1 Introduction

Fiber-reinforced plastics (FRPs) are used in a wide variety of engineering structures, especially for those where high mechanical performance under extreme loading combined with low density is desired (Ref 1, 2). These structures consist of the stacking of different plies of the same or different materials that can be oriented at different angles. This allows a high level of product and process flexibility with the production of tailorable product properties, optimized mechanical performance, and weight reduction. In addition, for some specific applications such as cryogenic tanks of H2 or turbojet engine fan blades, another important characteristic is the possibility to control the desired product thickness during the manufacturing without additional process steps, like rolling, compared with conventional structural materials such as metals.

These aspects highlight two relevant key points for FRP composites:

  • The possibility to produce hybrid structures by stacking plies of different materials. The composites resulting from such interaction are known as fiber–hybrid composites (FHCs), and their main advantages are to alleviate the inherent drawbacks of each of the fiber types and keep the benefits of both, while enabling synergistic effects that can lead to properties that neither of the constituents possess (Ref 3);

  • The possibility to obtain thin composite laminates that can be folded to very high curvatures. These structures are commonly used in deployable space structures because the laminates are elastically folded to allow a structure to be compactly stowed during launch and subsequently deployed to an operational state once in orbit. This basic technology enables operational systems containing structures such as antennas, solar arrays, reflectors, and booms to be efficiently packaged and launched within the payload envelope of conventional launch vehicle fairings (Ref 4, 5).

Considering the current environmental world crisis, new generation of composite systems have to minimize the environmental impact with an efficient use of energy resources and materials, waste management, and where possible replacing synthetic or petroleum materials with more eco-sustainable components (Ref 3, 6, 7). In this context, the use of vegetable fibers for producing natural fibers-reinforcing plastics (NFRPs) appears to be an interesting challenge that has gained the attention of the scientific community. Several advantages associated with the use of natural fibers justify this interest: relatively high specific strength and stiffness, low density, low emission of toxic fumes when subjected to heat and during incineration at end of life, lower manufacturing cost than synthetic fibers, less tool damage during processing, and the availability as a renewable resource (Ref 8).

Different vegetable fibers are available in nature (cotton, hemp, flax, jute, ramie, sisal, bamboo, etc.), but only few of them are attractive for structural applications. Among them, preliminary studies have demonstrated that hemp fibers are one of the most promising candidates for replacing traditional synthetic materials for the following reasons: They are characterized by good mechanical properties, high cellulose content, and lower production costs than other vegetable fibers like flax, they are easily grown on a wide range of soils, and they possess high biomass yield, high carbon storage potential, and low eutrophication (Ref 9). In addition, industrial hemp appears to be one of the most profitable crops since it produces an amount of oxygen during growth which is equal to the amount of CO2 released when burning the same amount of hemp. Not surprisingly, the hemp production increased of 62.4% from 94,120 tons in 2015 to 152,820 tons in 2019. Based on these considerations, hemp fibers can be successfully used to obtain biocomposites (Ref 10) and have attracted the attention not only of several researchers but also of multiple industrial sectors that have started looking at them as a potential solution to join performance and sustainability. Despite these interesting aspects, the mechanical properties of vegetable fiber like hemp cannot be compared with those of high-tech fibers like carbon and therefore currently NFRPs are mostly used for non-structural applications or to replace glass fibers. A possible solution for overcoming this limitation is their use in hybrid system configuration as demonstrated by the large number of publications on this topic in the literature (Ref 3, 11,12,13,14,15,16). Some applications are also available in the literature; for example, in the study of Amiri et al. (Ref 17) it was pointed out that it is possible to manufacture and use a carbon/flax hybridized bicycle frame ensuring the required performance.

Most of these works focus on the study of mechanical properties of vegetable/carbon hybrid laminate by varying the stacking sequence. For instance, Sarasini et al. (Ref 18) proved that flax/carbon hybrid composites show enhanced impact absorption performance when compared with traditional carbon fibers laminates, especially when the vegetable layers were positioned on the external laminate faces. Similar considerations were carried out by Pinto et al. (Ref 3) that investigated the mechanical properties and failure mechanism of hemp/carbon laminates. Other works focus on the use of vegetable fibers as core in sandwich structures using synthetic fibers as skins. These cores were produced using flax, as in the work of Sarwar et al. (Ref 19), or using hemp fiber composites, as in our previous work (Ref 11), and the results in terms of mechanical properties are very promising. On the basis of these considerations, it is clear that the hybridization of vegetable with synthetic fibers can open new engineering scenarios and future studies about hybrid composites in terms of manufacturing, improving properties, and product applications will be welcomed. Concerning this aspect, it appears interesting the study of the hybridization in thin composite structures that appeared to be an un-investigated field. Therefore, aiming to fill this gap and further extend the use of hemp fibers for structural applications where thin carbon composite structures are used, the aim of this work was to produce hybrid thin laminates combining the ductility of hemp fibers with the strength of carbon fibers developing a hybrid system and investigate their flexural and impact performances.

2 Materials and Methods

2.1 Materials

Carbon/hemp hybrid thin composite laminates under investigation were produced in sandwich configuration, where the hemp vegetable layer was located in the middle plane. Two types of woven hemp fabrics with a mean thickness of 1.2 mm (supplied by Fidia Srl) that differ in the areal density were adopted for this purpose: one characterized by an areal density (GMS) of 380 g/m2 (labeled H) and the other by a GMS of 190 g/m2 (labeled L), both shown in Fig. 1. As carbon fibers, a woven carbon fabric with a mean thickness of 0.25 mm and an areal density of 200 g/m2 (supplied by Toray International Srl) was used, while the matrix used for the manufacturing of the thin laminates was the SX10 epoxy resin supplied by Mates Srl.

Fig. 1
figure 1

Magnification of H (a) and L (b) woven hemp fabric

Full size image

2.2 Composite Manufacturing

Hybrid thin composite laminates consisting of one central impregnated hemp fabric and two external carbon fabrics were manufactured in three steps: First, the vegetable composite layers are prepared using an ad-hoc manufacturing process, then, the carbon layers are impregnated with the epoxy resin and layered in the desired configuration, and finally, the thin laminate is subjected to a heath treatment to activate the curing reaction. In detail, the production system used to produce the bio-based vegetable layer (Ref 11) is schematized in Fig. 2, and it is divided in the following steps: impregnation, resin content reduction, resin absorption, and cutting.

Fig. 2
figure 2

Schematization of the production system of the hemp plies

Full size image

In the first step, a caterpillar apparatus (localized at the end of the production system) pulls and forces the vegetable fabric in the impregnation resin system consisting of two silicon belts and a resin container. The resin falls on the fabric that is impregnated passing between the belts. At this point, the impregnated fabric passes under an air jet flow that blows through the mesh of the fabric and then between a couple of rolls covered with absorbent paper layers; thanks to the large mesh that characterizes both the hemp fabrics, the air flow is able to remove the excess of resin between the spaces reaching the desired fiber content. By controlling the amount of resin in the resin bath and by tailoring the resin removal process, the ratio between the weight of the impregnated and unimpregnated layers was kept constant to 2.5 and the fiber weight fraction content of each layer at almost 40%. Then, the final areal density of impregnated hemp fabrics was within the range of 475 and 950 g/m2. Finally, the impregnated fabric was cut in layers 400 × 300 mm.

Simultaneously, carbon fabrics (400 × 300 mm) were impregnated and the hybrid composites were manufactured by the combination of hand lay-up and vacuum compression molding technique. Impregnated fabrics were placed in a mold in sandwich configuration with the hemp fabric in the midplane, as schematized in Fig. 3. At the end of the stacking phase, the mold was sealed in an elastomeric bag under vacuum and placed in a hydraulic press under a pressure of 8 bar at 55 °C for 2 h, and then, the laminates were left to cure at room temperature under a pressure of 8 bar for 24 h.

Fig. 3
figure 3

Hybrid sample type schematization

Full size image

In addition to the hybrid thin laminates, pure carbon laminates were manufactured and tested as reference. In detail, two types of carbon laminates were considered, the first one characterized by the same number of carbon fabrics of the hybrid samples (i.e., two plies) and another type with an increased number of carbon plies (i.e., four plies) to obtain samples with the same weight of the heaviest hybrid sample type (i.e., the same areal density). Therefore, in each test, CHC and C4 specimens have the same weight. All details about the sample types under investigation are reported in Table 1. For evaluating the volumetric fiber content, fiber density values of 1.5 and 1.8 g/cm3 for hemp and carbon fiber were, respectively, considered. Due to the considerable difference between the thickness of carbon and hemp fabric and also considering that (different from carbon) the hemp tow, for its intrinsic characteristics, does not show a significant thickness reduction, it is justified the different thickness between pure carbon and hybrid laminates as listed in Table 1.

Table 1 Main laminates properties
Full size table

2.3 Mechanical Characterization

The mechanical characterization of the samples was achieved via means of low-velocity impact (LVI) and flexural tests. In detail, LVIs were carried out according to ASTM D7136 standard, using a drop-weight tower apparatus equipped with a 15-mm hemispherical tip and moving on a shuttle with a mass of 2.66 kg. As per the standard employed, five samples per typology were tested, with their dimensions being 150 × 100 mm, at two selected impact energy levels of 2 and 5 J. In order to select the testing impact energy levels, preliminary penetration impact tests up to 10 J were carried out according to the procedure reported in (Ref 20,21,22,23,24,25). These tests were conducted on C2 and C4 samples as these were considered the baseline against which the hybridized samples should be compared.

All samples were clamped between two steel plates to avoid any movement during testing, and the energy levels were achieved by adjusting the height from which the shuttle was released. Finally, Infrared thermography was used to analyze the extent of the internal damage using a Hensel TRIA 6000 S generator (rated at 6000 J energy), two Hensel EH Pro 6000 S flash heads (50% each of the total energy), and a Longwave IR Xenics Bobcat 640 GigE IR camera with a resolution of 640 × 512 pixels and a max frame rate of 50 Hz.

The images were then postprocessed with a MATLAB routine in order to have an estimation of the extent of the internal damage.

Three-point bending tests were performed according to ASTM D790 standards, using a universal testing machine (MTS Alliance RT 50 equipped with a 1-kN load cell). The flexural tests were performed according to ASTM D790-17 standard on a total of five specimens for each sample configuration. Following the standard, in order to test the samples in pure bending, different span lengths were adopted; in particular, sample types C2 and C4 (materials less than 1.6 mm in thickness) were tested using a span length of 25.4 mm, while CHC and CLC configurations were tested using a span-to-thickness ratio of 16:1.

3 Results

3.1 Impact Tests

The results of preliminary tests on the C2 and C4 samples are reported in Fig. 4. In detail, it was graphically evaluated the energy required for damage and perforation of the C2 samples when these were subjected to an impact energy of 5 J. It has been well established in the literature (Ref 20,21,22) that in order to evaluate the damage energy in laminates subjected to low-velocity impact, these must have gone complete perforation, with partial penetration not being adequate to provide the necessary information for the suggested values. A force–displacement graph overlapped with the evolution of absorbed energy over displacement is illustrated in Fig. 4(a) where an evaluation of the damage energy can be made. As the impactor hits the target, a sharp rise to recorded forces is seen up until the damage initiation point as matrix failure occurs there; on the energy–displacement curve, the corresponding point indicates the required energy to induce this damage on the laminate. On this occasion, the energy required to induce damage on the C2 sample was evaluated at 1.8 J. Past this point on the energy—displacement curve (orange line) and up until the values of energy remaining unchanged, it is considered energy consumed for propagation with matrix and fiber failures that take place at this stage. At the point where the perforation energy has been marked, the energy remains virtually unchanged due to complete fiber failure. The summation of these two energies is equal to the total amount of energy absorbed by the sample and used for perforation (Ref 23), which was evaluated approximately as 4.73 J on this occasion.

Fig. 4
figure 4

Typical force–displacement curve of: completely perforated C2 sample for test at 5 J (a) and C4 sample for test at 10 J (b). Orange curve indicates energy evolution over displacement until perforation. Points of intersections and interest are marked on the graph (Color figure online)

Full size image

Similarly, the damage initiation and perforation energy for the C4 samples were evaluated with impact tests at 10 J. As can be appreciated from the evolution of the curves and their intersection points in Fig. 4(b), the energy required to induce damage on the C4 samples was approximately 2.5 J, an increase of 25% compared to the C2 ones, with the perforation energy on this occasion at 9.8 J. Based on these results, it was set the first impact energy level at 2 J, a level which would induce damage at least on the C2 sample and would allow to compare its performance with the hybridized samples. The upper impact level, that of 5 J, was selected as this was approximately the energy threshold required for perforation on the C2 typology; higher energy levels were excluded as these would enforce to exclude the C2 typology from these, due to their inability to withstand loads at these energy levels.

Therefore, Fig. 5 and 6 shows typical force versus displacement curves and thermography images related to samples impacted at 2 and 5 J, respectively. Before the analysis of these results, a consideration needs to be done; as the impact energy levels were achieved by changing the release height of the shuttle, a change in the response of the samples could be expected. However, comparing typical curves carried out at 2 and 5 J on the same sample type, the initial response of the samples in the elastic range remained virtually the same with the small discrepancies observed attributed to filtering of the signal during the post-processing of the acquired data. A significant change can be observed for only the C2 typology though, which can be attributed to bending effects prior to complete perforation this typology sustained.

Fig. 5
figure 5

Typical impact curves (force versus displacement) carried out at 2 J of energy level for each sample type (a) and typical thermography images of impacted specimens (b)

Full size image
Fig. 6
figure 6

Typical impact curves (force versus displacement) carried out at 5 J of energy level for each sample type (a) and typical thermography images of impacted specimens (b)

Full size image

Looking at Fig. 5 and comparing C2 sample type with hybrid ones, it is possible to note that the use of an additional hemp layer between two carbon plies led to a considerable increase in the laminate stiffness which overcame also the rigidity achieved when four carbon plies were adopted (i.e., C4 sample type). The use of the high-density hemp fabric between two carbon plies (i.e., CHC hybrid sample type) allowed to reach a maximum force close to that of C4 (− 5.9%), which means that it can be pointed out that the hybridization process allows to replace two carbon plies with one hemp fabric ply in terms of maximum force. On the other hand, it has to be considered that all hybrid composite specimens show a load drop at the maximum displacement which testify the occurrence of damage. However, these defects appear to be only sub-critical cracks located exclusively on the impact side of the laminate (Fig. 7), while no trace of internal damage was detected with the thermographic inspection, as shown in Fig. 5(b). This means that 2 J was not enough to emphasize the hybridization effects by involving different absorption energy mechanisms in respect of pure carbon laminates. Indeed, as listed in Table 2 C4 sample type adsorbed less energy without the presence of a relevant damage, showing an almost pure elastic behavior.

Fig. 7
figure 7

Representative top surface of CHC sample type tested at 2 J

Full size image
Table 2 Mean values of LVI tests carried out at 2 J
Full size table

Differently, when the impacted energy was raised up to 5 J, the presence of the hemp layer allowed to reach more interesting results. Indeed, looking at Fig. 6 it is possible to observe that the presence of the hemp layer between carbon ones allowed to a significant improvement in the impact performances in respect of the pure carbon laminate with two plies (C2 sample type), which was instead completely perforated, offering almost no resistance to the impact load. Both the impact curves (Fig. 6a) and the thermography images (Fig. 6b) highlighted this aspect; indeed, the C2 curve presents a defined load drop in correspondence of the maximum load and the thermography revealed a well-defined sign of the penetrator tip. On the other hand, also the hybrid configurations showed a load drop when the maximum load is reached but they do not exhibit any sign of failure at penetration. Comparing the hybrid laminates with configuration C4, it appears that when the high-density hemp layer was adopted (CHC sample type), the elastic rigidity improved without sacrificing the maximum displacement while the maximum force decreased by about 13%. Looking at Table 3, it is also evident that although C4 and CHC sample types adsorb the same amount of impact energy, the hybridization allowed a small reduction in the internal damage size (− 6.9%). This is an interesting result since C4 and CHC are characterized by the same areal density; hence, they have the same weight and then directly comparable. Different results were obtained when the low-density hemp layer was adopted (i.e., CLC sample type), because the reduced hemp fiber content was not able to reinforce the midplane of the laminate by absorbing the impact energy maintaining at the same time a reduced damaged area. These results are highlighted in Table 3, where the CLC sample type, in comparison with the C4 and CHC samples, adsorbed a larger impact energy for the creation of new fracture surfaces. Therefore, even if the CLC sample revealed an increase in the damage area with respect to C4 and CHC typologies, it still showed better performance than the pure carbon layer with two plies (i.e., C2 sample type).

Table 3 Mean values of LVI tests carried out at 5 J
Full size table

3.2 Three-Point Bending Tests

Figure 8 shows the typical stress strain curves obtained using the span length suggested by ASTM standard, i.e., 25.4 mm for C2 and C4 samples and 28.8 mm for CHC and CLC samples. The stress and the strain were calculated by Eq 1 and 2, respectively:

$${\sigma }_{f}=\frac{3FL}{2b{t}^{2}}$$
(1)
$${\varepsilon }_{f}=\frac{6ft}{{L}^{2}}$$
(2)

where F is the load, L is the support span, f is the deflection of the center of the specimen, and b and s are the width and the thickness of the specimen, respectively. The flexural elastic modulus, E, was evaluated using Eq (3):

$$E= \frac{{L}^{3}m}{4b{t}^{3}}$$
(3)

where m is the slope of the tangent to the initial straight-line portion of the load–deflection curve.

Fig. 8
figure 8

Typical flexural stress–strain curves

Full size image

The results highlight that although in this testing configuration both the hybrid solutions show lower flexural properties in terms of flexural strength and modulus in respect of pure carbon samples, the hybridization allowed to reach higher values of maximum strain. By observing the typical curves of the hybrid samples, it is possible to note the presence of a stress drop at around 1.4 × 10−2 mm/mm of strain. This is due to the failure of carbon ply on the compression side, but, contrary to what observed in pure carbon samples, the crack does not propagate throughout the entire thickness of the sample, but it stops when it reaches the hemp layers and bends at the interface between the carbon and the hemp layers resulting in a plateau. Therefore, after the stress drops, only the hemp and carbon ply of the bottom side support the load, resulting in an additional energy absorption contribution. As for the effect of the density of the hemp layer, no significant difference in the flexural strength was detected between CHC and CLC (this indicates that up to the maximum load, the flexural behavior was mainly conferred by the top carbon ply) but after the stress drop, the CHC sample showed a better resistance due to the presence of a hemp layer with an higher fiber volume fraction than CLC (this indicates that in the second tract of the curve, the flexural behavior was mainly conferred by the hemp and bottom carbon plies). The small difference in the maximum flexural stress between CHC and CLC can be attributed to the fewer points of contact between the hemp layer and the carbon plies that characterize CLC sample type. This is clear by observing the different fabric mesh size in Fig. 1. The results of the bending tests are listed in Table 4.

Table 4 Results of three-point bending tests for each sample type
Full size table

A further observation can be done regarding the correct use of Eq (1) to evaluate the stress for hybrid samples, as this equation might underestimate the true stress since it assumes that the entire laminate is made of carbon while in realty this accounts for only around one-third of the thickness. Considering the configuration of these samples, a more suitable estimation could be obtained by using Eq 4,

$${\sigma }_{f}=\frac{FL}{4(t-{t}_{f})bt}$$
(4)

where tf is the skin thickness that in this case corresponds to the carbon ply thickness. Equation 4 was extrapolated from ASTM C393, and it is the equation usually used for sandwich panels; however, according to this standard, it should be used only when the ratio between the skins (carbon ply in this study) and core thickness (hemp ply in this study) is lower than 0.1. Although this hypothesis is not satisfied in the present case and therefore Eq 4 would not be appropriate, Fig. 9 shows the stress–strain curves obtained by using both Eq 1 and 4. Another interesting aspect can be highlighted considering the comparison between force versus displacement curves carried out on longer span. To this scope, additionally three-point bending tests were carried out fixing a value of span length equal to 64 times the thickness of the thicker specimen, i.e., the hybrid ones. The results are reported in Fig. 10 from which it is possible to observe how in terms of force the use of the hemp layer might be beneficial: supposing a structure of fixed length, a hybrid structure, like CLC, not only would be lighter than the C4 carbon laminate but it would also support higher loads with less deformation, hence joining performance and sustainability.

Fig. 9
figure 9

Comparison between the typical flexural stress–strain curves for hybrid samples using Eq 1 and 4

Full size image
Fig. 10
figure 10

Typical load–displacement curves for tests carried out at 116 mm of span length

Full size image

4 Conclusions

This work investigated the possibility to produce thin hemp/carbon hybrid laminates (1.8 mm thick) using an ad-hoc manufacturing process for manufacturing impregnated woven hemp fabrics characterized by low weight (areal density ranging from 475 to 950 g/m2) and located at the midplane of carbon laminates. By comparing impact and flexural performances between hybrid and pure carbon laminates, it is possible to draw the following conclusions:

  • For the 2 J impact tests, the use of the high-density hemp fabric between two carbon plies allowed to reach a maximum force close to the pure carbon samples of the same weight but with an higher absorbed energy associated with the formation of small sub-critical cracks on the top surface that was not detected for pure carbon samples;

  • Comparing hybrid samples with pure carbon ones with the same weight, it appeared from the results of the 5 J impact tests that the hemp layer was able to improve the elastic rigidity without sacrificing the maximum displacement by reaching similar maximum force values of the C4 configuration. In addition, the hybrid sample type adsorbed the same energy with a smaller damaged area then pure carbon sample. This is an interesting aspect because, under this point of view, the hybridization allowed to replace two carbon plies with one hemp fabric ply.

  • The results of flexural tests proved that by comparing stress–strain curves the presence of a hemp ply in the midplane and the use of only one carbon ply on the external face did not allow to take advantage of carbon flexural resistances. However, it was difficult to perform a comparison in terms of stress due to the difference in the span length used in the tests and to the use of conventional equations to evaluate the flexural stress that could not be appropriate. However, by comparing force vs displacement curves obtained from tests carried out on the same span length, the hybridization allowed to reach higher loads and lower deformation, but lowered the total weight of the laminate in comparison with traditional carbon laminates.