Bulk MoS2 has a layered structure similar to that of graphite and is built up from monolayer MoS2, which consists of a single layer of molybdenum atoms sandwiched between two layers of sulfur atoms in a trigonal prismatic structure. As the MoS2 monolayers in bulk MoS2 are weakly bonded by van der Waals interactions [49], monolayer and multi-layer MDNSs can be peeled from bulk MoS2 through chemical, liquid, or mechanical exfoliation technologies [47]. In this work, a liquid-phase exfoliation technique was used to prepare MDNSs because they provide simple routes for the large-scale synthesis of 2D materials at room temperature [3, 50, 51]. Figure 1(a) shows a photograph of suspensions containing the starting MoS2 powders and the as-prepared MDNSs in N-vinyl-pyrrolidinone (NVP). The suspension of starting MoS2 powders in NVP is silver-grey, and it becomes transparent and shows evident precipitation even after a short period of storage. After liquid exfoliation and extensive purification, a yellow-green dispersion of MoS2 nanomaterials was obtained. The suspension contains a high concentration of MDNSs. It is highly stable and shows no precipitation after being stored for one week. SEM images reveal that the starting MoS2 powders have flake-like structures with a lateral size of several micrometres, whereas the MDNSs have no specific structure due to their ultra-thin nature (Fig. (S1)). TEM investigations reveal that the MDNSs are electron transparent, and the majority of them are two-dimensional nanosheets with lateral sizes of hundreds to thousands of nanometres (Fig. 1(b)). In addition, if the nanosheet is locally parallel to the electron beam at a fold, high-resolution TEM (HR-TEM) analysis of folding at the edges can give the number of layers by direct visualization. In this work, it is proposed that the five dark lines shown in the HR-TEM of a MDNS fold reveal that the prepared MDNSs are five-layer nanosheets (Fig. 1(c)) [52]. To confirm the layer number of MDNSs, atomic force microscopy (AFM) analysis of the highly unfolded nanosheets was also carried out. AFM shows that the prepared MDNSs have a uniform thickness distribution, which indicates that the liquid-exfoliation process is highly efficient for obtaining few-layer nanosheets (Fig. 1(d)). The apparent thicknesses of two representative pieces of MDNSs measured by AFM are 3.51 and 3.44 nm, respectively (Fig. 1(e)). The apparent thickness of monolayer MoS2 was previously reported to be 0.65–0.8 nm. The layer number of the MDNSs is calculated to be approximately five, which is in good agreement with the HR-TEM analyses [53, 54].
The as-prepared MDNSs were uniformly dispersed in PVB with the aid of an ultrasonic bath, obtaining silver-grey paints. As shown in Fig. 2(a), MDNS/PVB paints remained unchanged in colour, and they exhibited no precipitation after storage for one week. Free-standing MDNS/PVB films were prepared for gas transmittance measurements. As shown in Fig. 2(b), even with very small loadings of MDNSs (~ 0.1 wt.%), a dense network of overlapping MDNSs was formed within the polymer matrix, forming a uniformly light silver grey film. This phenomenon indicates that the MDNS/PVB composite film can serve as an excellent barrier material for light, which implies that it may also hardly be permeable to gases, such as O2 and water vapour. The oxygen and water vapour transmittance rates (OTR and WVTR) of the blank and MDNS-reinforced PVB films were measured, and the results are presented in Fig. 3. It is clear that the blank PVB film (MDNS/PVB0) possesses considerably higher OTR and WVTR values of ~ 105 cm3/m2/day and 218.3 g/m2/day, respectively, revealing its poor barrier properties. This result is attributed to the porous nature of the MDNS/PVB0 film. Figure 2(c) reveals numerous nanosized pinholes at the cross-section of MDNSs/PVB0. The rapid evaporation of volatile methanol causes these pinholes. The pinholes can allow oxygen molecules to conveniently diffuse through the MDNS/PVB0 film. However, compared with MDNS/PVB0, significant decreases in the OTR and WVTR values for MDNS/PVB films are observed. Gas transmittance tests show that the OTR of MDNS/PVB films is < 1150 cm3/m2/day, and the WVTR is < 150 g/m2/day. The barrier properties of MDNS/PVB films are enhanced because, on the one hand, by forcing the molecules to permeate through the film matrices in an extremely "tortuous pathway.” MNDSs with a large aspect ratio can slow down the permeation of oxygen and water molecules through the composites; on the other hand, compared with MNDS/PVB0, the MNDS/PVB films possess denser structures (Table S1, Tafel polarization results reveal that MNDS/PVB films have lower porosity). SEM images of the cross-section of MDNS/PVB films reveal no pinholes in the matrix of PVB (Fig. 2(d-i)). The pinholes are eliminated because MDNSs can prolong the diffusion pathway of solvent molecules in the matrix of PVB, which can remarkably slow the evaporation of the solvent. Therefore, it is concluded that MBNSs are excellent barrier enhancers in reducing the gas and water vapour transmittance rates of PVB films.
To study the corrosion protection performance of the MDNS/PVB composites, the as-prepared paints were coated on brass substrates through dip-coating. The prepared samples were immersed in 3.5 wt.% NaCl solution and Tafel polarization data were collected after 4 months of immersion in 3.5 wt.% NaCl aqueous solution (Fig. 4(a)). The corrosion potential (Ecorr) shifted gradually to a more positive value, and the polarization current density tended to decrease when more MDNSs were incorporated into the PVB coatings. Through Tafel fitting, the Tafel polarization data can provide important information on the kinetics of corrosion reactions occurring at the metal/coating interface. According to the analysis of the polarization data, the corrosion current density (Icorr) of different samples can be calculated using the Stern-Geary equation:
Cathodic and anodic Tafel slopes (bc and ba, respectively) were fitted from the Tafel regions (the over-potential is more than ± 100 mV) of the cathodic and anodic branches of the polarization curves, respectively. The linear polarization resistance (Rp) was calculated from the linear regions (the overpotential was less than ± 20 mV) of the polarization curves. In addition, the corrosion rate Rcorr (mm/year) is calculated by:
where
k is a constant (3,268.5 mol/A),
Mm is the molecular weight of brass (g/mol),
n is the number of charge transfers, and
ρm denotes the density of brass (g/cm
3). The fitting results of the polarization curves show that the
Ecorr of the MDNS/PVB
0 coating is -0.285 V vs. SCE, and the
Ecorr of the MDNS/PVB coatings shifts positively from − 0.034 to 25.5 V vs. SCE with increasing MDNS incorporation in the coatings (Table S1). Furthermore, the
Icorr of MDNSs/PVB
0 is 2.34 µA/cm
2, while that for MDNSs/PVB
0.1 decreases dramatically to 3.12×10
− 4 µA/cm
2. The
Icorr decreases with increasing MDNS loading. Remarkably, the
Icorr decreases to 1.15×10
− 6 µA/cm
2 when 1.0 wt.% MDNSs are incorporated into the PVB coating. The
Rcorr of MDNS/PVB
0 is 2.75×10
− 2 mm/year, while it decreases approximately 13,000 times when only 0.1 wt.% MDNSs are used (3.66×10
− 6 mm/year, Fig.
4(b)). The
Rcorr of MDNS/PVB
1.0 is as low as 1.35×10
− 8 mm/year. The positive shift of the
Ecorr of the MDNS/PVB coatings to nobler values and the decrease in
Icorr or
Rcorr demonstrate an effective protective coating on the brass surface.
Electrochemical impedance spectroscopy (EIS) was further utilized to investigate the protection performance of the composite coatings. During the initial immersion period, the EIS spectra reveal that both the pristine PVB coating and MDNS/PVB composite coatings possess a coating resistance as high as ~ 1010-1011 Ω·cm2, indicating their excellent corrosion protection performance (Fig. 2(S)). Figure 5 shows the Bode plots of the MDNS/PVB coatings after 4 months of immersion in 3.5 wt.% NaCl. The Bode modulus plots show that the modulus |Z| at low frequencies increases with increasing MDNS incorporation (Fig. 5(a)). The |Z| value remains almost unchanged when incorporating more than 0.4 wt.% MDNSs into the matrix of PVB. Furthermore, Fig. 5(b) shows that the Bode phase plot of MDNS/PVB0 includes a peak at high frequencies and a peak located in the low-frequency region (two-time constants), attributed to the coating and corrosion response, respectively. However, only one peak can be seen from the Bode phase plot of the coating reinforced with MDNSs (one-time constant), which indicates that no or extremely slight corrosion occurs at the coating/brass interface. The inhibited corrosion suggests that little electrolyte can permeate through the MDNS/PVB coatings. Therefore, an effective barrier can be fabricated at the brass surface to completely hinder the access of the electrolyte to the brass substrates when sufficient amounts of MDNSs are incorporated into the matrix of PVB. To quantitatively compare the corrosion protection performance of MDNS/PVB coatings, the Bode plots were also fitted by the electrical models presented in Fig. 5(c). With the help of ZsimpWin software, different parameters, including Rct (charge transfer resistance), Rsol (solution resistance), Rc (coating resistance), Cdl (double layer capacitance), and Cc (coating capacitance), can be calculated based on the EIS data. Among all the parameters, Rc is used to evaluate the corrosion protection performance of the MDNS/PVB coatings because the corrosion protection performance of a polymeric coating is generally monotonically related to its Rc value. The fitting results of Rc are shown in Fig. 5(d). The Rc of MDNS/PVB0 is 4.28×103 Ω·cm2, while it dramatically increases to 2.22×108 Ω·cm2 when incorporating only 0.1 wt.% MDNSs. Furthermore, with increasing MDNS loading, the Rc of the MDNS/PVB coatings gradually increases to ~ 5×1010 Ω·cm2 (MDNSs > 0.4 wt.%). The EIS results reveal that the MDNS/PVB coatings can provide better protection for metals than the blank PVB coating.
To further confirm whether MDNS/PVB coatings can serve as corrosion protective coatings, we removed all the coatings and characterized the underlying brass substrates after 4 months of immersion in 3.5 wt.% NaCl aqueous solution at room temperature. Figure 6 shows a photograph of a freshly prepared brass sample (without coating, exposed to air for several hours only) and all the immersed brass substrates. Before immersion, all brass samples show a characteristic brass lustre to the naked eye after being polished by 1,000 SiC papers (Fig. 6(a)). After immersion, a notable colour change is observed on the substrate surface of MDNS/PVB0 due to the formation of corrosion products (Fig. 6(b)). However, for MNDS/PVB coatings, the formation of corrosion products is related to MNDS loadings in the PVB matrix. For MNDSs/PVB0.1 and MNDSs/PVB0.2, visible corrosion spots can be observed on the brass surface (Fig. 6(c,d)). However, brass substrates protected by MNDSs/PVB0.3, MDNSs/PVB0.4, MDNSs/PVB0.5, and MDNSs/PVB1.0 maintain their metallic lustre, and no noticeable colour change was observed on their surfaces (Fig. 6(e-h)). When inspected under a metallographic microscope, the corroded region on the brass substrates exhibits a red and black appearance (Fig. 7(a-d)). Nevertheless, similar to newly polished brass, uncorroded brass substrates appeared highly reflective and yellow (Fig. 7(e-h)).
Raman spectroscopy was utilized to characterize the corrosion products further. Three Raman peaks can be observed in the Raman spectra of brass substrates beneath PVB reinforced with no more than 0.3 wt.% MDNSs (Fig. 8). The peak at 283 cm− 1 is assigned to the Ag mode of CuO, and the peaks at 326 and 615 cm− 1 can be indexed to the Bg modes of CuO. The Raman spectra of the corrosion products also clearly show the Raman-active modes of ZnO, indicating the existence of ZnO in the products. The peak at 438 cm− 1 is attributed to the E2 (high) mode, a first-order optical mode of wurtzite ZnO [55]. The peaks at 380, 410, and 575 cm− 1 are attributed to the A1(TO), E1(TO), and A1(LO) modes of ZnO, which are also ascribed to the first-order optical modes of wurtzite ZnO [56]. In addition, the peaks located at 203 and 334 cm− 1 are 2E2(low) and E2(high)-E2(low) modes, which are second-order features of ZnO crystals caused by multi-phonon processes [57]. The peak located at 1,072 cm− 1 is assigned to TO + LO combinations, and the peak appearing at 1,158 cm− 1 is attributed to the 2A1(LO), 2E1(LO), and 2LO modes of ZnO. Therefore, the corrosion products formed on the brass surface mainly consist of CuO and ZnO. The Raman spectra of MDNSs/PVB composite coatings with more than 0.3 wt.% MDNS loading are similar to those of a newly prepared brass substrate and exhibit no evident Raman-active vibration modes, indicating that there are few corrosion products formed on the surface of the underlying brass substrates. These results demonstrate that the MDNS/PVB coating can effectively prevent the corrosion of brass at room temperature in 3.5 wt.% NaCl over a long time scale [57–59].