Atmospheric Pollution Is Known To Induce Corrosion Effects

Published Date: 02 Nov 2017

corrosion effects on various materials. For Greece,

stone deterioration could emerge severe costs in the

case of damaging cultural monuments. This work

aims to investigate the corrosion process on materials

of archaeological importance (marble, limestone, and

sandstone) in the Greater Athens Area (GAA) by

using sophisticated geoanalytical methods together

with dose–response functions for selected materials,

in order to derive corrosion maps for GAA in the

period 2000–2009. Also, a corrosion trend analysis is

performed, which can be a very helpful tool for the

prediction of potential risks to monuments of cultural

heritage due to atmospheric pollution. The corrosion

effects on the selected materials are generally weak.

Nevertheless, increasing corrosion trends are found in

the eastern regions of GAA for all sheltered materials

and in the northern parts of GAA for unsheltered

marble. The technique is finally applied to 12

locations in GAA, which include some of the most

important archaeological monuments of Athens, and

provides comprehensive results for the estimation of

the impact of atmospheric corrosion on the structural

materials of these archaeological sites.

Keywords Air pollution . Corrosion of materials .

Dose–response . Archaeological sites . Athens . Greece

1 Introduction

It is well established that atmospheric pollution causes

various problems to human health, such as chronic

diseases (Becker et al. 2002; Bell et al. 2011; Kampa

and Castanas 2008; Lechón et al. 2002; Wang et al.

2008; Wanner 1990; Yang and Omaye 2009), it

contributes to forest decline and plant elimination

(Bussotti and Ferretti 1998; Bytnerowicz et al. 2007;

Oszlanyi 1997; Paoletti et al. 2010), and in many

cases, it presents significant corrosion effects on

various materials like metals, plastics, wood, building

materials, and cultural heritage monuments (Graedel

and McGill 1986; Johansson 1990; O’Brien et al.

1995; Schuster and Reddy 1994; Van Grieken et al.

1998). Many of these materials need special attention

because of their great social, historical, and economical

value (Bell et al. 2011; Graedel and Leygraf 2001;

Kucera and Fitz 1995; Kucera 2002; Mirasgedis et al.

2008; MULTI-ASSESS 2005; Screpantia and de

Marco 2009). After the adoption of the Convention

on Long-Range Transboundary Air Pollution within

the United Nation Economic Commission for Europe

in 1979, a series of International Cooperative

Programs (ICP) was initiated for assessing the effects

of atmospheric pollutants on several materials of

major interest (Kucera 2002; Kucera et al. 2007;

Mikhailov 2001; MULTI-ASSESS 2005; Tidblad et

al. 1998, 2001; Tidblad 2009).

For a long time, sulfur has been considered as the

main pollutant; nevertheless, many researchers have

recently concluded that nitric oxides and ozone,

together with favorable meteorological parameters,

such as temperature, relative humidity, and precipitation,

play an important role in the corrosion effects

deduced on materials (Johansson 1990; Kucera and

Fitz 1995; Lan et al. 2005; Lipfert 1989; O’Brien et

al. 1995; Roots 2008; Schuster and Reddy 1994;

Screpantia and de Marco 2009; Tidblad et al. 1998;

Van Grieken et al. 1998). To study these effects,

dose–response functions (DRF) for corrosion on

materials have been derived and applied (Kucera

2002; Kucera et al. 2007; Mikhailov 2001; MULTIASSESS

2005; Tidblad et al. 1998; Tidblad et al.

2001). These scientific tools represent the relationships

between climate and air pollutants on the one

hand and the resulted deterioration of structural materials

on the other; they have been mainly used by the ICP

on Materials for more than 20 years to determine the

mass loss/increase of the materials under corrosion

attack in sheltered and unsheltered locations with very

promising results. These DRFs have been adopted in

the present work for a quantitative analysis of the

materials under investigation.

For Greece, a nation synonymous with ancient

civilization and cultural heritage formore than 2,500 years,

the protection and preservation of such monuments are

more than a national necessity (Mirasgedis et al. 2008;

Moropoulou et al. 1998). Athens, the cradle of

civilization, has numerous such historical monuments.

This work is, therefore, focused on the corrosion

effects on materials of archaeological value, i.e.,

marble, limestone, and sandstone, which are dominant

in ancient monuments in the metropolitan area of

Athens; the study also aims to quantify the potential

risks to these materials from their diachronic exposure

to the atmospheric pollution of the city under

sheltered and unsheltered conditions.

The above aims are implemented by producing

annual DRF maps over the Greater Athens Area

(GAA) in the period 2000–2009 for each of the

selected materials. The results show weak corrosion

for all materials, with greater spatial variations for

marble. These corrosion maps can be considered a

useful tool in predicting corrosion effects over the

region and can thus give to the Directorate of

Conservation of Ancient Monuments, Ministry of

Culture and Tourism, the necessary information

about the degree of corrosion of the cultural

heritage monuments in GAA. Corrosion trends are

also calculated for the selected materials by using

linear trend analysis for the corresponding DRF

values. For all the materials under investigation, an

increasing corrosion trend is observed in the east of

GAA, except for unsheltered marble where the

greatest corrosion rate occurs in the northern region

of GAA. These corrosion trend maps thus serve as

a potential protection tool for materials of important

archaeological value (Graedel and Leygraf 2001;

Kucera and Fitz 1995; Tidblad et al. 1998; Tidblad

2009). The above techniques are applied to some

important cultural heritage monuments in GAA, in

order to provide a quantitative estimation of the

atmospheric corrosion effects on the materials of

these monuments.

2 Materials and Methods

The mandatory parameters used in this study are the

mean annual ambient temperature (temperature in

degree Celsius), the mean annual relative humidity

(RH in percent), the total annual precipitation (PR in

millimeter), the mean annual time of wetness (TOW,

defined as the time fraction of the days with T >0°C

and RH >80%), and the mean annual concentration of

SO2 (in micrograms per cubic meter). For the

calculation of the mean annual TOW, daily values of

ambient temperature and relative humidity have been

used. The annual values of these parameters were

taken from various monitoring stations within GAA.

Table 1 presents these stations. This monitoring

network includes 18 environmental monitoring stations

(air pollution and meteorology) operated by the Ministry

of Environment, Energy and Climate Change

(DEARTH network), the 2 meteorological stations of

the National Observatory of Athens (NOA), and the 10

hydrological–meteorological stations operated by the

National Technical University of Athens (HOA

Table 1 The monitoring

stations in GAA ID Station Longitude (deg) Latitude (deg) Altitude (m above

mean sea level)

network). The location of all the above stations is shown

in Fig. 1.

Due to technical problems and/or lack of data,

some of the collected data series were incomplete for

some years and stations. To obtain a complete data

series for the problematic meteorological and/or air

pollutant parameter(s) at each station of Table 1, the

kriging geostatistical analysis was applied to the

whole region of GAA (as depicted in Fig. 1) for the

parameter(s) in question and the year(s) presenting the

gap(s), using the ArcGIS 10 software. This way, the

values of the parameters for those years, which had

missing data at some stations, were identified and the

gaps were filled. Kriging/Cokriging is an advanced

geostatistical tool that generates an estimated surface

from a scattered set of points. The procedure assumes

that the distance or direction between sample points

reflects a spatial correlation that can be used to

explain variation in the surface. The kriging tool fits

a mathematical function to a specified number of

points, or all points within a specified radius, to

determine the output value for each location. The

result of the kriging method gives optimal and

unbiased estimates.

Table 2 gives a list of all data used in this study

(measured and computed by kriging). Though meteorological

data were available at many locations in

GAA in the period of 2000–2009, additional precipitation

data at other locations than those of Table 2

only started in late 2005. Table 3 shows the additional

annual precipitation values used in the present work

from the HOA network. As in the case of air pollutant

and meteorological data in Table 2, the kriging

methodology was also applied to precipitation station

H2 in year 2006 to replace the missing value and to

fill the gaps at the stations of Table 1 (see derived

values for the DEARTH and NOA stations in the

second rows of Table 2 for years 2006–2009).

For sheltered materials (marble, limestone, and

sandstone), the DRFs were calculated for the whole

period of investigation. Table 4 presents the

corresponding DRFs used in this study. In the

literature, there also exist DRFs for limestone and

sandstone with equations that include the concentration

of ions (H+, Cl−

) in the precipitation or the

concentration of particulate matter (Lan et al. 2005;

Kucera 2002; Kucera et al. 2007; MULTI-ASSESS

2005; Tidblad 2009); these specific DRFs were not

taken into account in this work as measurements of

the concentration of ions are not available in GAA.

By using the DRFs of Table 4, the geostatistical

program ArcGIS 10 was applied to the whole area of

GAA to derive corrosion maps for every selected

material and each individual year as well as for the

total period of investigation. These maps indicate the

DRF values in all GAA, including the selected sites of

archaeological interest, which are presented in Table 5,

together with the main material that these monuments

are made of. Thus, for the whole period under

investigation, the degree of corrosion could be determined

for the 12 archaeological sites from the predicted

DRF values (by means of mass increase for limestone

and sandstone and of surface recession for marble).

In the next step, the linear trends of both sheltered

and unsheltered materials for all years in the period of

investigation were derived at each archaeological site.

The corrosion (DRF) trend is expressed as mass

increase/surface recession per year for each material.

Following the geostatistical technique for producing

DRF maps above, DRF-trend maps over GAA were

also derived. These maps can determine the degree of

the risk for corrosion attack on the selected materials

in GAA due to the dominant air pollutants and the

climatology in the area.

Having finalized the DRF and DRF-trend maps for

marble, limestone, and sandstone in GAA, special

attention was paid to the corrosion effects at the 12

sites of Table 5, where some important cultural

heritage monuments exist. Table 6 presents the DRF

trends at the archaeological sites of Table 5 for marble

(in micrometers per year) and limestone/sandstone (in

grams per square meter per year).

3 Results and Discussion

3.1 SO2 Concentration Profile in GAA

The sulfur dioxide database formed from measurements

and data gap filling (using kriging analysis) resulted in a complete SO2 mapping over GAA. As

one can see from the dose–response functions

(Table 4), this gas pollutant plays a significant role

in the corrosion impact on stone materials. For the

period 2000–2009, maximum SO2 concentrations

were found at Patission station (#D14 in Table 2)

and Piraeus-1 station (#D15) for all years, showing a

peak in 2003 with concentrations of 42.88 and

31.2 μg m−3, respectively. For the whole period,

minimum SO2 concentration was observed in the east

of GAA, with values of about three to four times

smaller than those at the stations mentioned before; an

exception was year 2009 when the situation was

totally different and the SO2 concentration became

highest at Koropi station (in the SE of GAA, #D18 in

Table 2), having a value of 18.05 μg m−3, the same

time that the SO2 concentration was about 14 μg m−3

at Patission and Piraeus stations and half of that at the

other sites of GAA. By comparing the mean concentration

values of SO2 over the whole GAA, one can

observe a negative (decreasing) trend for the specific

pollutant since in 2000 its mean concentration was

∼19 μg m−3 and in 2009 only half of that.

3.2 Materials Behavior and Corrosion Trends

3.2.1 Marble (Sheltered–Unsheltered)

In the case of marble, which is the predominant material

used in ancient monuments in Greece, the DRF

mapping of GAA reveals some very interesting results.

For the sheltered locations, following the SO2 concentration

profile, atmospheric corrosion has led to

maximum surface recession (SR) at Patission station

(#D14) for the period 2000–2008,with an SR value about

twice the average of all stations of Table 1, showing also

a secondary maximum at Piraeus-1 station (#D15)

after 2006. In general, the marble corrosion is weak,

with greater values of SR of about 1–1.5 μm in the

northern–northwestern areas of GAA, comparing to

values of about 0.3–0.6 μm in the southern–southeastern

regions of GAA (except from year 2000 onwards

when corrosion becomes greater in south–southwest

GAA and smaller in north–northeast GAA).

For unsheltered places and starting from 2006, marble

corrosion is about three to four times greater for the

whole area under investigation than for the sheltered ones

(about 3.7 μm for the maximum and 2 μm for the

minimum SR values). In 2009, as is shown in Fig. 2, the

corrosion seems to move eastward. Taking, therefore,

into consideration the archaeological sites of Table 5,

it is seen that marble corrosion is almost constant at

all locations, but Panathinaiko stadium (#11 in Fig. 2)

is affected the most and Dimitra’s Sanctuary (#7 in

Fig. 2) the least by atmospheric pollution, with SR

values of 2.71 and 2.42 μm, respectively.

For the corrosion trends, some differences exist

between sheltered and unsheltered marble. For the sheltered material, an increasing corrosion trend in the

east and a decreasing corrosion trend in the south

(having also a secondary minimum in the center of

Athens) are observed. For Patission (#D14) and

Piraeus-1 (#D15) stations, where marble corrosion

was previously determined to reach its maximum, the

corrosion variation per year is decreasing. For the

unsheltered marble, with relative data corresponding

to the period 2006–2009 as mentioned before, the

trends are the same as for the sheltered, but three to

four times greater. The important difference compared

to the sheltered case is that the greatest corrosion

trend occurs now in the north of GAA, while the

greatest SR decrease is observed in the south of GAA.

At Patission station (#D14), the corrosion trend is

again decreasing.

By applying kriging analysis, prediction of the

corrosion trend for the archaeological locations of

Table 5 was possible. For sheltered marble, as is

shown in Table 6, an almost constant decrease

corrosion rate for all sites occurs (−0.04 μm year−1).

For unsheltered marble, the situation changes since

the annual DRF variation is still decreasing at all places,

but not at the same rate. So, at Pnyx (#2 in Table 6) the

marble DRF decreases with a rate −0.06 μm year−1,

while at Dimitra’s Sanctuary (#7), it decreases with a

rate −0.19 μm year−1. The corrosion trend over GAA,

including the 12 archaeological sites, is presented in

Fig. 3. Here, dark blue areas in west and south GAA

correspond to decreasing DRF rates (as shown e.g. at

Dimitra’s Sanctuary, #7 in Fig. 3), while red-colored

areas, such as the eastern part of GAA, correspond to

an annual increasing trend of marble corrosion. Near

the historical center of Athens (including the archaeological

places of Acropolis, Pnyx, Kerameikos,

Ancient Agora, and Akadimia Platonos, #1, #2, #5,

#6, and #8 in Fig. 3, respectively), the trend is slightly

decreasing. Limestone corrosion has a more smooth variation within

GAA than marble, showing a difference of ∼60%

between the observed maximum and minimum values

of mass increase (MI) for the whole period under

investigation. The maximum corrosion effects on limestone,

occurring from the corresponding DRF, are

observed in western GAA (year 2000), moving to

northern GAA (years 2001–2007) and then to eastern

GAA (years 2008–2009), while the minimum is in the

east (year 2000), the south–southwest GAA (years 2001–

2005) and in the center of Athens (years 2006–2009).

The corresponding corrosion map for sheltered limestone

in GAA for 2009 is presented in Fig. 4 for all the

archaeological locations of Table 5; the blue areas correspond to minimum and the red areas to maximum

MI values.

As far as the corrosion trends are concerned, Fig. 5

presents the corrosion trend map for the period 2000–

2009 in GAA for sheltered limestone. The annual

variation is small, with the red areas in Fig. 5

representing an increase of limestone’s MI trend in

the east of GAA, while the blue areas in the north of GAA

correspond to negative corrosion rates with time. For the

archaeological locations of Table 5, a negative corrosion

trend is observed, with an average decreasing rate in

limestone’s MI of about −0.01 g m−2 year−1. 3.2.3 Sandstone (Sheltered)

Sandstone behaves in almost the same manner as

limestone for the total period and all areas of GAA.

The only difference is that the MI values are about

10–20% greater than the respective values for

limestone at each site. The corrosion trend for

sandstone follows that of limestone, with the maximum

SR (increasing corrosion) observed in the east

GAA and the minimum (decreasing corrosion) in the

north GAA. As for the DRF trend, as one can see

from Table 6 it is again decreasing for all the locations

of archaeological importance of Table 5, with an

almost identical profile with sheltered limestone.

4 Conclusions

Athens consists of many archaeological monuments,

with the majority of them consisting of marble,

limestone, and sandstone. Since atmospheric pollution

has been proved to cause corrosion effects on these

materials, this work was focused on presenting a

quantitative method for determining the potential

risks from corrosion on marble, limestone, and

sandstone over the Greater Athens Area, in order to

preserve and protect the cultural heritage monuments.

The use of experimental data from a wide network of

meteorological stations, together with dose–response

functions for each material to quantify corrosion

effects and sophisticated analysis methods (kriging),

resulted to corrosion maps for the three materials in

sheltered and unsheltered conditions. So, annual

profiles for the corrosion behavior for each material

were deduced.

For the sheltered marble, an increasing corrosion

trend in the east and a decreasing one in the south

were observed in GAA for the period 2000–2009. For

the unsheltered marble, the trends were identical with

those of the sheltered, but with three to four times

greater absolute values. For the sheltered limestone,

an increasing corrosion trend in the east and a

decreasing one in the north were observed in GAA

for the same period. For the sheltered sandstone, an

almost identical profile with that for the sheltered

limestone was found in GAA for 2000–2009.

Also, corrosion trends were evaluated for the period

2000–2009 for the materials under investigation, leading

to the production of corrosion trend maps over

GAA, which can be used as a guide to predict corrosion

impact on the archaeological sites in GAA.

Acknowledgment The authors would like to thank the

Hydrological Observatory of Athens (hoa.ntua.gr) for providing

some of the precipitation data used in this study.