Abstract
M.Sc.
The origin of a wine plays a key role in establishing the quality and the price the
consumer is prepared to pay. Fingerprinting techniques based on multi-element data
combined with multivariate statistical analysis as well as isotope ratio data for certain
elements such as boron (11B/10B) and strontium (87Sr/86Sr) are being developed and
have been used for provenance determination of wine with varying degrees of
success.
The aim of this study was to develop a method to determine boron isotope ratios
(11B/10B) with the required precision using ICP-MS (inductively coupled plasma mass
spectrometry) in soil and wine samples and applying this method to establish the
origin of South African wines. Analytical difficulties such as the boron memory effect,
dead time, mass bias drift and matrix effects were investigated. Although the
memory effect, dead time and mass bias drift were satisfactorily resolved, it was not
possible to determine what the cause of all the observed matrix effects was during
this study.
The method was used to categorise wines from the Robertson, Swartland and
Stellenbosch regions and an attempt was made to link the measured boron isotope
ratios with that obtained from the corresponding provenance soils. The 11B/10B
isotope ratios for the wine samples (Robertson: 4.202 ± 0.014, Swartland: 4.173 ±
0.013 and Stellenbosch: 4.174 ± 0.028) were, however, higher than the ratios
obtained for the soil samples (Robertson: 4.108 ± 0.020, Swartland: 4.070 ± 0.023
and Stellenbosch: 4.124 ± 0.039). It was possible to distinguish, using the boron
isotope ratios (wine and soil samples), between the Robertson area (Breede River
region) and the Swartland area (Coastal region). The wine and soil 11B/10B isotope
ratios obtained for the Stellenbosch area (Coastal region) overlapped with the
11B/10B isotope ratios of the Robertson and Swartland regions making it impossible to
differentiate it from these two regions.