34358

Towards modelling of calibration models for sparkling apple wines. Part I: the unique cultivar properties behind taste.

Mabel Martínez Vega 1 ; Thomas Skov 2 ; Torben Toldam-Andersen 3
Download
Favorite this paper

INTRODUCTION
The Danish cold-climate apple juices are particularly interesting to the sparkling wine industry, because of their characteristic lower sugar values, higher acids and malic acid levels. Juice’s organic sugars and acids(taste) are routinely analyzed by winemakers. There is a global growing demand for FTIR technologies, which provide fast analysis solutions for such monitoring-routines. Together with reference, spectral data and PLS models, FTIR data can provide valuable information about juice taste. This article explores the sugars and acids profile as well as their relationships with the spectral signature of a 28 new cultivars- juice set that was used for sparkling wine elaboration. This is an exploratory study, its modelling efforts prioritize the calibration part only.
MATERIALS AND METHODS
28 mid-late-season Danish cultivars were harvested in September 2014 at the Pometum, Copenhagen_University. They were stored at 4oC in cold rooms for one month. When ripen, each cultivar was transported to room_temperature (20oC) for few hours. The fruits were milled and_then juiced in a hydro press. Immediately after, a 50 ml sample of each cultivar was collected and frozen at -10o C. The remaining juice was reserved for sparkling-wine making.
After 1-2 months the samples were taken out from the freezers and thawed-out in a fridge at 10oC for 24-hours. Total soluble solids(tss) (brixmeter: Metler Toledo ‘Quick brix 60’), total acidity(Titrino ,719 S Titrino Metrohm)(g/l); fructose, sucrose, glucose, citric and malic acid(g/l) (Ion Chromatography; Metrohm AG, Herisau, Switzerland)were measured. The mid-near infrared range(1235 - 7330cm-1), absorbance of each sample was collected from a winescan_OenoFoss™(FOSS,Hillerød,Denmark). All samples were analyzed in duplicates. Reference data and the spectrum of interest were modelled using PLS(Principal least squares). MSC and autoscale pre-processing_techniques, in the X and Y matrix respectively, were used. The models were calculated in LatentixTM software(Version 2.12. Copenhagen, Denmark). Descriptive statistics_were calculated in Excel.
RESULTS AND DISCUSSION
The total averaged tss values were 10,9% Brix and total acids of 4,6 g/l. Malic Acid ranged from 2,6-15 g/l, citric_from 0-0,26 g/l, whereas total_acids ranges were 2,1-11,3 g/l.
Only ‘Vallekildeæble’ showed typical ph values for a challenging_fermentation. Most of cultivars showed malic-acid levels over 8 g/l. Some have higher-amounts than other commercial apple cider cultivars reported from southern_Europe.

Cultivars Hillested, Bedstefars or Dronning Luise had values over 12% Brix and total-acids between 3,8 and 3,9 g/l, so higher than commercial cultivars Royal Gala and Golden Delicious, which measured at similar conditions had 11,3 and 11,6% Brix and total-acids of only 5,4 and 5,1 g/l respectively.

Most cultivars showed higher-levels of fructose than glucose and sucrose. Ørdings apple had the lowest-fructose concentration (23 g/l) and Blankholm had the highest (83 g/l). The fructose_amounts should be further investigated in- relation-to fermentation, since problems with slow fermentations due_ to_high_fructose concentrations may be expected.
PLS1&2 described 94,91% of the model variance. Scores(cultivars) were well separated by their unique_taste_attributes. The loading-plot showed that modelling of sucrose may perform better than glucose and flucose. The most_relevant_variables were: taste, tss, sucrose, acidity ph and organic-sugars. Citric_acid does not appear to-be relevant for our model