Degradation of Monumental Bronzes

Author(s): John D. Meakin, David L. Ames and Donald A. Dolske

Source: APT Bulletin: The Journal of Preservation Technology , 1991, Vol. 23, No. 4,
Historic Structures in Contemporary Atmospheres (1991), pp. 58-63
Published by: Association for Preservation Technology International (APT)

Stable URL: https://www.jstor.org/stable/1504371

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/1504371?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Association for Preservation Technology International (APT) is collaborating with JSTOR to

digitize, preserve and extend access to APT Bulletin: The Journal of Preservation Technology

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms
 Degradation of Monumental Bronzes
JOHN D. MEAKIN, DAVID L. AMES and DONALD A. DOLSKE
Corrosion of the Kitson
Corrosion studies of bronze repli-
cas of the popular Hiker statue by
T. A. R. Kitson and of bronze tablets
at Gettysburg National Military
Park are presented.
58
160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
can be viewed as multiple exposure
Hiker Statues tests of essentially the same test
specimen, with exposure times rang-
Introduction. The many replicas of
ing from three-quarters of a century
the Hiker statue, which commemor-to essentially no exposure. To gain
ate Spanish-American War veterans
insight into metal loss rates, a repli-
in various locations across the United
cation technique was developed
States, provide a unique resource for which allows laboratory study of the
assessing the impact of environmen- surface morphology resulting from
tal conditions on the degradation of corrosion. Identification of the cor-
a monumental bronze. Theo Alice rosion products remaining on the
Ruggles Kitson sculpted one version bronze surface was attempted using
of a Hiker statue (Fig. 1), which was X-ray diffraction from surface scrap-
cast by the Gorham Company of ings. The surfaces of corroded stat-
Providence, Rhode Island between ues were also compared to a statue
1921 and 1956 in more than 52 in Montgomery, Alabama, which
replicas (Panhorst 1988). has always been housed indoors.
In 1965, the final copy of the
Kitson Hiker was erected in Wash- Metallurgical analysis. Sample
ington, D.C. It has been postulated drillings were taken from the under-
that this statue was assembled from side of the gun butt on five Hikers in
the metal masters from which earlier the Boston group and from a range
Hiker castings had been made of locations on the Shamokin and
(Panhorst 1988). Fig. 2 shows the Allentown statues to test the hypoth-
geographical distribution of the esis that the group of statues was
Kitson Hikers. Following an initial metallurgically identical. Analysis
study of 25 Hikers, statues in the by X-ray emission spectroscopy re-
greater Boston area and in eastern vealed that the alloy used was ounce
Pennsylvania were selected for more metal, generally now known as
intensive study. Analyses have been leaded red brass, alloy C83600,
conducted on the composition of the which is nominally 85% copper and
alloy used, the overall features of the 5% each of tin, zinc and lead. Trace
corrosion patterns, the composition amounts of silver (-0.02%) were
of the corrosion products and the also detected in the samples. Fig. 3
morphologies of the surfaces in the shows that although there is signifi-
as-corroded state or conserved state. cant variability in the zinc content,
In this paper, attention is focused the most volatile component of the
on microscopic studies of the corro- alloy, there is no discernable trend in
sion that were carried out with a alloy composition over the time pe-
view to determining, if possible, the riod considered.
corrosion mechanisms and the rela- The tin and zinc are expected to
tion between corrosion and atmos- be in solid solution, although there
pheric pollutants. The Kitson Hikers may be some segregation of the tin
This content downloaded from
All use subject to https://about.jstor.org/terms
 BRONZE CORROSION 59

color slides has been made using

35mm and 2'/4" formats.
Samples were scraped from a
number of Kitson Hikers and sub-
jected to powder diffraction analysis
using both film and standard diffrac-
tometer techniques. Samples from
the Allentown Hiker were analyzed
-is
at the Smithsonian Institution by
4L;h. ? .-H. Daphne
.x.ir iR. Ross using diffractom-
Ii etry, and samples from the Malden,
IF is W?~aia s?~;
Chelsea, Everett and Medford
.i- : . - .... Hikers were analyzed with a film
technique. In many of the samples,
S`i'v''Mt. ...M.. .... , ";
......_=1
"F",iI....
$iN.
? ., .. .. : . .
's:N the bronze X-ray line tended to be
dominant. Further difficulties arise
?00

as the major lines for some of the

copper compounds of interest tend
p* iii . ,?
:s 41s~f~?PB~~?i~;:~%"b3
~ ~:-???%' X

to overlap.
N.....I.F....
i:M" ? ;l
As would be anticipated, the most
': i ..;:::;.;;;ll~idi ;::il'ii, ' , ::.1.,-,,,
??N. is.:
NO W.....:.:. , '.:-, , generally detected corrosion product
. ........... l , is ... . .. was brochantite, a hydrated copper
; . ......4.: sulfate, CuSO4 -3 Cu(OH)2 (Graedel
.02OF 1987). There was also evidence for
fir
cuprite, Cu20, malachite, CuCO3
r. s. a w?
. . ......... J Cu(OH)2, and antlerite, CuSO4
?e .. "" .? .:.. .. - 2 Cu(OH)2. The latter material is of
?* ".??r.
:::: .. :::: ?? ..? '"," ; particular interest as it has been pro-
:r:X
posed that in low pH conditions (less
than pH 4) antlerite should be the
favored form of copper sulfate
e~N.

-4?uc

(Matyi and Baboian 1986). In order

to draw firm conclusions about the
4?;!:WI:
X's i~ii~i i

chemistry of the corrosion products,

and its variation from statue to
statue or at different sites on the
same statue, knowledge of the struc-
ture of the surface layers from un-
affected bronze to the outermost sur-
face would be needed. Such
Fig. 1. Hiker statue in Providence, Rhode Island. information would be impossible to
derive from scrapings.

Corrosion measurements. Direct

due to dendritic formation; the lead
will be rejected as isolated spheres
(ASM 1985). The overall size scale
of the microstructure is likely to be
less than 50 gm. The quenched
outer skin of the castings should be
essentially the same within and be-
tween statues. The overall high cop-
per content common to all statues
will render their corrosion behavior
very similar.
Surface corrosion chemistry. Prior
to erection, bronze statues are usu- measurements using calipers of the
ally patinated to create a relatively dimensions of the gun barrel near
uniform surface color that can range the front sight are fairly reprodu-
from a light to a dark brown. Cor- cible. However, the corrosion rate
rosion causes regions to become es- of bronzes is sufficiently slow as to
sentially black and others to become make this an unfruitful approach to
various shades of green (Lins 1985). comparing differences in corrosion
between Hikers. In an outdoor envi-
Streaking occurs widely, and corro-
sion pitting, resulting in a splotched ronment, the general rate of loss of
or speckled appearance, is also com- bronze over time is reported to be of
monly observed. A collection of the order of 1-3 1m/year (Mattsson

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms
60 APT BULLETIN

i2
J T- i I -

1 ......-f" t 0,
I Is

', I
( ,. i I ; 4 .,." ,-' .- _ /
7 7

",i i i L l - -,...-, ;r
\ - i t. i ir " ..v t !G ETRBSO
~ 3

"- !
Fig. 2. Distribution of Hiker statues. .... -, t : r I ..
! ? i . . ...

,i- i_. i :_ ..?. . .

atesresh n?heaonsv
Fig.2. istrbuton f Hiersatue. ! L c~

Dates are shown in hexagons.

will yield data of the required preci- peeled from the surface. Areas of
sion. In contrast, the surface pitting about 20 cm2 have been routinely
is occurring at rates up to 5 gLm/year,replicated on more than half of the
and relocation of precisely the same Hikers. Measurements were made
7::
pit at a later date is perfectly feasibleon a trial replica to establish the
using the technique described below.degree of dimensional stability.
Microscopic and surface charac- Over about a month the change in
terization techniques are much moredimension was + 0.002" in 0.58" or
readily carried out in the laboratory < +0.5%. A check after almost a
than at outdoor Hiker sites. Some year showed that the moulding ma-
Statue

photography can be carried out in terial had remained completely sta-

the field but the enlargement that ble.
can be achieved is very limited. Ac- Replicas were taken on the rifle
Fig. 3. Composition of five Hikers cordingly, a replica technique was
in Massa- barrel immediately under the front
chusetts and two Hikers in Allentown and
developed and applied to selected sight. This position is free of com-
Shamokin, Pennsylvania. Note that the scale features on a number of statues.
plex runoff effects and over a rel-
for copper is 1/10 of actual.
The replicating material selected is a atively short distance (about 3 cm)
vinyl polysiloxane, widely used for includes surfaces that vary from sky-
dental purposes and available from ward to ground-facing. A qualita-
and Holm 1982). Thus over approx- dental supply houses. The two com- tive assessment of the surface rough-
imately 20 years the total overall loss ponents are mixed by hand to yield a ness and pitting can be obtained by
will be about 50 pm or 0.002". As putty-like material which can be microscopic examination of the
there are no precisely parallel fea- pressed firmly against the surface to replica surface. Optical microscopy
tures on the statues because of hand be replicated. At temperatures is useable but is limited by the very
finishes, it is unlikely that repeated above about 500F., the material sets small depth of field. More flexible
measurements on one statue, or up to a firm, but slightly elastic solid microscopy, using magnifications be-
cross-comparisons between statues, in about 15 minutes, when it can be tween 10x and 1000x, can be carried

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms
 BRONZE CORROSION 61

3:? ting. For most of its exposure it has

N t! . ; ...N!.. ...
been coated, first, by the foundry
.. W N-"
and then by routine waxing by Na-
tional Park Service conservators be-
4 . ? ... . ,.... ..
ginning in 1985. The Lynn and
Chelsea statues are very heavily pit-
.:94i
!.. f::
, .. .
N:... ..?....,,
ted, and the traces show pit depths
up to 0.3 mm. The replica of the
.....:..... "
Fitchburg statue was taken after con-
servation with walnut shells air abra-
f? . . fB:'i .. ",
sive cleaning, repatination, and wax
coating. The pitting is still quite evi-
dent, but there appears to have been
some reduction (rounding) of the
fine structuring of the surface. Prob-
A A ably the
-. ..most valuable
.,. aspect of the.: .
: ! . ? . N...
a";~*a
. i;. .,.0 . . W
..0:. :,v . . replication procedure is the ability to
a ft , T monitor the
". . change in. surface condi-
. .
I-NEr RFAT:7~" tion over a period of time. Repeated
. .. .. . . .,:. . ...... .. replicas
... could be stored, and the
t:? " , ,Mr growth of selected pits, or the initia-
; ., , ..%. .,. 4",
tion of new pits, monitored. Repli-
cation also provides a quantitative
measure of the morphological
changes caused by any conservation
Fig. 4. Scanningtreatment.
electron

Bronze Runoff Measurements at

out using scanning
Gettysburg National Militaryele
Park
scopy (SEM).
Montgomery The rep
lightly gold coated
Introduction. The Gettysburg Na-b
tion. Fig. Washlngton
4 shows
tional Military Park is located th
in a
the replicas from the
rural area of Pennsylvania although
of the gun there is substantial vehicular
barrel imm traffic
the front
Lynn associated with more than 1.5 mil-
sight.
The contrast between
lion visitors per year. In about
tially un-pitted Wash
1906, a series of bronze brigade
and the markers were erected
heavily at more than
pitted
and Chelsea Hikers
60 locations throughout the Park. i
apparent. Although
These markers are cast bronze
sible to use
S.1 mm Chelsea these mic
plaques with raised letters and are
measure the fraction
mounted at 300 to the vertical on
1 mm o
area that is pitted,
granite supports. The markers were su
information Fitchburg on the
cast by Bureau Brothers of Philadel- d
From a degradation phia and, according to a letter to the
latter is an important
War Office dated October 6th,
age. Using the
1906, were tosame
be made of "Govern-
Fig. 5. Profilometer traces for replicas from
face profilometer
various statues
ment Standard Bronze Metal ofhas
the
develop a quantitativ best quality" (National Archives
the elevation1906).along Qualitative EDS analysis in- a
along the replica (sta
dicates a copper alloy containing
Fig. 5 detritus collected from the inside air. small amounts of
presents trace
tin, zinc and lead,
range The
of Washington statue has not been suggesting that these markers,
surface deg like
Montgomery
exposed for a sufficient period of statue
the Hikers, are ounce metal, nomi-
small asperities,
time to develop any significant pit- prob
nally 5% zinc, tin and lead in copper.

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms
62 APT BULLETIN

Bronze runoff collection. A collec-
tion system was designed to sample
the incident rain and the monument
runoff solution for a series of indi-
vidual rain events. Measurements of
about 40 storms were made during
the period 1986-88. The runoff
all marker runoff samples. The least
square fit shown in Fig. 7 has a slope
of 0.05 and a correlation coefficient,
R2, of 0.45, excluding nine extreme
outlying samples, probably repre-
6
senting defective analyses. These
data indicate that dissolution of the
bronze plaque occurs without prefer-
ential removal of zinc from the cop-
per base. Further, the data agree

samples were analyzed for pH and

for copper and zinc content, as well
as for the concentration of ions typi- 5

cally found in rain - sulfate, nitrate, 0 o

chloride, ammonium, magnesium,

potassium, and calcium. 4 0

Three markers were selected for

E [E
measurements and were fitted with a a-

channel device that fed into an acid- oo

33
S40 0
leached polyethylene sample bottle
(Fig. 6). The size of the sample bot- 2
c

tle, 500 ml, is such that a relatively

small rainfall (about 0.1 mm or
0.04 ") will fill the bottle. As a result
most of the runoff samples provide
an averaged measure of the pH and
ion content of the solution leaving 0
0 20 40 60 80 100
the tablet. It would be expected that
the first runoff, particularly follow-
ing a prolonged dry period, will be Copper pp
particularly acidic and rich in
leached species.
Fig. 7. Zinc vs. copp
Analysis. Initial correlations have
been tested for zinc versus copper for
100

80

60
E 0
oo

S40 0

20 1

20

96 data points

0 50 100 150

Sulfate

Fig. 6. Example brigade marker used in

runoff study Fig. 8. Copper v

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms
 BRONZE CORROSION 63

References
100

American Society of Metals. 1985.

Metals Handbook, vol. 9, 9th ed.,
637-45.
loO- 80
Graedel T.E., K. Nassau and J.P. Franey.
1987. Copper patinas formed in the
atmosphere. Corrosion Science
27(1):639-57.
E 60 Lins, Andrew. 1985. Outdoor bronzes:
some basic metallurgical considera-
tions. In Sculptural monuments in an
outdoor environment, ed. V.N.
0 40 - 0 Naude, 8-20. Philadelphia: Pennsyl-
O o 0ho E vania Academy of the Fine Arts.
Mattsson, E., and R. Holm. 1982.
Atmospheric corrosion of copper and
20 1 [ E its alloys. Atmospheric Corrosion, ed.
B3f [
00 %
W.H. Ailor, 365-82. New York:
Wiley-Interscience.
Matyi, R.J., and R. Baboian. 1986.
0 13 0 13
An X-ray diffraction analysis of the
0 20 40 60 80 100 patina of the Statue of Liberty. Pow-
der Diffraction 1:299-304.
National Archives and Records Adminis-
Nitrate ppm
tration. 1906. Letter from Gettysburg
Park Commission, October 16, 1906,
to the Secretary of War in 7393/46.
Fig. 9. Copper vs. nitrate
Panhorst, Michael inW. 1988. runof
The popu-
larity, patronage, production and
placement of monumental bronze
statue replicas of the Hiker. In Ames,
Conclusion et al., Project Report to the National
with a 5% zinc content
Park Service. for t
Fig. 8 shows an excellent d
correlation The effects of environmental
between expo- the cop
sulfate ion concentrations. The least sure on the corrosion of monumental Acknowledgements
square fit line shown has a slope of bronzes has been studied using the
0.62 and a correlation coefficient, This paper is an abbreviated version of
series of Hiker statues. Substantially
an article published in Atmospheric En-
R2, of 0.90. If the copper is dis- different degrees of attack have been
vironment B, 1992, and the permission
solved by ambient rain in the stoich- recorded. A molding technique has by the editors to publish in the APT Bul-
iometric ratio for cupric sulfate, been developed that has been used to letin is acknowledged. The research has
CuSO4, then the anticipated slope compare quantitatively the corrosion been supported by the National Park
would be 0.66. Thus, these results on different statues. The technique Service under Cooperative Agreements
CA 0424-6-8003 and CA 0424-6-8001.
strongly suggest that the amount of could also be used to monitor corro-
The support and encouragement of
available sulfate determines the sion over an extended period of time Susan I. Sherwood of the National Park
bronze dissolution rate. The corre- and to measure the effect of conser-
Service is gratefully acknowledged.
sponding plot of copper against NO3 vation treatments. Runoff studies
(Fig. 9) shows a much poorer corre- on bronze tablets at the Gettysburg
lation than Fig. 8. The least square National Military Park indicate that JOHN D. MEAKIN is Professor and
fit line has a slope of 0.82, a signifi- the dissolution rate of bronze by rain Chairman, Department of Mechanical
cant intercept of 11 ppm copper and is controlled by the availability of Engineering, University of Delaware.
a correlation coefficient, R2, of only sulfate ions.
DAVID L. AMES is with the Center for
0.5. The slope of 0.82 is very differ-
Historic Architecture and Engineering,
ent from 0.51, which is what is an- University of Delaware.
ticipated if Cu(N03)2 is the species
controlling the dissolution. This fur- DONALD A. DOLSKE is Professional
ther supports the interpretation Scientist, Office of Air Quality, Illinois
above that sulfate ion is the control- State Water Survey.
ling influence for dissolution of the
bronze surface.

This content downloaded from

160.75.86.201 on Wed, 23 Dec 2020 19:57:40 UTC
All use subject to https://about.jstor.org/terms