Properties and fibre characterisation of bleached hemp, birch and pine pulps: a comparison

The development of the world economy, the increase of demand for wood and the deterioration of forests have seriously threatened natural wood resources. Therefore, the pulp and paper industry, which produces about 40% of its products directly from wood, takes action to obtain fibres from forest plantations (Tarnawski 2006; Christersson 2010), producing papers and paperboards with a growing proportion of waste paper (Miranda et al. 2010) and mechanical pulps (Ford and Sharman 1996) and seeking possibilities to increase the yield of kraft pulps from wood (Bach and Fiehn 1972; Hart and Rudie 2014). Another possibility to reduce the dependence of the industry on wood is to use non-wood plant fibrous raw materials (NPFRMs) to produce papermaking pulps (Guadalix et al. 1996; Paavilainen 1998; Vargas et al. 2012; Danielewicz et al. 2015).

For NPFRM that can be used to make papers and boards, one can include primarily cereal straw, straw of plants grown for industrial purposes, as well as stalks of wild plants (Sadawarte 1995; Meeusen-van 1996; Potter 1996; Paavilainen 1998; Edelman 2000; Kane 2000; Schott et al. 2001; Jain et al. 2005; Xiaorong and Oinonen 2007; Zhu 2008). In 2010, virgin pulps from NPFRM accounted for 9.1% of the production worldwide of all such pulps (Bride 2012). Atchison (1996) and Paavilainen (1998) reported that the most important NPFRMs are cereal straw, bamboo and bagasse, the rest being reeds, cotton linters, flax and hemp, kenaf, esparto, jute, abaca and sisal. However, the possibility of making pulps from many of these NPFRMs in countries of moderate and boreal climates is limited, because they simply do not grow in these climate conditions.

One of the plants that grows in moderate or boreal climate areas is industrial hemp (Kane 2000; Pahkala et al. 2008). The advantage of hemp over cereal straw is a higher biomass yield per hectare of cultivation (Bawyer 2001; Burczyk et al. 2009; Prade et al. 2011; Consentino et al. 2013; Finnan and Styles 2013; Jankauskiené and Gruzdiené 2015) and lower content of mineral substances than in wheat straw. For papermakers, the yield of dry hemp stalks per hectare is more useful information than the yield of hemp biomass. The ranges of values of this index are quite different, depending on the cultivar of hemp and the geographical area where the cultivation of hemp is located, from a minimum of about 5 t/ha/y to as much as 22 t/ha/y (Struik et al. 2000; Bawyer 2001; Pahkala et al. 2008; Angelini et al. 2016).

As shown in the literature, the use of industrial hemp has been considered for paper production. Some Canadian researchers have been very active in this area. They studied the chemical composition of a Canadian cultivar of hemp, the kinetics of delignification of hemp stalks, hemp bast and hemp woody-core using the soda method, together with comparison of some morphological properties of pulps (Correia et al. 1998, 2001). In subsequent work, they also compared the results of processing of hemp stalks using soda, soda-anthraquinone (AQ), kraft and kraft-AQ methods into pulps (Correia et al. 2003).

Studies of the application of industrial hemp for papermaking purposes has also been conducted in Holland, Hungary, Australia, Spain, China, India, Czech Republic and Poland. They studies were concerned with the processing of hemp raw materials (stalks, woody-core or bast) into soda pulps (Van der Werf et al. 1994; De Groot et al. 1995, 1999; Dutt et al. 2005; Fišerova et al. 2013), organosolv pulps (Zomers et al. 1995; Barberà et al. 2011), thermomechanical pulp (TMP), cold caustic soda (CCS) pulp (Lisson 2001), alkaline peroxide extrusion pulp (APXP; DeJong et al. 1999), enzymatic-mechanical pulp (Ekblad et al. 2005), chemithermomechanical pulp (CTMP), sulphite pulps (Bączyńska et al. 1993; Stupińska and Madaj 1996; Stupińska et al. 2005; Dutt et al. 2005) and kraft pulps (Danielewicz and Surma-Ślusarska 2010a, 2011; Dutt et al. 2005). Also studied were the kinetics of the delignification of hemp stalks and hemp woody-core (Abdul-Karim et al. 1995; Dang and Nguyen 2006) and properties of bleached pulps from hemp roots (Miao et al. 2014). Kovacs et al. (1992) also pulped hemp stalks using kraft, bisulphite, neutral sulphite and alkaline oxygen methods, but they did not present the properties of the handsheets prepared from these pulps.

The presented review of the literature shows that hemp stalks can be processed into pulps by various methods. However, the properties of cellulosic pulps made from hemp stalks have been examined only in a few of the above-mentioned works. For example, De Groot et al. (1999) and Fišerova et al. (2013) presented the properties of hemp woody-core and hemp stalk soda pulps, respectively, Zomers et al. (1995) studied the density, tensile, tear and burst indices of organosolv hemp stalks, woody-core and bast pulps, DeJong et al. (1999) and Ekblad et al. (2005) presented the same properties of APXP and ezymatic-mechanical pulp, respectively, while Dutt et al. (2005) showed tensile, tear, burst and double fold analysis results of unbleached hemp stalks and hemp bast pulps obtained using different methods.

The results of these works indicate that hemp stalks and unbleached hemp bast pulps obtained using these methods possessed high bulk and tear, while these characteristics in hemp woody-core pulps had low values.

However, the economic effects, e.g. the profitability, of hemp pulp production using some of mentioned methods, especially soda and organosolv, are of great importance. It is necessary to take into account the high investment costs of building a separate mill for the processing of hemp stalks, for example, into unbleached and especially bleached pulps.

A method commonly used in the world’s paper industry to produce bleached cellulosic pulps consists of subjecting plant fibrous raw materials (PFRM) to kraft pulping, oxygen delignification and bleaching using elemental chlorine-free (ECF) technology. So, it would be a good idea to compare the properties of hemp stalk, hemp woody-core and hemp bast fibre pulps obtained using the sequence of these processes, especially since there is a lack of data the literature on the properties of bleached kraft pulps processed from these NPFRMs. An enlargement of knowledge in this area could help in evaluating the possibilities of joint processing of birch and pine wood with a small addition (a small percentage) of any of these hemp raw materials, which seems to be the best method of introducing hemp into the pulp and paper industry considering the possible use of existing industrial equipment.

This most economically justified solution would fit with the EU’s environmental directives, which aim to reduce the consumption of wood fibre in paper and board products, and replace it with other plant biomass, which is produced annually in huge quantities worldwide in a much shorter growing cycle than wood. If the resulting pulp could be characterised by a higher yield from the PFRM and better properties than pulps from wood, then replacing the wood with NPFRM could improve both the final yield and some of the final properties of fibrous semi-manufactured products. Due to the large-scale production of the papermaking industry, this could have both economic and qualitative importance for this industry and the properties of some kinds of papers and boards, respectively.

Based on these assumptions, studies were conducted to evaluate the different properties of bleached hemp stalk, hemp woody-core and hemp bast fibre kraft pulps, which were compared with the properties of bleached kraft pulp from birch and pine. The first two hemp pulps were made from hemp stalks and from hemp woody-core, respectively, according to the sequence: kraft pulping, oxygen delignification and ECF bleaching, while the third pulp was obtained using only oxygen delignification and bleaching (i.e. bypassing the kraft pulping process) due to low lignin content in hemp bast fibres.

The large differences of the properties of these pulps prompted us also to investigate the morphological properties of hemp fibres, their strength and chemical composition, which was necessary in order to explain the observed relationships.

Figure 2 shows the Schopper–Riegler freeness (SR-freeness) values of bleached hemp pulps and bleached birch and pine pulps after various beating times.

The SR-freeness of bleached kraft pulps made from hemp stalks (H _s ) and hemp woody-core (H _wc ) in an unbeaten state amounted to 24 and 29 °SR, respectively. The beating process of these pulps significantly increased the value of this feature. For example, after 10 min of beating, it increased to 38°SR (H _s ) and 51°SR (H _wc ), respectively whereas in the case of the birch and pine pulps, it increased from 13 to 17 °SR. Thus, the beatability of hemp stalks and hemp woody-core pulp (measured by °SR) was much better than birch or pine pulps. However, the former pulps are likely to show much worse drainage in the papermaking process than pulps from wood. On the other hand, susceptibility of pulp made from hemp bast fibres (H _bf ) for the beating was much worse than other hemp pulps studied. After 45 min of beating, the SR-freeness of this pulp hardly exceeded 23°SR.

Dry strength of paper is responsible for its good runnability in printing presses. This property of pulps is also important for the possibility of increasing the filler content in paper to achieve the required opacity. Figure 3 shows the comparison of the tensile index of handsheets of hemp, birch and pine pulps.

As seen in this Figure, bleached pulp made from hemp stalks has high tensile strength (approximately 60 N·m/g) even without beating. It is reported that high initial tensile strength of pulps is beneficial because such pulps can be beaten with lower input of energy or used without beating (Wandelt 2007). Kraft pulps made from hemp stalks and hemp woody-core may reach a high tensile index after beating (approximately 90 N·m/g). The maximum tensile index of H _s and H _wc pulps was comparable with the maximum tensile index of pine pulp and lower by approximately 10 N·m/g, in comparison with birch pulp. The obtained maximal values of tensile index of bleached H _s and H _wc kraft pulps are somewhat higher than those of unbleached hemp stalk kraft pulp, bleached hemp woody-core soda and organosolv pulps presented by other authors (Zomers et al. 1995; De Groot et al. 1999; Dutt et al. 2005; Fišerova et al. 2013).

The tensile index of hemp bast fibre pulp (H _bf ) turned out to be low and much lower than the tensile index of birch and pine pulps even after quite a long beating (45 min) despite reaching similar SR-freeness values compared with pine pulp. The tensile index of hemp bast fibre pulps was also determined by Ekblad et al. (2005) and DeJong et al. (1999). However, they were enzymatic-mechanical and APXP pulps, respectively. The values of this index of these pulps were also on a relatively low level (46-51 N·m/g), but higher than in our study. In the literature, one can also find the data concerning the tensile index of hemp bast organosolv, soda and kraft pulps (Zomers et al. 1995; Dutt et al. 2005). Interestingly, the tensile index of these pulps was higher (62–66 N·m/g) than that of H _bf pulp. The reason for this may be the content of secondary hemp bast fibres and other cells present in hemp bast, such as parenchyma cells and sieve tubes, which probably were absent in our pulp, composed mainly of primary hemp bast fibres.

The tear and tensile relationship shows the strength potential of pulp. This relationship for bleached hemp, birch and pine pulps is presented in Fig. 4.

Figure 4 shows that tear index of unbeaten and beaten hemp stalk pulp is high (15.3–17.4 mN·m²/g). Comparing the values of the tear index of bleached kraft pulps prepared from hemp stalks, birch and pine (at tensile index ranging from 70–90 N·m/g) one can conclude that the former pulp was characterised by a much higher tear index than birch pulp and even somewhat higher than pine pulp. The tear index of hemp woody-core pulp was, in turn, much lower after beating than that of birch and pine pulps (4.5–6 mN·m²/g). Unbeaten H _bf pulp is characterised by a tear index similar to that of birch and pine pulps beaten to 90–100 N·m/g. As seen in Fig. 4, the tear strength of this pulp increased during beating. After 45 min of beating in the Jokro mill, it increased to 21 mN·m²/g. The high tear index of pulp from hemp bast fibers indicates that the hemp stalk pulp owes its relatively high tear strength to the presence of hemp bast fibres. It seems that the high tear strength of H _s and H _bf pulps can be used in practice. For example, adding some amount of these pulps to low-tear strength pulp, e.g. waste paper pulp, should improve this feature of the latter pulp. The tear index of unbleached organosolv, kraft and soda-AQ hemp stalk pulps was also determined by such authors like Zomers et al. (1995), Dutt et al. (2005) and Fišerova et al. (2013), respectively. The index values of pulps obtained by two former research teams were lower than the ones obtained in our work. This can be a result of the method of processing of hemp stalks into pulps, and the hemp cultivars used by these authors. In the world, a lot of cultivars of hemp are known (De Meijer 2002; Grabowska 2004; Jankauskiené and Gruzdiené 2009; Consentino et al. 2013; Angelini et al. 2016). On the other hand, the tear index of handsheets of hemp stalk soda-AQ pulp of freeness up to around 40°SR obtained by Fišerova et al. (2013) was similar with the one obtained in our study. On the other hand, the values of tear index of bleached hemp woody-core pulp and hemp bast fibre pulp obtained in our work are in good agreement with that presented by other authors for hemp woody-core pulp (Zomers et al. 1995; De Groot et al. 1999) or hemp bast pulp (De Groot et al. 1999; Dutt et al. 2005).

One of the most important structural properties of pulp is its bulk. Higher bulk enables much stronger calendering of paper, influences positively the stiffness of paper, water absorbency of tissue products, the height of books and the height of tissue rolls. The bulk of hemp pulps was compared with the bulk of birch and pine pulps in Fig. 5.

Figure 5 shows that the pulp produced from hemp stalks (H _s ) at the level of 60–90 N·m/g of the tensile index had relatively high values of this ratio (1.60–1.95 cm³/g). In contrast, the bulk of the hemp woody-core pulp was low (1.25–1.35 cm³/g). As for the bulk of the hemp bast fibre pulp (H _bf ) it was high in both the unbeaten and beaten state. In the first case, it amounted to approximately 3.2 cm³/g, while in the second case (after 45 min of beating in the Jokro mill), it was 1.9 cm³/g. One of the reasons for the high bulk of H _bf pulp is certainly the low degree of bonding of its fibers (tensile index of 10–27 N·m/g). It seems that high bulk of H _bf pulp can be used in practice similarly as the high tear index of this pulp.

Out of optical properties, light-scattering ability and opacity were determined. The light-scattering ability and opacity of H _s , H _wc , H _bf , birch and pine pulps are presented in Figs. 6 and 7, respectively.

The highest values of light-scattering coefficient and opacity were obtained in the case of handsheets prepared from hemp woody-core pulp, and the lowest values were obtained from the hemp bast fibre pulp. The hemp stalk pulp at 80–90 N·m/g had similar or slightly better optical characteristics in comparison with birch pulp.

An important feature of bleached pulps, especially for hygienic tissue products, is their water-absorption time (WAT) and water-absorption capacity (WAC). A comparison of these properties of unbeaten bleached pulps is shown in Fig. 8.

Figure 8 shows that the average value of the WAT (left axis in Fig. 8) was lowest for bleached pine pulp and highest for hemp woody-core pulp. This time is similar for H _s , H _bf and birch pulps. WAC (right axis in Fig. 8) was higher for bleached birch and pine pulps than for hemp pulps by approximately 20%. The results suggest that the indices studied are more favourable for pulps that are characterised by longer fibres and/or more uniform distribution of fibre length. The lower WAC of H _s and H _wc pulp than birch pulp can be explained by a lower bulk of the former pulps in fluffed state than birch pulp, which results from higher fines content in hemp pulps and the presence of a fraction of short fibres in them. On the other hand, a lower WAC of H _bf pulp as compared to birch pulp when fluffed may result from too large of pores in the case of the former pulp, which were unable to hold water that leaked out more readily from this pulp.

In order to clarify the differences between the properties of hemp pulp and pulp from wood, further analyses, calculations and experiments were performed. The results of these analyses are presented in Tables 3, 4 and Figs. 9–11.

The data of Table 3 show that the hemp woody-core pulp contained approximately 32 × 10⁶ fibres per gram of pulp. This is around threefold higher than birch pulp. The number of fibres in hemp bast fibre pulp was only about 2.6 × 10⁶ fibres per gram of pulp, i.e., less by about one million than in pine pulp. The number of fibres in H _s pulp has an intermediate value of about 17 × 10⁶/g. A similar trend can be observed in the case of fines content. Hemp woody-core pulp has a fines content about threefold higher than birch pulp, while for the hemp bast fibre pulp, it was lowest among the pulps studied.

The differences in fibre content between pulps can be explained by differences in fibre length and coarseness. Hemp woody-core fibres (H _wc ) have low length and low coarseness, and these are the main reasons for high numbers of fibres in 1 g of this pulp. On the other hand, hemp bast fibre pulp has high average coarseness and average fibre length similar to pine pulp. Thus, the number of fibres in 1 g of this pulp is lower than in the case of pine pulp.

Differences in the number of fibres in 1 g of pulp explain significant differences in optical properties of pulps, while the number of fibres in 1 g of pulp and fines content explains the SR-freeness variation amongst pulps.

The calculated Runkel ratio indicates that H _bf pulp has approximately several times higher fibre stiffness in the unbeaten state than the fibres of birch and pine pulps. This is largely due to high cell wall thickness of hemp bast fibres, but the high degree of crystallinity of hemp bast fibres cellulose, reported by Thygesen et al. (2005), can also have some influence on hemp fibre stiffness. Cell wall thickness of fibres from hemp bast was also determined by Dutt et al. (2005). They reported a mean value of this index of 6.7 μm, i.e., 23% more than in our research. On the other hand, H _wc fibres are characterised by a Runkel ratio about twelve times lower than that of hemp bast fibres. This results from low cell wall thickness of hemp woody-core fibres (0.95 μm).

The Runkel ratios and cell wall thicknesses partly explain the good tensile strength of H _s and H _wc pulps, the low tensile index of H _bf pulp and the differences in bulk between pulps from hemp. As for the cell wall thickness of fibres of pulp from hemp woody-core, Dutt et al. (2005) reported quite different value of the feature (7.05 μm). This indicates the possibility of significant differences in cell wall thickness of hemp woody-core amongst different hemp cultivars. The higher values of cell wall thickness of fibres in hemp bast and hemp woody-core obtained by Dutt et al. (2005) seem to explain the lower tensile index of handsheets prepared from hemp stalk kraft pulp obtained by these authors than that obtained in our work.

Stiffness of fibres is not the only factor that influences the properties of pulp handsheets. For example, tensile strength depends also on the ZSTS and shear bond strength per unit of bonded area. The first property depends mainly on intrinsic viscosity, but the second one is primarily influenced by the hemicellulose content of fibres and the degree of the fibres’ external fibrillation (Page 1969; I’Anson et al. 2006). The strength of fibres of studied pulps was compared using the ZSTS results, which are presented in Fig. 9.

The data of Fig. 9 indicate that fibres of H _bf pulp are strongest among the studied pulps. This partly results from the higher intrinsic viscosity of H _bf pulp than other ones. Thus, the fiber strength of H _bf pulp is not the reason for the lower tensile strength of this pulp.

The hemicellulose content in pulp may also have a significant effect on its tensile index, and this was determined indirectly through determination of the α-cellulose content. These results show that the hemicellulose content in H _bf pulp is lowest among the pulps (probably around 2%), because of the highest α-cellulose content in this pulp (Table 4).

High α-cellulose content in H _bf pulp together with the high yield of this pulp (85%) are very promising results from the point of view of using it as a pulp destined for chemical processing, because the yield of such pulps obtained from wood with the use of a kraft method is usually low (about 35%; Surewicz and Rutkowski 1958).

As mentioned, the second factor that influences the shear bond strength of fibres in a sheet is the degree of their external fibrillation. In order to compare the ability for fibrillation of H _bf pulp fibre and pulps from wood, experiments consisting of beating pulp samples in the Jokro mill for 60 min and subsequent observation of fibres in the optical microscope were performed. The photographs of fibres after being beated are presented in Figs. 10 and 11.

From Figs. 10 and 11, it can be seen that fibres of H _bf pulp, after the same beating, are more fibrillated than fibres of pine pulp. This conclusion was also confirmed by the results of computer image analysis in the MorFi apparatus. The participation rate of the fibrils in the total length of bast fibres after 60 min of beating in the Jokro mill was 2.5%, while for the pine pulp, it was only 1.0%. Thus, the susceptibility of fibres of H _bf for fibrillation can’t be the reason for the low tensile index of handsheets of this pulp.