Cell Wall Pore Structures of Bamboo as Relation to Species and Tissue Types


 Efficient convention of bamboo biomass into biofuel and biomaterials, as well as chemical treatment are both highly related to the porosity of cell wall. The present work characterizes the micropore and mesopore structure in cell walls of six different bamboo species and tissue types using CO2 and N2 adsorption. Two plantation wood species were also tested for comparison. Bamboo species normally showed lower cell wall porosity (2.64%-3.75%) than wood species (3.98%-5.06%), indicating a more compact structure for bamboo than wood. A distinct species dependence of cell wall pore structures and porosity was also observed. Furthermore, the cell wall pore structure and porosity are shown to be tissue-specific, as the parenchyma cells exhibit higher pore volume and porosity compared to bamboo fibers. The obtained results give new explanations on the known facts that both bamboo and bamboo fibers exhibit higher biomass recalcitrance as compared to wood and bamboo parenchyma cells, constructing the base of pretreatment optimization and subsequent processing for bamboo-derived biofuels and biomaterials.


Introduction
As one of the most important lignocellulosic resources, bamboo has been receiving increased attention because of its fast growth, high biomass yield, and superior mechanical properties compared with wood (Sarah et al., 2019;Chen et al., 2020). Interest has arisen in the biomass conversion of bamboo to biofuels (Ye et al., 2017) and biomaterials, such as nanocellulose (Wang et al., 2015), activated carbons (Negara et al., 2019), carbonized solar vapor-generation device (Bian et al., 2018), transparent bamboo composite . To convert bamboo into all these newly developed products, the rst essential step is to disrupt the inherently recalcitrant plant cell walls by mechanical and/or chemical pretreatment technologies, including grinding, acid, alkaline, solvent, thermal pretreatment, and so on (Payne et al., 2015). In most cases, pretreatments applied for bamboo are basically borrowed from those used in wood based on their similar cell wall architecture of cellulose micro brils embedded in a matrix of hemicellulose and lignin (Chundawat et al., 2011). However, recent works indicated that bamboo needed a longer time and more chemicals during pretreatment due to its poor permeability. For example, it is di cult to achieve the desirable transmissivity of transparent bamboo similar to those transparent wood using the same deligni cation pretreatment method (Wu et al., 2020). It appears that the inferior pretreatment effectiveness of bamboo was due to its poor permeability caused by hierarchical structure . However, cell wall structure of bamboo, alongside its underlying relationship related to the mechanisms of pretreatments has not yet been fully understood compared to wood.
Pore structure of plant cell wall plays a vital role during biomass pretreatments, as they mostly affect the behavior of porous media such as the movement and ow of uid within biomass materials (Negara et al., 2019;Liu et al., 2021). Although bamboo and wood are both porous materials, their pore structure differ distinctly in two aspects: (1) bamboo mainly consists of two types of plant tissues, namely sclerenchyma ber and parenchyma tissue, whereas wood contains various plant tissues, including vessels, tracheids, wood bers, longitudinal parenchyma and wood ray (Yin et al., 2015); and (2) the secondary cell wall of bamboo exhibits a typical multi-layered wall structure whereas that of in wood had a three-layered structure (Zhai et al., 2014). Previous studies have focus on the quantitative determination of pore structure (e.g., pore shape, surface area, pore size distribution, pore volume) in micropore (<2 nm), mesopore (2-50 nm), and macropore (>50 nm) in native or pretreated wood cell (Chang et al., 2009;Yin et al., 2015). Very recently, the in uence of different treatments on the pore structure of bamboo has gaining much attention (Ye et al., 2020;Liu et al., 2021;Su et al., 2021a, b), but the native pore structure, especally the microspores, has not been fully characterized. Various techniques, such as gas sorption isotherms, mercury intrusion porosimetry, solute exclusion, thermoporosimetry, microscopic observations are available for determining pore structure of cell walls of plant materials (Papadopoulos et al., 2003). The gas adsorption method can quantitatively analyze the pore size distribution, pore volume, and speci c surface area without changing the pore structure of the material (Yin et al., 2015). At present, the most commonly used N 2 adsorption at 77 K can provide pore structure information with pore diameter in the range of 0.5-200 nm. However, the stronger quadrupole moment of N 2 molecules enhances the force between N 2 and the non-uniform surface of the adsorbent, which makes it di cult to distinguish certain micropores. Moreover, the adsorption of N 2 in the micropores occurs under low pressure and requires a very high degree of vacuum, for these two reasons, N 2 adsorption cannot measure some microporous in the materials. Therefore, CO 2 adsorption was used to test the micropores in wood cell walls due to its faster diffusion rates and better penetration ability (Nakatani et al., 2008).
In this study, the pore shape, pore volume, and pore size distribution of bamboo, wood, bamboo bers and parenchyma cells were measured using N 2 and CO 2 adsorption. Comparison of pore structure of six bamboo species to two commonly wood plantation species, namely Fir (Cunning lanceolate) and Poplar (Populus tomentosa), were also conducted. The main purposes of this article are to give new explanation to the known facts that bamboo is harder to be bio-conversed to biofuels or nanocellulose than wood, as well as the lower biomass recalcitrance of bamboo parenchyma cells as compared to bers, in terms of cell wall pore structure. The obtained results are able to provide a useful insight into development of optimized processing techniques speci cally for bamboo and its subsequent new product design. species were prepared from Fir (F) (Cunning lanceolate) and Poplar (P) (Populus tomentosa) as softwood and hardwood, respectively. Samples for gas adsorption characterization were ground to bamboo or wood powders between 30 and 60 mesh (screen size: 0.6-0.3 mm). In order to study the difference of cell wall pore structure characteristics between bamboo bers and parenchyma cells, a simple water separation method was used to physically separate bamboo bers and parenchyma cells from Moso bamboo (Phyllostachys edulis). Figure1 depicts the separation process of bamboo bers and parenchyma cells, a certain amount of 30-60 mesh bamboo powder was dispersed in 20 times volume of water, stirred and placed for 10-20 seconds, and the parenchyma cells oated atop water while bamboo bers sank to the bottom due to their signi cant density difference. Separated samples were oven-dried for further use.

Microscopic observation
Morphology of Moso bamboo was carried out using a eld emission scanning electron microscope (FE-SEM, SU8020, Hitachi). Sections of samples (4 × 4× 4 mm 3 ) were cut with a sliding microtome along the transverse directions. All the surface polished samples were sprayed with Pt using a vacuum sputter coater before subjected to the SEM analysis.

Pycnometer test
Density of dry cell wall (ρ cw ) and bamboo/wood substance (ρ s ) of sample was evaluated using the classic pycnometric method (Stamm, 1964) using water (polar solvent) and xylene (non-polar solvent) as replacing liquids, respectively. 2 g 30-60 mesh power samples were added into a pre-weighed 25 mL pycnometer, and then pycnometer was then lled with water or xylene and de-gassed under vacuum to remove air bubbles. After the vacuum was released, the liquid was added to the hydrometer, and the capillary hole on the hydrometer and the plug were lled with replacement liquid. The excess liquid was wiped off with lter paper and the nal weigh was recorded. After that, the liquid and sample in the pycnometer were poured out, then washed and dried the pycnometer. Finally, the pycnometer was re lled with liquid and weighed. ρ s , ρ cw, ,.and the porosity (C) were calculated as following: Where: m 1 was the mass of sample, m 2 was the total mass of pycnometer, sample and liquid, m 3 was the mass of pycnometer and liquid, ρ w was the density of water (0.9977735 g/cm 3 , Harris, 2003), ρ x was the density of xylene.

Gas adsorption measurement
Mesopore and micropore size distribution in cell walls of bamboo and wood were analyzed from N 2 and CO 2 adsorption isotherm, respectively. N 2 (77 K) adsorption measurement was conducted using an automatic gas adsorption instrument (AUTOSORB-1, Quantachrome, USA) and CO 2 (273 K) measurements on Micromeritics ASAP 2020. 1 g powder samples were added into a 6 mm bulb cell and outgassed for 10 h at 80°C before mesopore and micropore tests. Density functional theory (DFT) method was applied to provide pore size distribution for the entire micro-mesopore range. The Barrett-Joyner-Halenda (BJH) model and Horvath-Kawazoe (HK) model was applied to calculate mesopore and micropore volume, respectively. Using the density functional theory (DFT) model to analyze the measured N 2 adsorption isotherms and CO 2 adsorption isotherms, we can obtain detailed information about the micropore and mesopore structure of carbon materials. The DFT can correctly describe the uid structure close to the real pore wall (Puziy et al., 2015). Although the BJH method is based on a simple cylindrical pore model, it can be used to analyze the pore volume of pulp bers (Kimura et al., 2016). Therefore, we can analyze the mesopores in the cell wall of bamboo and wood. The Horvath-Kawazoe (HK) method is used to obtain the pore volume of the micropores, which is suitable for measuring the structure of micropores less than 2 nm (Kojiro et al., 2008).
It is generally believed that the pore shape is related to the type of hysteresis ring. According to the classi cation of International Union of pure and Applied Chemistry (IUPAC), six isotherms and four hysteresis loops are determined (Sing, 1985). The adsorption isotherm of the microporous material is type I adsorption isotherm. The gas adsorption isotherm produced by non-porous or macroporous materials presents a reversible type II isotherm, and the type III isotherm also belongs to the curve of nonporous or macroporous materials. Type IV isotherm is shown by mesoporous materials. In mesopores, the rst single-multilayer adsorption on the mesoporous pore wall is similar to the type II isotherm, and then capillary condensation occurs in the pores, so that the adsorption and desorption curves do not overlap to form a hysteresis loop. Type V isotherms are rarely encountered because of the relatively weak interaction between the adsorbed material and the adsorbed gas. Type VI isotherms are known for their step like reversible adsorption processes, which are due to the successive multilayer adsorption of homogeneous non porous surfaces. The H1 hysteresis ring is a uniform pore model, which can be regarded as a straight pore. The H2 hysteresis ring is generally considered to be caused by porous adsorbent or uniform particle accumulation pores. The shape of the pore is mostly considered to be the ink bottle. After the liquid nitrogen in the bottle neck of the small pore size is desorbed, it will escape suddenly when bound in the bottle. Compared with H4, H3 has a larger adsorption capacity at the high pressure, which is considered to be a slit pore formed by the accumulation of ake particles. H4 is also a slit pore, which is different from the particle accumulation and is similar to the pore produced by the layered structure.

Results And Discussion
3.1 Surface pore morphology of bamboo SEM revealed apparent differences in the surface pore morphology among different cell wall types in Moso bamboo, as shown in Figure 2. Pore structures of different sizes can be seen in the cross section of Moso bamboo, including lumen of parenchyma cell, lumen of bamboo ber, pits and intercellular spaces (Figure 2a, c). Bamboo bers exhibited a dense cell wall structure with small lumen and thick wall ( Figure  2a), and cellulose micro brils were tightly packed with scarce micropores were observed in ber cell wall (marked by white arrow, Figure 2b). In contrast, parenchyma cells showed more porous structure with large lumens, thin walls, and abundant pits (Figure 2c), and more pores formed between cellulose micro brils can be observed under high-magni cation SEM image (marked by white, arrow, Figure 2d).   (Zauer et al., 2013). The lignin content of bamboo was close to that of hardwood and lower than that of softwood (Itoh, 1990;Xin et al., 2015). Therefore, the density of substance of bamboo was close to that of Poplar (hardwood) and higher than that of Chinese r (softwood).

Cell wall density, density of substance and cell wall porosity
Bamboo species normally showed lower cell wall porosity (2.64%-3.75%) than wood species (3.98%-5.06%), which partially explained the phenomenon that bamboo exhibited poorer permeability and more di culty in impregnation process (Wu et al., 2020;. The calculated wood porosity was in the range of those reported in the literature, ranging from 1.64 to 4.76% for 18 wood species (Kellogg et al., 1969). In addition, parenchyma cells exhibited signi cantly higher cell wall porosity (5.07%) than bamboo bers (2.91%), which indicated a more loosen structure of parenchyma cells.

Cell wall mesoporous structure
To further analyze the pore structure of samples among different species and cell wall types, the nitrogen adsorption-desorption isotherms were conducted. As shown in Figure 4a, c, e, isotherms of wood and bamboo samples were similar, indicating intermediate between type II and type curve with H3 hysteresis loops (slit-shaped pores), which were typical of mesoporous structure as classi ed in the IUPAC (Sing et al., 1985). At present, many scholars have used this method to study the wood cell wall. Yin et al. (2015) used the N 2 adsorption method to study the heartwood and sapwood of Chinese r (Cunninghamia lanceolata) and showed that the pores of the heartwood and sapwood are all slit-shaped. Chang et al. (2011) used the same method to characterize the pore structure of tensile wood of poplar (Populus sp.) and found that there are ink bottle-like pores and slit-like pores. Compared with bamboo species, wood species showed a bigger hysteresis loop, suggesting that there is larger mesoporous pore diameter. Using the DFT method, bamboo and wood showed mesopores in the ranges of 2-35 nm diameter with the distribution peaks centered at 2.3 nm for F and P species, 2.0-2.5 nm for bamboo species, 2.0 nm for bamboo bers, and 2.3 nm for parenchyma cells (as shown in Figure 4b, d, f).
Moreover, the mesopore structural parameters of different samples were shown in Figure 5. The mesopore volume of C species (0.0028 cm 3 g −1 ) was 1.87 times higher than of the H species (0.0015 cm 3 g −1 ), indicating the bamboo species played some roles in the mesopore structure. Compared with bamboo, the cell wall in the wood exhibited larger mesopore volume. Largest amount of mesopores exited in softwood Fir, and 63.03% of mesopores were found in diameters of 2-10 nm. The trend of pore size distribution was similar between hardwood Poplar and bamboo species, as mesopores of 2-10 nm are the predominant pore size diameters and accounted for 47.92-56.06% of the total pore volume depending on the plant species. These results agree well with data previously reported for crops of 2-8 nm as typical pore diameters using the same Nitrogen adsorption method (Adani et al., 2011).

Cell wall microporous structure
The CO 2 gas adsorption revealed the presence of micropores in all bamboo and wood samples (0.45-0.90 nm), as shown in Figure 6. Both bamboo and wood samples showed large open hysteresis in the CO 2 sorption isotherms, indicating that bamboo and wood have a complex microporous structure and selective capture of CO 2 (Shi et al., 2017;Mulfort et al., 2010). As shown in Figure 7, all samples exhibited much larger micropore volume than mesopore volume (Figure 6), for example, micropore volume (0.0074 cm 3 g −1 ) was 4.1 times higher than mesopore volume (0.0018 cm 3 g −1 ) for H species. Previous studies reported 10 times larger of micropore volume than mesopore volume for dry wood (Kojiro et al., 2010).
In regardless of plant species, ultramicropores with diameters of 0.4-0.6 nm were the main pore class size diameters, i.e., 50.55%-64.87% of the total micropore volume of samples. It should be noted that around 80% pores having diameters below 0.8 nm, which were consistent with previous ndings reported for crops cell wall (Adani et al., 2010). As bamboo and wood depicted similar cell wall architecture of cellulose micro brils embedded in matrix of hemicellulose and lignin, and the presence of micropores below 0.8 nm arised from the space between cellulose brils and matrix . Previous studies have revealed that cell wall chemical composition has a strong in uence on pore volume, for example, extractives largely affect the micropores ranged between 0.4-0.6 nm (Yin et al., 2015). Bamboo has high extractive content (four times higher than that of wood, Subekti et al., 2015) which could coat the pits and cell walls , resulting in a decreased pore volume of bamboo. Interestingly, micropore and mesopore volume for the separated parenchyma cells are very close to that of two wood species, indicating their fairly porous structure. Recent works found that parenchyma cell require less energy input to product biomaterials, such as nanocellulose (Wang et al., 2015) and bioethanol (Jin et al., 2019). We have reasons to believe that higher cell wall porosity of parenchyma cells should contribute to that. Our investigations supported bamboo parenchyma cells may represent an underexploited important resource, which can be converted into valuable products with lower energy consumption.

Conclusion
In this work, the micropore and mesopore structure of bamboo and wood was characterized by CO 2 and N 2 sorption isotherm, respectively. Both bamboo and wood exhibited a mixed of adsorption type II and type IV and the H3 type hysterisis loop, but a less porous structure was reported for bamboo. Bamboo exhibited variable pore size distribution and pore volume among different species. Separated parenchyma cells have signi cantly higher meso/micropore volume and cell wall porosity than bers. The obtained results will promote the understanding of effect and mechanism concerning bamboo biomass pretreatment and subsequent processing for the development of advanced value-added bamboo products.   The N2 adsorption isotherm (a, c, e) and mesopore size distribution curve (b, d, f) of wood, bamboo, bamboo bers and parenchyma cells of Moso bamboo Note: data of M was cited from our previous study (Cao et  The mesopore volume (a) and pore volume proportion of different size (b, c, d) of wood, bamboo, bamboo bers and parenchyma cells of Moso bamboo Note: data of M was cited from our previous study (Cao et al., 2021) Page 16/17 The CO2 adsorption isotherm (a, c, e) and micropore size distribution curve (b, d, f) of wood, bamboo, bamboo bers and parenchyma cells of Moso bamboo Note: data of M was cited from our previous study (Cao et al., 2021)