Problem and objective of the study
Given the significant development of FTIR instruments, we wanted to investigate the possibility of using our TITRIVINs to calibrate these instruments.
We therefore set up a study whose main objective was to verify the feasibility of using TITRIVINs as calibration standards for FOSS brand FTIR instruments.
As the Oenocentres of the Chambre d'Agriculture de la Gironde group own FT120 (old generation - to be phased out) and WineScan Auto devices, we were able to carry out a study on 5 different devices for each model.
The tests were also carried out on an OenoFoss type device. As the results concern only one device, they will not be presented in the rest of this page.
Study principle
To do this, we verified the performance of a calibration carried out with TITRIVIN by studying the results obtained by FTIR devices on a population of wines selected to be representative of the Bordeaux region, whose main parameter values were established by an inter-laboratory comparison between 4 accredited laboratories.
Study set-up
Selection of the 30 wines representative of the Bordeaux region
The number of wines was set at 30, spread over 3 colors (white, red and rosé), and over the last three vintages (2012, 2011 and 2010).
The choice of these wines had to be able to cover, as far as possible, for each selected parameter, the field of application of the FTIR devices and the range covered by the TITRIVIN, i.e.:
- Alcohol Volumetric Title: 10 to 15%vol.
- Glucose+Fructose: 0.8 to 10 g/L.
- Total acidity: 2 to 5 g/L H2SO4.
- Volatile acidity: 0.2 to 1 g/L H2SO4.
- Malic acid: 0.2 to 4 g/L
Defining reference values for the 30 selected wines
To define highly reliable reference values for these 30 wines, we organized an inter-laboratory comparison between the 7 Oenocentres of the Gironde Chamber of Agriculture group. For each parameter, the Oenocentres carried out analyses in duplicate using reference or routine methods under accreditation.
Analysis of the 30 wines and TITRIVIN using FTIR instruments
All TITRIVIN and the 30 wines were analyzed in duplicate using the 5 WineScan Auto FTIR instruments in our Oenocentres. A control wine was also analyzed at regular intervals to check the stability of the instrument during the analysis session. As we had samples that had been used in inter-laboratory comparisons organized by the Œnologues de France, these were also analyzed.
Processing the results
For each instrument, we established a linear regression between the raw FTIR results and the TITRIVIN reference values. In this way, we were able to define the slope and bias corresponding to a calibration performed using TITRIVIN.
For ease of understanding, we chose to present the results obtained in the form of residual deviation. Thus, on the graphs of results that we propose, each point corresponds to the residual deviation between the value obtained with the TITRIVIN calibration and the reference value defined during the inter-laboratory comparison.
From all these residual deviations, we have been able to calculate the uncertainty obtained on the measurement of the various parameters studied when the calibration of the devices is carried out using the TITRIVINs.
To gain an appreciation of the performance obtained, we compared this uncertainty with the expanded uncertainty (with k=2) of interlaboratory reproducibility published by the Œnologues de France based on a study of the results of their analysis chain.
Results obtained on the WineScan Auto
For TAV
As the FT120 instruments are no longer renewed, we have chosen to present the results obtained only on the WineScan Auto model.
The following graph shows the results obtained by the 5 WineScan Auto FTIR instruments tested during the study.
The expanded uncertainty (with k=2) obtained with the calibration performed using TITRIVIN is 0.11% vol.
By way of comparison, the expanded uncertainty (with k=2) of interlaboratory reproducibility published by UOEF for the determination of TAV by all methods except FTIR is 0.14% vol.
The calibration was carried out with the TITRIVIN batches currently on the market:
AA2 lot A 04111216 2
AA3 lot A 04111216 3
AA4 lot A 04101215 4
AA5 lot A 03121217 5
BTA lot B 04101215 A
BTB lot B 03121217 T
For Total Acidity
For this parameter, the expanded uncertainty (with k=2) obtained is 0.109 g/L H2SO4.
By way of comparison, the expanded uncertainty (with k=2) of interlaboratory reproducibility published by UOEF for the determination of Total Acidity by all methods except FTIR is 5.4%.
Calibration was performed with TITRIVIN batches currently on the market:
AA1 lot A 03121217 1
AA2 lot A 04111216 2
AA5 lot A 03121217 5
BTA lot B 04101215 A
BTB lot B 03121217 T
For the other parameters
The parameters Glucose+Fructose, Malic Acid and pH were also evaluated in the same way. But in these 3 cases, the performances obtained did not allow us to obtain an uncertainty better than the interlaboratory reproducibility uncertainty.
This study was carried out in collaboration with the Oenocentres of Soussac, St Savin, Pauillac, Cadillac, Grézillac and Coutras, and Madame Evelyne Chanson (EC consulting). Tests on the OenoFoss were carried out on the Wine and Tools device.
Data processing was carried out by Vianney Bellegrade as part of his internship for the "Métiers de la Vigne et du Vin - option Oenocontrôleur" professional degree.
.