【正文】 na of the melt pool. Thus, the S number is considered to be an important parameter in relating liquid homogenisation of the melt pool, and can be used to assess the degree of dilution in laser cladding. In fact, the parameters of Eq. (1) can be grouped into two functional terms. One focuses on the material properties that is M。 the otherfocuses on the laser cladding parameters that is P: Therefore as far as material properties are concerned, parameter M can be used to study the dilution problem。 a low value of M means the condition of weak convection is expected. To assess the dilution tendency of the various coating materials of this study, their corresponding M value was determined and is given in Table 2. The results showed that Cu has a relatively low value of M and therefore a low dilution tendency. On the other hand, the relatively large calculated value of M for Al and Mg supports the position measurements that the Al layer was seriously diluted. In fact, the low dilution tendency of Cu is important for ensuring good corrosion resistance is obtained for the top Ni layer. If a large amount of Al that is present in the Cu layer should enter the Ni layer, then various intermetallic pounds will form and this could cause serious electro chemical attack. Fig. 5 shows a series of XRD patterns of the coating along the positional gradient of the sections at various depths. The different sections were obtained by carefully polishing the specimen in a direction parallel to the surface of the coating. The results suggest a series of phase evolutions has occurred across the coating and resulted in a number of phases: (Mg),Al 12 Mg 17 , T, λ 2 , λ 1 , γ 1 , (Cu Ni), (Ni). The T phase is a ternary phase of (Cu 1?x Al x ) 49 Mg 32 [11] with some mutual replacement between Mg and Cu + Al. Whereas, λ 1 and λ 2 are Mgbased ternary LavesFriauf phases [12], and the γ 1 phase has a Cu 9 Al 4 typestructure with as mallsubs tituti on between Al and Cu atoms. The (Cu), (Ni) and (CuNi) phases are solid solutions. The results also confirmed that only (Ni) is present in the Ni layer and no intermetallic phases have formed. The microstructures corresponding to the various phases are shown in Fig. hardness profile across the coating is presented in Fig. graded microstructure gave rise to a discrete hardness profile, in that the lowest hardness value was obtained for the Mg substrate, while the maximum hardness peak was recorded in the Al–Cu–Mg ternary mix region. The significant rise in hardness is believed to be due to the presence of γ 1 , λ 1 and λ 2 phases. For the Ni layer, as was expected, the hardness only reached a moderate level of 150HV, since no intermetallic pounds were detected.Fig. 2. A crosssection of the graded coating.Fig. 4. Results of the EDS analysis of the graded coating.Fig. 3. Microstructure of the graded coating.. Corrosion properties Fig. 7 pares the potentiodynamic anodic polarisation test results of the uncoated and coated specimens. The results show that the corrosion potential (E corr ) of the laserclad graded coating was about ?580 mV, which was 1020 mV more noble than the uncoated material. Whereas, the opencircuit corrosion current density (i corr ) of the coated specimen was mA/cm 2 , which was some two orders of magnitude lower than that of the magnesium substrate. Moreover, the laserclad specimen exhibited a large span of passivity of over 500 mV, while the magnesium substrate showed no sign of passivation. These improvements were significantly higher than those obtained for the magnesiumbasedsubstrate that was surface alloyed with Al+Mn [13]. For a surface alloy of 45Al–55Mn, the values of E corr and i corr were measured to be about ?1437 mV and 2 mA/cm 2 ,respectively [13]. The cause for the inferior corrosion properties of the surface alloyed coating is considered to be due to the presence of a high concentration of Mg in the coating. When the corrosion properties of the graded coating were pared to that of a Zrbased amorphous coating on Mg substrate tested under a similar corrosion condition [14], it was found that the pitting potential of the former is 110mV higher than the latter. This study thus showed that laser cladding of graded Ni/Cu/Al coating on magnesium significantly improved its inherently poor corrosion resistance. It is considered that plete solid solubility of Cu in Ni is important for achieving good corrosion resistance of the coating. Since second phase particles are absent from the solid solution phase, electrochemical reactions will be minimised. Perhaps, more importantly is the fact that the (Ni) phase of the top layer could form a particular protective passive film with a multilayer structure under corrosion. Druska et al.[15,16] has studied the multilayer structure of Ni/Cu alloys and found that under the corrosion conditions in both alkaline and acidic solutions, Ni forms a hydroxide layer at the outmost surface, and preferential Nioxidation causes Cu enrichment to occur in the (Ni) phase layer. As a result, a copper oxide layer forms between the hydroxide layer and the copper enriched (Ni) layer. This kind of multilayer passive film structure has shown to be able to guarantee a sufficient density of Ni atoms at the alloy surface to form a continuous passive film which suppresses Cudissolution in various electrolytes, especially in high pH solutions [16]. This perhaps explains why good corrosion resistance was obtained from the (Ni) layer. In fact, when alloyed with copper, nickel is known to have very good pitting resistance in sea water。 copper enhances its corrosion resistance under a reducing environment and reduces pitting attack in seawater. In addition, these alloys, while retaining the natural corrosion resistance of pure nickel to alkalis, are known to improve the resistance to chloride ions and reducing media[17].Fig. 5. XRD patterns of the graded coating Fig. 6. Microhardness