1 Introduction

Lignocellulosic biomass (LCB) is a readily available, inexpensive, and renewable feedstock for fermentation to biofuels and other value-added products [1]. A variety of plant/agricultural residues have been identified as potential feedstocks for valorisation. These include, among others, sugar beet pulp, pinewood needles, sugarcane bagasse, sorghum bagasse, and rice straw [2,3,4,5,6]. The component ratios of LCB vary from crop-to crop. Typical values range from 35–45% cellulose, 25–30% hemicellulose, and 15–30% lignin [7]. Lignocellulose is recalcitrant in nature and requires costly pre-treatment to release substrates for bioprocessing. Currently, biofuels produced from LCB cannot compete economically with those from conventional feedstocks such as sugar cane (bioethanol) and oilseeds (biodiesel), despite extensive research in this field [8]. Other biorefinery products from LCB with potentially higher market values include biochemicals (e.g., butanol, acetone) and polyhydroxyalkanote (PHA) precursors for bioplastics [9]. Pre-treatment and hydrolysis increases the porosity of LCB, removes lignin, and reduces cellulosic crystallinity, resulting in the release of the simpler soluble sugars such as glucose, xylose, and arabinose into the hydrolysate for downstream bioprocessing [10].

The most effective pre-treatment of LCB requires a combination of physical, chemical, and hydrolytic reactions. Ideal treatment options should result in the highest sugar yields with minimal release of inhibitors and lowest energy inputs [11]. The first pre-treatment step increases the surface area of the LCB by physically cutting, milling and/or grinding the LCB, usually followed by exposure to high temperature and pressure via steam explosion (SE) to increase the porosity of the feedstock [12]. Chemical pre-treatment relies on acids such as dilute sulfuric acid (dil. H2SO4), ammonia/ammonium (NH3/NH4+), or alkalis such as sodium hydroxide (NaOH) to delignify the LCB and reduce the crystallinity and polymerization of cellulose [13]. While chemical pre-treatment enhances the biodegradability of LCB, higher chemical concentrations may result in the release of microbial inhibitors such as furfurals, salts, and phenolic compounds which may negatively affect the efficiency of downstream biocatalytic reactions [14, 15]. Following chemical pre-treatment, sugar release is often enzymatically enhanced by application of glycoside hydrolases (most commonly cellulases) and lytic polysaccharide monooxygenases [16, 17].

Canola (Canadian oil, low acid) is a rapeseed cultivar originally from Canada that contains negligible erucic acid, a monounsaturated omega-9 fatty acid that is potentially toxic when consumed over long periods [18]. Canola oil has become popular as a healthier alternative to other seed oils, and commercial Canola seed production in South Africa (SA) has increased from 14 to 131452 metric tons between 1999 and 2021 ((South African Grain Industry (SAGIS) [19]). Among other organic agricultural residues, Canola seed processing residues can be considered as feedstocks for valorization into higher value products [20, 21]. During primary and secondary sorting, straw-like LCB residues and other debris known as Canola fines (CF) are separated from the seeds. In SA, Canola crops are silaged upon receipt and then processed over the entire year. The CF and other residues and by-products are therefore non-seasonal in nature, which is a desirable trait when considering downstream usage [21].

This study focussed on finding a simple, effective process for pre-treating CF for maximum release of sugars for downstream fermentation such as the production of PHA or speciality chemicals. Batches of CF were characterized, and pre-treatment was optimized using experiments based on central composite design (CCD) with three factors under consideration viz, temperature, biomass loading, and chemical loading with dil. H2SO4. Physical and chemical (acid) pre-treatment was followed by enzymatic hydrolysis with a commercial cellulase cocktail. Results were assessed by qualitative and quantitative analysis of the sugars released per unit biomass and structural changes in the CF was studied using scanning electron microscopy (SEM), SEM energy dispersive X-ray (SEM-EDX), powder X-ray diffraction (PXRD), and Fourier transform infrared spectroscopy (FTIR).

2 Materials and method

2.1 Sample collection

A batch of CF was collected from a large edible oil processing facility in SA. The samples consisted of fines from an undetermined mixture of Canola cultivars grown in the Western Cape of SA which include Varola 44, 50, and 54, Monty, Insignia, Scoop, Hylite 200 TT, Hyola 60, Oscar, and Rainbow. The material consisted of all matter taken out by the sieves after offloading and prior to pressing of the seed.

2.2 Characterization of biomass

The chemical composition of CF was determined according to the Laboratory Analytical Procedure (LAP) described by the National Renewable Energy Laboratory (NREL) [22]. This involved analytical acid hydrolysis of 300 mg of test sample followed by autoclaving and incubation in 4% H2SO4 in a closed system. The LAP involves standard preparations, neutralization, HPLC, and acetyl analysis. According to the method, each analysis was conducted in quadruplet.

2.3 Pre-treatment and hydrolysis of Canola fines

2.3.1 Physico-mechanical pre-treatment

To reduce the size and increase the area of the CF, they were manually cut into 19 mm then 4.75 mm lengths. Thereafter, they were sieved through a 2.36 mm sieve and allowed to dry overnight in a hot air oven at 50 °C. The dried biomass (pan) <2.36 mm was then ground using a 1200 W Bennett Read extractor (Model: WBL-001H, P.R.C.). Samples of CF after the various mechanical preparation steps are shown in Fig. 1.

Fig. 1
figure 1

Samples of Canola fines after sequential physico-mechanical size reductions

Full size image

2.3.2 Physico-chemical pre-treatment and enzymatic hydrolysis

The pre-treatment of CF was carried out in 250 mL Erlenmeyer flasks, with a working volume of 100 mL. The biomass content and the acid strength were kept as per the DOE (Table 1). Flasks were continuously agitated on an orbital shaker at 150 rpm. Pre-treatment was performed in three sequential stages: (a) chemical hydrolysis, which involves acid pre-treatment of CF with dil. H2SO4 for 24 h at 55 °C, (b) SAAH (steam-assisted acid hydrolysis) where, acid treated CF were further steam exploded (SE) at 15 psi pressure and 120 °C temperature for 15 min, and finally (c) AAEH (acid-assisted enzyme hydrolysis): where SAAH was further hydrolysed enzymatically. Prior to enzymatic hydrolysis, the pH of the reaction mixture was adjusted to pH 6.0–7.0 using NaOH pellets/1M H2SO4. Commercial cellulases, namely cellulase from Trichoderma reesei (cat no: C2730, Sigma-Aldrich, St. Loius, USA), and Viscozyme®, a cellulolytic enzyme mixture from Aspergillus sp. (cat no: V2021, Sigma-Aldrich) were added to the flasks at concentrations of 2 μl/ml each. The flasks were then incubated in a shaking incubator at 150 rpm for 24 h at 50 °C. After each stage, sample aliquots were drawn for quantitative and qualitative analysis of sugars (Section 2.5). The aliquots were centrifuged at 6000 rpm for 10 min, filtered using 0.45 μm sterile filters and stored at −20 °C until required.

Table 1 Statistical design of experiment with three factors studied at three levels
Full size table

2.4 Experimental design

After preliminary testing to inform the levels for each factor (data not shown), the process was optimized in terms of 3 factors: temperature (25–55 °C), biomass loading (2 to 8% weight/volume (w/v)) and chemical loading (0 to 2% dil. H2SO4 volume/volume (v/v)) based on central composite design (CCD) using the Design Expert® software version 13 (Stat-Ease, Inc. Minneapolis, USA). Each numeric factor was set to 3 levels: plus and minus 1 (factorial points) and the centre point. The 20 experimental runs that were generated by the software (Table 1) were each performed in duplicate.

2.5 Analytical methods

2.5.1 High-performance liquid chromatography

The soluble sugars in the filtrates (Section 2.3) were identified and quantified using high pressure liquid chromatography (HPLC) with a Phenomenex (Torrance, USA) Rezex RHM monosaccharide Hþ (8% cross-linkage) column and an Agilent Technologies 1100 series instrument (Santa Clara, USA) as previously described [23].

2.5.2 Scanning electron microscopy and SEM energy-dispersive X-ray spectroscopy

The SEM images and EDX were obtained using a Zeiss (Oberkochen, Germany) MERLIN Field SEM microscope. Prior to imaging, the samples were mounted on aluminum stubs with double-sided carbon tape. The samples were then coated with a thin (~10 nm thick) layer of gold, using a Leica (Wetzlar, Germany) EM ACE200 Gold sputter coater. The beam conditions during the quantitative analysis and the backscattered electron image analysis were set at 20 kV accelerating voltage, 1.4 nA probe current, 9.5 mm working distance, and a beam current of 11 nA. For the EDX, the X-ray counts and energies were quantified using the Oxford (Abingdom, United Kingdom) Aztec software.

2.5.3 Powder X-ray diffraction

Powder X-ray diffraction (PXRD) analysis was performed using a Bruker D8 advanced diffractometer using CuKα1 radiation (λ= 1.5418 Ǻ). The diffractograms were recorded in the range between 5° and 50° of diffraction angle (2θ). A step size of 0.05° was used with 1-s accumulation time. The crystallinity index (CrI) of raw and pre-treated samples was measured using the Segal method (Segal et al., 1959). In this method, the crystalline contribution is determined by the intensity of the 002 peak at 2θ=22.6° and the amorphous by the intensity at 2θ= 16.2° according to Eq. 1.

$$rI= \frac{{I}_{002}-{I}_{am}}{{I}_{002}} \times 100$$
(1)

where I002 is the maximum intensity of the (002) lattice diffraction peak and Iam is the intensity scattered by the amorphous part of the sample.

2.5.4 Fourier transform infrared spectroscopy

FTIR spectrum was recorded between 4000 and 500 cm-1 using a Perkin Elmer 1000 series Fourier transform infrared spectrometer with detector at 4 cm-1 resolution and 25 scan per sample in the transmission mode.

3 Results and discussion

3.1 Characterization of biomass

The major biopolymer components of CF are cellulose, lignin, and hemicellulose, with minor quantities of proteins, pectins, starch, and fatty acids [16, 24]. Lignin from rapeseed is composed of guaiacyl and syringyl aromatics commonly found in hardwoods and cellulosic fibre lengths are also similar to those found in hardwoods [25] In terms of hemicelluloses, xylans typically dominate in hardwoods, while glucomannans dominate in softwoods [8]. Lignin in hardwood is less thermally stable than softwood lignin [8], which is an advantage in terms of pre-treatment using heat to break down lignin.

For biorefinery applications based on fermentation of sugars, pre-treatment of CF to release soluble sugars is required because of the relatively high lignin and hemicellulose contents. In this study, the major macromolecular structural components in the raw CF (%weight/weight (w/w)) were quantified, mainly to assist with the selection of the most suitable pre-treatment method/s. Lignin was present in the highest amount, followed by cellulose and hemicellulose (Fig. 2). Extractives, which include non-structural components of lignocellulose such as fats, phenolics, resin acid, pigments, and waxes constituted 30.5% ± 0.06% w/w of the remaining biomass.

Fig. 2
figure 2

Macromolecular composition and ash measured in raw Canola fines

Full size image

The characterization results were compared with those from previous studies where Canola and rapeseed straws/fines were characterized (Table 2). The high variability, especially in lignin content, indicates that the biomass from each location needs to be characterized separately, viz. results should not be extrapolated from one study to the next. While some variation may be due to methodological inconsistencies, it is more likely to be associated with differences in plant varietals, climate, and/or the nutritional and chemical status of the terrain [18]. For example, Canola plants tend to accumulate reducing sugars in their stems and leaves which act as the major source of carbon (C) for transformation to fatty acids in the (oil)seeds [26, 27]. While Canola crop yields are enhanced, translocation of fatty acids from the stems to the seeds is disrupted in nitrogen (N)-rich environments. This has a negative effect on oil yield and elevates the sugar content of the stems and leaves, and by inference, the straw [28].

Table 2 Literature comparison of lignin, cellulose, and hemicellulose fractions of raw Canola and rapeseed straw/fines
Full size table

In this study, the lignin content of the raw CF fines was similar to a previous study (Table 2) where the final lignin and cellulose contents were <1% and 81.4% w/w after alkaline pre-treatment with 15% NaOH [29]. In a similar study, a cellulose concentration of 79.7% w/w was achieved by pre-treating raw CF with 0.75% NaOH for 6 h [30]. These studies indicate that pre-treatment of CF is needed to release cellulose from lignin and hemicellulose. Pre-treatment with acids or alkaline solutions are the most widely used and cost-effective methods. If high chemical concentrations and/or temperatures are employed, degradation products such as furfural and 5-hydroxymethylfurfural are released in concentrations that may inhibit downstream microbial and/or enzymatic processes and reduce product yields [31,32,33]. For environmental sustainability, and to reduce the generation of inhibitors, low concentrations of chemicals are recommended for pre-treatment [34, 35].

This study therefore focussed on maximizing the recovery of soluble sugars from CF with minimal chemical usage. This was achieved by limiting the concentration of H2SO4 to a maximum of 2% v/v and enhancing the efficiency of the process by incorporating SE and enzymatic saccharification. While it has been shown that there is an inverse relationship between the size of Canola straw and pre-treatment efficiency [30], this was not a factor to consider in this study as the size of the straw was mechanically reduced (Section 2.3).

3.2 The effects of biomass loading, acid concentration, and temperature on sugar yield

The total sugar (TS) yields (% w/w) were determined from the sum of the sugars identified and quantified in the hydrolysates. In this manuscript, the combination of SE and acid hydrolysis is defined as steam-assisted acid hydrolysis (SAAH) and the combination of SAAH and downstream enzyme hydrolysis is defined as acid-assisted enzyme hydrolysis (AAEH). The results obtained after SAAH and AAEH are presented and discussed separately and compared with one another.

3.2.1 Steam-assisted acid hydrolysis of Canola fines

A general quadratic polynomial with sequential sum of squares for two-factor interactions was used to model the total sugar yields from SAAH of the CF (Table 3). The variation between the factor means viz., temperature (A), acid concentration (B), and biomass loading (C) relative to the variation within the factors generated a higher F statistic and lower p-value, thereby rejecting the null hypothesis of the analysis of variance (ANOVA) and concluding the model to be statistically valid. For the fitted model, the F-value and p-values of 38.42 and <0.0001, respectively, indicated that there was only 0.01% chance that the higher F-value was due to noise. The lack of fit (0.61, p>0.05) indicated that the model is a moderate to good predictor of TS yield. However, the acid concentration was the only factor that had a significant direct effect on sugar yield (p <0.0001).

Table 3 Analysis of variance for reduced quadratic model for sugar yield from steam-assisted acid hydrolysis of Canola fines
Full size table

The predicted R2 of 0.7678 for the model was in reasonable agreement with the adjusted R2 of 0.9466 (difference < less than 0.2), while the adequate precision ratio of 16.944, used to measure signal to noise was greater than the desired value of 4. Thus, the model (Eq. 2) can be applied to navigate the design space and predict the TS yield under the optimized set of conditions.

$$TS\;\mathrm{yield}\;\left(Y\right)=3.36-0.07A+1.7B+0.04C-0.08AB+0.05AC+0.18BC+0.34A^2-1.35B^2-0.13C^2$$
(2)

The negligible effect of temperature and biomass loading and highly significant effect of acid concentration on TS yield determined using ANOVA can be seen visually in (Fig. 3). The optimal H2SO4 concentration range determined using RSM for SAAH of CF is between 1.5 and 2.0% v/v (Fig. 3).

Fig 3
figure 3

3D response surface plots showing the effects and inter-relationships of factors on total sugar yield: temperature and H2SO4 concentration (a), temperature and biomass loading (b), and biomass loading and H2SO4 concentration (c)

Full size image

The sugars in lignocellulosic biomass emanate from the cellulosic and hemicellulosic fractions that can be physically, chemically, and enzymatically broken down into their respective monomeric units [11]. Xylose and arabinose are the aldo-pentoses derived from the hydrolysis of hemicellulose, while glucose typically originates from the breakdown of cellulose [20, 38]. It has previously been shown that acid hydrolysis of rapeseed results in dissolution of glucan and xylans from hemicellulose into glucose and xylose, respectively [39]. In this study, the maximum TS yield of 18.57% w/w was achieved with a chemical loading of 2.0% v/v H2SO4 and biomass loading of 8.0%w/v at 25 °C (Fig. 4). Xylose > arabinose > glucose were the major soluble sugars recovered from SAAH, and only negligible amounts of sugars were released from the experiments conducted with no added H2SO4 (Fig. 3). These results indicate that SAAH mainly triggered the breakdown of hemicellulose from the CF.

Fig. 4
figure 4

Sugars measured in the hydrolysates after steam-assisted acid hydrolysis of Canola fines

Full size image

At all biomass loading rates, there was an increase in TS yield when the H2SO4 concentration was increased from 0 to 1% v/v and 1 to 2% v/v, validating the ANOVA results. Biomass loading appeared to have a positive effect on TS yield when increased from 2 to 5% w/v, but not between 5 and 8% w/v. When the biomass loading was increased from 5 to 8% w/v at 1% v/v H2SO4, there was a decreased yield of TS, particularly xylose. This strongly suggests that as the biomass loading increased, there was insufficient acid for efficient hydrolysis at this low (1% v/v) H2SO4 concentration. Although the TS yields at 2% v/v H2SO4 were similar for biomass loading rates of 5% and 8% w/v (Fig. 4), the actual concentration of the sugars in the hydrolysate would be higher at 8% w/v. As the cost of the substrate (waste CF) is not a factor, in a “real-world” scenario, it may be desirable to apply the higher loading rate to achieve higher sugar concentrations without having to increase the volumetric capacity of the reactor. No temperature dependant trends in sugar yields were noted. This was a positive finding because SAAH of CF could conceivably be conducted at ambient temperatures, obviating the need for heating and preventing the release of inhibitors for any downstream microbial and/or enzymatic processes [32, 33]

The maximum TS yield of 18.5 % w/w from SAAH was notably higher than reported in previous studies reporting on hydrolysis of Canola residues. Fermentable sugar yields of 10.2% w/w and 11.5% w/w were obtained from subcritical water hydrolysis of Canola silique and stalks, respectively [40], and a TS yield of 11.8% w/w consisting of 2.5% w/w xylose and 9.3% w/w glucose were obtained using similar methodology (SAAH) to this study [41]. In a third study, the authors [32] did not report on the TS yield, only the sugar concentration (7.22 g/L) after hydrothermal and acid treatment of rapeseed straw. However, the authors also reported dominance of pentose sugars, in keeping with the results of this study. Overall, the results indicated that the SAAH method applied is a suitable pre-treatment method for obtaining high TS yields.

3.2.2 Acid-assisted enzyme hydrolysis of Canola fines

A polynomial quadratic model was obtained from ANOVA of the AAEH data using the TS yield as the response and H2SO4 concentration and biomass loading as input factors. Temperature did not have a significant effect on TS yield after SAAH, and all the AAEH runs were conducted at 40 °C. Temperature was therefore discounted as a factor for analysis of the AAEH results. A high F-value (20.61) and low p-value (<0.0001) were obtained, with only a 0.01% chance that the F-value obtained was due to noise. The null hypothesis was therefore rejected, and it was concluded that the model was valid. As with the SAAH, biomass loading (2 to 5% w/v) was insignificant as a stand-alone factor, while H2SO4 concentration (0 to 2% v/v) was highly significant. However, the model showed that both factors (BC, Table 4) were significant when combined.

Table 4 Analysis of variance for reduced quadratic model for sugar yield from acid-assisted enzyme hydrolysis of Canola fines
Full size table

The predicted R2 of 0.6802 for the model was in reasonable agreement with the Adjusted R2 of 0.8377, with an adequate signal to noise ratio. Akaike’s information criterion (AIC) of 47.58 and Bayesian Information Criterion (BIC) of 47.09 showed the quadratic model to be a best fit. The model may be used to predict the TS yield from AAEH at different H2SO4 concentrations and biomass loading rates within the ranges of the parameters used in the experiments. The quadratic model was adjusted to the experimental data to produce Eq. 3.

$$\mathrm{TS}\;\mathrm{yield}\;\left(Y\right)=4.39+1.38B+0.41C+0.51BC-1.22B^2-0.65C^2$$
(3)

The effects of biomass loading and concentration on TS yield can be seen visually in Fig. 5. The optimal (most desirable) position (0.927) for maximum TS yield (25.15% w/w) are shown in Fig 5a and 5b, respectively. From the coordinates, increases in TS yield should be obtained at biomass loading up until the optimal ~6.8% w/v and H2SO4 concentration up until the optimal ~1.7%.

Fig. 5
figure 5

Contour graphs for desirability (a) and total sugar yield (b) from acid-assisted enzyme hydrolysis

Full size image

Hemicellulases include xylanases, mannases, xylosidases, and arabinofuranosidases [8]. Cellulases and polysaccharide monooxygenases trigger cellulose hydrolysis through oxidative chain breakages. Complete hydrolysis of cellulose results in the release of glucose, but incomplete hydrolysis may result in release of cellobiose which is a disaccharide consisting of two glucose units joined through β-(1,4′) glycosidic bonds.

In this study, enzymatic saccharification resulted in additional release of soluble sugars from cellulose and hemicellulose after AAEH of CF. A notable increase in TS yield between SAAH and AAEH was seen for most of the experimental runs (Fig. 6a), with additional release of monosaccharides as well as cellobiose (Fig. 4, Fig. 6b). A maximum TS yield of 28.3% w/w was obtained from AAEH with a biomass loading of 8% w/v and H2SO4 concentration of 2% v/v for the preceding SAAH step. Under these conditions, the sugars in the hydrolysate consisted of 12.5% w/w xylose, 7.4% w/w glucose, 5.4% w/w cellobiose, and 3.0% w/w arabinose, with respective increases in (w/w) yields of 2.6%, 4.7%, 5.4%, and 1.1%. The abundance distribution of the sugars changed from xylose>arabinose>glucose after SAAH to xylose>glucose>cellobiose>arabinose after AAEH. As xylans dominate the make-up of hemicelluloses in hardwoods [8, 42], the high relative xylose content found in the hydrolysates supports previous research where it was shown that the chemical structure of rapeseed straw is more akin to hardwood than softwood [25, 43]. No cellobiose was found in any of the hydrolysates after SAAH (Fig. 4), but was found in all of the AAEH hydrolysates with biomass loading >2% w/v and H2SO4 concentrations of 1% and 2% v/v. The most plausible explanation for the absence of cellobiose at low biomass loading (2% w/v) was that there was greater enzyme accessibility with less biomass present, and that any cellobiose that was released was hydrolysed to glucose. This could form a case for increased incubation time (>24 h) and/or the use of reduced biomass loading for the second (AAEH) stage. Indeed, increased incubation may result in additional hydrolysis of cellulose to glucose.

Fig. 6
figure 6

Comparison of total sugar yields in the hydrolysates obtained from steam-assisted acid hydrolysis (SAAH) and acid-assisted enzyme hydrolysis (AAEH) (a), and sugars measured in the final hydrolysates after AAEH (b)

Full size image

4 Scanning electron microscopy analysis

Scanning electron microscopy was used to study the morphology of the raw and pre-treated CF (Fig. 7). Smooth and continuous surfaces were noted in the images of the raw CF (Fig. 7a, b). Sequential physical structural damage was noted after each pre-treatment step (Fig. 7a–g) as previously described [44, 45], but some fragments were left intact. This indicated that the lignin was not completely degraded, but that the pre-treatment increased the structural surface area available for acid and then enzyme activity for release of sugars [46].

Fig. 7
figure 7

SEM images at ×100 (left) and ×500 (right) magnification of raw Canola fines (a, b), and Canola fines after steam explosion (SE) (c, d), acid-assisted steam explosion (SAAH) (e, f), and acid-assisted enzyme hydrolysis (AAEH) (g, h)

Full size image

5 Scanning electron microscopy energy-dispersive X-ray spectroscopy analysis

Changes in the elemental composition of the CF were monitored using EDX (Table 5). In all samples, carbon (C) and oxygen (O) were the main atomic elements, and both decreased after each sequential step of pre-treatment. The C content decreased from 62.92% w/w in the raw CF to 48.77% w/w after AAEH, while only minimal decreases in the elemental O occurred (Table 5). Lignin contains high concentrations of C and decreasing C content has been ascribed to delignification of biomass [47]. It was therefore assumed that the pre-treated CF contained less lignin than the raw fines. Furthermore, the decreasing O content could indicate hydrolysis of cellulose and associated release of sugars. The residues of other minor inorganic elements, namely, magnesium (Mg), silica (Si), potassium (K), and calcium (Ca) were also reduced by pre-treatment, suggesting a reduction in ash content (Table 5).

Table 5 Percentage of elements in raw and pre-treated Canola fines detected using SEM-EDX
Full size table

5.1 Powder X-ray diffraction analysis

The CrI provides a measure of the relative fraction of crystalline cellulose in the CF. It is influenced by the presence of the amorphous lignin and hemicelluloses. All the diffractograms from the PXRD analyses of raw and pre-treated CF show the typical XRD peaks of cellulose (Fig. 8). The CrI’s of all the pre-treated CF samples were lower than that of the raw CF as reported in Table 6. Decreased CrI values after pre-treatment of LCB have been observed in previous studies[48,49,50]. The decreasing crystallinity was associated with the disruption of the intra- and inter-molecular hydrogen (H) bonds in the cellulose due to the applied pre-treatment methods which increased the LCB surface area [51]. After the first pre-treatment stages, the CrI decreased from 32.9 % (raw CF) to 15.8 % (after SE) and 13.8% (after SAAH). However, the CrI increased to 21.5 % after AAEH, which is still lower than the CrI of the raw CF. This increase could be attributed to the removal or degradation of a large portion of the amorphous regions of the CF. The presence of sharp peaks following AAEH are due to the crystalline cellulose structure [47].

Fig. 8
figure 8

X-ray diffraction patterns of raw canola fines (raw) and canola after steam explosion (SE), steam-assisted acid hydrolysis (SAAH), and acid-assisted enzyme hydrolysis (AAEH)

Full size image
Table 6 Crystallinity index of raw canola and canola after steam explosion, chemical pre-treatment, and enzyme hydrolysis
Full size table

5.2 Fourier transform infrared spectroscopy analysis

Fournier transform infrared spectroscopy was employed to elucidate the effects of pre-treatment on the chemical structures of the CF (Fig. 9). This was achieved by comparing the infrared (IR) spectra of the raw CF and the CF sampled after each sequential step of pre-treatment. The band detected at 3308 cm-1 of the spectra is associated with the stretching of O–H bonds, which includes both aliphatic and aromatic hydroxyl groups in cellulose, hemicellulose, and lignin structures [52]. Compared to the raw CF, this peak was altered in the pre-treated samples due to the breaking of the ether bonds between the cellulose and lignin resulting in the formation of additional hydroxyl groups. The peaks at 2919 and 2845 cm-1 can be attributed to C–H stretching vibration from the methoxy and methyl groups [53]. The presence of lignin was confirmed by the appearance of stretching vibrations of the aromatic structure at wavenumbers of 1421, 1537, and 1603 cm−1 [54, 55]. After pre-treatment, there was no significant change in these vibrations, which suggests that the main structure of lignin was not completely broken as also evidenced by the SEM images (Fig. 7). The shoulder peak at 1737 cm-1 is associated with the stretching of C=O in hemicellulose [56]. The absorption at this wavenumber decreased after pre-treatment, indicating that the linkages in the hemicellulose were broken. The bands observed at 1034 to 1103 cm-1 can be attributed to the vibration of C–O and C–C stretch in cellulose, hemicellulose, and lignin [57, 58]. These peaks became more pronounced in the pre-treated samples, which indicated the removal of hemicellulose [56, 59]. A stronger effect was observed after each sequential effect, indicating the effectiveness of the pre-treatment in removing hemicellulose. A decrease in the intensity of the 1034 to 1103 cm-1 absorptions following the AAEH pre-treatment is associated with delignification [60, 61]. The peak at 1319 cm−1 corresponds to the C–O vibration of the syringyl ring in the structure of lignin [62, 63]. The small peak at the fingerprint region at 892 is related to the cellulosic β-glycosidic linkages [62, 64]. The peak at 609 cm−1 observed in the pre-treated canola fines is typical of pure cellulose [65].

Fig. 9
figure 9

Fourier transform infrared spectroscopy spectra of (i) untreated canola fines (raw), (ii) steam exploded (SE), (iii) steam-assisted acid hydrolysis (SAAH), (iv) acid-assisted enzyme hydrolysis (AAEH) at wavenumbers of 500–4000 cm-1(a) and the fingerprint region of 500–1800 cm-1 (b)

Full size image

6 Conclusions and recommendations

In this study, the maximum TS yields in the hydrolysates were 18.6% w/w after SAAH and 28.3% w/w after AAEH. To the best of our knowledge, these values are higher than those obtained in other studies reporting on sugar release from Canola or rapeseed straw/fines (10.2–11.8% w/w). The sugar profiles in the hydrolysates after steam-assisted acid hydrolysis (xylose>arabinose>glucose) differed from those obtained after acid-assisted enzyme hydrolysis ((xylose (12.5% w/w) >glucose (7.4% w/w) >cellobiose (5.4% w/w)>arabinose (3.0% w/w)).

For application of this technology, no heating would be required for acid pre-treatment because, within the range tested (20–55°C), temperature was not a significant variable. In addition, the optimal H2SO4 concentration was low (1.7% v/v), minimizing the use of hazardous and costly chemicals as well as the risk of releasing microbial and/or enzyme inhibitors during the pre-treatment process. As relatively high concentrations of sugars were obtained from AAEH, this hydrolysate could be used “as is.” If higher concentrations of sugars, or different sugar profiles are required for a particular downstream application, the AAEH could be followed by enzyme hydrolysis, which could be further optimized to obtain even higher sugar concentrations.

Future work will consist of optimizing the growth of microbial species in the hydrolysates for the production of value-added chemicals and compounds.