Countercurrent Extraction of Soluble Sugars from Almond Hulls and

Article pubs.acs.org/JAFC

Countercurrent Extraction of Soluble Sugars from Almond Hulls and Assessment of the Bioenergy Potential Kevin M. Holtman,* Richard D. Offeman, Diana Franqui-Villanueva, Andre K. Bayati, and William J. Orts PWA, WRRC, BCE, Agricultural Research Service, U.S. Department of Agriculture, 800 Buchanan Street, Albany, California 94710, United States ABSTRACT: Almond hulls contain considerable proportions (37% by dry weight) of water-soluble, fermentable sugars (sucrose, glucose, and fructose), which can be extracted for industrial purposes. The maximum optimal solids loading was determined to be 20% for sugar extraction, and the addition of 0.5% (w/v) pectinase aided in maintaining a sufficient free water volume for sugar recovery. A laboratory countercurrent extraction experiment utilizing a 1 h steep followed by three extraction (wash) stages produced a high-concentration (131 g/L fermentable sugar) syrup. Overall, sugar recovery efficiency was 88%. The inner stage washing efficiencies were compatible with solution equilibrium calculations, indicating that efficiency was high. The concentrated sugar syrup was fermented to ethanol at high efficiency (86% conversion), and ethanol concentrations in the broth were 7.4% (v/v). Thin stillage contained 233 g SCOD/L, which was converted to biomethane at an efficiency of 90% with a biomethane potential of 297 mL/g SCODdestroyed. Overall, results suggested that a minima of 49 gal (185 L) ethanol and 75 m3 methane/t hulls (dry whole hull basis) are achievable. KEYWORDS: almond hulls, carbohydrates, fermentable sugars, countercurrent extraction, fermentation, ethanol, biogas

■

INTRODUCTION

the biomethane potential of the thin stillage from ethanol production using almond hulls. There exists the possibility that almond hulls can support a true, niche biofuel market unique to California or be utilized in combination with an energy crop such as sweet sorghum to support a year-round biofuels facility. There also exists the potential to produce a concentrated sugar solution that can be used for nutraceutical or biochemical applications, although energy aspects will be specifically addressed here. The goal here is to provide extraction data that gives rough estimates for designing facilities that produce high sugar concentrate syrups from almond hulls.

The U.S. state of California produces ∼80% of the world’s almonds, resulting in 1.4 million tonnes annually of almond hulls as byproduct.1 Almond hulls therefore represent a large volume of agricultural residue exclusive to the Central and Sacramento Valleys of California and could support local bioenergy production. Almond hulls are known to be rich in water-soluble, free fermentable sugars and other nonfermentable sugars and sugar alcohols2−8 and hence could be good candidates for locally produced bioethanol. The free sugars could also be utilized for edible purposes9 and as feedstocks for the production of higher value chemical bioproducts such as butanol, 3-hydroxypropionic acid, succinic and lactic acids, butanediol, and polyesters.10 The extraction of sugar from almond hulls can be considered a leaching process that involves diffusion of solvent (water) into the substrate, allowing the substrate to swell and the soluble components to dissolve, followed by displacement of the soluble components from the insoluble matrix.11 In the sugar industry, extraction is performed in a unit known as a diffuser, which can be a batch-type Robert diffuser or a continuous diffuser (tower, slope, drum) to perform countercurrent extraction of the soluble materials.12 Two patents mention countercurrent extraction of almond hulls for sugar recovery but do not provide details of the extraction.7,13 The purpose of this paper is to provide data on countercurrent extraction of sugars from almond hulls and to assess the bioenergy potential of the soluble fraction of the almond hull, including (1) production of bioethanol from the free fermentable sugars and (2) production of biomethane from the thin stillage remaining after ethanol fermentation and distillation. Bioethanol potential has been published previously but with little detail.14 There is no reference in the literature to This article not subject to U.S. Copyright. Published 2015 by the American Chemical Society

■

MATERIALS AND METHODS

Almond Hulls. Nonpareil almond (Prunus dulcis (Miller) D.A. Webb) hulls were obtained by the Almond Hullers and Processors Association (AHPA) directly from a hullers operation in Colusa county, California, and delivered to USDA on August 21, 2011. Nonpareil hulls have been shown to contain 30% higher than reported in the previous literature (reported in Table 1 for comparison). Literature values for sucrose contents range from 5.3 to 10.6%,2,3 comparable to or higher than the current results (5.2%, Table 1). Inositol values are similar to the reported values (∼2%), whereas the sorbitol levels measured here are slightly higher (5.9 vs 2.7−4.6%). The quantitative difference between techniques is expressed in this analysis as marked increases in the contents of glucose and fructose, 16.3 and 15.9%, respectively. Reported literature values range from 5.8 to 10.4% for glucose and from 4.8 to 8.8% for fructose.2,3 Enhancement of the glucose and fructose contents could in part be a result of the decomposition of sucrose (a dimer of glucose and fructose) into its monomeric components but does not completely account for the difference. The difference may also be attributed to variety or breeding, changes in agricultural approaches, location, environmental conditions, sensitivity of analysis, higher degree of quantitation, or other factors. In total for the almond hulls analyzed here, fermentable sugar (FS) content is 37% and total sugar plus sugar alcohol content is 46%. Total extractable sugar plus sugar alcohol contents reported by the previous studies range from 25.6 to 31.5%2,3,18 (Table 1). The free fermentable sugars comprise 75% of the soluble components of the almond hulls with the remainder being soluble ash (3.4%) and unknown compounds (21.3%), comprised in part by the following categories: tannins, polyphenols, fats, proteins, pectins, gums, and other polysaccharides (Table 1).4 Characterization of the Insoluble Fraction (Spent Almond Hulls). The structural carbohydrate composition was determined by acid hydrolysis of the spent almond hulls (residual almond cake) after hot water extraction. Analysis of the residue assuming no losses during extraction indicates that the dry raw hulls contain 6.6% cellulose, 6.0% hemicellulose, and 12.3% lignin (acid-insoluble fraction). The cellulose and hemicellulose contents are slightly lower than previous results from dairy forage analysis (12.8 and 9.0%, respectively, on average), and the lignin content is slightly higher in this analysis (12.3%) versus an average of 10% reported previously.4,19−22 In total, 68% of the spent almond hulls is readily identifiable based upon cell wall analysis and ash content and, overall, 74.7% of the total almond hull content is identified by the procedures utilized here (Table 1). Practical Solids Content and Water Recovery. Commercial ethanol production requires a high sugar concentration for fermentation, because high sugar levels result in sufficient alcohol content to make distillation economical.23 The hulls were first subjected to a steeping process at 1 h and overnight time periods to determine maximum practical solids content and maximum achievable sugar recovery. The extent to which the fermentable sugars may be recovered is a function of the free water availability due to the absorptive capacity of the hulls and the equilibrium sugar concentration. Almond hulls are highly hydrophilic and can absorb and retain 10−15 times their weight in water. The initial steeping step involves wetting of the hulls and, as a result a substantial proportion of the solvent (and solute), remains with the solid phase, thus reducing extraction efficiency. Figure 2 shows the results at various solids content for the 1 h and overnight periods. As can be seen, low solids (5%) and overnight extraction result in 74% recovery of the free sugars but at very low concentration (17.6 g/L FS or

Figure 2. Effect of solids content and extraction duration on free fermentable sugar and free water recovery.

22.4 g/L TS). In all other cases, overnight sugar recovery was equal to or lower than that for the 1 h steep duration due to lower water recovery resulting from complete saturation of the almond hulls. Additionally, sugar concentrations were equivalent over the two time frames, suggesting a 1 h steep was sufficient to achieve equilibration of the free and absorbed solutes. At 20% solids content and a 1 h steep, sugar concentration (65 g/L FS) was sufficiently high to use this as a basis for a multistage extraction process. Pectinase Application. It has been reported previously that almond hulls contain pectin4,24 or pectin-like hydrocolloids. Many fruit-processing operations include the use of pectinase to control the viscosity, improve liquids/solids separation, and clarify juice products.25 Pectin is located in the primary cell wall and middle lamella and provides firmness or compression rigidity to the fruit.26 Almond hulls have a tremendous propensity to absorb and hold water, in part due to the pectin molecules, making free sugar extraction and disposal of the residuals more difficult. As a result, pectinase applications were explored to improve sugar extraction and water recovery. The enzyme cocktail used in this study contains methyl esterases and polygalacturonases as well as small amounts of hemicellulases and cellulases, all of which can contribute to the breakdown of the cell wall to facilitate fermentable sugar extraction. Almond hulls were steeped at 20% solids loading at 50 °C for 1 h with pectinase enzymes at application rates ranging from 0 to 0.75% (w/v). After a 1 h steep with no enzyme, the entire volume of water was absorbed by the control (i.e., no steep liquor available). Steep liquor recovery improves to 26% of applied with 0.25% pectinase application and then increases only slightly (34%) at the higher enzyme dosages. It is likely that the enzyme application removes the rigid, hydrophilic pectin material, decreasing the absorptivity of the hull, making more free water available, and facilitating equilibration of the free and absorbed solutes. Figure 3a illustrates how the maximum sugar extraction levels are directly tied to water recovery. Higher sugar recovery rates require additional extraction stages. Figure 3a also shows that the sugar yield after steep liquor recovery and first extraction (wash) improves 36% over the control with just 0.10% enzyme application and improves to 76% at 0.75% application. In total, at sufficient 2493

DOI: 10.1021/jf5048332 J. Agric. Food Chem. 2015, 63, 2490−2498

Journal of Agricultural and Food Chemistry

Article

Figure 4. Free sugar extraction with and without pectinase enzyme application and steeping. When applied, enzyme application was 0.5% (w/v) and steep time was 60 min at 50 °C.

no enzyme produces a more concentrated product (88 g/L FS), but very little free water is available after a single stage and the total sugar recovery is similar to the control. Only slightly more sugar is recovered by steeping without pectinase application (38% recovery vs 34%) than for no equilibration time. The two curves mirror each other in slope as the hulls become saturated and sugar extraction is equivalent because structurally the hulls remain similar. After three extraction stages, 96 and 93% of sugar was recovered, respectively, for steeping without enzyme and no steep options. An additional stage recovers 99% of the sugar in both of these cases. The pectinase steeped hulls release significant volumes of free water and fermentable sugars, resulting in a high sugar concentration in the steep liquor (74 g/L FS). Hence, much more sugar is recovered in the first stage of extraction. Figure 4 also shows that enzyme application under ideal conditions (fresh wash water) results in complete sugar recovery in three filtration steps and that 96% recovery is already achieved after only two filtrations. Enhanced sugar recovery is important in commercial processing to reduce the number of stages required for recovery of product and also to maximize the product concentration. As a result, the application of pectinase is beneficial to accelerate sugar recovery, produce higher volumes of free water at high solids processing conditions, and produce higher fermentable sugar concentrations. Countercurrent Extraction of Free Fermentable Sugars. On the basis of the above results a series of experiments were devised to mimic countercurrent extraction of soluble sugars, analogous to a diffuser, which is used in the food-processing industry for extraction of sugars. The counter flowing water is fresh at its inlet point and progressively increases in sugar concentration as it contacts sugar-rich hulls (Figure 1). The process involves (1) diffusion of the lower strength solvent (water) into the substrate, (2) dissolution and equilibration of soluble components, and (3) forced displacement of the free water and solubles by low-pressure filtration (extraction). Five rounds of extraction were performed to reproduce the concentration of sugars analogous to the countercurrent extraction process. Round 1 utilized fresh water as the wash water and produced a steep liquor, S1 (84 g/L total sugar and sugar alcohol concentration) and a battery of filtrates (S2, S3,

Figure 3. (a) Effect of pectinase application on the efficiency of sugar recovery through steep and first wash stage. Steeping was performed at 50 °C for 60 min. (b) Improvement in sugar recovery with pectinase treatments (0.5% w/v) over time.

enzyme application rates, sugar recovery after steeping and firststage washing is ∼70%. To examine the benefits of steep time, enzyme application was maintained at 0.5% (w/v) and steep time varied. Figure 3b shows sugar yields in the steep liquor increase only slightly (from 20 to 25%) over steep time varying from 0 to 60 min. The data indicate that equilibration occurs rather quickly, at least during the initial contact of the almond hulls with the solvent as milling in the blender reduces particle size. The apparent benefit of the pectinase is that it reduces the swelling capacity of the hulls and maintains a free volume during the wetting process so that recovery of free sugars can occur after saturation of the hulls is achieved. In comparison to the control (60 min with no enzyme), sugar recovery after steep liquor recovery and first-stage extraction increases 60%, and this is due to the reduction in the water-retaining capacity of the hulls. Over the time curve the enhancement of sugar recovery is less because equilibration between free and absorbed solute is quickly achieved. As a result, it may be acceptable to shorten steep time and reduce overall processing time. Figure 4 shows the effects of steeping and pectinase additions versus the control (no enzyme, no equilibration time). The control solubilizes the readily available soluble sugars, but contact time is not sufficient for diffusion of water into hulls. As such, a fair volume of water is recovered through filtration but with fairly low sugar concentration (46 g/L FS). Steeping with 2494

DOI: 10.1021/jf5048332 J. Agric. Food Chem. 2015, 63, 2490−2498

Journal of Agricultural and Food Chemistry

Article

four 1 min milling treatments and a 1 h enzyme steep. The deviation from the water recovery relationship may indicate that displacement is incomplete, the residual sugar is difficult to completely extract, or the final autoclave cycle causes some minor depolymerization of hemicelluloses. Figure 5 also depicts cumulative extraction efficiency defined as the total cumulative sugar recovery combined over the steep and three extraction stages divided by the total sugar input with the solid hulls and recycled wash water. On average, the cumulative recovery was 88% and varied between 83 and 93% over the five repetitions. The data show that although the extraction efficiencies per repetition do not decrease with increasing sugar concentrations (i.e., rounds 4 and 5), larger amounts of sugar are not extracted through three wash stages as the feedback wash loop contains larger amounts of sugar than are actually input with the hulls. Including the recycled sugars in the efficiency calculation, repetitions 4 and 5 are 83 and 88% efficient, respectively. However, using the calculation of losses divided by input with the fresh hulls would return values of 78 and 69%. The lingering sugar with the residue after three extraction stages and high sugar concentrations throughout indicates that a fourth extraction stage may be useful to achieve better sugar recovery. Five consecutive countercurrent extractions were performed to concentrate the soluble sugars in the product; however, equilibrium was not quite reached by this point. After a single round of extraction, the concentrated sugar product contained 65 g/L FS (84 g/L total sugars), and after round 5, the product contained 131 g/L FS (169 g/L total sugars) (Figure 6). The

S4) to serve as wash water in the subsequent rounds. S2 served as the dilution water for the fresh almond hulls in round 2, carrying along with it a total sugar concentration of 39 g/L total sugars (30 g/L FS). By mass balance, the equilibrium sugar content in the round 2 steep liquor (S6) was calculated as the sugar input divided by the total of the free and absorbed water and should be 115 g/L total sugars. By measurement of the filtrate (S6, Figure 1), the actual concentration was 117 g/L (91 g/L FS), indicating that the steep process was adequate for equilibration. The first wash stage consists of suspension of the almond hull cake in the second filtrate from round 1 extraction (S3, Figure 1) for the duration of milling (1 min) followed by filtration. Interestingly, the observed sugar concentration in S7 (56 g/L total sugar, 44 g/L FS) was predicted by the equilibrium sugar calculation, indicating that the contact time during milling in the blender was sufficient for equilibration. Observed values were consistently in sync with predicted values (± 10%) throughout, especially with the more concentrated filtrates. Dilute filtrates had a larger percent difference but were close in value for g/L total sugar. Throughout the entirety of the study, the free fermentable sugar content was 0.78 g per g of total sugar extracted, which demonstrates that no preferential extraction of one type of sugar or sugar alcohol is occurring. Figure 5 shows the average efficiency of sugar recovery by each individual extraction stage and the cumulative average

Figure 5. Sugar recovery efficiency for individual extraction stages within countercurrent extraction repetitions (gray bars); average cumulative sugar recovery within repetition (black bars). Figure 6. Total sugars and free fermentable sugars extracted per countercurrent repetition. Solid line represents laboratory data; dashed lines represent theoretical yields based upon modeling using measured extraction efficiencies.

across the three wash stages. Efficiency is lowest during the steep stage as the hulls are saturating, and the free water recovery yields are lower. On average, 70% of the steep water is retained within the hulls. Because the yield of extractable sugar is dependent upon free water recovery, it follows that sugar recovery is only ∼30% on average (see Figure 2a). After steeping, the hulls are saturated and the process equilibrates, whereby extraction stage efficiencies vary from 51 to 49 to 43% in the first to third stages, respectively. The efficiency is defined as the total sugar in the filtrate divided by the total sugar in the solid matrix and the wash solution. The first and second stage sugar recoveries are closely related to water recovery yields; however, the third stage strays somewhat from this relationship. Specifically, the average water recovery from the third wash stage was 60%; however, the average sugar recovery yield was only 43%. At this point, the average particle size is small after

results demonstrate that countercurrent extraction is efficient for producing a high free fermentable sugar concentration for fermentation to alcohol. In practice, the sugar concentration would increase until equilibrium is achieved through the system, and this would be the operational target. On the basis of the extraction efficiencies, the point of equilibrium was estimated as depicted by the dashed line extensions in Figure 6. Calculations indicate that two additional rounds of extraction would bring the concentrated product close to an equilibrium point of 184 g/L TS (144 g/L FS), which would mark the highest concentration that can be achieved through countercurrent extraction. 2495

DOI: 10.1021/jf5048332 J. Agric. Food Chem. 2015, 63, 2490−2498

Journal of Agricultural and Food Chemistry

Article

Figure 7. Fermentation of high sugar concentrate steep liquor to ethanol. Glucose and fructose were completely converted, and sucrose was 87 % converted. Ethanol yields were 86% of theoretical. Beer concentration averaged 7.4% (v/v).

Table 2. Data from the Anaerobic Fermentation of Almond Hull Thin Stillage to Biogas feed

SCOD added (g)

SCOD destroyed (g)

% removal eff

CH4 yield (mL/g SCODdestroyed)

SCOD effluent (mg SCOD/L)

1 2 3 4 5 6 7

5.71 4.74 4.97 4.97 4.13 4.34 4.30

5.15 4.24 4.40 4.44 3.97 3.68 3.92

90 90 88 89 96 85 91

263 280 278 291 392 276 300

1230 1253 1377 1429 1069 1312 1238

av σ SE

4.74

4.26

90 3.3 1.3

297 43 16

1273 117 44

shows the average results of fermentation of the sugar solution after a 24 h conversion period. The yields were an average of 86% of the theoretical ethanol (using four replicate experiments) and an average of 7.4% (v/v) ethanol concentration in the resulting beer. Furthermore, mass balance on the fermentable sugars shows a nearly complete utilization of both glucose and fructose and an average of 87% utilization of the sucrose in solution. Inhibitors such as furfural, HMF, and acetic acid were tracked but were not present in any appreciable concentration. Anaerobic Fermentation of Thin Stillage to Methane. The beer was subjected to overhead distillation to separate and purify the ethanol but also to create a thin stillage for further analysis. The thin stillage was analyzed for SCOD content and stored in the refrigerator for further use. A total of 60−70% of the almond hull was solubilized during the extraction process, whereas the fermentable sugars utilizable by the yeast comprise only slightly more than half of this content. As a result, the beer has an amber color and a very high SCOD content (233,250

Filter cake residues were collected and analyzed for moisture content and percent of original almond hull remaining. The average solids content of the spent almond hull cake was 9.2%, and overall the average recovery of solids was 28.5% of the original almond hull input, about 10% lower than the quantification assay. Discrepancies could be related to pectinase application or additional handling. Fermentation of Soluble Sugars to Alcohol. The countercurrent extraction experiment demonstrates that the sugar concentration in the steep liquor and wash filtrates can be progressively enhanced providing data for estimation of the maximum producible sugar concentrations. The countercurrent experiment, however, utilized small product volumes, with insufficient quantity to carry out a significant, reproducible fermentation experiment. Accordingly, a large extraction starting with 1 kg of almond hulls was undertaken to ultimately produce a large volume of high sugar product for fermentation experiments. Sugar content was monitored until the total sugar content was 178 g/L (140 g/L fermentable sugars). Figure 7 2496

DOI: 10.1021/jf5048332 J. Agric. Food Chem. 2015, 63, 2490−2498

Journal of Agricultural and Food Chemistry

Article

from 65 g/L fermentable sugars after one round of extraction to 131 g/L after five rounds of extraction as a result of the applied countercurrent approach. Extraction efficiency was high, 88% overall; however, the number of extraction stages within each round of extraction would ideally be extended one or two wash stages as the sugars concentrate in the recycle loop. The beer produced from fermentation of the concentrated steep liquor had an average ethanol content of 7.4% (v/v), well within a range that is acceptable for economic distillation. Fermentation efficiencies were high (86%), and the fermentation was complete in