Supplementary Material

Full text

1
Supplementary Information
Surface chromium on Terracotta Army bronze weapons is neither an
ancient anti-rust treatment nor the reason for their good preservation
Marcos Martinón-Torres1*, Xiuzhen Li2,3, Yin Xia3, Agnese Benzonelli2, Andrew
Bevan2, Shengtao Ma3, Jianhua Huang3, Liang Wang3, Desheng Lan3, Jiangwei
Liu3, Siran Liu4, Zhen Zhao3, Kun Zhao3, Thilo Rehren2,5
1Department of Archaeology, University of Cambridge, United Kingdom.
2UCL Institute of Archaeology, London, United Kingdom.
3Emperor Qin Shihuang’s Mausoleum Site Museum, Xi’an, P.R. China.
4Institute of Historical Metallurgy and Materials, University of Science and Technology
Beijing, Beijing, P. R. China.
5Science and Technology in Archaeology and Culture Research Center, The Cyprus
Institute, Nicosia, Cyprus
*Correspondence to: m.martinon-torres@arch.cam.ac.uk
This file includes:
Experimental design
Materials
Tables S1 to S6
Figure S1
Chromate conversion coating and accelerated ageing experiments

2
Experimental design
This research is part of a co-operative project concerned with the reverse engineering and
production logistics of the Terracotta Army and the broader mausoleum of Qin Shihuang.
For this paper, we analysed the typology, chemical composition and microstructure of
bronze weapons recovered in Pit 1, to characterise their materials, technology and
corrosion. We also considered the location where the artefacts were recovered within the
pit during excavations. In addition, we analysed soil samples from Pits 1 and 2, and
lacquer samples from terracotta warrior surfaces, to determine their chromium content
and characterise soil properties that could have aided metal preservation during burial. In
parallel, we carried out chromate conversion coating (CCC) experiments on bronze
tokens using different protocols for comparative purposes. We conducted accelerated
ageing experiments to compare the corrosion potential of the archaeological soil from Pit
1 compared to controls, and of CCC bronzes compared to controls, as well as to test the
possible migration of chromium to bronze surfaces under burial conditions.
Materials
We analysed a total of 464 weapons or weapon parts by pXRF, trying to cover the spatial
and typological span of the population (Table S1). Nine of these were additionally
analysed by SEM-EDS, including arrowheads, ferrules, blades and a sword fitting. In
addition, we analysed five samples of lacquer (Table S4) and five samples of soil (Table
S3) recovered during excavations in Pits 1 and 2. We also report results of analyses of
five ceramic warrior fragments (Table S5). The sampling of lacquer, soil and warrior
fragments was opportunistic, based on random samples from the ongoing excavation and
restoration works that could be made available to us. Additional sampling to address
variability and spatial patterns in these is underway, but we believe that the results
presented here suffice to support our claims.
Table S1.
Comparison between reference values for chromium and analytical results on three NIST
reference materials. All values in ppm.
Standard
Reference Cr
pXRF Cr
NIST 2702
352±22
362±6
NIST 2781
202±14
195±5
NIST 2710
23±6
46±4

3
Table S2.
Complete list of metal artefacts analysed by pXRF, classified by typology and Museum
registration number. An asterisk (*) denotes that the object or part thereof was
additionally analysed by SEM-EDS. Underlined codes denote that Cr ≥0.1% was
detected on the surface pXRF.
Arrowhead
bundles
(5-20 arrows per
bundle analysed,
probing head
and tang
separately)
Heads
1152, 1169, 1174, 1936, 1985, 1989, 1994,
2029, 2031, 2059, 2063, 2118, 2164, 2165,
2219 (3 out of 10), 2222, 2223, 2229
Tangs
1152*, 1169, 1174, 1936, 1985, 1989, 1994,
2029*, 2031, 2059 (1 out of 10), 2063, 2118*,
2164, 2165, 2219, 2222*, 2223*, 2229
Crossbow
triggers
Part A (handle)
915, 918, 926, 929, 931, 933, 954, 968, 976,
2238, 2256, 2277
Part B (tumbler)
915, 918, 926, 929, 931, 933, 954, 968, 976,
2238, 2256, 2277
Part C (rocking
lever)
915, 918, 926, 929, 931, 933, 954, 968, 976,
2238, 2256, 2277
Part D (bolt)
915, 918, 926, 929, 931, 933, 2256, 2277
Part E (bolt)
915, 918, 926, 929, 931, 933, 2256, 2277
Ferrules
1677, 1678, 1686, 2383*, 2388*, 4581
Swords and
lances
Blades
854, 857, 860, 997, 999, 2571, 3724*, 4099*
Fittings
854-1, 854-2, 854-3, 857-1, 857-2, 860, 2571,
3801*

4
Table S3.
pH and Cr content (by pXRF) of five samples of soil from Pits 1 and 2 of the Terracotta
Army site (samples 1-5), as well as organic-rich soil and Cr-spiked Terracotta Army site
soil used for ageing experiments (samples 6-7). Sample 5 was not analysed for Cr.
Sample No.
pH
±
Cr (ppm)
±
Soil1 (Pit 2)
8.3
0.03
71
5
Soil2 (Pit 2)
8.1
0.01
76
5
Soil3 (Pit 1)
8.3
0.03
76
5
Soil4 (Pit 2)
8.1
0.02
65
5
Soil5 (Pit 1)
8.5
0.02
Soil6 (organic-rich soil)
5.9
0.02
<LOD
Soil7 (Pit 1+chromite)
8.5
0.02
1997
155
Table S4.
Cr content of the five lacquer samples analysed by pXRF.
Sample No.
Cr (ppm)
±
L1 (Pit 2)
26802
1017
L2 (Pit 2)
6120
114
L3 (Pit 1)
7582
120
L4 (Pit 2)
48410
1031
L6 (Pit 2)
916
18
Table S5.
Cr content of five warrior samples analysed, showing higher values for the outer surface
(once covered with lacquer) compared to the fracture surface.
Sample No.
Location
Cr (ppm)
±
W1
outer surface
114
6
W1
fracture surface
82
6
W2
outer surface
350
8
W2
fracture surface
84
5
W3
outer surface
611
11
W3
fracture surface
92
6
W4
outer surface
575
10
W4
fracture surface
84
5
W5
outer surface
169
7
W5
fracture surface
92
6

5
Chromate conversion coating and accelerated ageing experiments
Pure metals were mixed to create 15 ingots of 40 g each of an alloy with composition
88Cu10Sn2Pb wt%. The composition was verified by pXRF. Each token was roughly
polished and cut in 5 pieces. The resulting 75 tokens of ca. 1x1 cm were polished with
600 grit (26 µm) SiC paper. At least three replicates were created under each
experimental set up.
Two different solutions were employed for chromate conversion coating (CCC)
experiments:
 Solution 1: Potassium dichromate + acetic acid (s1).
10 g of K2Cr2O7 was mixed in 100 ml of water. On an analytical scale, acetic acid
(C2H4O2) was added to the solution until obtaining pH 2.
 Solution 2: Potassium dichromate + sulphuric acid (s2).
10 g of K2Cr2O7 was mixed in 100 ml of water. On an analytical scale, sulphuric
acid (H2SO4) was added to the solution until obtaining pH 2.
The bronze tokens were treated in solutions 1 and 2 as summarised in Table S6.
Table S6.
Temperature (°C) and duration (minutes) of the CCC experiments and sample numbers
for the resulting metal tokens. For example, sample B200s2 was treated in solution 2 for
200 minutes at 80°C.
After CCC treatment, one token from each experiment was kept as reference, and two
placed in a beaker buried in soil from Pit 1 of the Terracotta Army site (pH 8.5). The
accelerated ageing was performed in an environmental chamber (Carbolite PF 60) with
controlled relative humidity (90%) and temperature (60°C) for four months.
In addition, three control bronzes that did not undergo CCC were subjected to accelerated
ageing in the same conditions, in the following media:
 the same Terracotta Army soil (pH 8.5) (Ctrl1)
 an organic-rich soil (pH 5.9) (Ctrl2)
 Terracotta Army soil spiked with ground chromite (FeCr2O4) (Ctrl3)
An assessment of the effectiveness of the different CCC treatments is beyond the scope
of this paper. Overall, CCC led to slight darkening of the surfaces, particularly those
performed with solution s2 and at higher temperatures. After accelerated ageing, all
samples buried in Terracotta Army soil (including the control Ctrl1) appeared darker in
colour but with no obvious patina or active corrosion. Conversely, the control sample
buried in low pH soil (Ctrl2) appeared noticeably corroded, with a black patina and clear
pitting of the surface (figure S1).
The surface of the metal sample token aged in chromite-enriched soil (Ctrl3) was
subsequently analysed by pXRF and SEM-EDS. No chromium was detected on the
surface.
100’
200’
500’
80°C
B100
B200
20°C
D100
D500

6
Figure S1.
Visual summary of the results of the chromate conversion coating and accelerated ageing
experiments.