Introduction

Bamboo, with the rapid growth rate and high strength, has been extensively used in architecture, bridges, furniture, pulp and paper, bamboo pipes among others1,2,3,4,5. Bamboo is a hygroscopic material and its mechanical properties, dimensional stability, and mold resistance are all affected by moisture content6,7,8,9. As in wood, water in bamboo exists in both forms of free water in the cell lumens and bound water in the cell walls. An accurate knowledge of water transport in bamboo is important to its drying process and chemical treatment. Of particular interest, water transport in bamboo is a multidimensional phenomenon due to its complex structure.

Bamboo has an anisotropic, gradient, and hierarchical structure10,11,12,13,14,15. As an anisotropic material, bamboo has variable water vapor diffusivities which could be analyzed in a cylindrical coordinate system. According to Huang et al.16, the water vapor diffusion resistance was markedly lower in the longitudinal direction than in the radial and tangential direction. Chen et al.17 measured the equilibrium moisture content (EMC) of bamboo at various radial positions across the culm and found that the EMC of the inner part was higher than that of the outer part especially at high relative humidity (RH) levels, and the outer part provided more sorption sites for water but fewer water clusters than the inner part. Besides the fore-mentioned studies on the anisotropic and gradient behavior, the literature to-date has provided little knowledge about the systematic effect of hierarchical structure on moisture sorption behavior.

In this study, we used the saturated salt solution method and dynamic vapor sorption (DVS) to measure the time-dependent mass response of cell wall substances, cells and blocks of bamboo. Two sorption directions were measured at these scales, and the parallel exponential kinetics (PEK) model was used to describe the sorption kinetics behavior. Emil et al.18 mentioned that the PEK model was highly dependent on the specific RH step and hold time. Here, by maintaining the same RH step and the hold time, PEK model was used to compare the scale and directional effects of bamboo. While improving the understanding of the micro and macro sorption mechanisms, this study was carried out with the following objectives:

  • To investigate the bound-water sorption behavior of cell walls of bamboo in both longitudinal and transverse directions;

  • To evaluate the effect of pits on water vapor sorption; and

  • To investigate the water vapor sorption behavior of bulk bamboo in both longitudinal and transverse directions.

Materials and methods

Materials

Moso bamboo (Phyllostachys edulis), with an age of 5 years, was obtained from Zhejiang Province, China. We confirm that we have the permissions to obtain the Moso bamboo from Zhejiang Province, China. Besides, all local, national or international guidelines and legislation were adhered to in the production of this study. Bamboo culm was selected from the internode sections located at a height between 1.5 and 2.5 m, and was air-dried to constant weight before testing. Bamboo blocks and bamboo slices were prepared for tests.

Bamboo blocks with two different sorption surfaces were cut from the stalks (inner and outer portion were removed). When the sorption direction was longitudinal (Fig. 1a), the dimension of specimens was 5 × 2 × L mm (tangential (T) × radial (R) × longitudinal (L)), where L was the sorption thickness (1 and 5 mm). TL and RL surfaces were covered by epoxy resin. When the sorption direction was transverse (Fig. 1a), the dimension of specimens was T × 2 × 5 mm (T × R × L), where T was the sorption thickness (1 and 5 mm). RT and TL surfaces were covered by epoxy resin (Fig. 1a). Figure 1b,c show the L and T specimens with the edges sealed tightly with resin coating, respectively. Very little resin penetrated into the samples.

Figure 1
figure 1

(a) Schematic of specimens with longitudinal and transverse sorption direction, where the blue arrows represent moisture sorption directions and red surfaces represent the surfaces that were covered by epoxy resin. (b) SEM picture of L specimen with resin edge sealing, (c) SEM picture of T specimen with resin edge sealing.

Full size image

Bamboo slices (Fig. 1a) with two sorption directions were cut from blocks using a microtome. The thicknesses of the slices were 30 and 60 μm. Without epoxy coating, the water vapor diffusion was assumed to occur mainly through the sorption surfaces because the width and length of slices were 2 and 5 mm, respectively, which were far greater than the thickness.

Saturated salt solution method

Five specimens were prepared for each thickness. The specimens (bamboo blocks and slices) were placed in the oven at 102 ℃ for drying till they reached a constant weight, and were then placed into desiccators with 90% RH (BaCl2 salt solution). Mass of the specimens was measured after 4, 24, 48, 72, 96, 120, 144, 168, and 216 h until equilibrium. For 1 mm and 5 mm samples, the MC was:

$$ {\text{MC}} = \frac{{{\text{W}}_{{\text{t}}} - {\text{W}}_{0} }}{{{\text{W}}_{{\text{B}}} }} $$

where \({\text{W}}_{\text{t}}\) is the weight of the sample (bamboo and resin) at time t, and \({\text{W}}_{\text{0}}\) is the weight of the sample at the beginning, and \({\text{W}}_{\text{B}}\) is the weight of the dry bamboo.

Dynamic vapor sorption apparatus

The bamboo slices were tested using a dynamic vapor sorption apparatus (DVS Intrinsic, Surface Measurement Systems Ltd., UK). The sample mass of each set was approximately 35–40 mg, and the sorption processes were run at a constant temperature of 25 °C. The RH was preset to increase from 0 to 90% with an increment of 10% and to 95% with an increment of 5%, and then decreased from 95 to 90% with a decrement of 5% and to 0% with a decrement of 10%. At each stage, samples were kept at the constant RH until the weight change per minute (dm/dt) was less than 0.01% in 10 min, and the mass measurement accuracy of samples was 0.0001 mg. According to previous studies30,31,32. On the other hand, water sorption rate is also directly related to the ability to treat bamboo, e.g., mold resistant treatment using borate solution, both in terms of the treatment time and uniformity33. In addition, the water sorption responses at different structure levels will have an impact on the multiscale mechanical properties of bamboo34,35,36,37, through the effect of moisture content or EMC38,39 and dynamic moisture diffusivity.

Conclusions

Water sorption process in bamboo was a complex transport response occurring in multiple scales. At the macro scale, the transport of moisture took place mainly along the longitudinal direction; whereas at the cellular scale, the moisture needed to transport in the transverse direction. Based on this study, the water vapor sorption behavior of bamboo at various hierarchical scales can be summarized as follows.

  1. 1.

    At the cell wall level (sample thickness 30 μm), the bound water diffused in cellular solids, and the sorption rate and EMC of cell walls were greater in the longitudinal specimens than in the transverse specimens. This was due to the looser packing of cell wall constituents and the more exposed sorption sites for interaction with water.

  2. 2.

    As the specimen thickness increased to 60 μm, the inter-cellular sorption involved more transport in the space of pits. Since the vapor diffusivity in the void space is much greater than bound water diffusivity within the cell walls27, the pits play a more dominant role in governing the sorption rate. The pits in parenchyma cells were only distributed in the lateral walls, and thus the pits increased moisture sorption in the transverse specimens.

  3. 3.

    At the macro scale with thick (1 and 5 mm) specimens, the space of lumens of vessel cells formed the main pathways in transporting moisture along the longitudinal direction of bamboo block. Thus, the sorption rate and EMC were greater in the longitudinal specimens than the transverse specimens. The effect of lumens of vessel cells on sorption rates between longitudinal and transverse directions increased with the thickness of specimen.