UEF//eRepository
DSpace https://erepo.uef.fi
Rinnakkaistallenteet Luonnontieteiden ja metsätieteiden tiedekunta
2019
Effect of steam treatment on the chemical composition of hemp
(Cannabis sativa L.) and identification of the extracted carbohydrates and other compounds
Väisänen, Taneli
Elsevier BV
Tieteelliset aikakauslehtiartikkelit
© Elsevier B.V.
CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/
http://dx.doi.org/10.1016/j.indcrop.2019.01.055
https://erepo.uef.fi/handle/123456789/7491
Downloaded from University of Eastern Finland's eRepository
1 1
2 3 4
Effect of steam treatment on the chemical composition of hemp
5
(Cannabis sativa L.) and identification of the extracted
6
carbohydrates and other compounds
7
Taneli Väisänen
1, Petri Kilpeläinen
2, Veikko Kitunen
2, Reijo Lappalainen
1,3, Laura
8
Tomppo
1,a9
1 Department of Applied Physics, University of Eastern Finland, 70211 Kuopio, Finland 10
2 Natural Resources Institute Finland, 02150 Espoo, Finland 11
3 SIB Labs, University of Eastern Finland, 70211 Kuopio, Finland 12
13
a Corresponding author 14
Corresponding author e-mail address: laura.tomppo@uef.fi 15
16
2 A B S T R A C T
17
The overarching aim of this study was to investigate the possibility of extracting carbohydrates and other 18
compounds from hemp (Cannabis sativa L.) using a continuous steam treatment that would also separate the 19
fibres from the hurd. Different parts of hemp, such as stalk, leaves, and decorticated hemp with hurd, were 20
extracted by a steam treatment. After processing, the fibers in the stalk could be easily separated. The 21
products obtained at different extraction temperatures were characterized using multiple techniques. The 22
hemicellulose content of untreated dry hemp was reduced from 275 mg/g to 237 mg/g at 160 °C treatment 23
while the corresponding cellulose content increased from 376 mg/g to 418 mg/g. For example, the iron (Fe) 24
content of the extracts increased with elevated treatment temperatures; for dry hemp at 100 °C, the Fe 25
content in extracts was 1.33 mg/kg, whereas at 160 °C, it was much higher, 16.58 mg/kg. The results 26
demonstrate that the extraction temperature influences the composition of the extracts. Compounds with 27
potentially useful applications, such as in medicine and in the chemical industry, were also identified in the 28
extracts. However, more research will be needed to optimize the treatment and the further processing of the 29
products in order to estimate the commercial potential of this technique.
30
31
GRAPHICAL ABSTRACT 32
33 34
3 Highlights
35
Hemp fibers were retted with continuous steam treatment while extracting compounds 36
The extraction temperature affects the composition of the hemp extracts obtained 37
With the continuous steam treatment the hemicellulose content of hemp decreases 38
The extracts contained compounds with potential to be used by the chemical industry 39
40
Keywords:
41
1. Hemp 42
2. Pressurized hot water extraction 43
3. Hemicelluloses 44
4. Flow-through 45
5. Steam treatment 46
4 1 Introduction
47
Fiber hemp (Cannabis sativa L.) is a rapidly growing plant that can be cultivated in most parts of the world. As 48
its growth requires very little maintenance and its fibers can be separated relatively easily from the other parts 49
of the plant, it has long been used for a variety of applications, such as in ropes and textiles (Shahzad, 2012). Its 50
high specific strength, low density, inexpensive production, non-abrasive nature and biodegradability are all 51
factors that make hemp an attractive material as a composite reinforcement (Dhakal and Zhang, 2015). Hemp 52
has also been used for medical purposes since some varieties contain bioactive compounds that may exert 53
beneficial effects on human health (Zuardi, 2006). Considerable amounts of by-products and residues, such as 54
leaves, pieces of woody core (hurd) and roots, are generated during the production of hemp fibers, and these 55
materials can be either returned to the soil or used as raw material in rather low value applications. Novel 56
ways to exploit these under-used hemp materials could increase their value.
57
Today, there is a global growing interest for bio-based chemicals (de Jong et al., 2012). There are several ways 58
in which biomass waste or residue can be converted into these compounds e.g. thermochemical conversion 59
and microbially-mediated transformations. There are problems associated with the microbially-mediated 60
transformations, including the release of greenhouse gases such as methane (CH4) and carbon dioxide (CO2) 61
into the atmosphere and the difficulties in the process control. Therefore, there is increasing interest in 62
identifying new ways to treat biomass waste (Liu, W. et al., 2017).On the other hand, several research teams 63
have adapted traditional retting process of hemp for producing high-quality fibers under more controllable 64
retting conditions (e.g. Liu, M., Silva et al., 2016). Compounds with high commercial potential are available in 65
different forms of the biomass. For example, high molecular weight galactoglucomannans (GGMs) may be used 66
as raw materials in the production of biofilms (Hartman, Albertsson et al., 2006a; Hartman, Albertsson et al., 67
2006b), biopolymers (Ebringerová, 2006) and hydrogels (Gabrielii et al., 2000). Wood hemicelluloses, softwood 68
GGM and hardwood xylans could potentially replace some of today’s chemical food emulsifiers (Bhattarai et 69
al., 2019; Mikkonen et al., 2016). Xylo-oligosaccharides have also been proposed as food additives and 70
5 nutraceuticals (Moure et al., 2006). After further processing, GGMs can also be utilized as a raw material for 71
ethanol production. Hemicelluloses may also find applications as additives in animal feed (Herrick et al., 2012).
72
It has also been proposed that other commercially interesting compounds, such as acetic acid, citric acid and 73
vanillin could be produced from biomass (de Wild, 2011).
74
Previously, hemp has been treated with short steam treatments involving one or two steps at 170 – 220 °C 75
(Lavoie and Beauchet, 2012) and with steam explosion to achieve the separation and delignification of woody 76
fibres (Vignon et al., 1995). Steam pretreatment (200 – 220 °C) of hemp has been used in the production of 77
ethanol with the addition of SO2 following a short steam treatment (Sipos et al., 2010). The durations of the 78
published high temperature steam treatments were only a few minutes i.e. they were really a short and rapid 79
pretreatment. If lower temperatures and longer treatment times are used, the method can be considered as a 80
continuous steam treatment with similarities to hot water extraction.
81
In the pressurized hot water extraction (PHWE) process, the temperature of water is above boiling point (100 82
°C) but below the critical temperature (374 °C). The pressure in the extraction system is sufficiently high to 83
keep the water in a liquid form. PHWE can be operated in dynamic (flow-through) or static (batch) mode 84
(Leppänen et al., 2011). PHWE has been successfully applied for the extraction of hemicelluloses from trees 85
(Bobleter, 1994; Hasegawa et al., 2004; Sattler et al., 2008), extraction of antioxidants such as isoflavones from 86
soybean flakes (Li-Hsun et al., 2004), extraction of essential oils and oil components from plants (Özel et al., 87
2005) and the extraction of fluorescent whitening agents and azo dyes from paper (de los Santos et al., 2005).
88
Steam treatment induced carbohydrate reactions have been examined with aspen (Li et al., 2005) and it has 89
been used to extract hemicellulosic oligosaccharides from spruce (Palm and Zacchi, 2003). The extraction 90
methods have been scaled-up from the laboratory to the pilot scale (Gallina et al., 2018; Kilpeläinenet al., 91
2014) pointing to potential industrial applications. There is a need to develop environmentally friendly 92
processes where the desired chemical products are easily separated at high purities and large quantities from 93
6 mixtures. In order to maximize the benefit, it would be advantageous if the chemical compounds could be 94
extracted from hemp during retting for fiber separation. In general, the extraction result for hemp would be 95
hypothesized to follow the results obtained for wood materials since the main constituents, i.e. hemicelluloses, 96
cellulose and lignin, and their ratios in hemp are similar to those present in wood. Nonetheless, there are 97
certain distinctive differences between hemp and trees, for example, the presence of cannabinoids, but it is 98
not yet known which of these would manifest themselves in the steam treatment.
99
For clarity, hereafter the term retting refers to any method for separating the fiber from the hurd, i.e. a 100
nomenclature adopted by several other authors (e.g., Hurren et al., 2002; Sisti et al., 2018; Tahir et al., 2011).
101
This use is different from the original use of retting, which has previously referred to a biological process (e.g.
102
Fuller and Norman, 1946).
103
There is a limited amount of information on the characterization of chemical compounds extracted from 104
industrial hemp using PHWE or continuous steam treatment. The overarching aim of this study was to evaluate 105
the retting of hemp fibers with a continuous steam treatment while extracting the chemical compounds. This 106
kind of process could improve the utilization of hemp; fibers, extracted carbohydrates and other compounds 107
could all be exploited by manufacturers. In addition, the extraction potential of the technique was examined 108
with respect to the different parts of hemp. The analyses of retted hemp fibers have been published previously 109
(Väisänen et al., 2018); the current study describes the detailed analyses of the chemical constituents. These 110
analyses provide a basis for evaluating the feasibility of this novel retting process for the extraction of 111
carbohydrates and other commercially valuable compounds.
112
113
7 2 Materials and methods
114
2.1 Hemp 115
Cannabis sativa L. (Futura 75) was sown in Juankoski, Finland, (63.0693889 N, 27.9991111 E) at the end of May 116
2016 in an organically farmed field. The plants were harvested 127 days after sowing and the hemp stalks were 117
immediately collected for further processing. Different parts of the hemp were separated (Fig. 1.). First, the 118
hemp leaves (HL) were removed from the fresh hemp stalk (FH). The fresh hemp stalk was further processed 119
into dried hemp stalk (DH) and mechanically decorticated hemp fiber with residue hurd (DCH). The drying took 120
place at room temperature i.e. about 20 C. Each batch was studied independently. In the chemical analyses, 121
the samples were freeze-dried and then ground in a Retsch MM400 Laboratory Ball Mill. Chemical analyses 122
were selected to provide information on most of the chemical compounds in hemp before and after the 123
treatment, and to explore the compounds that could be extracted should the hemp fibers be separated for 124
composite use by steam treatment.
125
126
Figure 1. Different parts of the plant were separated for the study: (a) fresh hemp stalk (FH) and dried hemp 127
stalk (DH), (b) decorticated fibers with residue hurd (DCH) and (c) leaves (HL).
128
8 129
2.2 Hemp steam treatment 130
Steam treatment was chosen to defibrillate fibers since hot water and steam are used commonly in industry.
131
Unlike organic solvents, water is nonflammable, non-toxic and there are many industrially-applied techniques 132
to concentrate and purify compounds from a water phase such as evaporation, ultrafiltration and high-volume 133
chromatographic methods which are used to isolate sugars. A large 3 L vessel was used so that larger hemp 134
stalks could be placed inside the reactor. Different parts of hemp plants were processed at 100–160 °C for 1 h.
135
Temperatures were higher than 100 °C to obtain steam with the temperature being elevated to increase 136
treatment severity and to determine how these procedures would affect the fibers and the extracted 137
compounds. In a smaller 50 ml laboratory scale vessel, more repetitions could have been done with the stalks 138
cut into small pieces. However, these would not be representative of the treatment at a larger scale. Hemp 139
samples were treated using a laboratory-built flow-through extraction system having a 15 kW preheater for 140
water flow and a 3 L extraction vessel with an in-line temperature measurement (Figure 2). The extraction was 141
conducted by a steaming downwards flow in the extraction system. The extraction system contained an in-line 142
temperature controller to reach the designated processing temperature. The sample was collected in a 40 L 143
vessel. Approximately 24 L of extracts were collected in each extraction and 1 L sample of extracts was frozen 144
at – 20 °C and stored in the dark before the analyses.
145
9 146
Figure 2. Simplified scheme of a steam treatment system.
147 148
2.3 Determination of Klason lignin from dry solids 149
Before the Klason lignin analysis, extractives were removed from the milled samples using ASE-350 (Dionex, 150
USA) with ethanol/water (90/10 v/v) extraction at 90 °C. The Klason lignin content in the dry solid residues was 151
determined using the KCL method (No. 115b:82, (KCL, 1982)). A two-stage acid hydrolysis in an ultrasonic bath 152
was applied to dilute lignin from the solid residues. The first step included the hydrolysis of the samples in 72 % 153
H2SO4 for 1 h. In the second stage, the samples were treated in 3-4% H2SO4 for 4 h followed by refluxing for 4 h.
154
The solid lignin was then filtered, washed and dried to determine the lignin content in the samples. The 155
extinction coefficient of 119 l g-1 cm-1 was applied to determine the acid-soluble lignin content. This method 156
also included a correction due to the UV absorption of carbohydrate degradation products.
157
2.4 Total noncellulosic carbohydrates 158
The acid methanolysis gas chromatography (GC) at 105°C for 3 h was carried out. A freeze-dried extract (2 mg) 159
was treated with 2 ml methanolysis reagent containing 2 M solution of HCl in anhydrous MeOH, which was 160
prepared by dissolution of acetyl chloride in MeOH (Kilpeläinen, Petri et al., 2012). The samples were then 161
cooled down and 100 μl of pyridine was added to the samples in order to neutralize them. A 1 ml volume of 162
10 internal standards (IS) containing 0.1 mg ml-1 resorcinol in MeOH and 0.1 mg ml-1 sorbitol in MeOH were added 163
and a portion of the solution was dried under N2 gas and further dried in a vacuum oven at 40 °C for 15 164
minutes. After the drying, the sample was dissolved in 150 μl pyridine and derivatized with 70 μl 165
trimethylchlorosilane (TMCS) and 150 μl hexamethyldisilazane (HMDS). The sample was then analyzed using 166
GC. The following sugars and sugar acids were analyzed: mannose (Man), glucose (Glc), galactose (Gal), xylose 167
(Xyl), arabinose (Ara), rhamnose (Rha), glucuronic acid (GlcA), galacturonic acid (GalA) and 4-O-methyl 168
glucuronic acid (4-O-Me-GlcA).
169
2.5 The determination of cellulose content 170
The cellulose content was analyzed after the acid hydrolysis and the amount of glucose determined in the 171
methanolysis was subtracted from the result since the crystalline cellulose remains unaffected during the 172
methanolysis (Kilpeläinen, Petri et al., 2012).
173
2.6 Carbon, nitrogen and protein content 174
The carbon and nitrogen contents provide information on the extent to which certain materials were extracted 175
from different hemp parts. Information can be valuable, for example, for life cycle analyses when hemp is 176
grown and processed in industrial processes. Some hemp proteins are used in the food and feed industry.
177
Therefore, the amount of extracted crude protein (N) could indicate whether steam treatment is a viable 178
method to isolate protein from fibrous hemp. Before the analyses, the samples were dried in an oven at 105 °C.
179
The carbon, nitrogen and protein contents were evaluated using a CHN elemental analyzer (CHN-1000, LECO, 180
USA). Approximately 100–150 mg of samples were weighed in a tin foil cup, and subsequently the samples 181
were incinerated at 1050 °C in a flow of oxygen, and the amount of released nitrogen was measured using a 182
thermal conductivity cell. The content of crude protein was then determined according to the procedure of 183
McDonald (1977); the percentage of nitrogen was converted to protein content by multiplying it with a factor 184
of 6.25.
185
11 2.7 Nutrient analysis
186
Inorganic constituents were analyzed in order to determine whether the steam treatment would remove them 187
from the hemp samples. In some scenarios, bio-refineries could utilize fibers that have a low amount of 188
inorganics. If hemp were pulped and bleached after steam treatment, the removal of inorganic compounds 189
would reduce the usage of chemicals. Furthermore, if the solid residue from the extraction is pyrolyzed into 190
biochar, the inorganic content may play a key role in its applications. The samples were wet-digested (HNO3- 191
H2O2) in a microwave (CEM MDS 2000, Charlotte, USA) and the extract was analyzed by an iCAP 6500 DUO ICP- 192
emission spectrometer (Thermo Scientific, UK). The concentrations (mg kg.1) of aluminum (Al), boron (B), 193
calcium (Ca), cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), potassium (K), magnesium (Mg), manganese 194
(Mn), sodium (Na), nickel (Ni), phosphorus (P), lead (Pb), sulfur (S) and zinc (Zn) in dry matter were determined 195
on an inductively coupled plasma atomic emission spectrophotometer (ICP/AES).
196
2.8 Characterization of the extracts by gas chromatography-mass spectrometry (GC-MS) 197
A sample of 1 ml from the 24 L extract was liquid-liquid extracted with methyl tert-butyl ether (MTBE) (Örsa 198
and Holmbom, 1994). Briefly, the sample pH was adjusted to 3, and 2 ml of 0.02 mg/ml MTBE solution with 199
internal standards betulinol and henecosaic acid were added. The sample was mixed thoroughly and 200
centrifuged for 5 min. The MTBE phase was transferred to other test tubes. Sample was further extracted twice 201
with 2 ml of MTBE and all of the MTBE was collected in the same sample. MTBE phase was evaporated in 202
heating block under nitrogen flow at 60 °C and further dried for 15 min in a vacuum oven at 40 °C. The sample 203
was silylated by adding 100 µl pyridine, 100 µl of N,O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) and 50 µl of 204
trimethylsilyl chloride (TMCS). The samples were kept in a heating block for 20 minutes at 70 °C, after which 205
they were cooled down and transferred to GC vials. The samples were analyzed using Agilent 7890 B GC-MS 206
instrumentation gas chromatograph and a 5977 A mass spectrometer. For the separations, two HP-5ms Ultra 207
Inert columns (15 m × 0.25 mm ID, 0.25 μm; Agilent Technologies, Santa Clara, CA, USA) were used. Helium was 208
used as the carrier gas at 1.8157 ml min-1. The temperature of the oven was programmed from 230 to 290 °C at 209
12 4 °C min-1 and finally held at 290 °C for 13 min. The MS was operated in electron ionization (EI) mode and the 210
m/z range from 50 to 800 was scanned with a cycle time of 0.5 s. The temperature of the transfer line was held 211
at 300 °C. The compounds were identified using computer matching of the mass spectra with the NIST 2014 212
library.
213
214
13 3 Results and discussion
215
Figure 3 shows a set of DH specimens after different levels of steam treatment. With the 100 C treatment, the 216
original green color was nearly lost and the bark could be easily peeled from the stalk. At a higher treatment 217
temperature, the stalk color turned brown. The mechanically decorticated fibers with steam treatment at 100 218
C were tested for their mechanical properties and water absorption behavior with the results being reported 219
in detail by Väisänen et al. (2018); the steam treatment reduced the water absorption but deteriorated the 220
fibers tensile strength as compared with untreated and otherwise treated fibers. In the future, the treatment 221
should be optimized in order to conserve better the fiber strength, for example, by reducing the treatment 222
time.
223
224
Figure 3. Dried hemp stalks after the steam treatment. (a) 100 C, (b) 120 C and (c) 160 C.
225 226
Figure 4 shows that the protein content in hemp leaves was considerably elevated when compared with other 227
hemp fractions. On the other hand, other parts of hemp (DH, FH and DCH) contained a greater amount of 228
cellulose and Klason lignin. The hemicellulose content was similar regardless of which part of the plant was 229
14 examined. In the untreated leaves (HL), the relative hemicellulose content was 254 mg g-1. The cellulose
230
content was 96 mg g-1, lignin (Klason and soluble) 166 mg g-1 and protein 247 mg g-1. For DH, the corresponding 231
values were 275 mg g-1, 376 mg g-1, 241 mg g-1 and 29 mg g-1, respectively. FH contained 269 mg g-1 232
hemicelluloses, 396 mg g-1 celluloses, 218 mg g-1 lignin and 35 mg g-1 proteins whereas the relative amounts of 233
these constituents in DCH were 262 mg g-1, 388 mg g-1, 244 mg g-1 and 42 mg g-1, respectively. This data 234
suggests that the highest cellulose content was present in FH whereas most lignin was found in DCH. A similar 235
amount of hemicelluloses was found in all of the studied material types. For example, Stevulova et al. (2014) 236
have reported that the cellulose content of 44.5 %, the hemicellulose content of 32.78 % and the lignin content 237
of 21.03 % in their hurd specimens. On the other hand, the approximate chemical composition of hemp fibers 238
in the literature was as follows: 15 % of hemicelluloses, 60–77 % of cellulose, 3–10 % of lignin and 12–21 % of 239
extractives (Bogoeva‐Gaceva et al., 2007; Faruk et al., 2012; Lilholt and Lawther, 2000; Liu, D. et al., 2012;
240
Rowell et al., 2000). These observations for the untreated materials in this study are closer to those of hurd 241
than processed fibers and similar to the findings of Gandolfi et al. (2013), Lavoie and Beauchet (2012) and Sipos 242
et al. (2010), as expected. The maturity of the plant also affects the chemical composition as the hemp fibers 243
become lignified after the flowering of the plant (Liu, M. et al., 2015).
244
15 245
Figure 4. Analytical data of extraction residues and untreated materials (HL = hemp stalks and leaves, DH = 246
dried hemp, FH = fresh hemp and DCH = decorticated hemp). The amount of proteins was determined by 247
multiplying the nitrogen content ( Fig. 5(b)) by a factor of 6.25.
248 249
Tables 1 and 2 show that the elevation in the extraction temperature generally increased the amount of 250
extracted hemicelluloses, which led to a decrease in the remaining hemicellulose content. For most of the 251
steam-treated hemp materials, the most abundantly extracted hemicelluloses and pectin were Glc, Xyl and 252
GalA. For HL, the amount of extracted Gal was also high as compared with other components. The extracted 253
hemicelluloses and pectins (galactouronic acid and rhamnose) from hemp after steam treatment are presented 254
in Table 2. Extracts of hemp leaves contained mainly glucose and galactouronic acid at 120 C, but when the 255
temperature was increased to 160 C, more arabinoxylan and galactose were extracted from the leaves. Dry 256
hemp stalks behaved similarly when the temperature was increased. At 100 C and 120 C, glucose and 257
galactouronic acid were the main sugars. When the temperature was increased to 160 C, more of 258
hemicelluloses as well as arabinoxylan and galactose, were extracted. A similar trend can be seen for fresh 259
hemp extractions. These results indicate that the treatment temperature should be increased to 160 C if one 260
16 intends to extract hemicelluloses from hemp; at lower temperatures, mainly glucose and pectins are extracted.
261
The nutrient concentrations (Table 3) were the highest in the top part of hemp (HL). The only exception was K 262
whose concentration was higher in DCH. Angelini et al. analyzed the mineral composition (i.e. N, S, P, Mg, Ca, K, 263
Fe, Mn, Cu, Zn, Ni, molybdenum (Mo) and B) in three different plant organs of five hemp cultivars. Table 4 264
shows a comparison between the nutrient concentrations observed in this study and the values reported by 265
Angelini et al. (2014). In most cases, especially for leaves, the nutrient concentrations observed in this study 266
were higher than the concentrations reported by Angelini et al. (2014). The highest percentual differences 267
were observed for the S, Mn, Mg and P concentrations in leaves (800 %, 181 %, 141 % and 125 %, respectively).
268
The differences in the nutrient concentrations may be attributable to the differences in the habitats in which 269
the hemp plants had been cultivated. In the current study, cattle manure was used as the fertilizer whereas 270
urea, triple superphosphate and potassium sulphate were applied by Angelini et al. (2014). The Cannabis sativa 271
L. varieties were also different, possibly causing the discrepancies in the results.
272
273
17 Table 1. Carbohydrate compositions (mg g-1, dry matter) of treated (PHWE, t = 1 h) and untreated hemp.
274
Sample Treatment
temperature (°C)
Man Glc Gal Xyl Ara Rha GlcA GalA 4-O-Me- GlcA
Total
HL Untreated 9 65 36 28 21 12 6 76 1 254
120 10 32 35 48 23 14 3 71 1 238
160 12 37 21 31 7 2 1 28 1 140
DH Untreated 16 74 15 105 5 9 7 42 3 275
100 21 80 16 86 5 9 5 44 2 267
120 16 83 14 95 4 8 6 32 2 258
160 14 69 8 121 2 4 5 11 3 237
FH Untreated 17 67 14 105 5 9 7 42 2 269
120 11 64 12 109 3 8 7 32 3 248
160 14 71 9 129 2 4 6 8 3 246
DCH Untreated 16 84 15 87 5 8 6 38 2 262
100 21 111 17 63 5 10 3 40 2 270
275 276
Table 2. Extracted carbohydrates (mg g-1, dry matter) of original hemp samples.
277
Sample
Extraction temperature
(°C) Man Glc Gal Xyl Ara Rha GlcA GalA 4-O- Me-
GlcA Total
HL 120 1 13 3 1 2 1 2 8 0 30
HL 160 3 16 13 4 10 6 3 13 0 68
DH 100 1 5 1 0 0 0 0 2 0 9
DH 120 2 10 1 0 1 1 0 5 0 19
DH 160 4 8 5 9 2 4 3 8 0 45
FH 120 2 5 1 1 1 1 0 6 0 17
FH 160 7 13 9 25 4 7 5 12 0 82
DCH 100 1 2 1 0 0 0 0 3 0 7
278 279 280
18 281
Table 3. Mean total elemental concentrations (mg kg-1) in the different types of hemp.
282
Element HL DH FH DCH
Al 112 20 15.9 4.48
B 41.1 7.48 8.42 10.3
Ca 24 100 3 420 3 400 4 420
Cd < 0.07 < 0.07 < 0.07 < 0.07
Cr 3.61 0.986 1.25 1.68
Cu 14.3 4.75 4.76 6.02
Fe 265 28.3 26.2 24.8
K 19 400 19 000 18 400 23 000
Mg 6 980 582 517 1 200
Mn 69.6 9.34 6.27 10.7
Na 45.4 9.09 7.84 < 6.33
Ni 3.1 1.06 0.846 0.612
P 6 280 1 700 914 3 310
Pb < 1.05 < 1.05 < 1.03 < 1.05
S 2 730 394 417 667
Zn 65.9 8.42 6.01 13.7
Total 60 110.01 25 185.43 23 725.5 32 669.29 283
284
Table 4. A comparison of the nutrient concentrations between this study and data in the literature.
285
HL vs. leaves FH, DH & DCH vs. Bark and core Nutrient This study Angelini et al.
(2014)
Difference (%) This study Angelini et al.
(2014)
Difference (%)
N (g kg-1) 39.6 33.4 19 5.7 6.0 -5
S (g kg-1) 2.7 0.3 800 0.49 0.32 53
P (g kg-1) 6.3 2.8 125 2.0 1.0 100
Mg (g kg-1) 7.0 2.9 141 0.77 0.55 40
Ca (g kg-1) 24.1 20.3 19 3.7 4.1 -7
K (g kg-1) 19.4 22.1 -12 20.1 11.7 72
Fe (mg kg-1) 265.0 210.7 26 26.4 371.0 -93
Mn (mg kg-1) 69.6 24.8 181 8.8 11.9 -26
Cu (mg kg-1) 14.3 8.8 63 5.2 4.8 8
Zn (mg kg-1) 65.9 31.2 111 9.4 9.4 0
Ni (mg kg-1) 3.1 3.3 -6 0.84 8.4 -90
B (mg kg-1) 41.1 63.6 -35 8.7 20.5 -58
286 287
To the best of the authors’ knowledge, this is the first study to have investigated hemicellulose extraction from 288
industrial hemp. Therefore, the comparison of the results with literature data has to be limited to other types 289
of biomasses used in hemicellulose extraction. The findings of this study are in accordance with the 290
19 observations of Kilpeläinen et al. (2012) and Leppänen et al. (2011); an increase in the extraction temperature 291
leads to a decrease in the hemicellulose content in the process residues. In other words, more hemicelluloses 292
were extracted from the materials as the extraction temperature was increased. Kilpeläinen et al. (2012) 293
extracted hemicelluloses from birch sawdust using a similar system as applied in this study although they used 294
higher extraction temperatures (160–200 °C).
295
296
The relative carbon content in the materials studied was between 43.0–49.8 % (Fig. 5). Poiša et al. (2010) 297
analyzed Cannabis sativa L. to evaluate whether it would be a suitable plant for biomass production in Latvia 298
and Lithuania. Carbohydrates contain 40 % of carbon and lignin aromatic structures approximately 60 %.
299
According to their analyses, the C content in hemp was between 38 % and 41 %, which is close to the values 300
obtained in this study. In the case of the relative N content, hemp leaves contained the highest levels of N (3.5–
301
4.0 %) whereas the N content in other hemp fractions was usually below 0.5 %. Angelini et al. (2014) obtained 302
identical results and Garcia-Jaldon et al. (1998) observed that the N content in hemp bast fibers and hurd was 303
below 0.5 %. As the extraction temperature was elevated, the relative C content in all parts of hemp increased 304
consistently (Fig. 5). Contradictory observations were made for N, i.e. there were inconsistent changes as the 305
extraction temperature was changed. The N content was highest for HL whereas the C content was somewhat 306
similar for all types of hemp. The results from Angelini et al. (2014) confirm the accumulation of N in the leaves;
307
the content is considerably higher when compared with the other parts of the plant.
308
309
20
310
Figure 5. (a) Carbon (C) and (b) nitrogen (N) contents (mg g-1) in treated (PHWE, t = 1 h) and untreated hemp.
311
312
When the extraction temperature was 120 °C, the highest concentration for the dissolved C was present in HL 313
(Table 5). When the extraction temperature was elevated to 160 °C, FH contained the highest concentration of 314
dissolvable C. This phenomenon was not observed for dissolved N; the highest concentration for dissolved N 315
was detected for HL, regardless of the extraction temperature. The dissolved carbon was mainly organic.
316
317
21 The mean total concentrations of the elements in DH and DCH extracts (100 °C) were higher for DH except for 318
Mg, Na and Zn (Table 5). With respect to the extracts obtained at 120 °C, DH contained the highest 319
concentration of Cu and Ni. In general, HL contained the highest concentrations of the elements. When the 320
extraction temperature was increased to 160 °C, the concentrations of the elements generally increased for all 321
the material types. HL contained the highest concentrations B, Cd, Mg, S, Si and Zn. Steam treatment removed 322
silica from hemp samples that could be useful for further processing, for example pulping. The concentrations 323
of Al, Cd, Cu, Fe, Mn and Ni were at their highest in DH whereas FH contained the highest concentrations of Cr, 324
K, Na and S.
325
326
The calculation of the relative amounts of extracted C and N revealed that the elevation of the extraction 327
temperature increased the amount of extracted C and N (Table 6). The highest amount of extracted C at 100 °C 328
was observed for DH whereas at 120 °C, HL released most C. When the temperature was further elevated to 329
160 °C, FH released most C. In addition, we detected correlations between the amounts of extracted 330
hemicelluloses (Table 2), total dissolved solids (TDS) as well as with the amount of extracted carbon. The 331
highest level of N release at 100 °C was observed in DH. At higher temperatures, FH released most N. With 332
respect to the extracted dry matter, the highest amount of dry matter was extracted from DH at 100 °C. When 333
the temperature was increased to 120 °C, the extraction of the dry matter was most efficient for hemp stalks 334
and leaves. At 160 °C, dry matter was most efficiently extracted from FH. The amount of extracted N did not 335
correlate with the dry matter content in the extracts although there was a correlation with the amount of 336
extracted C.
337
338
22 Table 5. Mean total element concentrations in the different types of hemp extracts. DC = dissolved carbon, DIC 339
= dissolved inorganic carbon, DN = dissolved nitrogen and DOC = dissolved organic carbon.
340
HL DH FH DCH
Temperature (°C) 120 160 100 120 160 120 160 100
Sample size (g) 700 270 270 270
Dry matter (%) 35.51 93.31 80.94 93.72
Dry sample (g) 248.57 251.937 218.538 253.044
DC (mg kg-1) 51 500 87 700 18 700 29 100 63 500 28 400 98 600 12 300 DIC (mg kg-1) 879 < 225 372 < 222 < 222 < 256 < 256 408 DN (mg kg-1) 4 470 7 950 1 610 1 700 2 530 2 240 3 310 1 270 DOC (mg kg-1) 50 600 87 400 18 300 28 900 63 400 28 100 98 400 11 900
Al (mg kg-1) 1.74 1.06 < 0.8 1.24 7.53 2.75 5.82 < 0.8
B (mg kg-1) 19.60 22.01 4.38 8.19 7.91 8.79 10.54 3.41
Ca (mg kg-1) 4 270 7 200 1 240 1 150 2 110 1 470 2 410 1 000 Cd (mg kg-1) < 0.11 < 0.11 < 0.11 0.08 0.21 < 0.13 0.15 < 0.11
Cr (mg kg-1) 0.10 0.28 < 0.10 0.16 0.38 0.34 0.63 < 0.09
Cu (mg kg-1) 5.18 0.85 5.98 5.22 5.16 4.67 1.38 4.89
Fe (mg kg-1) 4.15 4.63 1.33 2.29 16.58 3.40 9.33 1.04
K (mg kg-1) 9 530 7 050 15 150 18 290 19 430 19 880 23 720 9 960 Mg (mg kg-1) 2 047 2 858 393 420 583 470 671 406
Mn (mg kg-1) 4.34 5.89 1.62 4.29 8.00 1.21 3.18 1.04
Na (mg kg-1) 630 679 558 573 618 679 730 602
Ni (mg kg-1) 1.16 0.84 0.81 1.16 1.62 0.89 0.99 0.42
P (mg kg-1) 483 557 470 541 471 559 638 348
Pb (mg kg-1) < 0.48 < 0.48 < 0.48 < 0.48 < 0.48 < 0.55 < 0.48 < 0.47 S (mg kg-1) 1 622 1 709 1 248 1 248 1 248 1 505 1 548 1 214 Si (mg kg-1) 1 883 1 612 268 192 123 203 246 198
Zn (mg kg-1) 16.70 7.40 4.21 4.13 4.03 4.07 3.12 4.37
341 342
23 Table 6. The original carbon and nitrogen contents in treated hemp. The extracted C and N are calculated by dividing DOC 343
and DN (Table 5) with the original C and N contents.
344
HL DH FH DCH
Temperature (°C) 120 160 100 120 160 120 160 100
C (mg g-1) 421.87 421.87 466.85 466.85 466.85 465.75 465.75 449.03
N (mg g-1) 38.80 38.80 4.56 4.56 4.56 5.51 5.51 6.45
Extracted C (%) 12.0 20.7 3.9 6.2 13.6 6.0 21.1 2.6
Extracted N (%) 11.5 20.5 35.3 37.2 55.6 40.7 60.2 19.7
Extracted dry matter (%) 14.97 24.23 7.81 10.57 17.53 10.54 25.7 5.50 345
The GC-MS analyses of the hemp extracts are summarized in Tables 7 and 8, showing the compounds identified 346
with a probability of 70 % or higher. Generally, the same compounds were identified from the DH, FH and DCH 347
extractions. The extracts from HL contained several compounds not identified in other fractions, for example, 348
protocatechoic acid and cannabidiol. Terephthalic acid and the other phthalates detected in the extracts 349
probably originated from the plastic containers used in the experiments (Thiruvenkatachari et al., 2007).
350
351
The chemical compositions of the extracts obtained from DH at different temperatures were also rather similar 352
(Tables 7 and 8). All the compounds identified from the extracts that had been produced at 100 °C were also 353
detected in the extracts that were produced at 120 °C and 160 °C. When the temperature of the extraction was 354
elevated from 100 °C to 120 °C, 1-monopalmin was formed. A further increase of temperature from 120 °C to 355
160 °C led to the formation of vanillin, syringealdehyde, sinapaldehyde, 5-allyl-1-methoxy-2,3- 356
dihydroxybenzene and ferulic acid. These compounds are usually considered to be lignin degradation products 357
indicating that at 160 °C, also lignin starts to degrade at the same time as more hemicellulose is extracted at 358
this temperature (Table 2). All these compounds were identified with good probability (P). For example, vanillin 359
is a valuable chemical widely used in the food industry.
360
361
24 Similar observations were made when FH was extracted at 120 °C and 160 °C (Table 8); the compounds found 362
in the extracts formed at 120 °C were also present in the extracts formed at 160 °C. Similar to the findings for 363
DH, the extracts obtained at 160 °C contained syringealdehyde, sinapaldehyde, ferulic acid and 1-monopalmin.
364
365
The extracts produced from DCH at 100 °C exhibited a similar composition as the extracts produced from other 366
parts of hemp. One interesting finding is that only the extracts produced from DCH at 100 °C and HL at 120 °C 367
contained asparagine, which is one of 20 standard amino acids that are common in animal and plant proteins.
368
369
Gutiérrez et al. (2006) analyzed the chemical composition of industrial hemp bast fibers using pyrolysis-GC/MS 370
(Py-GC/MS). Their findings are in accordance with the observations of this study; bast fibers contained vanillin, 371
coumaric acid, sinapaldehyde, palmitic acid, ferulic acids and stearic acid, which were identified here in most of 372
the extracts.
373
Table 7. The chemical compounds of the extracts (N=2) obtained from DH and DCH at 100 °C temperature. Probability (ID) 374
over 90% named S, over 70% P.
375
Chemical name provided by the NIST library
DH DCH
4-Coumaric acid P P
Asparagine P
Citric acid P P
Henecosaic acida S S
Palmitic acid S S
Stearic acid S S
a Internal standard, see text.
376 377 378 379
25 Table 8. The chemical compounds of the extracts (N=2) obtained from DH at different extraction temperatures.
380
Probability (ID) over 90% named S, over 70% P.
381
Temperature (°C) Chemical name provided by the NIST library DH FH HL
120 1-Monopalmin P P
4-Coumaric acid P P
alpha-Linolenic acid P
Asparagine P
Citric acid P P P
D-Pinitol P
Henecosoic acida S S
Palmitic acid S S S
Stearic acid S S S
Terephthalic acid, isophthalic acid S S S
Vanillin P
Vanillin, Tartaric acid P
160 1-Hydroxy-2-methylanthraquinone, isophtalic acid P
1-Monopalmin P P P
4-Coumaric acid P P P
5-Allyl-1-methoxy-2,3-dihydroxybenzene P P
Alizarin yellow P
alpha-Linolenic acid P
Cannabidiol S
Citric acid P P P
D-Pinitol P
Ferulic acid P P
Henecosaic acida S S S
Palmitic acid S S S
Protocatechoic acid P
Sinapaldehyde P P
Stearic acid S S S
Syringaldehyde P P
Terephthalic acid, isophthalic acid* S S
Vanillin P P P
a Internal standard, see text
bLikely artifact, see text 382
Hemp has proven to be an excellent reinforcing material in polymer composites (Hautala et al., 2004; Liu, M. et 383
al., 2016; Maslinda et al., 2017; Väisänen et al., 2018) and it also can be found in textiles but also the other 384
parts of the plants could well have high-value applications. This is the first time that the chemical composition 385
of the extracts obtained from different parts of hemp has been investigated, and thus it can form the basis for 386
future research. The current study reveals that it is possible to separate the fiber from the hurd by continuous 387
steam treatment while simultaneously collecting extracts. The separation of fiber was visually detected and in 388
26 addition, extractives included GalA, which is the backbone of pectin. However, the fiber quality was lower than 389
can be achieved with other separation techniques (Väisänen et al., 2018), and thus more investigations will be 390
required to determine whether the fiber quality could be improved by adjusting the process parameters or 391
whether the value of extracted compounds can compensate for the loss of fiber quality. Furthermore, in the 392
next phase, the economic and environmental costs of the further processing will need to be assessed; in 393
particular, it will be important to conduct a cost-benefit analysis of the separation and purification of the 394
extracted compounds as well as evaluating the utilization potential of the extraction residue. Furthermore, it 395
would be interesting to examine other methods, such as pyrolytic processes, in the production of different 396
chemical compounds from hemp parts. Other techniques should also be tested to identify the compounds from 397
the extracts. In addition, methods to separate certain chemicals from the extracts should be explored and 398
developed.
399
A similar processing technique might be suitable for other fiber-producing crops such as kenaf. However, the 400
competitive advantages of extraction of different crops will depend on which compounds can be extracted as 401
well as determining their concentrations and estimating the costs of purification and the market value of the 402
compounds obtained..
403
404
4 Conclusions 405
In this pioneering study, a potential way to utilize industrial hemp more comprehensively was evaluated. The 406
aim of this research project was to use steam treatment in a flow-through reactor for retting of hemp fibers 407
and to characterize the products obtained from the different parts of hemp (Cannabis sativa L.) at different 408
extraction temperatures. The products were characterized using multiple characterization techniques. The 409
results of this study demonstrate that the fibers could be retted with steam treatment and that the extraction 410
temperature exerted a considerable effect on the composition of the extracts obtained from different parts of 411
hemp. Many compounds were identified from the extracts; these have a good potential to be utilized in further 412
27 applications, such as in medicine or by the chemical industry. However, more research will be needed to 413
optimize the treatment as well as further developing the processing of the products in order to determine the 414
technique’s true commercial potential.
415
416 417
28 Acknowledgements
418
This study was partly funded by the European Agricultural Fund for Rural Development (EAFRD) administered 419
by the Centre for Economic Development, Transport and the Environment, (North Savo, project 16664). We 420
thank Jorma Heikkinen and Terttu Turpeinen for their assistance with the raw material and Arja Tervahauta 421
and Satu Repo for laboratory analyses and technical assistance. We also thank Ewen MacDonald for linguistic 422
help.
423
29 References
424
Angelini, L.G., Tavarini, S., Cestone, B., Beni, C., 2014. Variation in mineral composition in three different plant 425
organs of five fibre hemp (Cannabis sativa L.) cultivars, Agrochimica 58, 1-18.
426
Bhattarai, M., Pitkänen, L., Kitunen, V., Korpinen, R., Ilvesniemi, H., Kilpeläinen, P.O., Lehtonen, M., Mikkonen, 427
K.S., 2019. Functionality of spruce galactoglucomannans in oil-in-water emulsions, Food Hydrocoll. 86, 154-161.
428
Bobleter, O., 1994. Hydrothermal degradation of polymers derived from plants, Prog. Polym. Sci. 19, 797-841.
429
Bogoeva‐Gaceva, G., Avella, M., Malinconico, M., Buzarovska, A., Grozdanov, A., Gentile, G., Errico, M.E., 2007.
430
Natural fiber eco‐composites, Polym. Compos. 28, 98-107.
431
de Jong, E., Higson, A., Walsh, P., Wellisch, M., 2012. Product developments in the bio‐based chemicals arena, 432
Biofuel. Bioprod. Bior. 6, 606-624.
433
de los Santos, M., Batlle, R., Salafranca, J., Nerín, C., 2005. Subcritical water and dynamic sonication-assisted 434
solvent extraction of fluorescent whitening agents and azo dyes in paper samples, J. Chrom. A 1064, 135-141.
435
de Wild, P., Reith, H., & Heeres, E., 2011. Biomass pyrolysis for chemicals. Biofuels, 2(2), 185-208.
436
Dhakal, H.N., Zhang, Z., 2015. The use of hemp fibers as reinforcement in composites. In: Faruk, O., Sain, M.
437
(Eds.), Biofiber Reinforcement in Composite Materials. Woodhead Publishing, Cambridge, UK, pp. 86-103.
438
Ebringerová, A., 2006. Structural Diversity and Application Potential of Hemicelluloses, Macromol. Symp. 232, 439
1-12.
440
Faruk, O., Bledzki, A.K., Fink, H., Sain, M., 2012. Biocomposites reinforced with natural fibers: 2000–2010, Prog.
441
Polym. Sci. 37, 1552-1596.
442
Fuller, W.H., Norman, A.G., 1946. The retting of hemp II. Controlled retting of hemp, Research Bulletin (Iowa 443
Agriculture and Home Economics Experiment Station) 27, 1.
444
Gabrielii, I., Gatenholm, P., Glasser, W.G., Jain, R.K., Kenne, L., 2000. Separation, characterization and hydrogel- 445
formation of hemicellulose from aspen wood, Carbohydr. Polym. 43, 367-374.
446
Gallina, G., Alfageme, E.R., Biasi, P., Garcia-Serna, J., 2018. Hydrothermal extraction of hemicellulose: from lab 447
to pilot scale, Bioresour. Technol. 247, 980-991.
448
Gandolfi, S., Ottolina, G., Riva, S., Fantoni, G.P., Patel, I., 2013. Complete chemical analysis of carmagnola hemp 449
hurds and structural features of its components, BioResources 8, 2641-2656.
450
Garcia-Jaldon, C., Dupeyre, D., Vignon, M.R., 1998. Fibres from semi-retted hemp bundles by steam explosion 451
treatment, Biomass Bioenergy 14, 251-260.
452
30 Gutiérrez, A., Rodríguez, I.M., del Río, J.C., 2006. Chemical characterization of lignin and lipid fractions in 453
industrial hemp bast fibers used for manufacturing high-quality paper pulps, J. Agric. Food Chem. 54, 2138- 454
2144.
455
Hartman, J., Albertsson, A., Lindblad, M.S., Sjöberg, J., 2006a. Oxygen barrier materials from renewable 456
sources: Material properties of softwood hemicellulose‐based films, J. Appl. Polym. Sci. 100, 2985-2991.
457
Hartman, J., Albertsson, A., Sjöberg, J., 2006b. Surface-and bulk-modified galactoglucomannan hemicellulose 458
films and film laminates for versatile oxygen barriers, Biomacromolecules 7, 1983-1989.
459
Hasegawa, I., Tabata, K., Okuma, O., Mae, K., 2004. New pretreatment methods combining a hot water 460
treatment and water/acetone extraction for thermo-chemical conversion of biomass, Energy Fuels 18, 755-760.
461
Hautala, M., Pasila, A., Pirilä, J., 2004. Use of hemp and flax in composite manufacture: a search for new 462
production methods, Compos. Part A Appl. Sci. Manuf. 35, 11-16.
463
Herrick, K.J., Hippen, A.R., Kalscheur, K.F., Anderson, J.L., Ranathunga, S.D., Patton, R.S., Abdullah, M., 2012.
464
Lactation performance and digestibility of forages and diets in dairy cows fed a hemicellulose extract, J. Dairy 465
Sci. 95, 3342-3353.
466
Hurren, C., Wang, X., Dennis, H.G., Clarke, A.F.K., 2002. Evaluation of bast fibre retting systems on hemp, 467
Through the eye of a needle, advances in fibre, fashion and fabric: Textile Institute: proceedings of the 82nd 468
World Conference. The Textile Institute.
469
KCL, 1982. Massan ja puun kokonaisligniinipitoisuus (In Finnish, “Total lignin content of wood and pulp”), KCL 470
(Finnish Pulp and Paper Research Institute), Reports. 115b. Espoo, Finland.
471
Kilpeläinen, P.O., Hautala, S.S., Byman, O.O., Tanner, L.J., Korpinen, R.I., Lillandt, M.K., Pranovich, A.V., Kitunen, 472
V.H., Willför, S.M., Ilvesniemi, H.S., 2014. Pressurized hot water flow-through extraction system scale up from 473
the laboratory to the pilot scale, Green Chem. 16, 3186-3194.
474
Kilpeläinen, P., Leppänen, K., Spetz, P., Kitunen, V., Ilvesniemi, H., Pranovich, A., Willför, S., 2012. Pressurised 475
hot water extraction of acetylated xylan from birch sawdust, Nord Pulp Pap. Res. J. 27, 680-688.
476
Lavoie, J., Beauchet, R., 2012. Biorefinery of Cannabis sativa using one-and two-step steam treatments for the 477
production of high quality fibres, Ind. Crop. Prod. 37, 275-283.
478
Leppänen, K., Spetz, P., Pranovich, A., Hartonen, K., Kitunen, V., Ilvesniemi, H., 2011. Pressurized hot water 479
extraction of Norway spruce hemicelluloses using a flow-through system, Wood Sci. Technol. 45, 223-236.
480
Li, J., Henriksson, G., Gellerstedt, G., 2005. Carbohydrate reactions during high-temperature steam treatment 481
of aspen wood, Appl. Biochem. Biotechnol. 125, 175.
482
31 Li-Hsun, C., Ya-Chuan, C., Chieh-Ming, C., 2004. Extracting and purifying isoflavones from defatted soybean 483
flakes using superheated water at elevated pressures, Food Chem. 84, 279-285.
484
Lilholt, H., Lawther, J.M., 2000. Natural Organic Fibers. In: Kelly, A., Zweben, C. (Eds.), Comprehensive 485
Composite Materials. Elsevier, New York, US, pp. 303-325.
486
Liu, D., Song, J., Anderson, D.P., Chang, P.R., Hua, Y., 2012. Bamboo fiber and its reinforced composites:
487
structure and properties, Cellulose 19, 1449-1480.2 488
Liu, M., Fernando, D., Daniel, G., Madsen, B., Meyer, A.S., Ale, M.T., Thygesen, A., 2015. Effect of harvest time 489
and field retting duration on the chemical composition, morphology and mechanical properties of hemp fibers, 490
Ind. Crop. Prod. 69, 29-39. doi: //doi.org/10.1016/j.indcrop.2015.02.010.
491
Liu, M., Meyer, A.S., Fernando, D., Silva, D.A.S., Daniel, G., Thygesen, A., 2016. Effect of pectin and 492
hemicellulose removal from hemp fibres on the mechanical properties of unidirectional hemp/epoxy 493
composites, Compos. Part A Appl. Sci. Manuf. 90, 724-735. doi: //doi.org/10.1016/j.compositesa.2016.08.037.
494
Liu, M., Silva, D.A.S., Fernando, D., Meyer, A.S., Madsen, B., Daniel, G., Thygesen, A., 2016. Controlled retting of 495
hemp fibres: Effect of hydrothermal pre-treatment and enzymatic retting on the mechanical properties of 496
unidirectional hemp/epoxy composites, Compos. Part A Appl. Sci. Manuf. 88, 253-262. doi:
497
//doi.org/10.1016/j.compositesa.2016.06.003.
498
Liu, W., Li, W., Jiang, H., Yu, H., 2017. Fates of Chemical Elements in Biomass during Its Pyrolysis, Chem. Rev.
499
117, 6367-6398.
500
Maslinda, A.B., Majid, M.A., Ridzuan, M., Afendi, M., Gibson, A.G., 2017. Effect of water absorption on the 501
mechanical properties of hybrid interwoven cellulosic-cellulosic fibre reinforced epoxy composites, Compos.
502
Struct. 167, 227-237.
503
McDonald, C.E., 1977. Methods of Protein Analysis and Variation in Protein Results. Farm. Res. 34, 3-7.
504
Mikkonen, K.S., Merger, D., Kilpeläinen, P., Murtomäki, L., Schmidt, U.S., Wilhelm, M., 2016. Determination of 505
physical emulsion stabilization mechanisms of wood hemicelluloses via rheological and interfacial 506
characterization, Soft matter 12, 8690-8700.
507
Moure, A., Gullón, P., Domínguez, H., Parajó, J.C., 2006. Advances in the manufacture, purification and 508
applications of xylo-oligosaccharides as food additives and nutraceuticals, Process Biochem. 41, 1913-1923.
509
Örsa, F., Holmbom, B., 1994. A convenient method for the determination of wood extractives in papermaking 510
process waters and effluents, J. Pulp Paper Sci. 20, 361-366.
511
Özel, M.Z., Göğüş, F., Hamilton, J.F., Lewis, A.C., 2005. Analysis of volatile components from Ziziphora taurica 512
subsp. taurica by steam distillation, superheated-water extraction, and direct thermal desorption with GC×GC–
513
TOFMS, Anal. Bioanal. Chem. 382, 115-119.
514