Changes in acceleration rate of chloride ions depending on clinnatic conditions . Influence of rain

Mild steel, copper and aluminium samples were exposed outdoors in two atmospheric test stations located in Havana, Cuba and Medellin, Colombia. Two parallel group of samples were formed, one for each station. They were submitted to accelerated outdoor test by intermittent spraying of a salt solution (SCAB test) according to ISO 11474:98, receiving also the influence of the open atmosphere. The acceleration of corrosion rate of the three metals caused by the presence of chloride ions in both stations was determined. As expected, steel shows the higher corrosion rate and acceleration by chlorides, particularly at Cuban corrosion station. A remarkable difference in the acceleration rate of chloride ions for mild steel and copper between Cuban and Colombian acceleration rate of chloride ions of steel and copper. Steel corrosion products were analysed by Mossbauer Spectroscopy. Water absorption was also studied. The presence of magnetite, goethite and other Iron compounds was determined.


INTRODUCTION
Chloride ions are one of the most important natural contaminants having influence on corrosion of metals.However, its role may change depending on climate.A very humid climate can cause a fast dissolution of chloride ions and diminish its role in the acceleration of corrosion rate.A different acceleration mechanism has been reported for chloride ions between Eastern and Western Caribbean^^.
In a single climate the influence of rain could change the acceleration of chloride ions on metal corrosion.Differences between rainy and dry seasons could be important in determining the acceleration cause by chloride ions.The electrochemical mechanism of corrosion by chloride ions do not changes, but the time and the concentration during which chloride ions acts on the metallic surface may appreciable change depending on climatic conditions and the nature of corrosion products.
Atmospheric corrosion occurs in an alternating growing and decreasing layer of electrolyte, which is usually very thin.In the case of existing corrosion products at metal-atmosphere interface, the sorption properties of these products and salts (usually hygroscopic) deposited or formed by corrosion determine the possibilities of presence of superficial wetness^ \ The amount of water on the corroding surface is of great importance for the corrosion rate.
Results from a very limited number of runoff studies are available in the literature^ .The understanding of air pollution effects on metallic corrosion has advanced considerably over the past several decades, there have been very few attempts to investigate the relation between precipitation features and metal damage.Besides the chemical composition of rainwater, the physical characteristics of precipitation (rainfall rate and amount, frequency and length of dry and wet events) also play a role in the corrosion process.It is reported^^ that the time evolution of chloride ion concentration in the steel runoff is approximately twice that presented in copper runoffs.The high values of steel over those from rainwater, suggest that chloride participate in the metal attack without forming corrosion products.
To study the rol of rainfall and other parameters in the corrosion acceleration of chloride ions is the main objective of the present paper.A cooperation between Corrosion Group from Antioquia University, Medellin, Colombia, and Corrosion Department, National Centre for Scientific Research, Havana, Cuba has made possible to obtain and process data at the same time in test stations of both countries having climatic differences.Samples were exposed facing south at 45-to the horizontal.Samples were submitted to a salt spray twice a week according to ISO 11474 .The average quantity of 3 % NaCl solution added to each sample was measured.It was about 2.37 ml by sample, equivalent to a chloride deposition rate of about 850 mg/m^d.Weight loss of steel, copper and aluminium was evaluated for different exposure times (See Table II and chapter II).Samples size was 100 mm x 150 mm X 1-2 mm.Chloride and SO2 deposition rate were evaluated monthly.The determination was carried out by the wet candle method in the case of chloride ions and by alkaline sulfhation plates in the case of SO2.Time of wetness was calculated using temperature-humidity data obtained from meteorological stations.Time of wetness at different intervals of temperature and relative humidity was calculated.Rain data was also obtained from meteorological stations.Adsoption isotherms were determined by the liquid-volumetric method^ .Rust samples were activated at 300 and 423 K and 0.13 Pa during 15 h.Water was supplied in quantities of 0.174 mmol every time, waiting for the necessary time up to the establishment of the equilibrium.Experimental isotherms were fitted to the Dubinin model.Based on this model maximum adsorption, microporous volume and characteristics energies were calculated.

Samples of AISI
Mossbauer spectra at room temperature of rust samples were obtained in order to determine phase composition of corrosion products.

Corrosion of metals at CNIC's and Antioquia stations
The influence of chloride ions at Antioquia and CNIC corrosion station was studied by submitting samples of steel, copper and aluminium to salt spray during exposure to the open atmosphere.
From tables II and III it can be observed that up to 6 months of exposure, the acceleration caused by chloride ions added in equal quantity is notably higher at CNIC's station for steel and copper.In the case of steel, it increases with time.Aluminium behaves in a different way, although there is only data corresponding to 6 months, because there is more acceleration at Antioquia than at CNIC.It could be due, perhaps, to a higher cleansing effect of rain in Cuba, causing a more effective elimination of chloride from the aluminium surface.As a passive metal aluminium is much more sensible to surface contamination than steel and copper.As can be seen from table IV, the influence of rain quantity is significant in changes of corrosion rate of steel and copper; however, the influence is different.In the case of steel an increase in rain quantity causes a lower corrosion {b < 0) while for copper rain increases corrosion rate (fc> > 0).This different behavior could be due to the solubility of corrosion products, because iron chlorides are much more soluble than copper chlorides.Although the above regression equations are very significant, they only explain less than 50 % of the cases.It means that there are other factors having influence on metals corrosion.It got to be taken into account that for small times of exposure there is difficult to obtain good correlation coefficients^ .
The term KCI-KJ^^^Q has been considered as the acceleration rate, because it is the difference between the corrosion of samples exposed in the natural atmosphere under the influence of a salt spray and those without this influence.

Behaviour at longer times
Data corresponding to 1, 3 and 6 months of exposure of steel at CNIC's station were processed and the following regression equations were obtained: Kci = -614.8+ 18.8 K,"ci R^ = 74.0% (1) This regression equation shows that there is a good agreement between steel corrosion rate submitted to a saline spray and without it exposed on the same atmospheric conditions.A similar behaviour is determined for aluminium at 6 and 12 months of exposure and a different one for copper at 3 and 6 months of exposure.Aluminium: Kc, =-1.16 +14.6 K""ci R^ = 80.5 % Copper Kci= 11.98 +0.55 K""ci R2 = 4.8 %

(2) (3)
There is almost no relationship between copper weight loss in addition of chloride ions and without it.An analysis of the data set shows that weight loss of samples without addition of chlorides appreciable changes, while in the presence of these anions weight loss is almost independent of changes on climatic season.

Influence of time of wetness
Previous results have shown that time of wetness in Cuban conditions could be divided into three main categories^ % time at 80-100 % relative humidity and temperature in the range 5-25 °C, time at 80-100 % relative humidity and temperature in the range 25-35 °C and time of rain.In this case the days of rain has been considered as time of rain.
As can be seen on table V, a different behaviour on the influence of the categories of time of wetness can be observed for steel and copper.In the case of steel, a higher influence of time of wetness at temperature in the range 5-25 °C (coefficient h) respecting to 125,35 Is obtained.The role of rain changes in presence of the salt spray.Weight loss increases with days of rain in presence of high salinity.It is not opposed to the above results obtained for one month period because in this case it is studied the influence of rain at longer and not constant time of exposure.
In the case of copper, the parameter with higher influence on corrosion is time of rain.A negative sign affecting coefficients b and c is not logic, because corrosion should always occur at relative humidity 80-100 % and temperature intervals 5-25 °C and 25-35 °C.This result could be due to the influence of other variables not taken into account in the regression equations or to the existence of a given correlation between variables considered as independent.

Corrosion products phase characterization
Mossbauer and IR Spectra shows the presence of Magnetite and Goethite as main phases in corrosion products of steel submitted to salt spray in natural conditions after 6 months of exposure.A paramagnetic phase, still unidentified is also reported according to the results of Mossbauer Spectra (See Table VI) Water adsorption of corrosion products.The Dubinin Theoryf'^^^^^^i^l (Vol umetric Microporous Fitting) was used to study activated carbon, but at The adsorption isotherms obtained show that activated rusts have a heterogeneous porous structure due to the presence of two homogeneous zones.In this case adsorption is described by the Dubinin-Radushkievich equation.Samples without chloride addition show higher maximum adsorption and porous volume than those with addition of chlorides (see Table VII).However, the characteristic energy of the samples with chloride addition is higher.It means that water is stronger linked due to the hydration heat of the chlorides.
Maximum adsorption in the sample without chloride addition is higher at 300 °C activation temperature, showing that desorption is easier at that temperature.Relatively lower values of the characteristic energies show that water adsorption has a physical link, but it should be noted a marked difference between the energies of the second zone, higher in more than 70 % in the sample with chloride addition, so this second zone has porous radii significantly smaller than sample without chlorides, showing that metal corrosion in Kci = Weight loss (g/m^) for samples with application of salt spray ' ^noCl = Weight loss (g/m^) for samples without application or salt spray R (mm) = Rain (mm) a = Intercept of the regression equation b = Hope of the regression equation r = Correlation coefficient r^ = Percent of variables explanation

Table I .
Main characteristics of pollution of corrosion stations (chloride and sulphur compounds deposition rate).

Table II .
Weight loss (g/m^) for steel, copper and aluminium submitted to salt spray and exposed at Antioquia and CNIC stations Tabla II.Pérdidas de peso (g/m^j del acero, cobre y aluminio sometidos a niebla salina y expuestos en las estaciones de Antioquia y el CNIC

Table ill .
Average relation Weight Loss Antioquia/WeightLoss CNIC for steel, copper and aluminium at different exposure times

Table IV .
Regression equations obtained between steel and copper weight loss and rain quantity (mm).

Table VII .
Activation temperotureTa (°K), maximum adsorption-Am (mmol/g), microporous volume -Vp (cm^/g) and characteristic energy -Ec (Id/mol) of water adsorption in rust formed in the atmosphere after six month of exposure