Roles of paper composition and humidity on the adhesion between paper sheet and glass: a molecular dynamics study

Understanding adhesion behavior between paper materials and inorganic substrate is important to minimize surface contamination by paper fragments. In this work, we investigate adhesion mechanisms between paper sheet and glass in terms of molecular interaction. Molecular dynamics simulations are employed to calculate adhesion force between paper film and the silica glass surfaces. Pulling and sliding tests are simulated to find the effects of the paper compositions such as glucan, mannan, and glucuronoxylan (xylan) and humidity on the adhesion. Simulation results reveal that artificially constructed film of mannan unit shows higher adhesion than that of cellulose film which consists of glucan unit due to more amount of hydroxyl groups on the silica surface. Adhesion force of xylan-cellulose composite case by pulling test is 40% lower than cellulose-only case, whereas during the sliding test, adhesion force of the former is 100% higher than the latter. It turns out that this difference comes from the existence of nanocavity around the xylan molecule when the composite is adsorbed on the silica surface. In addition, introduction of humidity leads to a further increase of adhesion due to hydrogen bonds bridged by water molecules. It is found that the adhesion force is maximized around surface density of 10 H2O/nm2. It is discussed that consideration of the capillary force for paper may result in different adhesion response that reflects more realistic situation.


Introduction
Packaging paper is used to protect underlying substrates during transportation and storage. To preserve substrate surface cleanliness, particularly for high tech applications whose surface quality is of great importance, it is ideal that the paper survives a wide range of humidity with minimal interaction with the substrate. Deposition of paper moieties onto the substrate surface during use and sticking, even at the micrometer-scale, may result in severe deterioration of surface quality. In this regard, the fundamental solution for those issues should start with an understanding of adhesion behavior at the interface from the molecular scale. Furthermore, it is expected that this understanding can contribute to chemical optimization of paper components or control of ambient conditions for minimization of the adhesion strength.
In general, interfacial adhesion is dependent on many factors including the nature of the interfacial interaction, mechanical locking, chemical reaction, and inter-diffusion (Awaja et al. 2009;Lacomb 2006). In ambient conditions, it is generally accepted that a major factor in the adhesion between paper material and relatively hard substrates, such as glasses or metals, is a non-bonded type of interaction which includes Van der Waals and/or Coulomb interactions Miwa et al. 1993;Yarovsky 1997). Thus, we limit our focus on the non-bonded interactions between the paper and glass in the present work. Nevertheless, under a normal laboratory type of environment, many other kinds of interactions due to relative humidity and organic contaminant levels are added. These factors make the fundamental understanding of relevant adhesion mechanisms all the more difficult, but equally critical.
There have been hundreds of studies on the mechanical properties of cellulose, hemicellulose and lignin in the plant cell wall as raw materials for the paper Gibson 2012;Mazeau and Charlier 2012;Molnar et al. 2018;Paavilainen et al. 2012;Zhang et al. 2015). However, studies of adhesion phenomena of the paper material highly modified from the living cell structure through the pulping process have been rare. As a relevant experiment, Goswami et al. discuss the adhesion between cellulose nanocrystal (CNC) and glass fiber/ epoxy composite (Goswami et al. 2019). Authors show that the CNC enhances the interfacial shear strength in glass fiber/epoxy composites due to interphase toughening with interpenetrating network formation. However, they did not consider a paper sheet which is a complex entanglement of cellulose microfibrils, nor the pure glass structure. There have been several computational works that focused on the fundamental aspects of adhesion between the silica glass and various organic polymer systems, from a single molecule all the way to complex thin film networks Min et al. 2016Min et al. , 2018. However, the systematic study of adhesion behavior between the paper material and the glass substrate is still lacking.
In this regard, we aim to understand the mechanism of adhesion behavior between the paper and fused glassy silica via molecular dynamics simulations. In order to elucidate the effect of main components of the paper on the adhesion, crystalline cellulose film and hemicellulose-cellulose composite are generated, and the interfaces with the silica glass are prepared for the adhesion calculation. Steered molecular dynamics (SMD) technique is employed to mimic experimental pulling and sliding tests on the interfaces, and failure behavior of the film is monitored. Based on the temporal change of free energy during test, the resultant adhesion forces and energies are calculated. Adhesion trends of cellulose and cellulose-hemicellulose composite are compared, and underlying mechanism is suggested. Finally, the effects of relative humidity on the adhesion are investigated considering it as a major environmental factor. The amount of interfacial water content is controlled, and the resultant trends of adhesion are analyzed for both kinds of interfaces.

Materials
The interface in this work is composed of paper film and a glass substrate. The basic molecular building block of paper is cellulose, which is a linear chain made of D-glucose unit and connected by b-1, 4-linked glycosidic bonds. The chemical formula of the cellulose is (C 6 H 10 O 5 ) n . Specifically, cellulose Ib monoclinic structure was chosen as a dominant crystalline form of cellulose in plants and woods (Nishiyama 2009;Zugenmaier 2008). Hemi-cellulose, the second major ingredient of the paper, contains sugar monomers such as glucose, xylose, or mannose and so on; thus, its structure is different from cellulose in the sense that it has a random and amorphous structure. We consider Glucuronoxylan (xylan) from hardwood, which consists of b-1, 4-linked D-xylopyranose backbone with 4-O-Me-a-D-Glucuronic acid residues at O-2 of one of every 6 xylans as suggested in Zhang et al (Zhang et al. 2015). Figure 1a and b show the chemical structures of cellulose Ib and the xylan molecule. For the glass substrate, amorphous silica structure is considered as an archetype of the glass whose surface is fully covered by hydroxyl groups with the surface density of 4.6 OH/nm 2 (Zuravlev 2000).

Construction of interfaces
The linear chains of cellulose Ib make strong intraand inter-molecular hydrogen bonds with each other due to a large amount of hydroxyl groups in the D-glucose unit. In the case of living plants, hemicellulose and lignin components are essential ingredients for the structure of the plant cell as well as cellulose (Gibson 2012). However, typical paper manufacturing process under high temperature and pressure destroys the structure of cell wall and removes most of the lignin components, leaving some fragments of fibril shape. Those entangled mixtures of the cellulose fragments with some hemicellulose molecules, called microfibril bundles, are the main feature of the paper material of our interest (Chinga-Carrasco et al. 2011). To focus on the molecular interactions between microfibril and glass surface, we chose a flat cellulose layer for excluding surface morphology effect such as roughness or porosity. Also, typical size of microfibrils may be up to 30 nm which can be considered to be almost flat in a few nanometer scale of our MD simulation (Kargarzadeh et al. 2017).
For construction of cellulose film structure, cellulose chains are infinitely extended to the crystallographic direction of [100], and (001) plane is exposed to the direction of silica. In terms of the coordination system shown in Fig. 1c, one monolayer is formed with 7 cellulose chains, and 4 monolayers are piled up in the z direction with a height of 7 nm. For the silica substrate, the protocol suggested by Lee et al. are used to generate the amorphous silica substrate with the lateral size of 6 Â 12 nm 2 . For the building of cellulose-silica interface, the cellulose film is initially placed at a 1 nm height above the silica substrate. The periodic boundary condition is applied to x and y directions while z direction remains fixed to prevent any neighboring interactions between the top and bottom of the simulation box. Through a thermal annealing protocol up to 600 K, the cellulose film is uniformly adsorbed to the silica to form a cellulose-  Fig. 1c.
We also consider xylan molecules which are adsorbed on the surface of the cellulose film. However, information of their precise configurations is also needed for the construction of the cellulose-xylan composite film because conformational degrees of freedom of xylan is very high due to its high flexibility. Recently, it has been reported that backbones of the xylan molecules are adsorbed as maintaining fully stretched configuration . Thus, we introduce xylan molecules of a stretched shape just above the silica surfaces so that they can be adsorbed while maintaining the stretched configuration. We consider two cases of this interface where 2 and 6 xylan molecules are adsorbed on the cellulose film. Figure 1d displays a top-down view of the cellulose surface with 6 xylans adsorbed.
To examine the impact of humidity on the papersilica adhesion, water molecules are introduced to the interface between paper and silica. For interface building, a layer consisting of water molecules with specific density is initially placed between the cellulose-xylan composite film and the silica substrate. Then the whole system is relaxed at room temperature for 3 ns, forming the composite-water-silica multilayer system. Since main adsorption sites of the molecular water on the silica surface are thought to be surface silanol groups, (Stolper 1982;Zuravlev 2000) the amount of water molecules introduced in this simulation starts from 5 H 2 O/nm 2 which is similar to hydroxyl density of 4.6 OH/nm 2 , and includes cases: 10, 15, and 20 H 2 O/nm 2 . For this range of water densities, the thicknesses of water layers relaxed between the composite and silica were calculated to be around 0.4, 0.7, 0.9, and 1.1 nm, respectively. As a relevant experiment, Verdaguer et al. measured the thickness of water layers on SiO 2 substrate, and reported that thicknesses of 0.6 nm to 1.3 nm correspond to relative humidity of 15% and 75%, respectively (Verdaguer et al. 2007).
All of the initial structures shown above except for silica substrate were constructed with the configuration biased Monte Carlo technique (Amorphous Cell module from Materials Studio 8.0 Package) (Materials Studio, Software for Technical Computation 2006). Thermal annealing and relaxation procedures were performed by using LAMMPS simulation package (Plimpton 1995) with the Interface-PCFF forcefield (Heinz et al. 2013). NVT ensembles with Nosé-Hoover thermostat and barostat were used throughout all simulations and the time step was set to be 1 fs. The cutoff for van der Waals interactions is 12 Å , and the Particle -Particle Particle -Mesh (PPPM) solver is used for the summation of long-range Coulomb interactions with the precision of 10 -4 . Lennard-Jones forcefield was used for the interaction between the water molecules whose charges are assigned using the simple point charge (SPC) model (Berendsen et al. 1987). For the interactions between water and other layers, the same Interface-PCFF forcefield is applied.

Simulation technique for adhesion calculation
To calculate adhesion properties and monitor the failure behavior at the paper-silica interface, we use steered molecular dynamics (SMD) technique to mimic the experimental pulling and sliding tests. SMD in the mode of pulling test has been utilized to estimate the interfacial adhesion and analyze conformational changes of the polymer systems Lee et al. 2017;Min et al. 2016Min et al. , 2018 and biomaterials, (Chabria et al. Dec. 2010;Davis et al. 2009) and it was also reported that SMD may be successfully applied to mimic peeling and sliding test . In this technique, a fictitious atom is connected to the center of mass (COM) of the system of interest by a virtual spring with a spring constant k, and the atom pulls it along a specific direction from the silica surface with a constant velocity v. During the process, the total force and the potential energy are calculated as following equations: where R t ð Þ is the current position of the COM of the system, R 0 is the initial center of mass of the system, and n is a unit vector along the direction in which the spring is pulled. Then the total work done is calculated as: where R f is the final position of the COM of the system. It is well known that the ensemble average of total work done can be regarded as the potential of mean force (PMF) using Jarzynski equality (Jarzynski 1997;Park and Schulten 2004): is the Boltzmann contant, and T is the temperature. The applied pulling velocity for SMD simulation is 0.75 m/s, which has been validated as slow enough to calculate the adhesion precisely in the previous reports. Lee et al.2017).
Throughout the present work, pulling and sliding tests indicate the situation where the atom is pulled in the normal and parallel direction to the interface, respectively. Figure 2 illustrates basic scheme of pulling and sliding tests and the evolution of PMF and relevant quantities which are calculable during the tests. In the case of pulling test, PMF is saturated when the paper film is completely detached from the silica substrate. The pulling distance, saturated PMF value, and maximum height of the pulling force are denoted as D det , E adh , and F max , respectively. On the other hand, since PMF just continuously increases without saturation during the sliding test, only the value of sliding force is considered as a meaningful adhesion quantity. The sliding force value is averaged over the sliding distance of 5 Å after the force is saturated and is denoted as F avg . For all kinds of interface systems considered in this work, every atom that constitute the upper and lower layer of the interface is pulled during SMD simulation. For example, in the case of interface between xylan-cellulose composite and silica substrate, COM of a composite molecule is pulled upward and that of silica substrate is pulled downward simultaneously using 'couple' command in LAMMPS. All the adhesion quantities obtained by SMD simulations were averaged over 5 samples.

Cellulose-glass interface
Adhesion at cellulose-glass interface: pulling test We start our study with the adhesion behavior between cellulose film and silica substrate. To clarify the effect of molecular interaction on the adhesion as well as estimate adhesion level, one more kind of film is additionally prepared in such a way all glucan units in the cellulose structure are substituted to mannan units. The molecular structures of two monomer units are shown in Fig. 3. The mannan unit shows a chair conformation which is identical to glucan unit except for equatorial hydroxyl in position 2. With respect to (001) plane along which the film is adsorbed to the silica surface, in the case of mannan-based film all hydroxyl groups may take a direction toward silica surface. On the other hand, in the case of cellulose film, only some of the hydroxyl groups take a direction toward the surface as indicated by green arrows in Fig. 3. Thus, one expects that mannan-based film can be directly compared to cellulose film with respect to effect of hydroxyl group on the adhesion. We note that mannan-based film is introduced only for the simulation study and is not present in the nature. Figure 4 exhibits a typical evolution of pulling force and PMF curves as a function of pulling distance for two interfaces where silica substrate forms with cellulose and mannan-based film, which shall be referred to as glucan-silica and mannan-silica interfaces, respectively. One can immediately find that both E adh and F max for the mannan-silica interface are higher than those for the glucan-silica interface. Adhesion values averaged over 5 samples show the differences in the level of E adh and F max are 40% and 15%, respectively. It is easily understood that the structural difference between the cellulose film and mannan-based film is a critical factor for their difference in adhesion. In other words, more hydrogen bonds can be simultaneously formed at the mannansilica than glucan-silica interface. One can imagine that if the other surface of the mannan-based film were adsorbed to silica surface, then the adhesion would be smaller than glucan-silica case.

Adhesion at cellulose-glass interface: sliding test
To mimic the rubbing motion between the paper sheet and the silica, the sliding test is also performed for the glucan-silica and mannan-silica interface. Figure 5 shows the temporal evolution of the sliding force applied along the x and y directions. It is observed that sliding forces result in oscillatory curves, which reflect the crystallinity of the film. Also, amplitude of the sliding force in the y direction is higher than the x direction. As mentioned in Computational details, the cellulose chain is linearly extended with covalent bonding along y direction and thus sliding plane in this direction is relatively flat. On the other hand, the chains make weak van der Waals interaction with each other in x direction, which makes for the cellulose film to experience friction motion during sliding as an Angstrom-scale 'bump'. Averaged over the x and y directions, the resultant F avg for glucan-and mannan-silica interfaces are found to be 17.2 and 64.7 kcal/mol/Å , respectively. It is not surprising that F avg for manna-silica interface should be higher than that for glucan-silica interface considering the results of our pulling test. However, it is remarkable that the difference in F avg between the two interfaces is more than three times. This result indicates that angular and torsional interactions between the surface hydroxyl groups of the cellulose and silica provide more contribution to F avg than the pulling test. Since the sliding process increases the probability of contacts between the functional groups of cellulose and glass, molecular interactions between them are also expected to increase. Overall, by comparing the pulling and sliding simulations, we can conclude that the surface density of the polar functional groups such as hydroxyl groups significantly affect the adhesion at the cellulose-silica interface.
Addition of hemicellulose Figure 6 compares adhesion forces between the cellulose only case and the cellulose-xylan composite case for pulling and sliding tests. Figure 6a shows that F max of the cellulose-xylan composite case is much lower than the cellulose only case during the pulling test. On the other hand, in the sliding test, F avg of the cellulose-xylan composite case is much higher than Fig. 4 Comparison of adhesion properties between glucan-and mannan-silica interfaces during pulling test: a E adh and b F max Fig. 5 Comparison of sliding forces between glucan-and mannan-silica interfaces during sliding test along a x and b y directions. F avg values were averaged over the sliding distance from 5 Å the cellulose only case as shown in Fig. 6b. Before analysis on the difference in adhesion, we note that it is well known that critical interaction regarding adhesion on the glass is hydrogen bonding for different kinds of polymeric materials because of the presence of much amount of silanol groups on the silica surface Lee et al. 2017;Mckenzie et al. 2017;Min et al. 2016;Wei et al. 2016;Zhang and Jiang 2002). Therefore, in this work, we can expect that hydrogen bonding between carboxylic acid or hydroxyl groups in xylan with hydroxyl groups on silica surface will play a role to enhance adhesion.
The apparent difference in the trend can be understood from analyzing the detailed geometry of the interface. When thermal annealing is performed to generate the interface between cellulose-xylan composite and the silica substrate, the lower part of the surface at which xylan molecules are adsorbed uniformly tries to cover the silica. However, since the cellulose film is relatively rigid, some cavities at the nanometer scale are created around the adsorbed xylan molecules due to their volume. Figure 7 displays nanocavities which are formed around xylan molecules at the interface between cellulose-xylan composite and silica surface where 2 xylans are added. One can recognize that vacant regions are distributed along the xylan molecules at the interface, while in other regions cellulose chains are well adsorbed to the silica so that no cavity can be found. These nanocavities weaken the strength of inter-molecular interactions at the interface, which leads to a decrease of F max when compared to the cellulose only case where the whole film is uniformly adsorbed to the silica. In the sliding test, on the contrary, it is noteworthy that the cellulose-xylan composite and silica keep their molecular interactions during the entire sliding process, and thus the interface has more chances to form strong interactions between polar functional groups. Indeed, it was observed that the xylan molecules originally adsorbed to the composite move with the silica substrate, in the opposite direction that the cellulose part of the composite film moves. This implies that the carboxylic acid groups in the xylan molecule may easily find proper binding sites to form very strong hydrogen bonds with the hydroxyl groups on the silica surface during sliding test. It also suggests that a few polar functional groups may play a role of critical factor for high sliding force as exemplified in Fig. 6b.
It is also remarkable that for both pulling and sliding tests, one can observe higher adhesion for 2 xylan added case than that for 6 xylan added case. This can be also explained by imagining the concept of nanocavity at the interface. Since 2 xylan case has more spacing between the xylans than 6 xylan case, more part of the cellulose film will be adhered to the silica substrate. Then more nanocavities will be collapsed for 2 xylan cases during thermal annealing, which leads to more cellulose-silica interactions and finally higher adhesion.

Adhesion at a low humidity
Since paper material is highly hydrophilic due to noncrystalline hemicellulose structure and the porosity between loosely entangled microfibrils, (Kargarzadeh et al. 2017) one can imagine that the adhesion level Fig. 6 Comparison of adhesion forces between cellulose only case and cellulose-xylan composite case for a pulling and b sliding tests should increase even at low humidity. To first check the effect of humidity on the paper-glass adhesion, we calculated adhesion under humid condition with 5 H 2 O/nm 2 to compare with the dry condition. The snapshots in Fig. 8 show how including water at the interface is different from the interface without water. The cellulose-xylan composite and silica substrate in the interface are denoted as small dots, and only carboxylic acid in xylan and water molecules are displayed as circles for clarity. At the low humidity interface, water molecules occupy nanocavities which caused low adhesion of the interface in the dry case. The water molecules also diffuse into both the cellulose chains and silica surface to form small clusters bridging polar functional groups such as carboxylic acid in the xylan or hydroxyl groups in cellulose and silica surface. Therefore, one can imagine that the water molecules may reinforce interactions between the composite and the silica by playing a role of 'glue' to adhere cellulose-xylan composite and silica (Zhang and Jiang 2002).
Indeed, Fig. 9 clearly shows that the adhesion levels for both pulling and sliding tests are significantly increased once the water molecules are introduced. Figure 9a shows that in the pulling test, the contribution of water molecules to the adhesion is Only water molecules and carboxylic acid groups (COOH) in xylan were emphasized with circles large enough to overcome previous low adhesion observed for the cellulose-xylan composite film. Also, the adhesion level in the sliding test strikingly increases more than three times as shown in Fig. 9c. It is thought that the sliding action evenly distributes the water molecules over the whole interface to increase adhesion further. To conclude, even low humidity conditions may significantly enhance adhesion strength between paper and glass by creating additional hydrogen bonded bridges between the surface hydroxyl groups of the silica and polar functional groups of the cellulose and xylan.

Trend of adhesion for humidity increase
Next, we consider more humid conditions up to 20 H 2 O/nm 2 to calculate the change of paper-glass adhesion. The resultant trend of adhesion is displayed in Fig. 10. Interestingly, the trend approaches a maximum around 10 H 2 O/nm 2 for both pulling and sliding tests. The trend of increasing adhesion until 10 H 2 O/nm 2 is easily understood considering the contribution of additional hydrogen bonds by more water molecules. However, it is notable that water molecules can form a well-defined layer once they cover the whole glass surface as indicated by Fig. 11a where water density is 20 H 2 O/nm 2 . Figure 11b shows that the center-of-mass distance between the cellulosexylan composite and the silica continuously increases with increasing humidity. This continuous increase of interfacial distance due to the thicker water layer leads to a reduction of the molecular interactions at the interface, and finally lower adhesion. It is expected that the tendency of adhesion increase by additional water molecules is saturated at some humidity level, whereas the tendency of adhesion reduction due to increasing separation distance will be maintained and continuously decrease. Since the former tendency will dominate the latter from 0 to 10 H 2 O/nm 2 , the adhesion energy and forces increase. Once the additional adhesion due to water molecules is saturated at 10 H 2 O/nm 2 , the decreasing strength due to increasing separation distance dominates which causes declining adhesion energy and forces after a humidity level of 10 H 2 O/nm 2 . we can conclude that the adhesion curves in Fig. 10 draw a maximum at that point. At the macroscopic scale, it is natural to expect some capillary action at the interface, especially for the high humidity scenario. However, in the current simulation the water meniscus which is qualitative evidence of capillary action was not observed even for 20 H 2 O/ nm 2 . Possible reasons for that would be twofold: first, paper fragments causing the surface contamination may include larger size fractions that cannot be modeled on the nanoscale of a MD simulation. Second, the amount of water which is supposed to be within the interface may not be enough to induce capillary behavior. Values of water densities used in this study have been estimated from the hydroxyl density of 4.6 OH/nm 2 on the bare silica surface exposed to ambient temperature and pressure. Since the paper itself can also absorb water molecules which will be distributed near the surface, there might be more interfacial water than the bare glass surface when the paper fragment is attached to it. This factor can be tested in the current model set up.
In this regard, we additionally consider two cases of water number density as 40 and 80 H 2 O/nm 2 to determine the trend of corresponding adhesion strengths. Figure 12 displays the results of adhesion calculation as a function of humidity. Both trends of E adh and F max clearly show that the change in adhesion as a function of humidity (i.e., the slope of the curve) diminishes as water number density increases.
Comparison with the force curve in Fig. 13 provides a clue for understanding this phenomenon: as the amount of water is increased, the height of maximum point of force curve becomes lower, whereas the tail of the curve at large pulling distance increases. It should be noticed the tail part represents the water interaction which resists detachment of the two layers. Therefore, one can conjecture that the capillary force does contribute to the adhesion strength to some extent and we expect that the adhesion may finally increase for even more amounts of interfacial water.

Conclusions
We investigated adhesion phenomena between paper and silica glass as an example of heterogeneous organic-inorganic interface. To understand the mechanism of molecular interactions in the order of increasing complexity, we adopted MD simulations to prepare two kinds of paper compositions to interface with the glass: pure cellulose film and a cellulose-xylan composite film. Corresponding interfaces were constructed, and relevant adhesion properties such as adhesion energy, maximum pulling force, and average sliding force were calculated by pulling and sliding tests using the SMD technique. It was found that hydrogen bonds formed by abundant hydroxyl groups in the cellulose film play a critical role in forming strong adhesion at the cellulose-silica Fig. 11 a water layer between cellulose film and silica surface at 20 H 2 O/nm 2 b Distance between cellulose and silica increases with increasing humidity, which implies molecular interactions between the layers are reduced interface. Furthermore, the interface of cellulosexylan composite and silica showed that even a small number of polar functional groups in xylan molecules significantly increases the adhesion strength. When comparing adhesion between two kinds of interfaces, adhesion force for the sliding test led to increased adhesion compared to the pulling test, which implies that rubbing motion of the paper on glass can induce significant contamination of the glass surface by paper fragments.
The effect of humidity on the adhesion of papersilica interfaces was also intensively studied. Various amount of water molecules in the range of 0 to 80 H 2 O/ nm 2 were introduced to the paper-silica interface, and the adhesion mechanism was analyzed. Consistent with the dry condition, only a low humidity significantly enhanced adhesion between cellulose-xylan composite and silica, which is responsible for another set of hydrogen bonding provided by interfacial water molecules. Interestingly, adhesion level reached a maximum point in the range of 0 to 20 H 2 O/nm 2 . We conclude that reduction of the interfacial interaction by thickening of the water layer finally leads to a decrease of adhesion at humidity levels upward of 10 H 2 O/nm 2 . Even though capillary action was not observed in the current scale of MD simulation, the possibility of adhesion increase was confirmed with increasing water content.
Ethical approval This article does not contain any studies with animals or human participants involvements by any of the authors. All authors declared that there were no ethical problems in this article.