1. Introduction. Keywords Alloys, Corrosion, Flow, Iron, 90Cu-10Ni alloys, NaCl solution, Iron content. Paper type Research paper

Transcription

1 The role of iron content on the corrosion behavior of 9Cu-1Ni alloys in 3.5% NaCl solutions College of Engineering, University of Bahrain, Sakheer, Bahrain Abstract Purpose The purpose of this paper is to investigate the effect of iron content (2% and up to 6% Fe) on the corrosion behavior of 9Cu-1Ni alloys in 3.5% NaCl at different temperatures (23, 5 and 88C) under stagnant conditions and fluid flow (with an agitation speed of 35 and 9 RPM). The laboratory study was conducted following a failure of high iron content (up to 6%) 9Cu-1Ni heat exchanger tubes in a desalination plant. Design/methodology/approach Potentiodynamic polarization measurement (DC) was used to estimate the corrosion rate of the 9Cu-1Ni alloys in NaCl solutions under stagnant and fluid flow conditions. Findings It was found that the higher iron content cupronickel material suffered higher corrosion rates in all tests. The intensity of the corrosion attack of both materials was increased significantly with increasing experimental temperature or flow velocity. The results support a previous prediction that the presence of excess iron (well above 2%) has played a major role in corrosion failure of 9Cu-1Ni heat exchanger tubing material in seawater. Originality/value This paper explains the role of iron content on the corrosion behavior of 9Cu-1Ni alloys in 3.5% NaCl under stagnant and fluid flow conditions. Keywords Alloys, Corrosion, Flow, Iron, 9Cu-1Ni alloys, NaCl solution, Iron content Paper type Research paper 1. Introduction Copper-nickel alloys, and in particular 9Cu-1Ni and 7Cu-3Ni types, have been used extensively in seawater piping systems for over five dec. Principal applications have been in marine vessel piping and heat exchanger condenser tubing. These alloys have excellent resistance to corrosion in seawater and possess excellent mechanical properties and microfouling resistance. The resistance to erosion corrosion of these alloys allows their economical use as piping materials (Powel and Michels, ; Tuthill, 1987; Sedricks, 1994; Todd, 1967). The reliable service performance of cupronickel alloys in marine environment depends largely on its metallurgical properties and design considerations. Other factors including seawater temperature and flow regime are reported to play an important role on the corrosion resistance of these alloys (Powel and Michels, ; Tuthill, 1987; Sedricks, 1994; Todd, 1967; Burleigh and Waldeck, 1999; Lo et al., 1987; Pearson, 1972; Ezuber, 9). Of the metallurgical properties, iron content and microstructure have been reported to play major roles affecting the corrosion performance of cupronickel alloys in seawater. Minor Fe additions are necessary in order to enhance the corrosion resistance of Cu-Ni alloys. The maximum iron content of 9Cu-1Ni alloys is recommended as 2 percent The current issue and full text archive of this journal is available at 59/4 (212) q Emerald Group Publishing Limited [ISSN ] [DOI 1.118/ ] Fe because more iron is difficult to keep in solid solution and its precipitation increases corrosion attack (Burleigh and Waldeck, 1999; Lo et al., 1987; Pearson, 1972). Commercial 9Cu-1Ni alloys complying with international standard contain percent Fe, which increases its resistance to general corrosion, erosion and impingement corrosion resulting from the turbulent flow of water containing air bubbles and silt flowing at a high velocity (Powel and Michels, ; Tuthill, 1987). In contrast, the addition of iron in excess of 2 percent to 9Cu-1Ni leads to severe segregation which is detrimental to the general corrosion resistance of the material (Burleigh and Waldeck, 1999; Lo et al., 1987; Pearson, 1972). The corrosion resistance of 9Cu-1Ni alloys changes with change in the solution temperature or flow velocity. Literature data have revealed an increase in corrosion rate with electrolyte temperature (Ezuber, 9; Habib, 1992; Badawy et al., 9). The 9Cu-1Ni alloys are found to be more susceptible to flow influenced attack than in stagnant solutions. Disturbed flows can lead to severe breakaway attack and rapid failures even at low Reynolds numbers (Re), and the intensity of such attack increases with increasing Re (Kear et al., 4b; Hodgkiess and Vassiliou, 5). The aim of this investigation was to study the effect of iron content (in excess of 2 percent) on the corrosion behavior of 9Cu-1Ni alloys in chloride solutions as a function of test temperature and under stagnant or fluid flow conditions (under turbulent flow regime). This study was conducted following a failure of high iron content (up to 6 percent) 9Cu-1Ni heat exchanger tubes in a desalination plant after one year of service. The main cause of the failure had been attributed to presence of high iron content (up to 6 percent Fe) in the 9Cu-1Ni material (Ragei et al., 21). 195

2 The role of iron content on corrosion of 9Cu-1Ni alloys 2. Experimental In order to substantiate the negative role of the high iron content on the corrosion failure of the 9Cu-1Ni tubes in the heat exchanger of reference (Ragei et al., 21), non-corroded samples of the failed heat exchanger tubes were used in this work. As outlined in that study, the chemical composition of the tubes, evaluated by SEM-EDX and confirmed by ferrite-scope meter, revealed the presence of up to 6 wt% iron. The corrosion behavior of the high iron content cupronickel was compared with 9Cu-1Ni alloy containing 1.95 wt% iron (, 2 percent Fe). To avoid any effect resulting from the shape of the material on the corrosion behavior, both materials were taken from 15-mm diameter heat exchanger pipes. The pipes were cut longitudinally and flattened with a roller. Samples, of size 2 2 mm, were mounted in epoxy with electrical connecting wires attached to the underside of the samples. The upper faces of test specimens (to be exposed to test solution) were prepared by wet abrasion on silicon carbide papers (down to 1, grit) washed in distilled water and air-dried. Prior to an experimental test the edges of the test samples were covered with a fresh layer of epoxy to minimize the risk of crevice attack. The electrochemical measurements were conducted using 3.5 percent NaCl solution made from analytical grade NaCl and distilled water. The saline solution was used to reduce variability resulting from conducting measurements using natural seawater. Tests were made at 23, 5 and 88C using a flat plate heater. A 5 ml beaker (open to air) with platinum counter electrode and a saturated calomel electrode (SCE) connected to potentiostat/sweep generator served as the electrochemical cell. Corrosion measurements were performed always after one hour of monitoring the open circuit potential (E OCP ) of the cupronickel specimens. Potentiodynamic scans at a rate of 1 mv/s were used to determine the anodic and cathodic Tafel slopes and to estimate corrosion rates of the test specimens. The Tafel slopes were determined by polarizing fresh samples starting at E OCP and going anodic or cathodic directions in a larger potential range. The polarization resistance was determined by polarizing the samples starting at 1 mv below the E OCP to 1 mv over this potential. The Tafel scans, repeated linear polarization measurements were undertaken under stagnant and fluid flow conditions and performed on all specimens. The instantaneous corrosion current was calculated from the polarization resistance and Tafel slopes using the Stern-Geary equation (Mansfeld, 1977): i corr ¼ B R p ¼ b a b c 2:33R p ðb a þ b c Þ where R p slope of the polarization resistance plot, V-cm 2, (E-i diagram at E E OCP ^1 mv), b a and b c anodic and cathodic Tafel constants, respectively, (V dec 2 1 ) and i corr corrosion current density (A cm 2 2 ). The penetration rate (in mmpy) was calculated using Faraday s law: CRðmmpyÞ ¼ Z i t 1 r n F where Z is average atomic weight (<63.6 g mol 2 1 ), n is number of electrons (two electrons), F is Faraday s constant (96,5 A s mol 2 1 ), t is time (s) (31,536, s yr 2 1 ), i is current ð1þ ð2þ density (A cm 2 2 ) and r is density of 9-1 cupronickel material (<8.945 g cm 2 3 ). The role of iron content on the corrosion rate of 9Cu-1Ni alloys under a turbulent flow regime was evaluated by agitation, i.e. stirring the test solution using agitator/heater (Type BIBBY B212) in the same electrochemical cell as mentioned above. The flow was driven solely by a magnetic stirrer positioned at the center of the beaker. The drive (agitator) had the capability of spanning the range of 1-1, rpm. The mean flow through the test section was controlled at 35 and 1, rpm, equivalent to Reynold s numbers (Re) of 12,88 and 33,12, respectively, by adjusting the agitator speed. The Reynolds number equation used to define the flow rate if it is laminar, turbulent or transient is given as (Geankoplis, 1993): Re ¼ D2 Nr m where Re is Reynolds number, D is diameter of the agitator ( ¼.35 m), N No. of rotations (rot/s), r and m are the density (kg/m 3 ) and viscosity (Pa s) of the solution, respectively. The flow was laminar in the container if Re, 1, turbulent for Re. 1 4, and for the range between 1 and 1 4, the flow was transitional being turbulent at the magnetic stirrer and laminar in remote parts of the vessel. The selected flow of 35 and 1, rpm (Re ¼ 12,88 and 33,12, respectively), therefore, lay in the turbulent flow regime. 3. Results Volume 59 Number Stagnant condition Open circuit potential measurements Open circuit potentials, E OCP, of Cu-1Ni-2Fe and Cu-1Ni- 6Fe materials were monitored over 1 h in 3.5 percent NaCl solutions at 23, 5 and 88C. Both alloys showed negative shifts in E OCP with time after initial immersion until a relatively stabilized value of potential was achieved. Steady state potential of both materials was established within 3 min of immersion and their values are given in Table I. E OCP values for the higher iron content material obtained at all test temperatures were almost similar to the 2 percent Fe cupronickel material. Generally, as the electrolyte temperature was increased E OCP became more negative and the relationship was approximately linear (Figure 1). Such a behavior has been reported for copper materials in aerated 1 or 3 percent NaCl solutions (Boden, 1971). The negative shift of E OCP with temperature is in good agreement with the literature (Ezuber, 9; Habib, 1992) and corresponds well with the reported increase in corrosion current of 9Cu-1Ni alloys in seawater Polarization measurements The anodic and potentiodynamic polarization curves, including R p measurements, were obtained to examine the influence of iron content on the potential-current profile and corrosion current of 9Cu-1Ni alloys in chloride solutions at different test temperatures (23, 5 and 88C). Figures 2 and 3 show typical sets of anodic and cathodic polarization curves for Cu-1Ni-2Fe and Cu-1Ni-6Fe materials subjected to temperature variations in chloride solutions under stagnant conditions. Generally, the anodic curve of both materials exhibited a typical Tafel-type active region, where log i was proportional to E, followed by a window for possible film formation, leading to maximum peak current ð3þ 196

3 The role of iron content on corrosion of 9Cu-1Ni alloys Volume 59 Number Table I Corrosion behavior of 9Cu-1Ni alloys in quiescent aerated 3.5 percent NaCl under solutions Temperature 238C 58C 88C Material Cu-1Ni-2Fe Cu-1Ni-6Fe Cu-1Ni-2Fe Cu-1Ni-6Fe Cu-1Ni-2Fe Cu-1Ni-6Fe E OCP (mv) SCE R p MVcm b a mv/dec b c mv/dec i corr (A cm 2 2 ) i crit (A cm 2 2 ) CR (mmpy) Figure 1 Effect of temperature on the open circuit potential (E OCP ), measured after 1 h of immersion, for Cu-1 Ni alloys in aerated 3.5 percent NaCl solutions Open Circuit potential mv (SCE) % NaCl solution Temperature ( C) Cu-1Ni-6Fe Cu-1Ni-2Fe Figure 2 Effect of test temperature on the potentiodynamic polarization curves of Cu-1Ni-2Fe in aerated 3.5 percent NaCl solutions Figure 3 Effect of test temperature on the potentiodynamic polarization curves of Cu-1Ni-6Fe in aerated 3.5 percent NaCl solutions Potential mv (SCE) Cu-1Ni-6Fe, 3.5% NaCl solution, Stagnant condition 5 C 23 C 8 C Current Density (A cm 2 ) Figure 4 Effect of test temperature on the anodic polarization curves of Cu-1Ni-2Fe prior to limiting current densities 6 4 Cu-1Ni-2Fe, 3.5% NaCl, Stagnant Condition 4 3 Cu-1Ni-2Fe, 3.5% NaCl, Stagnant Condition 5 C POTENTIAL mv (SCE) 23 C 4 8 C 6 23 C 5 C CURRENT DENSITY (A cm 2 ) POTENTIAL mv (SCE) 8 C 8 C CURRENT DENSITY (A cm 2 ) density and subsequent film or metal dissolution, giving a limiting current density at higher potentials (this is clearly shown in Figures 4 and 5). The peak maximum, minimum and limiting current densities were observed to be dependent on test temperature and iron content. As shown in Figures 4 and 5, the current densities (critical or limiting) of both materials attained higher values at higher test temperature. The peak current maximum and current minimum appeared to diminish with increasing electrolyte temperature. In terms of iron content, the high iron cupronickel material exhibited lower current densities (critical or limiting) by comparison with the 2 percent Fe cupronickel alloy at all test temperatures (for example, the magnitude of i crit for the 2 and 6 percent Fe cupronickel materials at 238C were.61 and.39 A cm 2 2, respectively. At 88C, the respective values were.86 and.68 A cm 2 2 ). 197

4 Figure 5 Effect of test temperature on the anodic polarization curves of Cu-1Ni-2Fe prior to limiting current densities Potential mv (SCE) The role of iron content on corrosion of 9Cu-1Ni alloys 6 4 Cu-1Ni-6Fe, 3.5% NaCl solution, Stagnant condition 23 C 5 C 8 C Current Density (A cm 2 ) Visual inspection of the sample surfaces after anodic polarization tests showed black and loose corrosion product films that scaled off easily. The film was thicker and more pronounced on the higher iron cupronickel material. This phenomenon has been reported in a previous study, which demonstrated that cupronickel materials with higher concentrations of iron (. 2 percent) caused continuous precipitates rich in Ni-Fe, resulting in the formation of black, thick and loose films that easily scaled off were formed on the alloy in seawater (Drolenga et al., 1983). Analysis of the anodic curves for both materials gave Tafel slopes of approximately V dec 2 1 for the temperature range 23-88C. As seen in Table I, the presence of high iron content seemed to have an insignificant effect on b a. The anodic Tafel slopes (of both alloys) at 238C corresponded well to 2.3RT/F (R, T and F are the molar gas constant (8.314 J mol 2 1 K 2 1 ), the absolute temperature (K) and the Faraday constant, respectively), indicating that concentration polarization is the dominating factor. The measured Tafel slopes were in good agreement with those of copper alloys or pure copper in chloride solutions (Kear et al., 4a, b, 7a). In the case of cathodic polarization, the measured cathodic Tafel slopes were found to be in the range of V dec 2 1. As in the case of b a, the cathodic Tafel slope seemed to be independent of iron content. Values of b c were in good agreement with those reported in the literature (Tuthill, 1987; Kear et al., 4b; Boden, 1971). Polarization resistance (R p ) values presented in Table I show lower values for the higher iron content material at all test temperatures. The negative effect of iron content on R p became more pronounced with increasing test temperature (i.e. value of R p for the 2 percent Fe cupronickel material was almost double that of the higher iron content alloy at 238C, becoming almost three times at 58C and about four times at 88C). Corrosion rates (CR) obtained by equations (1) and (2) and presented in Table I show values of.94 and.197 mmpy for the respective Cu-1Ni-2Fe and Cu1Ni-6Fe alloys at 238C. Both materials exhibited increased corrosion rates as the solution temperature was raised, with higher values for the higher iron content alloy. The effect of excess iron on corrosion rate became more pronounced with increasing test temperature (i.e. at 238C the CR of the higher iron content material was almost double that of the 2 percent Fe cupronickel alloy, becoming almost three times at 58C and about four times at 88C). This, however, corresponded well with the R p values. 3.2 Effect of fluid flow Figures 6 and 7 show the comparison of potentiodynamic polarization curves for the respective Cu-1Ni-2Fe and Cu- 1Ni-6Fe materials carried out in 3.5 percent NaCl solution under stagnant and as function of agitation speed of 35 and 1, rpm (corresponding to Re of 12,88 and 33,12, respectively) at 238C. The measured i corr and other electrochemical parameters are tabulated in Table II. Agitation caused a negative shift in the open circuit potential, measured after one hour of immersion, compared to stagnant conditions, and E OCP shifted from around 219 mv (SCE) in stagnant conditions to around 221 mv (SCE) at 35 rpm (Re ¼ 12,88). With further increase in agitation speed, E OCP became around 2225 mv (SCE). As in the case of stagnant conditions, the presence of excess iron content appeared to have negligible effect on E OCP. Figure 6 Effect of agitation speed on the potentiodynamic polarization curves of Cu-1Ni-2Fe in aerated 3.5 percent NaCl solutions at 258C Potential mv SCE Cu-1Ni-2Fe, 3.5 % NaCl RPM 1, RPM 6 35 RPM Current Density A cm 2 Figure 7 Effect of agitation speed on the potentiodynamic polarization curves of Cu-1Ni-6Fe in aerated 3.5 percent NaCl solutions at 258C Potential mv SCE Volume 59 Number Cu-1Ni-6Fe, 3.5 % NaCl RPM 35 RPM Current Density (A cm 2 ) 1, RPM 198

5 The role of iron content on corrosion of 9Cu-1Ni alloys Volume 59 Number Table II Corrosion behavior of Cu-1Ni alloys under stirred conditions in aerated 3.5 percent NaCl under solutions at 238C Agitation speed 35 rpm 1, rpm Material Cu-1Ni-2Fe Cu-1Ni-6Fe Cu-1Ni-2Fe Cu-1Ni-6Fe E OCP (mv) SCE R p MVcm b a mv/dec b a mv/dec i corr (A cm 2 2 ) CR (mmpy) Under fluid flow of 35 rpm (Re ¼ 12,88) the anodic part of both materials showed a rapid and abrupt increase in current, a region of peak current density and a limiting current density region (Figures 8 and 9). The general trend of the anodic curves at an agitation speed of 35 rpm (Re ¼ 12,88) was very similar to those in stagnant solutions (Figures 4 and 5). However, much more active behavior (of rapidly increasing current as the potential was forced positive to E OCP ) was Figure 8 Effect of agitation speed on the anodic curves of Cu-1Ni-2Fe prior to limiting current densities Potential mv SCE Cu-1Ni-2Fe, 3.5 % NaCl 35 RPM RPM.1 1 Current Density A cm 2 1, RPM Figure 9 Effect of agitation speed on the anodic curves of Cu-1Ni-6Fe prior to limiting current densities Potential mv SCE Cu-1Ni-6Fe, 3.5 % NaCl RPM 35 RPM.1 1 Current Density (A cm 2 ) 1, RPM obtained under agitation, compared to the stagnant solution, at identical potentials (i.e. at a given potential, the magnitude of i crit attained higher values under fluid flow and increased with increasing agitation speed, Table II). The peak current maximum and current minimum appeared to diminish with fluid flow. Further increase in agitation speed (from 35 to 1, rpm) resulted in rapidly increasing current densities at identical potentials. At 1, rpm (Re ¼ 33,12), however, the peak current maximum and the current minimum were not observed. Anodic Tafel slopes measured at an agitation speed of 35 rpm (Re ¼ 12,88) and 1, rpm (Re ¼ 33,12) are listed in Table II. Tafel slopes of approximately 7 mv dec 2 1 were measured at 35 rpm (Re ¼ 12,88). At higher agitation speeds of 1, rpm (Re ¼ 33,12) the Tafel slopes decreased to 63 mv dec 2 1. This, however, is consistent with the literature (Kear et al., 4a, 7a). On freshly polished cupronickels, the cathodic response of both materials, when measured using linear negative-going sweep from E OCP, exhibited a steady increase in current until a well-defined peak at which a maximum current was attained (Figures 8 and 9). After this, the current dropped to a minimum value, followed by a gradual increase with decreasing applied potential. The cathodic current density (at identical potential) and the magnitude of maximum peak current density were observed to increase with increasing agitation speed. Table II compares R p, i corr and the corrosion rate (CR) of the 2 and 6 percent Fe cupronickel materials in aerated 3.5 percent NaCl under stirred (magnetic stirrer of 35 or 1, rpm) conditions. The results presented in Tables I and II show clearly that the corrosion rate of both materials increased by a factor of about 2.7 percent when the solution was agitated at 35 rpm (Re ¼ 12,88). Further increase in agitation speed to 1, rpm (Re ¼ 33,12) resulted in a significant increase in corrosion rate (the corrosion rate of the cupronickel alloys increased by more than five times in comparison to the quiescent conditions). Regarding the influence of Fe content, the results presented in Table II clearly show a significant decrease in R p and, as a result, a remarkable increase in the corrosion rates, for the high iron content cupronickel material (the corrosion rate of the 6 percent Fe cupronickel material in stirred or quiescent conditions increased by a factor of almost two when compared to the 2 percent Fe cupronickel alloys under similar tests). 4. Discussion The E OCP behavior of the 9Cu-1Ni alloys immediately after immersion in chloride solutions may be related to a significant 199

6 The role of iron content on corrosion of 9Cu-1Ni alloys modification of the initial corrosion product film (mainly Cu 2 O/ CuO duplex film) formed during mechanical polishing (Kear et al., 4b). On contact with the chloride environment, the more soluble Cu 2 O component either could be dissolved directly into the bulk of the solution via the cuprous di-chloride complex (CuCl 2 2 ðaqþ ) or may be re-deposited as solid cuprous chloride (CuCl 2 ) (Kear et al., 7b). The E-log i profiles of both materials are typical of the general behavior of copper or cupronickel alloys in neutral solutions containing chlorides, as reported in the literature (Kear et al., 4a, 7b). It has been assumed that the anodic portion of polarization curves prior to the active metal dissolution corresponded to the formation of adsorbed intermediate species such as CuCl 2 ads on the electrode surface (Metikos-Hukovic et al., 21). The anodic polarization curves of copper alloys and pure copper in chloride solutions are assumed to be dominated by the dissolution of copper to soluble cuprous chloride ion complex (CuCl 2 2 ) (Kear et al., 4b, 7b) possibly according to the reaction: Cu þ 2C1 2 ¼ CuCl 2 2 þ e2 In general, it has been accepted that the corrosion of freshly polished copper and copper alloys in aerated chloride solutions involves the cathodic reaction: O 2 þ 2H 2 O þ 4e 2! 4OH 2 and the anodic reaction of copper dissolution and formation of CuCl 2 2 ions. Many investigators have reported that Tafel region is a mass-transport kinetics process, involving O 2 and CuCl 2 2 to and from the corroding surface (Kear et al., 7b; Metikos- Hukovic et al., 21; Alfantazi et al., 9; Bacarella and Griess, 1973). Reaction 5 results in an increase in ph near the electrode surface, which decreases the transport rate of CuCl 2 2 from the electrode surface and favors their hydrolysis and the formation of Cu(I) oxide precipitates on the copper surface (Metikos- Hukovic et al., 21): CuCl 2 2 þ 2H 2O! Cu 2 O þ 4H þ þ 4Cl 2 As E is increased above the Tafel region, the current exhibits a region of a peak current density and a limiting current density (Figures 4 and 5). It has been assumed that for neutral solutions, the rate of anodic reaction is controlled by the formation and dissolution of adsorbed precipitates of CuCl film formed at critical current densities (Kear et al., 4b; Trmans and Silva, 1997). The apex peak current, therefore, is followed by a current minimum as surface CuCl coverage reaches its maximum (Kear et al., 7b; Milosev and Kosec, 7; Barbucci et al., 1999). Further reaction of copper and chloride ions to form CuCl is reduced to a rate equal to the rate of Cl 2 diffusion from the bulk of the electrolyte to the electrode surface (Kear et al., 4a). The results presented in Table I show clearly the detrimental effect of excess iron (well above 2 percent) on the corrosion rate of 9Cu-1Ni materials. In general, the results are consistent with the published information; that is the addition of iron in excess of 2 percent to 9Cu-1Ni leads to decrease in corrosion resistance. It has been reported that commercial 9Cu-1Ni alloys complying with international standard should contain 2 percent Fe (max.). Addition of iron in excess of 2 percent iron is difficult to keep in solid solution and leads to severe segregation and coarseness of ð4þ ð5þ ð6þ Volume 59 Number microstructure. The Fe-rich precipitates are detrimental to general corrosion resistance of the material (Singh et al., 197; Drolenga et al., 1983; Lo et al., 1987). The effect of iron content on the corrosion properties of 9Cu-1Ni materials also has been linked to the characteristics of the passive film. It has been suggested that the presence of iron in cupronickel alloys promotes the formation of a film, consisting of corrosion product containing hydrated ferric oxide that is effective in reducing the corrosion rate of cupronickel alloys (Boden, 1971; Burleigh and Waldeck, 1999; Kear et al., 4a). Improvement of corrosion resistance of the passive film, as a result of iron addition, has been attributed to the decrease in electron holes resulting in an increased electrical resistivity (Burleigh and Waldeck, 1999). It has been pointed out that, in order for the iron to be effective in increasing the corrosion resistance of cupronickel alloys, it should be dissolved in solution (i.e. iron content should not exceed 2 percent) (Lo et al., 1987; Burleigh and Waldeck, 1999). The increase in the negativity of the corrosion potential with flow has been related to the influence of mass transport on the anodic reaction mechanism at potentials close to mixed potentials (Kear et al., 7b). The effect of fluid flow in increasing current densities (at identical potentials) has been attributed to an irreversibility of the anodic equilibrium reaction of equation (4), which can be depolarized via the removal of CuCl 2 2 from the equilibrium reaction (Kear et al., 4b), i.e. the potential of the anodic reaction is made more negative (depolarized) at larger value of fluid velocity, effectively leading to increasingly more positive overpotentials at identical working potentials relative to a reference electrode (Kear et al., 4b, 7a). The cathodic polarization curves under the flow regime displayed a large depolarization of the cathodic reaction under fluid flow conditions. The occurrence of a cathodic peak has been attributed to the reduction of CuCl 2 2 formed during the electrode equilibrium period (Deslouis et al., 1988) or to the doping of cuprous oxide surface film (Kear et al., 7b). Results from Figures 6 and 7 have confirmed previous data presented in the literature (Kato et al., 198; Al-Hajji and Reda, 1993) in which agitation resulted in an increased cathodic current in aerated seawater, which was the rate determining step for the corrosion of 9Cu-1Ni alloys in seawater. The increase in corrosion rate with Re has been related to accelerated outward diffusive flux of the anodic products (metal ions, or complexes ions (e.g. CuCl 2 2 ) through a diffusion boundary layer that decreases in thickness with increasing hydrodynamic severity (Hodgkiess and Vassiliou, 5). The change of corrosion rate with Re is associated with the formation of film on the surface of the alloy. Such films lead to a change of rate control mechanism to one of transport of anodic reaction products through the film. At higher Re the substantially increased corrosion rate is associated with the absence of films that are stable under less severe hydrodynamic conditions (Hodgkiess and Vassiliou, 5). It has been reported that stirring increases i corr as a result of a decrease of mass transport resistance of O 2 diffusion to the reaction zone near the surface and of an increase of anodic dissolution with stirring (Al-Hajji and Reda, 1993). study confirmed the negative role of high iron content (well above 2 percent) on the corrosion behavior of 9Cu-1Ni alloys in chloride solutions and supported the conclusion of previous

7 The role of iron content on corrosion of 9Cu-1Ni alloys work (Ragei et al., 21) that the presence of high iron content (well above 2 percent) has a major role in the corrosion failure of 9Cu-1Ni heat exchanger tubes in seawater. 5. Conclusions The presence of 6 percent Fe in cupronickel alloys appears to have an insignificant effect on the open circuit potential when compared to the 2 percent iron cupronickel alloys. The open circuit potential for both materials shifts to the negative side with increasing test temperature or increasing fluid velocity. The increase in the negativity of the corrosion potential with flow is related to the influence of mass transport on the anodic reaction mechanism at mixed potential. Corrosion rates derived from both Tafel extrapolation and polarization resistance measurement revealed higher values with increasing test temperature or agitation speed. The 6 wt% iron-cupronickel materials always exhibited higher corrosion rates in comparison to the traditional 2 wt% iron cupronickel alloys under similar tests. The negative effect of excess iron (above 2 percent) in 9Cu-1 Ni alloys is related to the presence of Fe-rich precipitates that are detrimental to the general corrosion resistance of the material. The results obtained during this study confirmed the negative role of high iron content (well above 2 percent) on the corrosion behavior of 9Cu-1Ni alloys in chloride solutions and support the conclusion of previous studies that the presence of high iron content (well above 2 percent) has a major role on the corrosion failures of 9Cu-1Ni heat exchanger tubes in seawater. References Alfantazi, A., Ahmed, T. and Tromans, D. (9), Corrosion behavior of copper alloys in chloride media, Materials and Design, Vol. 3, p Al-Hajji, J. and Reda, M. (1993), Corrosion of Cu-Ni alloys in sulfide-polluted seawater, Corrosion Science, Vol. 49, pp Bacarella, A. and Griess, J. (1973), Anodic dissolution of Cu in flowing NaCl solutions between 25 and 1758C, J. Electrochem. Soc., Vol. 12, pp Badawy, W., Ismail, K. and Fathi, A. (9), The influence of the copper/nickel ratio on the electrochemical behavior of Cu-Ni alloys in acidic sulfate solutions, Journal of Alloys and Compounds, Vol. 484, pp Barbucci, A., Farne, G., Matteazzi, P., Riccieri, R. and Cerisola, G. (1999), Corrosion behaviour of nanocrystalline Cu9Ni1 alloy in neutral solution containing chlorides, Corrosion Science, Vol. 41, pp Boden, P. (1971), Corrosion of Cu and Cu-base alloys under conditions of boiling heat transfer I: corrosion of Cu, Corrosion Science, Vol. 11, pp Burleigh, T. and Waldeck, D. (1999), Effect of alloying on the resistance of Cu-1% Ni alloys to seawater impingement, Corrosion, Vol. 55, pp Deslouis, C., Tribollet, B., Mengoli, G. and Musiani, M. (1988), Electrochemical behaviour of copper in neutral aerated chloride solution I: steady-state investigation, Journal of Applied Electrochemistry, Vol. 18, pp Volume 59 Number Drolenga, L., Ijsseling, F. and Kolster, B. (1983), The influence of alloy composition and microstructure on the corrosion behavior of Cu-Ni alloy in seawater, Materials and Corrosion, Vol. 34, pp Ezuber, H. (9), Effect of temperature and thiosulfate on the corrosion behavior of 9-1 copper-nickel alloys in seawater,, Vol. 56, pp Geankoplis, C. (1993), Transport Process and Unit Operations, 3rd ed., Prentice-Hall, Englewood Cliffs, NJ. Habib, K. (1992), Electrochemical behavior of coppernickel alloys in sulfide-polluted seawater at moderate temperatures, Desalination, Vol. 89, pp Hodgkiess, T. and Vassiliou, G. (5), Complexities in the erosion corrosion of copper-nice alloys in saline water, Desalination, Vol. 183, pp Kato, C., Ateya, C., Castle, J. and Pickering, H. (198), On the mechanism of corrosion of Cu-9.4Ni-1.7Fe alloy in air saturated aqueous NaCl solution, II: composition of the protective layer, J. Electrochem. Soc., Vol. 127, pp Kear, G., Barker, B. and Walsh, F. (4a), Electrochemical corrosion of unalloyed copper in chloride media a critical review, Corrosion Science, Vol. 46, pp Kear, G., Barker, B., Stokes, K. and Walsh, F. (4b), Electrochemical corrosion behavior of 9-1 Cu-Ni alloys in chloride-based electrolytes, Journal of Applied Electrochemistry, Vol. 34, pp Kear, G., Barker, B., Stokes, K. and Walsh, F. (7a), Electrochemistry of non-aged 9-1 copper-nickel alloy (UNS C761) as a function of fluid flow: part 1: cathodic and anodic characteristics, Electrochimica Acta, Vol. 52, pp Kear, G., Barker, B., Stokes, K. and Walsh, F. (7b), Electrochemistry of non-aged 9-1 copper-nickel alloy (UNS C761) as a function of fluid flow: part 2: cyclic voltammetry and characterization of the corrosion mechanism, Electrochimica Acta, Vol. 52, pp Lo, S., Gibbon, W. and Hollingshead, R. (1987), Corrosion resistance enhancement of marine alloys by rapid solidification, Journal of Material Science, Vol. 22, pp Mansfeld, F. (1977), Polarization resistance measurementsexperimental procedure and evaluation of test data, in Baboian, R. (Ed.), Electrochemical Techniques for Corrosion, NACE, Houston, TX, pp Metikos-Hukovic, M., Skugor, I., Grubac, Z. and Babic, R. (21), Complexities of corrosion behavior of copper-nickel alloys under liquid impingement conditions in saline water, Electrochimica Acta, Vol. 55, pp Milosev, I. and Kosec, T. (7), Study of Cu-18Ni-2Zn nickel silver and other Cu-based alloys in artificial sweet and physiological solution, Electrochimica Acta, Vol. 52, pp Pearson, C. (1972), Role of iron in the inhibition of corrosion of marine heat exchangers a review, Brit. Corros. J., Vol. 7, pp Powel, C. and Michels, H.T. (), Copper-nickel alloys for seawater corrosion resistance and antifouling a state of the art review, Corrosion, Paper No. 627, Houston, TX. 21

8 The role of iron content on corrosion of 9Cu-1Ni alloys Ragei, O., El-Shawesh, F. and Ezuber, H. (21), Corrosion failure 9/1 cupronickel tubes in a desalination plant, Desalination and Water Treatment, Vol. 21, pp Sedricks, A. (1994), Advanced materials in marine environments, Mater. Perfor., Vol. 33, pp Singh, S., Bardes, B. and Flemings, M. (197), Solution treatment of cast Al-4.5 pct Cu alloy, Metallurgical and Materials Transactions, Vol. 1, pp Todd, B. (1967), The corrosion materials in desalination plants, Desalination, Vol. 3, pp Tuthill, A. (1987), Guidelines for the use of copper alloys in seawater, Mater. Perfor., Vol. 26, pp Further reading Volume 59 Number Popplewell, J., Hart, R. and Ford, J. (1973), The effect of iron on the corrosion characteristics of 9-1 cupronickel in quiescent 3.4 NaCl solution, Corrosion Science, Vol. 13 No. 4, pp Tomashov, N. (1966), Theory of Corrosion and Protection of Metals, Macmillan, New York, NY, p Corresponding author can be contacted at: hzubeir@eng.uob.bh To purchase reprints of this article please reprints@emeraldinsight.com Or visit our web site for further details: 22