Evaluating the desorption of oxygen from wine

Sutton, Steven (2022-04)

Thesis (MScAgric)--Stellenbosch University, 2022.

Thesis

ENGLISH ABSTRACT: The removal of dissolved oxygen through desorption is commonly done in winemaking. Winemakers have indicated that under the same conditions, this process takes place at different rates, for different wines. The mass transfer of oxygen in six wines and various model wine solutions, was examined by evaluating the oxygen desorption volumetric mass transfer coefficient (𝑘𝐿 𝑎), the Sauter mean bubble diameter (D32), gas holdup (𝜀), the interfacial area (𝑎), and the oxygen mass transfer coefficient (𝑘𝐿). One of the wines used was split into two batches, with one half treated with bentonite, and the other not. A bubble column with a stone sparger was used for the experiments. The gassing out procedure and a 2nd order model was used to determine 𝑘𝐿𝑎. Bubble imaging was done to determine the interfacial area, and subsequently the 𝑘𝐿. During oxygen desorption within wine, the 𝑘𝐿𝑎 values varied between 0.0125 s-1 and 0.0275 s-1 depending on the wine. The 𝑘𝐿 𝑎 value during oxygen desorption within a 10 % ethanol solution was found to be 0.0275 s1. The addition of a small amount glycerol to this system reduced the 𝑘𝐿𝑎 to 0.0225 s-1. Further additions of organic acids did not affect the 𝑘𝐿𝑎, while the addition of protein in the form of BSA and yeast extract reduced the 𝑘𝐿𝑎 to approximately 0.0175 s-1. The 𝑘𝐿 𝑎 for during oxygen desorption within a wine that was protein unstable improved from 0.009 s1 to 0.015 s1 after being treated with bentonite. During desorption, there were no significant variations in the D32 and the interfacial area between systems containing wines or model wine solutions. Consequently, the variations between the 𝑘𝐿𝑎 values could all be ascribed to differences in the 𝑘𝐿. During oxygen desorption, the 𝑘𝐿 values were found to be between 0.015 and 0.045 mm/s within the different wines. The 𝑘𝐿 values were found to be between 0.03 and 0.04 mm/s within a 10% ethanol solution, and within the model wine solutions containing glycerol and organic acids. The 𝑘𝐿 values dropped between 0.02 and 0.03 mm/s with the addition of protein to the model wine solution. Treating a protein unstable wine with bentonite increased the 𝑘𝐿 value from 0.017 mm/s to 0.0225 mm/s. The combination of the reduction in the 𝑘𝐿 when protein was added to a model wine solution, and the improvement of the 𝑘𝐿 when wine was treated with bentonite, suggested that proteins in wine significantly affect oxygen desorption rates. It is suggested that winemakers can improve the oxygen 𝑘𝐿 within their system by operating at higher gas flowrates to increase the turbulence during desorption. However, the most effective way of improving the desorption rate is by using a sparger that produces smaller bubbles, so as to increase the interfacial area. It is suggested that desorption is performed after fining, as the 𝑘𝐿 will be greater.

AFRIKAANSE OPSOMMING: Die verwydering van opgeloste suurstof deur middel van desorpsie kom algemeen in wynbereiding voor. Wynmakers het aangedui dat, onder dieselfde toestande, hierdie proses teen verskillende tempo’s vir verskillende wyne plaasvind. Die massa-oordrag van suurstof in ses wyne en verskillende model-wynoplossings is ondersoek deur die volumetriese massa- oordragkoëffisiënt (𝑘𝐿 𝑎) van suurstof desorpsie, die Sauter gemiddelde borreldeursnee (D32), gasvertraging (gas holdup) (𝜀), die oppervlakte van die koppelvlak (𝑎) en die suurstof massa- oordragkoëffisiënt (𝑘𝐿) te evalueer. Een van die wyne wat gebruik is, is in twee verdeel, die een helfte waarvan met bentoniet behandel is en die ander nie. ’n Borrelkolom met ’n klip sprinkeltoestel (stone sparger) is vir die eksperimente gebruik. Die ontgassingsprosedure (gassing out) en ’n tweede-orde model is gebruik om 𝑘𝐿 𝑎 te bepaal. Borrel beelding (imaging) is gedoen om die oppervlakte van die koppelvlak, en gevolglik die 𝑘𝐿, te bepaal. Tydens suurstofdesorpsie in die wyn het die 𝑘𝐿 𝑎-waardes tussen 0.0125 s-1 en 0.0275 s-1 gewissel, afhangend van die wyn. Die 𝑘𝐿𝑎-waarde tydens suurstofdesorpsie met ’n 10% etanoloplossing was 0.0275 s1. Die toevoeging van ’n klein hoeveelheid gliserol aan hierdie stelsel het die 𝑘𝐿𝑎 tot 0.0225 s-1 verminder. Verdere toevoegings van organiese sure het nie die 𝑘𝐿𝑎 geaffekteer nie, terwyl die toevoeging van proteïen in die vorm van BSA en gis-ekstrak die 𝑘𝐿𝑎 tot ongeveer 0.0175 s-1 verminder het. Die 𝑘𝐿𝑎 tydens suurstofdesorpsie in ’n wyn wat proteïen-onstabiel was, het verbeter van 0.009 s1 tot 0.015 s1 ná behandeling met bentoniet. Tydens desorpsie was daar geen betekenisvolle verandering in die D32 en in die oppervlakte van die koppelvlak tussen stelsels wat wyn of model-wynoplossings bevat het nie. Gevolglik kon die verskille in die 𝑘𝐿𝑎-waardes almal aan verskille in die 𝑘𝐿 toegeskryf word. Tydens suurstofdesorpsie was die 𝑘𝐿-waardes in die verskillende wyne tussen 0.015 en 0.045 mm/s. Die 𝑘𝐿-waardes was tussen 0.03 en 0.04 mm/s in ’n 10% etanoloplossing, asook in die model- wynoplossings wat gliserol en organiese sure bevat het. Die 𝑘𝐿-waardes het met tussen 0.02 en 0.03 mm/s gedaal met die toevoeging van proteïen aan die model-wynoplossing. Die behandeling van ’n wyn wat proteïen-onstabiel was met bentoniet het die 𝑘𝐿-waarde van 0.017 mm/s tot 0.0225 mm/s verhoog. Die kombinasie van die afname van die 𝑘𝐿 toe proteïen by ’n model-wynoplossing gevoeg is en die verbetering van die 𝑘𝐿 toe die wyn met bentoniet behandel is, suggereer dat proteïene in wyn die suurstofdesorpsie-tempo’s aansienlik beïnvloed. Daar word voorgestel dat wynmakers die suurstof-𝑘𝐿 in hulle stelsel kan verbeter deur teen hoër gasvloeitempo’s te werk om die turbulensie tydens desorpsie te verhoog. Die doeltreffendste manier om die desorpsietempo te verbeter, is egter om ’n sprinkeltoestel te gebruik wat kleiner borrels produseer om sodoende die oppervlakte van die koppelvlak te vergroot. Daar word ook voorgestel dat desorpsie ná brei uitgevoer word, aangesien die 𝑘𝐿 groter sal wees.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/124943
This item appears in the following collections: