Subsystem 1 involved hemp cultivation, which produced hemp grains and stalks in bales, as illustrated in Figure 2. The crop chosen for this study was planted on a 66-acre plot in 2023 and a 32-acre plot in 2024, both located on the main campus of the University of Waterloo (ON), Canada. This situation meant that some parameters could not be determined from primary data. This gap was filled by values gathered from the literature, which were similar to and consistent with those of the operations practiced. The agricultural process began with soil preparation for planting, incorporating cattle and/or horse manure.

This study excluded the environmental impacts of acquiring these primary macronutrients and organic matter sources, as this practice corresponds to repurposing livestock waste. Thus, only the consumption and emissions related to transporting materials for 29 km from the supplier to the field were considered. The application of manure and subsequent plowing were conducted in separate stages using agricultural machinery that operated on diesel fuel.

Ontario’s well-drained, slightly alkaline clay soils (7.0 < pH < 7.5) are ideal for growing hemp. However, additional N-P-K supplementation beyond what manure provides is necessary to enhance agricultural productivity. This was accomplished by dosing a compound chemical, monoammonium phosphate (MAP), with a (52–11–0) ratio. The environmental impacts of fertilizer processing were also modeled using data from the Ecoinvent^® Database 9.5, which was adapted to reflect the specific conditions under which they occur. As with the other inputs in (successive) hemp fiber processing, MAP was examined using a cradle-to-use approach.

Therefore, in addition to the application of this fertilizer, which involved plowing with agricultural machinery, the industrial production of this fertilizer was also taken into consideration in the analysis. A plowing stage using discs (discing) was carried out before sowing the seeds into the soil. Organic hemp seeds were sourced from Winnipeg (MB, Canada), located 1996 km from the cultivation site. Like the other inputs, they were transported by road.

The lack of consistent primary data on hemp seed production led to this stage of the Product System being modeled after organic corn production. According to records from Harvest New York (2017) [11], the replacement by the substitute is consistent and representative of the conditions under which processing occurs in Winnipeg (Canada). Sowing was carried out mechanically, with 33.6 kg of seeds per hectare being applied at this stage of cultivation.

The amount of CO₂ absorbed and stored in the hemp biomass during plant growth was determined using secondary data. For this purpose, seven studies published in the literature were utilized; the selection criteria were based on the similarity of the technological and operational aspects of the crops reported in the references to those practiced in Ontario. The CO₂ uptake rates achieved in each situation are presented in Table 1, along with their respective sources. This series estimated an average value of 1.69 kg of CO₂ fixed per kg of biomass produced. The stages following plant growth—grain and stem harvesting and baling—were performed with the assistance of agricultural machinery.

Since the harvesting process yielded two products, it established a multifunctional situation. In conceptual terms, the allocation procedure should address cases of multifunctionality in LCA applications within the attributional modality [34], as in the present case. The application of this approach will be discussed later in Section 2.2.

During harvesting, the biomass may be left in the field in windrows for a few days or weeks to facilitate the retting process. Retting may be used as a method to reduce the energy required for the traditional decortication of long hemp fibers, which is typically achieved using conventional methods with hammer mills. However, the technique used here was microdecortication. This will be further discussed in the next section. The GHG emissions of this operation were negligible when verified using a cumulative mass contribution criterion, MCC = 1.0%, which was defined for data treatment during Product System modeling [34]. The agricultural stage concluded with the formation of hemp bales.

The next phase of the production arrangement involves transforming hemp bases into pellets (Figure 3). Subsystem 2 essentially consists of three operations: milling, pelletization, and packaging. The milling operation involves particle size reduction, taking a bale of hemp stalks as input to produce short hemp fibers.

The method of particle size reduction (milling) used here is based on newly invented technology known as microdecortication [26]. It produces short hemp fibers (less than approximately 2.0 cm or smaller) utilizing the entire hemp stalk from the bale. Microdecortication utilizes cutting mills, which produce a high yield of short hemp fibers (>85%), and the use of retting may not be necessary. This is fundamentally different from the conventional method of decortication (using a hammer mill), which separates long hemp bast fibers (usually more than 20 cm long, approximately 20% (w/w)), hemp hurds (about 70% (w/w)), and dust. The newly invented method of microdecortication is far more efficient because the production yields of short fibers are in the range of 85% or above. It has already been demonstrated in a pilot plant (operating at less than 100 kg per hour) at the Tangho Green Canada Technology Center in Waterloo, ON, Canada; however, it has not yet been implemented in practical, commercial terms (hundreds of kilograms per hour). Therefore, this compelled us to use secondary data obtained from mass and energy balances conducted under assumed regular commercial operating conditions.

The second operation of Subsystem 2 is pelletization. This operation utilizes equipment that compresses biomass inside a die to produce a compacted pellet. An advantage of handling short hemp fiber pellets instead of losing short fibers is a reduction in bulk density.

This strategy lends a prospective character to the environmental diagnosis made in this study for Subsystem 2. The results obtained by applying an LCA can inform Design for Environment (DfE) actions, aiming to develop the hemp pellet production process to minimize impacts related to the Global Warming Potential.

All flows entering and exiting the Product System boundaries were designed to handle 200 kg/h of hemp bales. The unavailability of specific consumption and emission values necessitated their determination through mass and energy balances, considering the conditions and performance rates outlined in the literature. Figure 4 illustrates a process flow diagram, emphasizing the quantities of the main flows.

In this stage, the bales produced in Subsystem 1 are transported by road to Processing Site 1, located 1000 km from the plantation. The first operation is named cutting, which aims to disassemble the large bale into chips (>5.0 cm). The cutter’s reception, storage, and feeding operations are performed manually. The cutter converts the hemp bale into short stalks and chips, which are transported using a conveyor belt. Sorting and sieving separate waste mixed with biomass (stones and sand), which accounts for approximately 2.0% of the total mass of the bales. The fines generated from hemp stalks during sorting are collected for later reuse.

The sorted material is comminuted in successive stages of particle size reduction and fractionation (presented as cutting, classification, and milling in Figure 4), leading to the generation of short fibers with a desirable particle size distribution. The short hemp fibers are subsequently transferred for pelletization. Comminution also produces fine particles, which are combined with those generated during sorting, after being captured by a system that includes a collector and a bag filter, to be reintroduced into the pelletizing process. Pelletizing is conducted wet, and a binding agent may be added to enhance the process efficiency, depending on the specific conditions. For clarification, the pellets produced in Subsystem 2 contain short hemp fiber, fines, moisture, and a small quantity of additives. Note that the term “pellet” is used for the products of Subsystem 2 and Subsystem 3, but they differ from each other. Finally, the hemp pellets are cooled and bagged in polypropylene bags with a storage capacity of 2.0 m³ and an estimated unit mass of 2.5 kg.

The energy balance of Subsystem 2 was conducted based on the nominal capacities of the equipment utilized in the process arrangement, using data and information provided by Tangho Green Canada Inc. The differences between the dimensions of the values that guided the preparation of Subsystem 1 and those reported by the company were addressed by linearizing the latter set to align with the processing flow defined for the analysis (200 kg/h). The electrical energy consumed by each piece of equipment is shown in Table 2.

Subsystem 3 comprises the manufacture of a hybrid composite (Figure 5). This material is produced by extruding a mixture that includes the pellets processed in Subsystem 2, glass fiber, and polypropylene resin. The composite has been made on a pilot scale (several kilograms per hour). Still, it has not yet been manufactured on a large commercial scale, thereby allowing for a prospective assessment of Subsystem 3’s environmental performance. These results can also inform future Design for Environment (DfE) actions aimed at reducing its Global Warming Potential.

The design of the hybrid partially natural composite was founded on a fundamental premise: its technical performance should be compatible with that achieved by the existing substitute in the applications where those materials are utilized. The automotive sector, where polypropylene composites are often employed to manufacture interior and exterior parts for the reinforcement and protection of molded structures, has proven to be a compelling scenario for validating this requirement because their mechanical properties and specifications are rigorously evaluated before they become vehicle components.

Consequently, parameters such as the tensile and flexural moduli, impact notch, and density were established as criteria for selecting the natural composite. Following extensive research, we developed a hybrid composite comprising 70% (w/w) polypropylene and equal parts (15:15)% (w/w) of glass fiber and hemp pellets. If successful in both technical and environmental domains, this variant could replace a similar material used by automakers to produce car parts, whose formulation consists solely of polypropylene and glass fiber in a [70:30] mass ratio.

Table 3 illustrates the average performance obtained by both materials during the analysis. These results confirm the initial expectation that the mechanical properties of the alternative composite would be maintained despite 50% (w/w) of glass fiber being substituted with short hemp biomass. Consequently, the alternative composite became the product of Subsystem 3, and this step in the study focused on it. Furthermore, its environmental validity regarding the GWP could be compared to the performance of the existing composite, which was used as a reference for analysis.

This stage of the Product System began by transporting the three active ingredients in the formulation to the unit where the composite is processed. The plant is located 500 km from the hemp pellet and glass fiber supply depots and approximately 100 km from the PP resin distributor. Composite processing was modeled exclusively on the extrusion and pelletization of hybrid composites (final gate) but did not include injection molding. Figure 6 displays the values of the material streams used to estimate the Global Warming Potential of the process. A TSE-95 extruder from Nanjing Haisi Extrusion Equipment Co., Ltd. (Nanjing, China) was selected as a representative of the compounding stage. It features a twin screw, a variable load capacity of 800 to 1200 kg/h, and an electrical power of 315 kW, which could be used to estimate its environmental performance.

The selected TSE-95 model is also compatible with the anticipated installation size for producing the natural composite. The extruder feeding and product packaging were lumped with extruder inputs, which means that the environmental loads associated with these processes were no longer included in the environmental performance analysis.

The determination of the Global Warming Potential of the production of hybrid composites containing short hemp fibers, along with their intermediates—hemp in bales and hemp pellets—through the LCA was based on the methodological standardization outlined in the ISO 14040 and 14044 standards [34,35]. The technique was applied in an attributional modality and followed a “cradle-to-gate” scope. The intensities with which the transformation steps of the composite processing were executed were directly linked to the properties that this production good must demonstrate to fulfill the technical requirements for serving as a structural material in automotive panel manufacturing. Given its linear arrangement, the same conditions also guided the production of the intermediates. Similarly, a comparable approach was used to produce the reference composite made of PP and glass fiber.

The reference flows (RFs) adopted in each case to conduct the LCAs included 1.0 kg of hemp stalks in bales (Subsystem 1), 1.0 kg of hemp pellets in bags (Subsystem 2), and 1.0 kg of hybrid composite (Subsystem 3 and reference arrangement). As mentioned in Section 2.1, this study used secondary data. These data were either collected from the technical literature and datasets in the Ecoinvent^® Database 9.5 or provided by companies in the hemp fiber production and manufacturing sector. Following the guidelines of ISO 14044 [35], the modeling of the subsystems was carried out, considering the quantitative criteria of the cumulative mass contribution (CMC) and cumulative energy contribution (CEC) to exclude environmental aspects. For the arrangements under analysis, CMC and CCE values of 1.0% were applied. Additionally, the consumption of materials and energy, as well as emissions, below these limits was evaluated for its potential environmental relevance [35] and reintegrated into the analyses when necessary.

The environmental impacts associated with polypropylene resin and glass fiber production were quantified using secondary data from the Ecoinvent^® Database 9.5, which were adjusted for the research conditions. The same strategy was applied to the assets comprising the operational recipe of the analyzed composites, as well as the process utilities (e.g., electrical and thermal energy and water) and transport operations.

Regarding geographic coverage, the study encompassed the Canadian provinces of Manitoba, where hemp seeds are produced; Ontario, the region where the plant is cultivated; and a location 1000 km away from Waterloo, where the pellets and composites are made. We also assumed that the reference composite would be manufactured in the same areas to avoid disparities and distortions in the results. This study’s temporal coverage primarily consisted of 2023. The exception was the diesel oil consumption of the agricultural machinery and means of transport, which was recorded for 2019, as this period typically represented such demands. Finally, the technological coverage was specified based on the technological and technical aspects of each subsystem, as described in Section 2.1. Lastly, concerning utilities, data and information about the electricity sources required for the production processes were gathered from the Ecoinvent^® Database 9.5 and updated for temporal coverage. For the fuel used in the process, estimates of the consumption, origin, destination, and import volume were cross-referenced with data from the Canada Energy Regulator 2024 [36] and customized for the locations included in this study.

Whenever a given operation (a process or a set of processes) receives various inputs, generates co-products (outputs) that are intended for potentially different functions, or is even connected to a more extensive arrangement such as open-loop recycling (OLR) that serves more than one use simultaneously, the LCA practitioner will be faced with a multifunctional situation. Cases like this should be treated to distribute the environmental loads associated with the multifunctional transformation, and cases with processes involving co-products should be treated to distribute those generated in the operations before it [37]. The allocation procedure is adopted with artifice for attributional LCA studies to treat multifunctionalities. In general terms, this consists of applying a partition factor estimated from physical relationships or, when this is not possible, applying one estimated from the economic value of the inputs (or co-products) that coexist with the environmental loads related to the process under treatment [38].

In this study, this situation arose in the agricultural stage (Subsystem 1), specifically during harvesting, when hemp stalks and grains were produced. We decided to investigate the effects of the multifunctional situation between these co-products by applying the physical mass criterion. Consequently, the environmental loads generated were divided proportionally to the quantities of hemp stalks and grains obtained in this process stage.

Given the objectives of this study, we adopted the method proposed by the Intergovernmental Panel on Climate Change (IPCC) version 1.02 for 100 years, including CO₂ uptake [39], to quantify the Global Warming Potential associated with each analyzed situation. As the title indicates, the method considers the CO₂ absorption rates of the photosynthetic processes of various anthropogenic crops throughout their growth cycles.

Additionally, it addresses biogenic emissions related to the natural carbon cycle and those arising from anthropogenic activities, such as combustion, harvesting, digestion, and fermentation, which lead to the decomposition or processing of bio-based materials. Biogenic CO₂ is produced when the carbon in biological matter reaches its highest degree of oxidation due to these interventions [40,41]. The CO₂ absorption and emission rates during a short cycle, without deviations, will be complementary, resulting in a net zero impact on the GWP. However, this balance can be disturbed if biogenic carbon is sequestered by a product intended for use in another life cycle or when it is emitted in its reduced form (e.g., methane). Regardless of the established condition—equilibrium or instability—the IPCC 2021 GWP100 (including CO₂ uptake) can estimate the effects of these deviations in terms of the GWP [39].

In addition to the requirements for defining the scope outlined in Section 2.2.1, further assumptions were established to model the subsystems under analysis. These are detailed below:

The analysis excluded environmental loads from the production of capital goods and labor usage because their contribution rates were minimal, and the data associated with them were significantly uncertain.

After several adjustments, the subsystem models were found to be suitable for a prospective environmental performance assessment, as planned for a product still in the design and development phase.

After obtaining the environmental impact results, a sensitivity analysis was conducted for key variables of the Product System. These parameters were selected because they were under the control and management of the composite formulators. The criterion was established to reduce the deterministic nature of the modeling and to allow the sensitivity analysis to guide future actions aimed at reducing, minimizing, or even eliminating GHG emissions whenever possible.

The procedure was applied to the following parameters: (x) hemp productivity in the field, (y) the addition of binder to the pelleting process, and (z) the distance from the transportation of the PP resin production unit to the composite processing site. In addition to assessing the individual effect of each variable on the GWP, the sensitivity analysis also enabled an examination of the potential amplification or damping effects resulting from their associations.

For this purpose, a matrix system adapted from experimental planning practices was created, where the lowest contribution is coded as [−1], the intermediate stage as [0], and the most significant contribution as [+1]. This strategy yielded the indicator GWP (x, y, z), for which (x) and (z) can vary between [−1; +1], while the variable y can only range from [0; +1]. An explanatory summary of the foundations adopted for the sensitivity analysis is presented in Table 4.