2.3 MD simulation
The most stable molecular configuration for E adsis obtained by MD annealing simulation. Implicit-solvent atomistic MD annealing simulation are carried out at 1 K to 1800 K which can overcome the rotational barrier in flexible molecules to search for conformation in a wider range the number of annealing cycles is 4000, and the heating and cooling ramps per cycle are 1038. The simulation system is a 11.404 nm × 11.404 nm × 8.000 nm box. The COMPASSII force field39 modeled the Van der Waals interaction and the bonding interaction. The MD simulations were performed using an NVT ensemble, and the thermostat is Nose-Hoover40. All the simulation starts with the structure of energy minimization and are carried out for 50 ps in the NVT ensemble. Another 1000 ps MD annealing simulation was performed for data collection at a frequency of 1 fs.
The molecular mechanics (MM) method41 was utilized to calculated theE ads between the designed molecule and the -OH decorated SiO2 surface. The Van der Waals interaction and the bonding interaction are modeled by COMPASSII force field 39 . The electrostatic interaction is calculated by the Ewald summation method where the charges of each atom are calculated by the charge equilibrium (QEq) method42, 43.The Eads is calculated by:
where ESiO2/designed, ESiO2 and Edesignedmolecule are the energy of SiO2/designed molecule complex SiO2 and designed molecule, respectively. The implicit solvent model (εr = 78.5) is used to consider the solvation effect38. The accuracy of Ewald summation is set as 10-5 kcal/mol, and the long-range tail of the Lennard Jones potential is truncated at 1.85 nm with an extra 0.1 nm (cubic spline method) to guarantee the smoothness of the potential function. The system is considered to be equilibrium when the energy change is lower than 2×10-5 kcal/mol, force lower than 0.01 kcal/mol/nm, and displacement lower than 10-6 nm.
Results and Discussion
3.1 Multi-scale molecular design and molecular simulation forrapid underwater adhesion and long-term antifouling
The tail chain, scaffold, and head group of target molecules have different roles in practical applications. The tail chain is exposed to the water environment and plays an antifouling effect. Related studies have reported hydrophilic linear monomethoxy-terminated poly(ethylene glycol) (mPEG)44, poly(sulfobetaine methacrylate) (PSBMA)18 or poly(N,N-dimethylamino ethyl methacrylate) (PDMAEMA)45 can play a good anti-protein role in the aqueous environment. These three polymer segments are designed in Molecules 1, 2, and 3 to compare hydration degrees. The branching degree of the designed molecule will not only affect polymer aggregation morphology in solution, but also determine the surface electrostatic potential of the polymer molecule and the adsorption site with the substrate46. Therefore, four designed Molecules 1, 4, 5, and 6 were designed to be a single head-single tail, a double head-double tail, a triple head-triple tail, and a quadruple head-quadruple tail, corresponding to the branching degrees of one, two, three, four. The conjugated scaffold connects the head group and the tail chain and its rigidity plays a decisive role in polymer adhesion configuration on the substrate surface. For this reason, soft and rigid scaffolds with similar chemical groups were selected and designed as Molecules 5, 7, 8, and 9 to consider their functions in the polymer. The head group binds to various substrates in water to play an adhesion role. The DOPA has been a powerful anchor for surface modification and can adhere to virtually almost any material surface47-49. So the DOPA will be selected to be the main component of the head group, and its function will also be confirmed through analyzing the designed Molecule 5, 10, 11, and 12. To summarize, the optimal structure of the target polymer was screened by adjusting the branching numbers of polymer molecules, changing the type of head groups, tail chain, and scaffold. (Figure 1a ) In the process of theoretical calculation, molecules without dominant properties have been screened out and will not be compared in the next step.
After completing the molecule design, the calculatedE adsof all designed molecules on hydroxylated silicon dioxide surface andG solv of designed molecules with different branching degrees are adopted to screen the most suitable molecule for rapid underwater adhesion and long-term antifouling. TheE ads is calculated by the most stable adsorption structure from MD annealing simulation (Figure S1-12 ), and theG solv is acquired from DFT calculation. The calculatedE adsof 12 molecules on the hydroxylated silica is listed in Table S1 . The E ads of DOPA-PSBMA (Molecule 1), DOPA-mPEG (Molecule 2) and DOPA-PDMAEMA (Molecule 3) are -26.35, -6.47, -24.04 kcal/mol, respectively. We can find thatE ads (Molecule 1) >E ads (Molecule 2) >E ads (Molecule 3), indicating that Molecule 1 is expected to be selected as the dominant monomer for subsequent molecular design (Figure 2a ). The adsorption energy in unit kcal/mol and J/g can express the adsorption capacity and adsorption efficiency of the molecules, respectively. TheE ads of Molecule 1, 4, 5 and 6 are -213.48, -86.10, -239.32, and -197.50 J/g, respectively. It is indicated that the adsorption efficiency of these molecules on hydroxylated silica surface is in the order of E ads (Molecule 5) > E ads (Molecule 6) >E ads (Molecule 1) >E ads(Molecule 4) (Figure 2b ), which beyond our expectation. From the optimal adsorption configurations of these molecules (Figure S1 , S4-6 ), we found that the catechol groups of Molecule 1 and 5 are all attached to the substrate surface, however, those of Molecule 4 and Molecule 6 are only one in two and only three in four, respectively. It means that whether the molecule is suitable for the target molecule depends on the adsorption configuration rather than the degree of branching. The difference in molecular adsorption configuration is the discrepancy inG solv. The calculatedG solv of Molecules 1, 4, 5, and 6 are -619.60, -435.99, -502.79, and -407.36 J/g, respectively (Table S2 ). The order of G solv is G solv(Molecule 1) > G solv (Molecule 5) > G solv (Molecule 4) >G solv (Molecule 6). AsG solv increases, the hydrophilic part of the molecule prefers to be in the stretched state in solution. Combining the two effects of E ads andG solv, we found that tri-DOPA-PSBMA (Molecule 5) with three branching degrees will be the most robust adhesive polymer molecule of the four molecules. According to the reports of Li et al.50, the scaffolds connecting the head group and tail chain are crucial for the configuration of the modified surface. Molecules 5, 7, 8, and 9, all else being equal, have scaffolds of varying degrees of flexibility. Molecule 7 has a rigid connecting scaffold, whereas the scaffolds of Molecules 8 and 9 are soft with longer chain lengths. TheE adsof Molecules 7, 8, and 9 are 151.2, -163.72, and -212.99 J/g, respectively. Since Molecule 5 has the largestE ads, it continued to be selected as the dominant target molecule. The possible reason is that the scaffold with too much flexibility is not conducive to the molecular support on the substrate, which results in a close distance between the hydrophilic group and the silica substrate (Figure S8-9 ). In contrast, the rigid scaffold will not ensure the simultaneous adsorption of multiple adsorption groups on the substrate, which will reduce theE ads (Figure S7 ). TREN achieves a balance between the flexibility and rigidity of molecules, thus it is a suitable molecular scaffold for adhesion and antifouling. In addition to the scaffold connection, the molecule’s headgroup also plays a significant role in the adhesion process. TheE ads of Molecules 10, 11, and 12 are -42.46, -14.80, and -165.77 J/g, respectively. From theE ads values of Molecules 10, 11, 12 and 5, it can be inferred that the benzene ring does not contribute to the surface adsorption, but hinders it. Although the E ads of Molecule 12 with one phenolic hydroxyl group is much larger than that of Molecules 10 and 11, it is still much smaller than that of Molecule 5 which has two phenolic hydroxyls. Hence, we found that the hydroxyl groups bind firmly to the hydroxylated silica surface (Figs. 2 a and b ). Based on these analyses of E ads andG solv on the surface of hydroxylated silica substrate, Molecule 5 stands out since it owns the significant advantages of adsorption and solvation properties and is chosen to be the most suitable molecule for rapid underwater adhesion and long-term antifouling.
The reason that four designed molecules (Molecules 1, 4, 5, and 6) showed different properties in their E ads andG solv on the silica surface lies in their diverse surface properties. The molecular polarity index (MPI)51 was employed to analyze the hydrophilicity of the designed molecules. The greater the polarity and the stronger hydrophilicity of the molecule is, the larger the MPI of the molecule is. The MPIs of the single head-single tail molecule (Molecule 1), double head-double tail molecule (Molecule 4), triple head-triple tail molecule (Molecule 5), and quadruple head-quadruple tail (Molecule 6) are 34.73, 23.91, 32.27 and 26.72 kcal/mol, respectively. Molecule 5 has the largest MPI, while molecule 4 has the smallest MPI, and the difference between them was 10.82 kcal/mol. Molecules 1 and 5 with MPI greater than 30 kcal/mol contain 1 and 3 sulfonate groups exposed to water, respectively. The number of odd sulfonate groups may result in a more stretched morphology of the molecule in water, which causes a larger polar surface area of the molecule. While the MPI of molecules 4 and 6 is less than 30 kcal/mol, the possible reason is the even number of sulfonate groups cause the molecules to curl up, resulting in a reduction in the polar surface area of the molecules (Figure 2 and Table S3 ). The surfaces modified with hydrophilic molecules will lead to the good antifouling ability. Therefore, Molecule 5 is expected to be a favorite candidate for antifouling surface modifier due to its properE ads and G solv and the suitable hydrophilicity.
To further elucidate the superiority of Molecule 5 in modifying the surface antifouling properties, the HOMO-LUMO and electrostatic potential distribution (ESP) of molecules 1, 4, 5, and 6 were analyzed (Figure 2 d-e ). The calculated HOMO-LUMO gaps of Molecules 1, 4, 5, and 6 were 5.677 eV, 5.646 eV, 5.344 eV, and 5.443 eV, respectively. It is worth noting that Molecule 5 owns the smallest HOMO-LUMO gap. The HOMO energy is -5.450 eV and the LUMO energy is -0.106 eV, indicating that the electron excitation energy of Molecule 5 is quite small, which will enlarge the polarizability of Molecule 552. Due to the high polarizability of Molecule 5, the electron in the molecule is delocalized, which facilitates the modification of the hydrophilicity of the silica surface. The ESP results reinforce our speculation. The head group and scaffold are positively charged, whereas the hydrophilic group sulfonate is negatively charged. Furthermore, as the electron withdrawing ability of groups bound to the sulfur element increased, the relative electrostatic potential absolute value decreased. The isoelectric point of silica is at a pH of 2.5, so silica surfaces immersed in water are known to exhibit a negative charge density53. The adsorption of target molecules on the silica surface is mainly dominated by positively charged benzene ring moieties adsorbed on the substrate, while negatively charged sulfonate tail chains are far away from the substrate, which is consistent with the structure we speculated. The negatively-charged sulfonate group surrounded with highly hydrated layer will repel the protein contaminants in the solution and play an antifouling effect. The surface modified by Molecule 5 has a lot of negative electricity, which makes its antifouling performance greatly improved (Figure 2e andTable S3 ). MD simulation was performed on the surface of silica substrate modified by the target molecule with a polymerization degree of 10. The designed molecules quickly adhered to the substrate surface as shown from the snapshots and animation (Figure 3a andVideo 1 ). There was no structural damage to the silica surface and modified molecules during MD simulation. Physical adsorption occurred between DOPA with catechol groups and hydroxylated silicon wafers, close to the silica surface, while PSBMA stretched away from the silica surface, which could play an antifouling role. The thermodynamic and kinetic stability of the silica surface modified by Molecule 5 during MD simulation indicates that the modification method is feasible. It is consistent with all the above analysis results, demonstrating that the designed molecules can be used for rapid underwater adhesion and long-term antifouling resistance.