WINETECH Technical Yearbook 2021

discovered in 1984 and salvaged in the 1980s and 1990s. It is difficult to know whether the wine was cargo or for private use. The wreck was exploited by treasure hunters and not archaeologically excavated; as such, very little contextual information is known. We could analyse seven samples recovered from the ship. To expand the sample pool with more examples of old wine, we have included another sample (coded M), recovered from the wreck of the SS Maori, sunk in 1909 near Llandudno. CHEMICAL ANALYSES A minimal volume of wine (less than 50 mL) was analysed using untargeted analyses, elemental analysis, sugars and organic acids. The analyses were performed at the Chemical Analytical Lab of the South African Grape and Wine Research Institute, the Mass Spectrometry Unit, Nuclear Magnetic Resonance Unit, and the ICP- MS Unit of the Central Analytical Facility of Stellenbosch University. Unexpected chemical characteristics regard- ing the metal concentrations, sugar and acid composition, as well as the aroma profile were found. For example, compared to cur- rent export regulations, the levels of heavy metals were considerably higher in most samples (figure 1). Trace element analysis

revealed high levels of As, Mn, Sr, Mo and Pb in all samples from the wrecks. This was not necessarily due to the wine itself, but possibly due to the composition of the glass bottle and leaching over time. X-ray diffrac- tion could show if the source of heavy metals were the bottles. One sample (NC) showed very high levels of Ca and Na probably due to seawater entering the bottle. When the bottles were brought to the museum for storage after salvaging, some of the recovered wines were transferred into new bottles, while others were left in the original ones. This could explain the differences in organic acids and sugars (storage in nonsterile environment) and alcohol levels (volatilisation?) (figure 2). The untargeted analyses showed various compounds including fermentation-derived products (esters, alcohols, aldehydes and fatty acids), terpenoids (linalool and fenchone) and “woody” compounds (furfural and guaiacol). The in-depth analysis of these results is still taking place. INFORMAL SENSORY EVALUATION The evaluation was done by four experi- enced researchers using free description and is based only on aroma. Incredibly, some of the samples still presented wine- like features related to aroma in both the chemical and the sensory evaluation. The

10000 9000 8000 7000 6000 5000 4000 3000 2000

OC

M

1000

M63

M68

0

B Al Mo Cd Sn Sb Ba Hg Pb Ca K Mg Na P Si ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l ug/l mg/l mg/l mg/l mg/l mg/l mg/l FIGURE 1. Elemental analysis for the wine samples recovered. V Cr Mn Fe Co Ni Cu Zn As Se Sr

60

50

40

30

RW

OC

20

NC

M68

M64

10

M63

M62

M

0 Citric Acid Tartaric Malic Acid Succinic Lac � c Acid Ace �c Acid Sucrose Glucose Fructose Glycerol

Methanol Ethanol (%)

(g/L)

Acid (g/L)

(g/L)

Acid (g/L)

(g/L)

(g/L)

(g/L)

(g/L)

(g/L)

(g/L)

(g/L)

FIGURE 2. Organic acids, sugars and alcohol composition of the samples. The missing columns indicate the compounds not detected.

WINETECH TECHNICAL YEARBOOK 2021 | 98