ArticlePDF Available

Abstract

Hemp is a sustainable source of natural fibres that can contribute to meet the increasing demand for technical applications in the textile and the composite sectors. Continuous reinforcements can be produced using the existing flax machinery, initially developed for textile purposes. To achieve competitive and economically viable fibre yields and a fibre quality suitable for secondary processing and composite application, hemp needs to be adequately selected and prepared and the flax machinery and settings have to be adapted to the hemp specificities. In this context, this paper studies the influence of agronomic features and processing stages and settings on the effective tensile properties of fibres extracted from two hemp varieties determined using impregnated fibre bundle tests. Results show that the effective properties of fibres are maintained and even improved during processing, in particular during the hackling and stretching steps. Hemp can achieve properties comparable to high quality long flax fibres.
1
Influence of industrial processing parameters on the effective properties of long aligned
1
European hemp fibres in composite materials
2
Xavier Gabrion1, Gilles Koolen2, Marie Grégoire3, Salvatore Musio4, Mahadev Bar3, Debora Botturi5,
3
Giorgio Rondi5, Emmanuel de Luycker3, Stefano Amaducci4, Pierre Ouagne3, Aart Van Vuure2 and
4
Vincent Placet1*
5
1 University of Bourgogne Franche-Comté, FEMTO-ST Institute, CNRS/UFC/ENSMM/UTBM,
6
Department of Applied Mechanics, F-25000 Besançon, France
7
2 Department of Materials Engineering, KU Leuven, B-3001, Heverlee, Belgium
8
3 Laboratoire Génie de Production, LGP, Université de Toulouse, INP-ENIT, Tarbes, France
9
4 Department of Sustainable Crop Production, Università Cattolica del Sacro Cuore, Piacenza, Italy
10
5 Linificio e Canapificio Nazionale, Villa d’Almè, Italy
11
* Corresponding author
12
E-mail address: xavier.gabrion@femto-st.fr
13
Abstract
14
Hemp is a sustainable source of natural fibres that can contribute to meet the increasing demand for
15
technical applications in the textile and the composite sectors. Continuous reinforcements can be
16
produced using the existing flax machinery, initially developed for textile purposes. To achieve
17
competitive and economically viable fibre yields and a fibre quality suitable for secondary processing
18
and composite application, hemp needs to be adequately selected and prepared and the flax machinery
19
and settings have to be adapted to the hemp specificities. In this context, this paper studies the
20
influence of agronomic features and processing stages and settings on the effective tensile properties
21
of fibres extracted from two hemp varieties determined using impregnated fibre bundle tests. Results
22
show that the effective properties of fibres are maintained and even improved during processing, in
23
particular during the hackling and stretching steps. Hemp can achieve properties comparable to high
24
quality long flax fibres.
25
Keywords: A. Natural fibers; A. Biocomposites; B. Mechanical properties
26
27
28
2
1 Introduction
29
Hemp (Cannabis sativa L.) is a multiuse, multifunctional crop that provides raw material to a large
30
number of traditional and innovative industrial applications. Traditionally, its main product was the
31
long bast fibre; now it is cultivated as a dual-purpose crop, for the fibre and the seed [1], or as a
32
multipurpose crop when also the flowers of threshing residues are used to extract high value
33
cannabinoids [2].
34
The use of hemp bast fibre, traditionally linked to the production of textiles, ropes, twines and paper
35
pulp, is now considered for the production of insulation materials or to reinforce composites. Hemp
36
fibres during processing are separated from the woody core (shives), a by-product that has a wide
37
range of applications from the production of MDF to bio-building material, even though its main
38
application is for animal bedding.
39
Hemp, which is well adapted to a wide range of environments, is cultivated all over the globe and its
40
acreage is increasing in China, Europe and North America.
41
Hemp is an environmentally friendly and fast-growing annual crop. It is thereby a substantial
42
consumer of carbon dioxide with an absorption of approximately 1.4 to 1.6 t of CO2 per tonne of hemp
43
[3, 4]. With a yield average of 5.5 to 8 t ha-1, this represents 9 to 13 tonnes of CO2 absorption per
44
hectare harvested [5]. In that respect, hemp provides a carbon-negative material for engineering. Hemp
45
also requires limited amount of water to be produced [6]. Due to its vigorous growth, shading capacity
46
and disease resistance, hemp can be grown without the use of herbicide, pesticide or fungicide. It also
47
regenerates and improves the quality of soils. Inputs of fertilisers are low [7] and the interventions and
48
manpower requirements for farming are limited. The resulting energy cost for raw hemp fibre
49
production is estimated at approximately 5 GJ t-1, about 7 times less than for glass fibres [3].
50
Expressed in CO2 equivalents, the approximate production cost is only 680 kg eq. CO2 t-1 for hemp
51
fibres in comparison to the 2500 and 4000 kg eq.CO2 t-1 required for example for the production of
52
respectively PP and PET fibres [3]. Thus, it constitutes an interesting alternative to mineral and
53
synthetic fibres.
54
In Europe, hemp is currently processed using mechanical systems, based on beating (hammer mills)
55
and or fast rotating nailed rollers (referred to as decorticators), which provide fibres in the form of
56
3
short and medium-length fibres from disordered straws [8]. These very efficient but also very
57
aggressive processing methods are very damaging for the fibres as very high loads are transferred to
58
them. As a consequence, a high number of defects such as kink-bands is observed and this number
59
increases as a function of the process severity [9]. The mean values of their resulting tensile properties
60
are generally significantly lower than for textile flax [10-12]. An alternative to these processing routes
61
(hammer mills or decorticators) is to use aligned straws and the scutching and hackling machinery
62
dedicated to textile flax [8]. Several authors have investigated the production of long hemp processed
63
with such machinery in view of textile and high-added value applications [13-15]. Musio et al. [13]
64
demonstrated that using such flax machinery and a well-controlled retting, hemp can achieve
65
properties comparable to high quality long flax fibres for high performance composites, with tensile
66
stiffness and strength reaching more than 55 GPa and 450 MPa, respectively, to be compared to 59.8
67
GPa and 527 MPa measured for industrially hackled flax [16]. In this study, they obtained quite low
68
scutching yields of long-aligned fibres and high amounts of scutching tows. Grégoire et al. [17]
69
demonstrated recently, at a laboratory scale, that the process parameters can be tuned to significantly
70
improve the long line hackled fibre yields up to 18% of the straw mass and thus obtain values
71
competitive to the flax ones if one considers that hemp shows generally a larger stem yield than flax.
72
Vandepitte et al. [14] also used industrial flax scutching facilities with some of the process parameters
73
changed for hemp extraction purposes with a wide range of European hemp varieties. When compared
74
to Musio et al. [13], higher levels of long fibre scutching yields were obtained but these ones were
75
dependent on the batches, varieties and levels of dew retting. So, there is a great interest and need for
76
optimising the industrial processing of hemp using the scutching and hackling flax machinery. The
77
optimisation work is dedicated to the scutching process settings and not to any change of the different
78
tools such as breaking rollers or beating turbines. Of course, this question would be interesting to raise
79
as the hemp stem diameters are globally larger than the flax ones, especially when the hemp is grown
80
for biomass purpose. When hemp is grown for fibre purpose and especially for garment textile or load
81
bearing composite with fibre extraction performed using scutching equipment, the hemp fibre stem
82
diameters may be considerably reduced to levels lower than 5 mm by increasing the plant density.
83
This is still higher than the flax stem diameter and the flax breaking rollers are probably not perfect,
84
4
but at the present time, it is not possible to change the design of industrial scutching plants as these
85
process hemp only for experiments or small-scale productions and textile flax the very vast majority of
86
the time.
87
The question of the fibre properties to be considered for such optimisation can be raised. Indeed, for
88
textile applications, fibre tenacity, fineness, cleanness and colour are often selected. For composite
89
applications, the choice of designers is mainly driven by the fibres’ strength and rigidity (stiffness). As
90
of today, there are mainly three different experimental methods to determine the tensile properties of
91
fibres: (i) single fibre tensile test (SFTT), (ii) dry fibre bundle test (FBTT) and (iii) Impregnated Fibre
92
Bundle Test (IFBT).
93
SFTT is the most widely applied method for the measurement of the tensile properties of synthetic
94
fibres. For plant fibres, the test is more challenging and time-consuming. For flax, the test is
95
standardized (NFT25-501-2 and NFT25-501-3). The accuracy of the measurement is directly
96
dependent on the fibre preparation (extraction, handling), the experimental settings (fibre alignment
97
and clamping, gauge length, strain rate, hygrothermal conditioning…), data collection (measurement
98
of load and displacement or strain) and post-processing (machine compliance correction,
99
determination of stress and modulus, loading history, assumptions related to isotropy and
100
homogeneity…) [18, 19]. For the strength and stiffness, the most influential parameter, generally
101
leading to large error and scattering in the measurements, is the determination of the cross-sectional
102
area of the fibre [20-27]. A large quantity of fibres (from at least 50 to few hundreds) has to be tested
103
to ensure a reliable analysis of results. A source of confusion and uncertainty also comes from the fact
104
that SFTT can be applied at the scale of individual (elementary) fibres and fibre bundles (‘technical’
105
fibres). These two cases are not always distinguished and the impact of the pectic interface on the
106
measured properties can be significant. For the design of composites, the nature of the tensile
107
properties to be used can also be questioned since, whatever the type of continuous reinforcement
108
used, the resulting composite is reinforced by both individual fibres and bundles of fibres.
109
In FBTT [28], a collection of several fibres are connected in parallel with both ends clamped to a support.
110
While quite easy to implement for synthetic fibres which are produced in the form of continuous
111
5
monofilaments, this test is very challenging for plant fibres due to their finite length and the resulting
112
discontinuities within rovings and yarns.
113
The IFBT is a well-established method for carbon and glass fibres, standardised for continuous and
114
staple-carbon fibre yarns (ISO 10618:2004). It was also adapted a few years ago for natural fibres and
115
validated by a round robin exercise on flax fibres [16]. In this method, an unidirectional composite is
116
manufactured and loaded in the fibre direction. The fibre stiffness and strength are then identified by
117
inverse method from the measured composite properties and a micromechanical model, generally the
118
rule of mixtures. A good impregnation quality with a negligible content of residual voids is required
119
[29]. If initially conceived and used for tensile loading, its use under compressive loading has also
120
been investigated more recently for flax [30].
121
Literature pointed out that the properties measured using these different testing methods can be
122
significantly different [31-34]. Shah et al. [32] explored the potential sources of the observed
123
discrepancy and concluded that the more likely origins relate to both measurement uncertainties and
124
inaccuracy in predictions based on the rule-of-mixtures. However, the main advantages of IFBT lie in
125
the simplicity of the preparation of the specimens and in the implementation of the test and also in the
126
fact that a large quantity of fibres is tested simultaneously, including individual fibres and bundles of
127
fibres. It also gives access to the effective properties of the fibres, i.e. the reinforcing potential of the
128
fibre in the matrix, resulting from the fibre properties but also from the fibre-matrix interfacial
129
bonding, the fibre individualisation and spatial distribution in the resin. So, it is considered so far as
130
the most efficient method to determine quickly and reliably the effective properties of fibres as they
131
behave in the composite.
132
The objective of this study is to provide pieces of knowledge for the optimisation of the industrial
133
processing of aligned hemp straws using the scutching and hackling flax machinery. The influence of
134
the processing stages (including scutching, hackling, sliver forming, doubling and stretching) and
135
settings on the effective properties of fibres is investigated on two hemp (Cannabis sativa L.) varieties,
136
namely Futura 75 and Fibror 79 cultivated in Italy in the frame of the BBI-JU project SSUCHY
137
(www.ssuchy.eu). Hemp is a widespread crop, well adapted to a broad range of environmental
138
conditions. Traditionally Italy was the European leader in the production of fibre hemp with up to
139
6
130.000 ha of cultivations and still now most of the European varieties have been bred from Italian
140
genotypes. The two selected varieties are both bred in France by Hemp-It and are relatively similar for
141
their cycle, habitus and productivity and they are both well adapted to Italian conditions. Interestingly,
142
Fibror 79 is a « yellow » variety, which is easier to process than traditional ones [35].
143
Tests were realised on hemp straws harvested and retted in 2018 and 2019. The sowing and harvest
144
times as well as the stem portion are considered in the analysis. The mechanical properties of the long-
145
aligned fibres obtained for the different batches at the different stages of the processing are evaluated
146
using IFBT.
147
2 Materials and methods
148
2.1 Materials and processing
149
2.1.1 Hemp stems: field trials, varieties and harvesting
150
Hemp straws were obtained from large-scale field trials carried out in Piacenza, Italy (45° 3' 9.436" N
151
9° 41' 34.742" E) in 2018 and 2019 with two monoecious varieties, a green one Futura 75 (FUT) and a
152
yellow one Fibror 79 (FIB). They were sowed with a target plant density of 150 plants m-2. The field
153
was fertilized with 60 kg of nitrogen per ha.
154
In 2018, the sowing was carried out on 25th April with a seed rate of 50 kg/ha. Stems were harvested at
155
two times: at the end of flowering on 11th August, and at seed maturity on 28th September as a dual
156
valorisation increases the income of the farmer (H2). Stems from H1 were dew-retted for 5 weeks (until
157
18th September), which was judged as appropriate for a good retting level according to the colour of the
158
stems.
159
In 2019, a similar protocol than in 2018 was followed with an early sowing time in April and two
160
harvesting times (H1 early august) and H2 (early September). Stems were left on the field until mid-
161
October to obtain a complete and optimum level of retting.
162
For the harvest, a prototype machine was used. It enabled an efficient cut of the stems and the
163
formation of swaths of aligned stems. The harvested plants were then laid in the field for the dew-
164
retting. A second prototype was used to cut the plants in 1-meter stem portions and to turn the swath to
165
improve retting homogeneity. The bottom and middle part of the plant, cut in 1-meter portion, were
166
7
baled separately. The meteorological data (rainfall and temperature) are added in the supplementary
167
information.
168
2.1.2 Fibre processing
169
170
Figure 1: Schematic representation of the processing stages and resulting products
171
Long aligned fibres were extracted using industrial flax machinery, specifically, an industrial
172
Depoortere scutching device and a Linimpianti hackling machine located at the Terre de Lin company
173
(Saint-Pierre-Le-Viger, France). The scutching machine is composed of two distinct devices: a
174
breaking system composed of a succession of horizontal fluted rollers and a beating stage which
175
consists of successive pairs of rotating turbines, with each turbine rotating in opposite direction.
176
For the hemp straw harvested in 2019, two different settings were used for the scutching, breaking and
177
beating steps, labelled R1 and R2. R1 corresponds to the high speed (settings used for flax processing)
178
while R2 corresponds to a lower speed. The exact values of speed and settings are confidential and
179
cannot be given here.
180
At the end of the hackling line a continuous sliver with a large count (linear mass of about 15,000 tex)
181
was realised. This sliver was then processed into rovings at an industrial scale in Linificio e
182
Canapificio Nazionale (Villa d’Almè, Italy). The sliver was drawn and doubled several times and
183
slightly twisted to obtain at the end a roving of 350 Tex with a twist level of approximately 35
184
turn/meter. This process was also performed at the lab scale drawing system from Linimpianti
185
company (called Mini-Sytem), usually used in the industrial flax processing to evaluate the spinning
186
ability of scutched and hackled fibres. This was used in this study to evaluate if any difference in the
187
effective properties of the fibres can be detected between the matters processed with the Mini-System
188
8
(supposed to mimic the industrial process) and the industrial process itself. Indeed, the Mini-system
189
simulates, at a reduced scale, the six drawing/doubling stages used in the flax spinning industry to
190
prepare the slivers into rovings that will be used at the spinning stage. It consists of six drawing stages
191
during which the linear mass of the sliver is decreased up to 150 Tex. During the different stages of
192
this process, six parallel “Gill type” systems perform the different drawing operations. During this
193
stage, the sliver mass is reduced but it is also homogenised as between each drawing stage, six drawn
194
slivers are each time grouped together before the following drawing. During these operations, the
195
technical fibre diameter may also be reduced when the technical fibres are pulled from the Gill system
196
pins.
197
The whole process is schematically represented in Figure 1. To facilitate the identification of the
198
matter, the following label will be used (See Figure 2). The label fields and entries are described in
199
Table 1.
200
201
Figure 2. Identification of the matter.
202
Table 1: Label fields and entries used for the nomenclature of the fibre samples
203
Label fields
Label entries
Year
2018
2019
Variety (VAR)
FUT =
Futura 75
FIB
Fibror 79
Sowing time (S)
1 = first
sowing
time
2 = second
sowing
time
Harvesting (H)
1 = Full
flowering
2 = seed
maturity
Stem portion
1 = 1st
meter
2 = 2nd
meter
year_variety_sowing time_harvesting time_stem portion_scutching parameter_process step
2018_FIB_1_2_1_R1_Sc
9
Scutching
parameter settings
R1 = flax
settings
R2 =
reduced
speed
Process stage
Sc =
scutching
Ha =
hackling
MS =
doubling/stretching
using MiniSystem
R =
doubling/stretching
using industrial
equipment
204
The production of continuous reinforcement from the discontinuous hemp fibres requires a certain
205
minimum quantity. This quantity has not been systematically reached for all the batches. Therefore,
206
for the two tested years (2018 and 2019), some of the batches produced the same year were mixed
207
together to form a single sliver or roving. They were labelled 2018_MIX_Sl, 2018_MIX_MS,
208
2018_MIX_R and 2019_MIX_R, respectively.
209
2.2 IFBT specimens manufacturing
210
IFBT specimens were prepared following the technical document “Impregnated Fibre Bundle Test
211
IFBT Methodology of uses published by the European Confederation of Flax and Hemp (CELC)
212
[36]. This work being carried out within the framework of a European collaborative project, the
213
specimens were prepared in two different laboratories, labelled A and B, using the same above-
214
mentioned protocol and some small adjustments related to the know-how and previous experiences of
215
the involved research teams. The exact protocols are described below. For both, the IFBT specimens
216
were manufactured by aligning the long fibres obtained at the different stages of the processing and
217
impregnating them with an epoxy system.
218
Protocol A (UFC)
219
The specimens were polymerized from the GreenPoxy 56 resin and the SD 7561 hardener provided by
220
Sicomin company. The fibre samples were placed in the mould cavity after being conditioned at a
221
temperature of 23°C and a relative humidity of 50% during few hours to reach equilibrium. The resin
222
was poured on top of the fibres as the fibres were placed in the mould. The quantity of resin used was
223
calculated to reach a fibre volume fraction target in composites of approximately 40%. The counter-
224
mould was then placed on the top. No spacer was used to limit the porosity level. The specimens were
225
10
then cured at 60°C during 1h under a pressure of 2 bars and demoulded. A post-curing at 130°C was
226
then realised during 1h. The dimensions of the manufactured specimens were approximately 200 mm
227
x 16 mm x 1 mm. The manufactured IFBT specimens were then stored in a climatic chamber at 23°C
228
and 50% RH during a minimum of four weeks to reach the equilibrium moisture content. The
229
dimensions and mass of each specimen were measured. At least, six specimens were manufactured for
230
each tested condition.
231
The fibre volume fraction
, matrix volume fraction
and void content
were determined using
232
the following equations:
233
 
(1)
234
  
 
(2)
235
   

(3)
236
where
is fibre weight fraction determined as the ratio of the measured mass of fibres to the
237
measured mass of composite, , and specific gravity of composite, fibre and matrix,
238
respectively. was determined as the ratio of the measured mass of composite to the measured
239
volume of composite and and were previously determined by pycnometry with values equal to
240
1.503 g/cm3 and 1.17g/cm3, respectively.
241
Protocol B (KU Leuven)
242
Protocol B is similar to A except for the following points. An Epikote 828 LVEL/ Dytek DCH-99
243
epoxy system was used. The fibres were dried during at least 24h at 60°C before the manufacturing of
244
the IFBT specimens. The specimens were cured at 150°C during 2h in a manual hydraulic press.
245
Spacers were used in the mould to approach a specimen thickness of 2mm. They were then
246
conditioned for at least one month at 21°C and 54% RH above a salt solution. The dimensions of the
247
manufactured specimens were approximately 200 mm x 10 mm with a thickness varying between 1.6
248
and 2 mm.
249
2.3 Testing methods
250
11
2.3.1 Fibre fineness
251
The fineness was measured using a FiberShape device developed by IST AG (Vilters, Switzerland). It
252
consists of a high-precision Reflecta MF 5000 scanner (Reflecta, Eutingen im Gaü, Germany),
253
associated with the Silverfast fibre recognition software developed by Lasersoft Imaging (LaserSoft
254
Imaging, Kiel, Germany).
255
No particular pre-treatment (temperature or humidity stabilization, etc.) was carried out on the fibres.
256
The technical fibres were only cut to a fixed length before being fed into the scanner.
257
The parameters used are as follows: bundle length: 2cm; measurement accuracy: 3200 dpi; number of
258
measurements: 4 000-10 000. This number, corresponding to the number of scanned technical fibres is
259
necessary to establish a good representation of the fibrous population. The number of scans depends
260
on the number of fibres placed on the scanner.
261
2.3.2 IFBT
262
Tensile tests were done on the produced IFBT specimens. As for the manufacturing, the testing of the
263
IFBT specimens was realised in the two different laboratories. The respective protocols used are
264
described below.
265
Protocol A (UFC)
266
For each condition, tensile tests were conducted on at least five IFBT specimens using a MTS
267
Criterion 45 universal machine, with a crosshead displacement rate of 1 mm/min and a load cell of
268
100kN. The longitudinal strain was measured with an Instron 2620-601 extensometer with a gauge
269
length of 50 mm.
270
Protocol B (KU Leuven)
271
Tensile tests were performed on at least five specimens using a Zwick/Roell Z100 universal
272
testing machine equipped with a 100kN load cell and a displacement rate of 2 mm/min. The
273
longitudinal strain was measured with optical and clip on extensometers with a gauge length from 50
274
to 80 mm.
275
The small difference in displacement rate (factor 2) is supposed to have neglectable effect on the
276
measured properties. Indeed, it was previously demonstrated for unidirectional flax epoxy composite
277
12
that the measured tensile properties are significantly different only when the displacement rate is
278
changed by a factor of 10, in the considered displacement range [37].
279
2.3.3 Back-calculation of fibre properties from IFBT tests
280
To correctly implement and exploit the IFBT tests, the selected matrix should have a high ductility so
281
that the failure strain of the matrix is higher than that of the fibres. The mechanical properties of the
282
epoxy systems used are synthetized in Table 2.
283
Table 2: Mechanical properties of the epoxy systems used for the IFBT specimens
284
Epoxy systems
E-modulus
(GPa)
Stress at failure
(MPa)
Strain at failure
(%)
GreenPoxy 56 / SD 7561
2.5
60
5
Epikote 828 LVEL + Dytek DCH-99
2.7
70
4
285
Protocol A (UFC)
286
The effective longitudinal modulus of the fibres (Ef ) was obtained by back-calculation using the rule
287
of mixtures proposed by Madsen et al. [38] for plant fibre composites (Eq. 4):
288

 (4)
289
where EC is the composite modulus, Em the matrix modulus, η0 the fibre orientation factor, η1 the fibre
290
length factor. In this work, the fibre length factor (η1) was considered equal to 1 (which is generally
291
the case when the length to diameter ratio of the fibres is higher than 50). is equal to 1 for all the
292
specimens except for the roving which has a twist level of approximately 35 turns/meter. In this case,
293
the fibre orientation factor is calculated using equation 5.
294
  with   (5)
295
where α is the surface twist angle, r the radius of roving and T the twist level of roving.
296
Taking into account the non-linear tensile behaviour generally observed for plant fibre
297
composites, the composite modulus was measured on two different strain ranges: Ec1 between 0 and
298
0.1% of longitudinal strain and Ec2 between 0.3 and 0.5%. The corresponding moduli determined by
299
back-calculation at the scale of the fibres were noted Ef1 and Ef2.
300
13
The effective longitudinal tensile strength of the fibres (
f ) was obtained by back-calculation using
301
the equation suggested in [39, 40] for plant fibre composites (Eq. 6):
302

 (6)
303
with
c the stress at failure of the composite,
m the stress in the matrix at the failure strain of the
304
composite, ηd the fibre diameter distribution considered to be equal to 1 in this study.
305
306
Protocol B (KU Leuven)
307
The effective stiffness and strength of the fibres were determined as in Bensadoun et al. [16]. The
308
equations used are given below. Again, stiffness was determined in two strain intervals.
309

(7)
310

(8)
311
2.3.4 Statistical analysis
312
ANOVA (Analysis of Variance) tests were performed to evaluate if the means of the measured
313
mechanical properties of the tested batches were significantly different from each other. Most of the
314
time, tests were realised on two batches to better discriminate the influence of one of the tested
315
features (i.e., variety, year, sowing time, harvesting time, stem portion, scutching parameter and
316
process stage). The confidence interval was fixed at 95%. For each test, a probability Pr was
317
calculated. The difference between means is considered to be significant when Pr is inferior to 0.05.
318
When more than two batches were compared at once, a single-step multiple comparison was preferred.
319
A Tukey’s test was used to evaluate if the means are significantly different from each other. It applies
320
simultaneously to the set of all pairwise comparisons. Letters (a, b and ab) are used to report the
321
results of the pairwise comparisons.
322
2.3.5 SEM observations
323
The cross-sections of the IFBT specimens were observed using a Scanning Electron Microscope
324
TESCAN Mira3 operating at 20 kV. The specimens were embedded in a PMMA resin and polished
325
with silicate paper (until fineness 2400).
326
14
3 Results and discussion
327
3.1 Tensile behaviour of IFBT specimens non-linearity and scattering
328
329
Figure 3: Tensile stress-strain curves of two tested batches of IFBT specimens pointing out the typical non-linear behaviour
330
and scattering observed for the different tested batches
331
Figure 3 shows the typical tensile responses obtained for the IFBT specimens. Two tested batches are
332
plotted. The shape of the tensile curves as well as the scattering of results within a same batch are
333
representative of those observed for all the tested specimens. The tensile response is clearly non-linear
334
with a linear response until a yield point located at a stress and strain level of approximately 40-50
335
MPa and 0.15-0.2%, respectively. This is typical of the unidirectional (UD) plant fibre composites and
336
it was often documented for flax fibres [41-47]. However, a significant difference can be observed
337
when compared to the typical bi-phasic behaviour reported for UD flax composites. Indeed, after the
338
yield point a decrease in the apparent modulus is generally observed and this one remains almost
339
constant up to failure. In the case of the tested UD hemp composites, a significant increase in the
340
apparent modulus can be observed in the last stage of the tensile test. The difference of morphology
341
(Figure 4), ultrastructure and the interface properties between hemp and flax fibres and the epoxy
342
matrix can explain this difference in behaviour. This behaviour is similar to the observed at the scale
343
of the individual hemp fibres [48, 49] often referred to as “type-3” behaviour. At the scale of
344
individual fibres, the non-linear behaviour was attributed to complex phenomena involving stick-slip
345
mechanisms and cellulose microfibrils reorientation in the fibre wall, and stress-induced crystallisation
346
of the amorphous or pseudo-crystalline cellulose [49]. A non-linear tensile behaviour was also recently
347
15
reported at the scale of the cellulose microfibrils themselves using molecular dynamics simulation
348
[50]. Previous studies also pointed out that the shape of the fibre cross-section and in particular the
349
degree of ellipticity has a strong effect on the shape of the nonlinearity of the tensile response [27].
350
This morphologic effect was demonstrated to be strongly related and coupled to structural parameters
351
and physical mechanisms, such as the cellulose microfibrillar angle and the viscoelastic behaviour of
352
the material of the fibre wall. The observed behaviour at the scale of the IFBT specimens could then
353
result from the fibre behaviour and the difference observed when compared to flax could be attributed
354
to the fibres’ geometry. Figure 4 clearly shows the complex morphology of the hemp fibres, in
355
particular when compared to flax.
356
Figure 3 also shows that the scattering within a same batch is quite limited in the first part of the curve
357
and then increases with the increasing strain. This is attributed to the initial defects in the composite
358
and the propagation of damage under tensile loading in the second part of the test. Indeed, initial
359
cracks are often observed in IFBT specimens (Figure 4) due to the presence of impurities and
360
remaining bark tissues.
361
362
Figure 4: SEM image of the cross-section of an IFBT specimen showing the typical microstructure of the hemp composite
363
(left) and SEM image of the cross-section of a specimen showing the typical microstructure of the flax composite (right). m:
364
matrix, sf: single fibre, bf: bundle of fibres, c: cracks
365
The tensile properties measured at the scale of the IFBT specimens are synthetized in the table in
366
supplementary information. The coefficients of variation of Ec1, Ec2 and c are in the range of 1-19%,
367
1-15% and 3-23%, respectively. This scattering in the tensile properties is attributed to the
368
heterogeneities in the spatial distribution of fibres within and between samples, the porosity level, the
369
m
bf
sf
c
16
presence of impurities in fibres (remaining shives and/or pieces of bark) and the geometrical defects of
370
IFBT specimens.
371
Table 3: Results of the statistical analyses on the back-calculated fibre properties Evaluation of the impact of IFBT protocol
372
(A and B). Letters (a, b and ab) are used to report the results of the pairwise comparisons.
373
LABORATORY
Ef1
Ef2
f
2018_FIB_1_2_1_R1_Sc A
50.5 a
28.2 b
369 a
2018_FIB_1_2_1_R1_Sc B
46.1 a
23.4 a
449 b
Pr > F(Model)
0.098
0.026
0.005
Significant
No
Yes
Yes
2018_FIB_1_2_1_R1_Ha A
56.8 a
32.1 b
423 a
2018_FIB_1_2_1_R1_Ha B
51.8 a
23.2 a
490 a
Pr > F(Model)
0.227
0.002
0.114
Significant
No
Yes
No
2018_FUT_1_2_1_R1_Sc A
50 a
24.8 a
420 a
2018_FUT_1_2_1_R1_Sc B
45 a
21.5 a
433 a
Pr > F(Model)
0.294
0.201
0.768
Significant
No
No
No
2018_FUT_1_2_1_R1_Ha A
53 b
27.2 b
443 a
2018_FUT_1_2_1_R1_Ha B
45.3 a
21.7 a
412 a
Pr > F(Model)
0.015
0.016
0.477
Significant
Yes
Yes
No
374
The values obtained for the same fibre batches by the two different laboratories were also compared.
375
Results of the statistical tests are presented in Table 3. The differences in the mean values are about 8
376
to 15% for Ef1, 13 to 28% for Ef2 and 3 to 22% for f. Most of the differences in the mean values are
377
not significant from a statistical point of view except for Ef2. For this latter quantity, some differences
378
17
may be related to the gauge length used for the strain measurement which is different for the two
379
laboratories. The appearance of heterogeneities in the strain fields during the last stage of the tensile
380
test due to the propagation of damage could then induce discrepancy in measurements in particular
381
when the monitored specimen’s length is different. A significant difference is also observed in the
382
fibre strength of the first batch (2018_FIB_1_2_2_R1_Sc), and in the Ef1 for the fourth batch
383
(2018_FUT_1_2_1_R1_Ha). It is worth mentioning that the fibre individualisation was not fully
384
achieved for this batch after scutching and that strong heterogeneities were observed between the fibre
385
packages. The influence of the laboratory on the IFBT results was also underlined in [16]. However,
386
the observed differences in this present study are lower and most of the time not significant from a
387
statistical point of view. The quantification of the scattering and uncertainties in the measurements
388
was a crucial step before investigating the influence of agronomic and processing parameters on the
389
back-calculated fibre properties.
390
3.2 Influence of agronomic parameters on the effective properties of fibres
391
The influence of the agronomic parameters on the fibre properties was investigated. The results of the
392
statistical analysis are presented in Table 4.
393
Interestingly, no significant difference (except for Ef2 for Fibror 79 variety) in the properties of the
394
hackled fibres (in the form of bundles of hackled fibres or slivers) was observed between the years
395
2018 and 2019, and this for the two tested varieties. The Ef1 modulus is comprised between 52 and 61
396
GPa, the Ef2 modulus between 27 and 36 GPa and the strength between 400 and 500 MPa, for the two
397
years. It shows that the mechanical performance of the processed fibres can be reproduced from year
398
to year. On the contrary, a significant difference in stiffness was observed between the years 2018 and
399
2019 at the scale of the mixed rovings. This can be attributed to the diversity of the very different
400
batches mixed, particularly in year 2018. However, this result should not be transposed directly to the
401
material which could be marketed later because the processors have a great know-how in the mixing
402
of materials during the production of rovings to ensure good homogeneity and quality and good
403
reproducibility over years.
404
18
No significant difference between the two varieties, Fibror 79 and Futura 75, was observed, as well in
405
2018 for the scutched and hackled fibres extracted from the first stem meter, and harvested at seed
406
maturity. On the contrary, a significant difference was observed between the two varieties in 2019, at
407
least after sliver forming; the Fibror 79 variety performing better. This could be due to the level of
408
retting. Indeed, Hendrickx [51] pointed out for flax that Ef2 is significantly influenced by the extent of
409
retting.
410
As already observed by Musio et al. [13], the harvest at full flowering provides slightly stiffer fibres
411
when compared to harvest at seed maturity. For all the tested batches except for the property Ef2 of the
412
batches “Ms”, the stem portion does not influence the fibre properties results.
413
Table 4: Results of the statistical analyses on the back-calculated fibre properties Evaluation of the impact of variety, year,
414
harvest time and stem portion. Letters (a, b and ab) are used to report the results of the pairwise comparisons.
415
YEAR
Ef1
Ef2
f
2018_FUT_1_2_1_R1_Ha
53 a
27.2 a
443 a
2019_FUT_1_2_1_R1_Sl
52.2 a
29.2 a
397 a
Pr > F(Model)
0.744
0.499
0.168
Significant
No
No
No
2018_FIB_1_2_1_R1_Sl
61.4 a
33.5 a
421 a
2019_FIB_1_2_1_R1_Sl
57.8 a
36.3 b
508 a
Pr > F(Model)
0.07
0.022
0.067
Significant
No
Yes
No
2018_MIX_R_A
59.6 b
43.2 b
616 a
2019_MIX_R_B
50.6 a
32.7 a
615 a
Pr > F(Model)
<0.0001
<0.0001
0.945
Significant
Yes
Yes
No
VARIETY
Ef1
Ef2
f
2018_FIB_1_2_1_R1_Sc
50.5 a
28.2 a
368.7 a
19
2018_FUT_1_2_1_R1_Sc
50 a
24.8 a
420 a
Pr > F(Model)
0.834
0.204
0.175
Significant
No
No
No
2018_FIB_1_2_1_R1_Ha
56.8 a
32.1 a
422.9 a
2018_FUT_1_2_1_R1_Ha
52.9 a
27.2 a
442.9 a
Pr > F(Model)
0.302
0.081
0.606
Significant
No
No
No
2019_FIB_1_2_1_R1_Sl
57.8 a
36.3 a
508 a
2019_FUT_1_2_1_R1_Sl
52.2 b
29.2 b
397 b
Pr > F(Model)
0.037
0.028
0.022
Significant
Yes
Yes
Yes
HARVEST TIME
Ef1
Ef2
f
2018_FUT_1_2_1_R1_Sc
50.0 a
24.8 a
420 a
2018_FUT_1_1_1_R1_Sc
58.4 a
32.3 b
318 a
Pr > F(Model)
0.06
0.036
0.07
Significant
No
Yes
No
2018_FUT_1_2_1_R1_Ha
53.0 a
27.2 a
443 a
2018_FUT_1_1_1_R1_Ha
62.8 b
34.7 b
491 a
Pr > F(Model)
0.04
0.006
0.202
Significant
Yes
Yes
No
2019_FUT_1_2_1_R1_Sl
52.2 a
29.2 a
397 a
2019_FUT_2_1_1_R1_Sl
64.3 b
39.3 b
379 a
Pr > F(Model)
0.001
0.013
0.695
Significant
Yes
Yes
No
STEM PORTION
Ef1
Ef2
f
2018_FUT_1_1_1_R1_Sc
58.4 a
32.3 a
318 a
20
2018_FUT_1_1_2_R1_Sc
55.7 a
28.7 a
337 a
Pr > F(Model)
0.451
0.240
0.597
Significant
No
No
No
2018_FUT_1_1_1_R1_Ha
62.8 a
34.7 a
491 a
2018_FUT_1_1_2_R1_Ha
60 a
32.7 a
513 a
Pr > F(Model)
0.455
0.441
0.642
Significant
No
No
No
2019_FIB_1_2_1_R1_Sl
57.8 a
36.3 a
508 a
2019_FIB_1_2_2_R1_Sl
55.2 a
36 a
504 a
Pr > F(Model)
0.191
0.816
0.927
Significant
No
No
No
2019_FIB_1_2_1_R1_Ms
56.2 a
33.7 a
506 a
2019_FIB_1_2_2_R1_Ms
52.9 a
31.5 b
515 a
Pr > F(Model)
0.112
0.009
0.852
Significant
No
Yes
No
416
3.3 Influence of processing parameters on the effective properties of fibres
417
Table 5: Results of the statistical analyses on the back-calculated fibre properties Evaluation of the impact of scutching
418
speed. Letters (a, b and ab) are used to report the results of the pairwise comparisons.
419
SCUTCHING SPEED
Ef1
Ef2
f
2019_FUT_1_2_1_R1_Sl
52.2 a
29.2 b
397 a
2019_FUT_1_2_1_R2_Sl
58.4 a
38.4 a
445 a
2019_FUT_1_2_1_R2_Ms
51.1 a
32.2 ab
510 a
Pr > F(Model)
0.065
0.016
0.169
Significant
No
Yes
No
2019_FIB_1_2_1_R1_Sl
57.8 b
36.3 b
508 a
21
2019_FIB_1_2_1_R2_Sl
64.1 a
39.9 a
420 a
Pr > F(Model)
0.010
0.037
0.054
Significant
Yes
Yes
No
2019_FIB_1_2_1_R1_Ms
56.2 a
33.7 a
506 a
2019_FIB_1_2_1_R2_Ms
49.1 a
30.8 a
492 a
Pr > F(Model)
0.105
0.097
0.824
Significant
No
No
No
420
In 2019, the impact of the scutching speed was also evaluated for the two varieties after sliver forming
421
and doubling/stretching (Table 5). The reduction of the scutching speed led to a slight but significant
422
increase of Ef2 after sliver forming for Futura 75, while both Ef1 and Ef2 were increased for Fibror 79.
423
This difference is not anymore significant after doubling and stretching (certainly due to the increase
424
in the fibre fineness). In all cases, no significant effect is observed on the fibre strength.
425
The reduction of the scutching speed therefore does not really change the tensile properties of hemp
426
fibres. However, reducing the stem progression speed as well as the turbine speed was shown to be
427
particularly interesting regarding the long fibre yield. This reduces the generation of scutching tows
428
and as a consequence maximise the long fibre yield as shown by Gregoire et al. [14] at the laboratory
429
scale. It is therefore necessary to adjust the processing speed so that to minimise the generation of
430
tows, but it is also important to keep the processing speed relatively high to keep the fibre production
431
rate sufficiently high. A compromise between the long fibre yield and the production rate has to be
432
determined.
433
3.4 Influence of processing stages on the effective properties of fibres
434
Finally, the impact of the processing steps, including scutching, hackling, sliver forming, doubling and
435
stretching, was investigated. Results of the statistical analysis are synthetized in Table 6. For the
436
Fibror 79 variety cultivated and harvested in 2018, the Ef1 modulus increases from 50.4 GPa to 61 GPa
437
and the strength from 369 MPa to 513 MPa from the scutched state to the doubled/stretched one.
438
22
For the mixed batch realised in 2018, a significant improvement except for Ef1 was also observed with
439
a general increase of the effective properties, from 55.9 GPa to 60 GPa for Ef1, 37.6 GPa to 42.7 GPa
440
for Ef2 and the strength increased from 439 MPa to 616 MPa from the sliver state to the
441
doubled/stretched one. So, the effective properties are significantly increased during the processing in
442
particular during hackling and stretching steps. This is attributed to the increase in cleanness and
443
fineness of the fibres, while the mechanical properties of the fibres all along the process are not
444
globally decreased drastically. The cleanness and fineness of the fibre induce a better adhesion
445
between the fibres and the matrix, an increased adhesion surface, a better spatial distribution and less
446
initial damage in the composite following manufacturing. Gregoire et al. [16] demonstrated that an
447
equivalent number of kink-bands is globally present in hemp fibres extracted using a soft laboratory or
448
the more aggressive scutching/hackling process. However, the size of the kink-bands is larger in the
449
fibres extracted when submitted to more aggressive process parameters. If the kink-bands area is
450
larger, it may be expected that the zones of weakness in the fibres are increased and this has as effect
451
to reduce the tensile properties of the fibres. The fibres considered in the present paper were extracted
452
using the same industrial equipment and process parameters as in [17], and it is expected that they
453
contain similar number and similar surface area of defects. With such an amount of defects, the tensile
454
properties of the elementary fibres determined in [17] are still globally high with tensile strength of
455
600 MPa and tensile modulus of 40 GPa and the mechanical potential of the fibres sufficient for load
456
bearing composites. In a different study, Grégoire et al. [9] showed that the number of kink-bands
457
increases as a function of the process severity. As a consequence, the number of kink-bands and the
458
size of kink-bands after hackling is higher than after scutching. But as the technical fibres are more
459
separated, the reinforcement properties of the fibres determined from back calculation of composite
460
properties remain equivalent. Figure 5 shows a significant increase in fibre fineness during hackling
461
with an average fibre width coming down from 61.9 to 47 m for Futura 75 and from 67.4 to 48.2 m
462
for Fibror 79, and a less marked but still significant (p= 2.10-5) decrease during stretching with a mean
463
fibre width decreasing from 49.1 to 45.9 m.
464
Table 6: Results of the statistical analyses on the back-calculated fibre properties Evaluation of the impact of processing
465
steps: scutching, hackling, sliver forming, doubling and stretching. Letters (a, b and ab) are used to report the results of the
466
pairwise comparisons.
467
23
PROCESSING STAGES
Ef1
Ef2
f
2018_FIB_1_2_1_R1_Sc
50.4 b
28.2 a
369 b
2018_FIB_1_2_1_R1_Ha
56.8 ab
32.0 a
423 ab
2018_FIB_1_2_1_R1_Sl
61.4 a
33.5 a
421 ab
2018_FIB_1_2_1_R1_Ms
60.9 a
33.9 a
513 a
Pr > F(Model)
0.002
0.065
0.023
Significant
Yes
No
Yes
2018_MIX_Sl
55.9 a
37.6 a
439 a
2018_MIX_MS
59.7 a
34.9 a
486 a
2018_MIX_R
60 a
42.7 b
616 b
Pr > F(Model)
0.06
0.004
0.009
Significant
No
Yes
Yes
468
469
Figure 5: Influence of the processing stages on the fibre fineness. Influence of hackling evaluated on the Futura 75 (a.) and
470
Fibror 79 (b.) varieties and influence of stretching characterised on the mixed batch formed in year 2018 (c.).
471
The improvement of the effective properties of long aligned fibres over the processing sequence is a
472
major achievement since a detrimental effect of processing was observed in a previous work on long
473
aligned hemp fibre [52]. Thygesen et al. [52] reported a monotonic decreasing relationship between
474
the strength and the number of processing steps (including retting, scutching, carding and
475
cottonisation). The fibre bundle strength was reported to be on average reduced by 27% per processing
476
step at the applied conditions. It is also worth mentioning that the rovings type yarns obtained in the
477
present study are suitable for weaving, without any traditional spinning step involving high twist of the
478
a. b. c.
24
fibres. The demonstration was made recently by Corbin et al. [53, 54] at the scale of woven balanced,
479
unbalanced and quasi-unidirectional fabrics. Results also point out that the different stages of the
480
hackling and scutching route up to the obtention of a low-twisted roving, as usually implemented for
481
the production of yarns for textile applications can be simplified in view of composite applications.
482
483
Figure 6: Tensile properties of the hemp fibres identified by inverse method from the IFBT tests, for the different tested
484
conditions. a. Ef1, b. Ef2 and c. strength. The bar represents the mean value and the error bar the standard deviation; blue
485
and orange represent the tests at the two participating laboratories.
486
All the back-calculated fibre properties are finally synthetized in Figure 6. Results underline that
487
whatever the agronomic and processing parameters considered, the effective fibre tensile stiffness
488
(modulus Ef1) and strength are comprised between 45 and 64 GPa and 320 and 616 MPa, respectively.
489
0
10
20
30
40
50
60
70
80
Ef1 (GPa)
A B
0
5
10
15
20
25
30
35
40
45
50
Ef2 (GPa)
A B
0
100
200
300
400
500
600
700
σf(MPa)
A B
a.
b.
c.
25
These results can be compared to the ones published in literature for hackled flax and hemp fibres
490
from previous studies (Figure 7). It clearly points out that hemp can achieve properties comparable to
491
high quality long flax fibres for high performance composites, to be compared to on average 59.8 GPa
492
and 527 MPa previously measured for industrially hackled flax in the frame of a round robin test [16].
493
494
Figure 7: Effective tensile stiffness and strength of hemp fibres identified by inverse method from IFBT tests. Comparison to
495
data from literature for hemp and flax fibres.
496
4 Conclusion
497
In this study, the influence of the processing stages and settings on the effective properties of fibres
498
were investigated for the two cultivated hemp varieties, namely Futura 75 and Fibror 79, using textile
499
flax machinery. Tests were realised on hemp straws cultivated, harvested and retted in Italy in 2018
500
and 2019. The sowing and harvest times as well as the stem portion were considered in the analysis.
501
The mechanical properties of the long-aligned fibres obtained for the different batches at the different
502
stages of the processing were evaluated using IFBT (impregnated fibre bundle test).
503
No significant difference in the effective properties of the fibres extracted in the first and second meter
504
of the stems was observed. It means that the whole stem can be valorised for composite application. It
505
is an important output in particular to maximise the fibre yield through processing and to ensure an
506
economically viable cultivation and transformation of hemp straws.
507
Comparable properties were obtained for the two cultivation years demonstrating that a good fibre
508
quality can be achieved from year to year.
509
As already observed in literature, the harvest at full flowering provides slightly stiffer fibres when
510
compared to harvest at seed maturity. However, the effective properties obtained at seed maturity are
511
still suitable for composite applications. This option allowing the seed harvesting would guarantee
512
0
100
200
300
400
500
600
700
800
020 40 60 80
σf(MPa)
Young's Modulus (GPa)
Flax IFBT Bensadoun et al.
Hemp IFBT Musio et al.
Hemp IFBT [This work A & B]
26
more income to the farmers and thus a more prosperous and profitable valorisation of the hemp
513
production.
514
Interestingly, a good preservation of the effective fibre properties was observed over the processing
515
steps and even an improvement was seen during hackling and stretching. This is attributed to both the
516
conservation of the integrity of the fibres and the improvement of their individualization.
517
Overall, results point out that with a well-controlled retting and well-suited processing settings, hemp
518
can achieve effective properties comparable to high quality long flax fibres. The fibre quality is
519
suitable for the production of low-twisted rovings that can be further used for weaving. This work also
520
suggests that the different stages of the scutching and hackling route up to the manufacturing of a low-
521
twisted roving type yarn, as usually implemented for the production of yarns for textile applications
522
could be shortened in view of composite applications as the final spinning step involving high twist of
523
the fibres is not desired.
524
To validate the results presented in this work, future research should be carried out in different
525
environmental conditions as these can affect fibre quality during plant growth and particularly during
526
the phase of dew-retting.
527
Acknowledgements
528
This project has received funding from the BioBased Industries Joint Undertaking (JU) under the
529
European Union’s Horizon 2020 research and innovation program under grant agreement No 744349
530
(SSUCHY project). The JU receives support from the European Union’s Horizon 2020 research and
531
innovation programme and the Bio-based Industries Consortium. This work has also been supported
532
by the EIPHI Graduate school (contract "ANR-17-EURE-0002"). The authors would also like to thank
533
the Terre de Lin company for opening its scutching and hackling industrial facilities as well as some
534
of its characterization equipment. XG and VP thank Stani Carbillet for his support in SEM
535
observations.
536
References
537
[1] Tang K, Struik PC, Yin X, Thouminot C, Bjelková M, Stramkale V, et al. Comparing hemp
538
(Cannabis sativa L.) cultivars for dual-purpose production under contrasting environments. Industrial
539
Crops and Products. 2016;87:33-44.
540
27
[2] Calzolari D, Magagnini G, Lucini L, Grassi G, Appendino GB, Amaducci S. High added-value
541
compounds from Cannabis threshing residues. Industrial Crops and Products. 2017;108:558-63.
542
[3] Carus M. Overview on the European Hemp Industry. 12th International Conference of the
543
European Industrial Hemp Association (EIHA). Wesseling, Germany. 2015.
544
[4] Meirheage C. Evaluation de la disponibilité et de l'accessibilité de fibres végétales à usages
545
matériaux en France. In: FRD, editor.: Etude réalisée pour le compte de l'ADEME par FRD. Mars
546
2011. p. 84.
547
[5] Mirizzi F. Hemp cultivation & production in Europe in 2018. In: EIHA, editor.: European
548
Industrial Hemp Association. 2018.
549
[6] Duque Schumacher AG, Pequito S, Pazour J. Industrial hemp fiber: A sustainable and economical
550
alternative to cotton. Journal of Cleaner Production. 2020;268:122180.
551
[7] Amaducci S, Scordia D, Liu FH, Zhang Q, Guo H, Testa G, et al. Key cultivation techniques for
552
hemp in Europe and China. Industrial Crops and Products. 2015;68:2-16.
553
[8] Amaducci S, Gusovious HJ. Hemp cultivation, extraction and processing. In: Müssig J, editor.
554
Industrial Applications of Natural Fibres: Structure, Properties and Technical Applications. 2010.
555
[9] Grégoire M, Barthod-Malat B, Labonne L, Evon P, De Luycker E, Ouagne P. Investigation of the
556
potential of hemp fibre straws harvested using a combine machine for the production of technical load-
557
bearing textiles. Industrial Crops and Products. 2020;145:111988.
558
[10] Placet V, Day A, Beaugrand J. The influence of unintended field retting on the physicochemical
559
and mechanical properties of industrial hemp bast fibres. Journal of Materials Science. 2017;52:5759-
560
77.
561
[11] Réquilé S, Mazian B, Grégoire M, Musio S, Gautreau M, Nuez L, et al. Exploring the dew retting
562
feasibility of hemp in very contrasting European environments: Influence on the tensile mechanical
563
properties of fibres and composites. Industrial Crops and Products. 2021;164:113337.
564
[12] Bourmaud A, Beaugrand J, Shah DU, Placet V, Baley C. Towards the design of high-
565
performance plant fibre composites. Progress in Materials Science. 2018;97:347-408.
566
[13] Musio S, Müssig J, Amaducci S. Optimizing Hemp Fiber Production for High Performance
567
Composite Applications. Frontiers in Plant Science. 2018;9.
568
[14] Vandepitte K, Vasile S, Vermeire S, Vanderhoeven M, Van der Borght W, Latré J, et al. Hemp
569
(Cannabis sativa L.) for high-value textile applications: The effective long fiber yield and quality of
570
different hemp varieties, processed using industrial flax equipment. Industrial Crops and Products.
571
2020;158:112969.
572
[15] van der Werf HMG, Turunen L. The environmental impacts of the production of hemp and flax
573
textile yarn. Industrial Crops and Products. 2008;27:1-10.
574
[16] Bensadoun F, Verpoest I, Baets J, Müssig J, Graupner N, Davies P, et al. Impregnated fibre
575
bundle test for natural fibres used in composites. Journal of Reinforced Plastics and Composites.
576
2017;36:942-57.
577
[17] Grégoire M, Bar M, De Luycker E, Musio S, Amaducci S, Gabrion X, et al. Comparing flax and
578
hemp fibres yield and mechanical properties after scutching/hackling processing. Industrial Crops and
579
Products. 2021;172:114045.
580
[18] Bourmaud A, Placet V. Overview of plant fibres for bio-based composite reinforcement; main
581
specificities and key properties. 1st ESBBC - European Summer School on Bio-based Composites.
582
Online. 2021.
583
[19] Depuydt D, Hendrickx K, Biesmans W, Ivens J, Van Vuure AW. Digital image correlation as a
584
strain measurement technique for fibre tensile tests. Composites Part A: Applied Science and
585
Manufacturing. 2017;99:76-83.
586
[20] Virk A. Numerical models for natural fibre composites with stochastic properties. 2010.
587
[21] Aslan M, Chinga-Carrasco G, Sörensen B, F., Madsen B. Strength variability of single flax fibres.
588
Journal of Materials Science. 2011;46:6344-54.
589
[22] Thomason JL, Carruthers J, Kelly J, Johnson G. Fibre cross-section determination and variability
590
in sisal and flax and its effects on fibre performance characterisation. Composites Science and
591
Technology. 2011;71:1008-15.
592
[23] Mattrand C, Béakou A, Charlet K. Numerical modeling of the flax fiber morphology variability.
593
Composites Part A: Applied Science and Manufacturing. 2014;63:10-20.
594
28
[24] Haag K, Müssig J. Scatter in tensile properties of flax fibre bundles: influence of determination
595
and calculation of the cross-sectional area. Journal of Materials Science. 2016;51:7907-17.
596
[25] Garat W, Corn S, Le Moigne N, Beaugrand J, Bergeret A. Analysis of the morphometric
597
variations in natural fibres by automated laser scanning: Towards an efficient and reliable assessment
598
of the cross-sectional area. Composites Part A: Applied Science and Manufacturing. 2018;108:114-23.
599
[26] Del Masto A, Trivaudey F, Guicheret-Retel V, Placet V, Boubakar L. Investigation of the
600
possible origins of the differences in mechanical properties of hemp and flax fibres: A numerical study
601
based on sensitivity analysis. Composites Part A: Applied Science and Manufacturing.
602
2019;124:105488.
603
[27] Del Masto A, Trivaudey F, Guicheret-Retel V, Placet V, Boubakar L. Nonlinear tensile behaviour
604
of elementary hemp fibres: a numerical investigation of the relationships between 3D geometry and
605
tensile behaviour. Journal of Materials Science. 2017;52:6591-610.
606
[28] Hansen A, Hemmer PC, Srutarshi P. The Fiber Bundle Model: Modeling Failure in Materials.
607
2015.
608
[29] Pisupati A, Ayadi A, Deléglise-Lagardère M, Park CH. Influence of resin curing cycle on the
609
characterization of the tensile properties of flax fibers by impregnated fiber bundle test. Composites
610
Part A: Applied Science and Manufacturing. 2019;126:105572.
611
[30] Prapavesis A, Tojaga V, Östlund S, Willem van Vuure A. Back calculated compressive properties
612
of flax fibers utilizing the Impregnated Fiber Bundle Test (IFBT). Composites Part A: Applied
613
Science and Manufacturing. 2020;135:105930.
614
[31] Oksman K, Wallström L, Berglund LA, Filho RDT. Morphology and mechanical properties of
615
unidirectional sisal epoxy composites. J Appl Polym Sci. 2002;84:2358-65.
616
[32] Shah DU, Nag RK, Clifford MJ. Why do we observe significant differences between measured
617
and ‘back-calculated’ properties of natural fibres? Cellulose. 2016;23:1481-90.
618
[33] Charlet K, Jernot J-P, Gomina M, Bizet L, Bréard J. Mechanical Properties of Flax Fibers and of
619
the Derived Unidirectional Composites. Journal of Composite Materials. 2010;44:2887-96.
620
[34] Virk AS, Hall W, Summerscales J. Modulus and strength prediction for natural fibre composites.
621
Materials Science and Technology. 2012;28:864-71.
622
[35] Müssig J, Amaducci S, Bourmaud A, Beaugrand J, Shah DU. Transdisciplinary top-down review
623
of hemp fibre composites: From an advanced product design to crop variety selection. Composites
624
Part C: Open Access. 2020;2:100010.
625
[36] hemp ECofa. Impregnated Fibre Bundle Test - IFBT - Methodology of uses. 2015.
626
[37] Jeannin T, Gabrion X, Ramasso E, Placet V. About the fatigue endurance of unidirectional flax-
627
epoxy composite laminates. Composites Part B: Engineering. 2019;165:690-701.
628
[38] Madsen B, Thygesen A, Lilholt H. Plant fibre composites porosity and stiffness. Composites
629
Science and Technology. 2009;69:1057-69.
630
[39] Shah DU, Schubel PJ, Clifford MJ. Modelling the effect of yarn twist on the tensile strength of
631
unidirectional plant fibre yarn composites. Journal of Composite Materials. 2012;47:425-36.
632
[40] Summerscales J, Dissanayake N, Virk A, Hall W. A review of bast fibres and their composites.
633
Part 2 - Composites. Composites Part A: Applied Science and Manufacturing. 2010;41:1336-44.
634
[41] Strohrmann K, Hajek M. Bilinear approach to tensile properties of flax composites in finite
635
element analyses. Journal of Materials Science. 2019;54:1409-21.
636
[42] Liang S, Gning PB, Guillaumat L. A comparative study of fatigue behaviour of flax/epoxy and
637
glass/epoxy composites. Composites Science and Technology. 2012;72:535-43.
638
[43] Shah DU, Schubel PJ, Clifford MJ, Licence P. The tensile behavior of off-axis loaded plant fiber
639
composites: An insight on the nonlinear stressstrain response. Polym Compos. 2012;33:1494-504.
640
[44] Shah DU. Damage in biocomposites: Stiffness evolution of aligned plant fibre composites during
641
monotonic and cyclic fatigue loading. Composites Part A: Applied Science and Manufacturing.
642
2016;83:160-8.
643
[45] Berges M, Léger R, Placet V, Person V, Corn S, Gabrion X, et al. Influence of moisture uptake on
644
the static, cyclic and dynamic behaviour of unidirectional flax fibre-reinforced epoxy laminates.
645
Composites Part A: Applied Science and Manufacturing. 2016;88:165-77.
646
[46] Bensadoun F. In-service Behaviour of Flax Fibre Reinforced Composites for High Performance
647
Applications: KU Leuven; 2016.
648
29
[47] Poilâne C, Cherif ZE, Richard F, Vivet A, Ben Doudou B, Chen J. Polymer reinforced by flax
649
fibres as a viscoelastoplastic material. Composite Structures. 2014;112:100-12.
650
[48] Placet V, Cisse O, Boubakar L. Nonlinear tensile behaviour of elementary hemp fibres. Part I:
651
Investigation of the possible origins using repeated progressive loading with in situ microscopic
652
observations. Composites Part A: Applied Science and Manufacturing. 2014;56:319-27.
653
[49] Trivaudey F, Placet V, Guicheret-Retel V, Boubakar ML. Nonlinear tensile behaviour of
654
elementary hemp fibres. Part II: Modelling using an anisotropic viscoelastic constitutive law in a
655
material rotating frame. Composites Part A: Applied Science and Manufacturing. 2015;68:346-55.
656
[50] Khodayari A, Thielemans W, Hirn U, Van Vuure AW, Seveno D. Cellulose-hemicellulose
657
interactions - A nanoscale view. Carbohydrate Polymers. 2021;270:118364.
658
[51] Hendrickx K. Extraction optimisation for and hygroscopic behaviour of flax fibres in composite
659
applications: KU Leuven. 2019.
660
[52] Thygesen A, Madsen B, Bjerre AB, Lilholt H. Cellulosic Fibers: Effect of Processing on Fiber
661
Bundle Strength. Journal of Natural Fibers. 2011;8:161-75.
662
[53] Corbin A-C, Sala B, Soulat D, Ferreira M, Labanieh A-R, Placet V. Development of quasi-
663
unidirectional fabrics with hemp fiber: A competitive reinforcement for composite materials. Journal
664
of Composite Materials. 2020;55:551-64.
665
[54] Corbin A-C, Soulat D, Ferreira M, Labanieh A-R, Gabrion X, Malécot P, et al. Towards hemp
666
fabrics for high-performance composites: Influence of weave pattern and features. Composites Part B:
667
Engineering. 2020;181:107582.
668
669
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Increasing the production of high-performance natural fibres that minimise their impact on the environment is a challenge that flax (Linum usitatissinum L.) cannot address alone. In flax traditional production territories, hemp (Cannabis sativa L.) can be a complementary source of high added value fibres if their yield of long line fibres can be maximised to levels equivalent to the one of flax. The objective of the present work was to establish process parameters maximising the long line fibre yield using flax dedicated scutching and hackling devices. A lab-scale scutching/hackling device was used to establish sets of process parameters which best improve the long fibre scutching yield and as a consequence minimise the production of tow fibres. Decreases in straw processing transfer and beating speeds during scutching were necessary so that to be less aggressive on the straw and fibres. Very high long fibre yields were obtained after scutching and hackling at the laboratory scale (18 % of the hemp straw mass). These very high results, combined to high straw yield production in the field indicate that hemp can be a very productive source of high-performance fibres as these ones showed tensile properties completely suitable for a textile use as well as for load bearing composite materials. If the potential of high production yields and high mechanical and morphological properties was demonstrated at the lab-scale, this one should be improved at the industrial scale. Suggestions to reach this goal are provided to prevent too high transformation of long fibres into tows and to keep the mechanical potential maximum. When using optimised parameters and a lab-scale scutching/hackling device, it was demonstrated that hemp has the potential for providing equivalent amounts of long fibres per hectare than flax with tensile properties about 20 % lower than the ones of flax.
Article
Full-text available
In this work, we study interactions of five different hemicellulose models, i.e. Galactoglucomannan, O-Acetyl-Galactoglucomannan, Fuco-Galacto-Xyloglucan, 4-O-Methylglucuronoxylan, and 4-O-Methylglucuronoarabinoxylan, and their respective binding strength to cellulose nanocrystals by molecular dynamics simulations. Glucuronoarabinoxylan showed the highest free energy of binding, whereas Xyloglucan had the lowest interaction energies amongst the five models. We further performed simulated shear tests and concluded that failure mostly happens at the inter-molecular interaction level within the hemicellulose fraction, rather than at the interface with cellulose. The presence of water molecules seems to have a weakening effect on the interactions of hemicellulose and cellulose, taking up the available hydroxyl groups on the surface of the cellulose for hydrogen bonding. We believe that these studies can shed light on better understanding of plant cell walls, as well as providing evidence on variability of the structures of different plant sources for extractions, purification, and operation of biorefineries.
Article
Full-text available
Given the vast amount of available research in the area of natural fibre composites, a significant step forward in the development of next-generation plant fibre-based products would be to devise a framework for rational design. The authors use a top-down approach, starting with an example final product to define the product specifications for high-performance hemp fibre-reinforced composites. Thereafter, all process steps are critically analysed: from textile preform and reinforcement yarn production, to fibre extraction and the agricultural process chain, to the microbiology of field retting, to cultivation and selection of crop variety. The aim of the analysis is to determine how far the current state of knowledge and process technologies are in order to use hemp fibres in high-performance composites. Based on this critical evaluation of the state-of-the-art, it can be stated that hemp will be found in high-performance composites in the short-to-medium term. There is, however, a need for performance optimisation especially through the selection of crop variety, best practices in retting, and effective fibre extraction methods to obtain more consistent fibre qualities suitable for reinforcement spinning and composite preform manufacturing processes.
Article
Full-text available
The textile industry is one of the highest polluting industries in the world. Recent studies have explored the introduction of environmentally friendly textiles to address this issue. One of these textiles is fiber derived from industrial hemp, which was recently approved for growth in the United States through the 2018 Farm Bill legislation. Motivated by hemp's potential to have a lower ecological footprint than cotton, the objective of this study is to determine if industrial hemp fiber can be produced in an economically competitive manner. Through the lenses of sustainable development and systems engineering, the basic design of the research assesses material selection decisions economically by taking a holistic supply chain view of the agricultural activities associated with industrial hemp compared to its largest competitor (i.e., cotton). With both fibers being comparable in performance, the production process of both textiles is juxtaposed, to account for interdependencies among stages with key economic and environmental considerations. As the economic cost of agricultural activities for hemp is currently uncertain in the United States, our methodology considers four main data inputs to capture the agricultural activities. First, fertilization costs are regarded as part of the cost associated with field preparations. Second, we assess the seed costs associated with cultivation. Third, the cost of irrigation (i.e., water consumption) and fourth, pest control cost represents the cost of field operations. These costs, combined with fiber yield, are used to estimate and compare the two fibers in USD per metric ton of final fiber produced. Industrial hemp is a high yield crop with (on average) 3 times more metric tons of fiber produced per hectare cultivated. Therefore, the adoption of hemp enables a reduction in cost associated with agricultural activities of 77.63%, when compared to cotton for medium total agricultural activity cost and medium yield estimates. In summary, our results suggest that industrial hemp fiber is economically viable and has the potential to be a more environmentally friendly alternative material than cotton within the textile industry.
Article
Full-text available
In this study, the back calculated compressive properties of flax fibers utilizing the Impregnated Fiber Bundle Test (IFBT) were investigated. The back calculated stress-strain response can be described by the Ramberg-Osgood model. The back calculated compressive modulus of the fiber is similar to its back calculated tensile modulus. The compressive strength of the fiber is approximately 45 % of its tensile strength. Considering the presence of local fiber kinking within the elementary fibers as well as global fiber kinking due to fiber misalignments and plastic shear deformation in the matrix material, this is a remarkably high value for the compressive strength. Our results indicate that local fiber kinking precedes global fiber kinking. We show that IFBT is a promising method for determining the compressive properties of flax fibers and provides necessary input data for finite element analysis of the compressive failure mechanisms in unidirectional flax fiber reinforced composites.
Article
Retting of fibrous plants such as flax is an essential step in the extraction of fibre bundles and their transformation into textiles and reinforcement fabrics for use in garments and composites. Dew-retting is traditionally performed from Northwest France to the Netherlands, as the climate is highly favourable for this process. Hemp is a plant that can be grown almost all over Europe with a low environmental impact. A retting step is also required to facilitate the separation of the hemp fibres before their transformation into textiles for garments or for 1D to 3D reinforcement composites, which requires thoroughly separated fibres. Dew-retting is currently used in flax production zones. The present work demonstrates that dew retting can be conducted under different climates on different soils, from dry Mediterranean environments up to the cooler regions of eastern France. If the ternary combination of moisture, temperature and solar radiation is appropriate, field retting (dew-retting) can be as short as about three weeks. In less favourable conditions, such as in dryer areas or when retting is performed late in the season after seed maturity (cooler temperatures), it lasts longer, but it can reach suitable levels. When conducted with care and with proper monitoring of the retting level, the dew-retting process does impact neither the tensile properties of elementary hemp fibres (by degrading crystalline cellulose I) nor the tensile properties of unidirectional and injected composite materials. Consequently, if extracted with a suitable process such as scutching and hackling, fibres suitable for load-bearing composites can be produced from dew-retted hemp stems produced in a wide range of climates and locations, therefore not limited to the conventional “dew retting zone” of flax production areas.
Article
Industrial hemp (Cannabis sativa L.) has great potential as a sustainable source of textile fiber; yet, to develop a viable European hemp-for-textile chain, agronomic practices and primary processing need optimization to current industrial standards. A straightforward approach is to process hemp using existing, modern equipment for flax (linen). Here we extensively evaluated the quantity and quality of fiber extracted from field-retted hemp stems, scutched on the industrial flax processing line. Varieties from diverse European origin (USO 31, Dacia Secuieni, Bialobrzeskie, Futura 75, Carmagnola Selezionata, Santhica 27 and Santhica 70) were sown in ran-domized field experiments in Belgium, which has a rich long-standing tradition in high-quality linen production. Biomass yield and the quantity of long fiber processed were assessed across three growing seasons (2017-2019; plot size: 15-45 m 2). In 2018, we also determined the quantity of tow (short fiber) and, the quality of long fiber in terms of fiber tenacity and elongation. The quantity of total fiber extracted (i.e. long fiber plus tow) accounted for 36.2% of the initial straw yield, indicating high processing efficiency. Approximately equal amounts of tow and long fiber were extracted. Mean long fiber yield approximated one ton per hectare; yet yield variation between varieties was considerable (range long fiber yield: 0.6-1.4 ton/hectare). Despite significant variation between harvest years in straw yield, the quantity of long fiber extracted held relatively constant. Fiber tenacity of long hemp was overall high and comparable to flax (range: 37.6-45.3 cN/tex). Results indicate that field-retted hemp has potential to be processed into quality fiber on the industrial flax line and, that fiber yield can likely further be improved by genotype selection. Harvest mechanization, focused on the collection of parallel hemp stem portions of appropriate length for the flax scutching line (ca. 1 m), seems warranted to make this approach economically viable. Additional research on the fiber properties following hackling and wet-spinning will be needed to fully explore the potentiality of long hemp as a flax supplement for textile applications.
Article
This study focuses on the development and mechanical testing of a quasi-unidirectional woven hemp fabric for composite applications. The fabric is designed to combine the advantages of easy handling, impregnability by resin and fiber alignment. In this sense, low twisted rovings, low crimping characteristics and a high roving density in the main direction were used. The results show that the manufacturing process proposed for this highly unbalanced woven fabric is less aggressive than that for a balanced pattern, and the effective tensile properties of the fibers are preserved. This innovative hemp fabric is also used to manufacture unidirectional and cross-ply laminated composites. The results indicate competitive properties relative to those of commercial flax-based reinforcements with rigidities similar to those of composites made with quasi-unidirectional woven fabrics, non-crimp fabrics and tapes.
Article
In western Europe, hemp is mainly cultivated for the seeds and the fibre rich straws, randomly aligned following the harvesting with a combine harvester. The straws are mainly valorised for low added value products such as pulp for paper and for insulation. They are also valorised in low property short fibre composites. With the view to use hemp fibres extracted from randomly aligned straws for higher added values, this work proposes to study a different process (fibre opener) for extracting the fibres than the one which is classically used in the industry (hammer mill) and to investigate the extraction performances and the impact of the process on the hemp morphological and mechanical properties. The morphological and mechanical properties measured at different moments of the extraction indicate that, even in the less favourable case, the length of the fibres (∼ 5 cm) is high enough for textile processing via the carded route. The tensile strength (∼ 660 MPa) and modulus of elasticity of (38 GPa) of the individual hemp fibres are situated above most of the ones of fibres extracted mechanically using a hammer mill (630 MPa and 25 GPa for strength and modulus respectively) despite the fact that large quantities of kink band defects are observed on each fibre. This therefore shows that the “all fibre” opener is suitable for the production of fibres that can be considered for the manufacturing of technical textiles such as load-bearing woven geotextiles or midrange load-bearing composite reinforcements.
Article
Recent developments in the field of bio-based composite materials are mainly focused on the use of unidirectional reinforcements. The production of woven fabrics and required yarns or rovings is still complex for composite applications due to the finite length of plant fibers and to the high number of process parameters which can be tuned. This study focused on the influence of weave pattern and process parameters on the resulting material properties at different scales. Results from mechanical characterizations and X-ray nanotomography show that very competitive tensile properties can be obtained for woven hemp fabric composites made from low-twisted rovings, in particular when compared to the front-runner flax cross-ply laminate. The authors wish to thank the Italian Company “Linificio e Canapificio Nazionale” for providing the hemp rovings used in this study. This project has received funding from the Bio-Based Industries Joint Undertaking under the European Union’s Horizon 2020 research and innovation program under grant agreement No 744349 (Ssuchy Project).