Food Research International 116 (2019) 717–723

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Dynamic flavor release from chewing gum: Mechanisms of release T

a,b a b a,⁎
Emma B.A. Hinderink , Shane Avison , Remko Boom , Igor Bodnár
a
Firmenich S.A., Rue de la Bergère 7, CH-1217 Meyrin 2, Geneva, Switzerland
b
Food Process Engineering, Wageningen University, PO Box 17, 6700, AA, Wageningen, the Netherlands

A R T I C LE I N FO A B S T R A C T

Keywords: Dynamic flavor release curves from chewing gum were measured using an Artificial Mouth coupled to the
Flavor release AFFIRM®. A flavor distribution model for chewing gum is proposed, where flavor is present as droplets in both
Chewing gum the hydrophilic (water-soluble) and the hydrophobic (water insoluble) parts of the chewing gum and as mole-
Food physics cularly dissolved in the hydrophobic part of the gum. During mastication, the flavor droplets in the water-soluble
APCI-MS
phase are released and responsible for an initial burst release. The flavor droplets captured in the gum-base are
Artificial mouth
pushed towards the interface by mastication and are responsible for the subsequent release. The flavor molecules
dissolved in the gum-base, released by diffusion, are only responsible for the release at very long time scales. It
was found that the oil-water partition constant is an important parameter to explain the flavor release, where
hydrophobic components show slower and longer release, while more hydrophilic components show more burst
release.

1. Introduction
The dynamic release of flavors is a major aspect of the sensory ex-
perience of consuming foods. While the formulation of an enjoyable
flavor profile is very important, there is little knowledge on the kinetics
of the release of individual flavor molecules during mastication. Even if
the equilibrium partitioning of a range of flavor components between
food and air is known, dynamic effects may cause different ratios be-
tween the components than intended, and one may even see a reversal
in the sequence of release, due to kinetic effects (Weterings, Bodnár,
Boom, & Beyrer, 2016). Thus, it is important to quantify the kinetics of
flavor release from food products during mastication. Since chewing
gum is masticated for a longer period of time compared to other food
products, it is a suitable and practically relevant matrix to study flavor
release. While not a major source of nourishment, chewing gum is used
just for enjoyment, the release kinetics of flavors are the essence of the
product. A high quality chewing gum is generally associated with a
long-lasting perception of the flavor, and it is known that even when the
flavor concentration declines in time, the flavors have been released
only in small amounts (Potineni & Peterson, 2008a; Raithore &
Peterson, 2016). This makes it of even more importance to understand
and improve the flavor release mechanism.
Chewing gum is primarily composed of a hydrophobic water-in-
soluble gum base and a water-soluble sugar or polyol phase in a ratio of
roughly 1:3 (Raithore & Peterson, 2016). The water-insoluble part
remains in the mouth during the whole mastication period, while the
water-soluble substances dissolve and are ingested. The flavor compo-
nents are released during mastication from the chewing gum, via the
saliva, into the gas phase in the mouth, throat and nose. The compo-
nents are then detected in the olfactory epithelium (Hodgson, Linforth,
& Taylor, 2003). Two important factors that can be coupled to the
flavor release are the Henry's constant and logF. The Henry's constant is
the equilibrium constant associated with air-water partition of the fla-
vors, while logF describes the equilibrium partition constant between
triglyceride oil and water and is a measure for the hydrophobicity of the
flavor components. LogF is comparable, but not identical, to the fre-
quently used logP that describes the partitioning between n-octanol and
water. The release rates of the flavor components determine how long a
flavor is perceived as pleasant and long lasting (De Roos, 2008).
However, the flavor perception in sweet foods is heavily influenced by
cross modality with tastants such as acid, sugar and-/or artificial
sweeteners. Sweetness enhances the perception of the flavors and the
loss of sweetness gives the impression that the flavor has been con-
sumed, even when the amount of flavor being released has not yet
decreased (Davidson, Linforth, Taylor, & Hollowood, 1999).
Flavor release from chewing gum has been studied before, but no
conclusive release mechanism was found. The release of the flavor
components menthone (Davidson et al., 1999), L-carvone and jasmone
(Potineni & Peterson, 2008b) from chewing gum was reported to be
constant over time; however for cinnamaldehyde, a peak release was

⁎
Corresponding author.
E-mail address: igor.bodnar@firmenich.com (I. Bodnár).

https://doi.org/10.1016/j.foodres.2018.09.002
Received 13 July 2018; Received in revised form 22 August 2018; Accepted 4 September 2018
Available online 07 September 2018
0963-9969/ © 2018 Elsevier Ltd. All rights reserved.
E.B.A. Hinderink et al.
reported in the first 2.5 min (Potineni & Peterson, 2008a, 2008b). This
release showed the same trend as the water-soluble sweetener sorbitol,
suggesting co-extraction of the flavor and the sweetener (Potineni &
Peterson, 2008a). We therefore studied the combined dynamic release
of flavors and sweeteners during mastication and the corresponding
release mechanisms from chewing gum into the water phase and sub-
sequently into the air.
The flavor release during the consumption of foods depends upon
parameters such as mouth warming, saliva composition, frictional
forces, pH, chewing rate and flow rates of the air and saliva (Burdach &
Doty, 1987; Haahr et al., 2004; Harrison, 1998; Overbosch, Afterof, &
Haring, 1991; Taylor & Roozen, 1996). Intra- and inter-individual
variabilities lead to variations during in-vivo flavor release studies
(Salles et al., 2007). We therefore employ an in-vitro system, the Arti-
ficial Mouth, which reproducibly simulates the mastication and ex-
change processes. The Artificial Mouth enables us to analyze the flavor
and sweetener release in liquid (saliva) and gas (nose) phases, by
coupling it with an AFFIRM® that determines the flavor concentration
in the headspace on-line, while the concentrations in the liquid phase
are determined separately off-line. A model flavor is used to flavor the
gum and is composed of isopropyl acetate, amyl acetate, menthyl
acetate, terpineol and amyl butyrate, which are chosen based on their
differences in hydrophobicity. The objective of this study is to combine
the release mechanisms of the sweetener and the flavor components
into a comprehensive model, to better understand and quantify flavor
release from chewing gum.
2. Materials and methods
2.1. Materials
Isopropyl acetate (CAS 108-21-4), amyl acetate (CAS 123-92-2),
menthyl acetate (CAS 16409-453), terpineol (CAS 98-55-5), amyl bu-
tyrate (CAS 540-18-1) were obtained internally (Firmenich SA).
Medium Chain Triglycerides (MCT) was used as flavor solvent in the
model flavor and was obtained internally. Dimethyl sulfoxide (DMSO;
CAS 67-68-5, Sigma-Aldrich) or ethyl acetate (CAS 141-78-6, Sigma-
Aldrich) was used to help solubilize the components for the logF and
Henry's constant measurements. Cyclopentane (CAS 287-92-3, Sigma-
Aldrich) was used for the calibration curve. Triglyceride oil (Sunflower
oil, local supermarket) was used to determine logF of the flavor com-
ponents.
Use is made of a standard chewing gum base, consisting of resins,
waxes, fillers, emulsifiers, anti-oxidants, according to FDA standard
guidelines – paragraph 172.615 (Electronic Code of Federal
Regulations, 2018). Sorbitol and Hydrogenated glucose syrup (Ro-
quette), glycerine (Ileon), Aspartame and Acesulfame K (Brenntag)
were added to the gum base to form the chewing gum blank.
2.2. Methods
2.2.1. AFFIRM®
The AFFIRM® was used to determine the logF and Henry's constants
of the individual flavor components, and to determine the flavor re-
leased from chewing gum during mastication. The AFFIRM® consists of
an Atmospheric Pressure Chemical Ionization Mass Spectrometer -APCI-
MS - (Micromass ZQ, Mancheser UK; APCI-MS (quadrupole)) with a
patented Venturi interface (Linforth & Taylor, 1998) and was operated
with the following settings: corona voltage 3.8 V; cone voltage 18 V;
source temperature 105 °C; heated sample transfer line was a deacti-
vated fused silica (SGE Analytical) operating at 130 °C; a heated in-
terface gas temperature at 105 °C. The AFFIRM® was calibrated by di-
rect injection of a continuous flow (1 μL/min) of the individual flavors
(Henry's constant and logF) or the model flavors (chewing gum ex-
periments) dissolved in three different concentrations in cyclopentane,
which was subsequently evaporated into a controlled airflow (10 L/
718
Food Research International 116 (2019) 717–723
Table 1
Volatile concentration that was used to determine Henry's constants and the fat-
air partition constants.
Component Aqueous concentration (mg/L) Oil concentration (mg/L)
Isopropyl acetate 0.00864 0.168
Amyl acetate 0.00868 2.14
Terpineol 0.5055 970
Menthyl acetate 0.00927 250
Amyl butyrate 0.00842 12.78
min). A correction was made for the signal of pure cyclopentane from
the standard signal.
2.2.2. Henry's constant
Henry's constants were measured for the five molecules that form
the model flavor using the method described by Avison et al. (Avison,
Van Gruijthuijsen, Pascu, Parker, & Bodnár, 2015). The flavors were
first individually dissolved in dimethyl sulfoxide, 0.05 mL of the DMSO
solution was transferred into 100 mL water and this was diluted further
to the concentrations as listed in Table 1. The aqueous solution
(100 mL) was transferred to a 250 mL Pyrex bottle containing a Teflon-
coated magnetic stirrer. This bottle was closed with a Teflon cap
equipped with a small fan (2.5 × 2.5 × 0.5 cm, Conrad, CH) hanging in
the headspace to ensure good mixing of the headspace. The sample was
left to equilibrate for 20 min. Subsequently the volatiles in the head-
space were sampled via a heated transfer line to the AFFIRM® for 1 min
at 25 mL/min. The signals were corrected for the signal from the
headspace of distilled water.
The obtained headspace concentration was used to calculate Henry's
constant H with Eq. (1) (Avison et al., 2015).
cHS
H=
c − cHS ( )
VHS
Vaq (1)
herein cHS (mg/L) is the concentration in the headspace, c (mg/L) the
initial concentration in the aqueous phase, VHS (L) the headspace vo-
lume and Vaq (L) the liquid volume.
We define the LogF value as the (10 base) logarithm of the parti-
tioning constant of a flavor component between water and triglyceride
oil, which was calculated via the ratio of the experimental value for
Henry's constant over the experimentally determined air-fat partition
constant (Eq. (2)).
H
log10 F = log10
maf (2)
in which maf is the air-fat partition constant and H, Henry's constant.
Dilution in DMSO was used for isopropyl acetate, amyl acetate and
amyl butyrate. The DMSO solution (0.05 mL) was transferred into
250 mL Pyrex bottles containing 100 mL of sunflower oil. Due to the
high concentrations needed for terpineol and menthyl acetate to obtain
a significant signal with the AFFIRM®, these molecules were added
directly into 100 mL sunflower oil in the Pyrex bottles. The bottle was
closed with a Teflon cap equipped with a small fan (2.5 × 2.5 × 0.5 cm,
Conrad) suspended in the headspace to ensure good mixing of the
headspace, while a Teflon-coated magnetic stirrer was used to mix the
oil phase. The sample was left to equilibrate for 20 min; subsequently
the volatiles in the headspace were sampled via the heated transfer line
to the AFFIRM® for 1 min at 25 mL/min. The background signal in-
tensity of pure sunflower oil was subtracted from the oil-flavor signal
intensity. Calibration was done in the same way as described in Section
2.2.1.
2.2.3. Chewing gum production
The unflavored chewing gum blank was produced with the in-
gredients shown in Table 2. The gum bas was melted at a temperature
E.B.A. Hinderink et al.
Table 2
Bland chewing gum ingredients.
Ingredients Concentration (wt%)
Standard gum base 25
Sorbitol 57
Hydrogenated glucose syrup 11
Glycerine 7
Aspartame 0.08
Acesulfame k 0.04
of 65 °C in the Z-blade mixer. The molten gum base was then mixed
with the hydrogenated glucose syrup, glycerine and one third of the
sweeteners (sorbitol, aspartame and acesulfame-k) and mixed for 5 min.
Additionally another one third of the sweeteners was added and after
5 min of stirring the remaining sweetener was added and stirred for 5
more minutes. The unflavored chewing gum was stored at room tem-
perature in air tight bags.
To add flavors the unflavored chewing gum was heated in a mi-
crowave until it was pliable. The model flavor (1 v/w %) was pipetted
onto the chewing gum and kneaded into the mass. The model flavor
consisted of 40 v/v % MCT and 12 v/v % of each of the five flavor
components; Isopropyl acetate, amyl acetate, menthyl acetate, terpineol
and amyl butyrate. The chewing gum was rolled out to a thickness of
3.5 mm using a laminating machine and cut into samples of 1.1 by
4.2 cm. These samples were stored and allowed to equilibrate for
2 months before analysis.
2.2.4. Artificial mouth
A schematic overview of the Artificial Mouth is shown in Fig. 1. It is
an in-house built device, based on aspects of different artificial mouths
published (Arvisenet et al., 2008; Peyron & Woda, 2016; Salles et al.,
2006; Woda et al., 2010). The main cavity is formed by two discs
containing raised surfaces that act like human teeth. The discs are pu-
shed together replicating the biting aspect of chewing and then one disc
is rotated to, replicate the shearing aspect of chewing. The water flow
through the Artificial Mouth cavity was set at a speed of 0.36 L/h,
matching stimulated flow rates (Humphrey & Williamson, 2001; Woda
et al., 2010) and the airflow at 600 L/h both at room temperature of
21 °C. Water flow was started into the cavity 30s before the mastication
started, to mimic the saliva present in vivo. Chewing gum was masti-
cated for 15 min in the Artificial Mouth, by moving a piston in and out
of the mouth cavity (15 chewing actions per minute where shear is part
Fig. 1. Schematic drawing of the Artificial Mouth. Water flow = 0.36 L/h, air
flow = 600 L/h, volume cavity = 1.57 × 10−4 m3.
719
Food Research International 116 (2019) 717–723
of the motion). The concentration of the flavors in the exiting gas flow
was measured online with an AFFIRM® using a sampling flow rate of
75 mL/min.
The water (saliva phase) flowing out of the Artificial Mouth was
collected over a period of one minute at t = 0, 1, 2, 3, 4, 7, 10 and
13 min into 250 mL Pyrex bottles which were closed with Teflon caps
that were equipped with a small fan. The sampled water was left to
equilibrate overnight at 21 °C. Before measuring, the headspace was
homogenized with the fan for 20 min. The flavor concentrations in the
headspace were then measured using the AFFIRM®. The concentrations
in the water phase were determined by measuring the equilibrium
concentration in the headspace and then back-calculating the corre-
sponding equilibrium concentration in the water phase, using Henry's
constants.
The sweetener release was assumed to be in proportion to the dry
matter release. This was determined by the weight from the water
samples after drying overnight in a 105 °C drying chamber.
All experiments were performed in triplicate. The reproducibility of
the headspace measurements is within 10–30% and of the water sam-
pling 10–50%, which can be considered rather accurate for the type of
systems under study. Furthermore, it needs to be considered that the
variation between the different release curves for the different com-
ponents is in the order of 2–3 magnitudes.
3. Results
3.1. Flavor components characterization
LogF and Henry's constants were measured for the individual flavor
components and shown in Table 3. The Henry's constants are of the
same order of magnitude for isopropyl acetate, amyl acetate, menthyl
acetate and amyl butyrate indicating that they have a similar driving
force for transfer from the water into the air phase. Terpineol has a
Henry's coefficient significantly smaller than the other four molecules,
indicating that it has a smaller driving force for transfer from water into
the air phase. A range of LogF values (0.70–4.31) is found for the five
components. Isopropyl acetate is the least hydrophobic component;
menthyl acetate the most hydrophobic. It is expected that the isopropyl
acetate is therefore released more easily from the hydrophobic gum
base compared to the other molecules, and menthyl acetate least easily.
3.2. Artificial mouth: Headspace measurements
The flavor release patterns in the headspace are shown in Fig. 2. A
peak was found in the first 1–5 min for all of the volatiles measured, but
they were not all at the same time. Closer inspection of the different
release curves reveals that some molecules show two peaks in their
release curve. Isopropyl acetate and terpineol show only one peak, but
amyl butyrate and menthyl acetate show a clear two-peak release while
amyl acetate shows only a small second peak. The first peak is at
0.7–0.8 min for all flavors where the second peak is at 2.5–2.6 min. An
explanation for the second peak will be given in the discussion section
of this paper but it has to be noted that the second peak is only found
for flavors with a higher logF-value compared to those that showed only
one peak. The release profiles of the flavors change into an exponential
Table 3
Experimental Henry's constants and LogF-values of the five individual flavor
molecules.
Component Hexp (−) LogF (−)
−2
Isopropyl acetate 1.03 × 10 0.70
Amyl acetate 1.39 × 10−2 1.70
Terpineol 2.26 × 10−4 1.54
Menthyl acetate 2.58 × 10−2 4.31
Amyl butyrate 1.39 × 10−2 1.92
E.B.A. Hinderink et al. Food Research International 116 (2019) 717–723

Fig. 2. The average flavor release curves of the chewing gum in the headspace
for: ( ) isopropyl acetate, ( ) amyl acetate, ( ) menthyl acetate, ( ) terpi-
neol and ( ) amyl butyrate in the AM, chewed for 15 min.
decay after 5–7 min, with similar rates of decline (d ln c/dt) but not with
the same absolute concentrations. Isopropyl acetate has the highest
flavor release in the first few minutes while menthyl acetate has the
lowest, but the order in the rate of release after 10 min is reversed. For
the last 5 min menthyl acetate has the highest release and isopropyl
acetate the lowest.
3.3. Artificial mouth: Aqueous phase measurements
3.3.1. Flavor release
The aqueous concentrations of isopropyl acetate, amyl acetate, amyl
butyrate and menthyl acetate are shown in Fig. 3. A similar trend over
time can be seen as found for the gas phase in Fig. 2. However, here we
do not see the change in order between the different flavor components.
The terpineol has a Henry's constant an order of magnitude different
than the other four molecules, which results in a lower flavor con-
centration in the headspace and the highest release in the aqueous
phase. While the gas phase showed exponential decline of the con-
centration at longer times, we see here a levelling of the concentrations
in the aqueous phase, suggesting continuous leaching of these compo-
nents from the gum over time, and not merely washing out.
3.3.2. Dry matter release
The average release of dry matter over time shows a maximum in
the first 4 min (Fig. 4). After 5 min the release of dry matter tends
Fig. 4. The average dry matter content in the water outflow from the Artificial
Mouth over time.
towards zero and is virtually zero after 10 min. As mentioned before we
assume that the released dry matter is dominated by the release of
water soluble components.
4. Discussion
We here discuss the flavor release from the chewing gum and look
into more detail into the individual flavor release curves in the head-
space as well as in the aqueous phase and the release of the dry matter.
A flavor release mechanism for chewing gum will be proposed based on
the obtained observations.
4.1. Total flavor release
The total flavor released is the integration over time of the flavor
release in the headspace and water phase combined, expressed as a
percentage of the amount originally added to the chewing gum
(Table 4). The release of the esters is between 2 and 4%, which is close
to the 5% release found by Raithore and Peterson (Raithore & Peterson,
2016), and 11.4% for terpineol. The total flavor release for the 4 esters
is correlated to the logF of the components as shown in Fig. 5: esters
with a higher logF have a stronger preference for the hydrophobic gum
base, and therefore, these molecules are released less. The fact that only
2–4% was released from the chewing gum indicates that almost all the
flavor is still captured within the gum. This leads us to the conclusion
that diffusional release from the gum cannot be the dominant me-
chanism. The data in Fig. 2 shows a decrease of a factor 2 of the flavor
release in the last 5 min of mastication, while the concentration in the
gum has hardly changed. If diffusion were the mechanism of release,
the release would be directly proportional to the concentration present
and much more constant in time bearing in mind that the gum is
masticated quite intensely.
In total, 11.4% of the terpineol was released over time, which is

Table 4
Total release of the flavor components in the headspace and water phase over a
time frame of 15 min.
Component Release

Headspace Water Total As of total

(μg) (μg) (μg) (wt%)

Isopropyl acetate 52.1 ± 4.4 37.5 ± 3.5 89.6 ± 3.1 3.7 ± 0.1
Amyl acetate 60.4 ± 6.4 17.6 ± 2.2 78.0 ± 5.5 3.3 ± 0.2
Fig. 3. Aqueous concentration of the five molecules calculated from the Terpineol 34.8 ± 2.3 239.4 ± 23.9 274.2 ± 25.7 11.4 ± 1.1
Menthyl acetate 55.7 ± 5.0 1.8 ± 0.3 57.5 ± 4.8 2.4 ± 0.2
headspace concentration of the sampled water for (◆) isopropyl acetate, ( )
Amyl butyrate 59.7 ± 6.4 6.0 ± 1.5 65.7 ± 5.8 2.7 ± 0.2
amyl acetate, ( ) menthyl acetate and ( ) amyl butyrate. and ( ) terpineol.

720
E.B.A. Hinderink et al.
Fig. 5. Total flavor release, as the sum of the release through the headspace and
through the water phase for (◆) isopropyl acetate, ( ) amyl acetate, ( )
menthyl acetate ( ) amyl butyrate. and ( ) terpineol. The line guides the eye.
Fig. 6. ( ) The isopropyl acetate concentration in the aqueous phase com-
pared with ( ) the dry matter release over time.
significantly more than the esters. This difference in quantitative re-
lease cannot be explained solely by the hydrophobicity of the terpineol,
as is demonstrated in Fig. 2, and is probably due to its alcoholic nature.
Thus, interaction with the gum is more complex than just based on
hydrophobicity. Study of different molecular classes is desirable to
explain the high release of the terpineol.
Fig. 6 shows the combined graph of the isopropyl acetate release
into the water-phase and the washed out dry matter over time. The
release curves are almost completely proportional, indicating that the
flavor is released simultaneously with the water-soluble phase. This
Fig. 7. The aqueous concentration of the flavor molecules: (◆) isopropyl acetate, ( ) amyl acetate, (
a function of the dry matter release in the initial 4.5 min (left) and 3.5–13.5 min (right).
721
Food Research International 116 (2019) 717–723
Fig. 8. The maximum peak heights in the headspace as a function of the LogF of
the molecules for (◆) isopropyl acetate, ( ) amyl acetate, ( ) menthyl acetate
and ( ) amyl butyrate. and ( ) terpineol. The line guides the eye.
leads to the hypothesis that much of the flavor may be dispersed or
dissolved in the water-soluble phase of the chewing gum, and if this
phase is washed out due to the intensive mastication, both flavors and
sweeteners are washed out simultaneously.
If the flavor is washed out with the dry matter, one would expect
that the concentration of the flavor would be directly proportional to
the concentration of dry matter washed out. Indeed, by plotting the
flavor concentration as a function of the dry matter (DM), shown in
Fig. 7, we see a straight line for the first 4 data points between the DM
release and the flavor concentration – that comprise of the first five
minutes of mastication. Therefore, we confirm that for the first 5 min,
washing out may well be the dominant release mechanism of the fla-
vors. However at later times (and lower concentrations) - as shown in
Fig. 7, right graph - the linearity is not present anymore. Here, the dry
matter is depleted further over time, but the flavors are still released
disproportionally. This indicates that a mechanism other than washing
out becomes dominant in the later stages and will be discussed in the
following paragraphs.
4.2. Headspace release
The maximum peak heights - at t = 0.77 min - of the five different
molecules are nicely correlated with the logF values of the flavors,
again with the exception of terpineol, as shown in Fig. 8. The same
explanation as given for the total concentrations would be valid here.
The opposite order is seen in Fig. 2 for the last 5 min where iso-
propyl acetate has the lowest release and menthyl acetate the highest;
again terpineol behaves differently. Again, the decline in
) menthyl acetate and ( ) amyl butyrate. and ( ) terpineol as
E.B.A. Hinderink et al.
Fig. 9. The concentration in the headspace as a function of the concentration in the aqueous phase of the flavor molecules. On the Left: ( ) amyl acetate, (
menthyl acetate and ( ) amyl butyrate and on the (◆) isopropyl acetate ( ) and terpineol.
concentrations cannot be explained by mere diffusion from the gum,
since the concentrations in the gum only changes by a few percent, as
discussed before. A second release mechanism is proposed for the latter
release: part of the medium chain triglycerides (MCT) that was added to
the gum as flavor solvent, is not dissolved in the hydrophobic gum base
but dispersed in the water-soluble phase of the chewing gum and ‘en-
capsulated’ as droplets in the gum base. During mastication, the gum
base matrix opens up and releases the MCT droplets. This mechanical
release is then responsible for the second peak found for menthyl
acetate and amyl butyrate as shown in Fig. 2. The less hydrophobic
components are mainly released via the dissolution of the water soluble
phase whereas the more hydrophobic components may be released via
dissolution and the release of the oil droplets. We expect that molecular
diffusion from the hydrophobic gum base is only relevant over longer
time scales and leads to low concentrations in the headspace.
4.3. Flavor partitioning - aqueous phase and headspace
For the flavor to be perceived in vivo, it should not just be released
from the gum but also reach the olfactory epithelium (Overbosch et al.,
1991). The flavor needs to be transferred from the gum via the aqueous
phase into the headspace, or directly from the gum into the headspace.
Fig. 9 shows the flavor concentration in the gas phase as a function of
the flavor concentration in the aqueous phase. If the transfer of the
flavor from gum to water would be limiting, we would expect the
concentrations in the aqueous and gas phase to be at equilibrium. In
this case, a linear relation between the flavor concentrations in the two
phases would exist. Fig. 9 shows that this linearity is not observed in the
artificial mouth. Therefore, the transfer from the aqueous phase into the
gas phase is significant, and may even be limiting. One might expect
that molecular diffusion from a solid gum should be limiting. However,
in the previous section we have already made the point that the release
from the gum is probably dominated by mechanical release due to the
mastication. This is much faster than molecular diffusion; therefore the
transfer from aqueous to the headspace may indeed be limiting.
Therefore, to develop well perceived flavors the dynamic transfer of the
flavor compounds within the mouth has to be taken into account.
5. Conclusion
The initial phase of the aroma release from chewing gum follows the
same kinetics as the washing out of the water-soluble part of the
chewing gum. For terpineol a release of 11.5% was found indicating a
stronger preference for the water-soluble phase of the chewing gum
despite its relatively low logF. For esters the more hydrophilic com-
pounds are present at significantly higher concentrations in the head-
space than the more hydrophobic ones. The release on somewhat longer
722
Food Research International 116 (2019) 717–723
)
timescales may be caused by oil-droplets with flavor components that
are embedded in the water-soluble part as well as in the hydrophobic
part of the chewing gum and which are mechanically pushed out by the
mastication process. Molecular diffusion is probably insignificant
during the first 15 min.
The transfer from aqueous to gas phase was found to be significant
and perhaps limiting. Thus, to develop a well perceived flavor mixture,
the dynamic release within the mouth is highly relevant.
Acknowledgments
Mirela Pascu is warmly acknowledged for her assistance with the
AFFIRM® measurements.
References
Arvisenet, G., Billy, L., Poinot, P., Vigneau, E., Bertrand, D., & Prost, C. (2008). Effect of
apple particle state on the release of volatile compounds in a new artificial mouth
device. Journal of Agricultural and Food Chemistry, 56(9), 3245–3253. https://doi.org/
10.1021/jf073145z.
Avison, S., Van Gruijthuijsen, K., Pascu, M., Parker, A., & Bodnár, I. (2015). Novel
Methodology for Measuring Temperature-Dependent Henry's Constants of Flavor
Molecules. Journal of Agricultural and Food Chemistry, 63(28), 6313–6318. https://
doi.org/10.1021/acs.jafc.5b01517.
Burdach, K. J., & Doty, R. L. (1987). The effects of mouth movements, swallowing, and
spitting on retronasal odor perception. Physiology and Behavior, 41(4), 353–356.
https://doi.org/10.1016/0031-9384(87)90400-8.
Davidson, J. M., Linforth, R. S. T., Taylor, A. J., & Hollowood, T. A. (1999). Effect of
Sucrose on the Perceived Flavor Intensity of Chewing Gum. 47, 4336–4340.
De Roos, K. B. (2008). Flavourings for chewing gum. In D. Fritz (Ed.). Formulation and
production of chewing and bubble gum (pp. 205–231). Kennedy's Publications Limited.
Electronic Code of Federal Regulations (2018). Title 21: Food and Drugs; part 172—food
additives permitted for direct addition to food for human consumption; Subpart G—Gums,
Chewing Gum Bases and Related Substances; §172.615 Chewing gum base.
Haahr, A. M., Bardow, A., Thomsen, C. E., Jensen, S. B., Nauntofte, B., Bakke, M., &
Bredie, W. L. P. (2004). Release of peppermint flavour compounds from chewing
gum: effect of oral functions. Physiology and Behavior, 82(2–3), 531–540. https://doi.
org/10.1016/j.physbeh.2004.04.061.
Harrison, M. (1998). Effect of breathing and saliva flow on flavor release from liquid
foods. Journal of Agricultural and Food Chemistry, 8561(97), 2727–2735.
Hodgson, M., Linforth, R. S. T., & Taylor, A. J. (2003). Simultaneous real-time mea-
surements of mastication, swallowing, nasal airflow, and aroma release. Journal of
Agricultural and Food Chemistry, 51(17), 5052–5057. https://doi.org/10.1021/
jf030118+.
Humphrey, S. P., & Williamson, R. T. (2001). A review of saliva: Normal composition,
flow, and function. Journal of Prosthetic Dentistry, 85(2), 162–169. https://doi.org/10.
1067/mpr.2001.113778.
Linforth, R. S. T., & Taylor, A. J. (1998). Apparatus and methods for the analysis of trace
constituents in gases. 10.1016/j.(73).
Overbosch, P., Afterof, W. G. M., & Haring, P. G. M. (1991). Flavor release in the mouth.
Food Reviews International, 7(2), 137–184. https://doi.org/10.1080/
87559129109540906.
Peyron, M. A., & Woda, A. (2016). An update about artificial mastication. Current Opinion
in Food Science, 9, 21–28. https://doi.org/10.1016/j.cofs.2016.03.006.
Potineni, R. V., & Peterson, D. G. (2008a). Influence of flavor solvent on flavor release and
perception in sugar-free chewing gum. Journal of Agricultural and Food Chemistry,
56(9), 3254–3259. https://doi.org/10.1021/jf072783e.
E.B.A. Hinderink et al.
Potineni, R. V., & Peterson, D. G. (2008b). Mechanisms of flavor release in chewing gum:
Cinnamaldehyde. Journal of Agricultural and Food Chemistry, 56(9), 3260–3267.
https://doi.org/10.1021/jf0727847.
Raithore, S., & Peterson, D. G. (2016). Delivery of Taste and Aroma Components in Sugar-
Free Chewing Gum: Mass Balance Analysis. Chemosensory Perception, 9(4), 182–192.
https://doi.org/10.1007/s12078-016-9218-y.
Salles, C., Mielle, P., Le Quéré, J. L., Renaud, R., Maratray, J., Gorria, P., ... Liodenot, J. J.
(2006). A novel prototype to closely mimic mastication for in vitro dynamic mea-
surements of flavour release. Developments in Food Science, 43(C), 581–584. https://
doi.org/10.1016/S0167-4501(06)80137-8.
Salles, C., Tarrega, A., Mielle, P., Maratray, J., Gorria, P., Liaboeuf, J., & Liodenot, J. J.
(2007). Development of a chewing simulator for food breakdown and the analysis of
723
Food Research International 116 (2019) 717–723
in vitro flavor compound release in a mouth environment. Journal of Food Engineering,
82(2), 189–198. https://doi.org/10.1016/j.jfoodeng.2007.02.008.
Taylor, A. J., & Roozen, J. P. (1996). Volatile flavor release from foods during eating.
Critical Reviews in Food Science and Nutrition, 36(8), 765–784. https://doi.org/10.
1080/10408399609527749.
Weterings M., Bodnár I., Boom R.M. and Beyrer M., A classification scheme for interfacial
mass transfer and the kinetics of aroma release, under subm.
Woda, A., Mishellany-Dutour, A., Batier, L., François, O., Meunier, J. P., Reynaud, B., &
Peyron, M. A. (2010). Development and validation of a mastication simulator. Journal
of Biomechanics, 43(9), 1667–1673. https://doi.org/10.1016/j.jbiomech.2010.03.
002.