Electrochemical corrosion behaviour of Sn-Zn-xBi alloys used for miniature detonating cords

1. Introduction

Miniature detonating cords (MDCs) have been primarily for use in the aircraft canopy severance systems [1]. A miniature detonating cord is bonded to the canopy, and upon initiation of the cord, the resulting output clears an escape path for pilots or crews [2,3]. MDCs are manufactured in various configurations, but are typically flexible cylindrical cords with an explosive core and a robust sheath/cladding material to contain the explosive material and to provide mechanical strength. The explosive core generally includes pentaerythritol tetranitrate (PETN), cyclotrimethylenetrinitramine (RDX), cyclotetr-amethylenetetranitramine (HMX), and hexanitrostilbene (HNS) [4]. The sheath is often made of metallic materials with high density and suitable mechanical properties, in particularly high ductility and soft [5]. The metallic sheath is deemed an essential component of the cord, which can release momentum/energy in the form of shattering forces to sever the canopy in a more effective way.

Conventionally, the metallic sheath materials for MDCs were made by antimonial or non-antimonial lead alloys [6]. Nevertheless, due to the health and environmental concerns associated with lead, new materials have been developed as the suitable substitutes for lead alloys. For example, Rodney et al. [7] claimed that Sn-Sb-Cu, and Sn-Bi-Cu-Ag could be used as the outer sheath materials for ignition cords and mild detonating cords. Graham et al. [8] suggested Sn-(0.5-4) wt%Ag to be the sheathing of the mild detonating cords due to the good ductility, $\widetilde{8}$8% in elongation. Additionally, our previous research has shown that the Sn-Zn-xBi alloys possess excellent combinations of strength and ductility, proving them to be the appropriate materials for the sheath of MDCs [9]. Also, Bi addition (1-5 wt%) was demonstrated to enhance both strength and ductility of the Sn-3n alloy [9].

However, in addition to mechanical properties, the corrosion performance of sheath materials is deemed another essential factor which will influence the feasibility and reliability of the cords for the specific application. Practically, the cord sheath is directly in contact with the explosive material interior. This increases the corrosion susceptibility of sheaths and degrade the long-term reliability. Furthermore, the manufacturing process of cords often involves drawing and swaging of the sheath together with explosives [10], the deformation therefore could pose a high risk of corrosion that initiates at the interface of the explosive and the metallic sheath, due to the temperature rise resulted from plastic deformation.

Thus, the present work aims to study the feasibility of developing Sn-Zn-Bi alloys as the sheath of MDCs from the perspective of corrosion properties. To the best of our knowledge, very limited study has involved the influence of Bi on the corrosion properties of Sn-Zn alloys, despite that Ahmido et al. [11,12] have reported Bi addition had essential effects on the corrosion performance of Sn-9Zn, but limited explanation for the mechanism was provided. Therefore, Sn-3Zn-xBi (x=0, 1, 3, 5, 7 wt%) alloys were prepared, and the electrochemical behaviour was investigated using potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) techniques, to assess the effect of Bi on the corrosion properties of the Sn-Zn alloy. The corrosion mechanism of the Sn-Zn-Bi alloys was studied through microstructure examination on the surface and cross section of the alloys after corrosion measurements, using X-ray diffraction and scanning electron microscopy (SEM) equipped with an Energy Dispersive X-ray Analyser (EDS). The discussion is focused on the role of Bi on the corrosion performance.

2. Experimental

2.1. Materials preparation

Sn-3Zn and Sn-3Zn-xBi (x = 1, 3, 5, and 7 wt%) alloys were prepared in the present study (Hereafter the symbol of composition unit wt% is omitted). The pure Sn, Zn, and Bi ingots with commercial purity (99.9%) were used as the raw materials. Prior to melting, each element was weighed to a specified ratio with specified burning loss compensation. The melt was prepared in a stainless-steel crucible coated with Al₂O₃ coatings and the melting was conducted in an electric resistance furnace. After melting, the melt with temperature around 360 °C was manually poured into a steel mould (preheated at 200 °C) to form casting bars. All the casting bars were 300 mm long and had a trapezoid-shaped cross section of 20 mm × 16 mm × 16 mm. The chemical compositions of the Sn-Zn-Bi alloys were determined using inductively coupled plasma atomic emission spectrometry (ICP-AES, ARCOS, Simultaneous ICP Spectrometer, SPECTRO Analytical Instruments GmbH, Germany) and the actual compositions are shown in Table 1.

The specimens for electrochemical measurements were cut along the transverse cross section of casting bars, with a thickness of 8 mm. The working surface was mechanically ground with successive SiC paper from 400 to 4000 grits. The ground surface was then ultrasonically cleaned with distilled water and dried with high pressure air for corrosion tests.

2.2. Electrochemical measurements

Electrochemical measurements were performed using EZstat NuVant Systems Inc. in 0.5 M NaCl solution at room temperature (20 ± 2 °C). The NaCl solution was prepared using analytical grade chemicals and distilled water, and was aerated by direct contact with the laboratory atmosphere. A conventional three-electrode cell configuration was employed to conduct the electrochemical measurements with the Sn-Zn-Bi alloy as working electrode. A Pt spiral wire was used as the counter electrode, and the reference electrode used in this study was saturated calomel electrode (SCE, saturated KCl). Working surface of the electrode exposed to the solution was approximately 1.0 cm². Potentiodynamic polarization curves were acquired by stepping the potential at a scan rate of 0.166 mV/s, from -250 mV to + 1000 mV with respect to the open circuit potential (OCP). Electrochemical impedance spectroscopy (EIS) curves were obtained at the open circuit potential over a frequency range from 10 mHz -100 kHz with an applied sinusoidal perturbation of 10 mV RMS (root-mean-square) potential. Prior to potentiodynamic polarization and the EIS tests, a 60-min OCP test was conducted to ensure that the working surface has reached a relatively stable state. The experimental EIS spectra were interpreted on the basis of equivalent electrical circuit using the program Zview to obtain the fitting parameters.

2.3. Microstructure characterisation

Microstructure of as-prepared alloys and corrosion products after polarization measurements were examined using a Zeiss Supra 35 V P scanning electron microscope (SEM) equipped with an Energy Dispersive X-ray Analyser (EDS). Moreover, X-ray diffraction (XRD) for phase identification of the corrosion products was carried out using a Rigaku D/max 2550 diffractometer with Cu Kα radiation.

3. Results

3.1. Microstructure of Sn-3Zn-xBi alloys

Fig. 1 shows the typical microstructure of as-prepared Sn-3Zn-xBi alloys prior to electrochemical measurements. The microstructure of the hypoeutectic Sn-3Zn (Fig. 1(a)) mainly consisted of light grey β-Sn phase exhibiting dendrites morphology and Sn-Zn eutectic in the form of alternate distribution of Sn phase and dark Zn-rich needles with very small spacing (inset). A similar microstructure was reported in hypoeutectic Sn-6.5Zn alloy [13]. By adding 1-7 wt% of Bi, the microstructure was largely changed in terms of the morphology, size, and distribution of both primary β-Sn phase and eutectics, comprising β-Sn phase, Zn-rich phase and white Bi particles [14]. Notably, misaligned Zn-rich phase in the form of relatively coarse flakes instead of well-aligned small Zn-rich needles was largely observed in Bi-containing Sn-3Zn-xBi alloys. More Bi addition led to more Bi aggregates and those were located close to Zn-rich precipitates, with individual particle at a size of approximate 1 μm (Fig. 1(f)).

3.2. Potentiodynamic polarization curves

Fig. 2 shows the potentiodynamic polarization curves of Sn-3Zn and Sn-3Zn-xBi (x = 1, 3, 5, and 7 wt%) alloys in 0.5 M NaCl solution. It was seen that all the Sn-3Zn-xBi alloys exhibited similar corrosion behaviour, evidenced by similar polarization curves. Since polarization measurements were conducted in a stagnant and naturally aerated NaCl solution at room temperature, the cathodic branch (AB) of polarization could be ascribed to the reaction with the dissolved oxygen [15]: O₂ + 2H₂O + 4e- → 4OH-. On scanning in the anodic direction to BC stage, all the alloys exhibited sharp increases in anodic current density attributable mainly to the active dissolution of Zn phase [16]: Zn + 2OH- → Zn(OH)₂+ 2e- → ZnO + H₂O. OH- would react with Zn near the interface to produce porous Zn(OH)₂, covering the interface. For all the experimental alloys, active dissolution of zinc continued with increasing potential until zincate concentration reached a critical value (point C). Afterwards, in the range of CD, the insoluble zincate salts covered the surface of the corroded samples and formed a plateau region. At this stage, the current density was found to be independent of potential over a range of 450 mV. This could be attributed to the formation of Zn and/or Sn oxide or hydroxides [17,18]. The film formation began with precipitation of Zn(OH)₂ on the surface, which might transform into ZnO with further corrosion. It should be noted that the film formed in the CD region was not a protective film as the current density was high. Abayarthna et al. [19] reported that the zinc oxidation film formed in the CD region was not a passivation film and could protect the solder from further corrosion. After point D, a sharp increase in the current density was observed, corresponding to the breakdown of the non-protective film. This could be caused by the existence of Cl^- absorbed by corrosion products and by the oxygen evolution reactions.

The detailed electrochemical parameters are summarised in Table 2. The corrosion current density (i_corr) was obtained by extrapolating the cathodic Tafel region back to the corrosion potential (E_corr) [20]. It was seen that the Sn-3Zn-1Bi alloy possessed higher value of i_corr compared with that of the Sn-3Zn alloy, suggesting a tendency to lower the corrosion resistance after a small amount (1 wt%) of Bi addition. However, in the Bi-containing Sn-3Zn-xBi alloys, the corrosion resistance increased with increasing the Bi contents, which was evidenced by the fact that the value of i_corr decreased constantly from 17.4 μA cm^-2 to 8.7 μA·cm^-2 as Bi contents was increased from 1 wt% to 7 wt%. The Sn-3Zn-7Bi alloy showed the lowest i_corr which indicates the highest corrosion resistivity among these alloys. It is noted that the Sn-3Zn-5Bi exhibited nearly same level of value (13.4 μA cm^-2) of i_corr with that of Sn-3Zn alloy (13.5 μA cm^-2), indicating the similar capability of corrosion resistance. Hence, according to the potentiodynamic polarization results, it could be concluded that the trend of corrosion resistance decreased in the order of Sn-3Zn-7Bi > Sn-3Zn-5Bi ≥ Sn-3Zn > Sn-3Zn-3Bi > Sn-3Zn-1Bi.

3.3. Electrochemical impedance spectroscopy

Fig. 3 shows the Nyquist and Bode plots of Sn-3Zn-xBi (x = 0, 1, 3, 5 and 7 wt%) alloys in 0.5 M NaCl solution at their open circuit potential. It was observed that each Nyquist plot was composed of two depressed capacitive semi-arcs (Fig. 3(a)). The Sn-3Zn-1Bi alloy displayed a smallest arc radius, indicating the lowest corrosion resistance among these alloys, as shown in Fig. 3(b). Also, larger capacitive arc radius appeared in the Bi-containing Sn-3Zn-xBi alloy with higher Bi contents. Plus, the Bode plots of |Z| vs. frequency (Fig. 3(c)) revealed that higher Bi contents resulted in higher values of impedance in the Bi-containing alloys, suggesting that Bi addition (in the range of 1-7 wt%) could decrease the corrosion susceptibility of the Bi-containing Sn-3Zn-xBi alloys. |Z|-f Bode curves at low frequency range illustrate that Sn-3Zn-7Bi shows the biggest |Z| value, agreement with the largest capacitive semi-circle of it. It is noted that these phenomena were in line with the results shown by polarization measurements. Focusing on the phase angle plots in Fig. 3(d), clearly, two peaks present near 100 Hz and 0.1 Hz, respectively, which indicates the existence of two time constant, corresponding to two capacitive loops in Fig. 3(a). The time constant at low frequencies (0.1-0.3 Hz) can be attributed to the electrical double layer formed at the interface between the alloy and the corrosion product. The capacitive semi-circle at high frequency (80-400 Hz) is related to the formation of corrosion products [21].

Fig. 4 shows two different equivalent circuits (ECs) which were used in this study to fit the experimental points. The fitted EIS results are summarised in Table 3. The goodness of fit was evaluated with the chi-squared (χ²) values, which in all cases was in the order of 10^-4-10^-3. Also, a good agreement was observed between the experimental points and the fitting curves, denoted as scattered symbols and solid lines, respectively (Fig. 3). In the ECs, R_s represents the uncompensated electrolyte resistance. As can be seen in Fig. 3, real systems do not behave as an ideal capacitor, therefore, a constant phase element (CPE) instead of a pure capacitor in the equivalent circuit was used to fit the impedance behaviour more accurately. In general, the constant phase element (CPE) is associated with the distributed surface reactivity, inhomogeneity, roughness, adsorption of species and electrode porosity [22]. R₁ and CPE₁ represent the resistance and capacitance of the corrosion product layer. R_ct and CPE_dl represent the resistance and capacitance of the electrical double layer between the interface of alloy and corrosion products. W represents the Warburg impedance describing the interface diffusion of charge species.

The evidence in Fig. 3 and Table 3 confirmed that the charge transfer resistance (R_ct) of Sn-3Zn-1Bi was much lower than that of Sn-3Zn. Smaller capacitive arc radius means lower corrosion resistance. However, the pronounced increases of R_ct was observed in Bi-containing alloys with increasing the Bi contents from 1 wt% to 7 wt%, indicating that the charge transfer process occurred in higher difficulty by adding more Bi. Interestingly, adding 7 wt% Bi resulted in the presence of diffusion-controlled impedance, which might indicate that the corrosion mechanism of the modified alloy was controlled by both charge transfer and diffusion process [23,24]. Note that, the existence of transport resistance (R_w) showed that the transport process of charge species occurred much more difficult in the Sn-3Zn-7Bi alloy [25]. According to the ECs (neglect the solution resistance), total impedance (R_t) can be extracted from R₁, R_ct and R_w (diffusion resistance), to evaluate the overall corrosion resistance. The Sn-3Zn exhibited a large value of R_t, 970 Ω·cm². When small amount of Bi was added, the total impedance of Sn-3Zn-1Bi and Sn-3Zn-3Bi alloys showed lower values of R_t, indicating a worse corrosion resistance. However, the R_t value for Sn-3Zn-5Bi and Sn-3Zn-7Bi was relatively large, approximately 1179 and 2265 Ω·cm², respectively, suggesting an enhanced corrosion resistance with a higher amount of Bi addition. It was clear that the Sn-3Zn-7Bi exhibited the highest corrosion resistance among these alloys. Thus, based on the EIS results, it was further confirmed that the corrosion resistance increased in the sequence of Sn-3Zn-1Bi, Sn-3Zn-3Bi, Sn-3Zn, Sn-3Zn-5Bi, and Sn-3Zn-7Bi. This was in agreements with the results proved by polarization measurements.

3.4. Corrosion products characterization

3.4.1. Surface characterization

Fig. 5 shows the typical SEM micrograph and element mapping of the surface of Sn-3Zn after polarization measurement. The severely corroded area denoted by the dark contrast (Fig. 5(a)) was mainly composed of Cl, O, and Zn elements, indicating that corrosion products were predominantly made of Cl, O, and Zn. The slightly-/non- corroded area represented by the light contrast in Fig. 5(a) was covered with numerous Sn and a small amount of O. It was thus deduced that Zn phase instead of Sn was selectively and preferentially corroded under the engagement of Cl- and OH- anions.

Fig. 6 presents the XRD patterns of the surface of Sn-3Zn-xBi (x = 0, 1, 3, 5, and 7 wt%) alloys after polarization measurements, for the purposes of confirming the phase composition of corrosion products. It was seen that there was still β-Sn phase remained after polarization measurements in all cases, while Zn phase was scarcely detected, further revealing that Zn-rich phase was heavily consumed. Bi phase was extensively detected in the Bi-containing alloys with Bi contents being more than 1 wt%, confirming the difficulty of Bi in being corroded in 0.5 M NaCl solution due to its relatively higher corrosion potential compared with Sn and Zn [26]. The main corrosion product for all Sn-3Zn-xBi alloys was identified as a complexed Zn hydroxyl chloride hydrates, namely simonkolleite Zn₅(OH)₈Cl₂·H₂O. Plus, a trace amount of ZnO was discovered as well.

Fig. 7 presents the typical SEM micrograph and EDS analysis illustrating the surface morphology and corresponding chemical composition of the corrosion products of Sn-3Zn after polarization measurement. From Fig. 7(a) and (b), the corroded region was covered with sparsely aggregated plate-like structures, which were confirmed simonkolleite Zn₅(OH)₈Cl₂·H₂O by EDS analysis on spectrum S1. This was in agreements with the XRD results. Enlarging the initial eutectic region (Fig. 7(c)), it was seen that the initial Zn-rich needles were dissolved and depleted, leaving narrow channels on the surface, whereas the initial β-Sn phase in the eutectic region remained, as demonstrated by EDS analysis on spectrum S2, further indicating the easier destruction of Zn-rich phase compared to β-Sn phase. Furthermore, the corrosion product exhibiting a sphere-shaped morphology was observed, which was found the aggregate of the ZnO plates (EDS analysis on S3). These sphere-like ZnO products have been previously reported by other researchers on the study of Sn-Zn alloys after immersion corrosion [27].

Fig. 8(a) and (b) presents the SEM micrographs of the surface of the Bi-containing Sn-3Zn-1Bi and Sn-3Zn-5Bi alloys after polarization measurements in 0.5 M NaCl solution, respectively. From Fig. 8(a), many Bi particles were seen to remain on the surface, as shown by arrows. These particles were confirmed Bi particles by EDS analysis (not present here). Surrounding the Bi particle, Sn phase was selectively corroded, suggesting appearance of micro-galvanic couples between the β-Sn phase and the Bi phase. For certain sites (dashed circle), Bi particles were depleted because the neighbouring β-Sn matrix was severely consumed and thus the interface bonding was consequently damaged. Also, pits and micro-cracks were observed in the corroded surface of Sn-3Zn-5Bi alloy, which might indicate weak protection of the corrosion product layers (Fig. 8(b)). Note that, pits and micro-cracks were observed not only in the case of Sn-3Zn-5Bi but in all Sn-3Zn-xBi (x = 0, 1, 3, 5, and 7 wt%) alloys.

3.4.2. Cross section characterization

For a better understanding of pitting corrosion process, cross section characterization was further carried out after the corrosion measurements. Fig. 9 shows the cross-section SEM micrograph and the element mapping of a typical pit of the Sn-3Zn-5Bi alloy after polarization measurement. The result of element mapping confirmed the strong presence of Cl, O, and Zn in the pit, suggesting that anions including Cl- and OH- had migrated inwards the alloy along the initial Zn-rich precipitates. It was thus concluded that Zn-rich precipitates could provide preferential transport paths for Cl- and OH- penetration. Notably, Bi phase was scarcely identified by the element mapping (Fig. 9(f)), indicating that this individual pit initiated and propagated without the participation of Bi particles.

Closer inspection of the pit (Fig. 9(a1)), the major pitting product was found Zn₅(OH)₈Cl₂·H₂O, which was confirmed by the EDS results. Interestingly, the pitting product, Zn₅(OH)₈Cl₂·H₂O, exhibited a porous interlinked network structure. A similar porous network-structured corrosion product was also reported on the study of Sn-Zn [27]. The pores could provide ease of transport of Cl- and OH- from solution inwards the interior of the alloys. In this way, a continuous complement of poisonous anions (e.g. Cl-) was delivered, thus, promoting the pitting propagation. In the meantime, localized cracks/breakdown appeared as well. The crack on one hand provided the channel for the penetration of corrosive medium (Cl- and OH-); on the other hand, the crack could increase the chances of removal of pitting products away from the surface, increasing the possibility of anions in the solution diffusing towards the alloy, which promoted the pitting rate kinetically.

Fig. 10 shows the cross-section SEM micrograph and the element mapping of another typical type of pits in Sn-3Zn-5Bi after polarization measurement. It was observed that the pit was covered with Cl, O, Sn, and Bi elements. A similar conclusion could be drawn that the pitting initiation and subsequent propagation were resulted from the migration of attacking Cl^- and OH^- anions inwards the alloy. EDS results (Fig. 10(b) and (c)) confirmed that pitting products contained SnO and the Sn hydroxyl chloride hydrate compounds, Sn₃O(OH)₂Cl₂. Different from the previously mentioned pit, the Zn-rich precipitate was barely identified here, while numerous Bi particles were largely detected in this pit (Fig. 10(d)). Thus, the pitting susceptibility could be enhanced in the sites where large amounts of Bi particles were assembled. This could be related to the formation of the β-Sn/Bi galvanic couples which will be discussed next.

4. Discussion

Based on the polarization and EIS results, it has been confirmed that a small amount (1 and 3 wt%) of Bi addition to the Sn-3Zn alloy could cause decreased corrosion resistance. However, when Bi contents was further increased to 5 and 7 wt%, the Sn-3Zn-xBi (x = 5, and 7 wt%) alloys exhibited higher corrosion resistance than that of Sn-3Zn. Furthermore, for Bi-containing Sn-3Zn-xBi alloys the corrosion resistance increased with increasing Bi contents from 1 wt% to 7 wt% and the Sn-3Zn-7Bi alloy showed the best corrosion resistance among them. This could be associated with: (1) the microstructure modification through addition of Bi with respect to the morphology, size, and distribution of the Zn-rich precipitates; (2) the additional micro-galvanic effect caused by Bi particles; and (3) the barrier effect of Bi phase.

In the Sn-Zn system since Zn exhibits more negative electrode potential (-0.763 V vs. SHE) than Sn (-0.136 V vs. SHE) the Zn-rich precipitates act as active anodes and can be preferentially corroded under corrosive circumstances [28]. The corrosion process begins with the dissolution of Zn at anodic sites:

Zn → Zn²⁺+2e- (1)

In this way, the initial Zn-rich precipitates could be selectively consumed, leaving the β-Sn phase (cathode) remained (Fig. 7(c)). With the participation of water and oxygen further electrochemical reactions take place, forming zinc hydroxide and/or zinc oxide:

1/2O₂+H₂O+2e- → 2OH- (2)

Zn²⁺ +2OH- → Zn(OH)₂ (3)

Zn(OH)₂ → ZnO + H₂O (4)

In the presence of Cl-, especially when the chloride concentration is higher than 0.1 M, Cl- will move towards Zn dissolution sites, causing gradual formation of insoluble zinc hydroxychloride, Zn₅(OH)₈Cl₂·H₂O [29]:

ZnO+2Cl-+6H₂O → Zn₅(OH)₈Cl₂·H₂O + 2OH- (5)

These corrosion products, ZnO and Zn₅(OH)₈Cl₂·H₂O, were readily recognised by the XRD results (Fig. 6) and EDS composition analysis (Fig. 7(b)). In this sense, Zn₅(OH)₈Cl₂·H₂O is thus expected to precipitate close to the anodic sites, i.e., Zn-rich precipitates, which was in deed the case denoted by the pit shown in Fig. 9.

In the case of Sn-3Zn, the Zn-rich precipitates displayed as small-sized needle-shaped morphologies and these Zn-rich needles were located at the Sn-Zn eutectic cells/regions. Accordingly, the eutectic regions could be reasonably regarded as the corrosion-vulnerable sites at which anodic Zn-rich phase was dissolved and depleted, while the primary β-Sn cells could be considered as the cathodic noble locations where corrosion could rarely occur. This was supported by the surface examination and composition analysis of Sn-3Zn after the polarization measurement (Fig. 7(c)).

When Bi was added, i.e., in the case of Sn-3Zn-xBi (x = 1, 3, 5, and 7 wt%) alloys, Zn-rich precipitates were dramatically increased in dimensions, and its distribution became more uniform. Those in larger size could lead to degradation of corrosion performance due to the weak bond between the coarse Zn-rich precipitates and β-Sn matrix. This was based on the idea that defects (e.g. dislocations or voids) could be easily accumulated near the interface between the coarse Zn-rich precipitate and the Sn phase [30,31]. The defects-accumulated sites could thus benefit transport of Cl- or OH- inwards and promote chemical reaction between anions and the Zn-rich phase, increasing the corrosion rates. Meanwhile, the pitting can be highly prone to initiation and propagation along the vulnerable Zn/Sn interface, which was reflected from the pit microstructure shown in Fig. 9. A similar phenomenon was reported by Liu et al. [32] who has demonstrated that the corrosion resistance is enhanced after addition of trace amount of Ti to the Sn-9Zn alloy attributed to elimination of large Zn-rich precipitates.

Also, in the Sn-3Zn alloy the corrosion-vulnerable sites, i.e., Sn-Zn eutectic “islands”, were geographically and largely isolated by a large proportion of primary β-Sn dendrites serving as noble cathodes (Fig. 1(a)). This means that noble β-Sn phase can act as the barrier, causing obstruction of corrosion [33]. On the contrary, in the Bi-containing Sn-3Zn-xBi alloys, a much more uniform distribution of the Zn-rich precipitates appeared, which means that more proportions of corrosion-vulnerable sites presented and fewer fractions of effective anodic barriers exhibited. Consequently, the Sn-3Zn-xBi (x = 1, and 3 wt%) alloys presented higher corrosion susceptibility in comparison with the Sn-3Zn alloy.

It is worthy of note that in addition to the effect attributable to the change of the Zn-rich precipitate with respect to its size and distribution, the Bi particle itself played a crucial role in affecting the corrosion resistance of the alloys. The effect caused by Bi particles may involve two aspects: galvanic effect and anodic barrier. When a small amount of Bi (1 wt%) was added to Sn-3Zn, Bi particles were randomly aggregated and discretely dispersed in the β-Sn matrix and majority of those were located near Zn-rich precipitates, as shown in Fig. 1b and c. Extra micro-galvanic couples, i.e., β-Sn phase/Bi phase and Zn-rich phase/Bi phase, may form, due to Bi phase possessing the highest electrode potential (0.293 V vs. SHE) compared with Sn (-0.136 V vs. SHE) and Zn (-0.763 V vs. SHE). Therefore, in the β-Sn/Bi micro-galvanic couples the β-Sn could be selectively resolved as the anodes, leaving Bi remained (Fig. 8(a)), causing formation of corrosion products, SnO and Sn₃O(OH)₂Cl₂ (Fig. 10). Possible chemical reactions involve [34,35]:

Sn → Sn²⁺ + 2e- (6)

Sn²⁺ + 2OH- → Sn (OH)₂ (7)

Sn (OH)₂ → SnO + H₂O (8)

3SnO + 2Cl- + 2H₂O → Sn₃O(OH)₂Cl₂ +2OH- (9)

In this way, the additional galvanic effect could have a negative effect on the corrosion resistance of Sn-Zn alloys. However, it worthy noted that the detrimental galvanic effect caused by β-Sn/Bi micro-galvanic couple was believed to be minimal because the corrosion potential difference between Bi and Sn is not large [26]. This was also reflected by the fact that no clear peaks corresponding to the corrosion products, SnO and/or Sn₃O(OH)₂Cl₂, were identified in the XRD results of the corroded surface of the Bi-containing Sn-3Zn-xBi alloys. Therefore, the galvanic corrosion theory which gives the idea that larger ratio of cathodic to anodic would lead to a severer corrosion of the anode alloy, resulting more damage during electrochemical measurement, might be not the dominant mechanism operating in the present case of Bi-containing Sn-3Zn-xBi alloys.

On the contrary, the noble barrier effect of cathodic Bi phase operates more significantly than the galvanic cathode, leading to the increased corrosion resistance with more Bi particles. Similar phenomenon was reported in the Mg alloys where the Mg₁₇Al₁₂ (β phase) has two influences on corrosion, as a galvanic cathode and as a barrier, depending on the volume fraction of Mg₁₇Al₁₂ in the Mg matrix [36,37]. The Mg₁₇Al₁₂ phase mainly serves as a galvanic cathode and accelerates the corrosion process of matrix if the volume fraction of Mg₁₇Al₁₂ phase was small. However, for a higher volume fraction, the Mg₁₇Al₁₂ phase may act as an anodic barrier to inhibit the overall corrosion of the alloy. Also, Osorior et al. [38] reported that in the Al-1.5 wt% Fe alloy more extensive distribution of Al₆Fe particles provides a better protective effect with the nobler intermetallic Al₆Fe particles “enveloping” the anodic Al-rich phase, resulting in better corrosion resistance. Notably, with increasing the Bi contents from 1 wt% to 7 wt%, the volume fraction of Bi particles was observed to be largely increased and those became better connected from each other, forming the noble barriers or “enveloping” against corrosion. Based on this, it was deduced that a larger proportion of Bi networks acting as anodic barriers in Sn-3Zn-xBi alloys with higher Bi contents was the reason for the correspondingly increased corrosion resistance.

5. Conclusions

In the present study, the effect of Bi on the corrosion performance of the Sn-Zn alloy was investigated using potentiodynamic polarization and electrochemical impedance spectra (EIS) techniques. Based on the results obtained from electrochemical measurements and microstructure examination, main conclusions could be drawn as follows:

(1) The corrosion current density of the Sn-3Zn-xBi alloys was measured in the range of 8-15 μA·cm^-2, with corrosion resistance increasing in the order of Sn-3Zn-1Bi, Sn-3Zn-3Bi, Sn-3Zn, Sn-3Zn-5Bi, Sn-3Zn-7Bi.

(2) Bi addition (1-7 wt%) caused dramatic increases in the sizes of corrosion-vulnerable Zn-rich precipitates and led to more uniform distribution of these precipitates. This was the reason that addition of 1 wt% and 3 wt% Bi increased the corrosion susceptibility of the Sn-3Zn alloy.

(3) For Bi-containing Sn-3Zn-xBi alloys, the corrosion resistance was constantly increased with increasing the Bi contents from 1 wt% to 7 wt%, due to the enhanced barrier effect of Bi particles. When Bi concentration reached 5 wt% and 7 wt%, the barrier effect of Bi phase could counteract the detrimental effect resulted from the coarsened Zn-rich precipitates, resulting in the increased corrosion resistance of Sn-3Zn-5Bi and Sn-3Zn-7Bi in comparison with Sn-3Zn.

(4) From the perspective of corrosion performance, all the investigated Sn-3Zn-xBi alloys have great potential for use as the sheath for miniature detonating cords, due to relatively good corrosion resistance.

Acknowledgments

Financial support from the National Aerospace Technology Exploitation Programme (NATEP) and Chemring Energetics UK [grant number WEAF058] is gratefully acknowledged. Dr Yan Huang is sincerely thankful for his providing instruments for potentiodynamic polarization measurements.

The authors have declared that no competing interests exist.

---