the structural and physicochemical properties of organosolve soluble hemicelluloses from wheat straw. Ultrason. Sonochem., 9, 95101. Sun, R.C., Tomkinson, J., Wang, S.Q., Zhu, W. 2000. Characterization of lignins from wheat straw by alkaline peroxide treatment. Polym. Degrad. Stab., 67, 101-109. Sun, X.F., Sun, R.C., Fowler, P., Baird, M.S. 2005. Extraction and characterization of original lignin and hemicelluloses from wheat straw. J. Agric. Food Chem., 53, 860-870. Velmurugan, R., Muthukumar, K. 2012. Ultrasound-assisted alkaline pretreatment of sugarcane bagasse for fermentable sugar production: Optimization through response surface methodology. Bioresour. Technol., 112, 293-299. Velmurugan, R., Muthukumar, K. 2011. Utilization of sugarcane bagasse for bioethanol produciton: Sono-assisted acid hydrolysis approach. Bioresour. Technol., 102, 7119-7123.

24

1 2 3 4 5 6 7 8 9

Yachmenev, V., Condon, B., Klasson, T., Lambert, A. 2009. Acceleration of the enzymatic hydrolysis of corn stover and sugar cane bagasse celluloses by low intensity uniform utrasound. J. Biobased Mater. Bioenergy, 3, 25-31. Yoshioka, A., Seino, T., Tabata, M., Takai, M. 2000. Homolytic scission of interunitary bonds in lignin induced by ultrasonic irradiation of MWL dissolved in dimethylsulfoxide. Holzforschung, 54, 357-364. Yunus, R., Salleh, S.F., Abdullah, N., Biak, D.R.A. 2010. Effect of ultrasonic pre-treatement on low temperature acid hydrolysis of oil palm empty fruit bunch. Bioresour. Technol., 101, 9792-9796.

25

A

B Figure 1

26

Solid residue yields. A) Various biomass loadings; the 1/15, 1/20 and 1/50 represent the solid to liquid ratios in g/ml, and B) various particle size ranges. The 2,1 and 0.5 correspond to wheat straw screened between 1-2, 0.5-1 and 0-0.5 mm, respectively.

A

B Figure 2

27

Composition of the solid residues. A) Various biomass loadings; the 1/15, 1/20 and 1/50 represent the solid to liquid ratios in g/ml, and B) various particle size ranges. The 2,1 and 0.5 correspond to wheat straw screened between 1-2, 0.5-1 and 0-0.5 mm, respectively.

A

B Figure 3

28

Free and total glucose in the pretreatment liquor. A) Various biomass loadings; the 1/15, 1/20 and 1/50 represent the solid to liquid ratios in g/ml, and B) various particle size ranges. The 2,1 and 0.5 correspond to wheat straw screened between 1-2, 0.5-1 and 0-0.5 mm, respectively.

A

B Figure 4

29

Liquor pHs. A) Various biomass loadings; the 1/15, 1/20 and 1/50 represent the solid to liquid ratios in g/ml, and B) various particle size ranges. The 2,1 and 0.5 correspond to wheat straw screened between 1-2, 0.5-1 and 0-0.5 mm, respectively.

Table 1

FTIR assignments of the fingerprint region of the milled wood lignin isolated from the solid residues. Absorption Assignment (cm-1) Acetylation of HC – acetyl ester band C=O 1732 Carbonyl stretching not in conjugation with the aromatic ring 1717 Unconjugated ketone and carboxyl stretching 1703 Carbonyl stretching in conjugation with the aromatic ring 1686 Conjugated carbonyl 1652 Absorbed water 1617 Aromatic skeletal vibrations (C-H deformations) 1594 Absorbed water 1560 Absorbed water 1542 Aromatic skeletal vibrations 1508 Asymmetric C-H deformations and aromatic ring vibrations 1457 Aromatic skeletal vibrations 1423 Symmetric C-H bending 1360 Syringyl ring breathing with C-O stretching 1340 Guaiacyl ring breathing with C-O stretching 1262 Aromatic ring breathing with C-O and C=O 1232 p-coumaric ester 1166 Aromatic CH in plain deformation for syringal types 1125 C-O deformation, secondary alcohol and aliphatic esters 1079 Glycosidic linkage v(C-O-C) and C-O and C-C stretching of 1048 attached hemicellulose Aromatic CH in plain deformation for guaiacyl types 1030 Arabinosyl side chain of the attached hemicellulose 977 C-H out of plane in aromatic rings 916 Aromatic C-H out of bending 839

30

Table 2 Frequency (kHz) Control sta 40 40 st a 376 376 st a 995 995 st a

Ratios of free to total glucose for different experimental settings. . SLR Ratio of free to Fraction Ratio of free to (g/ml) total glucose (mm) total glucose 1/15 0.8 < 0.5 0.9 1/20 0.4 0.9 0.5 – 1 1/50 0.4 0.6 1 – 2 1/15 0.7 1/20 0.8 1/50 0.8 1/15 0.6 < 0.5 0.8 1/20 0.6 0.8 0.5 – 1 1/50 0.5 1.4 1 – 2 1/15 0.6 1/20 0.7 1/50 0.6 1/15 0.7 < 0.5 1.1 1/20 0.8 0.7 0.5 – 1 1/50 0.6 0.8 1 – 2 1/15 0.2 1/20 0.6 1/50 0.6 1/15 0.8 < 0.5 0.8 1/20 0.7 1 0.5 – 1 1/50 0.6 0.7 1 – 2 a ST denotes stirred

31

Highlights A comparison of ultrasonic pretreatment at various frequencies was presented Low frequency favored delignification attributed to mechanoacoustic effects High frequency favored carbohydrate solubilization Lignin condensation was reduced with stirring The particle size and loading influenced the efficacy of ultrasonic pretreatment

32