Flavor Release - American Chemical Society


Flavor Release - American Chemical Societyhttps://pubs.acs.org/doi/pdf/10.1021/bk-2000-0763.ch028Similarflavor release o...

0 downloads 28 Views 1MB Size

Chapter 28

Influence of Formulation and Structure of an Oil-in-Water Emulsion on Flavor Release 1

1

1

Marielle Charles , SandrineLambert ,Philippe Brondeur , Jean-LucCourthaudon ,and Elisabeth Guichard 2

1

Downloaded by FUDAN UNIV on January 19, 2017 | http://pubs.acs.org Publication Date: September 7, 2000 | doi: 10.1021/bk-2000-0763.ch028

2

1

INRA-LRSA, 17 rue Sully, 21034 Dijon Cedex, France ENSBANA, 1 Esplanade Erasme, 21000 Dijon, France

The influence of proteins, polysaccharides, and droplet size on flavor release of an oil-in-water emulsion was determined using aroma compounds with different hydrophobic characteristics and static headspace analysis. Flavor release of lipophilic compounds (ethyl hexanoate and allyl isothiocyanate) was influenced by the three factors. Using protein as an emulsifier induces a decrease of flavor release of aroma compounds which interact with the protein. Emulsions with small fat droplets lead to a better release. The influence of polysaccharides depends on the fat droplet size: salting-out effect with large droplets and retention with small droplets. For hydrophilic compounds, no effect of the nature of the protein was noticed and the influence of the droplet size and polysaccharides was a function of the aroma compounds.

The food industry must constantly adapt its products to the consumer's demand. In order to minimize the number of experiments needed to study the effects of formulation or process, it is necessary to better understand the different phenomena responsible for flavor release. Salad dressings are a good example of a complex food in which flavor release is governed by several factors. Salad dressings are oil-in-water emulsions stabilized by both surface active agents and thickeners. In this type of product, flavor release should depend on: (1.) the oil content which affects the partition of aroma compounds between the different emulsion phases (lipid, aqueous and vapor) (i), (2.) the surface active agent which interacts with aroma compounds at the interface or in the bulk phase when present in excess (2, 5), or limits the transfer of aroma compounds between the oil and the aqueous phases (4), (3.) the polysaccharides which interact with aroma compounds (5), entrap or reduce their mass transport in the bulk phase (6, 7), (4.) the structure of emulsion (8) or emulsion type (9).

342

© 2000 American Chemical Society

Roberts and Taylor; Flavor Release ACS Symposium Series; American Chemical Society: Washington, DC, 2000.

343

Downloaded by FUDAN UNIV on January 19, 2017 | http://pubs.acs.org Publication Date: September 7, 2000 | doi: 10.1021/bk-2000-0763.ch028

Previous results on salad dressings (50% oil) homogenized with whey protein showed that sensory perception and flavor release were affected by the modification of the droplet size: some hydrophobic compounds were better released from the finest emulsion whereas hydrophilic compounds were less released (10). But decrease of droplet size was associated with an increase of viscosity and a decrease of protein concentration at the interface. It was thus difficult to explain the differences in flavor release only by the droplet size. This chapter deals with a fundamental study performed on model emulsions in order to determine the respective influence of proteins, polysaccharides, and fat droplet size on flavor release in dressings. Effect of proteins and polysaccharides were also determined in aqueous solutions.

Materials and Methods Aroma Compounds Five aroma compounds present in dressings and with different hydrophobicity constants were chosen (Table I). They were kindly supplied by International Flavours and Fragances (I.F.F., Longvic, France).

Table I. Physico-chemical and Sensory Characteristics of Aromas Aroma compounds

Diacetyl Butan-l-ol 2-Methylbutan-l-ol A l l y l isothiocyanate Ethyl hexanoate a

Formula

C42H602 C4H10O C5H120 C4H5NS C8H12Q2

Mr

86 74 88 99 114

Solubility in water (g/L)

250 (15°C) 111(25°C) 1.90 (25°C) 0.597 (25°C)

Odor

3

LogP

-2.0 0.8 1.2 1.7 2.8

Buttery Fusel oil Fusel oil Pungent Fruity

Hydrophobicity constant calculated according to Rekker's method (77)

Model Emulsions Model emulsions were prepared to test the effect of proteins, polysaccharides and droplet size. The fat content (30%) was the same for all the emulsions. Proteins Three protein concentrates were tested as emulsifiers: oc-lactalbumin (PSDI 4200, MD-Food, Videbaek, Danemark), β-lactoglobulin (Besnier Bridel Aliments, Chateaulin, France), and a mixture of β-lactoglobulin and cc-lactalbumin (73:23, w:w). Proteins were solubilized in a citric acid / sodium citrate buffer (0.1 M ; NaCl 25 m M , p H 3) at a concentration determined in order to have 0.5% of protein in the

Roberts and Taylor; Flavor Release ACS Symposium Series; American Chemical Society: Washington, DC, 2000.

344

Downloaded by FUDAN UNIV on January 19, 2017 | http://pubs.acs.org Publication Date: September 7, 2000 | doi: 10.1021/bk-2000-0763.ch028

final emulsions. Flavor compounds (40 mg), protein solution (31.5 g) and sunflower oil (13.5 g) were equilibrated during l h before homogenization with an Ultra-Turax at 8000 rpm for 3 min at 5°C. Polysaccharides Two polysaccharides were used as thickeners : xanthan (Kelco International) and a modified starch (waxy corn chemically modified, National Starch, France). Four emulsions (two emulsification rates and with or without polysaccharides) were prepared. cc-Lactalbumin was used as an emulsifier at 0.5% (w:w) and emulsions with polysaccharides contained 0.2% xanthan and 0.7% modified starch. Aqueous solutions were prepared in NaCl 25mM, p H 6 and heated at 85° for 30 min. Two emulsification rates were applied (10000 rpm or 20000 rpm) for 1 min with a Polytron at 5°C. Emulsions were flavored with a mixture of three aroma compounds. Ethyl hexanoate was solubilized in the lipid phase before emulsification and 2methylbutan-l-ol and diacetyl were added to the emulsions with a syringe, in order to obtain concentrations in the final emulsion of respectively 500 ppm, 500 ppm, and 200 ppm. Emulsions were gently stirred l h before analysis. Droplet Size Two emulsions differing by their droplet size (volume surface average diameter) and their α-lactalbumin content were made: 7 pm and 0.4% protein; 15 pm and 0.2% protein. Solubilization of the protein was done in citrate buffer prior to emulsification. After emulsification with a Polytron, emulsions were diluted by a factor of 1.2 with the buffer for the fine emulsion and with a 1% α-lactalbumin solution for the coarse emulsion. The final emulsions have thus a same total protein content of 0.3%. Aroma compounds alone were added to the final emulsions to obtain a final concentration of 200 to 500 ppm. Emulsions were gently stirred for 1 h, (2 h for allyl isothiocyanate) before headspace analysis.

Polysaccharide Solutions Polysaccharide solutions (xanthan 0.2%, modified starch 0.7%, and a mixture of xanthan-modified starch 0.2%-0.7%) were prepared by solubilization in citric acid / sodium citrate buffer (0.1 M , NaCl 25 m M , pH3) and heating at 85°C for 30 min.

Emulsion Characterization Particle size distribution was determined with a Malvern Mastersizer laser diffractometer (Model S2-01; Malvern Instruments, Orsay, France) both before and after headspace measurement to verify the stability of the emulsions. The volume surface average diameter d and the specific surface area were measured. 3 2

Protein Content Determination Protein concentration of the aqueous phase was measured using the Kjeldahl

Roberts and Taylor; Flavor Release ACS Symposium Series; American Chemical Society: Washington, DC, 2000.

345 method (12). The amount of protein adsorbed at the interface was calculated from the difference of protein concentration between that in the solution used for making the emulsion and the one measured in the aqueous phase after emulsion centrifugation.

Downloaded by FUDAN UNIV on January 19, 2017 | http://pubs.acs.org Publication Date: September 7, 2000 | doi: 10.1021/bk-2000-0763.ch028

Measurement of Interactions between Proteins and Aroma by Affinity Chromatography The procedure used for the immobilization of the proteins was adapted from that described by Sostmann and Guichard (13): α-lactalbumin, β-lactoglobulin or the mixture of the two proteins were immobilized onto a silica-diol support in P E E K (PolyEterEtherKetone) columns (4.3 mm χ 5 cm). One column without protein was prepared in the same conditions. The H P L C system used was a Varian 9010 pump and a Shimadzu SPD-6AV UV-vis spectrometric detector. The system was equilibrated with eluent (water, NaCl 25 m M , p H 3) at a flow rate of 1 mL.min" . Flavor compounds were dispersed in the same aqueous solution, injected and detected at their maximum of absorption. Global affinity was calculated as follows (14): 1

K = (t —i)IC t where K is the global affinity ( M ) , t the retention time of 1

b

R

Q

b

R

compound on column with protein (min), t the retention time of compound on column without protein (min), C the protein concentration (mol.L ) and to the void time (min). 1

p

Headspace Analysis Amber flasks (40 mL) were filled with 10 ml solutions (emulsions, or 5 m L polysaccharide solutions + 5 mL aroma solution) closed with a mininert™ valve (Supelco, St Quentin Fallavier, France) and placed in a 30°C water-bath with stirring. Vapor-liquid equilibrium analysis was used to measure the influence of proteins (24 h) and polysaccharides (2 h). To test the influence of the droplet size, flavor release was determined as a function of time. For each study, 1 mL of vapor phase above the emulsion was injected into a chromatograph Carlo Erba GC8000 series (Fisons Instruments) equipped with a DB-Wax fused capillary column (30 m x 0.32 mm i.d., film thickness 5 pm) (J&W Scientific Inc., Folsom, USA). The operating conditions were as follows: hydrogen carrier gas velocity, 50 cm.s" ; H flow rate, 23 mL.min" ; air flow rate, 318 mL.min" ; FID injector temperature, 240°C; detector temperature, 250°C; oven temperature 120°C, 140°C or 160°C depending on the volatile compounds. 1

1

2

1

Statistical Analysis The statistical analyses were realized with the Statistical Analysis Systems software (SAS Institute, Inc., Cary, NC). A one-way analysis of variance (proc G L M ) followed by comparison of means by duncan test was applied except for results on effect of polysaccharides in aqueous solutions (t-test).

Roberts and Taylor; Flavor Release ACS Symposium Series; American Chemical Society: Washington, DC, 2000.

346

Results

Influence of the Proteins Whey proteins are often used to emulsify dressings. Flavor release from emulsions prepared with two whey proteins, α-lactalbumin and β-lactoglobulin were compared. A mixture of these two proteins in proportions found in a commercial emulsifier was also tested (Figure 1). Average diameters (d ) of emulsions were 15.3 pm for α-lactalbumin emulsions, 10.5 pm for β-lactoglobulin emulsions and 13.2 pm for mixture emulsions. Flavor release of ethyl hexanoate and allyl isothiocyanate from emulsions with β-lactoglobulin was significantly smaller (p