Abstract
In the alkaline geological environment, crystallization blockage of tunnel drainage systems is a common engineering problem and it is difficult to treat. The microscopic mechanism of crystallization and the environmental influence factors are still unclear. Based on the molecular dynamics (MD) technique, this study establishes nanoscale models of CaCO3 and the polyvinyl chloride (PVC) pipe, which is commonly used in drainage systems, under different alkaline environments. The goal is to study the interfacial interaction between CaCO3 and PVC and to reveal the effect of the alkaline environment on the adsorption of CaCO3 by PVC at the atomic scale. Analysis of the adsorption properties predicted by the CaCO3-PVC MD model reveals that CaCO3 molecules attract each other and form many atomic clusters at approximately 0.11 nm from the PVC interface. The peak difference between the strongly alkaline solution and the pure water solution at this distance reaches 31.6%. An in-depth exploration of the differences in adsorption between CaCO3 and PVC under different alkaline environments indicates that the mobility of CaCO3 rises gradually as the alkalinity of the solution increases. In particular, the mobility of CaCO3 in strongly alkaline solutions is approximately 60% higher than that in pure water. Moreover, as the alkalinity of the solution increases, the binding energy of the interface increases, the affinity of the interface increases, and the CaCO3 adsorption capacity gradually increases. The results of laboratory experiments were consistent with the MD simulation results, which indicates that MD simulation can play an important role in the design and evaluation of engineering practice. The innovation of this paper is to try to use the molecular dynamic (MD) technique in the field of materials to explain the practical problems in the field of traditional civil engineering, and the feasibility of molecular dynamic simulation is verified by indoor simulation experiments. The findings of this study can help for a better understanding of crystallization patterns in tunnel drainage pipes in alkaline geological environments and attempt to provide a theoretical basis and new ideas for solving this problem.
1. Introduction
The blockage of tunnel drainage systems due to crystallization occurs widely in various regions of China. Such blockages could cause the gradual rise of the groundwater level behind the lining and increased water pressure on the lining. Consequently, the tunnel drainage system may fail, along with cracking of the lining, water leakage from the lining, and severe deformation of the lining structure [1–4]. In the past two years, many studies have been carried out on drainage pipe blockages in China and abroad. Ye et al. proposed three stages of crystallization, internal and external factors affecting crystallization, and corresponding prevention measures [5]. Zhang et al. surveyed many tunnels in Chongqing, China, and found that tunnel drainage system blockage due to crystallization was widespread and that this phenomenon should be taken seriously [6]. Jung et al. used the magnetic method and quantum stick to treat drainage pipe blockages, and both methods achieved good results [7]. Liu et al. proposed that the use of flocking polyvinyl chloride (PVC) pipes has a certain effect on the removal of crystallization in the drainage pipe, and the villus length should be selected according to the water flow velocity or the water pressure to fully exert the effect of villi in removing the crystallization [8]. Tian et al. proposed that the crystallization problem of drainage pipes can be mitigated by optimizing the concrete material and mix ratio, reducing the contact between groundwater and the concrete, improving the drainage capacity of the drainage system, preventing CO2 from entering the drainage pipe, and establishing an inspection system [9]. Dietzel and Rinder et al. conducted field investigations and tests on multiple tunnels in Austria and classified the water quality flowing through the tunnel drainage system into three types, i.e., groundwater like solution, strong alkaline drainage solution, and solution with an alkalinity between the first two types. They believed that blockage due to crystallization is caused by two main processes, namely, the chemical reaction with the shotcrete for initial support of the tunnel before the groundwater enters the drainage pipe and the crystallization of the mixed solution in the drainage pipe due to environmental factors [10–12].
According to the study by Zhai, the acidity and alkalinity of the solution are one of the main factors affecting blockage due to crystallization [13]. Xiang et al. conducted an indoor experimental study on the crystallization of drainage pipes under alkaline conditions and concluded that the higher the pH of the solution is, the higher the crystallization efficiency in the test tube [14]. Ye et al. conducted a test on tunnel drainage system blockage due to crystallization and proposed that the formation and deposition of crystals were completed in a highly alkaline, high pH solution [15].
The abovementioned experiments have achieved good results and clarified that the alkaline environment could affect the formation of crystallization. However, recent studies on tunnel drainage system blockage due to crystallization are still at the macro level. The existing drainage systems are mostly concealed, and the crystallization-related problems involve the intersection of several disciplines, such as inorganic and organic chemistry. Hence, there is no unified understanding of the formation mechanism of crystals at the microscopic level.
In this paper, the commonly used Materials Studio software is employed to study the crystallization pattern of a drainage pipe in an alkaline environment from the perspective of molecular dynamics (MD). This study mainly focused on two aspects: firstly, the dynamic analyses of various particles in the system are investigated to clarify the adsorption characteristics and interaction between CaCO3 and PVC in an alkaline environment; secondly, the results of molecular dynamics were verified by macroscopical experiments, and it was proved that the alkaline environment was favorable for the adsorption of CaCO3 on PVC surface. The model proposed in this paper provides an atomic scale view for the adsorption of CaCO3 on PVC surface, which helps to understand the effect of an alkaline environment on CaCO3 adsorption at the micro-nano scale, it provides a theoretical basis for the follow-up treatment of crystal blockage in the drainage system.
2. Methods
2.1. Computational Method
2.1.1. Basic MD Principles
The basic idea of MD simulation is to treat the system to be studied as consisting of many interacting particles. The motion of each particle follows the classical equations of motion (Newton’s equations, Hamilton’s equations, or Lagrangian’s equations). By analyzing the forces on the individual particles in the system, the equations of motion of the multiple particles of the system are solved numerically, the coordinates and momentum of these particles are obtained at each moment, and the microscopic states consisting of the coordinates and momentum of the particles are averaged over time using statistical methods, and macroscopic properties, such as pressure, energy, and temperature of the multiple systems, are calculated [16].
Considering a system containing N molecules, the energy of the system is the sum of the total potential energy and the kinetic energy of the molecules of the system. The total potential energy is a function of the position of each atom in the molecule, represented as [17]. In general, the potential energy can be divided into two parts: as shown in the following formula:where is the potential energy, is the van der Waals interactions, is the internal potential energy.
The van der Waals interaction can generally be approximated as the sum of the van der Waals forces between atom pairs, as shown in
The internal potential energy of the molecule is the sum of the potential energy of various internal coordinates (e.g., bond stretching and bond angle bending) [18].
According to classical mechanics, the force on any atom i in the system is the potential energy gradient, namely,
In turn, according to Newton’s second law, the acceleration of atom i is given as follows:
By integrating Newton’s second law with respect to time, the velocity and position of atom i after time t can be predicted as follows:where is the velocity and “0” is the initial value of each physical quantity.
The basic principle of MD calculation is based on Newton’s laws of motion. That is, the potential energy of the system is first calculated from the position of each molecule in the system, the force and acceleration on each atom in the system are calculated from (3) and (4), and then in (5), , the position and velocity of each molecule after can be obtained. represents a very short time interval. The abovementioned steps are repeated to calculate the potential energy of the system from the new position, calculate the force and acceleration of each atom, and predict the position and velocity of each molecule after . Repeating this calculation cycle can yield information on the position, velocity, and acceleration of the molecules in the system at each time point.
2.1.2. Introduction of the Force Field
The molecular force field is the basis of the MD simulation. Choosing different force field models will result in different atomic topological structures and motion behaviors, thus determining the consistency between the simulation results and the actual behaviors. In simple terms, the molecular force field is the total potential energy of the system, which is composed of a set of potential functions and a set of force constants, and the potential functions and force constants obtained from different perspectives constitute different force field models. Therefore, there are various molecular force field models, such as MM, COMPASS, and MMFF.
The COMPASS force field is the first to unify the force fields of the organic molecular system and the inorganic molecular system that were previously treated separately. The COMPASS force field can simulate small molecules and polymers, some metal ions, metal oxides, and metals. The MD simulation in this paper mainly involves polymers and inorganic molecules. Therefore, we selected the COMPASS force field because it can obtain a reasonable model to describe the mixture of two types of systems. This force field can describe the interaction force of atom pairs on different molecular chains to capture the adsorption mechanism between CaCO3 and PVC pipes and between ions and CaCO3.
2.1.3. On-Site Investigation
To determine the parameters selected for molecular dynamics simulation and make them closer to the actual situation on the site, the crystals and water on the site were sampled and analyzed.
Crystals and water samples from the Miaoziping tunnel and Liaoshan tunnel were collected and analyzed Both tunnels are newly built tunnels that have not yet begun service, and the blockage on the site is shown in Figure 1. The collected samples were analyzed by scanning electron microscopy (SEM) and X-ray diffraction (XRD) provided by the Chongqing Mineral Resources Supervision and Testing Centre of the Ministry of Land and Resources. The SEM test conditions were as follows: Apreo S HiVac SEM, high voltage 20 kV, sample coated with platinum, temperature 23°C, and humidity 40%. The XRD test conditions were as follows: Cu target, Kα radiation, 1 mm/8 mm/2.5°/Ni filter, and the slit system (divergence slit (DS) 1°), wavelength 1.540 Å, working voltage 40 kV, working current 40 mA, starting angle 10°, ending angle 90°, step 0.01°, rate 2 deg/min, temperature 23°C, and humidity 41%. The test results were processed using JADE 6 software. The content of each group of samples after phase analysis is shown in Table 1. The microstructure and XRD diffraction patterns of some samples are shown in Figures 2 and 3. A comparison with the standard patterns identifies the main composition of the blockage material as calcium carbonate (CaCO3, calcite, and aragonite, Figure 3(a)), sample 16 is mixed with calcium vanadinite, a major product of concrete hydrolysis, at the time of collection (Figure 3(b)).
(a)
(b)
(c)
(d)
(a)
(b)
Statistical analysis of the water quality by the Analysis and Testing Centre of the 280 Research Institute of China National Nuclear Corporation, as shown in Figure 4, shows that the groundwater of the severely blocked tunnels is mainly alkaline water with a pH of 7 to 13, and the proportion of groundwater with a pH of ≥8.5 (strongly alkaline) is over 90%. At the initial stage of the tunnel construction process, the shotcrete for initial support of the tunnel is usually mixed and sprayed at the site, and more quick-setting agents are often added to ensure that the concrete can harden quickly. Since the concrete does not even reach the initial setting stage during the spraying process, cement hydration is incomplete, the degree of hardening is low, and the porosity is high. Therefore, in tunnels where groundwater is developed, water channels are easily formed within the shotcrete. Although construction requires the initial and second layers of shotcrete, inspection for water leakage, and timely addition of drainage measures, it is still impossible to completely fill all microcracks, which can form seepage channels for groundwater. Tunnel construction often uses Portland cement; therefore, its hydration products are alkaline. After groundwater contacts the initial support of the tunnel, even trace amounts of alkaline compounds can raise the pH of the groundwater, and the concrete and groundwater could undergo a series of reactions and form an alkaline solution that could seep into the circumferential blind drainage pipes.
2.1.4. Modelling Method
First, the interface interaction model between CaCO3 and a PVC pipe was constructed. The overall form, size, and structural distribution of the model are shown in Figure 5. In MS (Materials Studio) software, single PVC chains were first constructed, and after geometrical and energetical optimization, a polymer model with a degree of polymerization of 20, a density of 1.38 g/cm3, and a size of 3 nm × 2.5 nm × 3 nm was constructed, using 10 single PVC chains. The preset structure of CaCO3 (calcite) in the MS software was imported, the crystal plane with the highest exposure probability was calculated using the morphology tools function, and the results showed that the CaCO3 (1–12) crystal plane has the highest exposure probability, with an exposure probability of approximately 91.59%. The cleave surface function was used to obtain the 1 nm CaCO3 (1–12) crystal plane, and supercell construction of this crystal plane was performed to obtain the model of the CaCO3 crystal plane with a surface size of 3 nm × 2.5 nm. The MD calculations were performed on both the established polymer model and the CaCO3 crystal plane model. First, the MD calculation was performed for 100 ps under the velocity scale temperature control mode so the model would reach an equilibrium state. Constant temperature molecular dynamics (CTMD) was then run for 200 ps in an Andersen thermostat under canonical ensemble (NVT, 298 K), and the results were output every 1 fs; these data were used as a basis for subsequent calculations. The amorphous cell function was then used to create a box of 600 water molecules with a size of 3 nm × 2.5 nm × 3 nm, and the calculation showed that an alkaline solution with pH = 14 at this size should contain 10 hydroxide ions. To simulate the ability of PVC pipes to adsorb CaCO3 in different alkaline environments, 0, 2, 4, 6, 8, and 10 hydroxide ions were placed in the water molecule box, and the positions of water molecules and ions were randomly distributed. After geometric and energy optimization of the solution box, the build layers function was used to assemble the polymer model, the CaCO3 model, and the solution box. After fixing the polymer substructure, a periodic boundary in the x, y, and z directions was applied to the overall model. To ensure that the lower surface of the polymer is not affected by the solution, a 2 nm-thick vacuum layer was established on the upper part of the overall model. The final model is shown in Figure 5, where 0–3 nm in the z-direction is the polymer model, 3–4 nm is the CaCO3 model, 4–7 nm is the aqueous solution, and 7–9 nm is the vacuum layer. After geometric and energy optimization of the overall model, the MD calculation was performed for 100 ps under the velocity scale temperature control mode, and the CTMD was run for 200 ps in an Andersen thermostat under canonical ensemble (NVT, 298 K). By recording the displacement, trajectory, and energy changes of the components in the system, the subsequent analysis was performed.
(a)
(b)
2.2. Experimental Method
Xiang et al. conducted an experimental study on the crystallization of a tunnel drainage pipe in an alkaline environment [14], as shown in Figure 6. According to their conclusions, the main component of the crystals in the alkaline environment is CaCO3 (calcite); when the solution pH is different, the crystals have certain characteristics in terms of the packing method, crystal form, and crystal size; the higher the pH value is, the more crystals in the test tube, the tighter the crystal packing, and the smaller and more uniform the crystal grain size.
3. Data Processing Methods
The main reason for tunnel drainage pipe blockage due to CaCO3 crystallization is as follows: the synergistic effect generated by the reaction of groundwater with various minerals and the contact with the concrete structure leads to a groundwater environment that tends to generate CaCO3 crystals. These crystals are then gradually adsorbed and accumulate on the inner wall of the drainage pipe during the drainage process. Thus, the blockage grows from dot to block and then from block to accumulation. Characterizing the adsorption effect of CaCO3 molecules on polymers can help qualitatively analyze the spatial interaction between CaCO3 and polymer surfaces in different alkaline environments and quantitatively analyze the ability of polymers to adsorb CaCO3 in different alkaline environments. In this paper, the interface binding energies between different phases of the entire model system were calculated through the dynamic calculation of the forcite module in the MS, and this index can be used to characterize the adsorption effect of CaCO3 on the polymer in different alkaline environments. The spatial distribution of CaCO3 molecules can be used to analyze the adsorption effect of the polymer to CaCO3. The radial distribution function (RDF) can be used to analyze the spatial correlation of the different atoms in the model to probe the interaction mechanism between CaCO3 and the polymer.
4. Analysis of the Results
4.1. RDF Analysis
The RDF is a commonly used function to characterize the structure of molecules and represents the relative probability of finding a particle at a distance from a reference particle:where is the total number of atoms, is the time step, is the distance increment, and is the number of B (or A) atoms within the range of and around atom A or B. is the volume density.
As shown in (7), the calculation of the RDF function can effectively characterize the spatial distance characteristics between atoms, thereby revealing the mechanism at the interface between CaCO3 and PVC. As shown in Figure 7, CaCO3 exhibits a peak at approximately 0.11 nm from the PVC interface, indicating that the two substances attract each other and form many atomic clusters at this distance. The live image of the molecular motion at the interface in Figure 8 reflects this prediction of the RDF function. Halogen atoms are the functional groups of halogenated hydrocarbons. According to article 1 of the relevant rules for the structural properties of organic compounds, the main chemical properties of halogenated hydrocarbons are determined by the halogen atoms [19]. The functional group of PVC is a monoatomic functional group formed by replacing hydrogen on a methyl group with a chlorine atom; therefore, its main properties are expressed in the C-Cl bond. The C-Cl bond has a certain electronegativity difference, and the difference indicates that all C-Cl bonds are polar covalent bonds (the carbon atom is partially positively charged, and the chlorine atom is partially negatively charged), and the C-Cl bond is prone to heterogeneous cleavage under the influence of external factors. In terms of energy, since the formed chloride anions are much more stable than the formed carbon cations when this polar covalent bond is in reaction, the partially positively charged carbon atom always tends to interact with other atoms or groups (electron donors), thus freeing the chloride anion [20]. In an alkaline environment, the main reaction of the C-Cl bond is the substitution of chlorine atoms by hydroxide radicals to form alcohols, and the free chlorine anions exhibit an adsorption effect on CaCO3. In a strongly alkaline environment, the peak difference at approximately 0.11 nm from the PVC interface reaches 31.6%.
4.2. Mean Square Displacement (MSD) Analysis
To study the dynamics effect of CaCO3 in different alkaline environments, the MSD was used to estimate the motion trajectory within the last 100 ps of the MD simulation. The MSD is defined as the average distances of all particles from their respective initial points when the motion time is t:where is the position of the molecular mass center at time and is the position of the molecular mass center at time .
As shown in Figure 9, with the increase in hydroxide radicals, the alkalinity of the solution increases, and the mobility of CaCO3 gradually increases. Compared with pure water, the mobility of CaCO3 in a strongly alkaline solution increases by approximately 60%. Figure 9 demonstrates that with the increase in alkalinity, the average distance of CaCO3 molecules from the initial point increases, which provides CaCO3 more opportunities to come into contact with the surface of PVC; therefore, CaCO3 tends to accumulate in large quantities on the surface of PVC, thus increasing the possibility of tunnel drainage pipe blockage due to crystallization.
4.3. Interface Binding Energy Analysis
Calculating the interface binding energy can characterize the interface affinity between CaCO3 and PVC. When two substances are attracted to each other, the binding energy is negative, and the greater the absolute value of the binding energy, the greater the interface affinity [21]:where is the average interaction energy between the polymer and the crystal surface. is the total energy of the crystal surface and the polymer. is the single-point energy and binding energy of the crystal surface. is the single-point energy and binding energy of the polymer surface, respectively, and the is the negative value of interaction energy.
As shown in Figure 10, as the alkalinity of the solution increases, the interfacial binding energy continues to increase, indicating that the interfacial affinity increases as the alkalinity of the solution increase, and the CaCO3 adsorption capacity gradually increases. The binding energy calculated here is the binding energy when CaCO3 is first adsorbed on the PVC inner wall. When CaCO3 contacts the already formed CaCO3 crystal plane, the binding energy is very strong, approximately 1500 times the binding energy when CaCO3 is adsorbed on the PVC inner wall. The effect of the alkalinity of the solution on the adsorption of CaCO3 by the CaCO3 crystal plane is relatively small, indicating that the effect of the alkaline environment on drainage pipe blockage due to CaCO3 crystallization is dominant only for the adsorption of CaCO3 on the inner wall of the PVC at the initial stage. From the numerical perspective, as the alkalinity of the solution increases, the change in the interfacial binding energy is relatively small; that is, the phenomenon of adsorption of CaCO3 by PVC is common. From the perspective of interfacial binding energy, the increase in alkalinity of the solution has less effect on the change in the adsorption capacity.
4.4. Comparative Analysis of Laboratory Experiments
Indoor tests investigated the amount of crystallization adsorbed by PVC pipes at different alkalinities. To eliminate the effect of flow rate, the effects of different water filling states (semi-filled and fully filled) on the number of crystals were also investigated in laboratory experiments. The results for semi-filled and fully filled test pipes are shown in Figure 11. After 7 days × 8 cycles, crystals were formed, as shown in Figures 12 to 13.
(a)
(b)
The trends in the figures show that in the early stage of the experiment, the number of formed crystals was relatively large in the semi-filled pipe since a low water flow rate is conducive to the formation of crystals. In general, the lower the water flow rate is, the higher the number of formed crystals [9]. As the experiment proceeded, the influence of the water-filled state on the number of formed crystals decreased. As mentioned above, the binding energy of CaCO3 in contact with the formed CaCO3 crystal surface after CaCO3 is first adsorbed to the inner wall of the PVC pipe is extremely strong, approximately 1500 times the binding energy between the PVC inner wall and CaCO3. With the accumulation of CaCO3 crystals on the inner wall of the pipe, the solution in the fully filled pipe was in contact with the full section of the pipe wall, and the efficiency of adsorption and crystallization increased. Therefore, the number of formed crystals in the fully filled pipe gradually exceeded that in the semi-filled pipe. When the pH is high, the main effect of the water-filled state on crystallization is gradually replaced by the high pH value [14].
This experiment demonstrates that the pH value of the solution is one of the main factors affecting the number of formed crystals. In both semi-filled and fully filled pipes, the increased alkalinity is conducive to the formation of crystallization, which is consistent with the conclusions of the MD simulation.
5. Summary and Conclusions
This study conducted a field investigation on severely blocked tunnels in a drainage system, carried out SEM and composition analysis of crystals causing the blockage, and determined that the CaCO3 crystal (calcite) is the main component. Groundwater from different parts of the site was collected and tested extensively and found to be mainly alkaline. In this study, MD simulation and laboratory experiments showed that the alkaline environment is conducive to the formation of blockage due to crystallization. The main conclusions are given as follows:(1)As the alkalinity of the solution increases, the adsorption effect of the PVC pipe on CaCO3 increases. At approximately 0.11 nm from the PVC interface, CaCO3 molecules attract each other and form many atomic clusters, and the peak difference between the strongly alkaline solution and the pure water solution reaches 31.6%.(2)The alkalinity of the solution increases, and the mobility of CaCO3 gradually increases so that CaCO3 has more opportunities to contact the surface of PVC and tends to accumulate on the surface of PVC, which increases the possibility of tunnel drainage pipe blockage due to crystallization.(3)As the alkalinity of the solution increases, the interfacial binding energy of CaCO3 and PVC continues to increase, indicating that the interfacial affinity increases with increasing solution alkalinity, and the CaCO3 adsorption capacity also gradually increases. However, from a numerical perspective, the change in the interfacial binding energy is relatively small, indicating that from the perspective of the interfacial binding energy, the change in the pH value of the solution has a small effect on CaCO3 crystallization.(4)The laboratory experiments indicate that the pH value of the solution is one of the main factors affecting the formation of CaCO3 crystals, and the higher the pH value of the solution is, the greater the number of formed crystals, which is consistent with the results of the MD simulation.(5)The mechanism of tunnel drainage system blockage due to crystallization still needs to be studied. From the perspective of prevention, the amount of alkali mixed in the concrete should be reduced during construction, and the appropriate amount of fly ash, silica fume, quick-setting agent, and water reducer should be used to significantly reduce concrete seepage and crystallization. However, excessive mixing of the quick-setting agent and water reducer could aggravate concrete seepage and crystallization [22]. At the same time, the maintenance of the initial support should be improved to prevent groundwater from forming a water channel before the system reaches the design strength. For tunnels with blockage due to crystallization, from the perspective of changing the internal environment, high-pressure water guns or environmentally friendly acidic reagents can be used to dredge the pipes and thus avoid damage to the lining caused by drainage system blockage.
The disadvantage of this paper is that it is difficult to accurately simulate the alkaline environment in the field at the molecular scale, and only regular research can be carried out by changing the strength of the alkaline environment. The traditional MD method is used in this paper, some scholars have used DFTB/PCM and TD-DFTB/PCM methods to calculate the spectroscopy of large systems in solution, [23] it is believed that the DFTB calculation enhanced by CPU + GPU is a more efficient and accurate dynamic method, the simulation calculation in this paper can be further improved in accuracy and efficiency [24]. However, both molecular dynamic simulation and indoor tests can better show the regularity, and provide relevant theoretical basis and ideas for the problem of crystal blockage in the tunnel drainage systems.
Data Availability
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This study was supported by Research and Innovation Program for Graduate Students in Chongqing (2020S0002); Wuhan Shenzhen Expressway (WLX (202107) YS1-001); the Sub Project of National Key R&D Plan (2021YFB2600103-01); the Cooperation between Chongqing University; and the Institute of Chinese Academy of Sciences (HZ2021009).