Flavor Release Research Articles

Understanding the Link between Bolus Properties and Perceived Texture

AbstractThe relationships between food structure, texture perception and bolus properties during bolus formation are poorly understood. This preliminary single‐subject study on three biscuit recipes investigated the links between bolus macrostructure, perceived texture (using temporal dominance of sensations), and mechanical and rheological properties (using back extrusion and consecutive double compression) throughout mastication. The fracture properties of the biscuits dominated textural perception in the early stages of mastication. As mastication progressed, a sensory transition from dry to sticky was accompanied by an increase in “cohesiveness” and decreases in the consistency coefficient, peak force and “adhesiveness.” At the point of swallow, the boluses were shear thinning pastes and sticky dominated perception. Consistency coefficient and “cohesiveness” indicated different swallowing thresholds for the different recipes. This study found back extrusion to be a suitable method for characterizing biscuit bolus rheology in the later stages of mastication as it is unaffected by bolus dimensions.Practical ApplicationsMastication of a solid food involves the breakdown of the food and reassembly with saliva into a deformable bolus that can be swallowed safely. The relationships between the perceived texture and structural, mechanical and rheological properties throughout mastication are not well understood. An understanding of these relationships could aid in the design of foods that can be swallowed safely by the elderly or those with dysphagia, or foods that follow a desired oral processing path and thus have a desired texture profile or flavor release.

Comparison of direct mass spectrometry methods for the on‐line analysis of volatile compounds in foods

For the on-line monitoring of flavour compound release, atmospheric pressure chemical ionization (APCI) and proton transfer reaction (PTR) combined to mass spectrometry (MS) are the most often used ionization technologies. APCI-MS was questioned for the quantification of volatiles in complex mixtures, but direct comparisons of APCI and PTR techniques applied on the same samples remain scarce. The aim of this work was to compare the potentialities of both techniques for the study of in vitro and in vivo flavour release. Aroma release from flavoured aqueous solutions (in vitro measurements in Teflon bags and glass vials) or flavoured candies (in vivo measurements on six panellists) was studied using APCI- and PTR-MS. Very similar results were obtained with both techniques. Their sensitivities, expressed as limit of detection of 2,5-dimethylpyrazine, were found equivalent at 12 ng/l air. Analyses of Teflon bag headspace revealed a poor repeatability and important ionization competitions with both APCI- and PTR-MS, particularly between an ester and a secondary alcohol. These phenomena were attributed to dependency on moisture content, gas/liquid volume ratio, proton affinities and product ion distribution, together with inherent drawbacks of Teflon bags (adsorption, condensation of water and polar molecules). Concerning the analyses of vial headspace and in vivo analyses, similar results were obtained with both techniques, revealing no competition phenomena. This study highlighted the equivalent performances of APCI-MS and PTR-MS for in vitro and in vivo flavour release investigations and provided useful data on the problematic use of sample bags for headspace analyses.

Fat and salt contents affect the in-mouth temporal sodium release and saltiness perception of chicken sausages

In cooked meats, sodium chloride is involved in taste, texture and flavour release. So a reduction in the salt content may have an impact on overall perception and acceptability. The aim of this study was to evaluate the influence of composition on sodium release and saltiness intensity in chicken sausages. The rheological properties of the sausages differed according to composition. Temporal sodium release and temporal saltiness intensity were evaluated by four selected subjects when eating sausages. At each time point, the effect of the salt level in sausages on sodium release was positive and highly significant. The effect of lipids on sodium release was negative. Concerning perception, the amount of salt used had a positive effect on saltiness intensity, and lipids seemed to exert a masking effect. Generally, clear relationships between salt levels, sodium release and saltiness intensity were found but the masking effect of lipids on saltiness intensity probably also involved texture or fat perception mechanisms.

Influence of heating and acidification on the flavor of whey protein isolate

Previous studies have established that whey protein manufacture unit operations influence the flavor of dried whey proteins. Additionally, manufacturers generally instantize whey protein isolate (WPI; ≥90% protein) by agglomeration with lecithin to increase solubility and wettability. Whey protein isolate is often subjected to additional postprocessing steps in beverage manufacturing, including acidification and heat treatment. These postprocessing treatments may further influence formation or release of flavors. The objective of the first study was to characterize the effect of 2 processing steps inherent to manufacturing of acidic protein beverages (acidification and heat treatment) on the flavor of non-instant WPI. The second study sought to determine the effect of lecithin agglomeration, a common form of instantized (INST) WPI used in beverage manufacturing, on the flavor of WPI after acidification and heat treatment. In the first experiment, commercial non-instantized (NI) WPI were rehydrated and evaluated as is (control); acidified to pH 3.2; heated to 85°C for 5min in a benchtop high temperature, short time (HTST) pasteurizer; or acidified to 3.2 and heated to 85°C for 30s (AH-HTST). In the second experiment, INST and NI commercial WPI were subsequently evaluated as control, acidified, heated, or AH-HTST. All samples were evaluated by descriptive sensory analysis, solid-phase microextraction (SPME), and gas chromatography-mass spectrometry. Acidification of NI WPI produced higher concentrations of dimethyl disulfide (DMDS) and sensory detection of potato/brothy flavors, whereas heating increased cooked/sulfur flavors. Acidification and heating increased cardboard, potato/brothy, and malty flavors and produced higher concentrations of aldehydes, ketones, and sulfur compounds. Differences between INST and NI WPI existed before treatment; INST WPI displayed cucumber flavors not present in NI WPI. After acidification, INST WPI were distinguished by higher intensity of cucumber flavor and higher concentrations of E-2-nonenal. No perceivable differences were observed between INST and NI WPI after heating; sulfur and eggy flavors increased in both types of WPI. After treatment, AH-INST-HTST samples were differentiated from AH-NI-HTST by grassy/hay and grainy flavor and increased lipid oxidation products. Further processing of WPI in food applications has negative effects on the flavor contributions of WPI.

The flavor release of agarose gel

The flavor release of agarose gel

In VitroandIn VivoFlavor Release from Intact and Fresh‐Cut Apple in Relation with Genetic, Textural, and Physicochemical Parameters

Flavor release from 6 commercial apple cultivars (Fuji, Granny Smith, Golden Delicious, Jonagold, Morgen Dallago, and Red Delicious) under static conditions (intact or fresh-cut samples) and during consumption of fresh-cut samples (nosespace) was determined by proton transfer reaction mass spectrometry. Textural (firmness, fracturability, flesh elasticity, and rupture) and physicochemical (pH, acidity, and water content) properties of the apples were also measured. Static headspace analysis of intact fruits revealed Fuji and Granny Smith apples had the lowest concentration for all measured flavor compounds (esters, aldehydes, alcohols, and terpenes), whereas Red Delicious apples had the highest. Fresh-cut samples generally showed a significant increase in total volatile compounds with acetaldehyde being most abundant. However, compared to intact fruits, cut Golden and Red Delicious apples had a lower intensity for ester related peaks. Five parameters were extracted from the nosespace data of peaks related to esters (m/z 43, 61), acetaldehyde (m/z 45), and ethanol (m/z 47): 2 associated with mastication (duration of mastication-t(con); time required for first swallowing event-t(swal)), and 3 related with in-nose volatile compound concentration (area under the curve-AUC; maximum intensity-I(max); time for achieving I(max)-t(max)). Three different behaviors were identified in the nosespace data: a) firm samples with low AUC and t(swal) values (Granny Smith, Fuji), b) mealy samples with high AUC, I(max), t(swal) values, and low t(con) (Morgen Dallago, Golden Delicious), and c) firm samples with high AUC and I(max) values (Red Delicious). Strengths and limitations of the methodology are discussed. Volatile compounds play a fundamental role in the perceived quality of food. Using apple cultivars, this research showed that in vivo proton transfer reaction mass spectrometry (PTR-MS) could be used to determine the relationship between the release of volatile flavor compounds and the physicochemical parameters of a real food matrix. This finding suggests that in vivo PTR-MS coupled with traditional physicochemical measurements could be used to yield information on flavor release from a wide range of food matrices and help in the development of strategies to enhance food flavor and quality.

Evaluation of Coffee Roasting Degree by Using Electronic Nose and Artificial Neural Network for Off‐line Quality Control

An electronic nose (EN) based on an array of 10 metal oxide semiconductor sensors was used, jointly with an artificial neural network (ANN), to predict coffee roasting degree. The flavor release evolution and the main physicochemical modifications (weight loss, density, moisture content, and surface color: L*, a*), during the roasting process of coffee, were monitored at different cooking times (0, 6, 8, 10, 14, 19 min). Principal component analysis (PCA) was used to reduce the dimensionality of sensors data set (600 values per sensor). The selected PCs were used as ANN input variables. Two types of ANN methods (multilayer perceptron [MLP] and general regression neural network [GRNN]) were used in order to estimate the EN signals. For both neural networks the input values were represented by scores of sensors data set PCs, while the output values were the quality parameter at different roasting times. Both the ANNs were able to well predict coffee roasting degree, giving good prediction results for both roasting time and coffee quality parameters. In particular, GRNN showed the highest prediction reliability. Actually the evaluation of coffee roasting degree is mainly a manned operation, substantially based on the empirical final color observation. For this reason it requires well-trained operators with a long professional skill. The coupling of e-nose and artificial neural networks (ANNs) may represent an effective possibility to roasting process automation and to set up a more reproducible procedure for final coffee bean quality characterization.

Evaluation of Flavor Release From Spray-Dried Powder by Ramping with Dynamic Vapor Sorption–Gas Chromatography

The ramping method for humidity at constant temperature was used to evaluate the release of D-limonene from spray-dried powder using a handmade dynamic vapor sorption (DVS) instrument. The release fluxes of D-limonene from spray-dried powder at two different liquid feed temperatures, 40 and 80°C, were used to evaluate the flavor release properties of spray-dried powder. The flavor fluxes for 40°C of liquid feed temperature were faster than those for 80°C over the whole humidity range.

Response Surface Methodology Guided Release of Two Acetate Volatiles From an Oil-in-Water Emulsion

Optimization of the volatile release of two commonly used flavor compounds (isoamyl acetate and furfuryl acetate) from a food emulsion model system was evaluated using response surface methodology (RSM). Twenty-seven random order settings were established using a central composite face-centered (CCF) experimental design. The main and combined effects of four independent variables—concentration of isoamyl acetate (50–90 ppm), furfuryl acetate (20–30 ppm), salt (NaCl; 0.1–2%), and pH (5–7)—on the responses were examined. The main objective of the present study was to determine the optimal concentration level of the four variables leading to optimal release of the volatile compounds. Quantitative measurements were conducted using solid-phase microextraction (SPME) coupled with gas chromatography–mass spectrometry (GC-MS). Salt concentration and isoamyl acetate concentration were found to have significant positive effects (p ≤ 0.001) on the release of isoamyl acetate. The optimization procedure indicated that the optimal conditions leading to the desirable volatile release were isoamyl acetate 90 ppm, furfuryl acetate 30 ppm, salt 2% w/v, and pH 6.0176. The findings of this study can enable chefs and food manufacturers to optimize conditions for maximum flavor release from food emulsion products.

CONSEQUENCES OF INDIVIDUAL CHEWING STRATEGIES ON BOLUS RHEOLOGICAL PROPERTIES AT THE SWALLOWING THRESHOLD

ABSTRACTThe chewing process transforms food into bolus for a safe swallow. It is known that humans adapt their chewing behavior to food product characteristics. This study aimed at identifying individual chewing strategies of healthy consumers and determining the respective consequences on bolus properties. For that purpose, the chewing activity of 50 subjects was recorded during consumption of five model cheeses. Boluses were collected at the swallowing threshold for rheological analyses. We found that 30% of subjects showed only slight adaptation of chewing activity to product characteristics and thus produced boluses with different rheological properties. Among the 70% of subjects who adapted their chewing behavior, 57% adapted their behavior via chewing time and 40% adapted their behavior via chewing time and muscular contraction amplitude. Among the bolus rheological parameters, only consistency was not influenced by chewing strategies. Hence, it seemed to be a determinant factor of the swallowing threshold for these products.PRACTICAL APPLICATIONSFlavor compounds are released during the oral processing of food. It is thus important to understand how consumers form a bolus and to analyze the consequences of chewing behaviors on bolus properties at the swallowing threshold, at the end of the oral stage of food, to clarify the role of mastication in the release of stimuli responsible for perception. Chewing strategies and bolus properties are some of the key factors of flavor release and product acceptability that is of interest for food manufacturers who can use these factors as additional parameters for new solid product design.

Effect of monoglyceride self-assembled structure on emulsion properties and subsequent flavor release

When monoglyceride (MG) was dispersed in oil above melting point, it formed self-assembled structures (liquid crystals) on cooling, which have the potential to structure oil in emulsion systems. In this study, the effects of MG on emulsion properties and subsequent flavor release were investigated. Model oil-in-water (O/W) emulsions stabilized with tween 20 (1% w/w, TW) and whey protein isolate (1% w/w, WPI) were prepared using microfluidization. All emulsions, with droplet size ranging from 170nm to 387nm, showed good stability during storage. In TW emulsions, DSC results indicated that development of MG crystalline structure occurred within 3days of storage. The MG self-assembled structures largely modified the rheological behaviors of the emulsions, which showed an increased storage modulus on storage and the TW emulsion gave gel property (G′>G″). The crystallinity was highly dependent on the content of MG in the system, while independent on the content of oil. However, the gel property was affected by the content of oil and MG. For WPI emulsions, very little MG crystallized during storage, and the WPI emulsions remained Newtonian. When limonene was introduced into the oil phase of these emulsions, a delay in the release of limonene in MG structured emulsions was observed by headspace analysis. Changes in emulsifier types and oil contents greatly influenced limonene release. The results of this study demonstrated the potential of using MG structured emulsion for controlled release of flavors, possibly in fat-reduced system.

Development of a model mouth containing an artificial tongue to measure the release of volatile compounds

The perception of aroma during oral processing is a complex interplay of many factors, partly related to the food matrix as well as to the oral processing conditions. The role of the tongue in transporting the bolus through the mouth by pressing it against the palate has been widely studied, however, the relationship between tongue pressures generated and volatile organic compound (VOC) release is not clear. Pressure patterns during swallowing were found to be unique across subjects which may suggest personal flavor perception. The model mouth described here is capable of reproducing actual human tongue pressure patterns by a computer controlled artificial tongue driven by an actuator. The tongue functionality is monitored by pressure and force sensors. The model was designed to incorporate oral features and conditions (e.g. temperature, saliva flow, gas flow and appropriate oral cavity volumes) with on-line VOC measurement using proton transfer reaction–mass spectrometry. Tongue material from different elastomeric materials was tested for inertness. The model will be used in follow-up studies to investigate in vitro VOC-release. This paper describes the development of an innovative model mouth which has a potential industrial application as a product development evaluation tool for food and pharmaceutical companies. The model is capable of providing fast and accurate feedback on the product flavor release and its physical properties under oral conditions.

Production and characterization of antibodies against crosslinked gelatin nanoparticles and first steps toward developing an ELISA screening kit

Nanotechnologies are finding a growing range of applications in the food sector. Nanoparticles are used notably to add vitamins and other nutrients to foods and beverages without affecting taste and color. They are also used to develop new tastes, preserve food texture, control the release of flavors, improve the bioavailability of compounds such as antioxidants and vitamins, and monitor freshness with nanosensors. Crosslinked gelatin nanoparticles are a component of nano-sized carriers for nutrient and supplement delivery in foods and related products. This paper describes the production and characterization of polyclonal antibodies against gelatin nanoparticles. Two immunization schemes were investigated: subcutaneous injection with and without a first intravenous injection. Two enzyme-linked immunosorbent assay formats were used to characterize the antibodies: an inhibition format with an antigen-coated plate for detection of the immune response and a sandwich format for development of the method. The antibodies showed good sensitivity with an IC50 equal to 0.11 ng mL(-1) using indirect ELISA format and a good specificity for the nanomaterials, without significant cross-reactivity against native gelatin. The limit of detection was determined-0.42, 0.27, 0.26, and 0.24 μg mL(-1) for apple, orange juice, milk, and soft drink matrices, respectively. ELISA technology offers rapid, low-cost assays for screening foods, feeds, and beverages. We have studied a prototype ELISA for detection of gelatin-based nanocarrier systems. Fruit juices, milk, and a soft drink were the matrices selected for assay development.

Preparation of core–shell CaCO3 capsules via Pickering emulsion templates

Micron size and food grade pristine CaCO3 particles were used to stabilize an oil in water Pickering emulsion. The particles also acted as nucleation sites for the subsequent crystallization of CaCO3 with the addition of CaCl2 and CO2 gas as precursors. After the controllable crystallization process, a dense CaCO3 shell with a few microns in thickness was formed. The CaCO3 shell was proven to be calcite without the presence of crystallization modifiers. The crystallization speed and the shell integrity were controlled by manipulating the addition of CaCl2 amount during the different crystallization stages; therefore, the homogeneous nucleation in the bulk was almost inhibited, and the heterogeneous nucleation at the oil–water interface on pristine CaCO3 particles was the main contribution to the growth of the shell. The encapsulated limonene flavor in CaCO3 capsules showed a prolonged release in neutral water at 85°C, while a burst release at pH 2 water as expected. The method is a simple and scalable process for creating inorganic core–shell capsules and can be used for producing food grade capsules for controlling the flavor release or masking undesirable taste in mouth.

Characterization of the Volatility of Flavor Compounds in Alcoholic Beverages through Headspace Solid‐Phase Microextraction (HS‐SPME) and Mathematical Modeling

The volatility of flavor compounds (10 distinctive esters commonly found in alcoholic beverages) was characterized using headspace solid-phase microextraction (HS-SPME) analysis combined with mathematical modeling. The impacts of extrinsic factors (extraction time and temperature) and intrinsic parameters (ethanol content and concentration of these flavor compounds) were evaluated on their influences. From extraction profiles, different kinetic behaviors of flavor compounds revealed that volatility is influenced by chemical natures (that is, molecular weight and physicochemical properties). Moreover, volatility was also found to be interrelated with extraction temperature and absorption/adsorption on the fiber's surface. Through mathematical modeling, the kinetic constants of these volatile compounds were computed, and their release profiles were determined. Finally, it was observed that an increase of ethanol (a competitive interference compound to flavor compounds) could decrease the extraction efficiency. Our studies indicated that this approach might be a rapid and practical method that would provide a better understanding of flavor release behavior from alcoholic beverages. The proposed approach may provide a simple and fast method in predicting the performance of key aroma esters in different alcoholic beverages. It could also be a practical way in quality control during the production of alcoholic beverage by monitoring key aroma esters.

Analysis of Factors Affecting Volatile Compound Formation in Roasted Pumpkin Seeds with Selected Ion Flow Tube‐Mass Spectrometry (SIFT‐MS) and Sensory Analysis

Pumpkin (Cucurbita pepo and maxima) seeds are uniquely flavored and commonly consumed as a healthy roasted snack. The objective was to determine dominant volatiles in raw and roasted pumpkin seeds, and the effect of seed coat, moisture content, fatty acid ratio, total lipids, reducing sugars, and harvest year on volatile formation. Sensory was conducted to evaluate overall liking of seed variety and texture. Seed processing included extraction from the fruit, dehydration, and roasting (150 °C). Oil extraction was done using soxhlet, fatty acid profile using Gas Chromatography Flame Ionization Detector, and reducing sugars using 3,5-dinitrosalicylic acid and UV-spectroscopy. Headspace analysis of seeds was performed by selected ion flow tube-mass spectrometry (SIFT-MS). Volatiles dominating in raw pumpkin seeds were lipid aldehydes, ethyl acetate, 2,3-butandione, and dimethylsulfide. Compounds contributing to roasted aroma include alkylpyrazines and Strecker and lipid aldehydes. Overall, hull-less seeds had higher volatile lipid aldehydes and Strecker aldehydes. Seeds dehydrated to a moisture content of 6.5% before roasting had higher initial and final volatile concentrations than seeds starting at 50% moisture. Higher oil content resulted in higher lipid aldehyde formation during roasting with a moderate correlation between free fatty acid ratio and corresponding lipid aldehyde. Harvest year (2009 compared with 2010) had a significant impact on volatile formation in hull-less seeds, but not as much as variety differences. No significant correlation was found between reducing sugars and volatile formation. Sensory showed that hull-less seeds were liked significantly more than hulled seeds. Elucidation of aromatic flavor development during roasting with SIFT-MS provides information on flavor release and offers better control during processing. Knowledge of volatiles in raw and roasted pumpkin seeds and effects of seed coat, moisture content, seed composition, and harvest date will allow for better control over the production/storage/transportation process and a more educated decision during selection of a variety for production of pumpkin seeds in the snack food industry.

PTR-TOF-MS monitoring of in vitro and in vivo flavour release in cereal bars with varying sugar composition

In the present study, PTR-TOF-MS was applied to better understand the influence of sugar composition on flavour release in a strawberry flavoured cereal bar system. To achieve this, measurements were made both statically from the headspace above cereal bar samples (in vitro) and dynamically from flavour release in the nose space during consumption (in vivo). An artificial strawberry flavour of known constituents (17 flavour active volatile compounds) was used in the preparation of cereal bars. For in vitro measurements, eight samples varying in the glucose syrup solids 42DE to polydextrose ratio were assessed. Measurements clearly showed that the level of glucose syrup solids substitution by polydextrose influenced the release of the added flavour compounds. In addition, distinguishable differences were detected for the release of volatile compounds between samples with different levels of glucose syrup solids and polydextrose during in vivo measurements. The improved mass resolution, sensitivity and speed of PTR-TOF-MS enabled direct comparisons between the rate compounds reached the nose space, maximum nose space concentration of compounds, and the time after which compounds were no longer detected in the nose-space.

Mixture design approach as a tool to study in vitro flavor release and viscosity interactions in sugar-free polyol and bulking agent solutions

Flavor release from mixtures of maltitol, erythritol, polydextrose and oligofructose dissolved in water at concentrations of 43% (w/w) was investigated to analyze possible interactions between the viscous matrix and the volatiles using an experimental mixture design in combination with data analysis and response modeling tools. The dynamic release of four flavor compounds ( cis-3-hexenol, benzaldehyde, ethyl butanoate and butyl isovalerate) from the matrices at 37 °C was determined by headspace analyses using proton-transfer-reaction mass spectrometry (PTR-MS). The aim of the present study was to understand the complex release mechanisms from high-viscosity polyol and bulking agent solutions for the intention of conducting further studies on flavor release from sugar-free confectionary products. Each of the non-volatiles had a significant effect (p < 0.05) on flavor release with regression coefficients (R²) between 0.72 and 0.93 for the release parameters I max, t ½, and Δc/Δt, and 0.99 for viscosity. However, the magnitude of influence varied between the bulking agents and polyols due to their different molecular weights. A clear correlation between viscosity of the solutions and flavor release was found, revealing complex matrix-volatile interactions in the high-viscosity solutions. The release of all investigated flavor compounds decreased when the viscosity of the solutions increased. Therefore, it is assumed that the flavor release is significantly influenced by the non-volatiles when a critical concentration (c*) is exceeded. Interactions between the sugar substitutes were found to affect the viscosity of the matrices, whereas flavor release was not affected by interactions between the polyols and bulking agents investigated.

The dynamics of aroma compound transfer properties in cheeses during simulated eating conditions

In vivo aroma release during solid food consumption is a complex phenomenon that depends on food structure and composition, as well as on oral processing (combination of mastication and incorporation of saliva into the food product). The objective of this study was to understand and to predict the physico-chemical properties of aroma compounds through the dynamics of flavor release during in-mouth oral processing of food before bolus swallowing. Within this context, the evolution of two aroma compounds during bolus formation was explored by studying the two main properties that account for mass transfer: air/bolus partition and mass transfer coefficients. Four types of industrial cheese products (varying in fat and firmness) flavored with ethyl propanoate and 2-nonanone were chosen. Each matrix was mixed with various amounts of artificial saliva to mimic boluses at different stages of mastication. The air/bolus partition coefficient was determined by the static phase ratio variation method (PRV), while the mass transfer coefficient was obtained by non-linear regression from dynamic headspace experiments. Results showed that there is a dilution effect on the air/bolus partition coefficient and both a dilution and a product effect (firmness) on the mass transfer coefficient of ethyl propanoate in the bolus. These results were also validated with 2-nonanone for the low-fat cheeses.

Influence of polyols and bulking agents on flavour release from low-viscosity solutions

The release of four volatile flavour compounds (cis-3-hexen-1-ol, benzaldehyde, ethyl butanoate and butylisovalerate) from pure water and various low-viscosity aqueous solutions (sucrose, maltitol, erythritol, polydextrose and oligofructose, each at 20% (w/w)) was investigated. Dynamic headspace concentrations of the flavour compounds at thermodynamic equilibrium were monitored by proton-transfer-reaction mass spectrometry (PTR-MS). The rheological properties of the solutions were characterised by their viscosity. Flavour release from pure water increased with increasing hydrophobicity and volatility of the flavour compounds. The highly volatile compounds were retained more extensively in the presence of sucrose, polyols and bulking agents, compared to in pure water, whereas an increase in the release of the less volatile cis-3-hexen-1-ol was observed. All aqueous solutions had similar viscosities, although bulking agent solutions tended to have higher viscosities than polyol solutions of the same concentration. A correlation between viscosity and flavour release in the low-viscosity solutions was not evident.