Workshop 4

‘Ploughing’, an unusual form of corrosion of archaeological glass

Vasilios LAMPROPOULOS, Antigoni LEAKOU* and Athanasios KARAMPOTSOS

Summary

This paper is a preliminary study of a form of corrosion

of archaeological glass that until now has been record-
ed, but not studied. Ten samples of glass that exhibit
this type of corrosion have been examined with the
scanning electron microscope (S.E.M.) and analysed
by energy dispersive X-ray analysis (E.D.X.A.). In addi-
tion, three samples of the soil in which they were
found have been tested by X-ray diffraction (X.R.D.).
The present study contains a description of ‘plough-
ing’, the results of some analyses of the samples and
some preliminary conclusions.

Figure 2. Enlargement of Figure 3 in a stereoscopic micro-

Description scope. Grooves cutting all the way through the material.
This unusual form of corrosion appears in a glass sur-
face in the shape of channels with no specific direction.
These vary from shallow grooves to tunnels cutting all
the way through the material (Figures 1 and 2). Their
sides are almost perpendicular to the surface, while
their endings are circular. The grooves appear either
singly on the glass surface (Figure 3) or with branches
(Figure 4), thus creating a network of channels through-
out the object. Their direction and width is not specific.
This phenomenon was recorded for the first time by
Tomas Buechner [BUECHNER, 1960] in glass frag-
ments from the excavation in Tharra (southern Crete)
in 1959. It was he who gave the phenomenon the
name ‘ploughing’, and speculated that it might be
caused by leaching down from a scratch on the sur-
face or that it may be the effect of some kind of mi- Figure 3. Single grooves on a Rhodian bowl, dated early
cro-organism. The conservator Anita Moraitou 1st century BC.

Figure 1. Bowl from the central necropolis of the city of Figure 4. ‘Ploughing’ in a glass fragment from Rhodes.
Rhodes, dated 150-100BC (ΤΡΙΑΝΤΑΦΥΛΛΙ∆ΗΕΣ, 2000, p.47).

296
 The conservation of cultural heritage for sustainable development

a b c the difference in their concentration from one point

to another. Moreover, in all cases a ratio of concen-
tration of the network modifiers to the concentration
of the silicon was calculated, so as to estimate the
leaching of the elements. This proportion will be re-
ferred to further on as λ (λ= Calkalis/Csilicon).
d e f
In Figure 5 a wide area of a groove in a glass sample
from Rhodes is observed and the elements are
mapped using the scanning electron microscope. It
can be noticed that sodium has been almost totally
leached out in the interior of the groove. It can be
Figure 5. Mapping of the elements of a glass sample from also noticed that in the centre of the channel the sil-
Rhodes – a) picture of the area of the groove (x 75), b) sili- icon appears reduced. This is due to the gradual loss
con, c) calcium, d) aluminum, e) sodium, f) potassium. of the corroded material. In the rim of the groove and
for some distance to the interior there is evidence of
the presence of the network former, which proves
[ΜΩΡΑΙΤΟΥ, 1996] also agrees with this hypothesis,
the existence of glass in that area. At the same point
but remarks that the phenomenon needs to be stud-
one can notice that the concentration of calcium and
ied, so as to enable a verification of the assumption.
potassium has not changed, while sodium does not
After the examination with a stereoscopic micro- appear at all. The same conclusions can be reached
scope, it was observed that the decayed glass from by observing the pit that is next to the groove.
the interior of the grooves was loosely connected to
The area near the side of the groove appears in Fig-
the concrete mass of the uncorroded material and
ure 6 and the elementary analysis of this field is com-
that the degradation products are easily abstracted,
pared in Table 1 to that of uncorroded glass from the
leaving the sides of the groove almost perpendicular
same sample. The numbers in Table 1 confirm the
to the surface. One can also observe some kind of fo-
observations described above regarding the mapping
liated arrangement of the corroded glass, which be-
of the elements. Thus, the λ values for the elements
comes looser in the centre of the channel.
indicate that sodium has been leached out, while the
concentrations of calcium and potassium remain at
Results the same level in relation to the concentration of sili-
con. Moreover, the λ value for aluminum, which ap-
In all of the samples an area of uncorroded glass
pears to have increased, reveals that a small amount
was first selected to make an elementary analysis
of silicon has been leached out.
and then various areas in the interior of the grooves
to observe the morphology of this phenomenon and Two channels almost conjoined are shown in Figure
perform analyses. In the areas where it was judged 7. The glass surface is very corroded and exhibits iri-
necessary, a mapping of the elements was made to descence and pitting, in addition to ‘ploughing’. If the
create a clear image of their distribution, as well as groove is observed closer (Figure 8) the arrangement

Table 1.

Uncorroded glass ‘Ploughing’

El. Weight % λ Weight % λ
O 43.85 +/- 0.31 40.73 +/- 0.53
Si 22.39 +/- 0.08 16.33 +/- 0.12
C* 20.23 +/- 0.42 34.51 +/- 0.66
Na 8.07 +/- 0.07 0.36 0.65 +/- 0.05 0.04
Ca 3.05 +/- 0.05 0.13 3.07 +/- 0.07 0.19
K 0.83 +/- 0.03 0.04 0.96 +/- 0.05 0.06
Al 0.71 +/- 0.03 0.03 3.26 +/- 0.06 0.20
Cl 0.48 +/- 0.02 0.25 +/- 0.04
Mg 0.19 +/- 0.03 0.23 +/- 0.04

*The presence of carbon in the elementary analyses is due

to the graphite with which the samples were covered dur-
Figure 6. The edge of the groove (x500, S.E.M.). ing their preparation for the scanning electron microscope.

297
Workshop 4

Figure 7. ‘Ploughing’ in a glass sample from Rhodes Figure 8. Enlargement of the side of the groove of Figure 7
(x 3.2/0.06∞, metalographic microscope). (x 20/0.40∞, metalographic microscope).

of the decayed glass in layers is clearly seen. The el- of ‘ploughing’ and there has been no leaching out of
ementary analysis at this point (Table 2) indicates that silicon. Thus, it can be concluded that the image cre-
all of the elements appear reduced in the groove, in ated by the distribution of the elements, as revealed
comparison to uncorroded glass, due to the non-con- in the mapping, is similar to that of ‘ploughing’, but in
crete nature of the decayed material and the gradual the case of ‘ploughing’ the leaching out of the ele-
loss of it. The same conclusions as for the previous ments is more substantial than in the case of pitting.
sample can be deduced from the λ values for the el-
An examination of samples of the soil in which the
ements. Proceeding to the centre of the same groove
glass fragments were found has also been performed.
and performing an elementary analysis (Table 3), one
The samples were analysed by X-ray diffraction, and
can clearly see that in this area there are only grains
the main minerals traced were calcite (CaCO3) and
of glass and calcium depositions. This leads to the
quartz (SiO2). montmorilonite
conclusion that the loss of the material, which leads
((Al1,67Na0,33Mg0,33)(Si2O5)2(OH)2), kaolinite
to channels cutting through the glass, begins from
(Al4(Si2O5)2(OH)8) and calcium oxide (CaO) were traced
the centre and proceeds to the walls of the grooves.
in smaller amounts. Their conductibility was also
A large pit close to the borders of the ‘ploughing’ is measured and was found to be between 102 – 118 µS.
shown in Figure 9. The reduction of the concentra- The conductibility is relatively low, which allows one to
tion of sodium is obvious in this case too, while the assume that the glass samples were not being at-
remaining network modifiers do not seem to have tacked by soluble salts. The pH of the soil samples
been leached out. This is shown more clearly by the was found to be between 7.09 –7.21, which indicates
λ values for these elements in Table 4 containing re- that the glass fragments were buried in a slightly alka-
sults of the elementary analysis. The reduction in line environment, where silicon is not very stable.
sodium content in this case is less than in the case Overall, the original environment in which the glass

Table 2. Table 3.

Uncorroded glass ‘Ploughing’ Uncorroded glass ‘Ploughing’

El. Weight % λ Weight % λ El. Weight % λ Weight %
O 37.21 +/- 0.32 65.85 +/- 0.32 O 37.21 +/- 0.32 31.34 +/- 1.49
C* 28.98 +/- 0.43 24.63 +/- 0.36 C* 28.98 +/- 0.43 40.30 +/- 2.98
Si 21.41 +/- 0.08 6.04 +/- 0.04 Si 21.41 +/- 0.08 4.48 +/- 0.00
Na 6.54 +/- 0.06 0.30 0.20 +/- 0.02 0.03 Na 6.54 +/- 0.06 0.30 –
Ca 3.43 +/- 0.05 0.16 1.12 +/- 0.03 0.18 Ca 3.43 +/- 0.05 0.16 7.46 +/- 0.00
K 0.86 +/- 0.03 0.04 0.40 +/- 0.02 0.06 K 0.86 +/- 0.03 0.04 –
Al 0.70 +/- 0.03 0.03 1.28 +/- 0.02 0.21 Al 0.70 +/- 0.03 0.03 –
Cl 0.46 +/- 0.02 0.12 +/- 0.02 Cl 0.46 +/- 0.02 –
Fe 0.20 +/- 0.05 0.26 +/- 0.05 Fe 0.20 +/- 0.05 0.26 +/- 0.05

*see Table 1 *see Table 1

298
 The conservation of cultural heritage for sustainable development

a b c
Table 4.

Uncorroded glass ‘Ploughing’

El. Weight % λ Weight % λ
O 37.21 +/- 0.32 49.54 +/- 0.32
d e f C* 28.98 +/- 0.43 32.28 +/- 0.28
Si 21.41 +/- 0.08 12.79 +/- 0.05
Na 6.54 +/- 0.06 0.30 1.51 +/- 0.03 0.12
Ca 3.43 +/- 0.05 0.16 1.59 +/- 0.03 0.12
K 0.86 +/- 0.03 0.04 0.54 +/- 0.02 0.04
Al 0.70 +/- 0.03 0.03 1.12 +/- 0.02 0.09
Cl 0.46 +/- 0.02 0.23 +/- 0.02
Figure 9. Mapping of the elements in a large pit in a glass
Fe 0.20 +/- 0.05 0.13 +/- 0.03
sample from Rhodes – a) picture of the pit (x 75, S.E.M.), b)
silicon, c) calcium, d) sodium, e) aluminum, f) potassium. Mg 0.17 +/- 0.03 0.07 +/- 0.02

*see Table 1

samples were found, does not seem to have been ex- arrangement and the leaching out of the elements. In
ceedingly aggressive, since silicon is unstable when the case of ‘ploughing’, the leaching out is more in-
the pH is above 9. Nonetheless, even slightly alkaline tense, which indicates that it might be some kind of
circumstances can cause it to begin leaching out. evolution or conjunction of the large pits, but this as-
sumption requires further study to be confirmed. In
regard to the hypothesis that has been mentioned
Conclusions about the existence of some kind of microorganism,
Recapitulating the observations concerning the mor- no such proof or evidence has been found. There still
phology of glass, one can state that in all of the sam- remain many questions to be answered on the subject
ples the weathered glass inside the grooves appears of this unusual form of corrosion in archaeological
in the form of layers, perpendicular to the surface, glass. It is a type of weathering that has only just be-
which are loosely connected to the healthy glass and gun to be examined and which requires further study,
exhibit iridescence (Figure 10). so that specific and clear conclusions can be drawn.

The corroded material is easily removed and its loss

begins from the centre of the groove and moves on to
References
the walls, until the channel cuts all the way through.
BUECHNER T.S., ‘The Glass from Tarrha’, Hesperia XXIX,
From a chemical point of view it can be noted that in 109-117, plates 36-38, 1960.
the interior of the grooves sodium has been almost to- ΚΟΝΤΟΥ Ε.Κ., ΚΟΤΖΑΜΑΝΗ ∆.∆., ΛΑΜΙΠΡΟΠΟΥΛΟΣ Β.Ν.,
tally leached out, while at the same time the concen- Γυαλι: Τεχυολογια, ∆ιαβρωση και Συνιηρηση, Αθηνα, 1995
tration of silicon appears reduced. This type of corro- (KONTOU E.K., KOTZAMANI D.D. and LAMPROPOULOS
sion shows some similarities to pitting, regarding the V.N., Glass: Technology, Corrosion and Conservation,
Athens, 1995.
ΜΩΡΑΙΤΟΥ Α., ,Οργωµα: Ενα Ασυνηθιστο Φαινοµωνο σε
Αρχαια Γυαλια, IIC Ενηµερωτικο ∆ελτιο 1, Ιανουαριοσ 1996,
σελ. 14 (MORAITOU A., ‘“Ploughing”, An Unusual Phenome-
non of Ancient Glasses’, IIC Information Bulletin 1, January
1996, p.14.
ΤΡΙΑΝΤΑΦΥΛΛΙ∆ΗΣ Π., Ροδιακη υαλουπγια Ι, εκδ. Υπουργειο
Αιγαιου, Αθηνα, 2000 (TRIANTAFYLLIDIS P., Rhodian Glass-
ware I, ed. Aegean Ministry, Athens, 2000.

Vasilios Lampropoulos
Technological Educational Institute of Athens
Faculty of Graphic Arts and Design
Department of Restoration of Antiquities
Ag. Spiridonos & Pallikaridi, Egaleo, Athens, Greece

Figure 10. Vertical section of a groove. Layers of corroded Graduated from N.T.U. Athens as a chemical engi-
glass that exhibit iridescence. Glass sample from Rhodes neer in 1980. He received also a degree in Business
(x 10/0.20∞, metallographic microscope). Administration from the University of Piraeus in

299
Workshop 4

1993. He received a PhD from the Section of Geo- Athanasios Karampotsos

logical Sciences, Department of Mining Engineering
Technological Educational Institute of Athens
of N.T.U. of Athens in 1998. He has written 6 books
Faculty of Graphic Arts and Design
on the conservation of ceramics, glass, porcelain,
Department of Restoration of Antiquities
stone and the environment in museums. He has par-
Ag. Spiridonos & Pallikaridi, Egaleo, Athens, Greece
ticipated in research and conservation projects of
the Ministry of Culture and since 1988 has held the Mechanical engineer, conservator of antiquities and
position of associate professor at the Department of works of art. He has worked for the Ministry of Edu-
Conservation of Antiquities and Works of Art, Facul- cation since 1979. Since 1998 also responsible for
ty of Graphic Arts and Design of T.E.I., Athens. analytical instruments in the Department of Conser-
vation of Antiquities and Works of Art.

Antigoni Leakou*
Technological Educational Institute of Athens
Faculty of Graphic Arts and Design
Department of Restoration of Antiquities
Ag. Spiridonos & Pallikaridi, Egaleo, Athens, Greece
e-mail: [email protected]
Graduated from the Department of Conservation of
Antiquities and Works of Art, in the Faculty of Graph-
ic Arts and Design of T.E.I., Athens, in 2002 and
works as a conservator of antiquities.

300