CS 2811 No.

of Pages 14, Model 5+

ARTICLE IN PRESS
6 August 2007 Disk Used

Corrosion Science xxx (2007) xxx–xxx

www.elsevier.com/locate/corsci

2 Time of wetness in tropical climate:

3 Considerations on the estimation of TOW according

F
4 to ISO 9223 standard

OO
5 F. Corvo a, T. Pérez b,*, Y. Martin c, J. Reyes b, L.R. Dzib b,
6 J. González-Sánchez b, A. Castañeda c
a
7 Instituto de Ciencia y Tecnologı́a de Materiales (IMRE), Universidad de la Habana, Vedado, Plaza, Ciudad de la Habana, Cuba
b
8

PR
Centro de Investigación en Corrosión (CICORR), Universidad Autónoma de Campeche, Ave. Agustin Melgar s/n, Colonia Buenavista, San Francisco de
9 Campeche, Cam, C.P. 24030, Mexico
c
10 Centro Nacional de Investigaciones Cientı́ficas (CNIC), Apartado 6412, Ciudad de la Habana, Cuba
11 Received 13 February 2007; accepted 21 June 2007
12 ED
13 Abstract

14 This work was focused on the discrepancies between the TOW determined according ISO 9223 standard and data obtained for indoor
15 and outdoor conditions in humid tropical climate. It was observed that for temperatures higher than 25 °C the determination of TOW
CT

16 changed. Some proposals are made based on the results obtained in the present research conducted in Cuba and Mexico in order to
17 improve the estimation of TOW. Such proposals take into account the linear relationship between time of exposure and TOW, the effect
18 of rain, the role of air contaminants, temperature and the differences between outdoor and indoor exposure conditions.
19 Ó 2007 Published by Elsevier Ltd.
E

20 Keywords: C. Atmospheric corrosion; C. Rust; A. Steel

21
RR

22 Q3 1. Introduction Results of outdoor and indoor corrosion rate and corro- 33

sion aggressivity in tropical corrosion test stations of Cuba 34
23 Atmospheric corrosion is the most extended type of cor- and Mexico are reported. The results mainly concern to 35
CO

24 rosion in the World. During many years, several papers natural atmospheric corrosion tests obtained in the western 36
25 have been published in this subject; however, most of the side of the Isle of Cuba and in the Campeche State located 37
26 research has been made in non-tropical countries and at the Yucatán Peninsula in Mexico. The two regions are 38
27 under outdoor conditions. The tropical climate is typical located in the tropical climate and receive the influences 39
28 of equatorial and tropical regions and is characterized by of the waters of the Atlantic Ocean and the Mexican Gulf. 40
UN

29 permanently high temperatures and relative humidity with Data processed in this paper correspond to atmospheric 41
30 considerable precipitation, at least during part of the year. corrosion tests carried out during a long period of time, 42
31 A high corrosion rate of metals is usually reported for this about the last 20 years in Cuba and the last 10 years at 43
32 climate. Campeche up to the present. 44

2. Climate of Cuba and the Yucatán Peninsula 45

*
Corresponding author. Tel.: +52 9818119800/62800; fax: +52
Q1 9818119800/62899. The humid tropical climate of Cuba and the Yucatán 46
E-mail addresses: corvo@imre.oc.uh.cu (F. Corvo), tezperez@ua- Peninsula (México) is characterized by an average air tem- 47
Q4 cam.mx (T. Pérez), abel.castaneda@cnic.edu.cu (A. Castañeda). perature always higher than 15 °C, frequently high relative 48

0010-938X/$ - see front matter Ó 2007 Published by Elsevier Ltd.

doi:10.1016/j.corsci.2007.06.012

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

2 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

49 humidity, a summer or wet season (May–October) with fre- to use the concept of time of wetness (TOW). It is a concept 104
50 quent and heavy precipitations and a winter of dry season commonly used in atmospheric corrosion of metallic mate- 105
51 (November–April) with lower precipitations. rials and refers to the time when the metal is sufficiently wet 106
52 In the case of these two regions there is a natural source for corrosion reaction to occur, that is, when an electrolyte 107
53 of airborne salinity: the waters of the Atlantic Ocean, the is present on the metallic surface. Under the particular 108
54 Gulf of Mexico and the Caribbean Sea. Airborne salinity characteristics of atmospheric corrosion there are time 109
55 plays an important role in determining corrosion aggressiv- periods where corrosion could not occur due to the absence 110
56 ity in Cuba [1–4] and in the Yucatán Peninsula [2,5,6]. of an electrolyte in the metallic surface. The lowest outdoor 111
57 Other anthropogenic contaminants can be present also in TOW values are observed in the desert regions, as also in 112
58 this region, particularly sulfur compounds coming from the Antarctic and Arctic regions. Atmospheric corrosion 113

F
59 the oil production and manufacture industries and trans- rates of metals at these climatic conditions are also very 114
60 portation. Depending on the contaminants sources existing low and in the case of cold regions, the increase of temper- 115

OO
61 in different regions, the kind of contaminants May change, ature leads to the increase of TOW and corrosion rate [11]. 116
62 but in general, the main types of contaminants are Chlo- In principle, TOW is a parameter that depends upon both 117
63 rides and Sulfur compounds. the climatic conditions and in the characteristics of the 118
64 Air temperature in the Yucatán Peninsula reaches higher metallic surface. 119
65 values than in Cuba, the top average air temperature in The definition of TOW presented on ISO standard 9223 120
Cuba is 26 °C in the eastern shoreline [7]. In the Yucatán

PR
66 is the following: ‘‘The period during which a metallic sur- 121
67 Peninsula part of the territory can be classified as Tropical face is covered by adsorptive and/or liquid films of electro- 122
68 very warm with an average temperature over 26 °C [8]. lyte that are capable of causing atmospheric corrosion”. In 123
69 Absolute maximum temperature in Cuba is of 38.6 °C, addition, the new document ISO WD/9223 [11] defines: 124
70 meanwhile a temperature over 40 °C is frequently reported The wetting of surfaces is caused by many factors, for 125
71 in sites of the Yucatán Peninsula, particularly in Campeche. example, dew, rainfall, melting snow and a high humidity 126
ED
72 About 1/4 of the Cuban territory is composed by moun- level. The length of time when the relative humidity is 127
73 tains and the rest is mainly terrain flatness. The Yucatán greater than 80% at a temperature greater than 0 °C is used 128
74 Peninsula has no mountains. The lower temperatures and to estimate the calculated time of wetness (s) of corroding 129
75 higher precipitations rates in Cuba are localized in the surfaces. Information on calculated time of wetness is help- 130
76 mountain regions. Orography is a transformation factor ful for informative atmosphere corrosivity estimation. In 131
CT

77 of local wind regime. this standard, TOW is ‘‘estimated” based on the character- 132
78 Average relative humidity in the west side of Cuba is fre- istics of the temperature–humidity complex, independently 133
79 quently higher than in Campeche, the average is about 80% of the pollutant level and the nature of the metal or alloy. It 134
80 for Cuba and about 76% for Campeche. Daily maximum is not the actual time of wetness, but it is an estimation 135
E

81 relative humidity values are over 90% for the west part of based only on climatic factors (temperature and relative 136
82 Cuba and over 88% for Campeche. In spite of this, the clas- humidity) and independent of the nature and characteris- 137
RR

83 sification of TOW according to ISO-9223 [9] is the same for tics of the metallic surface. Some results obtained in the 138
84 the two territories: s4, corresponding to ‘‘Outdoor atmo- tropical regions of the Gulf of Mexico suggest that an 139
85 spheres at all climates, excepting dry and cold climates”. upper limit in temperature for the definition of TOW could 140
86 The influence of air temperature may change depending be established [3,4,12–14]. TOW as defined in ISO 9223 141
87 on the conditions. The effect of increasing temperature does not cover all the aspects of climate. According to 142
CO

88 appears to be an increase in corrosion rate under condi- the ISO definition, in the case of tropical climate where 143
89 tions of permanent surface wetting, such as those obtained air temperature never reaches 0 °C, TOW is estimated as 144
90 during precipitation. Under conditions of varying surface the time when relative humidity is over 80%. The diminu- 145
91 wetting, however, the corrosion rate increases with temper- tion of the electrolyte layer at relative humidity over 80% 146
92 ature up to a certain maximum value, and thereafter when temperature is over 25 °C has been reported for San- 147
UN

93 decreases [10]. ISO 9223 standard established that for pur- tiago de las Vegas rural station in Cuba [12] using TOW 148
94 poses of corrosivity determinations the use of temperature– sensors based in a circuit print covered with a gold layer. 149
95 humidity complex data is the recommended methodology, By other way, metal temperature and TOW were registered 150
96 because corrosion rate depends on this complex and not in a rural station at 30 km of the seashore located in Mer- 151
97 only on temperature or humidity independently. Taking ida, Yucatan Peninsula, using copper/gold sensors [14] 152
98 into account this report, a decrease in corrosion rate should according to ASTM Practice G84–89 (1999). It was 153
99 be expected in tropical climate for an increase in average reported the influence of the nature of the metal and their 154
100 air temperature. corrosion products, the orientation of the sample (skyward 155
or groundward), the direction of the winds and rain precip- 156
101 3. Time of wetness (TOW) and ISO 9223 definition itation on the measured TOW values. None of this factor is 157
taken into consideration by ISO 9223. A quite different 158
102 Taking into account the electrochemical nature of the temperature was determined for the metal surface respect- 159
103 atmospheric corrosion process it is absolutely necessary ing air temperature. It is perhaps an explanation to the very 160

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 3

161 possible increase in evaporation of the surface electrolyte products accelerate the corrosion process. Water adsorp- 218
162 when air temperature is over 25 °C. tion isotherms were determined to corrosion products 219
163 Water adsorption on silver surfaces exposed into a ven- formed in Cuban natural atmospheres [21]. Sorption prop- 220
164 tilated shed in an urban-rural site of Cuba was studied [13] erties of corrosion products (taking into account their salt 221
165 using quartz resonators covered with a silver layer. It was content-usually hygroscopic) determine the possibilities of 222
166 determined that in these indoor conditions water adsorp- surface adsorption and the possibility of development of 223
167 tion significantly diminishes when air temperature increases corrosion process. 224
168 over 25 °C at relative humidity ranges of 80–90% and 90– For a single type of climate it could be more practical 225
169 100%. All these results confirm the idea that an upper limit and easy the use of time instead of TOW, because it is 226
170 of temperature should be established for the estimation of expected the same category of TOW according to ISO 227

F
171 time of wetness. 9223. In other words, it makes no sense to work using 228
172 The presence of water does not only create conditions TOW-ISO in tropical climate, because in general it is in 229

OO
173 for the existence of an electrolyte, but it acts as a solvent the same category at all places, so its influence is the same. 230
174 for the dissolution of contaminants [10]. Oxygen plays an Using data for different climates or wide regions a signifi- 231
175 important role as oxidant element in the atmospheric cor- cant change in TOW-ISO could take place, including 232
176 rosion process. The thickness of the water layer determines changes in wet/dry cycles. In these conditions it should 233
177 the oxygen diffusion toward the metallic surface and also be recommendable to use TOW according to ISO 234

PR
178 the diffusion of the reaction products to the outside inter- definition. 235
179 face limited by the atmosphere. Another aspect of ISO def-
180 inition is that ‘‘a metallic surface is covered by adsorptive
181 and/or liquid films of electrolyte”. According to new 4. Time and TOW-ISO 236
182 results, the presence of adsorptive or liquid films of electro-
183 lyte perhaps could be not in the entire metallic surface, but 4.1. Outdoor TOW-ISO 237
ED
184 in places where there is formed a central anodic drop due to
185 the existence of hygroscopic particles or substances sur- Corrosion rate is a function of time of wetness, consid- 238
186 rounded by microdrops where the cathodic process takes ered as the time during which corrosion occurs, but in gen- 239
187 place. This phenomenon is particularly possible in indoor eral it should not be a linear function because corrosion 240
188 conditions [15–18]. rate changes with time. There are different factors influenc- 241
CT

189 Using ISOCORRAG, MICAT and Russian data, Tid- ing, for example, the protective properties of the corrosion 242
190 blad et al. [19] showed that the inclusion of temperature products, the increase or decrease of the acceleration 243
191 among the environmental parameters improves consider- caused by contaminants, increase or decrease of the thick- 244
192 ably the usefulness of the dose–response functions and ness and conductivity of the electrolyte layer, i.e. 245
E

193 should be adapted in the revision of ISO 9223 standard. If the definition of TOW established by ISO is used 246
194 It is reported an increase in corrosion rate with average (TOW-ISO), a linear relationship between time and TOW 247
RR

195 air temperature in the range of 15 to 30 °C. is obtained, in spite of the different possible changes in cor- 248
196 In the eighties of the last century the adsorption of water rosion rate caused by changes in the nature of TOW. It has 249
197 layers on surfaces was studied [20]. It was demonstrated to be remarked that it is not the same effect on corrosion 250
198 that the change in thickness of these layers depends on rate caused by a heavy rain than dew, fog or water adsorp- 251
199 the physico-chemical properties of water in these thin water tion, so for the same interval of relative humidity (80– 252
CO

200 layers. It is reported that on iron surfaces, the number of 100%), notable changes in nature of TOW-ISO and conse- 253
201 adsorbed water layers is about 15 at RH 55% and 90 at quently in corrosion rate could take place. 254
202 100%. Similar values are obtained for copper and zinc; If a linear regression is fitted between time (sum of time 255
203 however significant differences are reported for Platinum, in months) and TOW-ISO (sum of TOW corresponding to 256
204 gold, aluminum and silver. These monolayers have been every month up to 1 year), a perfect linear relationship is 257
UN

205 calculated, in general, only in presence of water (without obtained (see Tables 1–5). The results show that a perfect 258
206 oxygen) where the corrosion process is very slow and, con- linear relation exists between time and TOW-ISO. It means 259
207 sequently, in conditions far from the reality. that both variables are equivalent, so for purposes of long- 260
208 Almost all tests carried out to study the starting process
209 of atmospheric corrosion have been performed in a surface Table 1
210 without corrosion products; however, in real conditions, Simple regression of TOW-ISO data and time for Campeche PCGM
211 the metal is covered with corrosion products after a given coastal test station in a period of 2 years
212 time and these products begin to play its role as retardants Year a b r r2 P
213 of the corrosion process in almost all cases. Corrosion 2004 230.42 234.39 0.979 95.90 <0.0000
214 products act as a barrier for oxygen and contaminants dif- 2005 300.15 258.77 0.974 94.88 <0.0000
215 fusion, the free area for the occurrence of the corrosion is Equation t = a + b TOW-ISO, r = correlation coefficient, r2 = percentage
216 lower; however, the formation of the surface electrolyte is of variation explained by the independent variable, P = statistical prob-
217 enhanced. Only in highly polluted areas the corrosion ability Campeche PCGM.

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

4 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

Table 2 term prognosis, work carried out time give the same result 261
Simple regression of TOW-ISO data and time for Viriato coastal test than using TOW-ISO. It does not mean that the actual 262
station in a period of 10 years
time of wetness should be the same as estimated according 263
Year a b r r2 P to ISO, because as it has been pointed out above, the ISO 264
1987 487.20 406.48 0.987 97.49 <0.0000 definition does not take into account the nature of the dif- 265
1988 14.83 423.15 0.997 99.38 <0.0000 ferent components of TOW regarding climate (rain, dew, 266
1989 160.20 365.59 0.998 99.58 <0.0000
1990 89.30 299.48 0.996 99.11 <0.0000
fog, water adsorption) and the nature of the metal and 267
1991 167.38 395.20 0.998 99.52 <0.0000 the corrosion products. 268
1992 237.26 375.55 0.996 99.16 <0.0000
1993 91.18 348.37 0.996 99.29 <0.0000
4.2. TOW-ISO under heat trap conditions 269

F
1994 70.60 380.73 0.999 99.71 <0.0000
1995 428.05 413.16 0.994 98.73 <0.0000
1996 83.97 252.75 0.997 99.489 <0.0000 A simple regression was made between time and TOW- 270

OO
1997 101.30 390.60 0.999 99.73 <0.0000 ISO for the different test stations under heat trap condi- 271
Equation t = a + b TOW-ISO, r = correlation coefficient, r2 = percentage tions for a 2 year exposure time. The results were the fol- 272
of variation explained by the independent variable, P = statistical lowing: Heat trap conditions [22–24] refer to the exposure 273
probability.
of samples inside metallic boxes. In those conditions of 274
partially closed metallic boxes, sun irradiation causes a sig- 275

PR
Table 3 nificant increase in temperature and the maximum values 276
Simple regression of TOW-ISO data and time for Holguin coastal station of air temperature are obtained under these conditions. 277
(located in the eastern side of Cuba) in a period of 3 years Some electro-electronic items are used under these 278
Year a b r r2 P conditions. 279
1987 126.42 469.84 0.999 99.79 <0.0000 The results of the regression are the following: 280
1988 27.13 435.98 0.999 99.88 <0.0000
ED
1989 54.44 403.36 0.999 99.96 <0.0000 Quivican rural test station, Cuba: 281
t = 0.005 + 0.0026 TOW-ISO 282
r = 0.999 r2 = 99.94 P < 0.0000 283
284
Table 4
Simple regression of TOW-ISO data and time for Cojimar coastal station
Campeche PCGM coastal test station, Mexico: 285
CT

in a period of 10 years t = 0.177 + 0.0032 OW-ISO 286

Year a b r r 2
P
r = 0.999 r2 = 99.79 P < 0.0000 287
288
CNIC urban rural test station, Cuba: 289
1991 127.68 389.13 0.998 99.52 <0.0000
1992 172.08 367.73 0.996 99.14 <0.0000
t = 0.0011 + 0.003 TOW-ISO 290
r = 0.999 r2 = 99.84 P < 0.0000 291
292
E

1993 54.0 342.23 0.998 99.53 <0.0000

1994 51.32 381.34 0.999 99.78 <0.0000 Cojimar coastal test station, Cuba: 293
1995 289.58 389.58 0.993 98.60 <0.0000 t = 0.097 + 0.0035 TOW-ISO 294
296
RR

1996 164.67 376.75 0.997 99.41 <0.0000 r = 0.999 r2 = 99.9383 P = 0.0000 295
297
1997 51.44 369.95 0.999 99.83 <0.0000
1998 75.36 344.79 0.997 99.40 <0.0000
The results show that a perfect relation exists between 298
1999 286.80 352.04 0.989 97.89 <0.0000 time and TOW-ISO, so both variables are equivalent and 299
2000 159.21 371.14 0.995 99.02 <0.0000 can be used for purposes of long-term prognosis. 300
Equation t = a + b TOW-ISO, r = correlation coefficient, r2 = percentage There are some small differences between the metallic 301
CO

of variation explained by the independent variable, P = statistical boxes used in Cuba and Campeche. On Figs. 1 and 2 it 302
probability. can be observed that Cuban metallic boxes are relatively 303
more closed than the one used at Campeche PCGM corro- 304
sion station. 305
Table 5
UN

Simple regression of TOW-ISO data and time for Quivican rural station in
a period of 8 years 4.3. TOW-ISO under heat trap conditions at long-term 306
Year a b r r 2
P exposure 307

1990 89.37 405.78 0.997 99.34 <0.0000

1991 121.50 421.15 0.997 99.34 <0.0000 It is very well known the equation: 308
1992 66.07 440.34 0.998 99.59 <0.0000
K ¼ asb 310
1993 33.53 443.97 0.999 99.85 <0.0000
1994 75.86 448.22 0.999 99.73 <0.0000
where a is a constant, b is the coefficient that indicates the 311
1995 105.04 458.38 0.999 99.72 <0.0000
1996 62.50 444.04 0.999 99.53 <0.0000 protective properties of the corrosion products layer and s 312
1997 71.86 457.37 0.999 99.78 <0.0000 is the actual time of wetness (effective time during which 313
2
Equation t = a + b TOW-ISO, r = correlation coefficient, r = percentage corrosion process occurs). 314
of variation explained by the independent variable, P = statistical It is very wide used in studying the atmospheric corro- 315
probability. sion process and its development with time. As smaller will 316

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 5

Table 6
Statistical fitness of data of corrosion of copper, steel, nickel and tin inside
a metallic box (heat trap conditions) respecting time and TOW according
to ISO definition
Station Metal K = asb K = atb
a b r2 a b r2
Cojimar Copper 0.0009 0.847 48.93 0.138 0.805 49.76
(coastal) Steel 0.0038 1.185 96.27 4.284 1.455 97.11
Tin 0.0014 0.817 88.09 0.177 0.795 98.41
Nickel n.w.l. n.w.l. n.w.l. n.w.l. n.w.l. n.w.l.

F
CNIC Copper 1.1270 1.239 90.85 0.205 1.179 91.76
(urban) Steel 1.7930 1.596 94.39 2.823 1.532 95.79
Tin 0.0044 0.766 85.15 0.467 0.716 82.98

OO
Nickel 9.7180 1.216 97.54 0.015 1.163 95.98
Quivicán Copper 0.0017 0.986 98.31 0.453 0.976 97.81
Fig. 1. Metallic box exposed at Cojimar coastal station in the north shore (rural) Steel 0.0206 1.087 97.83 9.412 1.083 98.68
of the City of Havana. Similar metallic boxes are exposed at CNIC and Tin 0.0135 0.705 70.76 0.728 0.697 70.30
Quivican Cuban stations. Samples and contaminant detector are placed Nickel 7.2561 0.953 94.82 0.158 0.945 94.67
inside the metallic boxes.

PR
Campeche Copper 0.00003 1.528 93.68 0.158 1.628 90.68
PCGM Steel 0.0080 1.137 93.33 4.141 1.209 90.84
(coastal) Tin 0.0022 0.914 88.23 0.354 0.975 85.62
Nickel 0.0000002 1.851 88.25 0.006 1.974 85.55
ED n.w.l. = no weight loss detected.

corrosion to occur due to the electrochemical nature of 331

the atmospheric corrosion process; however, in indoor con- 332
ditions there are no precipitations and the presence of sur- 333
face water depends mainly on water content in the air and 334
CT

changes in temperature on the surface, as well as the pres- 335

ence of hygroscopic substances on the metallic surface. 336
Indoor temperature and relative humidity appreciably 337
depends on the ventilation level, on the use or not of air 338
E

conditioner or heating systems and of the thermal isola- 339

tion. Leygraf reports [25] that indoor relative humidity 340
RR

ranges from 15% to 85% with an average value of 50%. 341

Fig. 2. Metallic boxes exposed at Campeche PCGM coastal station at In Cuba [26] average values significantly different to 50% 342
300 m of the shoreline in the City of Campeche (front view). Samples and
contaminant detectors are placed inside the metallic boxes. As it can be
have been found, because inside a closed storehouse (no 343
observed, this metallic box has a wider window than those used in Cuban windows) during 1 year an average value of 89.5% is 344
stations. reported. In a storehouse having windows the average 345
CO

value is of 79.5%. In a storehouse having a dehumidifier 346

system the annual average relative humidity diminishes to 347
317 be coefficient b, it is supposed that the protective properties
74.5%, still significantly high respecting Leygraf report. 348
318 of the corrosion products layer are higher.
In two storehouses without climatic control it is reached 349
319 On Table 6 it is shown a comparison between data fit-
a 100% of relative humidity, while in the storehouse sub- 350
ness using TOW calculated according to ISO 9223 and
UN

320
mitted to dehumidification the relative humidity reaches 351
321 time. It can be observed that there are not significant differ-
85%. Changes in indoor relative humidity are lower than 352
322 ences in using TOW or time because r2 is almost the same
those occurring outdoors, without the strong change 353
323 in both cases. It could be explained supposing a direct rela-
caused by the difference between night and day. In this 354
324 tionship between time and TOW, that is, TOW increases
way, the most probable way the metallic surface can be 355
325 with time. If, in general, there should be differences in
humid is through water adsorption, although it can be 356
326 TOW for the different climatic seasons, these are not signif-
observed that a significant relative humidity is reported 357
327 icant for the total time.
indoors respecting Europe, so a higher corrosion should 358
be determined. The higher temperature of the Cuban cli- 359
328 5. Indoor humidity mate causes also the existence of higher water content 360
and increases the possibilities of water condensation. 361
329 Under indoor conditions, in the same way than out- Under indoor conditions, wind rate is lower than out- 362
330 doors, it is necessary the presence of surface humidity for doors [27]. In this way, outdoor conditions promotes a fast 363

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

6 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

364 surface drying due to the influence of air and sun radiation, increase in the oxygen diffusion current at the beginning, 421
365 and also a more frequent surface wetness, thicker water passivity of the surface in a second period and the dimin- 422
366 layers due to rainfall, factors almost nonexistent in indoor ishing in the oxygen reduction at the end of the drying 423
367 conditions. Another factor nonexistent in indoor condi- process. 424
368 tions is the washing effect on the metallic surface by Recent reports about the microdroplets formation in the 425
369 precipitations. starting periods of atmospheric corrosion [15–18] show 426
370 It is important to remark that indoors, TOW-ISO is that the idea of a thin uniform water layers is not com- 427
371 higher in storehouses without ventilation than outdoors pletely in accordance with the reality. It has been observed 428
372 in tropical humid climates. ISO 9223 classifies TOW for that when a water drop is on the metallic surface, formed in 429
373 no ventilated storehouses as s5 (more than 5500 h/y), the place where a salt deposit existed before, microdroplets 430

F
374 whereas for outdoor conditions it is s4 (2500–5500 h/y). are formed around this central drop. The cathodic process 431
375 Taking into account that indoors corrosion rate is signifi- takes place in these surrounding microdroplets, meanwhile 432

OO
376 cantly lower respecting outdoors, it can be perfectly under- the anodic process takes place in the central drop. This idea 433
377 stood that the development of corrosion under adsorbed is not consistent with the proposal of a uniform water layer 434
378 water is very low respecting outdoors. In these conditions, on the surface and it is very probable that this situation 435
379 the magnitude of corrosion will be determined by the pos- could be obtained under indoor conditions. It has been 436
380 sibilities of acceleration that could cause contaminants, determined that microdrops (about 1 lm diameter) clusters 437

PR
381 being the presence of water on the surface a requirement are formed around a central drop. An important influence 438
382 for the occurrence of corrosion, but not for determining of air relative humidity is reported on microdrops forma- 439
383 the rate of the corrosion reaction. tion. There is a critical value of relative humidity for the 440
384 When the wind velocity is lower, drying process could be formation of microdroplets. Under this value no microdro- 441
385 larger and could be a significant part of the dry/wet cycle. plets are formed. This value could be considered as the crit- 442
386 It also depends on the temperature existing on the metallic ical relative humidity. This situation is very similar to the 443
ED
387 surface. On ISO-9223 there are not considered changes in process of indoor atmospheric corrosion: presence of 444
388 the dry/wet cycles. humid air, deposition of hygroscopic contaminants in the 445
389 The deposition of contaminants in the surface usually surface, formation of microdrops. Water is necessary for 446
390 causes an acceleration of the corrosion process; however, corrosion reaction to occur, but the reaction rate depends 447
391 there should not be excluded the possibility that a given on the deposition rate and nature of contaminants. 448
CT

392 contaminant could diminish corrosion rate, as it could be ISO 11844-3 [28] establishes that the combination of dif- 449
393 the case of ammonia and its influence on steel (due to its ferent parameters is what determines the corrosivity of the 450
394 alkaline properties, it could induce the passivation of steel). atmosphere. Under indoor conditions, corrosion process 451
395 An important difference between outdoor and indoor depends on a more complex number of parameters than 452
E

396 conditions is that in the last one there is no washing of con- outdoors; however in the same way than outdoors two 453
397 taminants by precipitations. Contaminants deposition is types of parameters are considered: 454
RR

398 lower, but it increases with time. In these conditions, the

399 presence of adsorbed water is necessary for corrosion to  Air temperature and humidity. 455
400 occur, but the rate of the process depends on the accelera-  Air contaminants (gas and particulates). 456
401 tion cause by contaminants. The calculation of TOW-ISO 457
402 should not be as important as outdoors, because adsorbed The effect of these two groups of parameters is linked, 458
CO

403 water could be present all time. It has been reported a sig- because contaminants need a given level of humidity to 459
404 nificant decrease in adsorbed water as temperature act on the corrosion process. A combination of contami- 460
405 increases in Cuban climate [13]. nants could have different effects than the sum of the indi- 461
406 In indoor conditions, the role of airborne salinity signif- vidual effects. In ISO 11844-1 [31] it is explained that the 462
407 icantly diminishes, because the deposition rate is signifi- impact of temperature and relative humidity can not be 463
UN

408 cantly lower than outdoors [26]. expressed according to ISO-9223. It could be explained 464
409 Stratmann reports [28,29] polarization curves on metal- taking into account that the nature of TOW is very differ- 465
410 lic surfaces covered by electrolyte layers as thin as 2 lm ent to outdoors conditions. 466
411 without using Luggin capillary, but using a Kelvin probe Recent reports [32,33] on the use of atmospheric corro- 467
412 as reference (it is not necessary a contact to the surface). sion sensors based on changes in electrical resistance 468
413 In this way there is avoided the effect in the polarization showed that when there were no contaminants [31], in tests 469
414 curve caused by the presence of a capillary having aqueous of 100–110 h, corrosion rate was zero or insignificant. 470
415 solution inside. It was possible to study the drying process, These sensors can determine changes in metal thickness 471
416 observing that at the beginning oxygen reduction increases lower than one nanometer. However, in the presence of 472
417 due to a fast transport of this element through the thin elec- 0.08 ppm of SO2 or 200 mg/m2 of NaCl in the system, 473
418 trolyte layer. Corrosion rate was determined by the oxygen changes in thickness where always detected over 75% of rel- 474
419 consumption in a closed volume. It was shown that corro- ative humidity. Corrosion rate was determined at tempera- 475
420 sion rate of pure steel could be divided into three parts: the tures of 20, 30 and 40 °C and the Arrhenius equation was 476

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 7

477 used to calculate the activation energy of the reactions. evaporation of the electrolyte layer prevails (air tempera- 531
478 This method is very similar to the natural conditions. ture >25 °C) – in some cases this last variable should also 532
include rarely time of rain. This model was developed 533
479 5.1. TOW at different exposure conditions based on previous results indicating that when air temper- 534
ature is over 25 °C a significant diminution of the surface 535
480 TOW depends mainly on the meteorological parame- water layer takes place, possible due to a significant 536
481 ters, the nature of the metal, the properties of the metallic increase in evaporation of this water layer. 537
482 surface and of the corrosion products layer formed. The The suggestion of dividing the time of wetness into three 538
483 most important climatic factors on the corrosion process different contribution parts was made in order to get a more 539
484 are relative humidity, sunshine hours, temperature of the quantitative approach step to study the atmospheric corro- 540

F
485 air and the metal surface, wind velocity and duration and sion process. The influence of time and quantity of rain is 541
486 frequency of the rain, dew and fog. very important for characterizing differences between 542

OO
487 Dew or condensation of humidity is considered an indoor and outdoor corrosion. It was observed that in all 543
488 important cause of the corrosion of metals. Its formation cases the variable TOW 25–35 is affected by a negative sign, 544
489 depends on the relative humidity and on the changes of indicating a diminishing of corrosion rate when this vari- 545
490 temperature. Dew does not wash the metallic surface so ables increases, what is according to the fact that when tem- 546
491 the concentration of pollutants is relatively high and could perature is over 25 °C the electrolyte layer significantly 547

PR
492 be more aggressive than rain. Rain gives rise to the forma- diminishes instead of existing a high relative humidity. 548
493 tion of a thick layer of water and also adds corrosive agents The cleaning effect of rain is important in the corrosion 549
494 such as H+ and SO2 4 ; however it can wash away the con- process. In many cases this variable is significant, as well as 550
495 taminants as well. It will depend on the intensity and dura- the inclusion of time of rain as an independent variable in 551
496 tion of the rainfall. explaining the influence of different parameters in atmo- 552
497 The results of a wide evaluation carried out on different spheric corrosion rate of basic metals. 553
ED
498 exposure conditions in Cuba have already been published
499 [3,4]. The following Cuban corrosion test stations and con- 6. Air temperature, relative humidity, TOW-ISO, corrosion 554
500 ditions were considered: of steel and deposition of contaminants at different exposure 555
conditions 556
501 – Coastal test station under outdoor, sheltered and venti-
CT

502 lated shed conditions. A comparison of the behavior of air temperature for dif- 557
503 – Urban industrial test station under outdoor and shel- ferent exposure conditions in Cuba and Campeche using 558
504 tered conditions. data corresponding to 1 year is presented on Fig. 3. It is 559
505 – Rural test station under outdoor, sheltered, ventilated only an approximated behavior because there was not used 560
E

506 shed and closed space conditions. the same year for all exposure conditions. Average air tem- 561
507 perature is around 25 °C for all conditions, slightly higher 562
RR

508 As was expected, the higher corrosivity corresponds to in the storehouses in urban and rural sites, as also at 563
509 the coastal station. It was also noted that the higher the cor- Campeche outdoor coastal station. No data of average 564
510 rosivity of the atmosphere, the higher the difference between minimum and maximum temperatures were available for 565
511 the outdoor corrosion rate and the indoor corrosion rate. storehouses. It can be observed that higher temperatures 566
512 The atmospheric corrosion rate of metals depends are reported for conditions of heat trap and lowers for 567
CO

513 mainly on TOW and pollutants; however, if the differences storehouses. The wider interval of temperature is also 568
514 in the corrosion process between outdoor and indoor con- obtained at heat-trap conditions, probably due to the emis- 569
515 ditions are taken into account, the influence of direct pre- sion of heat of the painted galvanized steel used in the con- 570
516 cipitation such as rain is very important for outdoor and struction of the metallic boxes. The lower interval between 571
517 negligible for indoor conditions. The acceleration effect of maximum and minimum temperature is obtained in the 572
UN

518 pollutants could change depending on wetness conditions case of storehouse. It is due to the fact that these are wide 573
519 of the surface, so the influence of the rain time and quantity and close spaces with concrete walls where air temperature 574
520 should be very important in determining changes in corro- changes are relatively low. The lower temperature interval 575
521 sion rate. is reported for the storehouse with air conditioner. 576
522 A model was proposed which consider the influence of In outdoor, sheltered conditions and ventilated sheds 577
523 the relationship between rain quantity/time and the inter- there is reported a second wide interval of temperature. 578
524 action between pollutants at different TOW. TOW-ISO The lower and higher temperature is reported for the out- 579
525 was divided into two parts, when air temperature was lower door rural station, because it is relatively far from the sea- 580
526 than 25 °C and when air temperature was higher than shore and its influence is lower. 581
527 25 °C (up to 35 °C). In this way three different times of wet- The behavior of relative humidity is shown of Fig. 4. It 582
528 ness are considered: time due to rain (included indepen- is interesting to note that maximum relative humidity is 583
529 dently of TOW-ISO), time including rain, dew and fog obtained in almost all exposure conditions, excepting the 584
530 (TOW-ISO at air temperature <25 °C), and time when storehouse having air conditioner, rural ventilated shed 585

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

8 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

Outdoor and indoor air temperature

60

50
Temperature (oC) Air conditioned storehouse
40 Average T
Ave.min.T
30 Ave.max T
Min T
20 Max T

10

F
opened windows storehose
no windows storehouse
0

Urban

Urban
Rural

Rural

Rural

Rural
industrial

Coastal

Coastal

Coastal
Campeche

Campeche
coastal
Urban-

OO
Cuba

Cuba
Coastal

Coastal
Outdoor and sheltered Ventilated sheds Storehouses Heat trap conditions
Exposure conditions

Fig. 3. Changes in average, average maximum and minimum, maximum and minimum air temperature depending on exposure condition and type of

PR
atmosphere for test stations located in the western side of the Cuban Isle and Campeche PCGM coastal station.

Outdoor and indoor relative humidity

100

90
no windows storehouse
ED
80
Relative humidity (%)

70
opened windows Average RH
60 storehouse
Ave.min.RH
50 Ave.max.RH
Max RH
CT

40 air conditioned storehouse

Min RH
30

20

10

0
E

Coastall
Rural

Coastal

Campeche

Rural

coastal

Rural

Urban

Coastal

Rural

Urban

Campeche
industrial

Cuba
Urban-

Coastal

Cuba

Coastal
RR

Outdoor and sheltered Ventilated sheds Storehouses Heat trap conditions

Exposure conditions

Fig. 4. Changes in average, average maximum and minimum, maximum and minimum air relative humidity depending on exposure condition and type of
atmosphere for test stations located in the western side of the Cuban Isle and Campeche PCGM coastal station.
CO

586 and Campeche heat-trap conditions, the last two with a in general, relative humidity at Campeche is lower than in 602
587 very small difference. It means that regarding relative Cuba (and temperature higher). Both territories are classi- 603
588 humidity the possibilities of condensation of moisture are fied as humid tropical (s4), but Campeche presents lower 604
UN

589 significantly higher if the required conditions are available. values of relative humidity regarding Cuba. 605
590 Average relative humidity is around 80%, higher in the These two figures refer to the temperature-relative 606
591 storehouse having no windows and lower in the ventilated humidity complex, but there is an important difference in 607
592 shed located in the coastal site. It is very well known that in the environment between outdoor, sheltered, ventilated 608
593 a closed place relative humidity increases. No data of aver- shed and indoor corrosion. Precipitation is only possible 609
594 age maximum and minimum relative humidity for store- in outdoor conditions, although condensation of moisture 610
595 houses were available. The lower average in relative is possible in sheltered conditions and less probable in ven- 611
596 humidity between storehouses is reported for the store- tilated shed. In storehouses and heat trap conditions there 612
597 house having air conditioner as it should be expected. are no possibilities of precipitation and condensation of 613
598 The lower average RH is reported for Campeche PCGM moisture is more difficult. 614
599 (outdoor and indoor) coastal station. This station is A calculation of time of wetness according to ISO 9223, 615
600 located on a roof, where humidity should be lower regard- considered as the time when relative humidity is over 80% 616
601 ing a soil covered by grass as Cuban stations. By other way, and air temperature over 0 °C is presented on Fig. 5. 617

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 9

Outdoor and indoor time of wetness

8000
No direct water precipitations
7000

6000 Rain, dew, spray

Closed storehouse

5000
h/year

4000 t80-100

3000 Windows Storehouse

2000

1000

F
Air conditioned storehouse
0
Rural

Coastal

Campeche

Rural

coastal

Rural

Urban

Coastal

Rural

Urban

Coastal

Campeche
industrial
Urban-

OO
Cuba

Cuba
Coastal

Coastal
Outdoor and sheltered Ventilated sheds Storehouses Heat trap conditions

Exposure conditions

Fig. 5. Time of wetness at different exposure conditions and types of atmospheres calculated according to ISO 9223 for test stations of the western side of

PR
the Cuban Isle.

618 It can be noted that the lower TOW is obtained for the door conditions, the lower TOW-ISO value reported corre- 654
619 storehouse having air conditioner and the higher for the sponds to Campeche coastal station. 655
ED
620 storehouse located in the coastal site and the metallic box Average annual corrosion rate of steel reported for test 656
621 located in the rural site. According to these results, the stations depending on exposure conditions and type of 657
622 higher possibilities of corrosion should be in these last con- exposure (see Fig. 6) is significantly different to the behav- 658
623 ditions if we only consider the definition of TOW and the ior of time of wetness for the same conditions. It is per- 659
624 values obtained. However, it is very well known that the fectly explained based on the influence of contaminants. 660
CT

625 higher corrosion rate is usually obtained in outdoor or Chloride deposition rate causes a significant acceleration 661
626 sheltered conditions. This figure shows that the concept of corrosion rate, which is the reason why the maximum 662
627 of TOW established on ISO 9223 could not be used simply annual corrosion lost is reported for outdoor conditions 663
628 and many considerations have to be made. It should be in the coastal zone, following sheltered and ventilated shed 664
E

629 limited to outdoor conditions and some considerations in the same coastal zone. Urban industrial atmospheres 665
630 have to be made respecting the nature and characteristics (outdoor, sheltered and ventilated shed) are the second 666
RR

631 of different meteorological parameters such as rain, dew, one in weight lost and rural stations the less aggressive 667
632 fog, i.e. In other words, the concept of time of wetness atmosphere. Regarding indoor corrosion, the lower weight 668
633 (established in ISO 9223) shows limitations, because the lost is determined in the storehouses and heat trap condi- 669
634 nature and the changes occurred in the electrolyte formed tions increase respecting the last one. 670
635 on the metallic surface have clear differences between expo- A significant difference in steel corrosion exists between 671
CO

636 sure conditions. There is a significant difference between Campeche and Cuban coastal stations. It is mainly caused 672
637 outdoor time of wetness and sheltered or indoor TOW. by a significant difference in chloride deposition. 673
638 Actually, the general concept of TOW (time during As it can be observed, in the storehouse where there is 674
639 which corrosion reaction occurs) is, undoubtedly, correct determined the maximum time of wetness, it is reported 675
640 and generally applicable, taking into account the electro- the lower corrosion rate. It is explained based in the fact 676
UN

641 chemical nature of the atmospheric corrosion process, that in this case TOW consists mainly in air humidity with- 677
642 but the estimation made by ISO 9223 has to be limited to out the contribution of any precipitation or condensation 678
643 specific conditions, mainly outdoors and taking into con- (no rain, no dew, no significant condensation). It is a dem- 679
644 sideration the differences in the corrosion reaction as a onstration that the estimation of time of wetness according 680
645 function of the way the water is present on the surface, to ISO 9223 presents limitations. 681
646 because, for example, rain forms an electrolyte in the Changes in steel corrosion rate for different exposure 682
647 metallic surface different to dew, and the formation of conditions and types of atmosphere are better explained 683
648 the electrolyte during rain also causes a diminishing rate based on average deposition rate of Chlorides, as can be 684
649 of deposition of contaminants in the surface depending observed on Fig. 5. It can be note that chloride deposition 685
650 on the extent of rain. How can this complicated process is higher in Cuban coastal stations, as it should be 686
651 be represented? Obviously, not by the simple concept expected, and it is in agreement with a higher weight lost 687
652 established by ISO 9223. It should be also considered the for these stations. In the case of Campeche chloride depo- 688
653 nature of the metal and the corrosion products. Under out- sition is lower as also corrosion of steel. 689

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

10 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

F
OO
Fig. 6. Average corrosion rate (g/m2) depending on exposure condition and type of atmosphere for test stations located in the western part of the Cuban
Isle and Campeche PCGM coastal station.

PR
690 The lower values of chloride deposition are reported for coastal station is lower. Respecting chloride deposition, the 709
691 the storehouses, where coincidently, there are reported the determination at Campeche coastal station was carried out 710
692 lower weight losses. Under heat trap conditions there is a by the Wet Candle Method and at Cuba by a dry proce- 711
693 given increase in Chloride deposition that is why corrosion dure reported in [22]. Taking into account that a different 712
694 rate of steel is higher than in storehouses. method for determination of chloride deposition was used 713
ED
695 On Fig. 7 there is represented average chloride deposi- a correct comparison can not be made. 714
696 tion rate determined in Cuba and Campeche coastal station The average deposition of sulfur compounds found at 715
697 for different types of atmosphere and exposure conditions. the Cuban western side test station and Campeche PCGM 716
698 The results are very similar to those presented on Fig. 6, coastal station is presented on Fig. 8. As can be expected, 717
particularly respecting outdoor and ventilated shed condi- the higher sulfur compound deposition rate corresponded
CT

699 718
700 tions. Chloride deposition is determined under shelter, that to the urban industrial station in outdoor conditions. There 719
701 is why it is assumed the same value for outdoor and shel- is no report for ventilated shed. Part of the sulfur compound 720
702 tered conditions; however, it is very well known, the signif- deposition for coastal stations should not be attributed to 721
703 icant influence of precipitations upon outdoor corrosion gaseous SO2 or SO3, because it is well known that airborne 722
E

704 and its negligible effect under sheltered conditions. In store- salinity contains sulphate ions and the procedure used for 723
705 houses and heat trap conditions the influence of Chloride determining sulfur compounds deposition is sensible to sul- 724
RR

706 deposition significantly diminishes; however, under heat phate ions (see ISO 9225 [31]). A direct relationship between 725
707 trap conditions the higher corrosion of steel is reported average corrosion rate and sulfur compound deposition is 726
708 for the Cuban coastal test station. Corrosion at Campeche not observed as in the case of chloride deposition. It 727
CO

Chloride deposition rate at different exposure conditions

1000
Deposition rate (mg/m2d)

100
UN

Rural
Urban-industrial
10
Coastal Cuba
Coastal Campeche

0.1
Outdoor V.Shed Storehouse Heat-trap
Type of exposure

Fig. 7. Annual average Chloride deposition rate for test station located at the western side of the Cuban Isle and Campeche PCGM coastal station
depending on exposure conditions and type of atmosphere. Wet Candle method was used for determination of chloride deposition at Campeche and Dry
deposition methods for Cuba. A correct comparison can not be made under these conditions.

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 11

Sulphur compounds deposition rate at different exposure conditions

40

35

Deposition rate (mg/m2d)

30

25
Rural
Urban-industrial
20
Coastal Cuba
Coastal Campeche
15

10

F
5

OO
0
Outdoor V. Shed Storehouse Heat-trap
Type of exposure

Fig. 8. Annual average sulfur compound deposition rate for test stations located in the western side of the Isle of Cuba and Campeche PCGM coastal
station depending on exposure conditions and type of atmosphere.

PR
728 confirms the importance of Chloride deposition in deter- of corrosion rate at Havana. For copper, the difference 765
729 mining corrosion aggressivity in Cuba. does not appreciably change with time; but it does for steel. 766
A more aggressive action of chloride ions is observed in the 767

730 7. TOW-ISO and rain case of steel. 768

The remarkable difference in the acceleration caused by 769
chloride deposition rate in Havana and Medellin could be
ED
770
731 Chloride ion is one of the most important natural pollu-
732 tants influencing corrosion; particularly in the tropical due to the considerable difference in the rain regime 771

733 humid conditions of Cuba and Campeche in Mexico; how- between both sites. Other characteristic of the environment 772

734 ever, its role may change depending on climate. A very could also have influence, but rain should play an impor- 773

735 humid climate can cause a fast leaching of the chloride ions tant role. 774
CT

736 and decreases its effect on the acceleration of corrosion Samples were submitted to the same artificial chloride 775

737 rate. A different acceleration rate has been reported for deposition rate at Havana and Medellin stations (suppos- 776

738 chloride ions between Eastern and Western Caribbean ing that differences concerning natural values of chloride 777

739 [2]. In a given climate the influence of rain could change deposition rate are not very significant). Samples were pre- 778
pared from the same origin material, so the differences on
E

779
740 the acceleration caused by chloride ion on metal corrosion.
741 Differences between rainy and dry periods could be impor- the corrosion behavior can be associated only to climatic 780
characteristics. In the case of Medellin the washing effect 781
RR

742 tant in determining such acceleration. The electrochemical

743 mechanism of corrosion by chloride ions does not change, of rain should be higher than in Havana because the rain 782

744 but the time during which chloride ions acts on the metallic amount is more than four times than the reported for 783

745 surface and their concentration may appreciable change, Havana during the exposure period. It explains why a 784

746 depending on factors such as climatic conditions and nat- lower acceleration rate of chloride ions is obtained at the 785
former station. Based on this fact, a model for studying
CO

786
747 ure of corrosion products.
748 A research about the role of the rainfall characteristics the influence of the washing or cleaning effect of rain in 787

749 on the acceleration rate caused by chloride ions between the determination of the acceleration rate of chloride ions 788

750 two stations having noticeable differences in rain regime is proposed. 789

751 was already published [34]. The influence of chloride ions The bi-logarithm equation for atmospheric corrosion 790
UN

752 at Medellin and Havana corrosion stations was studied establishes that: 791

753 by submitting samples of steel and copper to salt spray dur- K ¼ atb 793
754 ing exposure to the open atmosphere. Up to 12 months of
755 exposure, the acceleration on the deterioration caused by where K is the mass loss; a and b are the constants; t is the 794
756 chloride ions was notably up to 12 months of exposure, time of exposure. 795
757 being higher at Havana station for steel and copper. In the presence of a given value of chloride deposition 796
758 The corrosivity category for steel and copper (C3) in rate, an acceleration of corrosion takes place; this acceler- 797
759 natural conditions in Havana increases to C5 for copper ation means that corrosion increases with time. At the 798
760 and over C5 for steel when salt spray is applied; however same time, the acceleration of corrosion caused by chlo- 799
761 at Medellin, copper corrosion increases to C4 for copper rides depends on the washing or cleaning effect of rain. 800
762 and C5 for steel under the application of salt spray. Under this condition the following model is proposed: 801
763 The acceleration rate caused by the addition of a salt
764 spray under natural conditions causes a higher acceleration K ¼ atb ½Clc ðW =DÞd 803

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

12 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

804 where K is the mass loss; a, b, c and d are the constants; 8. Corrosion in tropical coastal atmospheres. Role of TOW- 853
805 [Cl]is the chloride deposition rate; W is the rainfall (mm); ISO 854
806 D are the rainy days; t is the time of exposure.
807 It has been considered that the washing or cleaning Changes in TOW-ISO are negligible between all the test 855
808 effect of rain could be represented by the ratio W/D stations of the Cuban western side. Campeche coastal sta- 856
809 (amount of rain/frequency of rain). This washing effect tion report a lower value of TOW-ISO. All values are clas- 857
810 could affect the influence of chloride deposition rate on sified as s4 according to ISO 9223. In these conditions main 858
811 corrosion. changes depend on contaminants deposition. Due to the 859
812 A regression analysis including steel mass loss data for hygroscopic properties of the salts deposited in the surface 860
813 Havana and Medellin showed the following results: there is no doubt that corrosion should occur at a relative 861

F
humidity lower than 80%, so in this case the use of TOW- 862
0:4313 0:2447
K ¼ 18:8516t1:0947 ½Cl ðW =DÞ ISO has no sense. Only in coastal atmospheres having 863

OO
815
2
r ¼ 98:11 lower chloride deposition could be possible the use of 864
TOW-ISO. 865
2
816 where K is the mass loss (g/m ); t is the time of exposure
817 (months); [Cl] is the chloride deposition rate (mg/m2 d);
818 w is the rainfall amount (mm); D is the number of rainy 9. TOW and corrosion products of steel 866

PR
819 days.
820 A regression analysis including copper mass loss data Chloride ions move through the corrosion products 867

821 for Havana and Medellin showed the following results: layer and arrive at metal surfaces producing a notable 868
acceleration in corrosion rate. A catalyst role has been sug- 869

K ¼ 1:2466t0:7858½Cl
0:2616
ðW =DÞ
0:2379 gested for anions in the corrosion process. Chloride ions 870
are very soluble, that is why they are easily removed from 871
r2 ¼ 98:08
ED
823 metallic surfaces due to liquid precipitation. Usually, a low 872
concentration of chloride ions is determined in corrosion 873
824 A good data fit is also obtained for copper. It confirms products on outdoor exposures. A relation between chlo- 874
825 that the complex chloride deposition rate–rain regime is ride concentration present in corrosion products and cor- 875
826 important for determining mass loss of copper and steel. rosion rate has been reported. 876
CT

827 In this model TOW-ISO has not been used and a good fit- The adsorption isotherms for metallic surfaces are 877
828 ness has been obtained. reported in the literature; however, an important part of 878
829 It confirms that the acceleration rate caused by chloride the atmospheric corrosion process takes place under rust 879
830 ions on atmospheric corrosion of steel and copper depends layers, which play a decisive role in the long-term course 880
E

831 on the characteristics of rain regime. For a place having of corrosion because of its sorption capacity for water. 881
832 high amount and time of rain, a lower acceleration on cor- The influence of the chloride and sulfate anions has a real 882
833 rosion rate should be expected for a given chloride deposi-
RR

effect only when the corrosion products layer is already 883

834 tion rate formed. Thus, the adsorption isotherms of the steel corro- 884
835 Data of outdoor exposure of carbon steel and copper sion products formed in different atmospheres were deter- 885
836 during 2 years using TROPICORR samples (1  5 cm) mined [21]. 886
837 were processed together with data for the same exposure Water adsorption should be caused by salts content but, 887
CO

838 periods of TOW-ISO at temperature 25 °C or below, perhaps, the presence of high levels of contaminants can 888
839 TOW-ISO at temperature higher than 25 °C, time of rain determine the formation of a corrosion product having a 889
840 and amount of rain. The regression equation obtained more ‘‘hygroscopic” nature. To make clear this possibility, 890
841 was the following: adsorption isotherm of corrosion products from Viriato 891
corrosion station after elimination of salts were deter- 892
K steel ¼ 59:93 þ 0:021TOW-ISO T 6 25
UN

mined. The values of adsorption are lower after salt elimi- 893
þ 0:218TOW-ISO nation, but higher than those corresponding to Quivican. It 894
843 T > 25  2:05train  1:28 mm=train : could be due to a more ‘‘hygroscopic” nature of corrosion 895
products formed in coastal stations. 896
844 As it can be observed, the washing and cleaning effect of It has been reported [21] that the superficial characteris- 897
845 rain causes that time of rain and the ratio amount/time of tics of atmospheric corrosion products of steel depend on 898
846 rain show a negative sign, indicating a diminution of corro- the type of atmosphere where the sample has been exposed. 899
847 sion when time and ratio amount/time of rain increases, The way of adsorption of the corrosion products obtained 900
848 confirming the importance of washing and cleaning effect in the coastal atmosphere is polymolecular due to a higher 901
849 at Campeche tropical climate. content of salts. This makes easier the presence of water in 902
850 The washing and cleaning effect of rain is not included the metal–corrosion products interface and determines a 903
851 and is far from the concept of TOW established on ISO high corrosion rate. The adsorption of water of a corrosion 904
852 9223 standard. product formed in a rural zone obeys a Langmuir isotherm, 905

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx 13

906 i.e. a monomolecular adsorption takes place. It causes a 11. Uncited reference Q2 960
907 lower corrosion rate. All these results show the importance
908 of corrosion products in determining the presence of an [30]. 961
909 electrolyte on the metallic surface.
Acknowledgements 962

910 10. Conclusions

The authors appreciate the contribution of Dr. Lucien 963
Veleva, National Coordinator of TROPICORR project 964
911 TOW considered as the time during which corrosion
(CYTED), M.Sc. Miguel Ramón Sosa Baz (Corrosion Re- 965
912 occurs is an important parameter in atmospheric corrosion
search Center, University of Campeche, Campeche, Mex- 966
of metals. It defines the possibility for atmospheric corro-

F
913
ico) and Tec. Eva González Mellor and Julia Pérez 967
914 sion to occur based on its electrochemical nature, but cor-
Acosta (National Center for Scientific Research (CNIC), 968
rosion rate will depend mainly in the acceleration caused

OO
915
Havana City, Cuba). Part of this paper is related to results 969
916 by contaminants deposition and other factors.
of CYTED TROPICORR project. The authors thank this 970
917 TOW according to ISO definition shows several limita-
program for support. 971
918 tions in tropical climate:

919  It is a variable perfectly linked to time. It is not neces- References 972

PR
920 sary to calculate TOW-ISO when the development of
921 corrosion rate on time is studied because a linear rela- [1] F. Corvo, N. Betancourt, A. Mendoza, Corrosion Science 37 (12) 973
(1995) 1889–1901. 974
922 tionship exists between time and TOW-ISO. [2] F. Corvo, C. Haces, N. Betancourt, L. Maldonado, L. Veleva, M. 975
923  Its nature changes from outdoor to indoor conditions. Echevarria, O.T. de Rincón, A. Rincón, Corrosion Science 39 (50) 976
924 Higher TOW-ISO values are reported for indoor condi- (1997) 823–833. 977
978
ED
925 tions where corrosion rate is significantly lower. It is [3] A.R. Mendoza, F. Corvo, Corrosion Science 41 (1999) 75–86.
926 mainly due to the absence of precipitations. [4] A.R. Mendoza, F. Corvo, Corrosion Science 42 (2000) 1123–1147. 979
[5] L. Veleva, L. Maldonado, British Corrosion Journal 33 (1) (1998) 53– 980
927  It has no sense to calculate TOW-ISO for coastal trop- 57. 981
928 ical atmospheres having an appreciable amount of chlo- [6] D.C. Cook, A.C. Van Orden, J. Reyes, S.J. Oh, R. Balasubramanian, 982
929 ride deposition, because in those conditions corrosion J.J. Carpio, H.E. Townsend, Atmospheric corrosion in marine 983
CT

930 process occurs at relative humidity lower than 80%. It environments along the Gulf of México, in: S.W. Dean, G. Hernández 984
931 has been determined that water adsorption by corrosion Duque-Delgadillo, J. Bushman (Eds.), Marine Corrosion in Tropical 985
Environments, ASTM STP 1399, American Society for Testing and 986
932 products is polymolecular in these conditions. Materials, West Conshohocken, PA, 2000. 987
933  TOW-ISO does not take into account the washing and [7] M. Sosa, A. del Sol, Revista de Climatologı́a, Cuba 2 (2004) 1–4. 988
E

934 cleaning effect of rain, a very important aspect in atmo- [8] L. Mariaca, J. Genesca, J. Uruchurtu, L. Salvador. Corrosividad 989
935 spheric corrosion outdoors. Atmosférica (MICAT-México). Programa Iberoamericano de Ciencia 990
936  An upper limit for air temperature has not been included y Tecnologı́a para el Desarrollo (CYTED), 1999. Editorial Plaza y 991
RR

Valdés S.A. de C.V., México. 992

937 in TOW-ISO definition. Several results show that a sig- [9] ISO 9223:1992. Corrosion of metals and alloys. Corrosivity of 993
938 nificant change in TOW-ISO takes place at air tempera- atmospheres. Classification. 994
939 ture higher than 25 °C and RH >80%. [10] C. Leygraf, T.E. Graedel, Electrochemical Society Series, Wiley 995
940 Intersciences, 2000. 996
[11] ISO/WD 9223. Corrosion of metals and alloys. Corrosivity of 997
CO

941 In order to improve the estimation of TOW-ISO it is

atmospheres. Classification. 998
942 proposed: [12] R. Pascual, F. Corvo, Rev. Iberoamericana de Corrosión y Pro- 999
tección, mayo-junio, 1980. 1000
943  To establish an upper limit of temperature for TOW- [13] F. Corvo, J. Rocha, Characterization of environments and their 1001
944 ISO calculation, dividing TOW-ISO in two categories: influences on copper at different atmospheres and exposure conditions 1002
in tropical climate, International Corrosion Congress (ICC), Gra- 1003
UN

945 Air temperature 0 > 25 °C< and air temperature >25 °C.
nada, Spain, 2002. 1004
946  To include time and amount of rain as an additional [14] L. Veleva, M.A. Alpuches-Aviles, Outdoor atmospheric corrosion, in: 1005
947 variable, taking into account the washing and cleaning H.E. Townsend (Ed.), ASTM, STP 1421, American Society for 1006
948 effect of rain. Testing and Materials International, West Conshochoken, PA, 2002. 1007
949  To limit the use of TOW-ISO to outdoor and not highly [15] Tooru Tsuru, Ken-Ichiro Tamiya, Atsushi Nishikata, Electrochimica 1008
950 contaminated environments. Acta 49 (2004) 2709–2715. 1009
[16] Li Bian, Yongji Weng, Xiangyi Li, Electrochemistry Communications 1010
951  Always take into account the influence of contaminant 7 (2005) 1033–1038. 1011
952 deposition and its interaction with TOW. [17] Jibiao Zhang, Jia Wang, Yanhua Wang, Electrochemistry Commu- 1012
953  The concept of an adsorptive or liquid film of electrolyte nications 7 (2005) 443–448. 1013
954 on the metallic surface should include the possibility of [18] Wang Jia, Zhang Jibiao, The effect of micro-droplets formation 1014
955 the existence of localized presence of adsorptive or caused by the deliquescence of the deposited salt particle on 1015
958
atmospheric corrosion of metals, in: Proceedings of the 16th 1016
956
959 liquid films on the metallic surface, particularly in International Corrosion Congress, Beijing, China, September 19–24, 1017
957 indoor conditions. 2005. 1018

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012
 CS 2811 No. of Pages 14, Model 5+
ARTICLE IN PRESS
6 August 2007 Disk Used

14 F. Corvo et al. / Corrosion Science xxx (2007) xxx–xxx

1019 [19] J. Tidblad, A.A. Mikhailov, V. Kucera, Application of a model for [25] C. Leygraf, Indoor atmospheric corrosion. Proceedings 15th Inter- 1041
1020 prediction of atmospheric corrosion in tropical environments, in: national Corrosion Congreso, Granada, Spain, September 22–27, 1042
1021 S.W. Dean, G. Hernández Duque-Delgadillo, J.B. Bushman (Eds.), 2002. 1043
1022 Marine Corrosion in Tropical Environments, ASTM STP 1399, [26] F. Corvo, A.D. Torrens, N. Betancourt, J. Perez, E. Gonzalez, 1044
1023 American Society for Testing and Materials, Conshohocken, PA, Corrosion Science 49 (2) (2007) 418–435. 1045
1024 2000. [27] I.S. Cole, R. Holgate, P. Kao, W. Ganther, Corrosion Science 37 (3) 1046
1025 [20] Iu. N. Mikhailovskii, Atmospheric corrosion of metals and protection (1995) 455–465. 1047
1026 methods (in Russian). Editorial ‘‘Metallurgiia”, Moscow, 1989. [28] M. Stratmann, H. Streckel, Corrosion Science 30 (6/7) (1990) 697– 1048
1027 [21] F. Corvo, A.R. Mendoza, M. Autie, N. Betancourt, Corrosion 714. 1049
1028 Science 39 (4) (1997) 815–820. [29] M. Stratmann, H. Streckel, K.T. Tim, S. Crockett, Corrosion Science 1050
1029 [22] José Rocha Andrade da Silva Editor. Productos electro-electrónicos 30 (6/7) (1990) 715–734. 1051
1030 en ambientes tropicales, Campinas, Sao Paulo, Sitta Gráfica, 2003. [30] ISO 11844-3. Corrosion of metals and alloys. Classification of 1052

F
1031 [23] L. Veleva, L. Dzib-Perez, J. Gonzalez-Sanchez, T. Perez, Initial stages corrosivity of indoor atmospheres. Measurement of environmental 1053
1032 of indoor atmospheric corrosion of electronic contact metals in humid parameters affecting indoor corrosivity. 1054
1033 tropical climate: tin and nickel, Rev. Metalurgia (CENIM) 43 (2007) [31] ISO 11844-1. Corrosion of metals and alloys. Classification of 1055

OO
1034 101–110. corrosivity of indoor atmospheres. Determination and estimation of 1056
1035 [24] L. Dzib-Perez, J. Gonzalez-Sanchez, T. Perez, L. Veleva, M. Sosa, indoor corrosivity. 1057
1036 Indoor atmospheric corrosivity of metals used in electro-electronic [32] J.P. Cai, S.B. Lyon, Corrosion Science 47 (2005) 2956–2973. 1058
1037 components in humid tropical climate, in: J.A. Gonzalez-Sanchez, [33] Seon Yeob Li, Young-Geun Kim, Sungwon Jung, Hong-Seok Song, 1059
1038 Tezozomoc Perez Lopez (Eds.), IV International Materials Research Seong-Min Lee, Sensors and Actuators B 120 (10) (2007) 368–377. 1060
1039 Congress 2005, Simposium 19, Cancun, Quintana-Roo, Mexico, 21– [34] F. Corvo, J. Minotas, J. Delgado, C. Arroyave, Corrosion Science 47 1061

PR
1040 25 August 2005, 96p. (2005) 883–892. 1062
1063

ED
E CT
RR
CO
UN

Please cite this article in press as: F. Corvo et al., Time of wetness in tropical climate, Corros. Sci. (2007), doi:10.1016/
j.corsci.2007.06.012